JPRS ID: 10707 TRANSLATION SEMICONDUCTOR PRODUCTION EQUIPMENT BY P.N. MASLENNIKOV ET AL.

• Document Type: CREST
• Collection: JPRS
• Document Number (FOIA) /ESDN (CREST): CIA-RDP82-00850R000500090004-3
• Release Decision: RIF
• Original Classification: U
• Document Page Count: 346
• Document Creation Date: November 1, 2016
• Sequence Number: 4
• Content Type: REPORTS

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
JPRS L/ 10707
3 Aug!'st 1982
Translation
SEMICONDUCTAR' PRODUCTtON' EQUIPMENT
By
P.N. Maslenntkov et al. ,
FgI$ FOREfGN BROADCAST INFORMATION SERVICE
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
NOTE
JPRS publications contain information primarily from foreign
newspapers, periodicals and books, but also from news agency
transmissions and broadcasts. Materials from foreign-language
sources are translated; those from English-language sources
are transcribed or reprinted, with the original phrasing and
cr_her pharacteristics retained.
Headlines, editorial reports, and material enclosed in brackets
are supplied by JPRS. Processing indicators such as [Text]
or [Excerpt) in the first line of each item, or following the
last line of a brief, indicate how the original information was
processed. Where no processing indicator is given, the infor-
mation was summarized or extracted.
Unfamiliar names rendered phonetically or transliterated are
enclosed in parentheses. Words or names preceded by a ques-
tion mark and enclosed in parentheses were not clear in the
original but have been supplied as appropriate in context.
Other unattributed parenthetical notes within the body of an
item originate with the source. Times within items are as
given ny source.
The contents of this publication in no way represent the poli-
cies, views or attitudes of the U.S. Government.
COPYRIGHT LAWS AND REGUI..ATIONS GOVERNING OWNERSHIP OF
MATERIALS REPRODUCED HEREIN REQUIRE THAT DIGSEMINATION
OF THIS PUBLICATION BE RESTRICTED FOR OFFICIAL QSE 0NLY.
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
JPRS L/10707
3 Auqust 1982
SEMICONDUCTOR PRODUCTION E9UIPMENT
Moscow OBO'RUDOVANIYE POLUPROVODNIROVOGO PROIZVODSTVA in Russian 1981
(signed to press 14 Jun 81) pp 2-336
[Book by Pavel Nikolayevich Maslennikov, FConstantin Andreyevich
Lavrent'yev, Aleksandr Davydovich Gingis, V.I. Rononov, I.V.
Kirichenko, V.A. Nazarov, V.V. Rudnev, V.V. Stepanov, G.I. Rholin and
V.S. Scherbakov, Izdatel'stvo "Radio I Svyaz", 5,000 copies, 336 pages
UDC 621.382.2/.3.0021
Annotation....................... . 1
1
Pb reword...............................................................
Introduction........................................ I-1. 7he Development of Semiconductor Device Production 3
I-2. Semiconductor Production, Ite Oompler.ity. General
g,equirements Placed on the Performaace Level of Equipment 3
and Production
I-3. Semioonductor DeviceF. Structural Components of Some Zypes
~ �
of Semiconductor Devices................... o
I-4. Standard Production Process Schemes and the Major Steps in ~
the Product:ion of Certain 1~pes of Semiconductor Devices.........
Part I. Equipment for Fabricating Wafers and Producing Semi-
conductor Structures......�..�,....�.�..�...� 1 8
Chapter 1. Equipment for Mechaaical Pm cessing of Wafers 18
1-1.
Equipment
for
Crystallographic Orientation ot Waf,ers.........
22
1-2.
Equipment
for
Cutting Semiconductor Materials
26
_ a_ jI - USSR - G - FOUO]
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
1-3. Equipment for Grinding and Polishing Semiconductor
Materials................................................ 32
(hapter 2. Equipment for Chemical Processing of Wafers and
Controlling Their Quality 3$
~
2-1. Equipment for Etching Ingots and Wafers 39
2-2. Equipment for Cleaning and Drying Wafers 44
2-3. Equipment for Controlling the Quality of Wafers 47
Chapter 3. Equipment for Creating p-n Junctions 50
3-1. General Information on the Planar Process 50
3-2.
Diffusion
Equipment
53
3-3.
Equipment
for Ion-Implanted Doping Processes.................
61
3-4.
Equipment
for Producing Epitaxial Filme
72
3-5.
Equipment
for the Production of Alloy Junctions
81
Chapter
4. Equipment for Film Production
85
4-1. Vacuwn Film Deposition Equipment 85
. 4-2. Film Precipitation From a Gas Phase 100
(hapter 5. Equipment for Phatolithography Proces.ses 107
5-1. Equipment for Preparing the Surface of Wafers 112
5-2. Equipment for Producing a Photosensitive Layer 114
5-3. Equipment for Producing Relief in a P'hotosensitive Layer..... 118
5-4. Pattern Matching and Expos ure Equipment 119
~ 5-5. Equipment for Producing Zbpological Relief on a Subs trate.... 129
5-6. Equipment for Fabricating Photographic Templates............. 136
Part II: Equipment for the Assembly and Quality Control of
Finished Devices. Finishing Operations...................... 141
Qiapter Er. Equipment for Separating Wafers Into Qzips 141
6-1. Equipment for Separating Wafers by Means of Scribing......... 142
- b -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
6-2. Other Kinds of Equipment for Separating Wa#ers Into
(hips 147
(hepter 7. Equipment for Assembling Semtconductor Devices 149
7-1. Methods of Asaembling the Major Typea of Devices.
Requirements Placed on the Equipment 149
7-2. Equipment for Mcsunting the Chips of Planar Devices............ 152
7-3. Equipment tor the Attaching of Leads to Planar Devices..:..... 158
7-4.
Equipment
and Complexes for Mounting Semiconductor
_
Devices and Integrated Circuits on a Strip Conveyor...........
166
7-5.
Equipment
for the Aesembly of ?oint Contact Diodes............
171
7-6.
Equipment
for the Automated Assembly of Alloy Diodes..........
173
7-7.
Equipment
fcr the Assembly of Power Transistors
175
Chapter
8. Equipment for Hermetically Sealing Semiconductor
Devices

181
8-1.
Equipment
for Cold Welding
181
8-2. Equipment for Electrical Contact Resistance Welding........... 184
8-3. Equipment for Hermetic Encapsulation With Plastics............ 187
8-4. Equipment for Hermetic Sealing by Means of Soldering.......... 189
8-5. Equipment for Checking the Hermetic Seal of Semiconductor
Devices 189
Chapter 9. Equipment for Testing the Electrical Parameters of
Semiconductor Devices 193
9-1. Measuremen t Equipment . 193
9-2. Classification Equipment...................................... 20f
9-3. Automated Systems Using Computers for Parameter Testing....... 206
9-4. Cantacting Assemblies for (hecking the Parametere of Semi-
conductor Devices 211
- c -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ON1.Y
Chapter 10. Teat Equipment 218
107-1. Equipment for Mechanical Tests 218
10-2. Equipment for Climatic Tests 224 ,
10-3. Equipment for Aging and Reliability Testing 236
(hapter 11. Production Pr .~:.ess Equipment for the Finsl Operati.ons....... 241
- 11-1. Equipment for the Protective Coating of Finished Devices...... 241
11-2. Labeling Equipment 243
11-3. Packing Equipment 246
Part III. Lines and Systems for the Mass Production of Semi-
conductor Devices and Integrated Circuits 249
Chapter 12. The 7heoretical Principles of the Comprehensive
Mechanization and Automation of Semicanductor
Production 249
12-1. Pr.oblems of Comprehensive Automation and Specific Features
of Semiconductor Productinn 249
12-2. The Systems Approach do the Planning of Automated
P roduction 253
12-3. 1he Engineering Economic Analysis of a Technologi-cal
and Production Process 257
12-4. Some Methods of Determinipg the Optimal Parameters of
Semiconductor Production Lines and Systems 263
_ Chapter 13. Fquipment for Purifying Media and Providing th e
Microclimate in the Ma3or Operations of a Production
Process 272
13-1. Requirements Placed on Production Process Media. 1he .
Main Methods of Purifying Media 272
13-2. Equipment for Centralized Watpr Purification 278
13-3. Equipment for Finish Water Purification 284
13-4. Equipment for Gas Purification and Drying 290
13-5. Gas Purity Monitoring Instruments 294
13-6. Clean Roomg. Dustproof Chambers and Boxes..................... 297
- d -
FOR OF'F[C[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFI CIAL US E ONLX
fhapter 14. Production Process Control Sys-teme and Equipment in
the Production of Discrete Semiconductor Devices and
Integrated Circuits 300
14-1. General Information on the Control of Production Processes
in Semiconductor Production 300
14-2. Computers and Information Control Complexes.
Peripherals 303
14-3. Automated Production Process Control Systems for
Several Production Steps 309
Chapter 15. Comprehensively Mechanized Production Lines for
Certain Mass Produced Types of Semiconductor Devices
and Integrated Circuits 313
15-1. Product Packing and Placement Hardward for ttke Major
Steps in the Comprehensively Mechanized Production of
Semioonductor Devices 313
15-2. A Conprehensively Mechanized Line for the Assembly of
Pulse Diodes 320
15-3. Dhe Comprehensively Mechanized Assembly Line for D226
Diodes 322
15-4. Comprehensively Mechanized Lines for the Ma.Jor Steps in
the Productian of Planar Transisto rs and Integrated
Circuits 324
Bibliography 331
- e -
FOR OFFIQAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL U
Annotation
[Text] The structural designs are described and the major charac-
teristics are given in the book for the most widely used semicon-
ductor production equipment; the requirements placed on the equip-
ment are foYmulated, and practical recommendations are given for
the major types of equipment for the operational checking of its
good operating condition. The basic principles of comprehensive
mechanization and automation in modern semiconductor production
are set forth. The most characteristic production lines and pro-
duction systems are described.
The book is intended for engineers and scientific workers involved
in the production'and application of semiconductoc devices and
integrated circuits.
Foreword
The fast pace of growth in the production of discrete semiconductor devices and
integrated circuits poses one of the major problems in the sector: the contin-
uous refinement of production process and instrumentation equipment, comprehen-
sively mechanized and automated lines as well as systems as the major basis
for their mass production. Because of the qualitative changes which have taken
place in the production technology for devices in recent years, the demand for
literature devoted to semiconductor production equipment is felt especially
sharply.
The most characteristic domestic and foreign equipment used in the production
of semiconductor devices is described in this book. The greatest attention is
devo*ed to pro3uction process equipment for rnanufacturing mass produced types of
transistors and semiconductor integrated circuits, the development of which is
based on planar technology.
1
FOR OF'F[C[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
FOR OFFICIAL USE ONLY
The three concluding chapters of the book are devoted to comprehenaively mechan-
ized lines and systems for the mass production of semiconductor devices and
integrated circuits. Some of the general questions of comprehensive automation
- and mechanization of semiconductor production are treated here, including
questions of the systemic approach to ar_d optimization of the major parameters
of comprehensively mechanized lines and systems when planniag automated produc-
tion processes, as well as questions of automation of transport operations
between process cycles and control of the technological process and production
of semiconductor devices.
Working with the book presupposes the familiarity of the reader with the funda-
mentals of semiconductor production technology. For this reason, questions of
technology are not treated in the book and are touched on only in individual
cases: in connection with the necessity of more completely explaining the
operational principles or the structural design features of the equipment
being described.
The book, in the opinion of tHe authors, will be useful both to equipment design-
ers, production process engineers and other workers in the semiconductor indus-
try involved with its operation, as well students in the higher educational
institutes and technical schools in the appropriate specialties.
The introduction and ChaptQr 12 were written by P.N. Maslennikov; Chapter 1 by
I.V. Kirichenko and P.N. Maslennikov; Chapter 2 by I.V. Kirichenko and K.A.
Lavrent'yev; Chapters 3 and 4 by V.V. Rudnev; Chapter 5 by V.V. Stepanov;
Cliapter 6 by V.A. Nazarov; Chapter 7 by V.A. Nazarov and G.I. Kholin; Chapter
11 by G.I. Kholin; Chapters 8 and 13 by V.S. Shcherbakov; Chapter 9 by V.I.
Kononov; Chapter 10 by K.A. Lavrent'yev and V.I. Kononov; Chapter 14 by V.V.
Stepanov and A.D. Gingis; Chapter 15 by P.N. Maslennikov, V.A. Nazarov and
G.I. Kholin.
The authors would like to express their gratitude to the reviewer, candidate
of the technical sciences and lecturer I.N. Rubtsov and the editor, candidate
of the engineering sciences, Professor D.B. Zvorykin for the detailed analysis
and valuable comments on the manuscript, as well as to all persons who rendered
assistance in the selection of the'materials for the book.
The authors will gratefully accept all remarks and proposals by readers directed
towards the elimination of all possible deficiencies in the hook, which they
ask be sent to the following address: 101000, Moscow, Chistoprudnyy Boulevard,
2, Izdatel'stvo "Radio i Svyaz .
- 2 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2447102/09: CIA-RDP82-00850R000500494444-3
FOR OFFICIAL USE ONLY
IIGTRODUCT IOti
I-1. The Aevelopment of Semiconductor Device Production
The development of electronics has taken on a special role in the age of the
scientific and technical revolution. The most important achievements of science
and engirreering are related to the use of electronic products, and primarily
to the use of semiconductor devices (PP) and integrated circuits (IM) [IC1.
In bei.ng one cf the youngest sectors of industry, the semiconductor industry
has developed at an exceptionally fast pace. The development and industrial
production of semiconductor devices necessitated the creation of a large number
of new technological processes and techniques. The methods developed at the
dawn of the development of solid-state electronics for producing p-n junctions
made it possible to set up the production of extremely simple point junction
and later also alloy junction germanium devices. However, sillcon technology
was the basis for the modern semiconductor industry [1, 6]. The creation of
dirfusion techniques fcr producing p-n junctions and epitaxial methods of
fabricaeing semiconductor films was of especial importance for the development
of silicon semiconductor device technology. The development of oxide masking
and photolithography, which comprise the basis of planar technology, made it
possible to create integrated circuits [S].
The further development of semiconductor production was related to the continuous
improvement of the techniques and equipment for planar epitaxial technology,
direCted towards substantially reducing the dimensions of components and
increasing the level of integration of the devices being fabricated. The
techniques of electron and X-ray lithography, ion-plasma and plasmochemical
processing make fundamental improvements in silicon device technology. The
indicated techniques opened up the possibility of developinz the so-called
submicron technology, which in the immediate future should become the basis
for the production of. devices with an increased level of integration.
I-2. Semiconductor Production, Its Complexity. General Requirements Placed
on the Performance Level of Equipment and Production
Modern semiconductor production is a complex of complicated operations, frcm the
input quality control of the raw materials to the final assembly of the finished
device, ita testing and packaging. When manufacturing semiconductor devices,
it is neceasary to perform tens and hundreds of production process and test
and measurement operations, which require the use of special equipment. Thus,
to fabricate a relatively simple technological type of semiconductor device,
a silicon diffusion diode, it is necessary to use more than 80 pieces of
special production process equipment, not counting the general purpose and
typical hardware used in manu�acturing operations. With the transition to the
fabrication of more complex semiconductor devices and IC's, as well as the
comprehensive mechanization and automation of production, the quantity and
comDlexitv of special equipment are also rising [4].
The ma3ority of the processes known to modern engineering are used in the fabrica-
tion of semiconductor devices: metallurgical, chemical, electrophysical, thermal
- 3 -
FOR OFFICIAt USE ONLY
0%
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
FOR OF'FICIAL USE ONLY
and mechanical metal treatment, welding, soldering, precision assembly; vacuum,
~ electron, ion-beam, as well as diverse monitor, measurement and test equipment
is employed to apply various kinds of coatings; electrical and non-electrical
meaeurements, internal flaw detection techniques, etc.
The high requirements placed on the technology and equipment are determined by
the specific features of the semiconductor device as a product*. The most
important of them consists in the fact that the entlre fabrication process for
the device is performed on a si.ngle chip and within its volume, in which layers
are produced having special physical properties, governed by the various con-
centrations of the doping impur3.ties. A deviation from the specified production
process modes in one of the operations can lead to the final rejection of the
device as z whole. The fabrication complexity of a semiconductor device is also
due to the extrao:dinarily small dimensions of the components. For example,
when producing high frequency semiconductor devices, it is necessary to solder
- and weld electrode leads 8 to 10 um in diameter to pads with dimensions of
20 x 70 um, without disturbing the layers in this case which are located under-
neath them, the thickness of which is 3 to 5 um.
It is specifically these features which primarily dictate the exceptionally
high requirements placed on the overall technical level of the production: the
purity of the raw materials, the stability and reproducibility of the production
process parameters, on the orgar.ization of production, and as a result, on the
equipment with which the semiconductor devices are manufactured. For example,
germanium to satisfy the requirements of semiconductor production should con-
tain no more than 0.2 � 10-8 % foreign impurities, while pure silicon should
be of a purity 1,000 times greater. The precision with which the temperature
is maintained in diffusion furnar_es during the heat treatment process of the
original semiconductor material (at a level of 800 to 1,300� C) should be no
less than +0.5� C over the entire length of the working zone, etc.
The same high requirements are also placed on the purity and stabllity of the
composition of the microclimate in which the devices are fabricated. For example,
the dust content in gases should be no more than 2 dust particles per liter;
the dust content of ordinary air amounts to about 20,000 dust particles per
liter with a size of 0.5 um or more. When fabricating devices, more than 10
kinds of energy vehicles are needed (nitrogen, argon, helium, dried air, hydrogen,
hot gas, etc.).
The exceptionally high complexity of semiconductor production technology and
the necessity of ineeting the requirements cited above and many others are
responsible for r.he need to create fundamentally new methods and tools as well
develop and introduce a large number of types of special equipment, frequently
custom-made equipment having no counterparts in other sector.s of industry.
Thus, the average precision in the fabrication of semicondtictor equipment is
two to three classes higher than in general machine building, and in a number
of cases, exceeds the precision of the equipment used, for example, in watch
*See the following for more details on the specif ic features of semiconductor
production (Chapter 12, � 12-1).
- 4 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
making. The microscopically small dimensions of semiconductor device components
have necessitated the use of apecial optical instruments and devices, which make
it posaible to execute and observe micromotions with a precision of down to
- 0.5 to 1.0 um and less. Extraordinarily complex problems have been solved in
the design of equipment to carry out numerous chemical .:ngineering processes
related to the use of especially corrosive chemical reagents such as hydro-
fluoric, nitric, sulfuric, hydrochloric acids, hydrogen peroxide, etc., as well
as the use of various gases and mixtures of them.
High requirements are placed on the level of automation, and reliability of
semiconductor equipment (primarily on the reproducibility of the production
process parameters) as well as the stability of its operation. These require-
ments are substantially incressed because of the problems of comprehensive
mechanization and automation of the production of mass produced types of semi-
conductor devices and IC's.
1-3. Semiconductor Devices. Structural Components of Some Types of Semiconductor
Devices
Each kind of device, which differs from another in its structural design and
even more in its technology, requires the creation of a specialized set of
production process and monitor and measurement equipment to set up industrial
production. A schematic of the classification of semiconductor devices in terms
~ of structural design and technological criteria is given in Figure I-1. Although
it is not exhaustive, the indicated schematic assists in showing the manifold
character and diverse nature of the complexes and groups of equipment used just
for the production of semiconductor diodes and transistors.
Semiconductor devices consist of a number of elements which are common to prac-
tically all types in terms of their function. The tnajor component of a semi-
conductor device is the chip of either a rectangular or more rarely a circular
shape with the p-n junctions formed in it. For protection against external
exposure and to improve the heat sinking, the chip is housed in a hermetically
sealed package, the structural features of which are governed by the type
of device, or it is sealed in plastic. The devices have internal and external
leads for thg electrical connections. The chip is either soldered (or glued)
dirECtly to the socket base or to the crystal holder.
We shall briefly deal with the configuration and structural features of the
most widely used types of semiconductor devices.
Diodes. The most widespread groups of semiconductor diodes are point and sur-
face contact (alloy and diffusion) types. The group of point contact devices
includes high frequency and microwave diodes, as well as pulse and converter
diodes based on germanium and silicon. The group of surface contact diodes
includes 1ow frequency rectifier and pulse diodes based on germanium and
silicon, silicon zener diodes, varicaps based on germanium and gallium arsenide
as well as tunnel diodes based on germanium, silicon and gallium arsenide.
- 5 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
(1
(2
(3
(1)
(2)
(6)
(5)
Figure I-1. Th
Key: 1.
2.
3.
4.
5.
6.
(2)
Yeaa-~~~eroaere
02�"�o""OPlanar
enr-nausytme 3
~
0
~ eur:exoreasrme 4
n
9arorere
~
N
W
~ Sovevm+e o rit
0
N (2) Ysu-7t~Ys~osa~s
n.euapxVe Plana
I 9nw-aaaaepxme (3)
(6) 31IM?SKON6aj1W0
0
0
C
R

i
 4J
. ~
.g' cn

a
8 m
v ~
~
C
P.
O
a
m
cn
e classification of semiconductor devices.
Alloy diffusion;
Mesa diffusion;
Epi-planar;
Epitaxial;
Elionic [sic];
Epitaxial. Point contact diodes have become widespread because of the technological simpli-
city of their fabrication and low cost. Typical structural designs for point
contact diodes are shown schematically in Figure I-2. The diode consists of
the germanium or silicon chip 2, which is sealed to the crystal.holder 1, the
contact electrode 3 in the form of a thin sharpened metal needle and capsule 4.
The housing (capsule) for D2 and D104 type diodes take the form of a glass tube
with Fernico inserts 5 sealed to it in the end faces. The semiconductor chip
is soldered to the massive nickel crystal holder, which is inserted in the
capsule and soldered to the Fernico insert using low temperature solder. At
the opposite end, asimilar electrode is sealed in the capsule, which sunports
the contact needle. The external leads 6 are usually circular, and sometimes
ribbon shaped; they are fabricated from nickel or platinite.
- 6 -
FOR UFF(CIAL USE ONLY
Point
To..a~e
e
~
e
p
Coau~r~ Alloy
a
.
~
( 2 ~ Yese-1p~ri~oriw
o
.
Tov~~~s Point

~
ccsvuu. Alloy
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
I 2 J V S t Z J t S 6
0
. ~
r ~ 9 e. s s
(b)
J '1 6
Q + Z ~ Ol ~C~ .
Figure I-2. Structural designs of point contact diodes.
Key: a. Germanium type D2;
b. Silicon type D104;
c. Germanium type D4.
In the all-glass D9 devices, the chip is usually soldered using low temperature
solder directly to the end face of the platinic lead which is located inside
the capsule. Sometimes, a very fine Fernico washer is placed between the chip
and the end face of the lead to match the temperature coefficient of linear
expansion. The second electrode takes the form of a platinic lead which is
12 fused into the glass of the capsule.
0,5
1
2
3
4
S
6
7
e
9
IO
Figure I-3. Structural designs of alloy diodes.
a. The germanium D7 type;
b. The silicon D202 - D205 types.
, Al1oy semiconductor diodes are fabricated by melt3ng alloys containing acceptor
or donor impurities into the original semiconductor. Electron-hole junctions
are produced in the overwhelming ma3ority of cases in germanium surface contact
diodes by melting indium into n-type germanium, while junctions in silicon alloy
- 7 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
devices are fabricated by melting aluminum into n-type silicon or an alloy
of tin and phosphorus (or gold and antimony) into p-type silicon. Structural
designs of germanium and silicon allc.diodes are shown in Figure I-3. The
germanium diode chip 7 with the fused in indium electrode 6 is soldered to the
stamped steel chip holder 8. To protect the p-n junction against external
- exposure, the germanium chip is housed in a metal-glass capsule, consisting of
- the Fernico housing 5, glass insulator 4 and the Fernico tube 2 for the internal
lead 3. The external leads 1 are connected to the chip holder and the Fernico
tube.
The structure of a silicon alloy diode has much in common with the germanium
diode. The chip 8 with the fused-in electroda 7 in the form of a small aluminum
calumn is soldered to the copper base 9. The junction is housed in a capsule
consisting of the Fernico housing 5, glass insulator 4 and Fernico tube 2. The
interior lead 3 is connected to the aluminum column. External lead 1 is brazen
to the Fernico tube. To improve the heat sinking, the diode is fastened to
metal chassis 10, to which the current is fed by means of a threaded contact.
The electron-pole junction is protected with varnish 6.
Transistors are one of the most widespread components in electronic systems.
The advantages of silicon planar transistors [6] exerted an especial influence
on the expansion of their wide scale applications in electronics. The first
semiconductor triode, which was proposed in 1948 by Bardeen, Brattain and
Shockley, was a point contact device. The point contact transistor is prac-
- tically not used at all at the present time; the major type of transistor is
the surface device. Field-effect transistors have been finding wide scale
applications recently.
In terms of the maximum power dissipation, transistors are broken down into the
groups of low (up to 0.3 watts), medium (from 0.3 to 3 watts) and high power
(more than 3 watts) transistors; in terms of the maximum working frequency,
they are broken down into low frequency (up to 3 MHz), medium frequency (from
3 to 30 MHz), high frequency (from 30 to 300 MHz) and SHF (more than 300 MHz)
devices.
Surface contact transistors are broken down into alloy, diffusion, planar and
epitaxial types according to the methods of fabricating the p-n junctions.
Variants and combinations of these methods are also widely used (see Figure I-1).
Transistors are also broken down according to the material used (germanium,
silicon).
Without going into the structural design of alloy transistors, we shall move
directly on to epitaxial planar transistors; this book is primarily devoted to
the equipment for the production of this type of transistor.
The structural designs of
are shown in Figure I-4:
as in plastic.
the mass produced types of low power planar transistors
transistors encapsulated in metal and glass as well
- 8 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
029
1A ~ r. ~f�:
s ~r
. , .
, ZOJa y , z ~ w
ooa o
(b) 61
(a) ,
c)
Figure I-4. Structural designs of low
power planar transistors.
a. In a metal-glass pack-
age;
b. In a plastic package.
a. In a metal-glass
package;
b. In a plastic package.
For transistors having a metal-glass package (Figure I-4, a), the chip with the
p-n junctions 1 is soldered to the mounting base 4 and connected by the leads
2 to the cross-ties 3. The capsule 5 is either resistance or cold welded to
the mounting base 4.
The structural design of transistors in a plastic package is the one most suited
to the requirements of mass production (Figure I-4, b). The chip 1 is connected
~ to the exterior flat leads of transistor 2 by wire leads 3. The advantage of
such a structure consists not only in the low cost of the hermetic sealing
plastic 4, but in the reduction of the labor intensity of the assembly operations,
which are accomplished on a single traveling belt carrier, including the sealing
operation, something which m:ikes it possible to automate the assembly process.
High power planar transistors are shown in Figure I-5. A considerable power is
liberated in the collector junction during the operation of such transistors,
because of which it is necessary to improve the heat sink so that the temperature
of the transistor componentR does not exceed the permissible level for the
material being used. For this reason, a considerable massiveness of the package
elements, the mounting base 1 and the capsule 45 a greater cross-section of the
emitter, base and collector leads 3 as well as a special structural design for
the feed-through insulators 2 are characteristic of power transistors (Figure
I-5, a). The base of the package is made of copper or a copper insert is used.
The bottom surface of.the package is usually not painted so as to reduce the
thermal resistance and improve heat removal from the package to the chassis or
heat-sink. Power transistors usually also have differences in the geometrical shape of the
p-n 3unctions as compared to low power devices so as to not excessively increase
- 9 -
�
FOR OFFICIAL USE ONLY
(a)
Figure I-5. Structural designs of
high power planar tran-
. sistors.
IO,S
w
~
~
:
b )
6)
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
S / 2 J A
~ .
o a'
2
J
% -
(c)
a)
1
Figure I-6. Structural designs of integrated circuits.
a. In a flat metal-glass package;
b. In a circular metal-glass package;
c. In a plastic package;
Key: 1. Semiconductor chip;
2. Internal lead;
3. Package;
4. External lead;
5. Mounting base.
the emitter current density and at the same time, not increase the base resis-
tance� Complex configurations of the junctions are used, making them in the
form of strips or rings. A power tranaistor in a plastic package (Figure I-5, b) differs from a low power
type also in the spectal structural deaign of the collector lead and the complex
configuration of the junction.
The external appearance of some maes produced types of integrated circuits is
shown in Figure I-6. It is easy to see that the structural packaging of
integrated circuits in circular metal-glass and plaetic packages ia a natural
development from the similar structural design variants of transistor packages
described above. Thus, chip 1 in the IC in Figure I-6b is mounted on base 3
and connected by leads 2 to its cross-ties 4. The hermetic sealing of the
,
- 10 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
a)(8) 1 2

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
device, i.e., the cnnnection of the cap 5 to base 3 is accompliahed by resistance
welding. The integrated circuit in the plastic package (Figure I-6, c) ia in
practice a structural design variant of the transistor shown in Figure I-4b,
but with a higher degr.ee of complexity in accordance with the functional complex-
ity of the device. The detailed design of semiconductor diodes, transistors
. and integrated circuits is 3escribed in [1-3, 6]. , I-4. Standard Production Process Schemes and the Major Steps in the t:-oduction
of Certain Types of Semiconductor Devices
The technological production processes for semiconductor devices include a large
number of operations which are executed in various sequences and can be repeated
several times, forming a complete fabrication cycle. Standard production process
(A)
7lscson c6opn ltepsKms c Rprcraaaor
G~~Kprcrua, OQesarprsexrs. HawM 06e3aprsenre. 8emm- Cnra
urnooran n IxImtlOReN :s no~-
pepaasaas o0p~0osna 7te- Apro- o6paQoiea xep Rprc- pusu
(2 ~---J :uaa Raseaell o :u-
~3~f7lrci - (4) iaeen 2D2"uaor(6 T3 (8) L"Qrao~ImirJ
7vacto= c6opir iepaazeaLB~sossureoR eP7aror
Dp~septa 3aezrpo- Os~~za r kaorarr- OswsNa Pasepacos-
r~opros artrteoxoe eainaetsor pos~xre Ae ae'ee *e Qo
�s sor asocspeoe n 7Np=ase- Sorsus- ae~ o sae~rer~
rnrsadl roarax~sRaB ax o eos:~ . rat rorssx: n7t7
~9~ oPln~ . OP-llv~ ra~ ~lu- 12pP7Ud ~7asedl 14
11
(13`
(15)repaa:sar- rroso:o: ~Wire
L___J
7vacro� cQopxr r semoaornsoi, rcmreXril AXoAoB
C6opRa AeP- HactpoNRa Pu6pa- osep- T~pro- oiosui Uposep-
uxae c a ueYSpo- ~o~ra no s csa- ~sepe- speu- ~a
zpreseaaor foproata sre~re- raecs o~~ posxe sraec-
E 1 N oaaaoacr ~oxot 17 ry s"~p up~610: 20 (21 =e~rei'
� (22
-
(23)~0, 0N16s upsaoresel rAeps~ssas c io~:ersror np7~~~ (2 ~
- f2-4)---J
~
qriu ra cN- teczri csesrasa= xnoxos 02e11vec ras ~roxo
posa- ssReo- cm~:e- xa ue- ee uaro- x~ ne- uepeA ot-
rse trz na- ur p~ret- osortocis perespos pacx 2
(26) Pa's~� 1 o0i(99 ('A
oarpu
Orpsora Y~prrpq(3 a~~- J:Ipolepx& H Uposepsa !3A'U.',afa~ro
r cnee  cn~n us s~ crasneaxrz (33) A'oROS ~o~tos ) so~to~ '~e Qapaw:pos Figure I-7. The technological production scheme for point contact
diodes.
Key: A. The section for the aesembly of the holder and the chip;
B. The section for the assembly of the holder and the
contact spring;
- 11 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
FOR OFF[CIAL USE ONLY
Key [cont.]: C. The section for the assembly and production process testing
- of the diodes;
1. Chip; 22. Static parameter testing;
2. Holder; 23. Capsule;
3. Solder disk; . 24. SoldPr washer;
4. Degreasing, chemical treatment 25. Holder with contact spring;
of the parts; 26. Thermal cycling;
5. Soldering the chip to the 27. Static parameter tesving;
holder; 28. Mechanical tests;
6. Degreasing, chemical treatment 29. Static parameter testing;
of the holders with the chip; 30. Moisture immunity testing
7. Protective coating of the chipg; of the diodes;
8. Drying the coatings; 31. Static parameter testing;
9. Welding and shaping the contact 32. Degreasing of the diodes
spring; prior to painting;
~ 10. ElPctrolytic sharpening of the 33. Painting and drying the
contact spring; diodes;
11. Washing and putting together 34. Marking and drying the
the hulder with the contact diodes;
spring; 35. Tinning the leads;
12. Aluminizing the contact spring; 36. Check TB;
13. Washing the holders with the 37. Static parameter testing;
contact spring; 38. Checking the external
14. Quality control rejection based appearance.
on external appearance;
15. Holder;
16. Assembly of the holdei with the chip and the capsule;
17. Alignment and electroforming of
diodes;
18. Quality control rejection based
on external appearance;
19. Static parameter testing;
20. Heat conditioning;
21. Current burn-in;
schemes for typical semiconductor devices (Figures I-7 I-10) are given below.
Thus, the technological production acheme for point contact diodes using the
example of the D18 diodes is given in Figure I-7; given in Figure I-8 is the
scheme for alloy diodes using the example of the D226 and D814 diodes. Further
on, the production process scheme for planar epitaxial transistors using the
example of the 2T-312 device is shown in Figure I-9. The major production
process operations for planar epitaxial technology using photolithography are
also employed in the production of semiconductor IC's, something which can be
seen in Figure I-10.
It can be seen from the schematics given here that the methods of fabricating
various semiconductor devices are extremely diverse. However, in all cases the
semiconductor chip is subjected to a number of common basic production process
operations.
- 12 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
Section for the Production
of a p-n
Junction
Chip
- 7vsosw nos7wxrn p-n nspe:oAs
Lr~ecru~
Cnx~s~r LWMtA6Rt11 Cr:a
a o0p6o:xa aspeso~tos
Naaeceue
s4rseoro
noKp~T~x
Korspas '
acperes-
pos
~p~~XA
7te:uN
u
aPe:c,Aos nsPe=onos
Drying
�oip"tr"
3wos~e
(1) opo~asar~
3~ A`
he Jurtc
(G)
(7)
(8)
~
(5)
VaorraraW
�
saeKtpo71,
(2)
~
.
7aaarox caopn Assembly
Section
CsapKa CnRe 06~~ r
Qprsspra
Ilpo~epxe
Teprourx-
oa -
aaNss
Ha
aepe:o~
Ilpaiepte
saex=pol~-
6uaoee ~aKy7rxt~ nDurps a
roseposW
sep=eern
srsoue
:NVxocrr
ure
-
ne rpac:u
aoAepu-
roro
uposs
c spresy-
aeesr
Pos
(9
(10)
(11) (12) 13
14
1
yvscsox rezxoaorrvecer: i01012608M �oKOn"1eseaseoN caoprr (17)

Itorspws
Yexarr-
aaaor-
OOa~sr- fieer-
@81&4+e
aeoeoe
spKr-
posMe
oxrpoas
~+eY=P~-
Koy=p~s
sre~esro
sasxspr-
v~crrx
veoKrS
Mcrnss-
DrBanO N
onrs nompr:rO 1uNo~tos
"b
E
~p~=
ulls
p0
~
18) H
m(19)
(20) ~troxos (21) (23)
(22)
(24
p(25)
(26)
Figure I-8. Technological scheme for the production of alloy diodes.
Key: 1.
Gold washer;
16.
Thermal cycling;
2.
Aluminum electrode;
17.
The section for produc-
3.
Chemical treatment of the
tion process testing and
Junction parts;
final assembly;
4.
Alloying of the junctions;
18.
Checking the electrical
5.
Chemical treatment of the
parameters;
junctions;
19.
Piechanical tests;
6.
Application of a protec-
20.
Degreasing and drying;
tive coating;
21.
Gal.vanic coating of
7.
Drying the coating;
diodes;
8.
Checking the parameters;
22.
Moisture resista:ice
9.
Soldering ~he junction
testing;
to the crystal holder;
23.
Classification of the
10.
Welding the electrode
diodes;
lead;
24.
Marking and drying;
11.
Welding the capsule to
25.
Checking the electrical
the fitting;
parameters;
12.
Vacuum drying;
26.
Checking the external
13.
Cogging and welding the
appearance.
Fernico scrip;
14.
Welding the upper lead;
15.
Chezking the hermetic
seal;
The production technology for semiconductor devices can be broken down into
several main steps.
The Fabrication of the Wafers. This step includes three groups of operations:
mechanical machining of the semiconductar materials, the technical chemical
- 13 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
treatme:it of the warers and their quality contrcl. The appropriate equipr:ant
is used for these operations.
The Fabricatinn of Structures with p-n Junctions. The major operation in this
step is the oi,eration of producing the p-n junction, and for this reason, the
name of the semiconductor device is almost 3l.ways determineii by the designation
of the processing method. Besides the group of equipment with which the p-n
junction is produced directly, one can single out two groups of equipment which
participate in the formation of the structures: equipment for producing films,
~ i.e., for applying metallic, dielectric and semiconductor films to the wafer
when making ohmic contacts, insulating coatings, thin film elements of micro-
circuits, etc.; equipment for the photolithography processes, i.e., for local
etching, directed towards the formation of the microrelief in planar technology.
Section for Producing Struc:�ires
(~aac:URa c 91131vororuf= I Yvnctor noaywrre crPYMSyV
soY nieero~ _I
- '
OOpa6o:le flopsoe Ilepeae 06paQo7wr Ibpi&n 'C~rrs Bso~aS cta7lnx ObpaOoeMS Bsopan y~oso-
( Z) naeosre napell o~rca~- @o:o- uaectrx ctsxre 6CC ~,,o acieere~~ ar~NOlV~oto--A (srrssepn~e
A0~7-
ae40 parr otrc- rre rpasr- aepex
aeurer a~aosrb posxe Aroyarep awr aaa~ rpaleposeoR oNOni u orrc-
(10)-
~3) (4) (5) 6op~6 (7) (8) (9) ,e)
aaM-
itOYII@
p6paQo2ra
p1~0l~~ GO~~
4e2sop-
T~ll
r
06pa6otRa
I14a0TNI 0IPA
Hami-
AoIIMe
06pa3osra
1{As0tMt 40pes
~ ~as
ilesan
tOSO-
rpur-
06pa6o:na
aaaosrr nepsl~
urr~rrw
Bureare I
At
Mo
vossepsor
lorp
naiouteexer
At
a
nelo
rpurpostiW
po~r~
h
(15)
rpurposros
srposra
At
:2
A Z
~
20
21
22
-
-
yvacro osop.m� Assembly Sect' on
Uposap-
PasAeae
UpormsRa
Nosu-
daaeYra
xyram"sos
Teprr
oCpoOos-
O91100e0-
wue
Srs7uamd
=orspab
ierapc~Ke,
e71in`
Ibprerrasqre
HertnetiC
xe m-
puuspo
u$
oaaosut
sp
ios
xe eoaq
ta
moxos
mont
2
29
30) '(rxo~c-~ re
i:pouue ~sAw ~ rb~oa !
(34}
(391i
acuox reXuworrvaorat mcm+serW  orottasaasaa ouepaqd
TFO o-
F1
Se pro-
Ye:aarvsorra
mtzaesm
Upoeepxe
reprevrv
I'ouse-
ttvsolof
Toxou
zporr-
Laaeoa
@ota-
YepRrpor
ta� asia-
aa7eu-
rYp =ot-
apossp-
ms
as
-
t
eue
1
W~xaa-
poseue
ao
eoara
a lseru
44
Possa
(45
40
4n
(46)
Posn r
�r~l 47
40
s ots ra
e
~
a
saeRrp
~o

~
)
r(05
(40)
(41)
(42)
(43)
e
os r
p
)
N
u
(
,
Figure
I-9. The technological scheme for the production of planar
epitaxial transiators.
Key: 1. The wafer with the epitaxial film;
2. Processing of the wafers prior t: the first oxidation;
3. First oxidation of the wafers;
4. First photoengraving;
5. Treatment of the wafers prior to diffusion;
6. First boron diffusion stage;
7. Removal of the BSS [?quick setting mixtures?];
- 14 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
Key [cont.]: 8..
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Second boron diffusion stage
with oxidation of the base;
Treatment of the plates prior
to the second photoengraving;
Second photoE*!graving (emitter
windows in the oxide);
Check of the parameters;
Treatment of the wafers prior
to diffusion;
First phosphorus diffusion
stage;
Second phosphorus diffusion
stage with the oxidation of
the emitr.er;.
Deposition of Mo;
Treatment of the wafers
prior to the third engraving;
Third photoengraving;
Check of the parameters;
Treatment of the wafers prior
to the deposition of Mo;
Treatment of the wafers prior
to the fourth engraving;
Fourth photoengraving;
Treatment of the wafers prior
to the deposition A1;
Deposition of A1;
Treatment of the wafers prior
to `he fifth photoengraving;
Fifth photoengraving;
Treatment of the wafers prior
to the burning-in of the A1;
The burning-in of the A1;
Check of the parameters;
Separation of the wafers;
Washing the chips;
31. Mounting bases;
32. Soldering the chips to the
mounting base;
33. Heat treatment;
34. Leads;
35. Attachment of the leads; .
36. Visuai inspection of the
mounting bases;
37. Varnishing, drying;
38. Capsules;
39. Section for production pro-
cess tests and final opera-
tions;
40. Thermal conditioning
41. Thermal cycling;
42. Mechanical tests;
43. Checking the hermetic seal;
44. Galvanic tinning;
45. Current burn-in;
46. Classification;
47. P4arking, varnishing and dry-
ing;
48. Visual checking of the over-
all dimensions;
49. Checking of the electrical
parameters;
50. Check of the external appear-
ance.
The Assembly of the Semiconductor Devices. This stage of the production combines
three equipment graups: equipment for monitoring: the structures on the wafer
and separating the wafers into chips; equipment for mounting the crystal on the
mounting base or strip as well as equipment for: nermetically sealing the devices.
Measurement of the Electrical Parameters, Classification and Tests of the Devices.
Besides quality control and test operations, other auxiliary finishing operations
are performed in this stage of the production, including the marking and packag-
ing of the finished device. The breakdown into steps cited here most precisely corresponds to planar produc-
tion technology for such devices, where the group method is used to produce p-n
structures on a wafer. For other types of devices, the composition of the
- 15 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
1)
Figure I-10. Flow chart showiug the production process for the
fabrication of epitaxial planar structure integrated
circuits with a buried n+ layer.
Key: 1. Group treatment;
2. Oxidation of the p-type
silicon wafers;
3. First photolithography
(in Si02); '
4. Diffusion of the n} doping
impurity;
5. Removal of the oxide,
n-layer epitaxy;
6. Oxidation;
7. Second photolithography
(using Si02);
8. Fourth photo.lithography
(using Si02);
9. Base diffusion of the
p doping impurity;
10. Third photolithography
(using Si0);
11. Separate diffueion of the
p doping impurity;
12. Emitter diffusion of
the n doping impurity;
13. Fifth photolithography
(using Si02);
14. Metallization;
15. Sixth photolitt:ography
(in metal) ;
16. Burning-in the contac�t~s
17. Hermetic sealing of
the package;
18. Installation of the
leads;
19. Mounting tha chip in
the package;
20. Scribing, breaking the
wafers and quality
control sorting;
21. Probe testing of the
integrated circuits on
the wafer;
22. Tests of the integrated
circuit.
23. Individual processing.
production ateps changes somewhat. The step by step production process scheme
for silicon planar epitaxial tranaistors in a metal package is shown in Figure
i-11.
- 16 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
.
TMrp7mow oQPoOosn

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
~wi~i,r Lac~rrwt /lwrrm" ~p. "io
e a~r.wr~~u~~ww ~ 4 ~
~wws ~KrN~
Sc 0 ~2~ ~ ~ ~
Bsr L 1
) (9)
/lwina~n
(12 1 iomelM
R/N/// 0/# Rj 8brt~w COi~er fpt ~aam
NjwyirAr M//R/W/1<
Y J
/
I
0
9 ~ /0
~
c)
Figure.1-9. Simplified Mechanical Diagrams of Machine Tools for Cutting
by Means of Diamond Wheels with an Inside Cutting Edge
In machines belonging to the first type the shaft, 2, of the spindle is positioned
horizontally and rotates in bearing elements, 3, by means of a V-belt transmission,
1. On the shaft is fastened a drwn, 4, with rings for attaching a diamond cutting
wheel with an inside cutting edge, 5. The ingot, 6, is attached to a holder, 7,
which is.fastened to the machine tool's carriage, which accomplishes longitudinal--
along the R-axis--and transverse--along the Y-axis--feeding. The spindle's shaft
is solid~and of small diameter and therefore small-diameter bearings are used.
A disadvantage of machine tools of this type is the wear of the carriage's guides,
which is.responsible for a loss of preeision. In addition, there is a restriction
on the length of an ingot which can be cut (76 to 90 umt); however, this disadvan-
tage must be considexed ten�porary--the etap].oytaent of a vacuum remover for the cut
wafer makes it posstita7.e xo incxease the length of an ingot to 500 mm and more.
In machine tools be],Qnging to the second group the apindlefs shaft is hollow.
The shaft's diametQr ts selecCed ao ttAx the tngot pAsses fxeely inside tt. The
vertical pqsitian of saich a.shaft makes it possiA],e to cut tngots of any length.
However, the lax$e diametex of the aptnd7,e'l& shAt't xequtxe$ the use of precision
bearings of large size, which #nvolves Aa incxease. In 7,tnear velocity and, conse-
quently, the intensiRied vQatr of Deaxtngs. Tn add#tion, there are di#ficulties in
lubricating the bearings and protecting tRem from the cutting f1uid (SOZh). Tn
machines'of this type transverse and longitudinal awvement of the tngot are accom-
plished by moving the carriage a1ong the guidea. TRerefore, in Chem intensified
-27-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
wear takes, p].ace at khe end of: the statioRa.r}* guide closes,t tq the cuttlng wheel,
since the gxea:est hending i1omEnt from the cutti.ng !'orce, actS relative to the
edge of the guide,
The design qf machine xqol,s of the third type is the beat fxom the viewpoint of
operating chaxacteri,Stics. A rotating arm, 8, instat7.7.ed on beaxings, 9, is used
instead of txansvexse feed $utdes. TRe entire spind1e assemQ7.y wfith the drum, 4,
and cutting wheel,, 5: i$ fastened xo it. ' The nrat together wfi.th the spi,ndle is
ro tated axound A Staxionaxy ax1e, 10, attached to the bed pf tRe machine.
Machine tools constructed according to the firat desfgn inc7,ude the 2405,
"Almaz-4", TS-21, TS-23 models, etc.; according ta the second de$ig,n, the ASM--lU
and AS-350; and accoxding to the third design--tRe 3,5X7.8 and 4,QX24 and "A].maz-61f"
models.
Most models have been developed according to the first design. These matchines
are organized in the following tnannex (fig 1-10). On the upper base qf the bed,
10, is installed a cast p7,ate, 8, on which the sqain units of the machine are
p laced: the spind].e, 6, with a heAd, 5, ' for atfiaching the cutting wheel.; the
c arriage, 12, with a setting mechanfsm, 17, a microawitch, 15, and afeed control
unit, 16; a#eed mechanism, 3, with a reduction gear, 18, a control mechanism, 4,
a brake, 2, and a shi#t lever, 1; the sptndle drive, 7; and a protective enclosure,
14, with a system for supplying the cooling fluid. A rotatigg head, 13, is in-
s tal.led on the carxiage, making it possibl.e to turn the tngot in the horizontal
and -trtical p].anes, which makes it possible to cut the i.agot para17.e1 to a specific
crystallographic plane. The head's scale value is 121. A unit, 9, for feeding
the cooling fluid is installed on the lower hase of tlie bed. Tt consists of a
tank and a centrifugal pump. An e7.ectsical equipment unit, 11, is fastened to the
upper half of the bed behind a panel.
The interaction of inechanisms is illustxated i.n the mechani:cal diagx'am (.fig 1-11",
Rotation is transmitted from the motor, 10, through the V-belt transmissivn, 9,
to the spindle, 8, with the head, 7, installed on it. The carriage, 5, (its top
part) with the rotating head, 6, to which the ingot is fastened, is moved in the
transverse direction under the influence of a weight, 3, along ball bearing guide^,
11. The carriage returns to its original position by means of the shift lever, ,
which is driven into oscillating motion b}r means of an eccentric, 14, installed
on a cam, 36. Ia longitudinal feeding of the ingot the carxiage ia moved along the
guides, 12, by the action of the drive sct'ew, 1, and nut, 2, fastened to the lower
half of thA carriage. The dxi:ve scxew is dxiven into moti.on by a sqotor, 26,
through a system of gears and cxanks, 34.
Adjustment fox the requixed cutting epeed--the longitudi,nal feed rate--is accom-�
p lished by moving ae"tox 40, which covera part of the teeth of xatchet wheel '9,
d epending on the specified xate, The txanavexse feed xaCe is set by means of a
guide fastened to the fxame of an indtcator attached to the tah1e, and a ttmi,ng
d evice. The mechanigm tox contxolling the tranvexse #eed rate oQexates in the
following manner. When the table, 5, moves, piston 38 foxcea oi1 thxough the gap
between the bodv of hydraulic cyl.i.nder 44 and cvne 43 tnto the space of the lower
half of the carriage, 37. The rate of outflaw of the oi1 determines the transvexse
f eed rate and depends on the gap, wh3,ch is set by means of screw 42. When button
-28-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
41 is preesed the gap is expanded and, when necessary, the table is rapidly
delivered ta the extxene l,eft posttion. An indivi:dual wafex ts fi.fi at cut fxom
the ingot, wi-th which the coxrectness of rQtatton of zhe itt$ot is checked fax the
purpose of axriving at the spectlted crqstai],ographic p7.ane. A,fter the angl.e of
rotation of the i-ngat is coxreeted, a wafer is cut o#f aad fxena its xhickness the
cutttng rafie is coxrected. These wafers are cut of# sqanually and the following
ones autanAtica}.1.y. Lever 25 ia accordtnglp plnced 3,n the "Manual Feed" or "Auto-
mattc Feed" position.
Figure 1-10. General Appearance of Cutting Machine
In the manual mode cam 17 occupies a position makin$ it possible to xai.se and
hold lever 16 at tts top position. ATith this carriage 5 is fixed by means of stap
15. In order to free the carriage it is necessary to press on gusher 22. When
lever 24.'is tuxned cam 21 turns pin 20. Lever 18, following the pxofile of cam 23,
frees lever 19, which is raised by means of pin 20. Tab1e 5 is returned manually.
In the "Automatic Feed" position of lever 25, cam 17 by means of lever 18 fixes
lever 19'in a poeition whexeby lever 16 is 1et down along the pxo,file of cam 36.
The upper half of the carriage--tab7.e 5--under the effect of weight 3 is moved in
the transnexae dixecticm xelative to the tngot's axis.
Longitudin$l feedtng at the cuxting pace ts aecamplished automatical1y. At the end
of a cut a screw cJ.osea the conCacts of micxosvttcR 4, which tuxns on motor 26 and
electromAgnet 30. The electrosvAgnet turna block 29, wtt:Lch spreacls shoes 27 of
brake dxum 28. qfter shaft 35 turna 360 degrees, cam 33 closes-the contacts of
switch 32, whi:ch cuts oi'f motor 26 $nd e7,ectromagnet 30. Block 29 and brake shoes
27 assume t:.eir original position undex the action of sprtngs 31. One of the most
ideal models of this type-rthe xS--23 taachine---has longitudinal feed precision of
-29-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
a 0.001-ma step. xtis tqachine ta: dtsti:ngutshed by a high 7.evel qf autrnqation:
The automatic loading of cut wrafere into ho7,dex i,a, pravided far.
Figure 1-11. Mechanical Dtagram oi Diamond-Type Cutting Machine
Machine tools designed accordtng to the second type (cf. fig 1-9, b) are distin-
guished by a verticall}r posittoned spindle. This position makes it possible to
increase the length of the ingot and to remove wafers automatically through the
center of the spindle. However, increasing the diameter of the spindle and accord-
ingly the size of the shaft bearings results in rapid wear. One machine of this
type is the model ASM10A from the Okamoto firm (Japan). This machine makes it
possible to cut silicon ingots up to 99 mm in diameter and 350 mm long. The rota-
tional velocity of the spindle ia 5000 r.p.m. The longitudinal feed range is 0.2
to 2.5 mm. The thicknesa tolerance of wafera which can be cut off is + 0.005 mm
for a diameter of 50 mda. The nonflatness and nonparalleliam of wafers is not
greater than 0.005 mm.
The main disadvantage of tnAchines o# the fi,xst tWo types is the weax of the table
guides, which reaulta in time in a 1,ose in Initial accuxacy. In order to eliminate
this, in machinea o� the 4,OX24 and "A1maz-6M" tqpes the tranavexae feed is accom-
plished by rockiAg rhe entire cuxti~ng head, insta7.led on beaxings, whose wear is
insignificant even afxax 10 yegtaca of uim. Astxuctura7, diagxam of machines of this
type is shpwr4 in fig 1-12.
-30-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
Figure 1-12. Block Jiagram of Machine Tool for High-Precision Cutting of
an Ingot into Wafers
Key:
1. Block for controlling and in- 10. Drive for turning wheel
dicating transverse speed 11. Vacuum
2. Drive for rocking spindle 12.. Cutting wheel
(transverse feed element) 13. Ingot
3. Block for inputing machine 14. Block for controlling and moving
tool's operating program vacuum sucker
4. Up 15. Longitudinal feed unit
5. Down 16. Longitudinal feed control and in-
6. Longitudinal feed dicating unit
7. Transverse fped
8. Rotational velocity of wheel
9. Number of wafers
The drive and drum with the tightened cutting wheel are fastened to a rotating arm
and are rolled in roller bearings installed on an axle with preloading, i.e., with
zero clearance. The spindle unit is rocked by means of a hydraulic drive to accom-
plish transverse feed. The transverse feed rate is set by means of the indicator
of the unit for inputing the machine tool's operating program. Longitudin al feed
of the ingot by a step is accomplished by means of a precision drive screw and
a stepping motor. Control is accomplished by means of the longitudinal feed con-
trol and~indicating unit. The longitudinal feed is set and monitored with a direct
indicator. In inputing the operating program, in addition to the longitudinal
- 31 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFIC[AL USE ONLY
and transverse feed the totational velocity of the cutting wheel and the number of
wafers to be cut off are set. The machine is outfitted with special rings for
fastening the cutting wheel and for tightening it by means of a specisl hydraulic
system. The wheel is stretched by means of oil pumped into a special annular
space. Uniformity of tensile atresses and high-quality tightening of the wheel'e
fabric are thereby made possible.
All these design featUres of the mactsine make it possible to obtain high accuracy
with high reproducibility. Maximum deviation in the thickness of cut-off wafers
on the 4,OX24 machine is + 0.007 mm. The machine makes it possible to cut wafers
up to 101.6 mm in diameter and up to 609.6 mm long. The use of long ingots elimi-
nates the operations of cutting an ingot into billets and of orienting the faces of
ingots and reduces losses of expensive semiconductor materials on account of kerfs
and on account of the.reduction of waste when ad3usting the machine for a cut.
The rotational velocity of the cutting wheel can be set fr.om 2000 to 5000 r.p.m.
The machine is furnished with an automatic attachment for removing cut-off wafers
and placing them in a special container with water. The attachment includes a
vacuum sucker and a unit for controlling and moving the vacuum sucker mechanically
connected to the longitudinal feed control and indicating unit. When the ingot is
fed by a step the vacuum sucker moves close to the face of the ingot and is pulled
toward the ingot. After the cutting-off operation the sucker with the cut-off
wafer and ingot is removed from the cutting wheel by a distance of a feed step.
- The vacuum is shut off and the flexible elements of the sucker remove the wafer
to bar guides from which the wafer drops onto a conveyer which carries it to the
end of the machine and removes the wafer to a container with water. The automatic
removal of wafers and automatic stopping of the machine when a specific number of
wafers has been cut off make it possible for a single worker to attend simultane-
ously to five or more machines.
1-3. Equipment for Grinding and Polishing Semiconductor Materials
Modern grinding machines are divided into two basic types according to the type of
tool used: machines for grinding with a free abrasive--an abrasive suspension--
and machines for grinding with a bonded abrasive--abrasive wheels.
Machines for grinding with a free abrasive are in turn divided into two types:
for one-sidP and two-side machining. For one-side machining wafers are cemented
to heads in the form of inetal disks with a ground working surface and are machined
first on one side and then after recementing, on the other. With two-side machin-
ing both sides of the wafer are ground and polished at the same time.
With respect to the method of fastening wafers, the distinction is made between
machines employing cementing, vacuum fastening and free laying in special flat
separator holders [12].
In machines for one-side grinding, e.g., the V1M3 type (fig 1-13), rotary motion
is transmitted from a motor, 1, through a worm reducer, 8, to a grinder, 2, with
working heads, 4, installed on it. The heads, supported by roller bearings, 3,
rotate on their own axes, at the same time making possible conditions for even
grinding of wafers. A mixer, 6, with a motor, 7, serves the purpose of mixing the
abrasive suspension in it, which is supplied to the grinder by means of a dropper,
5, at a rate of 60 to 80 drops per minute. Three heads are usually installed on
- 32 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
the grinder. The grinder is made of cast iron or of glasa--materiuls which are
easily ground--and the heads of steel or Duralumin.
7
6
S
s
J 2 �
/ -
e.
Figure 1-13. Mechanical Diagram of Machine for One-Side Grinding
A machine for one-side grinding can at the same time serve the purpose of polishing
wafers. For this purpose by means of a surrounding rim a soft fabric, usually
artificial suede, is attached to the grinder, or artificial leather is cemented on.
A mixture of diamond paste with a grain size of ASM 3/2, ASM 2/1 or ASM 1/0 with
ethyl alcohol and transformer oil is used as the abrasive suspension.
When using chemically active suspensions consisting of micropowders of zirconium
or silicon oxide, water and an alkali, chemical-mechanical finishing polishing is
performed. In machines created specially for chemical1mechanical polishing, e.g.,
the SKhMP-1 and Yul MZ.105.004 models (fig 1-14), holders with the wafers are
clamped to the pclisher by means of special clamps, by means of which elevated
pressure is applied to them. The gressure is created bq a pneumatic cylinder and
is transmitted through a rod--an intermediate mechanical spindle--fastened to
bearing supporte. There are no roller bearings in these machines. The model
SKhMP-1, Yul MZ.105.004 and Speed FAM32 machines have tables with a.larger
; diameter--680, 860 and 800 mm, respectively. These machines make it possible to
' machine wafers up to 100 mm in diameter. Four holders are installed on these
machinea and a great number of wafers are polished simultaneously: on the SKhMP-i,
20 wafers and on the Yul MZ.105.004 and Speed FAM32, 28 wafers 75 m4a in diameter.
Machines for one-side machining make it possible to produce wafers having one
surface of very high quality. However, the cementing and recementing of wafers
severely worsens their geometry and precision characteristics.
The IO 19006, AL-2F, SDSh-100 and SDP-100 machines for two-side machining are used
to produce wafers with a preciae geometry. The main design feature of these ma-
chines is a planetary train which makes possible planetary movement of wafers be-
tween two grinders (fig 1-15), which also makes possible high plane-pare.llelism
and planeness of machined wafers. The wafers to be machined, 9, are laid in the
, openings of toothed separators, 7, which engage wfth a center gear, 6, and a
peripheral gear wheel, 10. Gear 6 is fastened to shaft 5. Gear wheel 10 and gear
6 are turned by ;a single drive in the same direction, but with different angular
-33-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE UNLY
' velocities. On account of this the separators move over the grinder and at the
same time rotate on their own axles. Wafers are installed with the upper grinder
raised. The abrasive suspension is fed to the grinding zone via openings, 3, in
the top grinder, 1.
Figure 1-14. Machine for Chemical-Mechanical Polishing of Wafers
~ z.1 v ss i r s .9 m
A � ~
/0 7 9
i.
TpaeKmopua dsuJvccNUA
ytnmpo nAaunuHdi
ti
Key:
; 1. View A, top grinder re-
~ moved
2. Mechanical trajectory of
center of wafer
Figure 1-15. Diagram of Machine for Simultaneous Two-Side Grinding and
Polishing of Wafers
-34-
FOR OFFiCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
1J Rad A
8rp.s11ar1 qonU/10/la~b~~tLK C/iam

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
In machines of the IO 19006 type the top and bottom grinders do not turn. In
SDP-100 machines rotation of onlp the bottom grinder is provided for the purpose
of reducing the wear of the separator's teeth and its defarmation. Special spa-
cers, 2, through which water circulates are provided in the body for the purpose of
cooling the grinders. Machines Por simultaneous two-side grinding are also used
, for polishing wafers. For this purpose recesses are made in the grinding wheels
and by means of outside and inside steel rings, 4, suede is stretched over them.
There are holes in the suede and in the top grinder for the purpose of feeding the
abrasive suspension to the polishing zone. Protective fabric-based-laminate rings,
8, are used at the edges of wafers for the purpose of eliminating splitting. The
clearance between the protective ring and the separator is 0.4 to 0.8 mm and be-
tween the ring and wafer 0.4 to 1.5 mm. .
Machines for grinding by means of a bonded abrastve are divided in terms of grind-
ing method into machines which operate by the surface (face) grinding method (fig
1-16, a) and by the infeed grinding method (fig 1-16, b). SASh-420M, SASh-100 and
SASh-I50 model machines operate according to the infeed grinding method. SPSh-1,
MSh-259, SPShP-1 and MPS-R600 model machines operate according to the surface
(face) grinding method.
1) Cmc, ~ 8 }
key:
y
� f = f,
b ) � .
Figure 1-16. Diagrams of Surface (Face) (a) and Infeed (b) Grinding
1. Table
2. Grinding Force
Machines of the SASh-3 and SASh-420M type (fig 1-17) are designed for grinding
wafers up to 60 mm in diameter and the SASh-150 to 150 mm in diameter. All these
_ machine models are basically organized according to the same principle: three
spindles with a rotational velocity of 9000 to 14,000 r.p.m. and a table with
vacuum suckers rotating at a speed of 0.5 to 5 r.p.m. SASh-420M and SASh-150 ma-
chines have an intermediate mechanical spindle to whose shaft is fastened an a-
brasive disk of the AChK type. High-frequency generators power the electric
spindles.
The bed is cast and massive and consists of two halves (an upper and lower) fast-
ened to one another by means of a bolted connection. The spindle units and their
drives, the oil film units, counterweights and other units are mounted on the to;
half and the turntable, turntable drive and power supplies on the bottom half.
The turntable is in the form of a flat thrust-type ball bearing with the balls
arranged in three rows. The lower race is installed on four bearings whose height
can be adjusted. The upper race is in the form of a base for installing
-35-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
replaceable housings with vacuum suckers. The working area of the turntable is
covered with a removable cover made of acrylic plastic.
Figure 1-17. Type SASh-420M Diamond-Type Grinding Machine
The spindle is in the form of a cylinder on whose outside surface a trapezoidal
thread has been cut. The spindle's shaft is installed on high-precision radial
thrust bearings. In the lower half of the spindle there is a device for supply'.
cooling fluid to the grinding zone through the inside space of the diamond wheel.
Each spindle is driven into rotary motion by high-frequency electric sp3ndles
connected coaxially wfth the mechanical spindles via a centrifugal-action flexible
coupling.
The vertical feed mechanism for the spindles is in the form of a screw-and-nut
pair. A worm wheel is fastened to the nut. The drive is accomplished via the
worm manually from hand wheels or automatically from a motor by means of a V-belt
transmission. The spindles are moved vertically along cylindrical ball guides.
For the purpose of reducing axial stresses in the spindle vertical feed mechanism
and for exercising the option of having all clearances in the non working direc-
tion, the spindles are balanced by means of counterweights with an exceas weight
of 25 to 35 kg. The bearings of the mechanical spindles, electrical spindles and
turntable are lubricated by means of an oil film produced in the oil film unit.
- 36 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
A vacuum unit consisting of two pumpa of the pNZ type is used for creating a
vacuum in the turntable's auckers. A trap in the form of a sealed container is
provided.for the removal of Water.
k
The electrical spindle drive congists of a type 12 GIS-2 high-frequency generator
aiid a P-42 motor installed on a single panel and connected bq means of a V-belt
transmission.
The machine's units are mechanicallq connected in the following manner (fig 1-18).
Table 9 is driven into rotarq motion bq means of d.c. motor 1, bq means of a V-belt
transmission and worm gearing. The spindles, 3, receive rotary motion from the
electric -spindles, 6, via a centrifugal coupling. The spindle unit is balanced
by means.of a counterweight, 5. The spindles with the abrasive wheel, 2, are fed
vertically by rotation of the hand wheel, 8, which is coupled via a worm gearing
with the drive screw, 7. Automatic feed is accomplished by means of a motor, 4.
Figure 1-18. Mechanical Diagram of Type SASh-420M Diamond-Type Grinding
Machine
Diamond wheels of the dish type (AChK). with grits of ASM 80/60, ASM 40/28, ASM
28/20 and ASM 14/10 are used for grinding on the SASh-420M machine. Using a fine-
grit ASM 14/10 wheel guarantees the production of a surface finish of class 11 to
12.
-37-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The SPSh-1 type machine has two spindles independent of one another whose rota-
tional velocity is 2400 r.p.m. The grindtng platforms rotate at a rate of 350
r.p.m. Preliminary grinding is performed with one spindle--with an AS 12 diamond
wheel for silicon or AS 5 for germanium--and finishing grinding on the other,
with an ASM 40 for silicon or ASM 28 for germanium. The wafers are cemented to
the platforms. The diamond wheel lies with its abrasive edge on the wafers.
The wheel is lowered under the influence of the spindle's weight, Fsh (cf. fig
1-16, a).
Removable platforms (holders) for cementing wafers are provided for in machines of
the SPSh-1, SPShP-1 and MSh-259 types. However, the use of a vacuum table is also
possible in them. In one of the latest models from the Georg Muller firm (FRG),
the MPS-R600, the table has vacuum suckers over its entire area, which makes it
possible drastically to increase the loading of wafers and consequently the pro-
ductivity of the grinding machine. With a table diameter of 600 mm, 36 wafers
75 mm in diameter or 92 wafers 50 mm in diameter are ground at the same time.
The rotational velocity of the table is 0.8 to 20 r.p.m. The diameter of the
grinding wheel is 300 mm. This machine makes it possible to produce wafers with
deviation from planeness of 0.002 mm, deviation from plane-parallelism of 0.0025
mm and a wafer thickness deviation of 0.0025 mm.
Machines operating according to the infeed grinding method have become widely used
for grinding off thick (up to 700 microns) polycrystalline layers in the production
of integrated microcircuits with dielectric iso1F'-j.on. A layer of silicon 50 to
90 microns thick (cf. fig 1-16, b) is removed by three wheels of different grits
during a single turn of the table. The entire allowance is ground off in 6 to 12
turns of the table, depending on the thicknesa of the wafer.
Chapter 2. Equipment for Chemical Processing of WafErs and Controlling Their
Quality
Chemical processing equipment is designed for performing the production operations
of etching, cleaning and washing wafers [13].
Etching equipment is used for performing the operations of etching ingots and
wafers, which are done for the purpose of removing the defect layer originating
in mechanical processing. Etching the defect layer on wafers makes it possible to
eliminate the buckling of wafers (the ltayman effect), ro reduce the allowance and
to reduce rejects in the following operation of grinding.*
Impurities are eliminated in the following sequence: mechanical particles--by
cleaning in an ultrasonic bath and by means of various brushes; organic compounds--
by treatment in boiling solvents of the trichloroethylene, acetone and benzine
type; salts and metals--by boiling in redox solutions with the conversion of in-
soluble substances into easily soluble ones which are removed with the solution.
At intermediate stages in the fabrication of wafers their surface is cleaned
partially, pursuing one of the following goals:
The removal of impurities which can influence the accuracy of testing the wafer.
The removal of inechanical particlea which-are larger than the abrasive used in
the following operation.
-38-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
Preparation of the surface of wafers for the possibi],ity of perfoxming opera-
tions sensitive to impurities, such as chemical etching and chemical-mechanical
polishing.
Cleaning to the full extent is performed only after the final mechariical operation.
The purpose of this cleaning, called finishing, is to reduce the level of impuri-
ties on the surface of wafers to the level of impurities in the original single
crystal., Usually the concentratfon of impurities in original single crqstals of
germanium and silicon is not greater than 0.0001 percent. This requirement is
the most important for the entire complex of technological equipment.
2-1. Equipment for Etching Ingots and Wafers
For the purpose of etching silicon and germanium wafers, a number of pieces of
equipment are used, the basic elements of which are the etching bath, the rinsing
bath, a wafer drying unit, a table and a pressurized chamber with an air hole.
Wafers are placed in special containers made of fluoroplastic.
A typical unit for chemical etching of wafers (fig 2-1) consists of three main
parts: a chemical cabinet, 1, a pressurized chambei-, 6, and a table, 3. On the
rear wall of the cabinet there is an exhaust nozzle for the purpose of drawing
off the etchant's vapors and in the top left half of the table, the etching vat,
5. During etching and preliminary rinsing of wafers r_.he vat is hermetically
sealed with a fluoroplastic lid. The etchant is forced by compressed nitrogen
from the cooling vat, 2, into a measuring tank s*_:d from there enters the vat
through a hand cock, 7. Nitrogen is fed by means of inechanism 12 when the pedal
is pressed. Containers with wafers are rotated by means of drive 4. After etch-
ing, deionized water is supplied to the vat. Secondary rinsing in the deionized
water takes place in a second vat, 8. Thie vat is supplied with a cover and an
electric.heater. The finishing washing of wafers takes place in vat 9, made of
acrylic plastic.
The deioni?ed wata* heater, 10, consists of a quartz tube inside of which have
been p"laced three electric coils enclosed in quartz tubes. The heater is supplied
with two float switches for checking the upper and luwer water levels. The tem-
perature of the water is regulated by means of a contact thermometer. On the out-
side the heater is covered with ajacket. In recent times the coil heaters have
been replaced by more efficient quartz heaters with a current-conducting film.
In these contamination of the deionized water by the coil is totally eliminatQd.
In table 3 there are two electromagnetic cocks, 13, and a mixing tank, 11. The
electromagnetic cocks serve the purpose of draining the etchant from the vat int_
the mixing tank, 11, in which it is diluted with water before being discharged
into a special waste-water disposal system. The front wall of the pressurized
chamber is made of acrylic plastic and has two openings with rubber gloves fast-
ened in them by means of rings. Locks for joining to other units are installed
on the side walls of the presaurized chamber. Purified compressed nitrogen (fiR
2-2) is supplled to both locks 1 and 9 through pipelines 23, 20 and 21 via cocks
I and XII. Nitrogen enters 4-way pipe union 32 through pipe 22 and through pipes
30 and 34 enters below the unit's pressurized chamber, and through pipe 31 and
cock VIII, into the blow-through lock, 6. Compressed air is fed through pipe 17.
Passing through the pressure regulator, 13, it is divided into three branches:
-39-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFF7CIAL USE ONLY
through electromagnetic cock XITI into tank 15 for cooling the mixture; through
an angle cock with a pedal control through pipe 18 into canister 12; and through
pipe 40 with the same kind of cock into canister 16 with the etchant. Deionized
water with resistivity of 2 to 3 MS2�em enters through pipe 24. From collector 11
the water is supplied to heatez 8 through pipe 29 and electromagnetic cock TX,
into vat 4 through pipe 33, cock VI and header 5, and into vat 3 through pipe 38,
electromagnetic cock II and T-joint 44.
Figure 2-1. Unit for Chemical Etching of Wafers
I
a)
oe-~'oe~oy~swa 6)
N- XOasN ~wtN7 ~
wa.Nrm.wt
nodeas lsaosr-
,roOaN1oC p-
1) �C,,,. 9rrryw. NiQIN0AS0/IIn
Figure 2-2. Technological Diagram of Unit for Etching Wafers
[Key on following page]
- 40--
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
Key:
l. Power line 5. Deianized finishing water supply
2. Drq sir supply 6. Manual cocks
3. Nitrogen supplp 7. Electromagnetic cocks
4. Deoinized water aupply 8. Drain for recpcling
Deionized water with resistivitp cf 12 to 15 MS2�cm is supplied to vat 7 for finish-
~ ing washing of the wafers. When cock R is opened, water enters through pipe 25
with a water resistivity seneor, 10. From the vat the water enters for recycling
through pipe 26, into which is built a second sensor.
The prepared etchant is supplted fram canister 16 through pipe 43 to measuring
tank 2, connected by pipe 45 with electromagnetic cock III and by T-joint 44 with
the etching vat, 3. Air and the excess etchant are removed from the measuring tank
through pipe 46 to mixing tank 14. The etchant is drained from the etching vat
into the mixing tank via electromagaetic cock III and pipe 39. The cooling mix-
ture enters the jacket of the cooling vat, 3, through pipe 42 and leaves it through
pipe 41.
Vat 4 is used for removing the protective coating (wax, chemical-resistant lacquer
(KhSL), photoresist, etc.), appl'_ed to the untreated side of the wafer. A solu-
tion of hydrogen peroxide is supplied to the vat through pipe 19 via cock V and
header 5 for the purpose of removing the protective coating. Hot deionized water
is supplied from the heater through pipe 28 through cock VTT and header 5. In
washing the vat the water is drained through pipes 36 and 37 into the mixing tank.
The sQlution of hydrogen peroxide is drained to the same place through pipe 35
via cock IV. The water ts drained from vat 7 through electramagnetic cock BI and
pipe 27 into mixing tank 14.
The wafers are loaded into a fluoroplastic container and are placed in the etching
bath. The wafers are etched and prewashed automatically according to a predeter-
mined program. After preliminary rinsing it is recommended that the wafers be
hold in an atmosphere of purified nitrogen for 2 to 3 min. An inner lock (cf.
fig 2-1) is provided in'the unit for this purpose. Prom this lock wafers are
transferred to the finishing washing vat. The degree of washing of wafers is
controlled sutomatically by the difference in resistivity of the deionized water
in the vat's inlet and outlet. After ffnishing washing the wafers are transferred
into a polystyrene transit container and are forwarded through a connecting lock
for the next operation.
Wafers are loaded into the fluoroplastic container and are transferred to the
transit container after etching both manually and automatically by means of
special wafer loading and unloading units.
Instead of the usual etching and washing vats, in the automatic unit for dynamic
chemical etching of wafers (fig 2-3, a) there are two chambers: an etching cham-
ber, 1, and washing chamber, 2(fig 2-3, b). Two fluoroplastic screws, 4, which
turn in the sawe direction are installed in each chamber. The etchant is pumped
by means of a pump into the etching chamber, 1, and is sprayed by means of jets,
3. The temperatuze o# the initial etchant is held sutomatically in the range of
- +25 to -F50 �C with an accuracy of + 5�C. The wafer, 5, is fed automatically by
means of an unloading unit from the contatneT 3nto a tray and through a slot into
- 41 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR OFFICIAL USE ONLY
the etching chamber. In the chamber wafers are moved along the threaded grooves
of -otating screws 4 and at the same time rotate on their own axes, ensuring uni-
formity of etching over the wafer's area. From the etching chamber wafers enter
the screws of the washing chamber where they are washed with deionized water and
then they are dried by hot air and are removed by conveyer 6 to the unit for auto-
matically loading them into containers. The etching process is performed complete-
ly automatically by the continuous-flow method. However, the etching method used
in the automatic unit ha s an important disadvantage--it is practically impossible
to control and maintafn at a specific level and with the required accuracy the
etching temperature over the entire area of the wafer and, consequently, it is
difficult to control the etching reaction.
=~s
1)
S 3
~ Tpolwnens ui 6aea / 2)
y., Eoda
~ - NQ10C/JJN(L ~
~ I CAUQ Q 6Qd ,
i Z s
-5)-- 1--_J
~or~yr,.a roae,rN~ ~ Mid- Cf/I/LMQ e.irpy,,Aa
"a b )
a) b)
Key:
Figure 2-3. Automatic Unit for Dynamic Chemical Etching of Wafers:
a--general view; b--schematic
1.
Etchant from tank
6.
Etching
2.
Water
7.
Washing
3.
Drain into tank
8.
Drying
4.
Heated air
9.
Unloading
5.
Loading
-42-
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R040500090004-3
In the etching unit (#ig 2-4) the wafers, 1, are p].aced in fluoroplastic holders,
2, installed on the fluoroplastic gears, 6, of a screw rotator, 3, and are manu-
ally loaded into the vat, 5, Wtth the etchant. The screw rotator, 3, is driven
into rotarq motion by means of shaft 4 around the vat's aicis. The gears, 6, en-
gage with gear wheel 7 and rotate the holders on their axes. The combined motion
of the wafera--around their oWn axes and around the axes of the vat--is conducive
to producing high-quality wafers. The screw rotator, 3, with the wafers is trans-
ferred manually from the etchtng bath to the washing vat, 8, with circulating de-
ionized water.
7 s s zp49j e
TpaBney~c b ~ nPOMb/~MQ 3 ~
Key:
Figure 2-4. Etching Unit: a--general view; b--schematic
1. Etching
2. Water
3. Washing
Another variety of this etching variant is moving wafers in the etching vat with
their simultaneous rotation around the axes of the vat and a multipocket holder
which rolls over the inclined conical bottom of the vat. This motion ensures the
most uniform etching of wafers. Good quality of wafers is achieved when etching in a unit in which the etching vat
rotates around its axis at an angle oP 15 to 20 degrees. Wafers are adhered to
- fluoroplastic disks by means of wax or chemical-resistant lacquer and are placed
on the bottam of the vat, called the bell. When the bell turns the disk rolls
along its wall and in addition rotates on its own axis, which makes possible the
uniform and controllable etching of wafers. A laboratory-type unit is installed
in a fume cabinet. Wafers are washed in a vat standing alongside it and drying
takes place in a'centrifuge or in a drying cabinet tirith predrying by means of
filter paper. .
-43-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFFICIAL USE ONLY
In addition to chemical methods of etchtng, other methods have become used ever
more often, such as the electrochemical, thermal, ion and electron bombardment
and plasma chemical methods (12, 281. Tn theae cases apecific equipment is em-
ployed which, as a rule, is desigaed for carrying out the prqcesses o# cleaning
the surface prior to oxidation and diffuston. Therefore, it is discuased in the
appropriate chapters.
2-2. Equipment for Cleaning and Drping Wa4ers
Specific methods and the appropriate equipment are used for cleaning wafers, de-
pending on the type of contamination of the surface of wafers.
After the diamond cutting of ingots, as wellas after grinding by means of free
micropowders, wafers are cleaned of sludge and abrasive particles in ultrasonic
cleaning units in circulating deionized water. Wafers polished by means of diamond
paste are cleaned of the paste and aludge tn a unit for washing in a washing solu-
tion. After finishing chemical-mechanical polishing, wafers are cleaned succes-
sively in units for washing in detergents, in peroxide solutions and in a unit for
the hydromechanical cleaning of wafers (a brush washer).
After each washing operation wafera are dried in a centrifuge which as a rule is
a separate unit of the entire washing unit.
The basic elements of ultrasonic cleaning units are a dustproof cabinet, a vat for
washing, a magnetostrictor and an oscillator. The magnetostrictor converts oscil-
lations of electrical current of the appropriate frequency (20 kHz), which flows
through the magnetostrictor's winding, into mechanical oscillations of the core.
The oscillator serves the purpose of producing electrical current of the requirei
frequency and power. The magnetostrictor's core is made of a permalloy or of
nickel possessing the magnetostriction effect.
The operating principle of an ultrasonic cleaning unit consists in the creation uf
enormous local alternating-sign pressure on the aurface of wafers as the result of
the cavitation phenomenon, which consists in a discontiuuity in a liquid on the
surface of a solid and in the formation and then collapse of cavities. In the
collapse of microcavities enormous local pressure pulses are created. The cyclic-
ity of the effect of these pulses is determined by the frequency of the ultra-
sound. As a result of their repeated effect on the surface, solid particles
found on it are separated and remov4d by the liquid medium from the surface of
wafers. The water is made to circulate for the purpose of removing particles of
contaminants from the vat.
Units for the mechanical cleaning of wafers are the most effective in removing
solid particles and dust particlea from the surface. A great number of types of
units can be reduced to five basic systema (fig 2-5):
Cleaning rotating wa�ers by means of rotating brushes (ftg 2-5, a).
Cleaning rotating wafers by means of a soft moving tape (fig 2-5, b).
Cleaning wafers laid on a rotating inclined table by means of a roller made of
soft fur (fig 2-5, c).
Simultaneous 2-side cleaning of wafers by means of rollera made of soft fur (fig
2-5, d).
- 44 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
Cleaning'wa#ers laid on a horizontal rotating table by means of a roller made
of soft fur tn combtnation With a atrong jet of water (ftg 2-5, e).
oooa
A
i)~ e)
Figure 2-5. Schematic Diagrams of Units for Hydromechanical Cleaning of
Wafers: a--brush; b--tape; c--roller on inclined table;
d--simultaneous 2-aide; e--roller on horizontal table
Key:
1. Water
,
Either rests or vacuum suckers are used to hold the wafers on the table. In the
case of simultaneous 2-side cleaning the wafer is in a suspended state between
two rollers. The wafer is fed to the cleaning position along guides.
Universal-type units for the chemical processing of wafers from the "Lada-1"
complex are discussed below. This equipmeat is distinguished by the use of a
uniform preseurized chamber, 1, with a block of filters, 2, making it possible to
create a,:laminar flow of dustfree sir from top to bottom, which makes it possible
to perform all technological operations of cleaning and controlling wafers in a
dustfree~atmosphere (fig 2-6). In these units there are fluoroplastic vats, 3,
(from one to three), for processing in corroaive media and one polypropylene
cascade Vat for washing in circulating deionized water. For example, in chemical
processing units of the 0,8 ChKhN-100-001 and 0,8 ChKhN-100-005 type there are
three fluoroplastic and one cascade vat each. The holders with the wafers are
moved sutomatically by means of a naving mechanism in the first of these or
manually in the aecond.
The fluoroplastic vat is furnished with a heater, 4. The maximum heating tempera-
ture is 120 �C and the accuracy of maintaining the temperature is + 5�C. These
units are designed for washing wafers tn peroxide-asomonia mixtures, in detergents
and in various corrosive media.
The 08 ChW-0008-002 ultrasonic waehing unit ia designed for washing wafers in an
ultrasonic ffeld, in corrosive media, e.g., in nitric acid, and for washing in
- 45 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL. USE ONLY
deionized water. It contatns an ultrasonfc vat, one fluoroplastic vat with a
heater and a single cascade vat.
~ ~ f ~ , ~ �
2.
J
~1)
n
� ~ 3
c y
c
Key:
Figure 2-6. Diagram of Chemical Proceasing Unit
1. Exhaust
The 08 ChPVS-0/1500-004 wafer washing and drying unit consists of a box and a
centrifuge with a turntable for placing holders with wafers on it. It is passible
to load two or four holders simultaneously. Cleaning of wafers with deionized
water prior to drying is provided for. The washing time of 50 to 240 s and drytn;
time of 50 to 240 s, as well as tlie rotational velocity of the centrifuge of
200 to 1400 x.p.m., are set on a control panel. After termination of the required
cycle the unit turns off automatically.
The-04ChShch-75/4-001 automatic hydromechanical washer consists of four tracks
which opera;.e independently of one another (fig 2-7). Each track has its own con-
trol unit.
~
Z J n s s 7 .
' .o
~
i- ~
J ~
f
Figure 2-7. Diagram of Autrnaatic Hydromechanical Washer
-46-
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
The holder, 1, with the wa.fers, 2, is placed on the platform of a loading mechan-
ism. By�tneans of a conveyer, 3, wafers are fed one by one to the platform of a
centrifuge, 4, to which the wafer ts fastened by means of a vacuum and undergoes
the operations of cleaning, washtng and drying. The surface of the wafer is
cleaned by means of a rotating cylindrical brush, 5, with the simultaneous feeding
of a washing solution or deionized water and with the wafer rotating on its own
axis. Drying of the wafer is accomplished with a considerable increase in the
rotational velocity o� the wafer. 7'he tinre for the perfortnance of each operation--
cleaning.with the brush, first washing, second washing, drying--is set over one-
second intervals over the range of 0 to 99 s. The washed and dried wafers enter
a receiving container, 7, through inclined guides, 6.
The visual inspection unit consists of a dustproof box and an M-2 microscope.
In addition to the above, the "Lada-1" complex includes a number of units for
serving the cleaning line, in particular, a unit for transporting and supplying
reagents, a unit for heating deionized water,'a water purification system, a
water recycling unit, etc.
Units are put together in a cleaning line as a function of the specific purpose and
technological process.
Still greater possibilities for putting together cleaning lines and sections are
opened up with the modular design of units based on a dustproof box. The indivi-
dual washing units are interchangeable and can be matched in ane and the same
pressurized chamber (box) in any combination. For example, if necessary, the.
first thiee boxes can be totally made up of ultrasonic cleaning units or can con-
tain all types of units for carrying out the most complicated cleaning cycle.
2-3. Equipment for Controlling the Quality of'Wafers
After finishing cleaning, wafers are checked for agreement with technical require-
ments or,specifications. Some parameters--the diameter of wafers, the orientation
of the surface with respect to the prescribed crystallographic plane, the absence
of a defect layer, the length and orientation of the base cut--are guaranteed by
the technological process. The remaining parameters are checked during a finish-
ing check. Here the thickness of wafers, nonplaneness, nonparallelness, buckling,
the degree of cleanliness of the surface, the roughness of surfaces, and the pre-
~ sence and length of marks, cracks, spots, bruises and chips are checked.
Thickness, nonplaneness, nonparallelness and the buckling of wafers are checked hy
means of clock-type indicators with a division value of 0.001 mm and S-III or S-.LV
equipment racks. More precise measurements of thickness are made with an IZV-2
optical measuring machine, which makes it possible to measure thickness with an
accuracy of + 0.0005 mm. For the purpose of obtaining reproducible results, indi-
cators and measuring machines must be checked systematically against end gages or
standard wafers.
In measuring thin wafers of large diameter this method results in considerable
errors and in this case it is necessary to perform measurements on an MII-4
microscope. Far this purpose a standard, e.g., a mask, is placed on the stage of
the microscope and the microscope is adjusted for a sharp image of the lower
- 47 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2447102/09: CIA-RDP82-00850R000500494444-3
FOR OFFICIAL USE ONLY
surface of the mask. A note is made of the position of the stage's lifting
screw. The mask is removed and the wafer being studied is put in its place.
A sharp image is obtained of the surface of the wafer. The difference in the posi-
tions of the stage's vertical raising screw multiplied by the division value gives
the amount of buckling of the wafer. In measuring buckling on the MII-4 micro-
scope it is necessary to make sure that the direction in which the screw is turned
when adjusting for the standard and studied surface is identical (i`rom the bottom
to ttie top). The accuracy of ineasurements of buckling is determined by the divi-
sion value of the micrometer screw and can be brougYit to -I- 0.0005 mm. More pre-
cise measurements of the geometrical parameters of a wafer are made on a laser
interferometer of the UKP-2 type, from interference bands [11].
Instruments for non-contact testing of the geometrfc parameters of wafers have
begun to be used,ever more extensively in practice in recent times. These instru-
ments make it possible to make measurements rapidly with high precision without
scratching and contaminating the working surf$ce of the wafer. Their operating
principle is based on the change in the drop in pressure of air coming out of a
nozzle or the change in the capacitance of an electrode-stage system when wafers
of various thicknesses are introduced into it. For example, a capacitive pickup
is used in the ADE Corp. (USA) model 6033 non-contact instrument for measuring
thickness; the division value is 0.0001 mm and the measurement range is 0 to 1 mm.
The instrument consists of a stage, a capacitive pickup and an electronics section
with digital readout of the measured quantity. The operator places wafers down,
moves them and removes them.
Units for automatically checking the geometrical parameters of wafers are being
used ever more extensively, e.g., the UKTP-1, SPT-1, ST-100, etc.
The quality of cleanliness of the surface, marks, spots, bruises and chips are
checked on units for the visual inspection of wafers. A typical visual inspect~r
unit, e.g., the SA-710 and SA-720 models (f ig 2-8), consists of the following matn
elements: cassettes, 1, an optical microscope, 2, a stage for the wafers, 3, a
conveyer, 5, a manipulator, 4, and an electrical unit, 6. A straight-line system
is used for feeding wafers for the loading, sorting and unloading of wafers. A
high-speed belt feed mechanism conveys wafers from the loading position to the
inspection stage and returns checked wafers to the appropriate classification
cassette.
In model SA-710 and SA-720 units there is one cassette for good wafers and three
cassettes for wafers which must be reprocessed. A counter is provided for each
group of wafers. The capacity of a single cassette is 25 wafers up to 101.6 mm in
diameter. The productivity of these units is 200 to 300 wafers per hour.
In all units visual inspection is performed by an operator by means of an optical
microscope. MBS-1, MBS-2, MBI-11, MM[1-3, MII-4 and MTM-7 microscopes are used in
domestic units. Foreign firms use the EPY model microscope from the Zeiss and
Nikon firms.
A so-called air cushion (cf. ch 5) is often used as a conveyer. Wafers are di-
rected to a cassette by means of an air jet without touching the conveyer, which
eliminates scratching them and contaminating the material of the conveyer. For
example, in the model 5500 unit when the operator presses a foot pedal wafers are
-48-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
moved on an air cushion from the caasezte being inspected to a stage with a
vacuum sucker. The movement of a wafer fqx the purpos.e of inspecting its surface
is accomplished by the opera,tor by means of a manipulator. Thexe is an automatic
mode for moving a wafer at two speeds. Upon an instruction from a Poot switch,
wafers are returned to a cassette and rejected wafers are moved on the air cushion
to the holding position and are transferred by the operator into the approprigte
individual cassette.
Figure 2-8. Unit for Visual Tnspection of Wafers
Projectors, which considerably reduce operator fatigue, are often used in inspec-
tion units instead of an optical microscope. For controlling the quality of the
cleanliness of wafers a projector is used for the purpose of ineasuring the wetting
angle. A wafer is put into a chamber. A drop of liquid is applied to the wafer
by means of a dispenser and this is then projected onto a screen. The wetting
angle is measured by means of a scale graduated in degrees. The productivity of
a unit of this type is not greater than 60 wafers per hour.
The amount of microcontaminants on the surface of a wafer can be estimated with an
instrument of the ICh=2 type. This instrument consists of an X-Y stage, a ball-
type sensor, a weight and a traction element.. The instrument's operatlon is based
on the dependence of the coefficient of static friction between two surfaces on
the amount of contaminants. A polished steel ball has point contact with the sur-
face being studied. The traction element links the ball with the core of an elec-
tromagnet by means of which the force of friction is estimated. The amount of
contaminants on the surface is estimated from the instrument's readings and from
-49-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
a calibration curve. The instrument functions in the contamination range of
10 b to 10 9 B/CtA .
Some surface defects (buckling, nonplaneness, cracks, holes, etc.) are checked by
means of laser interferometers--UKP-1 and UKP-2 units for inspecting polished
wafers. The quality of the surface is judged fram an tnterference pattern con-
sisting of black and white lines and bands. However, the productivity of this
kind of inspection is not high--fram 60 to 100 wafers per hour.
Chapter 3. Equipment for Creating p-n Junctions
3-1. General Information on the Planar Process
The planar process is the basis of the fabrication of semiconductor structures.
The planar process's principle is that all processes of the creation of semicon-
ductor devices are carried out on one plane of semiconductor substrates, including
ohmic contacts and protective coatings. The development of photolithography and
diffusion technology and equipment, as well as the discovery of the masking and
passivation properties of silicon dioxide, led to the invention of the planar pro-
cess.
The technological sequence of operations for creating structures of devices is
given in fig 3-1 and the process for producing transistors in fig 3-2 [5]. As is
obvious from these figures, the entire technological process of the creation of
transistors can be arbitarily divided into two groups of operations:
Operations by means of which doped layers are created by the deposition of epitax-
ial films and by the diffusion of impurities, masking and passivating coatings hy
oxidation or deposition from the vapor phase, and ohmic contacts by the depositic,,,
of inetallic films.
Operations by means of which high purity of the surface of semiconductor wafers tu
be processed by chemical and plasma chemical methods is made possible and a pattern
is created in an oxide film and metallic films by means of photolithography.
The subject of discussion of this chapter is the first group of operations (the
second is described in chs 2 and 5). Tt must be mentioned that the technology for
the fabrication of integrated microcircuits and large-scale integrated microcir-
cuits is of course more complicated than that of transistors and requires addi-
tional operations to prevent stray coupling of their elements and to make possible
the prescribed (necessary) electrical connections. For these purposes regions
electrically insulated from the bulk of the chfp, in which active elements are
produced, are created in a chip. The two most widelp used methods of creating
the isolation of circuit elements are presented in fig 3-3. They are the method
of isolation by means of p-n Junctions and the method of dielectric isolation.
The first method is simpler in terms of technology and it is less expensive. The
second makes possible greater reliability and longevity, as well as enhanced
radiation resistance.
- 50 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
9892usune 901arsr Wswno- ~D~so- p~*ro-
~narucwr rpy~e rpar~
sa~ na~ur 6afr sr~ssspa Koasuc~ gossersoN
1) 3) i 6) unc oXor _ 9)
oqrcTws 41 I I ovrcna I I I Oareste I I I oywczrs
rtepsoe I I��P� 5~^ I I I~~~ I J I S~xre~ I I I e~~
orMCae r~e 1
soaoss aoA
xoaaersop-
ern eon:awr
Key:
Pigure 3-1. Typical Technological Process for Fabricating a Transistor
Structure of the 'n-p-n Type ,
1. Growing epitaxial film
2. Primary oxidation
3. Photolithography of base
4. Cleaning
5. Diffusion of boron
6. Photolithography of emitter
+
a)
1'%.� . n
,,,/,;�,,i b)
s~oz .
�A
c>
1 ~ 6a,toQo~ e6~icm~
p-muna
. ,

d)
.9nummr.pNa~ a6sa:ms '
tcey: 1. p-type base region
e)
f)
7. Photolithography of contact windows
, 8. Spraying of aluminum
9. Photolithography of contacts
19. Brazing of aluminum
11. Spraying gold beneath collector
contact
Figure 3-2. Technol.ogi,ca1 Pxocess for
yabricatfon of n-p n Planar
Tranfstor: a--starting wafer
of n -type silicon; b--depo-
sition of n-tqpe epitaxial
film; c--first thermal oxida-
tion; d--photolithography of
base and creation of base by
difPusfon of boron; e--photo-
lfthography of emitter by
diffusion of phosphorus; f--
photolithography beneath con-
tacts and deposition of inet-
als
Z. n -type emitter region
- 51 -
FOR OF'F[CIAL. USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFF'ICIAL USE ONLY
Msoastlte P-n aepa:oltar
"
NCxo1{xse
p
naeCerx~
21
Si Ot
OKMCaerre 
~posoa~sorps@~e
aoA pesAearrer,r
71e A007aIo 3)
4) itaprse
5
AM~~arn bops
Atn o6pe3ossure
raw~P7o~rx �
repreaos
NaoaRuEN ANaaeriparor 6
,
n
NcxoAase nXac-
~r-n
u
:Mea c :aprM-
wecxtu oMmcaor
St 02
.
9) P%
i/i ~ v la)naec:~u c re-
eecexemr xe eee
I ~ron~rpMC~aa~-
i: ecMrr caoer
`(IoaNKpressaa
xaprea
06pa3osaxre
xapyrrexa caAr-
~axow IIABCTN-
/ IW
Key:
Figure 3-3. Two Methods of Isolating Active Elements of Integrated Micro-
circuits
1. Isolation by means of
p-n junction
2. Origiaal wafer
3. Oxidation and photolithography
under separation diffusion
4. Pocket
5. Diffusion of boron for forming
isolating pockets
6. Isolation by dielectric
7. n-type
8. Starting wafer with thernjal oxidf,
9. Polycrystal
10. Wafer with polycrystalline film
applied to it
11. Formation of pocket by grinding
wafer of#
The more complicated technology af fabxicating integrated nicrocircuits (IM's)
and large-scale integrated microcircuits (BTM's) includes the following technolo-
gical operations: thermal oxidation of silicon waPers; diffusion of impurities
for the purpose of creating doped filme, including p-n junctions (in addition to
diffusion, ion-implanted doping is also employed); epitaxial growing, as well as
the deposition of po1}rcrystalline silicon in creating electrically isolated re-
gions by the dielectric isolation method; application of inetalization, including
-52-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42109: CIA-RDP82-00850R000500090004-3
multilayer and taulti].eVe1; 1Qw,tetqpexatare depoaitian ag dielect7Cic �ilws from
the vapoY phese for the purpoae of i,aol.ation in tau7.ti7.eve1 mexa7.fzation and for
protecting finiahed atructures freta the iafluence of the envtronment.
The planar process for the #aAxicatfon of semiconductor devices and TM's has been
discussed by using as an example the process employing silicon wafers, since
silicon both now and in the future Wi11 retain its domtnant position in the elec-
tronics industry.
Thus, in order to create in silicon wafers active structures for discrete devices,
IM's and%BlM's, the following equipment is required: diffusion furaaces, ion-
~ implanted doping apparatus, apparatus for epitaxial growing and apparatus for
applying metallic and dielectric coatings.
3-2. Diffusion Equipment
In the production of semiconductors and integrated microcircuits a wafer is sub-
jected to a number of high-temperature processes, including: oxidation--for
forming on the surface of the substrate a film (silicon dioxide), which is used
for the passivation of p-n structures, for masking the surface of the semiconductor
from the>diffusion of impurities, and as a gate oxide for MOS [metal-oxide semi-
conductor] devices and integrated fnicrocircuits based on them; and diffusion--
for creating doped layers in semiconductor substrates in forming active p-n junc-
tions, isolation between elements, separating regions, etc.
Diffusion of Impurities
Diffusion is the process, caused by thermal motion, of the transfer of atoms in
any material regardless of its state of aggregation. If atoms are distributed
nonuniformly in the substance, i.e., a concentration gradient exists, then the
directional flaw of atoms takes place from a region with high concentration to a
region with low concentration of the atotas of the material in question. Direction-
al flow originates similarly to this also in the case of the origin of a tempera-
ture gradient in the material. In this case atoms diffuse from a region with a
higher temperature (with higher energy) into a region with a lower temperature.
Processes of diffusion with a concentration gradient with constant temperature of
the substrate are usually used in tfie production of semiconductor devices.
Let us discuss one important diffusion parameter which in fact determines equip-
ment specifications. This is the diffusian coefficient:
D = Do e-ee/RT,
where k= 8.63�10 5 eV/deg is the Boltzmann constant, T is the absolute
temperature, and DD and lSE are fundamental diffusion parameters
-53-
FOR OFFICIAL USE ONLY
(3-1)
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFF'ICIAL USE ONLY
(DD is a constant corresponding to the value o# D with unl.imited T and
AE is the activation energy).*
Let us discuss the requixement,s for the stAbi7,i;ty af the tqaintenance cf aapecific
temperature fox di,ffusian furnaces. if we anat7.yze equa.fiiQn (3-1), it can be ob�
served that a change in tempexafiure of afew degrees can result in a 2- and 3-fold
increase in the diffusion coefficient, i.e., in the depth of occurrence of the
doped layer. For example, the diffusion coe#ficient increases approximately
5-fold with every 100 �C increase in temperature from 900 �C.
In designing diffusion furnaces it fs necessary to take into account the fact that
the accuracy of maintaining the temperature in the furnace zone must be not worse
than + 0.5 �C with the condition of triviality of the time for transient tempera-
ture processes when putting wafers into the furnace as compared with the time for
the diffusion process itself. In this case the variation in the depth of occur-
rence of impurities, e.g., of boron and phosphorus in silicon, will be in the
range of one percent, which satisfies the technological requirements for producing
p-n junctions for the canplex class of microwave devices for transistors with a
very thin base region (within the range of 0.1 u), variation in which entails
strong variation of the frequency properties of devices.
The need to maintain high temperatures is dictated also by the following facts.
Of course, the solubility of an impurity in solids and, in particular, in semi-
conductors, is determined by the kind of impurity and the temperature of the
process. The higher the temperature, the higher the solubility. In creating an
emitter region for transistors it is necessary to take into account the fact that
the concentration of the added impurity must be on the order of 1020 to 1021 cm 3.
This concentration can be achieved only with high temperatures (on the order of
1000 to 1300 �C). Thus, in designing diffusion furnaces it is necessary to 1)
guarantee a temperature in the furnace over the range of 900 to 1300 �C, and 2)
to guarantee accuracy in maintaining the temperature of not worse than + 0.5 �C
over the entire diffusion temperature range employed.
In carrying out the diffusion process, on the basis of requirements for high
cleanliness in carrying out the process, quartz, alundum and polysilicon tubes
are employed as the diffusion process chamber, having high melting points (higher
than 1300 �C), a long operating life under high-temperature conditions, and high
("semiconductor") purity of their raw material.
Tubes made of quartz glass, which have high purity, low internal stresses, a
small number of large bubbles, and also high transparency for ultraviolet and
infrared radiation, are chiefly used for diffusion.
Tubes made of polycrystalline silicon are superior to quartz and alundum in
purity of the material and permeability for alkali metals. Furthermore, the
*AE --the activation energy in di.ffusiqn--corresponda to aW ---the height o� the
potential barrier which a particle must surmount in order to go Prom one position
of equilibrium in the lattice to another, e.g., #rom one site or interstitial site,
respectively, to another site or interstitial site.
-54-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
lifetime o� a tube made of polycrystalline silicon at a tempera[ure o!; 1300 �C
is aimost 5-fold longer than fox tubes made of quaxtz, since quartz crystallizes
at this temperature and looses rAchanical strength. The use of tubes made of
polpcrystalline silicon has been hantpered by the campleuity o� the equfpment for
productng them.
Tubes with a round and, recentlp. a'rectangular cross secCion are chfefly used
as process chambers #or dif#usivn processes. Tu6es made of quartz glass are
used chiefly. The size of the tubes' cross section depends on the diameter of
the wafers betng processed. lln recent fumace models (e.g., of the SDO-125/V-15
type) tubes up to 150 mm in diameter can he used or tubes with a rectangular
cross section with a diagonal of approximatelq the same size. The substrate
holder (dish) and other elements of the loading unit located in the furnace's
working space are also made chiefly of quartz glass. Let us discuss the process of performing the diffusion operation in a furnace in
which, consistent with the above, the tube and substrate holder are made of an
especially pure heat-resistant material, e.g., of quartz glass, and the temperature
in the diffusion zone is maintained with accuracy of + 0.5 �C over the range dif
900 to 1300 �C. Here it is necessary to take into account the following negative
factors which must be avoided in designing diffusion furnaces:
Adding to a furnace a substrate holder with semiconductor wafers at room tempera-
ture introduces a disturbance of the static temperature conditions of the diffu-
sion furnace, and a considerable time (10 to 15 min) is required for the purpose
of establishing working conditions. During the transition period the accuracy of
maintaining the temperature in the diffusion zone will vary, which will result in
variation in the depth and distribution profile of impurities in the substrate.
In addition, the rapid loading or unloading of wafers from the high-temperature
zone results in the appearance in them of stresses and strains as the result of
thermal shock.
Holding the substrate holder and silicon wafers at high temperatures for a pro-
longed period can result in bonding of substrates with one another and with the
substrate holder. This is especially characteristic of tubes and substrate hold-
ers made of quartz.
The depth and distribution profile of {mpurities in the substrate vary with an
unstable feed rate of the gaseous diifusant and its uneven distribution over the
tube's cross section.
Shortening the duration of t:rsnsient conditions in the furnace during loading is
achieved by using a special te=verature restoration unit which boosts the fur-
nace`s supply power during loading, or by using preheaters (a prechamber) in
which wafers to be loaded are heated to a temperature 180 to 200 �C below the
furnace's operating temperature.
Bonding of the substrate holdex with the xeactor tube can be prevented by a vi-
brating movement of the substrate holder, accomplished by means of an automatic
loader.
- 55 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
DiCfusion units have been created for difRuai.an pxocesaes which tnc].ude, as a
rule, 3- or 4-tube difRusion fuxnaces with duat#ree ce11,s wi,th lam4nax floW$,
automatic ],oaders, gas cabinets, and also progxamers for a minicomputer. Tf
~ there are several of these comp7.exes, then one taore cowputer is used kox con--
trolling processes in a11 diffusion compl,eaces, as well as a conveyer used fox
the purpose of making possible transport flows of containers with wafers to
each furnace and from them to tlie site of tAe process f1ow control statton. An
automated 4-tube diffusion system is presented tn fig 3-4, consisting of :a 4-tube
diffusion furnace of the SDO-125/4A type, 9; autrnaatic loaders mounted on a
single base, 7; a dustfree ce11 with laminar f1ow, 8; a cabinet for the automated
control of the delivery of gas to the diffusion furnace, 1, which in turn con-
sists of a computer interface, 2, a power supply, 3, a peripheral input console,
4, a programmer, 5, and a signaling unit, 6; and a cabinet for controlling the
system with a prograimner and camputer interface, 10, consisting of a channel con-
trol unit, 11, a flowrate sensor unit, 12, and a temperature sensor unit, 13.
io
890
889
~ o00
000
~ oo.o
~
J 'l 1
Figure 3-4. Automated Diffusion System
Let us discuss the components of diffusion equipment.
Diffusion Furnaces
i.; r,: u
Three- and 4-tube diffusion furnaces are used at the present time in the electron-
ics industry. In order for the diffusion process to be reproducible, a tempera-
ture zone with nonuniformity in distribution of the temperature of + 0.5 �C is
required. The length of this zone determines the furnace's produtivity and
usually equals 600 mm.
Single-zone furnaces with three (the SDO-125/3-12 and SDO-125/3-15) and four
(the SDO-125/4A) thermal modules are used in the industry at the present time.
Structurally,'a diffusion furnace has two parts: thermal heating chambers and a
frame-type base with electrical unita. The heating chambers are attached to the
frame-type base.
Located in the fratme-type base are the units for supplying electric power to the
heating elements and for automatic control of the furnace's operating mode with
-56-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
unified tetaperature cantFo7,e of the RYeYIA^1 ox BPRT^1 zypP-. The heating chamber
inciudes a heater caztridge And a q;tAxtz xeacfiion tuDe. The heater caxxridge can
be of two types: w~th a heating e1,eatent covered with a thin 1ayer of a cexamic
coaCing based on pure alundum, And w-ith a heating e7.ement without a coating. Two
t y pes o f heater cartrid ges are showm fn fi g 3-5 [4]. A heat-Tesistant Wire 5.9 mm
in diameter wound i~nto a coil is used as tRe heating element. Nhen electric
current passes through it it ia hented to a temperature taaking it possible to pro-
duce in the reaction tube a taaximum temperature on the order of 1350 �C. In order
to control the temperature distrtbution profile over the entire zone, the heater
is made of several sections, mainlp of three. Contacts for connecting to indivi-
dual electric power supplies are led outside Prom each section by means of wires
made of the same heat-resiatant material. The coil�s winding pitch is determined
by the installation oP'ceramic insulators, and their projecting ends serve as a
support for a ceramic muffle tube. �
.r.
~ I !'jee ff
a~`-1 Q r_

~ . _ : b)
Figure 3-5. Design of Heater Cartridges: a--with ceramic covering based
on alundum; b--without covering
Key:
1. Unit I The center heating section makes it possible to produce uniform heating at the
center of the zone, and the two end o:tes, equalization of temperature for the
ends of the center section.
Control of the temperature in a diffusion furnace is accomplished by means of
platinum - platinum-rhodium thermocouples and a temperature control unit of the
RYePID-1 or BPRT-1 type connected to the power supply. Five thermocouples are
used. The thermocouples are placed at the center of each section and at the ends
of the center section. Tt is possible to become acquainted in greater detail
with control of the temperature of diffusion furnaces in [4], for example.
In order to reduce the area occupied and the number of dustfree ce11s and of
other systems, di#fusion furnaces are arranged either side by side or end to end.
In this case the dustfree ce11s with laminar f].ow are united into a single one
in which automatic loaders are used. The gas systems are also united into a
single unit.
-57-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
9,ren ! 1~ , A-A ~

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
FOR OFFICIAL USE ONLY
Au[omatic Loaders
With an increase in the diameter of wafers being produced there wats an increase
in rejects uf wafers in diffusion operations, caused by the origin of high in-
;ternal stresses when loading and unloading wafers from high-temperature furnaces.
For the purpose of eliminating these rejects, as well as rejects associated with
bonding of the quartz substrate holder directly to the quartz tube of the furnace,
automatic loaders are used (fig 3-6) with a reversing mechanism making it possible
to add wafers to the furnace and remove them smoothly.
Figure 3-6. General View of Automatic Loader
Automatic loaders perform the following functions:
Putting wafers into a diffusion furnace at a certain speed which can be adjusted
by degrees, e.g., 100 to 200 mm per min.
Rocking the substrate holder (dish) in the high-temperature zone (roughly 2.5 mm/
/min over a range of 20 mm).
Unloading wafers from the high-temperature zone at a certain speed (e.g., 25 to
100 mm/min).
-58-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
In order not to introduce contamtnanta, the automatic loader'a pushex.rod is
made of quartz of the same purity ae the tube and dish. Thus, the use of auto-
matic loaders dxastic$l1y reduces rejects of wafers, especially of large dia-
meter, eliminates exror on the part of tAe operator and reduce$ the 1abor inten-
siveness of xhe diffusion Operaticm. In addt,tion, the repxoducibility of the
characteristics of doped lapers fraa batch to batch of Wa#exs is increased.
Dustfree Cells
Dustfree cells are installed at the tnlet to diffusion fuxnace reactors for the
purpose of ensuring cleanliness in transferring Wafers from a container into
substrate holders and in putting them into diffusion tubea (cf. ch 13).
Automated Gas Cabinets
The automated gas cabinet is designed for preparing, delivering, regulating and
monitoring flows of vapor-gas mixtures in the diffusion and oxidation of semi-
conductor wafers.
As a rule two cabinets are provided in diffusian systems for delivering to the
quartz tubes mixtures of vapors of the diffusant with argon and for supplying
dried or moist oxygen.
Mechanical diagrams for diffusion (fig 3-7, a) and oxidation (fig 3-7, b) are
prPsented in fig 3-7. There are three lines in the cabinet for the diffusion pro-
cess: for supplying inert gas for blowing tubes through and for freeing them from
the atmosphere's air; for supplying oxygen; and for delivering to the reactor
a vapor-gas *aixture of inert gas and diffusant. Electromagnetic valves for clos-
ing lines, rotameters for monitoring the gas flowrate, flow regulators, and sta-
bilizers for a specific RaR flowrate are uaed in the system. Liquid compounds,
e.g., BBr3, PC13 and POC132 or gaseous, B2H6 and PH3, are used as a rule for the
diffusion of boron and phosphorus into silicon. In individual cases methods of
diffusion from a parallel source or from a deposited surface source--a film con-
taining a diffusant--are used.
The principle of diffusion from a parallel source consists in the fact that the
source of the diffusant and the silicon wafers are placed parallel to one another:
The source is above horizontally lying wafers. A gas medium flow passes between
them. Vapors of the impurity diffuse through the g$s, strike the surface of the
silicon and form there a surface diffusant source--a film of liquid glass, e.g.,
borosilicate, B203�Si02.
The surface source--a film containing the diffusant--can be formed by the high-
temporature deposition of doped fi1ms of Si02 or by depositing films of a dis-
perser of a similar composition b}r the method of centrifuging followed by vitri-
fication.
The use of single-zone diffusion furnaces with a simplified gas system is re-
quired in all of the above-named methods.
-59-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFF7CIAL USE ONLY
1 ~
6noK ~P
Ar r N I
a)
1 .
i r N
M
Tul ~
7-
6ao~' �
oXrcnr#r,F i
EoaOyi~
2)
b) 3F
~
Key:
~
~Oqar
i 5)
r-----
I ~ I
L'"~'. . J I r
.wu~wuB
Oa~m~aaMm
4)
~
/n I
~
~
P
A R I ,f ~ 7/yQf 1
15) f-10) nQa ~I 6)
~ AMMOC CONI/$ 7 ~
K 6 . ~oaDy.~
I Nz0 I TI? ~2)
- ~----------~I
Figure 3-7. Diagram of Delivery of Gases and Diffusants to Diffusion
Furnaces: a--for diffusion; b--for oxidation; RD--pressure
regulator; M--manometer; K--electromagnetic valve; R--rotameter;
Tp--thermocouple; N--flow regulator; SM, E1--motor; Dr--choke
1.
Diffusion unit
9.
2.
Oxidation unit
10.
3.
Air
11.
4.
Liquid diffusant
12.
5.
Tube
13.
6.
Vapor
14.
7.
Atmospheric air
15.
8.
Pressure regulator
Manometer
Electromagnetic valve
Rotameter
Thermocouple
Flow regulator
Motor
Choke
Plasma chemical methods of cleaning, which have replaced liquid chemical, are
being used extensively in recent years. The equipment used for these purposes has
small overall dimensions and increases the percentage of the yield o� suitable
structures. The operating procedure� of a diffusion systam with a plasma chemical
cleaning unit (fig 3-8) is as follows: The holder with the wafers enters, from
the automatic loader, the plasma cleaning module and through it enters the diffu-
sion furnace. For the unimpeded travel of holders with waPers from the furnace
to the automatic loader and vice versa, swing-away sealing covers have been made
at the ends of a module's quartz tubee.
-60-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500490004-3
Ic should alsQ be mntivned t1'At Al,au prawiatn$ fox digfuaicm pxctce$%es is xhe
uae of radiant tAfraxed heating, vh;tch. PakAa poasib1e a slwFt time- for the fuxnace
to wnrm up and a mximwa degxee of cleAnl.#tes:s t,n caxxying aut tne pxocess.
Figure 3-8. General View of "Plasma-modul' FT" Predifrusion Plasma Chemical
Cleaning Module for 3-Tube Furnace
3-3. Equipment for Ion-Implanted Doping Proceases
Ion implantation is the introductfon of ionized atoms of an impurity into the
surface layer of a substrate as the result of imparting to these atoms high
kinetic energq (fzom keV to MeV). Aa the ion advances in the substrate it grad-
ually loses energy on account of electronic and nuclear deceleration and ulti-
mately stops in the substTate at the approprtate depth from the surface. The basic
distinction DetWeen the ion-implanCed doping method and thermal diffusion is in the
method of imparting energ}r to impurtty atows: Tn thernaal diffusion it is thermal,
on account of the htgh temperature (on tRe ordex of 900 to 1300 �C), and in ion-
implanted doptng tt is electrical on account of the ionization of vapors of the
- 61 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
impuri ty substance and of theix acce].eration with the application of a high
potential difference.
The advantages of ion-implanted doping ovex thermal difi'usion cqnsist in the
following capabilities:
Doping solid substrates with atoms of any substance regardless of its maximum
solubility.
Doping at any temperat ure right doQn to very low.
Creating in the substrate a concealed layer at some distance from the surface of
the substrate.
Produc ing non-deep (on the order of 0.1 and less) doped layers, including
multistage.
Producing a high degree of purity of an introduced impurity.
Doping a substrate through a protective layer.
Controlling with high precision the depth and distribution profile of impurities
in a substrate by changing the energy and dose of introduced impurity ions.
Among the disadvantabes can be numbered the complexity of the equipment and re-
sidua 1 radiatiun defects in the substrate.
The f o llowing are necessary to accomplish ion-implanted doping:
Ioniz ing the impurity substance.
Impa r ting to the impur ity ion the appropriate energy in order to introduce it
at th e prescribed depth in the substrate.
Separating impurity ions by mass from undesirable elements.
Direc ting ions to the surface of the substrate in order to introduce them.
As is obvious from th is sequence of operations, ion-implanted doping apparatus
must include the following main units: an ion source, an accelerating tube,
a mas s separator, and scanning and receiving equipment.
Ion-implanted doping units are classified by the method of acceleration (fig 3-9).
if the acceleration of ions is accomplished before the mass separator, then these
units are called units with pr,,acceleration (A). Tf the unit is based on the
acce leration of impurity ions b2ter the mass separator, then this unit is a unit
with postacceleration (B).
If low energy of ions is required (up to 50 keV), then an accelerating tube is not
used in the unit and the acceleration of ions is accomplished on account of an
extraction potential applied in the ion source (C). Tn these units the magnet and
-62-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
receiving unit axe undex ground potentia7., which ensures e],ectrical safety fox the
work of operating pexeonnel and woxk convenience,
.
.
A
r-----~
4)
I
NoexW
1
Yscc-arveps-
CReerpya~ea
Uprerxoe ~
Mciovarx
zop J
)orpopa:3a
7c:po~crso
~l_

J
y.- v.
~ r-----~

i
HoxxW
Y111oc-caaapa-
Cuexrp noqee
(ipreraoe
~
NCSOVNMII
top
yCSpoilCTSO
ycTpoNc'eso
!
L
J

J

~
�
I r-----,

I
~
~ i
BoerW
Ysoc-ceaape-
Cxerrp7n[ee
Dprersoe '
N
rcroaeax
sop
~
~cspo~ctso
~cspo
ctw
~~LL =~===J
. V'------~
Key:
Figure 3-9. Classification of Ion-Implanted Doping Units by Method of
Acceleration: Uo--ion extraction voltage; Up--accelerating
voltage
1. Ion source 3. Scanning unit
2.. Mass separator 4. Receiving unit
Units of type A are used for low energies and any ion mass, but the weight and
overall dimensions of the electromagnetic separator increase drastically with
an increase in energy. In units of type B the magnet is under ground potential
and they are thus electrically safe; therefore, it is possible to use large
magnets for the mass separation of ions of heavy elements and, by using an
accelerating tube before the receiver, to accelerate ions to high energtes. The
disadvantage of these units is that the receiver is under high potential.
Units of type C are used to produce beams of ions with low mass and high energy.
Here the magnet and ion source are under high potential and the receiver under
ground potential.
Ion Sources
An ion source is designed to ionize vapors of substances introduced into it and
to extract ions of atoms (molecules) of this substance into an ion conductor or
accelerating tube. It consists of the following elements: a discharge chamber
connected by an opening with the ion conductor or accelerating tube, and a unit
for extracting and focusing ions.
With the introduction into the chamber of a gaseous substance, a plasma forms in
the discharge chamber. The ionization of vapors or a gas takes place by the
_ collision of electrons with atoms (molecules) of the introduced substance. A1
discharge in the chamber is usually produeed at 1ow pressure of "(13 to 1)010 Pa,
in order to make possible the requixed density o# the plasma. The discharge
chamber is'usually placed in a magnetic fie1d for the purpose of increasing the
probability of the collision of an electron with atoms (molecules).
-63-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
With respect to the method o.f creating the plasma, ion soux'ces axe divided into
ion sources with a high-fxequency di.&charge, with dis.charge from a heated cathode,
znd from a cold cathode.
A schematic diagram of a source with a high-frequency discharge is shown in fig
3-10. It is economical, and simple in design. However, this type o� source is
characterized by a large spread in the energy of ions, its operation becomes un-
stable at high temperatures, it does not make it possible to pxoduce currents of
high density, and it also requires controlling a great number of parameters in the
process of its operation. As a rule it is designed for operating with gases.
The operating principle of a high-frequency source is as Pollows. A gas discharge
is ignited in the chamber, forming a plasma on account of a high-frequency magnet-
ic field with a frequency of 10 to 40 MHz created by coil 1. A potential differ-
ence on the order of 3 to S kV is applied between the anode in the upper half of
the chamber, 3, and the extracting electrode (the cathode), 2. High-frequency
ion sources make possible low current (hundreds of microamperes to single numbers
of milliamperes), are designed for low power, and are simple in design. ~
ti
+UE
raa 1)
3 - iZ
~
2)
Key:
Figure 3-10. High-Frequency Ion Source
1. Gas 2. Ion
A Penning ion source in which discharge with a cold cathode is employed is
presented in fig 3-11. In the discharge chamber of this source a self-maintained
glow discharge is created, originating as the result of the applied potential
difference (on the order of 5 to 7 kV) between a cold cathode, 4, and an orificed
anade, 6, an of the application of a longitudinal magnetic field created by coil
1. The anode, anticathode, 3, and cathode are electrically isolated by means of
insulators, 5. Aluminum, beryllium, uranium and other cathodes having a low
sputtering coefficient are usually used in order to increaee the service life of
the catiiode (reduce its sputtering).
The cathode is coated with an oxide film in order to inerease ion-electron emission.
The anticathode, 3, serves the purpose o# forming a plasma front and the extract-
ing electr.ode, 2, in addition to its main purpose, of focusing the beam.
Sometimes the transition of a Penning discharge into an axc discharge in the
absence of a magnetic fie1d is employed for the purpose of producing an arc
-64-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
discharge. WiCh thi,a it ia pcateiAle to tncreAse xhe extxAGtian of *an current
from the source.
S 6 121/
y +
~ l)raa -

+ 3
~
xey:
Figure 3-11. Structural Diagram of Pennfng Ion Source
i. ca$
A Nielsen ion source is shown in fig 3-12. By means of this source it is possible
to produce ions of substances from liquid and solid compounds, as well as from
gases. It consists of a tungsten hot cathode, 4, an electromagnet, 1, an extract-
ing electrode, 3, an anticathode, 2, and insulatora, 8. A crucible, 6, with a
heater, 7, making it posaible to heat the crucible from 170 to 900 �C, is em-
ployed in order to vaporize liquid and solid sublimating materials. Cylinder 5
serves as the anode. By means of this ion source it is possible to produce strong
beams of gaseous elements and chiefly of hard and high-melting elements including
boron, carbon, aluminum, silicon, iron and the like. . "
/'aj .
8 . t ,
I I I 1'
23
/
6 7 S ! 4
Key:
Figure 3-12. Nielsen Ion Source
1. Gas
The schematic diagram of a Moxozov s7,it-type i:on source, by means of which it is
possible to produce ion beams of the ribbon ty�pe Mrith a circulax cross section,
is shown in fig 3-13, a. A thermionic cathode, 3, with a screen, 4, heated by
a heater, 2, serves as the electron source. An applied magnettc fie1d of
-65-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFF7CIAL USE ONLY
H ti 4�105 A/m promotea Che formattvn of a Deam of electrons moving toward the
anode, 8. .
2
J
N + �
IIIII I S
IIII B
~ IIIII
1~ Illll pmOK
IIIII
Paa~ ` ~u~~ aoNOa
11111 I '-7
L=J
Key:
H~
_ b)
Figure 3-13. Morozov Ion Source: a--for gaseous substances; b--for solid
substances �
1. Gas
2. Ion beam
The plasma, 5, forms with the admission of gas through the gas distributor, 1.
In the diagram in fig 3-13, b, a crucible, 10, with the substance, which must be
vaporized into-the discharge zone by mearis of heater 9, serves as the anode.
The shape of the ion beam is determined by the shape of slit 7 and of the openinf;
in the extracting electrode, 6. By means of this source it is practically possible
to produce ions of any substance, including high-melting, as we11 as multiply
charged ions.
One of the most widespread is an ion source of the duoplasmatron type with an arc
discharge developed by Ardenne. Its fundamental difference from ion sources of
other types consists in double contraction of the plasma. (fig 3-14), i.e., con-
centration of the plasma in the required region of the discharge chamber by means
of electric and magnetic �ields, 6, of the appropriate configuration. The methods
of forming the plasma are as in the ereceding source.
Figure 3-14.
r'' ;ti;ra S t
41
z: �
a )
Diagram oP Electric and Magnetic Contxaction of Plaszna in
Duoplasmatron: a--electrtc contraction of plasma; b--~aragnetic
-66-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
An additional anode, 2, of the appropriate shape, the potential acxoss which
differs fraa the potenCial acrosa the main anode, 4, is introduced for the
purpose of contracting the pl$ema 6r means of the elentri.c field. As a xesult,
an additional potential, difference #s created inside the channel and between the
cathode, 1, and main anode, 4, an electric double layer, 3, is formed, which
focuses and accelerates electrons into the region of the opening of the main
anode. With this the ionization efftciency is increased, Which results in an in-
crease in the concenCration of ions extracted through the anode opening and in
the formation of a distinct plasma boundary, 5, in the outlet from anode 4.
Electric contraction of the plasma is illuatrated schematically in fig 3-14, a, and
in fig 3-14, b, magnettc contractton of the plasma on account of a strong in-
homogeneous magnetic field. Here the additional electrode, 2, and the main anode,
4, are the magnet's pole pieces. As a result, the concentration of electrons
and therefore the ionizatton ePficiency are increased. Inserts made of high-
melting materials--molybdenum or tungsten--are fastened inside the electrodes
�or the purpose of increasing their heat resistance.
Accelerating Tube
The accelerating tube is designed for accelerating ions and depending on the
type of unit it is placed before or after the mass separator. The additional
focusing of the ion beam is also accomplished by means of it.
As indicated previously, accelerating tubes are not used in units in which low
ion energies are required (on the order of 40 to 50 keV), since the acceleration
of ions in them takes place on account of the use of an accelerating electrode
at the outlet of the ion source. For the purpose of producing a uniform distribu-
tion of voltage along the length of the accelerating tube, it is usually made
sectional by alternating electrodes, 1, with insulating rings, 2, and by employing
a voltage divider, 3(fig 3-15).
Figure 3-15. Multisection Accelerating Tube
Key:
1. Ion beam
-67-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
As an insulator are u$ed uAteri$ls which have $ high bxeakdown Voltage and which
do not rel.ease and do not transwit gases, aince there tqusC be a high vacuum
inside [he tube. Ppxcelain and specia7, epoxy cc7mpounds axe chie,f1,y, used xor
insulating rings. The metal electrodes, 1, are made in a sttape which reduces
the space chaxge fornsed in the die7,ectric ineulating ring, 2, aad Which changes
the mechanical trajectory of ions.
Mass Separator
In order to facilltate the creation and control of ion beams it is best to intro-
duce the doping impurity in the gaseous atate into the source. But it is difficult
to obtain in the gaseous state such substances as phosphorus, boron, arsenic, etc.,
used for creating doped layers in silicon; therefore, gaseous compounds of these
substances are used, such as diborane, phosphine, arsine, etc.
The ion beam, which consists of a whole number of impurity elements, must be
cleaned of undesirable elements upon entry into the substrate. Mass separators
are employed for these purposes. They are divided into two types with respect to
operating principle: magnetic separators and electromagnetic filters.
Ions with the same charge and energy will be deflected at different angles in a
transverse magnetic field depending on the mass of the substance. By reversing the
magnetic field it is possible to produce beams of the prescribed composition in
the outlet of the sector magnets of mass separators. This method of separating
pure substances is considered the most effective. The relationship between the
parameters of the magnet (radius, r, angle of turn of the ion, 6, magnetic
induction, B), the mass of the ion,~M , its charge, q, and the accelerating
voltage, Vo , is presented below:
(M/q)U0 = Kr0 B2 ,
where K is determined by the angle of the sector (magnet).
Consequently, if the appropriate value of magnetic induction, B, is selected with
constant assigned design dimensions of the magnet, then in its outlet, with
radiue ro , it is possible to separate through the slit an ion beam with mass
of M/q .
Magnets of the sector type with an inhomogeneous magnetic field are usually used
for the purpose of increasing resolution. A diagram of an ion-implanted doping
unit with a magnetic mass separator of the sector type, 4(the sector angle
equals 90�), and of an ion source, 1, with an accelerating tube, 2, is presented
in fig 3-16. The beam.o,f ions of the doping substAnce, passing through the co7,-
limating opening, 3, is scanned by me$ns of potenCials supplied to the electrodes,
5, over the surface of the substratas p'laced in the receiving chamber, 7. The
beam current is measured periodically by nteans of a probe, 6. The radius of the
magnet is determined by the acceleratiag voltage and the mass of the substances
used for doping,
r0 = 2MV~ qB .
-68-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02/09: CIA-RDP82-00850R004500090004-3
Pigure 3-16. Design of Ton-Implanted Doping Unit with Magnetic Mass
Separator
2
- 11~ � 7
ro ~
.
�
' / .
NedmpaA~Nwit 9 S 6
nyvoK
xey:
1. Neutral beam
USE ONLY
The design of theae mass separators is very simple, but for accelerating voltages
of hundreds of kV and for great masses of ions of doping impurities (e.g.,
phosphorus, arsenic and the like) electromagnets with large dimensions and heavy
weight (several tons) are required.
Units with postacceleration have begun to be used in recent times. In this case
the preacceleration of ions before the mass separator is slight (15 to 20 kV);
therefore, the dimensions of the electromagnet are drastically reduced.
A similar effect for the mass separation of an ion beam can be accomplished if it
passes through a filter in which the magnetic field acts along the path of the
beam and the electric field acts orthogonally to the magnetic field and consequent-
ly to the ion beam. This is�a so-cAlled E X B filter (fig 3r17).
Z J k S 6 7 B
4
~ Z
9 /I
~

, 77
~
Figure 3-17. Design of Ton-Tmplanted Doptng Unit with Maes Separator of
the E XB gilter Type
The mass-separated beam will continue its straight-line motion and the ions of
other masses will be introduced into the walls of an ion conductor. This mass
separator is simple in design: It consista oP a amall-size permanent magnet, 4,
and electrodes, 5. The original ion beam of doping impurities, 1, passing through
-69-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02/09: CIA-RDP82-00850R004500090004-3
FOR OFFICIAL USE ONLY
the collimator, 2, is sepaxated Uy taeans of an E X Bf ilter, 3. The disadvantage
of th is filter is the fact that together with the ion beam, 6, a neutral beam,
8, of undesirab].e substances passes through it. Tn order to eliminate this, in
the f ilter's outl.et there is a system for deflecting the ion beam by means of an
electric field, 7, with a collimator, 2. As in the unit shown in fig 3-16, the
ion beam is scanned by means of electrodes, 9, over a substrate placed in a
receiving chamber, 11, for the purpose of producing a doped layer uniform over the
depth and surface of the substrate. A probe, 10, serves the purpose of periodic-
ally checking the ion beam's current. On the basis of the use of mass separators
oP one type or another, as we11 as of the convenience of putting together elements
of the entire unit, various designs of ion-implanted doping units are employed
(fig 3-18): of the horizontal type (fig 3-18, a and b), of the vertical type
(fig 3-18, c to e), with sector magnet mass separators, 2(fig 3-18, a, c, d),
and mass separators, 4, of the E X B Pilter type (fig 3-18, b and e). The ion
sources, 1, and receiving chambers, 3, are selected on the basis of the specific
application of the unit.
3 ~
1 i 3 ~ 4 3 2 ~
_ 2 3 31
a) b) c) d) e)
Figure 3-18. Design of Ion-Implanted Doping Units
Receiving Unit
The receiving unit is designed for loading semiconductor wafers, moving them
beneath the ion beam, doping and heating. The cross-sectional area of the ion
beam is as a rule several square millimeters and the beam is nonuniform over its
area. Therefore, for the purpose of doping wafers with high uniformity it is ne-
cessary to move either the beam over the wafer or the wafer relative to the beam.
The first method is used most often--scanning of the beam as in a television tube.
In the doping process, for the purpose of annealing radiation defects which ori-
ginate in wafers, they are heated by means of heaters to a temperature of 200
to 700 �C.
Add itional focusing lenses are employed to increase the intensity of the beam.
Depending on the design of the unit, they are placed in �ront of the mass separa-
tor or behind it in front of the beam scanning system. Electrical and magnetic
quadrupole lenses (duplex and triplex) are usually used.
Vacuum System
The vacuum system is an important part of the apparatus. Tt is necessary to main-
tain a high vacuum on the order of 1.3�10 4 to 1.3*10~6 Pa in the vacttum ion
conductor for the purpose of ensuring high purity of the ion beam. However, for
making it possible to produce an ion beam of high intensity it is necessary to
introduce into the ion source a high coYicentration of the doping substance,
- 70 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
which is accomplished at a pressure of 13.3 to 1.3 Pa. A differential evacua-
tion sYstem which incxeases the vacuum in the receiving chamber to 1.3�10^5 to
1.3�10 6 Pa is created between the ion source and the receiving chamber. As a
rule the receiving chamber is furnished wtth powerful high-vacuum evacuation
equipment with freezing or other traps which prevent the entrance of oil vapors
from oil-vapor and mechanical vacuum pumps, or with oilless evacuation equipment,
chiefly turbomolecular punips, as weil as with metal vacuum seals.
The "Vezuviy-2" [Vesuvius-2] ion-implanted doping apparatus is widely used fn the
series production of semiconductox devices (fig 3-19). The control rack is shown
on the left in the photograph. The postaccelerstion principle is used in this
apparatus; therefore, ehe magnetic mass separator has small overall dimensions and
low weight. The preliminary acceleratton is up to 20 keV and the subsequent up
to 130 keV on account oP the use oP a multisection accelerating tube. The total
maximum accelerating voltage is 150 kV and the minimum 20 kV. The receiving unit
is under high potenrial.
Figure 3-19. Ion-Implanted Doping Apparatus of the "Vezuviy-2" Type
The ion source is of the arc type with a transverse magnetic field. At the outlet
of the accelerating tube the boron and phosphorus ion beam current is a maximum of
100 uA and 300 uA, respectively. The number of wafers up to 80 am in diameter
which can be loaded at the same time is not greater than 40. Vertical scanning
of the beam is accomplished by means of the electric field and horizontal by
mechanical movement of the subatraCes. An integrating dosimeter ie used to monitor
the doping dose. The apparatus is compact and has sma11 overall dimensions. The
area occupied without auxiliary apace is 18.m2. The apparatus is simple and reli-
able.
- 71 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
3-4. Equipment .Eax Pxoduci.ng Epttaxi.A7. I'ilma
Methods for the deposi tion of atcnas fron the vapor or 1,iqu:Cd phase onza solid
substrates are eiqployed for the puxpose of producing monocrysta7.7.ine semicon-
ductor films. This pxocess is call.ed epitaxial iP crpstal7.ization of the de-
posited material takes place on oriented monocrystalline substrates with the
repetition of their structure. The distinction is made between Iiomoepitaxial,
autoepitaxial and heteroepitaxial grow3ng of tilms. Tn the first case the de-
posiCed substance and the substrate are identical and in the second different.
At the present time successes have alread}r been achieved in the depc+sition oP,
for example, silicon monocrystalltne films on sapphire and quartz substrates and
substrates made of spinels.
Two trends exist in the epitaxial growing of semiconductors on substrates of
a different substance, including on dfelectric substrates. The first is the
deposition of a thin layer of a molten semiconductor with its subaequent oriented
crystallization. The second is the epitaxial growing of layers of a semiconductor
during crystallization from a melt, a melt solution and the gas phase.
Methods of epitaxy from melts, solutions and the gas phase have become the most
widespread in the production of semiconductor devices. Epitaxy from melts makes
possible high structural peifection and purity of films, and by this method it is
possible to produce layers of such materials as indium antimonide and gallium
antimonide, which are difficult to produce by means of gas transport reactions.
Epitaxy from the gas phase includes the deposition of monocrystalline films by the
vacuum deposition method (thermal vaporization, ion-plasma spraying, electrical
firing) and by the method of gas transport reactions.
Vacuum methods make it possible to produce films of very high quality, but they
require a superhigh vacuum and are poorly productive.
The reduction of a tetrachloride (SiCl ) or a trichlorosilane (SiHCl ) and the
pyrolysis of a silane are widely used gas transport reaction methods3 A vapor-gas
mixture of the appropriate compositton passes over a substrate heated to the
appropriate temperature, being deposited in the form of monocrystalline layers
of the substance or compounds in question. These processes are usually accom- �
plished in quartz reactors.
The equipment created for carrying out epitaxy by these methods is relatively in-
expensive, uncomplicated and highly productive.
Depending on the composition of the reactive gases, deFosition processes can differ
in the type o� reaction, which has an influence on the individual designs of
reactors. For example, the basic difference between epitaxy processes for silicon
and gallium arsenide is the fatet that in the first case this reaction ts endo-
thermic and in the second exothermic.
Therefore, in growing silicort, where its deposition results in the absorption of
energy, for the repeated perPormance oP the epitaxy process without the wa11s of
reaction chamDera becoming overgrown, it is necessary to hold them at a tempera-
' [ure below the deposition temperature. For these purposes heating, either
-72-
FOR OFFiCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
high-�requency i,nduction ox by means of infxared lamps, is empl.ayed. In the
deposition of galliutn axaenide fxom chlorides accompanied bp the releaae of
energy it is neceswxy to heat the walls for the purpose indt,cated aAove. Tn
these cases it ta desixable Co uae res3sttve hesters making it possi:ble to heat
the walls of cRatabexs or to make reaction chafltbers out of heat-absorbing materialg.
Liquid epitaxy taethods axe also used for semicanductor coutpounds of the ATTTBV
type in the industrial production chiefly of optoelectroaic devices.
Equipment for Gaa Epitaxy,
The deeign of equipment designed #or epitaxy according to the method of gas
transport reactions is characterized by 1) the type of epftaxial reactor, 2) the
method of heating substrates, and 3) the gas distribution system.
Epitaxial reactors are divided into three types: horizontal, vertical and cylin-
drical. (A classification of reactors according to othez features--the direction
of gas streams relative to the subatrate, heating of substrates, etc.--is $iven
in [14].)
This division according to type of reactor is based on the difference in the posi-
tion of the substrate relative to the reactor's axis and the direction of flow of
the reaction gas and is tradittonal.
The three types of reactors are shown in fig 3-20. Horizontal reactors
(fig 3-20, a) are the simplest and do not have any moving parts inside. The
vapor-gas mixture stream in them is supplied parallel to the surface of the sub-
strate and the axis of the reactor. Usually inaide the tube there is a holder
made of graphite, 3, coated with silicon or silicon carbide, which is mounted on
quartz slides or rectangular supports with ,3 certain pitch, a. The quartz or
metal tube, 1, has a circular or rectangul.ar shape. The vapor-and-gas mixture
in the inlet nozzle passes through a gril.le, which forms a turbulent flow, as the
result of which good mixing of the work.ing mixture with the gaseous doping sub-
stance is achieved. Preliminary hesLing of the vapor-gas mixture is carried out
in order not to introduce a disturbance into the reactor's temperature zone. In
the outlet the exhaust mixture is cooled to 50 �C and is burned up in an outlet
unit (scrubber). The graphite substrate holder with the substrates, 4, in this
case is heated from a high-frequency oscillator through a work coi1, 2.
In a vertical reactor (fig 3-20, b) a quartz dome, 5, is used as the reaction
chamber and the aubstrate holder, 7, made of graphite coated with silicon or sili-
con carbide as in a horizontal reactor, ia mounted on a rotating platform, 9,
which is driven into motion by a shaft, 8. The vapor-gas mixture enters the
chamber through a rotating tube, 6, and, being repelled by the quartz dome, is
directed toward the substrates, 4. Tn this case the graphite substrate holder is
heated from a high-frequency oscitl,atox by means. of a woxk coil, 2. In this reactor
the vapor-gas mixtuxe can be also supplied directly through the top part of the
dome. In both cases it enters parpendicularly to the substrate and spreads over it.
The cyltndrical reactor (fig 3-20, c) consists of a quartz or steel chamber, 5,.'
in which there is a graphite substrate holder, 7. The substrates, 4, are placed
in it on an incline, which prevents them from falling out of the grooves. As a
-73-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
F'OR OFRICIAL USE ONLY
rule, the substrate haldex is drtven into rotatfion fox the puxQose oR uniform
contact between the vApox-gas mixture and the $urface of the substrate and of
mixing the mixtuxe; hexe it is heated by a high-�requency osci.llator by means of
a work cotl, 2, separateS from the reactton chamber by means of a quartz cylinder,
9. '
:'znoraaoBa,v / 2 3 9
C/IapozoaoiaA crec.
s S f
y Z
7
a 0 7
0 y
f c)~
/ArsajoW a~eeb
Key:
Figure 3-20. Construction of Epitaxial Reactors: a- horizontal type;
b--vertical type; c--cylindrical reactor
1. Vapor-gas mixture
The vapor-gas mixture is fed into the reactor either from below, as illustrated
in the drawing, or from above along the substrate holder. In some types of
cylindrical reactors the mixture is fed through a slit inlet perpendicularly to +l�~
axis of the substrate holder.
Substrates are heated by meana of-l) a high-frequency oscillator via a work coil,
2) infrared heating by means of quartz halogen lamps, and 3) resistive heaters.
The heating method plays an important role in choosing the type of reactor. A
high-frequency heater is used most wtdely, since then only the substrate holder
made of highly conductive material together with the wafers is heated, and the
walls of the quartz tube remain cold. Therefore, the reactor's walls do not be-
come overgrown with the reaction products of the vapor-gas mixture.
When outside resistive and lamp-type infrared heaters are used, the substrate
holder together with the wafers and the quartz reactor tube are heated, as in
diffusion Purnaces. Since a lamp--type infrared source has a higher temperature
than an incandescent souxce, the power radiated by it belongs to a shorter wave-
band (1 to 3It), in which quartz glass is practically tranaparent. This results
in less heating of the walls of the reactor tube.
High-frequency and tn#rared heating systems are lesa interial than resistive
heaters and easily make it possible to per#orm ntultilevel heating cycles of short
duration. A11 rhree kinds o# heating are approximately identical With respect to
the unit input oP power.
- 74 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
All three types o!' xeactors are curxently uaed in the induatry. Epitaxy units
with a hoxizontal reactor are simplex in design but are di.stinguished by htgher
consumptions of woxking gases and a greater spresd in the thickness and resisti--
vity of deposited i'i1,ms. Units with vertical aud cylindrical reactoxs are complex
in design but make possible a smaller spread in the thickness and resistivity of
deposited films. Cylindrical reactoxs make possible the highest productivity and
the lowest consumption of working gases.
The reproducibility of the thickness and resistivity of epitaxial layers depends
basically on the design of the reaction chamber and heattng system, as well as on
the reproducibility of gas and temperature parameters, i.e., on the gas distribu-
tion system and the system for controlling the temperature in the reactor.
A typical system for supplying gas to the reactor is described in the following
chapter (sec 4-2.).
The construction of the reaction chamber of a UNES-2P-V unit with a vertical re-
actor is shown in fig 3-21. The bottom flange, 15, is fastened by means of
clamps, 21, to a plate-type base, 1. SEals 16, 13 and 14 ensure airtightness in
~ the bottom half of the reaction chamber, which coiLsists of two coaxial tubes--an
inside quartz tube, 7, and an outside one made of acrylic plastic, 8. Water
circulates between them for the purpose of cooling the quartz tube, which is
heated by radiatian from a pyramid-type substrate holder, 10. The upper flange,
6, makes airtightness possiole in the upper half of the reactor by means of seals
2 and 3. A substrate holder made of graphite, 10, is fastened to a support, 11,
and a centering flange, 19, and is rotated on a shaft sealed by means of gaskets,
17. This entire unit is fastened in tlange 18.
Figure 3-21. Construction of Reaction Chamher of tJPTES-2P-V epitaxy unit
Key: 1. Discharge 2. Water
-75-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE JNLY
A high-frequency induction heater, 9, and the reaction chamher, tightened by
= means of �langes 6 and 12, are fastened to a bracket, 4. The temperature is
measured by means of an optical pyrometer, 5. If the reaction chamber has not
been fastened by means of screw clamps, microswitch 20 blocks the switching on
of the hi gh-frequsncy heating system.
The substrate holder, 10, is in the form of a po],yhedral truncated pyramid.
Wafers are Eastened to it at an angle of 5 to 7 degrees to the vertical axis.
lt is made of graphite coated with a carbide la}er. Uniform heating and equal-
ization oP the concentration of the vapor-gas mixture on the surPace of wafers
are produced by rotating the pyramid. This mixture is fed from above parallel to
the reactor's axis and is removed Prom below.
In a UNES-2P-V unit it is possible to perform the epitaxial growing of n- and
p-type layers, the deposition of films of silicon dioxide, as well as etching
witt hydrogen chloride. It is possible to bring heating of wafers up to 1300 �C.
The cylindrical reactor of the UNES-2P-KA industrial epitaxial growing unit is
shown in f ig 3-22.
Omeod p~aKU~oN-
Nozo zaaa Bada
/7 f IB 19 I
Bod~_ I 1 3 )Badopoa
2 )
-31 16 I Boaa
eo6a
2
0
15 ~ , O0 ,
I ~0
O
4
13
Boda Boda
S
11 ~ I ~ pp
eoaa 7 6 1 leoaa i~l f io B
f f f
9oaopoa I

Boda.
Boda
Key:
Figure 3-22. Construction of Reaction Chamber of UNES-2P-KA Epitaxy Unit
1. Discharge of reaction gas 3. Hydrogen
2. Water
-76-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
e.
The reactor's shape,forming parts are af1at uppex flange, a cylinder, 2, and
a lower flange, 5, Which forq the airttght space of Che xeactox, Ingide the
reactor there is a hollow cylindrical substrate holder, 3, with the~ wafers, 15,
placed on several Ciers. The vapor--gas ntixture is #'ed from below through a pipe
connection, 10, and discharging takes place through three water--cooled pipe con-
nections for discharging the reaction gas. The annular zone between the wa11 0�
the reaction chamber and the substrate, 3, is blown through through two pipe
connections, 11, in the lower flange of the reactox. In addition, the annular
spaces between the Cwo concentric rubber gaskets, 16 and 12, of the upper and
lower flanges, as we11 as the sealing rings, 7, of the rotating shaft, are blown
through with nitrogen through pipe connections 6, 9 and 17 for ensuring the relia-
bility of seals. A quartz bu1b, 14, inside oP which a high-frequency work coil,
13, is placed, is inserted into the inside space of the substrate holder, 3.
The substrate holder is placed on a quartz support, 4, placed on the flat disk
of a water-cooled rotating shaft, 8. The quartz bulb is sealed by means of a
rubber ring-type gasket, 18, and flange 19.
The temperature profile over the height of the substrate holder is equalized by
changing the spacing of the work coil's turns: a local increase in temperature
by compressing the turns, and a reduction by stretching them.
The UNES-2P-KA unit is designed for the large-lot production of silicon epitaxial
structures. It is possible to perform gas etching in it, as well as to produce
coatings of silicon dioxide and nitride. The operating temperature range is from
900 to 1300 �C.
Equipment for Liquid Epitaxy
Liquid epitaxy has assumed an industrial scale in the last five or six years.
The impetus fo r this development was the demand for solid-state 1 s rs, light
displays and other optoelectronic devices based on the use of AII~B~ compounds
and solid solutions based on them. This method makes it possible to produce
heterojunctions. The process occurs with the release of heat, i.e., is exothermic.
As a result, for the purpose of forming an epitaxial film it is necessary to lower
the temperature of the substrate with the solution melt layer on it. The tempera-
ture drop must take place in keeping with a specific law for various substances.
The process can be divided into the following process steps. The elementary
substance--a solid binary or ternary solution--is melted in an appropriate low-
melting substance which is chemically inactive with respect to the solution and
substrate, most often in a metal. Then the solution melt is brought into contact
with the substrate and after the establishment of thermal equilibrium between the
is cooled in keeping wj_th the appropriate law for the purpose of epitaxial deposi-
_ tion of a film.onto the substrate.
For some materials the epitaxial deposition of layers �rom a so7.ution melt is
performed with the existence of a temperature gradient created along the substrate.
Impurities for pxoducing doped epitaxial layers are introduced into the solution
melt either duxing pxeparation of the charge hefore its melting or from the gas
phase after melting. The excess solution melt is mechanically removed from the
substrate af ter deposition of the eoitaxial layer. Then the substrate with the
deposited layer is cooled to room temperature.
-77-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R040500090004-3
FOR OFFICtAL USE ONLY
On the basis of this zechnol.ogical pxoces-s., i.t is po$aib].e Go formu7.ate Xequire-
ments for epitaxia,l equipment:
1. The heating Purnace must be a quick-response one and have a heating temperature
range of 300 to 1100 �C, makin$ it possible tv satisfy requirements for me].ting
a charge of the various materials used for liquid epitaxy. The accuracy of main-
taining the temperature is + 0.75 �C. For untts with a continuous epitaxial
deposition cycle it is necessary to have several temperature zones: a zone for
melting the charge and heating the substrate; a zone for bringing the solution
melt into contact with the substrate and Por holding for a certain time; a zone
with a temperature gradient in the temperature reduction direction for making the
epitaxial deposition process possible; and a zone #or the cooling.of substrates.
2. A mechanism or device for bringing the solution melt into contact with the
substrate must be provided in the reaction chamber.
3. Before deposition onto the substrate, the solution melt must be produced in
a crucible which is chemically inactive with respect to it.
4. The material of the reaction chamber must a].so be chemically inactive with
respect to the substrate and solution melt.
5. Gases which are chemically active with respect to the substrate and solution
melt must not be used when carrying out these processes in the reaction tube.
Therefore, in the reactor either a flow of a purified neutral gas is created or
evacuation is created by means of.vacuum pumps.
The last requirement is satisfied by equipment which is distinguished by the method
of creating the appropriate atmosphere in the reaction chamber--of the open or
closed type. In a unit of the open type a neutral gas, most often mixed with
purified hydrogen, enters the reactor continuously throughout the entire process
and this gas, breaking down oxides on the surface of the solution melt, makes
possible the occurrence of a reduction reaction. In units of the closed type a
vacuum is created in the reactors. In some cases, before contact between the
solution melt and the substrate hydrogen is introduced into a reactor of the cl(,.
- type for the purpose of breaking down the oxide layer formed on the surface of
the melt, and then evacuation is again performed right up to the final cooling.,of
the substrates with the deposited layers.
Reaction cha.mbers are subdivided into the following types with respect to the
method of bringing the solution melt into contact with the surface of the sub-
strate:
A reactor with an inclined holder (fig 3-23, a), in which the solution melt, 2,
with the substrate, 1, placed in tYe holder, 3, which is streamlined by a neutral
(nitrogen or argon) or reducing (hydrogen) gas in a quartz tube, 4, are brought
into contact by changing the tilt of the reactor relative to the rest, 5.
A reactor in the form of a rotating cylinder (fig 3-23, b), in which the solution
melt, 2, is brought into contact with the substrate, 1, by turning a graphite
cylinder, 8, 180 degrees from the poaition shown i,n the drawing, and by lifting
-78-
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
the substxate by meaAs. of ].iftex 7, unttl it makes contact with the surface of
the solution melt.
_-y
~
fa,~ '
Merim/owwNd 4
J i J~~ /0 6 / 2.J4
~ .6
6 S _ ~ � - � 2 1~.
..a) b) c
Key:
Figure 3-23. Reactors for T.iquid Epttaxy: a---employing the rotating
cylinder method [as published]; b--employing the rotating
cylinder method; c--employing the wetting method; d---employing
a container of the cylindrical case type
1. Neutral gas
A reactor of the vertical type (fig 3-23, c), in which substrates, 1, fastened to
a holder, 9, are immersed by means of a moving rod into the solution melt, 2,
which is in a crucible, 10. This type of reactor is designed for working with
solution melts possessing heightened reactivity for the formation of oxides.
When the substrates are inunersed into the melt they rupture the oxide film and
make contact with the melt in bulk. A reactor with a container of the cylindrical case type (fig 3-23, d), in which
a pool, 2, is moved along a holder, 3, with substrates, until the solution -nelt
makes contact with the substrates.
All these reactors are put into a quartz tube, 4, and are furnished with a heater,
6, making possible the required temperature conditions.
There are also other varieties of reactors which differ in the principle of bring-
ing the substrate into contact with the solution melt. Reactors with a container
of the cylindrical case type have become most widespread in industry. It is
possible to cite as an example the continuous induatrial unit for producing epi-
taxial films of GaP.
-79-
FOR OFFICIdL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFFICIAL USE ONLY
The reaction chamber, by virtue of the sectional structure of the
heating furnace, has several temperature regions (Figure 3.24a): the region
of fusion of the Ga solvent from room temperature Tk to the point A; the region
of dissolution and saturation of the melt with polycrystalline GaP (AB); the
region of epitaxial growth of the GaP layer on the monocrystalline substrate of
GaP (BC) and after that, the region for substrate cool-down and the removal
of the melt from its surface (CTk).
The installation operates in the following manner (Figure 3.24b): the cylindri-
cal case type containers move from left to right along the quartz guide rails 8,
which are positioned in quartz tube 9; these containers consist of graphite
holder 2 with two substrates 3 positioned in the grooves and graphite tray 1
with the solution-melt 4, which has the ability to move in the graphite holder.
The containers, in passing through the two temperature regions of the furnace
TkA and AB, at point B lock the light beam of laser 5 on light disk 11 through
slot 10 in holder 2. As a result of this, mechanism 6 matches the tray to the
tiolder so that the melt-solution 4 comes in contact with substrates 3. The
container, in passing through the temperature region (BC), where the GaP
epitaxial f ilm grows on the monocrystalline GaP substrate, moves the tray to
the initial position by means of return mechanism 7, separating the melt 4
from the substrates 3 and removing the solution-melt which has not undergone
a reaction. Then the container is fed into the region for cooling and unloading
the wafers. The epitaxy process is performed in a flow of a mixture of nitrogen
and hydrogen gases, which are cleaned of oxygen and moisture. Tellurium is
used as the hardener to obtain the epitaxial layers of n-type GaP.
Gas screens 3(Figure 3.24c) are used in this installation to prevent the intru-
sion of atmospheric gas into the reaction chamber. Moreover, dust-free boxes 1
- 80 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
are used in it, which assure the high purity of the air environment in which the
loading and unloading of the wafers and charging with the solvent take place.
,4 e(iaso �c)
T I
CI~90'C)
;
�T ~
(1) Hr.np.Bntnuc J4uaceduA Kautiu ~ s 6 7 a/ (a)
~
IF-
-
.
1 2 J 9 ~l~ %0 B 61 (b)
i 1
i6 i2 J 7 J S 6
el (C )
Figure 3.24. The temperature cycle during the growing of epitaxial
GaP films (a); basic schematic of the reactor with the
cylindrical case containers (b); the continuous expo-
sure installation for the epitaxial growth of GaP layers
from the liquid phase (c).
Key: 1. Direction of holder travel.
To iucrease the productivity of the installation, the loading and unloading
process has been mechanized and sutomated by means of devices 2 and 5. The
control units for gas systems 6 and the control and regulation of the tempera-
ture 7 in the reaction chamber 4 are built into the housing of the installation.
To apply multilayer epitaxial films, several trays with the solution-melt are
placed in the container, where these trays are brought in contact with the
substrate in turns for the sequential deposition of the specified layers of
the substances.
3.5. Equipment for the Production of Alloy Junctions
The alloying technique is used at the present time primarily to obtain low
frequency semiconductor devices and ohmic contacts.
The alloy process takes place in three steps:
--Local wetting of the semiconductor surface with the metal;
--Dissolution of the surface layer of the semiconductor in the volume of the
melted metal;
--The formation of the junction layer (the p-n junction or the ohmic contact) as
a consequence of the crystallization during cooling of the semiconductor
dissolved in the melt.
- 81 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
A furnace with a controlled environment and holders in which the geometric match-
ing of the semiconductor substrates to the metal electrodes being fused in at
that point where it is necessary to produce the alloy region are needed to
produce an alloy Junction.
As an example of producing a p-n junc-
tion in germaniuir. chips by means of
melting in indium, we shall consider
the requirements placed on the tempera-
ture profile of a conveyor furnace
(Figure 3.25). To form a flat and even
front edge for the melting-in, it is
necessary to heat the semiconductor--
metal pair and cool them following the
formation of the melt at a definite
rate.
c 2)
Figure 3.25.
Key:
rr zv .'S minutes ~uN
As can be seen from the figure, the metal
Temperature profile of the is melted during the wetting process
process of inelting indium and it flows over the surface of the
into germanium to produce semiconductor in that part of it where
a p-junction. there is to be an alloy Junction. For
1. V= 3�C/min; this reason, the holder with the wafers
2. Region of heating and is kept in the furnace for 1 to 3 min-
utes at a temperature on the order of
wetting; 300� C(100 to 150� C higher than the
3. Dissolution region; melting point). Moreover, the presence
4. Region of slow cooling. of foreign films and especially oxides,
is impermissible for normal wettability
of the surface. For this purpose, th:
melting-in process is carried out in a reducing medium (pure hydrogen), having
beforehand subjected the internal portion of the heating furnace and holder to
a careful cleaning.
After the completion of the wetting step, the temperature is increased sharply
(up to 550� C) to dissolve the surface of the semiconductor with the metal melt.
In this case, because of diffusion at the boundary of these substances, there
is the formation of aJunction layer. The time needed to establish thermodynamic
equilibrium is governed by the dissolution rate of the semiconductor in the
metal and the speed of diffusion of the atoms in the melt, and for the given
case, fluctuates in a range of 5 to 12 minutes. Following this, the system is
slowly cooled initially with a temperature gradient on the order of 3� C/min
and thereafter at 8� C/min to recrystallize the formed alloy and to form the
p-n junction between the original n-germanium and the recrystallizing p-german-
ium. In the case of rapid cooling, because of the difference in the thermal
expansion coefficients of germanium and indium, the melting-in region can
develop cracks.
Thus, for a periodic exposure installation, the furnace should have a programmer
for the heating and cooling of the holders with the substrates, while in
-82-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
continuous exposure installations, the furnace must be made as a sectional de-
sign, where the temperature is maintained in accordance with the specified
alloying temperature cycle. To obtain germanium.semiconductor devices, a
maximum temperature on the order of 700� is needed, while for silicon devices,
the temperature is on the order of 1,000� C.
The thermal installations for the production of alloy semicon3uctor devices can
be classified according to the following parameters:
--According to the working atmosphere - as vacuum and gas installations;
--According to the type of heaters - with direct heating, in which the working
channel of the furnace is at the same time the heater, and with indirect
heating, in which silite [electrical insulating material] rods, wire and
other heaters heat the reactor tube;
--According to the operating principle - as periodic exposure furnaces, in
which the loading, the process and the unloading are performed after each
input into the working channel, and as continuous devices, for example, con-
veyor installations;
--According to the working temperature - as low temperature (up to 700� C) and
high temperature (up to 1,000� C).
Figure 3.26. Conveyor furnace for the production of germanium alloy
- semiconductor devices.
To obtain alloy devices, primarily low power diode matrices, electron beam
installations are used in addition to thermal installations, in which the
instantaneous local heating and melting-in are accomplished by a focused elec-
tron beam. When this method is used, the depth of the melting-in region
depends on the accelerated electron beam energy, while the geometric dimensions
of ttie p-n junction are governed by the diameter of the electron beam. However.
tlie tectinique of alloying in conveyor furnaces is used more often in industry,
which provides for high output and economic efficiency by virtue of the simpli-
city of the structural design of the installations and their low cost.
We shall consider the structural design of the SK 11/16�10 = 6 conveyor furnace,
intended for producing p-n junctions by means of alloying (Figure 3.26). The
installation consists of a heating chamber 5 and muffle tube chamber 4, the gas
feed system for nitrogen and hydrogen 3, the gas screen devices 2 as well as
the charging area 1 and the unloading area 8 as well as the refrigerator 7.
-83-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
v S 6 7 J t 6

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
F'OR OF'FICIAL USE ONLY
All of the assemblies of the installation are mounted in a welded metal housing
m �
rl +J
N
41
7 7
0
r~1 r-4
� w M"~ ~-i
r-i
}.I
P4 W
V /
CC
to 0
1.

~4
u V+
o u
~4
,H
4.+ ~ 4.4
ci
:J
11
L1 cO
rl cO CI
~
c:
al
00 ~
~
=
r
l
c
.
E
N c0 N
~
W~
1+
-
u
~
'b
~7 H
ti D[,
-
O
=
ri
-
i-i

-
~
c
a
,-4
=
E
c
y
.
m

~
>
w
~
3
u
O
O
1-1
0
c
9
44
0
>
F'
r
-1
f+

~
~
~
~
~
cC
0
c
9

~
M ~
o
~
~

~
~
~ ~
~
~
�
44
O
Cl N D,
r-1 r-I
�
~
'
rl
;
M
u
O
~
D'~
r-1
1

�I
y
c
u
u
I r-I
c0 co co
~

c0
7 D U
U�~
O
=
> >

~O
a)
W W 14
41 rl
M
~
JJ
tb
-ri -H
41 41 Gl
00 C'+
00
Y
U
r4 ~i
:1 0 r-I
cd c0
Gl
m
~
~
t/~ V)
U U W

"
~
N
3
do4
j 4 4i
r4 6
u
T
ea
tl'1
J
~
V3
N
$4
m.
~
T
~
1
rl
y.
W
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00854R000500090004-3
FOR OFFICIAL USE ONLY
The most widespread precipitation technique is a thermoct.emical reaction, for
example, the oxidation of the monosilane SiH4. Doped and undoped layers of
oxides can be applied using this method at low temperatures (200 to 250� C).
This technique makes it possible to attain high rates of deposition: up to
several thousands of angstroms per minute, a figure which is 50 to 100 times
greater than the precipitation rate of an oxide using the pyrolysis of tetra-
ethoxysilane at 600 to 700� C as wpll as improve the quality of the oxide and
the adhesion of the oxide layer to the substrate.
The method is based on the monitored oxidation of monosilane with oxygen in
dilute mixtures with inert gases, in accordance with the reaction:
SiH4 + 02 S102 + H20
and provides for the possibility of doping silicon dioxide with boron, phosphorus,
arsenic and other impurities during the precipitation process.
The following processes have been developed at the_�present time to obtain passi-
vating coatings:
--Precipitation from a gaseous phase at normal pressure;
--Precipitation from a gaseous phase at low pressure;
--Precipitation from a gaseous phase in a plasma.
Types of reactors in installations for chemical precipitation from a gaseous
phase are described in �3.4.
An example of an installation with a reactor operating at normal pressure and
at low temperature (less than 700� C) is the "Oksin-3" installation. The
structiral design of the reactor of this installation which is intended for
the precipitation of silicon dioxide films, both phosphorus doped and undoped
films, is shown in Figure 4.16. Three reactors, arranged one on top of the
other, are used in the installation to increase its productivity. Square
cross-section reactor tube 4, which is made from stainless steel or quartz,
has the flanged end secured in head 10 of plate 9 and is sealed with washers
8 and 11 made of silicone rubber. Diffuser 7 with twin gas injection to feed
the working gases from the gas distribution 7 into the working region through
connecting tube 12 is positioned in head 10. The open end of the tube is
secured in the clamping unit by means of moving clamps 6 with adhesive washers
5 made of silicone rubber.
The clamping device is fastened to the fan ventilation housing 2 0� the mounting
stand. The spend gases are exhausted into the exhaust ventilation through the
open end of the tube.
The electric heater 3 is made in the form of tubes 16, which are built into
graphite plate 1; along with thermocouple 15. The quartz plate 13 with the
substrates 14, on which the dielectric layers are applied, i.s placed on the
graphite plate. r -102-
F'OR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
Figure 4.16. The structural design of the reactor of the "Oksin-3"
low temperature silicon oxide precipitation installation
with an internal resistive heater.
NZ SiHy
~ (1) ~~(1)
.no (2) xv (2)
/ HiLl (3) xtu(3) .
PPr (4)
Reactor
' (7)
PcsKmop
(5) 6 _
Kn ~1117 (8)
r( 9 )
�tu Km ( 3 )
PP/' 10(i)
� /ro Ko ( 2 )
m PHj Dy
Figure 4.17. Pneumatic configuration of
the gas distribution sys-
tem for the "Oksin-3" in-
stallation.
Key: 1. Filter;
2. Check valve;
3. Ball valve;
4. Gas flow rate regulator;
5. Restrictor;
6. Direct reading flow
meters;
7, 8. Pneumatic valves;
9. Gas flow rate regulator.
The temperature in the reactor does
not exceed 600� C.
The gas distribution system for the
input and regulation of the feed of
the working gases into the reactor,
in particular, silane, oxygen, nitrogen
and the doping gas phosphine or
diborane, is shown in Figure 4.17. A
filter F is installed at the system
input to scrub the gases. In the
figure, one line is used to indicate
each gas. Where necessary, one can
increase the number of regulated
channels to feed in appropriate addi-
tional gases. The gas distribution
components are connected to each other
and to.the main delivery lines through
pipe filters made of stainless steel.
Gases can be fed into the installation
both from tanks and from a centralized
network. A specific feature of the
operation of the gas system of the
installation is the capability of the
simultaneous operation of all reactors.
The gas system consists of the nitro-
gen, oxygen, phosphine and monosilane
feed lines as well as the disposal line.
We shall consider the operation of the
gas distribution system. Nitrogen
is fed in through the filter, the check
- 103 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094004-3
FOR OFFICIAL USE ONLY
valve K0, which prevents the reverse flow of the gas, and tha ball valve KSh,
to the direct reading flow meters R, which serve for monitoring and regulating
the nitrogen consumption when diluting the monosilane and phosphine with it.
A choke Dr is inserted in parallel with the ball valve, where the choke is
intended for providing for the requisite gas rate of flow with the constant
f lushing of the reaction chamber.
The phosphine feed line also starts froiu filter F. Then follow the check valve
� KO to prevent the back flow of the gas, the [ball] valve KSh, the gas flow rate
regulator RRG with a digital display, a pneumatic valve (non-return) KP, which
allows or cuts off the gas access to the system. The oxygen and silane lines
are similar. The gas mixture is fed via two separate lines directly into the
reactor.
The installations for film precipitation at low pressures have a number of
advantages over the preceding type. At low pressure (1 to 66 Pa), the free
path length of the moleciiles of the regulating gas in the chamber is increased.
This makes it possib le, by placing the wafers vertically and close to each other,
to increase the productivity of the installation (from 80 to 150 wafers/cycle
when depositing polysilicon with a thickness or 50,000 X or silicon nitride with
a thickness of 1,000 X) and does not require a gas vehicle. The substrate holder
is heated simultaneously with the heating of rhP quartz tube, according to the
type of diffusion furnace. Dielectric films p:�oduced in such systems are dis-
tinguished by their high homogeneity and large coefficient of coverage of the
relief steps of the substrates. The precipitate adheres to the hot wall of
the tube, while the low pressure of the gas does not cause the particles to
circulate in the tube. For this reason, the films are distinguished by a
minimal number of defects (less than two pores in a wafer with a diameter of
75 mm). However, b ecause of the vertical arrangement of the wafers, the gas
flow to the substrates is encumbered and the film deposition rate is reduced.
I i u r
; qy.? /f Q
r;,, zpyJ n a
.7/.' (I L/771L N
(A)
l( NQCOC.q
-im.
To pump
Figure 4.18. Structural design of the reactor of an installation
for chemical precipitation from a gas phase at low
pressure.
Key: A. Wafer loading and unloading.
A schematic of a reactor with hot walls, which operates at low pressure, is
stiown in Figure 4.18.
- 104 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
~ Beisvd Gas Output
za3�

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
The reaction chamber is a circular quartz tuoe 2, heated by the three zone
resistance heater 3. The tube is secured at the two ends by the flanges 5.
The substrates are inserted at a spacing of 3 to 5 mm from each other in sub-
strate holder 4, which is set in the reaction chamber through the left flange
- 5, which serves for loading and unloading the wa�ers, and then the cover is
closed. The chambsr is exhausted from the opposite end of the tube by a mechan-
ical vacuum pump. There is an opening in the left flange for the admission of
the reaction gas, the flow rate of which is regulated by pressure transducer 1.
A comparatively n--w method of film precipitation from a gas phase is the plasma
chemical technique. A glow discharge plasma is usually employed. The chemical
activity of the reaction gasea increases in a glow discharge, as a result of
which, for example, silicon nitride films in the reaction of silane with
ammonia can precipitate on substrates which are heated up to 300 to 500 � C,
instead of 900� C with the high temperature interaction of the gases indicated
above.
x !a~ryyuNC.uy
.YQGOGf/
(A)
Figure 4.19. Structural design of the
reactor of a plasma
chemic;al r.recipitation
installation.
Key: A. To the vacuum pump.
One of the major factors which has an
impact on the uniformity of film thickness
is the homogeneity of the plasma density.
High frequency capacitor type induction
units are used for these purposes, where
the electrodes in the forms of discs are
arranged parallel to each other. One
of them (the lower one) is the substrate
holder.
The structural design of a capacitor type reactor for the plasma chemical
precipitation of silicon nitride films is shown in Figure 4.19. The working
chamber is fabricated from stainless steel. Substrate holder 4 with the plates
iS ofie of the high frequency electrodes and is fastened to shaft 2, through
which the working gas feed system passes. The substrate holder is rotated by
magnetic drive 3 to provide for uniform precipitation of the films on the wafers.
The substrate holder can be heated from heaters 1, which are located outside
the chamber, up to a temperarure of 200 to 300� C, so as to assure satisfactory
adhesion of the films in the aubstrate as well as their density. The pressure
in the chamber is maintained at a level of 26.6 Pa, which assures the stability
of the glow discharge. The nonuniformity in the film thickness runsaup to 5
to 7% and the rate of growth of silicon nitride films is 300 to 400 A/min at
a power dissipation of 0.5 KW. The silicon nitride films obtained in these
-105-
FOR OFFICIAL USE ONLY
High frequency radiators are used as the
plasma excitation sources. The gas pres-
sure, high frequency radiation po-wer,
plasma density distribution over the
substrate holder, composition of the gases
and the temperature of the substrate all
influence the uniformity of the thickness
and composition of the films.
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
reactors at low substrate temperatures are used for the passivation and protec-
tion of semiconductor devices. Silicon, silicon dioxide and other films are
pxoduced by the technique of plasma chemical precipitation from the gas phase.
-106-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500090004-3
CHAPTER FI'EYE EQUIPMENT FOR PHOTOLITHOGRAPHY PROC~ ESSES
Phololithography is one of the major steps in semiconductor production, which
governs the quality of the entire technological process as a whole.
Photolithography includes the following main operations:
--The ssrface preparation of the semicanductor wafer;
--The formation of the photoresist layer;
--The formation of the photoresist relief;
--The formation of the reli_ef in the oxide or metal;
--The removal of the photoresist layer.
In accordance with the steps in the photolithography production process, the
equipment is classified according to function:
--For processing the surface of a wafer;
--For applying and heat treating the photoresist;
--For matching and exposing, developing and heat treating;
--For the etching operations.
The processes and techniques of photolithography are in the stage of continuous
refinement; this is also related tc: the diversity of the equipment used for
the same processes. For example, the application of photoresist is accomplished
by means of centrifuging and atomization methods; developing uses immersion
and pulverizatioh methods; heat treatment uses convection and infrared heating,
as well as microwave energy in a vacuum and at elevated pressure (thermal
compression technique).
Besides the direct improvement of the production processes, the level of auto-
mation of photolithographic processes has a direct influence on the improvement
in the quality of processing the wafers and boosting productivity. Problems
of stabilizing the production process modes, transporting the wafers and elimin-
ating operator contact with the wafers are the ones being primarily solved here.
The first stage in the automation of the photolithographic prores. was the
construction of a series of automatic units which perform the individual pro-
duction process operations by the group method in accordance with a specified
program:
--The photoresist developer unit of the Kulicke and Soffa Industries Company
(U.S.); '
--The five-position installation for applying photoresist, t4acronetics Model 1201
(U.S.): the one and two position units for photoresist application of Plat-
General (U.S.).
- 107 -
FOR OFFiCiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
FOR OFFICIAL USE ONLY
However, these installations have a serious drawback: the loading, unloading and
transpcrting of the wafers from the working positions to other operations are
acr_omplished manually. Domestic equipment of similar type, with which the
"Taran" and "Korund" lines are equipped, also have the same drawbacks. The first
attemvt to design a domestic computer controlled production complex was the
APL automated flow line. The photolithographic operations are carried out on
the line using the group method. The photoresist is applied by means of cen-
trifuging from a group drip pan simultaneously for ten wafers, which are located
in the common spindle of the centrifuge (Figure 5.1). The developing is done
by atomization of the developer. The heat treatment is accomplished by the
thermal compression technique, something which has a positive effect: the
heat treatment operation is eliminated and the removal of the photoresist is
- facilitated.
However, all of the high prodiictivity equipment mentioned above, installations
with group processing of the c,*afers, do not meet the major requirement of
modern technology: absolute reproducibility of the production process modes
Eor each wafer in a batch.
Recent years in the field of semiconductor machine building have been character-
ized by the transition from installations which perform individual production
process operations to the development of automated lines and complexes.
u
F
.3
R-A

3
~A y
c=o
0
Figure 5.1. A device for applying
photoresist coatings by
means of group centrifuging.
Key: 1. Base;
2. Receiver;
3. Vacuum suction fitting;
4. Wafer;
5. Cover.
Yet another important trend is observed
as concerns photolithography lines: a
transition from group processing of the
wafers to individual processing, which
best meets the main requirement of the
production process: its reproducibility.
In step with the refinement of the tech-
nology using individual wafer processing,
and as a result of solving questions of
transporting them to the processing
position, it became possible to move on
to the next step in the design of auto-
mated photolithography lines. In 1971,
the Japanese company Toshiba designed
an automated photolithography line in
which the entire production process is
accomplished using a single cassette
holder. The American company Industrial
Modular Systems Co. [39] developed a
system for transporting the wafers on an
air cushion, and automatic units were
designed using this principle for the
application and developing of the photo-
resist, which had a high productivity
and devices for automatically loading
-108-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
and unloading the holders. The first attempts to design automated photolitho-
graphy lizes in domestic industry were *'i: UNT-80 and AFS-100 lines, where the
principle of transporting the wafers on an air cushion was employed and a single
cagseti.e with a capacity of 30 wafers was used. In the process of operatiing
these lines, a number of defir.iencies were ascertained, both structural and
production deficiencies, the main one of which was the insufficient reliability
of the equipment.
Photolithographic equipment sets and lines, which are based on the following
operational principles, most completely satisfy the requirements of the problems
posed:
--Individual treatment of the wafers using the "holder to holder" technique;
--Automatic feed of the wafer from the holder to the working position and its
reloading into a receiving cassette holder, something which precludes contam-
ination from hands and the damaging of the wafeis;
--The control of the sequence and duration of the production process operations
by means of a control unit or microcomputer;
--Operation in accordance with a specified program which assures absolute
_ reproducibil.ity of the production process modes, precluding the influence
of subjective factors on the production process.
Figure 5.2. A universal loading and unloading mechanism.
-109-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
FOR OF'FICIAL USE ONLY
Key to Figure 5.2:
1. Guides;
2. Push rod;
3. Pneumatic cylinder;
4. Electric motor;
5. Coupling;
6. Pusher;
7. Microswitch;
8. Pulley;
9. Belt;
10. Guide;
11. Motion screw;
12. Pulley;
13. Microswitch;
14. Pneumatic cylinder;
15. Support pogt;
16. Microswitch;
17. Pull bar;
18. Pneumatic tray;
19. Photocell;
20. Guides;
21. Microswitch;
22. Rod.
An important organizational component in the functioning of an automated line
is the set of cassette holders which make it possible to create universal loading
and unloading mechanisms and implement the organizational principle of the
production: the operator works only with the cassette holder. The standard
structural design of such a mechanism from the "Lada-125" line is shown in
Figure 5.2. The photo?_itnography lines which have been developed by various
companies make it possible to not only curtail the expenditures for manual labor,
but also to optimally limit the intervention of the operator in the production
process.
Macronetics and the III Companies put together a line from modular units for
hydromechanical washing, application of the photoresist, developing and IR
heat treatment. The Cobilt Company produces the Autofab-IV photolithography
line, in which there is rigid coupling between the installations. A cassette
holder with a multi-shelf configuration with'the starting wafers is installed
at the input to this line, and the holder is removed at the output with the
topological figure already on the plate.
Thus, there are two trends in the construction of photolithographic lines:
--The joining of modular units to iudividual loading and unloading posts for
wafers into holders, which can be rigidly joined in pairs, structurally and
in terms of the power supply [lines of the III Company and the Class-1000 line
of the Macronetics Company (U.S.), and the "Lada-elektronika" and "Lada-125"
(USSR) ] ;
--A rigid line in which the wafer loading and unloading posts are located at the
beginning and the end (the "Autofab-IV" line of the Cobilt Company).
The ilniplane 4 000 line of the Kasper Company (U.S.) occupies a special position
among those treated here; this line includes centrifuge cleaners, developers
and furnaces. The Uniplane 4 000 line is made from modules, each of which can
operate independently of the other under the control of its own microprocessor
or as part of an overall comprehensive system with complete coupling between
the modules in both the forward and reverse directions [18]. The lines
- 110 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
cfescrlbed here are distinguished by the transport systems for the wafers, which
can be conditionally broken down into three types:
--Wafer transport on an air cushion (Figure 5.3);
--Combination transport (on an air cushion and using a transport carriage)
(Figure 5.4);
--Wafer transport on polyu~.�e+:hane belts in a strictly horizontal plane with a
smooth change in the carriage travel apeed at the outset of the motion and
when stopping (Figure 5.5).
. o T
~

~
Figure 5.3. The wafer transport system
using an air cushion.
Key: 1. Wafer;
2. Carrying flow.
Key: 1. Sloped tray;
2. Wafer;
3. Cleaner;
The latter principle of moving the wafers 4. Holder;
being processed is the most expedient orLe, 5. Loader;
since it precludes shock contact of the 6. Drive;
end face of the wafer and the loading a~.id 7. Centrifuge cartridge.
manipulating devices: the edge of the
cassette holder, the carriage stops, clamps, guides, etc. Such microshocks have
been observed in the first two transport methods and have led to damage to wafers,
the formation of silicon crumbs and dust, and consequently, to the contamination
of the photoresist film, and the fouling and failure of moving mechanisms.
Figure 5.5. The transport system in the "Lada-125" line.
- 111 -
FOR OFFICIAL USE ONLY
.�i.Jl~~[lLll6/(l \ ~ \ .
\ \ \
numoK i
-
Air Flow
Figure 5.4. The carriage and unloading
mechanism.
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
i 2 3 y 5

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090044-3
FOR OFFICIAL USE ONLY
Key to Figure 5.5: 1. Wafer loader;
2. Transport carriage;
3. Carriage drive;
4. Wafer throw-off ejector;
5. Transporter;
6. Upper plate;
7. Support stand;
8. Lower plate.
5.1. Equipment for Preparing the Surface of Wafers
The quality of a photolithograpbic process is governed in many respects by the
preparation of the wafer suriace, and for this reason, cleaning is one of the
most important operations in semiconductor technology, on the effectiveness of
which the electrophysical properties and percentage output of good devices
depend.
Surface contamination can be broken down into physical-chemical and mechanical.
Physical-chemical contaminants are ionic or neutral impurities adsorbed on the
stirface which form monolaycrs and influence primarily the parameters and reliabi-
lity of the devices. Primarily chemical cleaning methods are used to remove
them where these methods are based on the desorption of the impurities when the
wafers are treated in solutions, gaseous media and in a plasma [5]. The par-
ticles take the form of clusters of the material with dimensions of 0.1 um and
greater.
The complete removal of conta.ninants is one of the difficult problems in the
processing of wafers, for the solution of which primarily physical cleaning
methods are employed, which in turn include ultrasonic and mechanical treatmetit.
Of the physical cleaning methods, hydromechanical washing is being successfully
used of late, which is coming to replace traditional techniques: polishing
with cambric fabric and washing with brushes. The function of a hydromechanical
washing installation is to remove mechanical formations: particles of silicon,
quartz, dust, etc. from the surface of the wafers.
To assure washing effectiveness, it is necessary to use a fluid with a high
degree of purity and to deionize the water, which is filtered through filters
with pores of 0.2 um or less, as the washing medium; the washing is accomplished
directly prior to the process which is sensitive to contamination; hydromech-
anical washing should follow chemical cleaning (in the case where two types of
cleaning are combined), since hydromechanical cleaning makes it possible to
eliminate those contaminants for which chemical cleaning is not effective. The
washing quality and the duration of the production process cycle are governed
not only by the reagents used, but also by the material and the structure of
the brushec. The brush material should meet the following requirements: it
should not change its initial properties in water; it should wash the wafers in
accordance with the production process requirements; it should not introduce
additional contaminants and defects which have an impact on the quality rf the
devices; and it should not permit mechanical damage to the wafers being processed.
- 112 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
Individual treatment of the wafers with a
rotating brush, where the wafers are lo-
cated in a centrifuge cartridge, is employed
in the hydtomechanical wafer cleaning in-
stallations of the "Lada-elektronika" and
"Lada-125" (Figure 5.6). The ejection of
wafers from the holders onto the trans-
porter, the transporting and placing of
a wafer in the working position, the
treatment of a wafer, the removal of a
treated wafer and its output - all of these
operations are performQd automatically in
a standardized unit for moving wafers,
which is the basis for automatic equipment
for the hydromechanical cleaning, applica-
tion and developing of the photoresist.
Between cleaning cycles, a brush is flushed
Fig~ -125" With deionized water, washing solution or
hydromechanical washer, another fluid which is used for treating
the wafers. A brush which is shiCted witli
respect to the center of the wafer and
which rotates counter to the wafer motion is used in washing installations with
brushes of the 1100 SD series of the So.litec Company. A stream of liquid con-
stantly fed from the center of the brush flushes away contaminant particles.
The washing of wafers with cylindrical and conical brushes is shown in Figure
5.7. Along with cleaning wafers with a brush, a number of companies, Macronetics,
. ~
a) (a)
r
. .
~
6) (b)
Figure 5.7. Schematics showing wafer washing.
' Key: a. With a cylindrical brush;
b. With conical brushes.
Cobilt and Kasper, use jet cleaning of wafers. This method is espeeially
eCfective when removing contaminants from etched channels, where a brush does
not reach. The spray cleaning system consists of a well protected atomizing
at[achment made of tungsten carbide and a stainless steel pump, which delivers
a tiigh pressure. The cleaning solution is filtered, and then fed to the
rotating surface of the wafer as a pulsed jet stream at a pressure 2.75 � 105
-113-
FOR OFF[CIAi. USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONI.Y
2.75 � 106 P2, depending on the adjustment.
During the cleaning cycle, the chamber is
hermetically sealed for operator safety
(Figure 5.8). In the Kasper hydromechan-
ical cleaner, which is incorporated in
the uniplane line, a combination of wash-
ing with brushes and atomization of a
washing solution under pressure is em-
ployed.
5.2. Equipment for ProdLCing a Photosen-
sitive Layer
The production of a photoresist layer
is the initial operation of the photo-
lithographic cycle itself, in which the
quality of the photolithographic process
of a hole is established. Tlie following
Figure 5.8. The automated Macronetics major requirements are placed on it:
(U.S.) jet wafer washer. high adhesion of the photoresist to the
surface of the wafer, uniformity of the
pliotoresist film thickness over the wafer and reproducibility of the thickness
from wafer to wafer, a minimal number of puncture holes and the absence of flows
of the photoresist to the back side of the wafer.
Figure 5.9. Schematic of the.unit for applying a photoresist by
means of atomization.
Key: 1. Device for loading the semiconductor wafers;
Tank for the photoresist;
3. Chamber for the application of the photoresist;
4. Net conveyor;
5. Wafer washer;
6. Pneumatic cabinet [sic];
7. Infrared furnace;
8. Infrared furnace conveyor;
9. Receiver for the processed wafers.
- 114 -
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
1 7. J v S 6 7 B 9

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
14odern equipment for producing a photosensitive layer, just as for hydromechanical
cleaning and developing, is based on the modular equipment principle and incor-
porates the module for applying the photosensitive layer and module I for the
heat treatment of the layer. The existing methods of applying the photoresiet
include: immersion, rolling, atomization and centrifuging; the latter two
techniques are the most useful in IC technology.
The mociel 265N unit of the In-Line Technology (U.S.) Company can serve as an
example of a unit for the application of a photose:.sitive layer by means of
atomization. The unit makes it poseible to produce pYaotoresist layers from 0.5
to 2 um thick with a uniformity of +6%. The unit is equipped with an automatic
loader and unloader. The loading and unloading positians are protected by a
plexiglass hood, under which a constant flow of air is provided which is filtered
through 0.3 um pores. The atomizer executes a reciprocating motion over the
transporter with the wafers at a v2riable frequency of up to 60 motions/min
[19] (Figure 5.9).
Atomization is the most universal technique for producing a photosensitive layer
and applying it to a wafer in the form of a finely dispersed aerosol. The
photoresist is broken up into small droplets by a gas flow, which flows around
the jet as it exits the nozzle of the in3ector (Figure 5.10).
Pacr?-pumo.
'es r.;Cn.
_:D:J�y
(a)
1
Figure 5.10. In3ector for the atomization
of photoresist.
Key: 1. Pneumatic valve;
2. Feed hole for the atow-
ized jet of the photo-
resist;
3. Holes which shape the
photoresist flare with
compressed air;
4. Filter.
a. Solvent for flushing the
injectors;
b. Photoresist;
c. Compressed air.
Ttie merits of the technique are the capability of producing coatings in a large
range of thicknesses with rather good reproducibility and a slight scatter in
the thickness, as well as the capability of applying the photoresist to pro-
fi.led surfaces. However, ttie most widespread method of applying photoresist,
as before, remains centrifuging.
During centrifuging, the boundary layer adjacent to the substrate is produced
by means of the equalization of the centrifugal and cohesion forces. With
a certain approximation, the layer thickness is governed by the viscosity of
the photoresist, so that:
~ - 115 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONI.Y
1
d = k~ v-/-w
(5.1)
wliere d is the thickness of the photoresist layer; k is a coefficient which takes
into account the concentration of the photoresist; v is the viscosity; and W
is the angular rotational speed of the centrifuge.
The centrifuge run-up time has an influence on the uniformity of the photoresist
layer [5]. To reduce this influence, it is necessary that:
trun < 12/w (5.2)
For the most useful centrifuge speeds, the run-up time is trun < O.J. sec [20].
Tite requirements placed on the production process equipment are deter.mined f rom
t.}ie requirements placed on the quality of the photosensitive layer:
--The drying of the wafers with nitrogen or with dried and cleaned air (with
a dew point of -65� C) prior to the application of the photoresist;
--A centrifuging speed stability in a working range of 500 to 6,000 r.p.m. (a
permissible instability of +5%);
--riinimal and fixed run-up time of the centrifuge (0.1 to 0.15 sec);
--Constancy of the dosage of the photoresist;
--Stnbil.ity of the centriEuging time.
These requirements are made more stringent for equipment which is intended for
processing large diameter wafers, for which the production of a uniform photo-
resist coating is a problem because of the high linear speed of tlie edges of
the wafers when they are centrifuged at the specified angular speed [21].
Figure 5.11. The "Lada-125" automated
unit for the application
of photoresist.
The modern photoresist application
equipment of Macronetics, III, In-Line
Technology, "Lada-elektronika" and "Lad:i-
125", though differing in the system
for wafer transport, tre number of tr.a('
and the diame[er of the wafers which cati
be treated, execute the processing cycle
using a common principle: the automatic
output feed of the wafers from the
cassette holders; the automatic trans-
portation of a wafer to the processing
position (centrifufie platform); nitrogen
flushing of a waCer; apportionecl feed
of the photoresist to a wafer; centri-
fuging a wafer at a specified speed;
and automatic transport to tlie next pro-
duction process operation. The sequence
of the operations, the time for their
execution and the centrifuging speed
are specified on the control panel of
- 116 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE'ONLY
LhL' auCumatic unit. The automatic unit for the photoresist application which is
incorporated in the "Lada-125" line is intended for processing wafers with
diameters of 75, 100 and 125 mm (Figure 5.11). The readjustment of the automated
unit from one diameter ta another reduces to replacing the cutter which provides
for centering the wafers on the centrifuge platform, since the loader is adapted
for operation with any of the three standard dimensions of the holders and
does not require readjustment, while the guides of the transport carriage move
to fit any size depending on the wafer diameter.
The control panel has three buttons in all.for simplicity in operating the auto-
mated unit ("stop", "start", "return"). Changing processing modes is accomplished
at the control console, which is coveri:d with a panel and is opened only when
Figure 5.12. Infrared treatment unit (German Democratic Republic).
setting up the automated units. The production process operations are carried
out in accordance with the program set on the control console or from a computer.
An important feature of the line is the presence of an upper annular exhaust
at the working positions oi the automatic units (bath-centrifuge), which pre-
vents the intrusion of spray and vapors of the production process media to the
wa�er during the processing, as well as their contaminati.on of the mechanisms
of: ttie automatic units. All of the equipment units in the line automats are
made from standardized modules and differ only in the use of a particular
assembly which determines their production process assignment. For example,
by hanging a bracket on a unit with a brush drive, we have a hydromechanical
washer; by hanging a bracket with a dropper, we make an applicator; by hanging
a bracket with injectors, we obtain a developer. The control units are the
same for any of the three automated units and differ only in the interchangeable
panel for the switching of the production process modes for treating the wafers.
Optimum productivity is achieved through the funetioning of one or two trac::s.
-117-
FOR OFFtCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ON1.Y
Hcat treatment operation I, which completes the process of producing a photo-
sensitive layer, has the purpose of removing volative components from the film.
To provide the masking properties, it is necessary that this operation does not
lead to a weakening or to point breaks in the photoresist layer, which can
occur during rapid evaporation of the solvent [5].
Infrared heating is used for the heat treatment in modern photolithography lines.
Infrared lamps (Figure 5.12) [the unit made by the UEB Elektromat Company (GDR)]
and dark infrared radiators are used as the heaters.
Figure 5.13. The "Lada-125" infrared
conveyor furnace.
The domestic heat treatment units incor-
porated in the "Lada-elektronika" and "Lada-125" lines are equipped with Figure 5.14. The automated "T.ada-125"
"dark" IR radiation sources in the photosensitive layer
form of a thin current: conducting film developer.
applied to a sheet of quartz .glass.
The heat treatment in IR conveyor fur-
naces of the "Lada" type (Figure 5.13) is accomplished in the relatively short
time of 2.5 to 5 minutes with continuous nitrogen flushing. The duration of -
the heat treatment is set by the travel speed of the conveyor. The temperatut-(
is maintained automatically in the furnace with an ultimate deviation in the
zones of the heaters of +5� C, which is permissible even for heat treatment II
of the photoresist, which requires greater precision in the maintenance of the
temperature. Infrared conveyor heat treatement installations, ,just as all
automatic "Lada-125" lines are made in a two track variant. Each track operates
independently of the ather. While in the IR conveyor heat treatment unit incor-
porated in the "Lada-elektronika" line the shutdown of the conveyor curtain
means the shutdown of the entire module, in the "Lada-125", with the shutdown
of one track of the furnace, the other can continue to operate.
5.3. Equipment for Producing Relief in a Photosensitive Layer
Developing a photosensitive layer is a process on which the precision of the
reproduction of the geometric dimensions of the topological elements depends.
- 118 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
In the developing process, because of the different rates of solution of the
exposed and unexposed portions of the photoresist film, a re]'-f image of the
topology is produced in the developer [20]. The technique of atomizing (or
pulverizing) the developing solution is the primary one in the modern equipment
of both the leading foreign companies (Macronetics, GCA, III, Kasper) and
domestic industry ("Lada-elektronika", "Lada-125") (Figure 5.14).
(A) n,79raQoa14uri ' ~ .?Developer
pacmCap ! /l,vn,r!Umenb
(B) ~�ir,iuwa 3o~dyafa~vmJ
,~eCmBopll
(C) wa3dy.s N
~a ~onil ~ _
3
4
~
Figure 5.15. The pulverization wafer
treatment technique.
Key: 1.
2.
3.
4.
5.
6.
Washing in,jector;
'Developing injector;
Nozzle;
Wafer;
Centrifuge holder;
Centrifuge.
The pulverization method (Figure 5.15) is
advantageously distinguished from the
obsolete method of immersion developing
in that it makes it possible to speed up
and automate the developing process. The
developing cycle in the "Lada-125" auto-
mated line consists in the sequential
execution of.the following operations:
developing; first washing; second washing;
and drying.
The time for the execution of each opera-
tion is adjusted in a range of from 0
to 99 seconds in discre.r_e steps of 1
second.
Provisions are made for operating the
automated units in three modes: automatic,
semiautomatic and manual.
5.4. Pattern Matching and Exposure Equip-
ment
A. Washing solution I;
B. Washing solution II; In the execution of the photolithography
C. Air (nitrogen). process, operations of transferring the
image of the IC components and semicon-
ductor device from the .photographic template to the wafer, coated with the
photoresist film, and the precise matching of the image of the IC components on
the photographic template to the image on the wafer are of great importance.
The main characteristics of semiconductor devices and integrated circuits, and
in the final analysis, the yield of good devices, depend on the quality of
the performance of the pattern matching and exposure operations.
Several methods exist for transferring the photographic template image to the
wafer.
The contact technique: the photographic template, after being matched, is
brought in contact with the wafer, after which the photoresist is exposed witlt
ultraviolet rays through transparent portions of the figure on the photographic
template. The exposure quality depends in many respects on how complete a
contact is made between the photographic template the wafer, and how precisely
the images of the photographic template and the wafer are matched up.
- 119 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
Photolithography with a constant gap between the photographic template and the
silicon wafer: this method is similar to the contact method, but following the
matching of the images of the photographic template and the wafer, a gap of
from 5 to 20 um is maintained between them which prevents damage to the photo-
graphic template.
The projection technique: the photographic template image is projected onto
the wafer through a special high resolution objective lens.
The electron lithography process for the generation of IC topology as a result
of the nonthermal action of an electron beam on the resist [sic].
The holographic technique is a process of photographically recording the itaage,
in which case, the wave pattern of the light scattered by the objective is
registered on the photoresist; the capability of reproducing the optical pattern
of the photographic object is assured in this case.
The X-ray radiography is the process of exposing the photoresist with soft X-rays
with a wavelength in a range of 5 to 50 A.
- A
7 8
~ y F77, / /7; iy,�:2 rif%srlr~~p~
3
Z S
(g) a) (b) 6)
Figure 5.16. Basic configuration of a
contact exposure and match-
ing unit.
Three methods of image transfer have found practical application in production
at the present time: contact, projection and electron beam exposure.
In light of the high requirements placed on the match-up precision (the error
shoulcl not exceed fractions of a micron), this process is carried out on
special precision equipment: image exposure and matching installations which
are complicated optical-mechanical complexes. The installations are character-
ized by the method and precision of matching, the resolution, the contact
quality, the productivity, the service life of the photographic templates (the
wear rate) and the permissible dimensions of a wafer.
The basic configuration of a contact unit for exposure and matching is shown
in Figure 5.16. The major components of the installation are the microscope
1 for visually monitoring the matching process, photographic template 2, at a
definite distance from which 3(the spacing is lesa than the depth of focus
of the microscope) wafer 4 is positioned. During the match-up process (Figure
5.16a), the wafer is moved along the X and Y coordinates and with respect to
-120-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-04850R000504090004-3
Figure 5.17. Basic configuration of a
ection exposure and
the angle ~ on coordinate table 5. After
the match-up has been made with the
requisite precision, the wafer comes in
complete contact with the photographic
template and it is exposed (Figure 5.16b)
by the high pressure mercurq vapor lamp
6 through shutter 8 and condenser lens 7,
which provides for the requisite illum-
ination uniformity over the ERtire sur-
face of the substrate. Projection match-
ing and exposing installations make it
possible to avoid contact between the
substrate and the photographic template,
which improves the durability of the
photographic template and promotes an
increase in image resolution.
proj
matching unit. Pro3ection photolithography systems are
subclassified as the following types
according to the method of generating the image on the substrate: with simul-
taneous transfer of the image in the f ield of the wafer; with sequential multi-
plicative transfer of the image; and with sequential scanning transfer of the
image. These systems (Figure 5.17) contain the illuminator 1, the matching and
focusing device (manual or automatic) 2, the match-up monitor unit 3, the match-
up manipulators 4, on which photographic template 5 i.s placed, which is pro3ected
through objective lens 6 onto wafer 7, which is positioned in the manipulator
(coordinate table) 8. The projection objectives can be the same for the
different photolithography systems. Images of the elements of a semiconductor
device to be exposed can be transferred by means of them from the photographic
template to a wafer with a working field of up to 50 to 80 mm.
r-----,
~ I
s j~ I

r--
i -
I
i
1S I
~
i
I
~

Figure 5.18. Basic configuration of an electron beam exposure
installation.
- 121 -
FOR OFFICIAI: USE ONLY

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-04850R000504090004-3
FOR OFF'ICIAL USE ONLY
Electron beam exposure installations consist of a large number of complex devices
[221. The typical electron beam exposure unit (Figure 5.18) consists of three
major assemblies: the photocopier or remote copier 1, the electron optics column
with the exposure chamber 2 and the video monitor 13. The electron beam from
kinescope 10, in passing through the focusing device, scans the topology on
photographic template 11 and transfers the information to photomultiplier 12.
The information from the photomultiplier is fed to the control unit 14 of the
video monitor with kinescope 15 and controls the electron beam which is generated
by electron gun 8 and electron lenses 7. The electron beam acts on the photo-
resistive layer on the surface of substrate 6. Block 9 serves to control the
beam deflection system and drives 3 and 4 for moving coordinate table 9.
In other types of installations, the control of the electron beam motion is
accomplished by means of a specialized computer, into which a program is fed
which provides for the requisite topology on the substrate.
The technique of matching the photographic template and substrate with visual
monitoring of their position relative to each other by means of base markers
(matching characters) has become the most widespread one in microelectronics.
The matching process can be carried out by the operator, who visually monitors
the mutual position of the photographic template and the substrate, and moves
them by means of the micromanipulator until they match. In the case of auto-
matic matching, a photoelectric device analyzes the position of the matching
marks on the substrate and the photographic template, generates an error signal,
which is fed to the micromanipulator drive and causes the substrate to move.
In the case of complete matching, the error signal disappears and the drive is
cut off.
The precision and productivity of matching installations with visual monitoring
depends in many respects on the subjective aspects of the operator also (visuai
acuity, etc.).
As a result of refining the structural designs for the major and auxiliary
mechanisms of contact matching and exposure installations using visual monitoring,
a productivity of more than 100 pieces/hr and a matching precision of 0.5 um
(Table 5.1) have been successfully achieved [23].
We shall consider the structural design of contact matching and exposure instal-
lations in more detail as well as the major requirements placed on their com-
ponent assemblies.
The major mechanisms of matching and exposure installations are the following:
--The match-up micromanipulator;
--The mechanism for orienting the plane of the substrate;
--The template holder;
--The loader;
--The contact exposure unit;
- 122 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFI(
TABLE 5.1. Comparative Characteristics of Contact Match-Up and Exposure Equipment
with Visual Monitoring
J&B2108
PLA-
CA-280 OH
VEB Elec-
CA-2020
CA-300
520A
PLA-500A
EM-576
Cobilt
tromat
Cobilt
Cobilt
Canon
Canon
Parametere
USSR
U.S.
GDR
U.S.
U.S.
Japan
Japan
Error in matching
0.5
0,5
0.3
0.125
0.75
-
-
the elements of
the phototemplate
and the substrate,
um
.
Size of the mini-
2
1
1
1
2
0.5
3
mum image element
on the substrate,
)In
Substrate dia-
60, 76,
up to
up to
up to
up to
up to
up to
125
meter, mm
100
100
76
100
76
100
Productivity,
160
-
130-150
-
-
90-100
100
pieces/hr
Microscope magni-
94, 257,
-
200
-
-
-
-
fication, times
208
--Microscopes for visual monitoring.
The Match-Up Micromanipulators. The matching process places a number of require-
ments on micromanipulators, the most important of which are the following: a
uiicromanipulator sl:oLld assure independenrr of the coordinate motions, sufficient
dimensions of the fields, high sensitivity, a definite algorithm for the muLions,
and motions in a plane parallel to the plane of the photographic template.
Because of this, there are various structural designs for match-up micromanipu-
lators.
One of the first structural designs of a match-up micromanipulator was a rotat-
ing coordlnate motion table with guides for rolling and a screw drive. The
sensitivity of the manipulations during the final matching in this type of
micromanipulator depends on the kind of drive and the stiffness of its coupling
to the table. A drive from a micrometer screw transmission is most frequently
used. To assure sensitivity of the microdrive down to tenths and hundredths
of a micron, a two stage drive with a lever transmission for the fine step is
employed [24J. A two stage lever drive is also used in micromanipulators with
a two-coordinate table. Besides the screw and lever drives, an eccentric drive
is used for rectilinear motions in micromanipulators with a two-coordinate
table on roller guides. A screw mechanism is used in the majority of cases as
the rotation drive in the manipulator, though sometimes a worm gear drive with
power locking.
- 123 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090044-3
FOR OFFICIAL USE ONLY
Drawbacks to the coordinate-rotational design of micromanipulator tables are
the complexity, cumbersome nature and difficulty in assuring that the three
motions are parallel in the plane of the photographic template.
In micromanipulators with a flat table and a pantograph drive, the lack of
parallel motions of the plane of the photographic template is reduced because
of the fact that the number of working surfaces is reduced (two). However,
micromanipulators with a pantograph drive do not provide for the requisite
independence of the coordinate notions during the matching [23].
The structural design of a micromanipulator with a magnetostrictive drive is
of interest. A drawback is the small motion field [23].
Manipulators are capable of providing a motion precision of + 0.1 um, but the
actual precision of a match-up with visual monitoring usually amounts to + 1 um.
The mechanisms for orienting the plane of the substrate perform two major
functions: they arrange the substrates strictly parallel to the working plane
of the photographic template and move the siibstrate with high precision when
contacting the photographic template. The major requirements placed on the
orientation mechanism are precision in the vertical motions of the substrate,
the preservation of the working surfaces of the photographic template and the
substrate during orientation, as well as assuring a complete contact between
the substrate and the photographic template and minimal displacements of the
matched substrate when it is contact with the photographic template. In the
majority of the well-known exposure and match-up units, the orientation of
[he plane of the substrate relative to the photographic template is accomplished
by means of a small spherical table. Following the equaliZation of the plane
oE the substrate, the position of the sphere is clamped by means of a vacuum.
A small table on three floating supports is also used for the orientation of
the plane, however, the friction in this case between the substrate and the
photographic template and the normal force is somewhat greater than when a
small spherical table is used. The precision of the vertical motions of the
substrate and the displacement of the substrate when in contact with the
photographic template depend on the precision and stiffness of the guides and
the points of application of the resulting force during contact. The result-
ing vertical load vector runs through the center of gravity of a triangle,
drawn through the points of contact of the substrate and the photographic
template, and rarely coincides with the center of the substrate.
The most diverse types of guides are used in the well-known installations:
cylindrical sliding guides, prismatic roller guides with power locking, guides
in the form of a parallelogram with flexible hinges [22, 24, etc.].
The orientation mechanism, following the match-up, presses the wafer against
the photographic template and assures contact between the surfaces. The quality
of the resulting figure during exposure depends to a great extent on how com-
plete the contact was between the substrate and the photographic template.
The change in the dimensions of the elements and the increase in the photometric
wedge, which occur as a result of diffraction phenomena and multiple light
- 124 -
FOR OFFICIAL USE ONI.Y
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500094444-3
reflections from the surfaces of the wafer and the metalized photographic tem-
plate, are in direct proportion to the gap between the template and the wafer.
The difficulties of producing a tight contact are due, first of all, to local
uneven places on the wafer with a height of up to 1 um, which occur during the
polishing process, during epitaxy and even during photolithography itself;
and secondly, they are due to distortions of the wafer because of exposure to
various production processes, especially heat treatment. However, the consider-
able ratio of 1/100 - 1/250 between the thickness (usually 0.2 to 0.3 mm) and
the diameter (20 to 75 mm) imparts adequate elasticity to semiconductor wafers
(2sl�
To assure contact between a wafer and a photographic template, a provision is
made in a number of orientation mechanisms.for power clamping of it by means
of pneumatic cylinders or a lever mechanism, in which the clamping force is
produced by an adjustable compression spring.
A more refined approach is to press the wafer against the photographic template
using air pressure following the creation of a vacuum between them.
Loaders and Template Holders. In installations intended for working with
emulsiun photographic templates, tl:c thickness of which changes little, the
latter are secured to a template :iol_:.er on the bottom, to the nonworking sur-
face by means of a vacuum. In installations which allow the use of photo-
graphic templates, the difference tn the thickness of which runs up to several
millimeters, mechanical fastening of the photographic template in a template
holder is employed (for example, by means of a bayonet lock).
The wafers are loaded by means of satellites, arranged on the rotating disk
of the installation, or by means of moving from the loading position into the
working zone, and then to the unloading table by means of a push rod. Further
automation of the wafer unloading and loading in pattern matching installations
led to the appearance of a pneumatic transport system with a device for the
preliminary orientation of the wafers to line up with a cut out segment or
a cut out groove [23].
The Contact Exposure Assembly. The contact exposure assembly includes a light
source, an optical device to produce a light flux, a mechanism for controlling
the transmission of the light flux and a housing for holding the light source.
The optical device in a contact exposure assembly is intended for producing a
uniform light flux with a parallel bundle of rays in a definite.,range of wave-
lengths over the entire exposure field. The exposure field diameter in modern
installations should amount to more than 75 to 100 mm, while the scatter in
the illumination over the entire field should not exceed 5 to 10%. Such para-
meters can be assured by special systems of quartz condensers, having from
1 to 5 lenses. Moreover, in order to segregate a particular wavelength which
is most suitable for the photoresist being used from the overall radiation
spectrum of the lamp, it is necessary to have a set of appropriate light filters
-125-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
[22]. Uniform illumination over the exposure field is one of the major require-
ments for a high quality photolithographic process. The illumination system
plays an important part in this case.
3 4 S G 7
1 2 l X xl
x ' ~
x k k X~ ~
X k,~y � z
d 9
Figure 5.19. Optical configuration of a scanning illumination
system.
A scanning raster illuminator satisfies this requirement (Figure 5.19) [26].
The light source (a DRSh-250 lamp, 2, in the focal plane of spherical mirror
1) projects through the thermal filter 3 with condenser 4 onto lens raster 5;
following magnification by the lenses of raster 6, the mutually superimposed
light spot images of the source (the number of spots corresponds to the number
of raster lenses) are projected through lens 7 to the plane of the photographic
template 9(close to lens 8).
In this case, the high uniformity of the illumination of the photographic
template is accompanied by a reduction in the influence of spatial and time
instability of the light flux from the DRSh lamp.
The mechanism to control the transmission of the light flux, the shutter, is
needed so as to set the requisite exposure time, which depends on the sensi-
tivity of the photoresist. The exposure time can fall in a range of from 1
second to 2 minutes and more, and for this reason, the major requirement placed
on the shutter consists in the fact: that the opening and closing time of the
blind be on the order of 0.05 to 0.1 sec, while the relative actuation error
does not exceed 10%. Electromagnetic shutters which meet these requirements
are used in modern installations.
The light, condenser and shutter ar-- t:,;used in a single block. The block makes
it possible to correctly adjust the lamp relative to the optical axis of the
quartz condenser and protect the operator and light sensitive materials against
the harmful effect of ultraviolet radiation. The structural design of the
housing should assure normal thermal conditions for lamp operation. In the
majority of installations, the housings for the lamps are air-cooled [22].
Visual Inspection Microscopes. The precise mutual positioning of the matched
structures is determined by meana of a microscope for the visual observation
and quality control of the matching.
The major requirement placed on the microscope consists in the fact that it
should provide for a clear image of the two structures being matched, which
- 126 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
are located in different parallel planes at a certain distance from each other
during the match-up time, as well as during the time when the quality of the
matching is monitored after the pressing of the wafer against the photographic
template. In contact exposure and matching installations, the followi.vg para-
meters characterize the microscope: the overall magnification determin,ss the
structure image scale; the resolution determines the smallest visible size of
a structural element; the depth of focus determines the working gap during the
match-up time; the field of view determines the observation area; and the
working distance determines the thicknese of the phototemplate glass and the
structure of its holder.
These parameters are interrelated, depend primarily on the characteristics of
the objective used and are designed using the laws of optics.
The basic design quantity is the numerical aperture of the objective, which is
estimated from the formula:
A = nsinu
where n is the index of refraction of the medium; u is the angle formed by
the rays from the point source, positioned on the optical axis of the objec-
tive in the first main focal plane, to the ends of the diameter of the objec-
tive.
The numerical aperture defines the overall useful resolution of the microscope
and the resolving power: with an increase in the aperture, the objective
resolution increases (d = 0.61a/A, where a is the wavelength of the light),
but the depth of focus decreases (I' a/2A2). The decrease in the depth
of focus is accompanied by more stringent requirements placed on the quality
of the substratea, on the precision of the positioning of the wafers, the
manipulation of them, etc. Thus, if an objective has an aperture on the order
of 0.5, then the resolution will be about 0.6 um, while the depth of focus
will be on the order to 2 to 3 um, which makes it very difficult to match-up
actual semiconductor wafers which frequently have a considerable curvature.
To match-up elements with dimensions of 2 to 5 um, it is necessary that the
objective have an aperture of at least 0.2, then the useful magnification will
be 200 x, the depth of focus will be on the order of 10 to 15 um while the
working field will be about 1 to 3 mm. However, such a working field for a
microscope is many times less than the size of a wafer. For this reason, it
is impossible to check the match-up quality over the entire wafer and it is
easy to allow an angular shift, because of which considerable linear error
occur in the matching of the elements remote from the working field. This
circumstance has led to the design of special microscopes, which make it
possible to observe in the field of view of the ocular two sections of the
wafer being matched-up at the same time where these sections are a certain
distance apart. The structural designs of two field microscopes can vary:
with a separate (split) field and with a double field.
-127-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The images from the two objectives are brought to the field of view of the ocular
independent of each other, whera this field is aplit into two sections so that
the image from the right objective falls in the right portion of the field of
view while that from the left ob3ective falls on the left side. The operator
has the capability of simultaneously observing both portions of the wafer being
matched (split field microscope).
7
4
J As a rule, microscopes are used which
have a magnification which changes con-
~ tinuously or discretely in a range of
from 40 to 80 x(survey) to 100 to 400 x
(preciae matching); the minimum image
size is 1 um.
In step with the miniaturization of
topology, obaerving using standard micro-
scopes having a magnification up to
400 x and a resolution in a range of 1
um is made diff icult. At the present
C~ ~time, equipment has been developed in
.v r which the microscope magnification can
Figure 5.20. Optical configuration of be switched in steps to 500, 1,000 and
a microscope with a 2,000 x. Such microscopes have a high
built in TV camera. resolving power of 0.45 um. Moreover,
a remote camera is incorporated in one
of the blocks of the microscope-optical
system, which makes it possible to observe the topology on a screen, determine
the presence of dust and defects in it, etc. The operational principle of a
microscope with a built in television camera can be seen from Figure 5.20.
Wafer 8 is located on stage 7, where the photographic template 5 with the
topological figure 6 is placed a certain distance from the wafer. The image of
the wafer surface and the topological figure of the photographic template
are perceived by the eye through the ob3ective 9, the field splitting priem 2,
intermediate lens 4, beam splitting prism 2 and ocular 1. A portion of the
rays is fed through the beam splitting prism to the vidicon of TV camera 3.
One such exposure and matching installation with a television screen is the
unit made by the American company III.
It must be noted that units for matching with visual monitoring and with contact
exposure are the most widespread at the present time.
A drawback to the contact technique is the rapid wear of the photographic tem-
plate.
The automated EM-576 unit can be cited as one of the industrial models of a
matching and exposure installation. An sutomated cassette !'iolder for transport-
ing the wafers and a device for their preliminary orientation are included in
the unit, wh ich makes it possible to use it both independently and incorporated
in automated photolithography lines. The installation makes it possible to use
both contact exposure and exposure with a gap.
- 128 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
Its main parameters: Productivity (without taking the match-up
time into account), wafers/hr 160
Match-up error, um 0.5
Minimum size of the componenta on a semi-
conductor wafer with contact exposure, um 2
The diameters of semiconductor wafers for
the case of exposure with a gap, mm 60, 76,
100
The installation can operate under conditions where vibrations act on it in
a frequency range of 1 to 5 Hz with an amplitude of no more than 5 um�
The installation inclucles:
--A precision manipulator and match-up microscope, by means of which a high
precision is obtained in the matching of images on a photographic template
and semiconductor wafer;
--An illuminator, which provides for good repzoducibility of the minimal image
elements over the field of the wafer;
--An automated cassette type transport unit for the wafers and a device for
their preliminary orientation;
--Interchangeable attachments for semiconductor plates and photographic templates
of different standard sizes.
The EM-576 installation can be used either independently or incorporated in
the equipment for automated photolithography lines.
5.5. Equipment for Producing Topological Relief on a Substrate
The formation of the topological relief on the substrate completes the photo-
lithographic cycle, which contains the following operations:
--Etching (masking, insulating, protective, conductive and other layers);
--Removing the photoresist;
--Washing (prior to diffusion, metalization and passivation).
These operations can be carried out based on the use of chemical or plasmo-
chemical methods, the specific differences in which are responsible for sub-
stantial structural design and functional differences in the equipment used
for these methods.
The chemical wafer processing technique is characterized by the following:
--The corrosiveness of the reagents and because of this, the necessity of using
closed working volumes, and corrosion resistant structural materials for the
functional blocks; �
- 129 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
FOR OFFICIAL USE ONLY
--Thermostatic
the necessity
- processes;
temperature control of thP working volumes, and because of this,
of using units for stabilizing and controlling the production
--The necessity of neutralizing the chemical reagents.
The plasmochemical method makes it possible to perform several, and in the
future even ail of the operations indicated above without using liquid reagents.
The wafers are treated in a low temperature oxygen or carbon halide gas plasma,
excited in the working volume of the chamber by means of a high frequency or
microwave discharge.
Plasmochemical etching provides for greater resolution and control of the etch-
ing profile, and reduces lateral undercut etching as compared to chemical fluid
etching, which brings about an improvement in the precision of the geometric
dimensions of the topology. Various plasmochemical processes may be carried
out sequentially in a single chamber. The moment of completion of each process
can be easily registered by means of optical spectrum, mass spectrometry, laser,
interferometric and other physical contactless testing methods. All of this
creates the conditions for the complete automation of the production processes
fo r surface treatment. Moreover, the plasmochemical method of surface pro-
cessing excludes or considerably reduces the use of chemical reagents, which
not only curtails capital expenditures for the construction of cleaning
facilities, but also eliminates the contamination of the environment by chem-
ical production wastes.
Universal installations which are incorporated in universal equipment for the
chemical treatment of wafers can be used to execute the operations enumerated
above for the formation of topology by chemical means.
The functions of the complex are:
--The cleaning of the wafers prior to the first oxidation;
--The etching of the oxide layers, the boron silicate and phosphorus silicate
glasses;
--The removal of the photoresist both from the oxide layers and from the
metallic surfaces;
--The cleaning of the oxide layers and the metallic surfaces following the
removal of the photoresist;
--The etching of the metal and silicon.
The complex consists of seven independent lines, where the package is put
together depending on the technical function of the installations for the
chemical treatment, washing and drying, processing in organic solvents, ultra-
sonic and hydromechanical washing as.well as with quality control units.
The system is filled with reagents from mobile transport blocks and the chemical
reagents and deionized water for washing are fed from ultrapure water systems
with recycling. All of the equipment of the complex is of a modular block
- 130 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
design and can be put together in production process lines which are tied
together with the power supply systems.' The number of installations which can
be assembled into a line is governed by the technological processes and the
requisite productivity.
6
S
4
3
2
Figure 5.21. A unit for the chemical
treatment of wafers.
The most typical is the structural design
of the 084KhP-100-004 chemical wafer pro-
cessing unit (Figure 5.21).
The unit consists of production process
and dust removal blocks.
The wafer treatment technique is a
cassette holder group type system and
has two standard cassette sizes for
products either 65 or 75 mm in diameter.
Production process block 1 on base 2 is
placed on a support for the mounting of
the dust removal unit 6. The production
process unit is made in the form of a
bath made of wood particle board lined
with polypropylene. A perforated grate
8 with holes for the placement of the
production process baths is located on
the flange edge of the bath.
The wafers are chemically treated in
various reagents in teflon bath 4, which has an immersion heater. Teflon baths
4 are equipped with ejectors to exhaust the oxygen after the production process
is completed. Acid diluted with tap water is drained out at the bottome of
production process unit 1 and then through drain pipe 3 to the station for
cleaning the industrial effluents. The next step is the washing of the wafers
with deionized water and is accomplished in the cascade washing bath 7. The
cassette holders with the wafers are moved manually.
For the purpose of preventing dust getting onto the wafers being processed
from the production room, the production process blocks are placed in dust
removal units 6, which create a laminar flow of dust free air.
The requisite temperature modes are automatically maintained and signal light
5 signals the conclusion of the production process operations.
The controls are located on the front panel of the production process unit.
The technical data on the unit
The number of wafers which can be processed simul-
taneously, pieces
Reagent heating time (volume of 3 liters) to 100� C,
minutes
- 131 -
FOR OFFICIAL USE ONLY
50
30 + 5
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
Reagent working temperature, �C
Power cons umption, KW
Overall dimensions, mm
+50 to +120, +5
2.3
1,915 x 945 x 470
The remaining units of the complex are likewise of a modular block design with
maximum utiliz ation of standardized modules.
Plasmochemical Treatment Equipment. The major requirements placed on installa-
tions for plasmochemical processing are [27]:
--High uniformity of the distribution of the electrically neutral radicals, atoms
and charged particles throughout the entire volume of the reaction space;
--The reproducibility of the energy state of the plasma with respect to time
and from pro cess to process;
--High stability of heat and mass exchange in the reaction space;
--Monitoring of the moment of completion of the etching process for one material.
Only when equipment meets these requirements can a high uniformity of surface
treatment be ob tained.
To meet the requirements indicated above, plasmochemical processing equipment
should be controlled with respect to many parameters: pressure, gas flow rate,
discharge power, frequency of the RF electrical field and temperature of the
wafers.
A high frequency electrical discharge in a low pressure gas is used to obtain
a low temperature gas plasma in plasmochemical processing units.
Depending on the kind of operation performed, various techniques can be used
to obtain and sustain a plasma as well as various structural designs for the
reaction discharge chambers.
For example, a high frequency electrodeless discharge in a cylindrical quartz
chamber is used for the removal of photoresist, while a low pressure electrode
diode discharge is used in planar metallic reactors for etching.
Plasmochemical processing is carried out with group loading of the wafers, which
are placed in the discharge chamber, which is hermetically sealed with a cap
by means of a vacuum produced by an initial vacuum pump. Upon reaching a
pressure of 40 to 60 Pa, while the initial vacuum pump is running constantly,
the variable working gas infiltration is actuated, i.e., a dynamic vacuum is
- produced which is governed by the geometry and structural design of the chamber
and electrodes, as well as the method of excitation and the intensity of the
applied RF electric field.
Reaction-discharge chambers of installations for the electrodeless excitation
of a discharge are shown in Figure 5.22. Planar reactors with a diode excita-
tion system ar e shown in Figure 5.23.
- 132 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
Gas Input A�~~~~
za.,rc
o '
1 2 .l
Type A
Tzn A
(A) I aeisad zaya (A)
� Busod i
taaa 1 2
~ 3 4
. . ~ Tun 6
(B) I Type B
Brod raaa
Bsod zaaa
000
- Tu^ fl.
000
 3 Type C
(A~~,soa ~
ra~a
� Beizod ' Z Type D
� raa� - - y- J
. Trcn
raja
(A) ~I II I ~
Bei,zod Z ~
r.aia o vee
- � J Tt~nQ
6~ aomoom ype E
a, z } d'r t G a s Input
Figure 5.22. Types of reaction dis-
charge chambers.
Key: 1.
2.
3.
4.
5.
6.
Reactor;
Cap seal;
Cap for the reactor;
Linings;
Receiver;
Cap.
A. Gas output;
B. Gas input.
The homogeneity and stability of the
plasma in the reaction discharge chamber
depend on the pressure in the chamber
at the working frequency of the RF field
and the structural design of the elec-
trodes [28].
The power of the RF genergtore in plasmo-
chemical treatment units in cylindrical
reaction discharge chambers with elec-
trodeless excitation of the discharge
usually amounts to 300 to 1,000 W, while
in planar reactors with diode excitation
of the discharge, it is 1.5 to 3 KW, and
in some cases, 6 KW.
We shall treat the structural design and
operational principle of the domestic
"Plazma-600" plasmochemical unit for
photoresist removal and prediffusion
cleaning of the surface of semiconductor
wafers (Figure 5.24) with electrodeless
excitation of the discharge by means of
inductive coupling.
1,
Figure 5.23. A planar reactor with an
electronic P.F discharge
excitation system.
Key: 1.
2.
3.
4.
5.
6.
-133-
FOR OFFICIAL USE ONLY
Vacuum pump;
Restrictor;
Manometer;
Lower electrode and
stage for the wafers;
Upper electrode;
Wafers.
n �RF
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The Technical Data for the Installation
Productivity (per photoresist removal operation),
wafers/hr
Simultaneous loading of the reaction chamber,
wafers
Diameter of the wafers which can be treated, mm
RF generator:
Output power, watts
Working frequency, MHz
Working pressure of the oxygen in the chamber, Pa
Oxygen rate of flow, liters/min
300
up to 50
up to 80
600
13.56 + 1%
(0.6-1.3) � 102
0.36
The major assemblies of the installation are (Figure 5.24): the vacuum ex-
haust block 1, the reaction discharge chsmber and RF generator block 2, and
the control block 3.
J E 1 =-.Ov
o~
7,
~
I
O
'
rigure 5.24. The "Plazma-600" plasmo-
chemical unit for photo-
resist removal.
The low temperature plasma is formed in
the chamber at a vacuum on the order of
60 to 133 Pa, when a RF voltage is fed
to the indicator placed around the
chamber.
Wafers can be sequentially treated in two
gas plasma media in the installation.
When oxygen is admitted into the chamber
at a pressure of up to 133 Pa, atomic
oxygen and oxygen ions are formed with
the action of the discharge, which have
a high chemical reactivity. They
oxidize the photoresist, forming vola-
tile compounds as a result of the chem-
ical reactions.
The final products of the photoresist
decomposition reaction are C0, CN, C02
and H20, which are exhausted from the
chamber by a vacuum pump.
When freon-14 is admitted into the cham-
ber at a pressure of from 40 to 70 Pa, the action of the discharge forms
atomic fluorine, which as a result of the chemical reaction cleans the surface
of the wafers of oxides and contaminants.
Tlie chamber and generator block includes the reaction-discharge chambers, the
RF generator, the vacuum gauge and the ventilation system for cooling the
cli:imbers and generator units.
- 134 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090044-3
The chamber takes the form of a quartz tube, which is hermetically sealed in
front with a door, and is connected to the vacuum system at the rear. The tube
is ringed with the RF induction unit coils.
The process is automatically controlled (creating the specified vacuum, feeding
in the working gas, exciting the discharge and breaking the seal). Both
automatic and manual operating modes are provided in the installation using
two programs: with a single gas (photoresist removal) and with two gases
(photoresist removal and prediffuaion cleaning).
Installations with planar reactors,
. having diode electrode excitation of
~y the discharge, make it possible to
~ Z dCNff0700
efficiently utilize the plasma to etch
~ 3 --iiiqh silicon dioxide and aluminum with a
Fxequency high level of rarefaction. The Plasa-
Generator fab 3200 unit of the Electrotech Com-
~ , pany (England) can serve as an example
Ornaxdcliue of such equipment. The installation
Cooling is an automated rotor type unit with
Figure 5.25. Schematic showing a diode 8 diode system for plasma excitation.
discharge in a reaction The wafers are positioned horizontally
chamber. in the reaction chamber. A schematic
of a diode discharge in a reaction
chamber is shown in Figure 5.25. Wafer
1 is placed between the electrodes in a stainless steel vacuum chamber. One
electrode 3 is a water-cooled wafer holder; the gas which forms the plasma
between the high frequency field electrodes is admitted to the chamber through
the upper electrode 2. The active radicals of the dissociated gas react with
the wafer surface, forming a volative compound. Water-cooled holder 3 makes
it possible to maintain a constant temperature for the wafers and protect the
photoresist against destruction. The productivity of the unit is 200 wafers/hr
with diameters of 76 or 100 mm.
Besides the plasmochemical technique for treating the surface of semiconductor
wafers, another method of precision processing is also being developed at
the present time: ion-chemical. It is distinguished from the plasmochemical
method by the higher energy of the plasma beam of ions and the radicals of
fluorine and chlorine containing gases, which is directed onto the wafer being
processed at a pressure of (1 to 5) � 10-2 Pa. This technique has a great
future with the transition to micron and submicron topology.
The development of plasma and ion chemical techniques for precision treatment
of the surface of semiconductor materials makes it poseible to create a com-
pletely "dry" photolithographic process.
Quality Control During Photolithography
Checking the quality of photolithographic treatment in the production of semi-
conductor devices and integrated circuits is accomplished visually.
- 135 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
7
The use of optical and mechanical means for monitoring the geometric dimen-
sions of micron and submicron structures is difficult because of the inadequate
resolution and depth of focus of light optical instruments. Thus, the uneven-
ness of the edge of an element of semiconductor structures 1 um wide amounts
to about 0.1 um. The resolving power of the test equipment should be at
least 3 to 5 times greater, i.e., should amount to 0.02 to 0.03 um [30]. When
checking microtopology using a conventional microscope, the measurement error
is + 0.5 um� This estimate shows the unsuitability of optical and mechanical
means for quality control of semiconductor structures with micron and sub-
micron dimensions of the elements.
Scanning electron microscopes [31, 32] are being increasingly utilized for this
purpose at the present time, where these microscopes make it possible to non-
destructively test objects with a resolving power of 100 to 300 A. However,
considering the comparative complexity of electron optical quality control
techniques, it is expedient to perform quality control operations at that
stage in the technology which has the greatest impact on the production process
as a whole. For this reason, it is important to check the quality of struc-
tural elements on the photographic template. Quality control of inetalized
photographic templates using a scanning electron microscope makes it possible
to determine the existing distortions of the image, the unevenness of the
edge of elements, breaks and pores in the chromium layer, unetched areas and
other defects, and thereby reduce the number of rejects in photolithographic
operations.
5.6. Equipment for Fabricating Photographic Templates*
The basic tool for producing spatial topology for semiconductor devices and
integrated circuits in the photolithographic process is the photographic tem-
plate. A photographic template takes the form of a transparent substrate,
coated on one side with a film which does not pass the actinic radiation, in
accordance with the regions which form the specified topological figure. The
substrate can be made from optical glass, while the masking film can be made
from chromium, metal oxides and other materials.
Photographic templates are manufactured in accordance with the following pro-
duction process scheme:
--The execution of the original of the topological drawing to an increa:,ed
scale (200:1, 500:1 and more);
--One-time or multiple reduction of the primary original;
--The execution of the master photographic template by means of printing the
images on its working surface (multiplication);
*The description of the equipment for the fabrication of precision photographic
templates as a support production process goes beyond the scope ot the sub-
jcct of this book, and for this reason, only some individual explanatory data
are given below based on materials from the literature [23, 33], which contains
considerable material on the question touc'1ed upon here.
- 136 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
--The fabrication of photographic copies from the master template (working
photographic templates).
The primary original of the drawing of the topology for the photographic tem-
plate can be produced in two ways:
--By cutting the figure into an opaque varnish film applied to sheet glass;
--By means of phototypesetting using image generators.
In the first case, precision coordinatographs are used, and in the second,
phototypesetters. The intermediate reduction of the original is accomplished
with reduction cameras, while the final reduction of the photographic template
topology and the multiplication are carried out using a scanning camera or
photographic duplicator.
To fabricate working photographic templates, the equipment described above as
well as photolithographic techniques are employed.
Coordinatographs are broken down according to the kind of control of the motion
of the cutting tool, into manual and automatic types, and can differ in the
coordinate system employed (cartesian, polar, mixed), as well as in the pre-
cision of the execution of the boundaries of the topological elements. The
productivity of a coordinatograph is usually related to the precision of
the execution of the topological figures. Considerable precision is achieved
with a comparatively low output productivity (slow motion of the cutting tool).
Originals of a photographic template topology figure with dimensions of 750 x
750 and 800 x 800 mm for a precision in setting the boundary of a figure
element of + 50 Um and a repeat precision of + 25 um can be made on manually
controlled coordinatographs, for example, the EM701 and EM707 respectively.
In automated coordinatographs, the working tool moves in accordance with a
specified program with computer control. Because of this, an automated
coordinatograph contains a set of equipment similar to a production process
automated control system of the first functional algorithmic level (see
Chapter One): data input-output peripherals, a computer, equipment for con-
trolling the motion of the drawing tool, a control console and various indi-
- cating instruments. The major functional unit of an automated coordinatograph
is the drawing table.
Primarily low productivity and the cumbersome nature of the equipment can be
numbered among the drawbacks to the technique of fabricating originals using
coordinatographs.
Characteristic Technical Specifications for Automatic Coordinatographs
(Given for the EM-703)
Maximum size of the original, mm
Minimum step, um
- 137 -
FOR OFF'ICIAL USE ONLY
1,200 x 1,200
25
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504090044-3
FOR OFF[CIAL USE ONLY
- Drawing precision, um + 50
Repeat precision, um + 25
Resolution, um 25
Travel speed along the coordinate axes, mm/sec 100
Data input from punched
tape
Data coding Binary-decimal
Data input rate, characters/sec 1,500
Area occupied, m2 30
Reproduction cameras for the intermediate reduction of a photographic template
original are photographic installations for precise photography and contain
the following major assemblies which are characteristic of such installations:
a system for uniform illumination of the original, fastening devices for the
original, an optical system, a cassette holder with the photographic plate and
functional support equipment. The major unita of the camera are mounted on
a massive cast support bed, which is usually mounted on shock absorbers.
Technical Specifications for the EM-513 Reduction Camera
Maximum size of the plane table of the original, in imn
Maximum working field of the original, mm
Dimensions of the photographic plate, mm
Original reduction scale
Conditions for transilluminating the original:
Screen brightness, Nit
Brightness nonuniformity, X
Motion of the photograghic template in the image
plane, in mm:
Horizontally
Vertically
The precision of photographic plate motion, mm
The precision of the repeat setting of the original, mm
The precision in checking the image dimensions, mm
ExposurP time, seconds
Overall dimensions, mm
Weight, kg
1,300 x 1,300
1,200 x 1,200
90 x 120, 60 x 90
30, 40, 50
5,000
up to 10
90
100
+ 0.002
+ 0.1
+ 0.002
0.5 to 999
8,600 x 1,790 x
2,165
2,700
Reproduction cameras, just as coordinatographs, are rather cumbersome.
- 138 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
New methods of producing images, which make it possible to replace the
operations of cutting the photographic template original and its intermediate
reduction with a single process are based on the utilization of scanning image
generators or microphotutypesetting installations. When generating images by
means of scanning, the light spot moves relative to the photographic plate,
for example, by means of using line by line (raster) scanning and modulating
the light intensity in accordance with a program corresponding to the photo-
graphic template topology. It is expedient to use a laser as the light source.
The realization of a universal scanning image generator involves difficulties
of assuring resolution and operational speed. Because of this, the scanning
method is usually combined with phototypesetting. Thus, the technique of
outline scanning is realized in the form of one of the operating modes of the
EM-539 image generator [34].
Microphototypesetting is based on the sequential exposure of fragments (type-
setting elements) of the photographic template on the photographic plate.
Microphototypesetters contain an automated coordinate table, program and type-
setting element generator controllers, an optical projection system and a
light source.
Technical Specifications for the EM-549 Phototypesetter
Coordinate table travel, mm
140 x 140
Error in the positioning of the coordinate table, um
+ 0.5
Rotation angle of the typesetting aperture stop, degrees
45
Average productivity, exposures/Hr
2,400
Maximum travel speed of the coordinate table, mm/sec
5
Reduction scale
1:10
Minimal width of the image lines, um
10
Maximum dimensions of the square, um
3,000 x 3,000
Deviation of the actual dimenaions from the nominal,
um
1
Edge unevenness, um
1
Radii at the corners of elements, um
1.5
Exposure time, sec
0.1 to 10
Current
Alternating
Voltage, volts
380 x 220
Frequency, Hz
SO
Air pressure, atm
4 to 8
Maximum power consumption, KW
2
Area occupied, m2
20
- 139 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500094404-3
F'OR OFFI('IAL USF: ONLY
The master photographic templates are made using optical or mechanical multi-
plication methods. The former employ lens or mirror raster scans while simul-
taneously photographirig a large number of reduced images of the topological
figure of the photographic template, while the latter are based on the sequen-
tial, step by step photographic printing of single images. An example of an
optical multiplication unit is the EM-514 scanning photographic camera. The
scanning objective of this photographic camera contains no less than 1,500
short focus lenses, positioned with a step of 0.8 mm. The working field of
the screen is 300 x 300 mm and the reduction scales of the camera are 200 x
and 300 x. The camera is used for the multiplication of images with minimum
element dimensions of about 20 um.
~ Photoduplicators based on sequential photographic printing with the moving of
the photographic plate contain an automatic coordinate table, an exposure
assembly (illuminator, optical system, etc.), measurement instruments, as
well as controls and structural support. We shall give the basic technical
specifications for such photoduplicators using the example of the EM-522 unit:
Maximum travel of the table, mm 80
Positioning precision, um + 0.2
Positioning reproducibility, ym + 0.2
Reduction scale, times 10
Intermediate photographic template dimensions, mm 50 x 50,
70 x 70
Master photographic template dimensions, mm 70 x 70,
100 x 100
Mechanical multiplication systems are the best developed and the most universal
at the present time. However, the development of precision electronic multi-
plication systems up to the level necessary for industrial applications will
be a reality in the next few years.
- 140 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
PART II EQUIPMENT FOR THE ASSEMBLY AND QUALITY CONTROL OF
FINISHED DEVICES. FINISHING OPERATIONS
CHAPTER SIX Equipment for Separating Wafers into Chips
The geometric shape of the chips used in the mass production of semiconductor
devices is the most diverse: circular, square, rectangular and polygonal;
however, square and rectangular chips are the most widely used. The side of
a chip varies from 0.25 mm up to several millimeters.
The following methods of separating wafere into chips have found application
at the present time: .
--Scribing with a diamond cutter;
--Laser cutting;
--Separation using diamond disks;
--Separation using cutting arrays.
Scribing with a diamond cutter is a rather high output operation as compared
to other types of wafer separation, especially when producing small chips (in
a range of 0.35 to 1.0 mm). The advantages of the technique are also the easy
resetting of the cutter for square and rectangular chips of various sizes and
the servicing simplicity. However, there are drawbacks inherent in the method
of scribing using diamond cutters:
--The difficulty of scribing Si02 and polysilicon;
--The occurrence of dust-like formations which get on the surface of the
structures being scribed (the dust is the disintegrated cut material);
--The influence of the quality and condition of the cutting edge of the diamond
cutter on the shape of the scriber groove, and as a consequence, on the
quality of the breaking of a wafer into chips;
--The necessaity of performing a supplemental breaking operation on the wafers
along the scriber lines;
--The presence of an improper geometrical shape of the cut surface (shear sur-
faces, violation of the rectangular shape, etc.);
--A deterioration of the break quality and an increase in rejects with a reduc-
tion in the ultimate ratio of chip length to thickness (this ratio should
be no less than 3:1 for germanium and 4:1 for silicon).
The expanding use laser cutting of semiconductor wafere in industry is due to
a number of advantages of this technique:,
--There is no cutting tool which wears;
--The electronic control of the laser beam makes it possible to adjust the
parameters of the cutting channel in a wide range: from producing the shape
-141-

FOR OFFICIAL USE-ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
FOR OFFICIAL USF ONLY
of a scriber "notch" to completely cutting the wafers apart and separating
the chips from each other.
Draw backs to this method are the necessity of using complex equipment for the
process and the possibility of droplets and vapors of the material getting on
to the surface being cut.
A promising method of separating semiconductor wafers into chips is notching
or completely cutting them apart using rapidly rotating thin disks, coated
with diamond dust. A merit of the method is the possibility of producing chips
with a good geometric shape and precise dimensions, as well as the possibility
of cutting apart structures on which beam or strip leads are formed. Moreoever,
where a polycrystalline layer is present at the surface of the semiconductor
wafer, cutting with rotating disks yields a more stable and high quality
separation than does scribing.
Separation by means of cutting arrays or wires is less productive than using
the methods indicated above. In all of the latter methods, it is necessary to
provide for feeding an abrasive suspension to the surface of the materials
being cut, something which is extremely undesirable in some cases. However,
good separation quality and the capability of cutting apart relatively thick
wafers provide for the rather wide scale use of these methods in industry.
At the present time, wafers of semiconductor materials with an increased diameter
(more than 100 mm) are being introduced into the technological process. The
requirements placed on the precision of the geometric dimensions of chips are
increasing, especially for devices with an increased level of integration. It
must be noted that the most promising method of separating large diameter wafers
into chips is cutting with diamond disks.
6.1. Equipment for Separating Wafers by Means of Scribing
The EM-201A and EM-201B diamond cutter scribers [35] (Figure 6.1) have found
wide applications in domestic industry. The unit has the following technical
specifications:
Range of steps, mm
Discrete setting step, mm
Maximum feed travel ?ength, mm
Length of a scribing run, mm
Number of double runs per minute
Pressure of the cutter on a wafer, N
Precision in making lines during a step, mm:
0.01 to 15 mm
15 to 30 mm
- 142 -
FOR OFFICIAL USE ONLY
0.01 - 39.99
0.01
75
40 - 85
from 15 to 60
0.125 - 2.5
+ 0.005
+ 0.006
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
Figure 6.1. General view of scribing units.
a. EM-201A;
b. EM-201B.
The scribing installation has the following specific design features. The
scribing speed is set by selecting interchangeable gears. The length of a
working run is regulated to match the diameter of the wafer being scribed. A
correcting linear rule is used in these models which improves the precision of
the step feed motion mechanism. The nut and screw of the feed mechanism have
forced lubrication. A sophisticated structural design of the cutting head pro-
vides it with a low inertia. A precision mechanism for setting the cutter loading
force has been introduced. The cutting head can be rapidly aligned outaide the
machine, where the changeover from the previously used cutters with a circular
tool post to cutters with a square tool post significantly curtails the setup
time and improves the scribing precision. The EM-201B model is provided with
a projector and is recommended when processing large diameter silicon wafers
and sitall [ceramic glass similar to pyroceram] substrates. Vacuum suction
provides for the restraint of the wafer being cut. One can scribe wafers held
on adhesion substrates in the installation also: polyxylaxane strip, viniproza
[sic], etc.
A significant factor which determines the quality of separation is the manner
of breaking the scribed wafer. The most widespread method of breaking is rolling
a small roller along the scribe lines with the application of a r.ertain force
on the surface of the wafer. Ir. this case, the wafer either moves between two
elastic plastic substrates, or in an envelope (from which one can exhaust the
air and which can be hermetically sealed), or on an elastic base. A number of
practical data on scribing and breaking silicon wafers with a thicknesa of no
more than 0.3 mm in the production of chips 0.45 x 0.45 mm for transistors is
given in [37].
It has been established from production experience that one can obtain a minimal
widtll scribe groove which makes it possible to have high quality separation of
a silicon wafer into chips using a diamond cutter with a sharpened angle of
150� and an angle of inclination of the cutting edge to the plane of 5 to 6�.
- 143 -
FOR OFF ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
Changing the scribing speed in a range of 10 to 50 double runs per minute has
little influence in practice on the scribing quality.
1'he orientation of the crystallographic planes of the wafer does not influence
the process of making the scriber groove, but can have a decisive importance
when sepprating the scribed wafer into chips. The best results are obtained
when scribing wafers grown in the (100) plane along the traces of the (100)
planes. The working of the back side of a wafer (polishing or etching) makes it
possible to improve the breakout quality following scribing. This is explained
by the removal of the surface layer of the wafer strengthened by polishing. Thus,
the determination of the correct combination of scribing modes (speed, pressure,
point angle and setting of the cutter) as a function of the initial conditions
(wafer material, coating thickness, etc.) makes it possible to produce a scriber
groove of minimal width (no more than 5 um) and depth (2 to 3 um), which assures
effective and high quality separation while preserving the specified geometric
parameters of the chips. The scriber notches are made with minimal destruction
of the wafer surface layer. The stresses are concentrated in the scriber grooves
and are governed by the boundaries of the next break. '
Dt
t 2 D, - � S '
ki, .
s y a) (a) (b) 0 "Z
Figure 6.2. A device for stretch ten-
sioning scribed wafers.
a. Before stretching;
b. After stretching.
' = s Vacuum
J IaKyy,w
2
~ Cacamari
eaays
a) e ~ sComnyes'sed
g f Air
o. ~ � s ,
, o p
00
To separate scribed wafers into chips, (b) 6J
an installation for breaking semicon-
ductor wafers can be used which provides
for the oriented separation of scribed Figure 6.3. Device for placing chips in
wafers positioned between flexible a chip holder.
transparent viniprose films, one of
which is the satellite carrier. The breaking is accomplished by rolling a
spring loaded steel roller into mutually perpendicular directions, corresponding
to the scriber lines. The breaking force is uniformly distributed over the
generatrix of the roller, which prevents mechanical damage to the chips. Wafers
with diameters of 25 to 60 mm and thicknesses of 0.1 to 0.3 mm can be broken
apart in the unit, with chip dimensions of from 0.45 x 0.45 to 3 x 3 mm. The
breaking force is adjusted in a range of 50 to 118 N, depending on the wafer and
the chip dimensions. A study of the process of breaking scribed wafers showed
that the quality of a break depends to a significant extent on the roller diameter
and material of the substrate.
- 144 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
Following the breaking, the satellite carrier with the oriented chips secured
to it is subjected to stretching tension. The film tensioning unit provides for
uniform stretching of the satellite carrier, something which makes it possible
to create the optimal spacing between the chips for their subsequent oriented
loading. The basis for the tensioning (Figure 6.2) is the plastic flow of the
film carrier material with exposure to temperature: the satellite carrier 1
made of viniprose with the broken wafer 2 placed on it is caught in the clamping
device of the tensioning mechanism; the stretching of the film is accomplished
by stage 3 with heater 4 mounted in it, the temperature of which is monitored
and adjusted. The small stage is secured to the push rod of a pneumatic cylinder,
the construction of which makes it possible to adjust the speed of the working
and no load motions. Following of the completion of the stretching, the heater
is turned off and the small stage is cooled by a flow of water fed into its
interior cavity. Cooling is provided to_speed up the shaping of the carrier.
Following cooling down to a temperature of 50� C, the film with the chips which
is formed in the shape of a small cup 5 is removed from the clamping device.
The placement of the chips with a specified orientation in multicell cassette
holders is a necessary operation for the subsequent automation of assembly
processes [36]. A desk top unit for placing chips with dimensions of 0.5 x 0.5
mm in an annular cassette holder with 180 cells is depicted in Figure 6.3a. An
operator f ills a cassette holder in 5 to 10 minutes (depending on the operator's
skill) using such a device.
In the device, the semiconductor material wafer 1 is fed to the manipulator
stage 2. The wafer is broken into chips which are separated from each other
and placed on a transparent substrate. The operator orients each chip with
respect to the intersection, which is observed in microscope 4. The vacuum
suction attachment 3, which executes an up and down motion and with lever 5
moves from position I to position II (Figure 6.3b), is precisely positioned
relative to the intersection; the suction attachment places the chip in the cell
of cassette holder 8, after which the cassette holder is rotated through the
angle B.
The device has an electric motor drive 6, a system of drum program mechanisms,
pipes to create the vacuum in the suction attachment when transporting the chip
as well as to feed compressed air at the moment the chips are placed in the
cell of the holder.
To prevent possible vibration when the mechanisms are operattng, the entire unit
is mounted on plate 7. Holder 8, when being set on the stage of the unit, is
centered in the internal annular groove with the bands of the three wide bearings
9, one of which moves while the other two are stationary. The holder is rotated
through the angle B by virtue of the precise pitch of the teeth of the outer
gear ring and the pawl which fits in the slot between the teeth during locking.
The automated all-purpose EM-202A installation is widely used in semiconductor
production for separating a scribed semiconductor into individual chips. The
operating speed of travel of the breaking rollers is regulated in the unit, and
there are interchangeable breaking rollers of different diameters and
- 145 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500094404-3
FOR OFFICIAL USE ONLY
interchangeable holders for securing wafers of various diameters. The major
technical specifications of the ;init are:
Maximum diameter of a wafer which can be broken
up, mm
Diameter of the breaking rollers, mm
Breaking roller pressure, Newtons
Rate of travel of the breaking rollers (con-
tinuously adjustable), mm/sec
60
3, 6, 10, 12, 16, 20
0.49 - 49
0.8 - 6
The automated EM-436 unit is used to place ch ips in a multirow rectangular
cassette holder. The installation is designed to remove chips which are arranged
with a specific orientation on an adhesive strip, and place them in the cassette
holder where the number of positioned chips is determined beforehand by means
of a photoelectric sensor.
A Z
Figure 6.4. A laser production process
installation
Figure 6.5. Structural design of a
laser scriber which pre-
vents contamination of the
wafers.
-146-
Wafer thickness, mm Up to
0.25
Wafer diameter, mm Up to
60
P ulse repetition rate,
pul.ses/sec 12, 25,
50, 100
FOR OFFICIAL USE ONLY
A number of installations have been
designed for laser scribing and cutting
of semiconductor wafers, which differ in
terms of their composition, the kind of
laser used and structural design of the
various devices. A domestically produced
laser production process unit (Figure
6.4) makes it possible to produce clean
square edges of chips, a deeper groove
than even with diamond scribing while
sparing silicon wafer area (the number
of chips on a wafer is increased by 5%).
A pulsed yttrium aluminum garnet with
neodimium laser is used in the installa-
tion.
The main technical specif ications of the
production process laser installation
are:
Cutting speed, mm/min 60
Width of a cut with the
defect region, mm 0.12
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
In accordance with another technique (Figure 6.5) [38], the laser beam, passing
through optical system 1, scribes the wafer 4; the melted silicon collects in
droplets, which adhere to the plastic film 2. Since the film is fed from
roller 5 and wound on roller 3, the adhering droplets do not fall on the wafer.
Because of the fact that the film is transparent to the laser beam, it does not
create any obstacle to the scribing.
Specialized equipment complexes have been designed for the operations of separat-
ing wafers into chips. One of these complexes is the TAS-1000 system of the
Teledyne Company (U.S.) [39]. It must be noted that such systems are most appli-
cable where there is a sufficiently large chip production volume, i.e., in cases
where entire plants specialize in the production of chips and these chips are
then sent to other enterprises.
The widely used technique of removing the chips by means of a?neumatic suction
attachment has significant drawbacks: the very small force for pulling a chip
away from the carrier and the necessity of aealing the edges of the suction
attachment� and the surface of the ehips so that a vacuum capture is accomplished.
The necessity for sealing requires precision methods in the fabrication of the
tip pieces of the vacuum attachments. For this reason, it is of interest to
remove chips following scribing using electromechanical tweezers [41]. To pick
up a chip with electromechanical tweeZers, following the scribing the wafer is
placed on an adhesive elastic strip and broken up by a tool moving along the
scribing line which produces the breaking pressure. Then wafer 1(Figure 6.6)
is placed on prismatic knife 3, from which the chip 2 is removed by the jaws 4
of electromechanical tweezers 5. The capture is accomplished by virtue of the
compression of annular spring 6 of the electromechanical tweezers with the action
of solenoid 7. Spindle 8 of the chip positioner has the capability of executing
cyclical turns through a certainangle to accomplish the "biting" when removing
a chip and execute a reciprocating motion to transfer the removed chip to the
placement position.
An advantage of electromecnanical tweez-
ers is the capability of producing a
considerable clamping force on the chip
when grasping it and as a consequence,
~P producing a considerable force to pull
the chip away from the substrate.
6.2. Other Kinds of Equipment for
' . Separating Wafers into Chips
Automated machine tools with rapidly
rotating disks (up to 45,000 r.p.m.)
Figure 6.6. Schematic showing electro- made of diamond chips in a rubber-like
mechanical tweezers. binder. The disks are assembled into a
composite tool like an agricultural disk
(for group cutting), are easily cleaned and are less inclined to clogging with
silicon dust. When completely cutting through the wafers, there is no necessity
far a breaking operation. The width of the cut fluctuates in a range of 40 to
- 147 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504090044-3
FOR OFF'ICIAL USE ONLY
100 um. The cutting speed runs up to 30 cm/sec (when working with a single disk).
The region of the destroyed layer does not exceed 12 Um. There are machine tools
which make it possible to cut wafers apart with diameters of up to 127 mn. The
securing of a wafer in the case where the cut does not go all the way through is
accomplished by means of an adhesion carrier or by means of a glue-on label. With
the installation using several disks, several tracks are cut immediately in a
single pass, something which improves the productivity. The service life of a
diamond cutting disk reaches 28,000 cuts when cutting wafers with a diameter of
50 mm to a depth of 0.25 mm. Reinforced disks are used to cut ceramic substrates.
The domestically produced SRP machine tool for cutting wafers into chips using a
wire which forms a grid matching the dimensions of the chip has the following
cutting scheme. Two cutting devices with the wire wound on are mounted on moving
slides. This makes it possible to simultaneously cut two wafers of semiconductor
material glued to a glass aubstrate. The matching of the wire to the separating
track is accomplished through an auxiliary optical device. The working pressure
during cutting is produced hydraulically. After the wafer is cut apart in one
direction, the wafer carriers are rotated through 90�.
Chemical methods of separating wafers into chips using photolithography techniques
proved effective in some cases. These are especially applicable in those cases
where projections extend out beyond the chip, such as, for example, in some
integrated circuits with beam leads. In this case, the following technology
is employed to place the chips in an oriented manner on a flat disk. The front
side of the wafer is coated with wax, and the wafer is secured to a flat disk
about 0.9 mm thick. The back of the wafer is polished down to a thickness of
50 um and coated with photoresist. Then, the photoresist is removed at those
points where necessary by means of masking. Infrared illumination is used to
provide for matching during the masking. The region from which the protective
layer is removed has the form of a grid, positioned just underneath the beam
leads. Following photolithographic processing, the wafer is etched through at
the points unprotected by the photoresist.
- 148 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
CHAPTER SEVEN EQUIPMENT FOR ASSENNBLING SEMICONDUCTOR DEVICES
7.1. Methods of Aasembling the Major Types of Devices. Requirements Placed On
the Equipment
The major assembly operations in planar technology are the mounting of the chips
in a package and the connection of the leads. The chip mounting operation con-
sists in seating it in a specified position and securing the chip in the package.
The operation of connecting the leads, or the making of the connections between
elements consists in creating an electrical path between the contact pads on
the chip and the external leads. The requirements placed on the equipment are governed by the technical require-
ments placed on the installation of the chips [42]:
--The stability of the electrical and thermal characteristics of the device at
the maximum operating temperatures;
--The adherence to conditions which assure the permissible limits for mechanical
and electrical stresses at specified current and voltage levels, as well as
installation methods, temperature ranges for storage and operation and the
vibration and shock ranges.
At the present time, two methods of chip installation are widespread [43]:
--The connection of the plane of the chip to a contact pad of the substrate or
frame;
--Connection using the inverted chip technique.
The connection can be realized in the following ways:
--Eutectic joining with the formation of an alloy of the two metals;
--Soldering, where a third component is used to join the two metals;
--Gluing;
--Ultrasonic welding [40].
Eutectic joining is widely used when installing silicon chips on a gold plated
bonding pad.
The successful use of soldering depends on the capability of the solder to wet
the metals being joined together and to create strong joints with them. Drawbacks
to the soldering method are the necessity of fluxing the surfaces being ,joined,
and since fluxes car cause corrosion, careful removal of the flux following the
completion of the soldering.
Gluing of semiconductor chips is used to mount them on nonmetal sur.faces (ceramic,
plastic, etc.). Current conducting materials are sometimes used for the adhesive.
The eutectic fastening, soldering and gluing processes can be intensified by the
-149-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
FOR OFFICIAL USE ONLY
application of oscillations at various frequencies, all the way up to the ultra-
sonic. The major methods of fastening chips are given in Table 7.1.
The use of ultrasonics to speed up the process is most widely employed when con-
necting a chip using the inverted chip or "flip chip" method, the essence of
which consists in joining the bonding pads of the inverted chip to the bonding
pads of the base (substrate, package, contact frame) with the repeated melting
of the solder from which these pads are made [44]. The flip chip techniques
combine the operations of fastening the chips and connecting the leads. To make
contacts on specified pads, their height is raised up on the chip or package.
The height of such projections is about 25 um. It is possible to connect several
leads simultaneously to a single chip in one batch processing pass by means of
bead and beam projections.
Besides the methods enumerated above for batch assembly of IC's without wires
using beam and bead leads, there are two more to be singled out: the combination
flip-chip and beam lead method and the strip technique, or the contact frame
method. The latter is quite promising, since it makes it possible to use strip
automation technology for the assembly process. Spiderwed leads which are
stamped out or etched out from aluminum foil make the connections between elements
when making batch connections in ultrasonic welding installations.
The most widespread method of connecting silicon IC chips to package leads is the
welding of wire jumpers. The advantage of the technique consists in the fact
that it does not place stringent requirements on the geometric dimeisions; a
drawback is the difficulty of automation and the high cost. Moreover, wire jum-
pers are at times a reason for IC failure, since the lead often breaks at the
base where with the action of the welding head there is the strengthening of the
wire by work hardening, which is accompanied by the formation of microcracks.
Wire leads are fabricated primarily from aluminum and gold.
Thermal compression is the major method of connection, which provides for joining
two parts by means of heating and pressure. Thermal compression overlap welding
provides for a strong connection of semicondu:.tor materials to leads made of
gold, aluminum, silver and other malleable metals; butt joint welding can be done
only with gold. The combination of plastic deformation and diffusion leads to
a close interaction of the molecules of the parts being joined, as a result of
which there is adhesion, although the joining of the parts is accomplished with
heating up to a temperature below the melting point of the metals being joined.
Ultrasonic microwelding is used both for batch connection of leads and for making
wire connections. It must be noted that it is necessary in mass production to
very carefully select the ultrasonic welding conditions to assure stability of
ttie quality indicators for the execution of the process. There is information
that ultrasonic welding causes diffusion and recrystallization of the metals
without their melting or with local melting in the contact zone. The formation
of cumpounds is also ascribed to processes similar to friction welding. The
majority of specialists feel that ultrasonic oscillations of the cool in the
initial welding period destroy oxide films and provide for conract of the
- 150 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
~
i

o
o
-W
a~ iom

ma
o
ih
a-W
o
u 0 a
ar ea co
ao
a~ a
~o
o~
a>,
-H
w
,H oo o
V. a
ac
0 V �
~o .a a
p
ca
0
m
p co
Y+ w
6"+ o
a~ o0 3
d
u " o
u
-H V
co Ac 4
-4 10 u m
3
co
"H
~o
~
W
~
~

~
a
m
c
H
.
ca
~

+
o
i,

a
,~m
a
i
o
o
c
i
~
m
a
~ mwp
a~ u
N~
u 11. 3
r+r-I w Co
a
G
-i
~
-W
a

d
~
w,a ~a
1
a u
I o a
(A
GI (A
.7
fA C+
lJ � ~ 01 N
Q~
C~ $4 Ol
L! O.r,.
L.1
Cl 00 lC 9C
c0 'A .C v'1
p r~
~ W-W
41
t~
O N L+ -W 00
c0
U R N~
1 1~ a~ vl
0
N
tA
'd
'O
(d
co '1~
~
~ '1'~ P'~ ~
M
M+ '1'+ ~
~ � '
A
11 CO
'1'4
a! O" m'C
W
$4
u
W
w
u
,"i
m
3
q
o
o
+

,
c0
-H W'C 'C N
8
N 'C Ol
O 8 w rl
q
00 :3
> G!
00
0
a
d.0
~
~
u
g ~

�
-~i
~
x 0
o u~
4
c
o
H o
,
~
a
a
,
ui
m
0
�y
o
mw
~
�
8
a
o
H
,
r'
w
u
a
~
>1
~n
c
cA.

a
0
+
um
a
a c
v
r-i r-I o$+ u
r-i c
o a~ ~n
.-i 0
0 co
u-H Co
44 o0 00
$4 a)
,a co
0 w
co
a 3++
.c
u
a+1
or-i
co r. Z
cn
o
co 7-4
u
a
a0 owo
a+~ Y+
woN
a0
a
co
~.C -r+ o u
Cd u�>
0 ~4 P.
co p
o0 0
U N
u
u
m r 'b
.r+ CL LJ
W �w Q)
~
'1"+ r"'I
d
a~ a w o o
a Z -H u
m oo w
w
x
rc
~0
�ub
w0~
Ho

H
a
~u
i
$4
i
a
a
~
�
ai
n
L
H U
.7
N
to
~O
0
00
1
I
~
I
~
r
l N
O
O
O 00
O
~
M
N
N
~T
O 41
r+
3 ~d
~
10
w b
~
~
i
~
~
i
u
~
d~
c0
co ,H
co "1
(d v

U
:3
O VJ
O ul
W 0)
0 r-4
~

0.01010
u o
u o
+1 co
u
cn d
+1 d~q
a
v-+
N
.c
c: rl ar O
r4 N P
r-i N i-i
tA
~ 1+
ri)
~
~
~
3
~
~
~
~V
u
t
l1
i
a
i

a
D,
G1
W
tC w
ri H O w
Cl H R1
N F+ c0
w
O~
o~
cao
u oao
~or-I
~0 ~4
o
uq
w
v
o)
o
.c w
s
o i
.c
41 o a
a~
u
oo ~
d
~
~
w
o~~~
0
a
~
d ~
o c
.~i
3
)
a, u
4+ u 0 w
41 o

,J 0 1+
m
Cd 44, a o
u u a u
"
uP-4 d
o
cu oo a
0 4 >,+1
-r+ r-I
r+
0 4+ u
u
,H c~
u u u u
�o ,a cc
3w
b co u0
Y+
c
W �ri v~
:3
93.~
d
A cU t0
f3
~
(1) 'b
'ti 13 'C 'b
O u W N
0 'C
04 +1 41
c0
4.+ G~
rl --1 r-1 C:
U"4 tA .14
O r-I
U 41 A N
1+
N
R7 O.G
O O O
O c~
44 O
*i O~rl
N
c~ rl oo u
d 3 a e
+1 m
d 3 u v,
c~
H

w 00
~
� 0
i
~
~
-H
� -r-I
~ t
A
u
O
b
O H~
N
ra
.C ~
C
l
'
'
~
0
~
~
~ +
W ~
V
1
v
U
Ch
- 151 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFF1CIAt, l1SF. ONI.Y
"juvenile" surfa-es. Soldering is used when fabricating high power planar tran-
sistors and alloy devices, while electron beam and laser welding have a small area
of applications at the present time.
The equipment for mounting chips on a bonding pad and for connecting the leads
has many similar structural and design elements, especially in the loading, unload-
ing and transport devices for the products during the process of executing the
operations, visual observation, etc.
The assembly operations when manufacturing mass produced alloy semiconductor
devices can be broken down into the following types: the assembly of the device
components in multiple cell cassette holders; the attachment of the chip holder
to the transition connection; the assembly of the transition connection and its
installation in the device package; the hermetic sealing of the device.
The assembly of the elements of a device in multiple cell holders is used to
improve the productivity of the operations performed by the batch method, for
example, such as melting, group soldering, etc. The connection of the chip holder
to the transition connection and the mounting of the transition connection and
its installation in the device package are the most labor production process
operations, which require high skill on the part of the operators and high pre-
cision actuating mechanisms.
Thermal compression and ultrasonic welding are widely used in the production of
alloy semiconductor devices to join the fine wire leads to the interconnection
electrodes and external leads, as well as in the production of planar devices
[451. The heat is delivered to the weld region in the following ways: by heat-
ing the mounting base of the device; direct heating of a needle or punch; indirect
heating of the weld region, by passing current through the tool; heating a needle
with simultaneous heating of the mounting base.
7.2. Equipment for Mounting the Chips of Planar Devices
With respect to the kind of product loading and unloading during mounting, a
distinction is drawn between equipment with manual piece by piece loading and
unloading and with automated loading and unloading mechanisms.
With respect to the type of feed of the chips for fastening, a distinction is
drawn between various equipment designs: with cassette holder feed of chips put
in them beforehand; with combined selection of chips suitable for attachment;
and when they are fed in as wafers, separated into chips while preserving their
orientation.
WiLh respect to the type of feed for the solder tablets or gold liners (in the
case of eutectic joining), a distinction is drawn between automated equipment
and equipment with manual loading. The semiautomatic PUN-700 unit for soldering
chips to the mounting base, which automatically sequentially joins the gold liner
and the chips has the following technical specifications:
- 152 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00854R000500090004-3
Productivity (depending on the chip size),
pieces/hr i 700
Size of chips which can be joined, mm 0.5 x 0.5 +
1.5 x 1.5
Connection temperature, �C e 460
Precision in coanecting the chips relative to
the center of the gold plated pad, mm + 0.2
Range of connection making times, sec 0.4 + 3.9
Range of working frequencies of the ultrasonic
generator, KHz < 60 + 5
Output power of the ultrasonic generator, watts 16 + 1
Control range for the tool pressure on the elements
being welded, Newtons 0�2 - 2
Figure 7.1. The PUN-700 semiautomated unit for sealing chips to
mounting bases.
The soldering of the chips of transistors with a metal-glass package and flexible
leads can be accomplished on the PUN-700 unit. The functions of the operator
when working with the semi-automatic unit reduce to periodically changing the
-153-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
ROR OFFICIAL USE ONI.Y
cylindrical cases with the belt carriers. In structural terms, the semi-automatic
unit consists of the following structural parts (Figure 7.1): the machine tool
bed S; the grab type feeder for the products 1; the chip 3oiner and feed device
3; the foil feed mechanism 2 and the control unit 4.
J
Figure 7.2. The grab type feed device for the products.
The bed takes the form of a welded structure. A plate is fastened on top of the
bed, on which the major assemblies of the semiautomatic unit are mounted.
The grab type feed device for the products (Figure 7.2) serves for the step feed
of the belt carrier with the mounting bases of the devices placed on it. The
device consists of the guides 1, which are fastened to two bases 2, the drawbar 3,
which joins the base and which imparts rigidity to the eystem; rods 5, which
hold the magazine. The entire system is suspended freely on the springs 4 from
the frame and is driven by a lever device.
Figure 7.3. The feed and attachment mechanism for the chips.
The chip feed and attachment mechanism (Figure 7.3) consists of two assemblies in
structural terms: the feed mechanism and the machanism for attaching the chips.
Thc transport mechanism 4 serves to transport the cassette holders with the chips
oriented on them to the position for grasping the chip by the working tool. The
- 154 -
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
mechanism consists of the following components: the bracket 5, platform 6, which
is rigidly fastened to the platforms; the cassette holders with the chips 7,
which lies freely on the platform and takes the form of a ring with teeth about
the periphery, where the chips are placed in nests in an oriented manner, where
the number of nests corresponds to the number of teeth; disk 8, which presses
against the cassette holder and can move in the ball guides along the platform.
The step motion of a cassette holder is produced by a rock.ing lever, moving
clamp 9 and finger 10. The mechanism for feeding and attaching the chips 11
consists of the carriage 1, which moves in the ball guides on the base with the
- magnetostrictive transducer secured to it, where this transducer has a vacuum
holder. Both mechanisms are mounted on the common base 2 and are driven by a
single drive 3.
The foil feed mechanism provides for feeding, cutting off and placing the gold
tablet on the mounting base and consists of two assemblies: the foil feed and
cut off assembly; the gold liner placement and connection assembly.
The latter assembly has a structural design and performs functions similar to
the chip feed and attachment mechanism (Figure 7.3). The foil feed and cut off
assembly has a spool with the gold foil wound on it, the feed of which is accom-
plished by two small rollers and the cutting is accomplished by a rocking knife.
The intermittent rotation of the rollers is realized by a ratchet mechanism. The
motion of the knife and the ratchet mechanism is accomplished from the cams of
the common drive for the mechanisms. The chip feed and attachment mechanism is
coupled by means of an elastic sleeve to the mechanism for placing and cutting
off the foil.
. _ Ge
Figure 7.4. Kinematic schematic of the ultrasonic attachment automat.
The mounting basea of a device are heated by tunnel heaters with built-in heating
elements.
-155-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
In accordance with the kinematic schematic (Figure 7.4), the operation is realized
as follows: electric motor 28 rotates the camshaft 25 through the elastic coup-
ling sleeve 27. The grab mechanism is moved in the vertical plane by cams 23 and
26, as well as by spring 4 through lever 29, the system of levers and connecting
rod 11; the grab mechanism is moved in the horizontal direction by cam 24, spring
30 and lever 21.
The belt carrier 32, which is filled with the mounting bases of the devices, is
moved by the grab mechanism in vertical and horizontal planes and clamped by
catches 33. The carrier moves from the cylindrical case 31 along the plate 41
to the position for the attachment of the foil and chips, where heater 42 is
located. Then the carrier goes into case 3, from which it can be transmitted to
the next production process operation.
Toothed belt 2 transmits the rotation from electric motor 1 to camshaft 46. Cam
34 transmits an intermittent motion through the lever and connecting rod 12 to
the ratchet 15, which is clamped by finger 14. The rotation is further communi-
cated through shaft 17 to disk 16, which feeds the gol.d foil from roller 13 to
the stationary guides 18. The foil is cut off by knife 19, which is driven by
cam 35, the lever and connecting rod 20. The cut-off foil (the gold liner) is
caught by vacuum suction attachment 44 and by means of cams 38 and 39 as well as
springs 43 and 40, a combined oblique motion with the gold liner to the mounting
base of the device is imparted through connecting pull rod 37 and lever 39 to
the suction attachment. After making the connection, the operational cycle is
repeated. Holder 7 with the chips is set on the stage 6 and is pressed by disk
8 and spring 9 against the stationary clamp 5. A stepped motion is imparted to
the cassette holder by lever 10, which receives its reciprocating motion through
the system of connecting pull rods 56 from cam 45 with both end face and radial
working profiles. Lever 10 moves the cassette holder out from stationary clamp
S and rotates it through one step. When the lever moves out to the initial
position, the cassette holder is clamped. The capture and transfer of the
oriented chips from the cassette holder are accomplished by vacuum attachment
54, to which a combined oblique motion is imparted by cams 38 and 48 as well as
springs 47 and 55 through connecting pull rod 49 and lever 51. The pressure of
the vacuum suction attachment on the chips during capture and when making the
connection is regulated by spring loaded levers 52 and 53. Programming devices
50 and 22 are mounted on the camshafts to synchronize the operating cycle of
the semiautomatic equipment.
A pravision is made in the semiautomatic unit for local feed of an inert gas into
the region of chip 4'achment.
The loading and unloading of the belt carriers in the cylindrical cases are accom-
plished without shutting down the semiautomatic unit; a stockpile of semi-finished
products can be stored in special magazines, the dimensions of which make it
possible to place them in the pedestals of the machine bed.
During operation of the semiautomatic unit, attention is
operation of the vacuum attachments, which can be fouled
of contamination of the mounting bases of the devices or
chips, a reduction in the connection quality is possible.
-156-
r ' .3, OFFICIAL USE ONLY
to be given to the
or wear. In the case
the back side of the
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
Depending on the combination of the material of bead leads of chips and the bond-
ing pads of substrates for integrated circuits, the following methods can be used
to make connections in the EM-431 installation:
--Thermal compression microwelding or soldering with the capability of constant
heating of the substrate and cHip (the working stage and tool), as well as
pulsed heating of the chip (or the tiol); --Combination microwelding or soldering (ultrasound with pulsed heating).
Figure 7.5. The assembly for the
placement of chips in
the EM-431 unit.
The assembly for placing chips of the EM-431
installation is depicted in Figure 7.5. An
integrated circuit substrate 2 is placed on
the working stage 1, which is equipped with
a spring holder. The chips are fed into the
unit in ring cassette holder 3 having a
capacity of 180 chips. In contrast to the
structural designs of the chip feed assembly
described previously (see Figure 7.3), the
ring cassette holder is mounted coaxially with
the working stage. The spindle 4 with the
tool 5 executes reciprocating motions from
the center of the working stage to the peri-
phery of the cassette holder and back.
The operator matches the chip up with the bonding pads of the substrates by means
of a micromanipulator and an optical system, consisting of a microscope and
semitransparent and reflecting mirrors. The installation can be put together
with two types of interchangeable working stages: one is heated; batch instal-
lation of substrates is possible on the other.
The structural design of the MS-602-1 semiautomatic unit for the attachment of
chips is of interest [46]. The semiautomatic unit is used in a set of equipment
intended for assembling multiple chip hybrid integrated circuits. The attach-
ment of chips with tinned leads-stubs on sitall and ceramic substrates is
accomplished by means of pulsed heating of the working tool with additional heat-
ing of the working stage while applying ultrasonic vibrations. Crystals of a
particular type (with dimensions of from 0.7 x 0.7 up to-1.8 x 1.8 mm) are placed
in 6 vertical holders with a capacity of 350 to 400 chips each. Selective auto-
mated selection of the chips from any holder is possible in accordance with a
program set on the control console. The specific features of the operation of
the semiautomatic unit are the automatic orientation of the chip on the working
tool and the matching of the bonding pads of the substrate to reference marks on
the screen of the projector. The dimensional tolerance of a chip should be kept
within + SO um for high quality mounting. The fabrication of the cassette holders
in the form of tubes, in the cavity of which the chips are stacked make it possi-
ble to create more compact feed assemblies than when using the cassette holders
shown in Figure 7.5. The flip chip method of attachment requires the use of
special devices for the observation of the matching of the contact projections
of the chip and the substrate [38]. Several methods of matching are well known:
using semitransparent and reflective mirrors (Figure 7.6).
- 157 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
~
~
z
~
X/f
.t k ~
M
FOR OFFICIAL USE ONLY
a) (a)
q
S
61 (b)
Figure 7.6. Schematic showing the matching of chips to the substrate.
When a semi-transparent mirror is used (Figure 7.6a), the weld.ing tool, which is
simultaneously the vacuum suction attachment, picks up the chip 2, which is
matched to the figure on the substrate 5 by means of semi-transparent mirror 4.
Then the transparent holder 3 is shifted in a horizontal direction, while tool 1
places the chip at the specified point on the substrate with a vertical motion.
Such a method of matching requires that the chip always be located above the sub-
strate.
When using a reflecting mirror (Figure, 7.6b) , the matching of the chip 2 to the
figure on the substrate is accomplished using the microscope crosshairs and using
the imaginary image of the chip 5. The end face of the tool 1, which has a
suction attachment for the chip, is aligned and guided in accordance with the
same crosshair lines. The chip is arranged on the moving plate 3 of the reflec-
tive mirror 4 below the projections.
Following matching, the chip is held by suction, the plate 3 is removed, the weld-
ing tool moves it and the attachment is made.
A needle with a special geometry having a central opening to produce a vacuum
over the chip can be used as the tool for catching and attaching the chips. The
needle is made of a solid tungsten carbide alloy or special steels.
7.3. Equipment for the Attaching of Leads to Planar Devices
1~,ro types of equipment exist in accordance with the technology for lead atCachment:
for wire mounting, and mounting without wire. The equipment for mounting without
wire as a rule, uses the principle of batch attachment of the leads to the chip;
some models of this equipment were described in the preceding section. The
structural design of installations for wire mounting are treated in the following.
- 158 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
TABLE 7.2
The Characteristics of a
Thermal Compression Bond
Wire material
Wire diameter, um
Bonding area (gold lead with
diameter of 25 um), um
Substrate temperature, �C
Method of cutting off the
wire
Output, welds/hr
Method
of Bonding
Butt with a
Overlap with a
Overlap with a
"Bead"
"Wedge"
"Bird's Beak"
Au
Au, A].
Au, A1
10 - 250
10 - 250
7.5 and more
Diameter
Diameter
90
40
50 x 100
250 - 400
250 - 400
250 - 400
Hydrogen
Moving the
Moving the
flame
wedge; small
weoge; small
knife
knife
= 2000
< 1000
= 2000
The types of thermal compression welding and the tool used in this case are given
in Table 7.2. Thermal compression overlap welding provides for a strong joint
between semiconductor materials and leads made of gold, aluminum, silver and
other malleable metals; butt welding can only be done with gold. The heating
temperature in the case of thermal compression welding should not exceed the
temperature for the formation of eutectic alloys of the materials being welded
and should not lead to the formation of dislocations�. The heat is delivered
to the weld region in the following ways:
--Heating the mounting base of the device;
--Direct heating of the needle or wedge;
--Indirect heating of the weld site, by passing current pulses through the tool;
--Heating the needle with simultaneous heating of the mounting base.
The characteristics of a few methods of thermal compression attachment are given
in Table 7.3.
Besides the kinds of thermal compression welding indicated in Table 7.2, combina-
tion methods are being widely used of late [47]. The execution of the process
using two methods is shown in the schematic (Figure 7.7): butt welding of gold
wire to the bonding pad of a device and welding by bonding to the bonding pad of
the substrate.
Ultrasonic bonding of wire leads is accomplished in two main ways at the present
time: using a wedge with an obliquely positioned opening (Figure 7.8) and with
a capillary. Steps in the process of bonding wire leads using a wedge tool are
depicted in Figure 7.8: 1
--The lead wire 1 is fed to the point of its bonding to the chip 2 and clamped
by clamp 4 (Figure 7.8a);
- 159 -
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000540090044-3
FOR OFFICIAL USE ONLY
Bead formation
19paao/arire
rrapoa[
3awomQA
npo0uoff
~old wire
Noecww.v:`� `~'~m~e�Sd
1 e0apirsQw
( ~ +owe0~a
(7) Om1911#99
c,vemvy~+esma
(6)
nvasQo.s 4
r~qr~t (8)
I
1 OQjaao4asas
t roi0a~iso~~nass
� m~snR~
. 0) IwiXu
/000000/II ~~~Q ~
�aos,,ra ( ,5 ) ai rwatn ( 9 )
(3) Claoza � ~po0owi~ry ' I1,ven~pqMi~in ~
�t/17/.M'AON noOsoMe"nae.
p~spu0a~
lapaso~rmaw~NOe n~ e~IC(sMat npOClwi~p �
(4) oMemppNisma o~~jysm r acmasisw .
ooeDa,rs~as maea.e~nsaaR" N1911Am18
RoNrrR
(11) '
A'i,v,n~,m~~i A~oaraO~Q ToNmo,~msaa ~eaa(aD,~[ ( lO~
npr0ipa rrOKOQ e,~tNM
Figure 7.7. Schematic showing the combination weld bonding of leads.
Key: 1. Capillary welding head;
2. Flame from a hydrogen burner;
3. "Seam" welding;
4. Horizontal motion of the tool forms the "seam" type connec-
tion;
5. The collet chuck clamps the wire;
. 6. Welding of the bead;
7. Withdrawal of the tool;
8. Formation of a"nail liead" type bond;
9. The tool is raised, breaking the wire and leaving a small
end;
10. Bonding pad of a flexible circuit;
11. Bonding pad of the device.
--The lead is welded on (Figure 7.8b);
--The tool 3 is brought to the end face of the cross piece of the lead of the
device 5, the second weld is made, the lead wire is clamped by clamp 4 and
broken off (Figure 7.8c). The following steps in the process of ultrasonic welding using a capillary are
differentiated:
--The attachment of the lead to the chip;
--The attachment to the end face of the cross piece;
---t;utting the wire and simultaneously forming the "whisker".
-160-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
~
2 y ;
a~ (a)
6)(b)
~
f
aI (c) I,
Figure 7.8. Steps in the process of
bonding wire leads using
a wedge tool.
The PW-0.8 semiautomatic unit for the
ultrasonic overlap bonding of leads to
the components of devices assembled on
TO-18, TO-5, etc. mounting bases uses the
latter method. The semiautomatic unit has
assemblies for loading, unloading and trans-
porting the devices, similar to those
used in the semiautomatic unit for attach-
ing chips (see Figure 7.2). All of the
operations with the exception of the load-
ing and unloading of the strip carriers
and the matching of the leads to the bond-
ing pads are performed automatically.
The position where the bonds are made is
shown in Figure 7.9. The wire is fed
from spool 1 into the weld capillary 2,
which is secured to the welding head 3.
The strip carrier 4 with the products
secured to it can be fed out in a stepped
fashion. Following welding at the end
faces of the cross piece, the wire lead is
cut off by knives 5.
Figure 7.9. The lead bonding position of the PW-0.8 semiautomatic unit.
Brief Technical Specifications for the PUV-0.8 Semiautomatic Unit
Output, welds/hr 1,000
Diameter of the wire leads, um 25 - 125
Welding time, seconds
Tool pressure on the elements being welded, N
- 161 -
FOR OFF[CIAL USE ONLY
0.05 - 1
0.1 - 1.5
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAI. USE ON1.1'
Travel of the manipulator in the horizontal plane, mm
Precision in setting the manipulator, um
Travel of the tool in the vertical plane, mm
8 x 8
+ 5
10
A provision is made for feeding an inert gas into the region of lead bonding.
The NPV unit for the ultra$onic bonding of leads makes the connection using a
wedge, is of a desk top design, (Figure 7.10) and has the following technical
specifications:
Kinematic productivity, cycles/hr 3,000
rfaterials which can be welded gold, aluminum
Wire lead diameter, um 24 - 60
Welding tool travel automatic
Tool pressure on the elements being welded, N 0.1 - 0.15
Planipulator travel in the horizontal plane along the
X and Y axes, mm 6 x 5
The forming of the jumpers automatic
The strip carrier feed stepwise,
discrete
The structural design of the unit is executed as follows (Figure 7.10). The
welding head 2 is secured with a screw in bracket fashion to base 3; a camshaft
is mounted inside the head with a drive, a welding mechanism 8, a lever system,
an electrical panel 4 and microscope 7. Ultrasonic generator 1 is mounted on
[he welding head. The manipulator 6 is installed inside the table. The worl.1i,
stage 5 is installed in the upper plate of the manipulator. The fastening of
the camshaft and programming unit makes it possible to easily disassemble them
for technical servicing.
Figure 7.10. Overall view of the NPV-1
unit for the ultrasonic
bonding of leads.
In accordance with the kinematic scheri,% c
(Figure 7.11), the unit operates as fol-
lows: by pressing the "start" button on
control 24 of manipulator 23, electric
motor 1 is started. The rotation is
transmitted from the electric motor
through the V-belt drive 46 Co the shaft
with cams 39, 40, 43 and 44. The tool
which is fastened to chassis 9 is lowered
to the first and second weld positions
respectively by means of cams 32 and 40
through levers 41 and 42, plate 34 and
rod 30. The tool lift is executed with
the action of spring 29 and is limited by
device 28. The tool executes a complex
motion in conjunction with the cllassis
- 162 -
FOR OFFICZAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
Figure 7.11. Kinematic schematic of the NPV-1 unit.
9 and bracket 8. The bracket 8, with Che chassis 9 which is fastened in it in
a hinged fashion, are drawn up to the stationary plate by spring 33 and have the
capability of moving back and forth by means of former plate 5, which is secured
with screw 4, and lever 45 and cam 44, rotating relative to roller 32. The wire
being welded is fed and broken off by cramps 16, which are given the following
motions: for tearing the wire, a feeding motion from electromagnets 11, mounted
on the bracket, which moves together with the chassis 9; when clamping the wire,
it is driven by electromagnet 31. The height of the wire lead when finishing a
weld is limited by means of cam 39 through lever 36 with an adjusting screw.
Screws 36 and 37 are used to adjust the height for the first and seconds welds,
while screws 12 set the amount of travel of the cramps 16 when breaking and
feeding the wire. The load on the elements being welded is set by weights 6 and
7, where weight 7 ad3usts the pressing force only when welding on a chip. The
strip carrier moves when the manipulator handle is inclined. In this case, the
electric motor of drive 19 is turned on. Rod 25 with the dog which catches the
strip carrier and moves it by one step is driven in motion by cam 18 through
lever 22. The strip carrier is clamped in the welding position by levers 20 by
means of cam 21, while the return travel of the rod with the dog is accomplished
by spring 26; in this case, the dog slips freely along the strip. The adjusting
devices 2, 3, 27 and 35 serve to align the corresponding assemblies.
The welding mechanism consists of converter 15, which is mounted on chassis 9 by
means of device 14; bracket 8; the wire feed and breaking device, which consists
of bracket 13, the plate for the electromagnets 11, the U-shaped retainer 16,
chassis 9 in which the spool with the wire 47 is placed as well as the shaft with
the load 6.
-163-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
FOR OFF[C1AL USE ONLY
The mechanism operates as follows: bracket 8 is kinematically coupled through
master form plate 5 to the lever system and has the capability of rotating about
the support stops 10, and moving in a vertical plane along with bracket 13. The
lead wire is unreeled from spool 47, passes through the opening in U-shaped brace
16 and is fed under the tool of transducer 17. When a voltage is applied to the
lower electromagnet 11, clamp 16 compresses and clamps the wire. Then the voltage
is applied to the upper electromagnet 11, and the plates which are coupled to
the U-shaped retainer 16 are pulled towards its core, the U-shaped holder rotates
counter-clockwise and breaks the wire off. The stage for securing the strip
carrier or product can be made as a driven unit: for assembling devices in metal-
glass and plastic packages, or without a drive: for the direct fastening of a
product package.
In the process of executing the operation, the stage can be manually rotated
through an angle of up to 360�. When strip caXriers are used, the unit can
operate in a single production process line with aemiautomatic units for bonding
the chips.
Automated installations using minicomputers have been developed for the thermal
compression bonding of leads [48, 49]. One of them is designed for the mounting
of LSI circuits having a maximum of 42 leads. The LSI chips are mounted in a
framework of leads, which are fed into the unit in the form of strips (tape
segments). The receiving magazine holds 40 such strips and in step with their
use, they are moved to another magazine. The frames for the leads are fixed
precisely on pins in the guide racks. The operator corrects the shifting of the
center of the chip and frame by means of an electrical device, where this shift
runs up to 0.5 mm along the X and V axes with an angular shift of up to 5�. The
operator views the chip being mounted on a television screen and by means of
two controls orients the edges of perpendicular pairs of bonding pads with res-
pect to the crossing lines on the screen. One control makes it possible to
rotate the bracket with the frame for the leads fastened in it through a maximum
of + 5� to correct the angular shift, while the other changes the position of
the thermal compression head by + 0.5 mm in the direction of the X and Y axes to
correct the shifts along these axes. The vidicon which serves for the determina-
tion of the chip position is attached to the thermal compression head. The
controller generates signals proportional to the mechanical displacements and
feeds them to the specialized computer of the system. Calculations are performed
using the data on the position of the bonding pads, which are stored in a pro-
grammable read-only memory, to determine the corrections along the X and V coor-
dinates. This makes it possible for the microcomputer to feed out the corrected
X and V coordinates for the welding points to step motors as signals for the
positioning of the thermal compression head with respect to the X and V coordi-
nates in steps of 10 um. No angular motions are required. Gold wires for the
leads with diameters of 20 to 30 um are bonded by the thermal compression head
with a capillary using a butt joint at a rate of 2 leads per seconds; up to 3
chips with 42 leads are install.3d per minute.
A great advantage of programmable units for attaching wire leads, including those
using computers, is the capability of setting up multiple machine tool servicing,
- 164 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
where one operator works several production procese machines simultaneously, some-
thing which achieves a reduction in the labor intensity of product manufacturing.
j ~ o i6c
~ lacnisMenue
i
1
r
~ .ti � ~
I ~~..ary
1
(2) NS00 /
drtnonNenae
a)
(a)
06 ~

a.
y I
l,SJSO,af ~ ~
~
~
.
M
\
y,~A I
(b)
(3) Paawepa, YRM
Q� Id, I d, I b I R
I
L. YY
CL3; %J1 - - PM3MlpY, NKl1
t ~
ai I � I b le,s 10
(4)
28f3
88
i IafS
20
20
7t0.3
20 23t2,5
c~3t2
10rtt 2
ao
40
38t4
113
150~�5
25
25
(nepedfi
sapxaxr)
25 28t3.0
5) 38-05
70f2
98t3
107t2
15,1t3
40
48t5
138
182t:
30
30
12t0.8
30 - 48:0,0
105t4
165t4
40
40
60t5
180
238f7
40
40
(sropoR
40 70t6,0
138t5
2 1 0t6
60
B8PN8HT)
100t8,0
245t8
300t8
70t5
220
293,-t:7
50
50
(6)
50 140It8,0
3.50t10
420t10
90
95
i
170t 10,0
420t 12
SOOt 10
taot12,0
440t13
515t13
ioo
Figure 7.12. The main dimensions of hard alloy needles.
Key: 1.
2.
3.
4.
5.
6.
Design I;
Design II;
Dimensions, um;
Wire diameter, Um;
First variant;
Second variant.

18 '
20
25
30
50 .
70
100
120
125
The series produced hard alloy microwelding tool has a complex microprofile for
the contacting working surface, a capillary opening with a diameter of 0.02 to
0.1 mn and is usually fabricated from metal-ceramic hard alloys of the VK-8
tungsten group. The capillary holes are produced by using the techniques of
sintering a hard powder alloy using a metal form into which a suspension of pow-
der and filler is poured under a slight preseure.
-165-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
FOR OF'FICIAL USE ONLY
The basic dimensional parameters of hard alloy needles used in lead bonding'are
shown in Figure 7.12: Figure 7.12a shows a cutting alloy capillary; Figure 7.12b
shows a wedge with a lateral opening.
7.4. Equipment and Complexes for Mounting Semiconductor Devices and Integrated
Circuits on a Strip Conveyor
i2 ~ Strip
/1eMma
-m-Xpucraen
Chip
(3) Strip
AtNm�
~ Qdeed
Lead
(4) /ltNma
vLa(1L~[1 Strip
0
Figure 7.13. The typical sequence of
operations to assemble
diodes using a strip.
Key: 1. Scribing and separation
of the wafers into chips;
2. Attachment of the chips
to the strip;
3. Bonding of the lead to
the strip and the chip;
4. Cutting the strip on the
chip side;
5. Hermetic sealing of the
p-n junctions;
6. Potting or compression
molding the p-n ,junctions;
7. Cutting the strip and se-
parating the diodes.
- 166 -
One of the ways of automating the
assembly of semiconductor devices and
IC's at the present time is the use
of a strip conveyor. Equipment is
described below for the installation
of transistors and IC's on a continu-
ous perforated strip.
A typical sequence for the performance
of assembly operations for discrete
devices (transistors, diodes) using a
strip carrier is shown in Figure 7.13.
The "Potok" comprehensively mechaniied
line [50] finds application in the
production of a mass produced high
frequency transistor for home enter-
tainment electronics equipment (the
KT-315 transistor). A continuous
perforated strip which is simultaneous-
ly also the structural component of
the transistor itself and the means of
transporting it during assembly is
used.
The use of a Fernico strip with partial
local striped gold plating, accomp-
lished by a continuous cladding tech-
nique, has made it possible to attacli
chips directly to the strip, without
using additional solder tablets.
The metalized pads for the emitter
and base leads of transistor struc-
tures are made in the form of two
concentric circles, something which
has made it possible to eliminate the
orienting of the chips relative to
the external leads during the assembly
process and to make the bond (thermal
compression) of the two leads simul-
taneously.
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
Figure 7.14. The KT315 mechanized flow line for the production of
transistors.
The mechanized flow line consists of units, a portion of which are shown in
Figure 7.14:
--The automated press for cutting out the perforations in the strip (Figure 7.14a).
It has a nominal force of 104 Newtons, and a unit output of 6,000 transistors
per hour. Several lines can be serviced;
--The unit for attaching the chips to the strip (Figure 7.14b). It has a machine
output of 600 connections/hour. A vibrational hopper with a device for iden-
tifying and orienting the side of the chip by a probe head is used for the
automated chip feed. The soldering of the chips is accomplished at a tempera-
ture of 400 to 420� C using ultrasonic generators with a power of 40 watts;
--The unit for attaching leads (Figure 7.14c). A double ruby capillary is used
in it for the simultaneous connection of two leads (two units in a complete set).
The thermal compression is realized using beads at a temperature of 300 to 36'0�
C. The unit performs two operations. The first operation is thermal compres-
sion of two gold wires, which terminate in beads, to the metalized pads on a
chip; in this case, all of the transitions in the operation are made automatic-
ally; only the precise matching is accomplished manually under a microscope.
The second operation is the automatic connection to external current connectors;
the operation is performed by means of resistance welding without the partici-
pation of the operator. The machine capacity is 300 to 350 devices per hour;
--The unit for preparing for hermetic sealing (Figure 7.14d). It has two working
positions: for the unreeling of the strip with the transistor assemblies from
- 167 -
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2047/02109: CIA-RDP82-00850R000504090004-3
FOR OFFICIAL USE ONLY
a magnetic drum and loading into a satellite cassette holder for subsequent her-
metic sealing; and for cutting off one of the jumpers on the strip. The capa-
city of one satellite holder is 20 devices. The output of the unit is 700
devices per hour;
--The hermetic sealing unit. It takes the form of a piston metering dispenser
for epoxy compound, which is apportioned by 20 nozzles simultaneously. The
working stage of the unit has a mechanism for automatically moving the potting
f orms by one regular step; the shifting is accomplished after each dosage;
--The KT-2-12 semiautomatic classifier. It is used for classifying the transis-
tors; its feed and connection devices have been modernized for the flat plastic
transistor package;
--The automatic marking unit. It is equipped with a vibrating feed hopper dnd a
spiral volute chamber type drying chamber. The machine output of the automatic
unit is 3,000 devices per hour;
--The devices are packaged in a polyethylene strip in two parallel rows. There
is a vibrating feed hopper. The output of this automatic unit is 3,000 devices
per hour;
--The magnetic storage drum is intended for accumulating the taFe in the individual
assembly operations and the subsequent transfer of the assemblies on the strip
tape to the next production process operation. The drum is made of Fernico and
bar magnets, which make it possible to heat treat the devices in an oxygen
atmosphere.
For the batch assembly of IC's without wires using the contact frame technique,
a set of equipment is employed in which the method of preliminary embossing of
aluminum foil to produce the outline of the leads with subsequent etching out of
the jumpers is used to produce the contact frames.
Some of the technical questions related to the use of strip carriers, in particu-
lar, additional information on the use of strip technology, known under the name
"Mi-ni-Mod" are described in [51]. This technique, Just as the variants of it,
is based on the use of a plastic tape, reminiscent of a motion picture film,
over the entire surface of which frames with leads produced by photolithography
are arranged. A polyimide tape cladded with copper foil is most widely used for
th is purpose. The assembly of the IC's consists in the batch attachment of the
chips, made with bead or beam leads, to the external leads of the frame. The
chip is held above a small window in the plate by means of the internal leads,
bonded by a thermal compression head to the bonding pads of the chip. Two narrow
rectangular gaps make it possible to stamp out the exterior leads, removing the
unnecessary edges of the film in this case. Then, these leads can be connected
by the batch method to the housing or to the printed circuit board. The perforated
holes at the edges of the films serve to move the f ilm following the bonding of
the chip and for the precise setting of the frame with the leads underneath the
chip.
The typical process sequence for the group bonding of leads is shown in Figure
7.15. Prior to the start of the welding, the strip is lowered so that the
internal leads formed beforehand match up with the bonding pads of the chip,
-168-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
(2) npedeapxr,mo Noqw&.-
Iarms-
rarM~u / IN~IIAPLNN4L an~MR~t (3 )
" krloau
.
CNor�p1lm
(a) a) A',~u:ma.ew 44moAa a,e0 .pucmQo,.oe ( 5
Du~oJer, Kacoioruaect Ma+maM-
GaS mnuz Qacmyno (6)
: : .
(b) 61 (8)
Ceau ` rl
r~/
(C)Ql
Figure 7.15. A typical process for bond-
leads by the group method.
Key: 1. Contact projection;
2. Preformed internal
leads;
3. Guide films;
4. Chip;
5. Support for the chips;
6. Leads which touch the
contact projections;
7. Layer of wax;
8. Current;
9. Bonded chip;
10. Chip in the matrix;
11. Direction of chip
motion.
which is =lued with wax to the sub-
strate (Figure 7.15a). The chip,
which is observed in a microscope, is
` 'y ' set in the requisite position either
(d)z) 1\12 ni,uQape~~Nari (9) manually or sutomatically. Then the
"P"ema~'� strip is lowered, bringing the internal
~
(10) ~ leads in contact with the contact pro-
Kprrcman ~�orPuae ~ jections of the chip (Figure 7.15b).
~A stream of inert gas protects the weld
~
(ed) (11) ,qy,T,#dqeW�� dsumarup ~uucmaena Site. The welding head is lowered,
pressing the internal leads against the
pads (7.15c). A pulsed electrical current heats the head, creating the bond. The
liberated heat melts the wax and frees the chip. Following welding, the strip 1d
lifted along with the bonded chip (Figure 7.15d), and then shifted to the left,
transporting the next batch of leads into the working position (Figure 7.15e).
The coordinate stage is shifted to the left, feeding the next chip for welding.
Assembly machines have been designed having a tape winding unit. The machines
are intended for soldering by means of inelting a dosed amount of solder; the film
is moved in steps of up to 127 um at a speed of 38 mm/sec. Each chip is manually
placed under a microscope. The requisite film tension is assured by an induction
motor. The welding head of the machine has an electric drive. It is moved into
the working position for 375 msec. The productivity of the machine is about 1,000
chips per hour.
A drawback to the technique consists in the fact that the polyimide strip, after
the bonding of the leads, proves to be practically unneceasary, something which
is not efficient from the viewpoint of material consumption, and moreover, the
plastic tape yields a shrinkage on the order of 25 Um, which makes it difficult
for automated equipment for connecting leads to operate reliably. Proposed as
a promising system is the assembly of integrated circuits on a plastic film in
rollers [42]. Using this method, instead of etching a polyimide film cladded
- 169 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500094404-3
FOR OFFICIAL USE ONLY
~
Figure 7.16. One of the fastentng methods in an integrated circuit
package in a film frame.
with copper, electrolytic and chemical deposition of copper or nickel is used, as
a result of which, high precision ia obtained in the figure for the lead layout.
The minimal width of a conductor is equal to the minimum spacing between the
conductors and amounts to 0.005 mm.
The walls of the conductors are vertical, while in circuits fabricated using foil
etching, they are inclined as a result of undercut etching. The new technique
is also favorably distinguished from earlier methods in that it does not require
the use of an adhesive material to join the plastic film to the metal. This
makes it possible to fully utilize the high thermal resistance of polyimide film:
it is capable of sustaining a temperature of 400� C for 15 seconds, which is
more than sufficient for the mounting of integrated circuits using electrical
contact heating or thermal compression welding. Gold beads or beads of solder
can also be deposited on the original film. The film, which is 16, 35 or 70 mm
wide, with the integrated circuits attached is protected by a special coating.
The packages within which the film frames are hermetically sealed have inspection
windows, and for this reason, when mounting IC's on a transparent polyimide film
using the flip chip technique, the resulting contacts are partially visible
through the package.
Frames of f ilm with integrated circuits can be mounted in packages in various
ways. One of them is shown in the figure (Figure 7.16). The edges of the film
are bent around a rigid dielectric plate, and the resulting semi-f inished piece
is inserted in the package so that the conductors on the film make a reliable
- 170 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
electrical contact with the package leads. The gluing of the frame to a dielectric
plate or aluminum heat sink is also practiced. The most widespread method does
not require hermetic sealing of the film and provides the capability of automating
the assembly process. A special machine cuts out the frames with the IC's from
a roller, sets them with their front side down on the printed circuit board and
completes the assembly by means of electrical contact heating.
7.5. Equipment for the Assembly of Point Contact Diodes
The most widespread type of point diode is the D9 device. The attachment of the
wire by means of welding to the needle holder of the D9 device, the cutting off
of the wire, the forming of the needle, the quality control rejection as well as
the placement of the needle holdere with the weld attached needle in a cassette
holder are accomplished on the automatic unit described below (Figure 7.17).
When operated by a single operator, the automatic unit provides for the assembly
of 3,600 to 4,500 devices per hour. It consists of three major assemblies: the
unit for the welding and shaping of the needle 3, the table 1 and the vibration
hopper 2. A pulse counter and other instruments, the readings of which are used
to choose the welding process mode, are arranged on the control panel 4. The
vibration hopper feeds the needle holders to the transport mechanism.
A kinematic schematic of the automatic
unit is shAwn in Figure 7.18, from which
one can trace the operation of the
assemblies and mechanisms. The mechanisms
are driven by drive 1, which provides
for the motion and rotation of all the
mechanisms. The needle holders which are
loaded manually into vibration hopper 2
A-A are fed via guides 3 and 5 to the trans-
port mechanism, which moves a needle
holder from one position to another and
consists of a reducer, stationary rack
9 and moving rack 8. The piecewise feed
mechanism 4 is mounted on the reducer,
where this mechanism executes the indivi-
dual feed of the needle holders to the
transport mechanism by means of four
levers, which are driven by a cam of the
Figure 7.17. The automated needle weld- transport mechanism. The transporting
ing and forming unit. element of the transport mechanism is
rack 8, which executes a reciprocating
moCion and feeds the needle holders from one position to the next. The needle
holders are moved via the rack to the region where the end of a needle holder is
cut off. The cut-off inechanism 13 consists of a guide, and a slide to which the
upper knife is fastened. The slide is driven by a cam, located on the reducer of
the transport mechanism by means of a bearing. The bottom knife is mounted in
the stationary rack of the transport mechanism. The needle holder is clamped in
a definite position by means of a special device which is fas*.ened to the slide.
- 171 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-04850R000500090004-3
FOR OFF[CIAL USE ONLY
Figure 7.18. Kinematic schematic of the automated needle welding and
forming unit.
The needle holders then move to the next position, where the needle holder is
clamped., the wire is fed in and bonded and the needle is cut off and shaped. The
clamping mechanism 6 consists of a slide, which is moved in the guide by means
of a cam and a tilting bearing. A clamp is fastened to the slide by means of
the chassis and a spring, where this clamps presses the needle holder against the
rack prior to the welding operation. A contact is located on the chassis, which
is used to check for the presence of the product and turn on the electrical power
when performing the welding. The mechanism for feeding and bonding the wire 10
is located on a post. Two guides are secured to the post; the vertical up and
down travel of the slide is accomplished via these guides using a screw. Two
horizontally running guides are fastened to this slide; the rack is moved alonF;
these guides by means of a screw. A bracket is rigidly fastened to the frame,
which also has two guides, along which the slide travels with the feed and welding
heads positioned on the slide. The feeding of the wire is accomplished by means
of a lever, which is driven by a cam of the transport and clamping mechanism.
The cutting and forming mechanism 7 consists of a bracket, guides and two slides.
The lower knife is secured to one of them, while the upper knife is fastened to
the other. The knives perform the operations of cutting off and shaping the wire.
The motion of the guides is realized by means of a spring and cams of the trans-
port mechanism through bearings. By virtue of grooves which exist in the bracket,
on which the cutting and forming mechanism is mounted, it can be moved in a
horizontal plane.
The welding process is accomplished by means of sector 12, which controls the
limit switch. The finished products are fed to the cassette holder loading
position. The final operation is carried out by means of the loading mechanism
11, which consists of a stationary housing and a moving carriage, on which the
cassette holder is placed. The carriage moves along the guides by virtue of the
motion of a stem and lug, which meshes with the toothed rack, fastened to the
carriage. When one cell of the caesette holder is filled, the pulse counter feeds
- 172 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
a signal to the electromagnet, which when turned on, the stem touching the cam
moves the cassette holder by one step. When the cassette holder is completely
full, the carriage moves to the microswitch, and the automatic unit is cut off.
The operator changes the cassette, returns the carriage to the initial position
and the cycle is repeated from the beginning.
7.6. Equipment for the Automated Assembly of Alloy Diodes
The assembly of D226 silicon alloy diodes is accomplished on an assembly unit
with a continuous nickel belt. The belt transport mechanism is fed to a press
fY-om the unreeling assembly; the fdrming of the chip holder and the cutting out
of the preparations are done automatically in the press. Oil residues are
removed from the strip in a degreasing bath with trichloroethylene. The degreased
strip is fed into the automatic unit for the weld attachment of the lower lead.
A moving electrode is brought from above against the strip, while from below, a
feed carriage feeds a nickel plated copper wire and presses it against the strip
and welds it, after which a knife cuts off the lead. Then the strip moves on,
going to a tinning mechanism, a washing bath and drying furnace.
The washed dry strip is transported in this way to a furnace for bonding the
junctions. The operator sets the cassette holder with the junctions on the strip.
The juncti.ons are wetted with solder, and after cooling form a reliable electrical
contact with the chip holder. Then a capsule is manually placed on the junction
lead. After this, the capsule is automatically ring welded to the chip holder.
The operator seats the cassette holder with the intermediate leads on the capsules
to center a capsule relative to the crystal holder. The clamping devices go in
the perforation holes, thereby centering the capsules. After performing the
operation, the operator removes the cassette holder with the bonded capsules and
the strip is transported to the automat for tube swaging. A die, which in press-
ing the tube against a stationary support, deforms it by 3.5 + 0.5 mm is fastened
to the rod of a pneumatic mechanism, which is operated by a valve controlled by
a cam. The swaged portion of the tube is then welded at 3+ 0.5 mm by the weld-
ing mechanism. The operation of the tube welding mechanism is similar to the
swaging mechanism, only electrodes are mounted in it instead of dies.
Following the welding of the tube, the electrodes are moved away and the sealed
device is fed to the position for welding the upper lead. The automatic welding
oE the lead is accomplished in the following manner. A clamp of the mechanism
catches the nickel wire and feeds it through an interception assembly and a
knife draw die to the small tube of the capsule. A moving electrode is brought
up by means of a pneumatic cylinder, where this electrode presses the lead from
the tube against another electrode which can be moved. With the approach of the
moving electrode to the tube of the device, a centering fork catches the tube
with its lower cutout, while with the upper one, centers the lead relative to
its axis. Following the welding, a knife cuts off the lead, while the intercep-
ting assembly retains the wire until the clamp returns to the upper position.
The moving electrode is returned to the initial position and the entire cycle is
repeated.
-173-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ON1,Y
The removal of the assembled device from the strip JAs accomplished in a press
for cutting out the finished devicQ. The press cuts out two devices for the
purpose of improving the dynamic characteristics.
In structural terms, the assembly unit combines several sections, which are joined
together in a line, which is described in Chapter Fifteen. The mechanisms of the
unit operate from thr.ee separate drives. The ring welding mechanism and the
mechanism for swaging and weld attachment of the tube, as well as the mechanisms
Eur welding the lower leads and tinning the strip, and the strip transport and
upper lead welding mechanisms are all driven by the main drive. The forming
press for the chip holders and the press for cutting out the finished device
operate from individual drives. The synchronization of the operation of all of
the drives is electrical.
The major assemblies and mechanisms which assemble a device are described in the
following.
-1
~

~
The assemblies for clamping, feeding
and cutting the wire, the upper elec-
trode, the pneumatic cylinders and
valves are located in the mechanism
for welding the upper lead (Figure
7.19). The parts of the mechanism
are driven by the main drive through
clutch 8, cylindrical gear 7, the shaft
for actuating the valves 5, cylindrical
gears, shaft 13, the conical spur gear
pair 6 and a vertical distribution
shaft.
Figure 7.19. Kinematic schematic of the
mechanism for welding the
lower lead and the tinning
mechanxsm.
The lead is clamped by means of jaws
which are driven by the rod of pneumatic
cylinder 9, and moved by pneumatic cylin-
der 12. The pneumatic cylinders are con-
trolled by valves 4. The pneumatic -
inder for feeding the lead provides for
the requisite clamping force for the lead
against the strip, which is regulated by
varying the air pressure in the system
by means of reducing valve 1 and is sta-
bilized by receiver 2. The feed brake
is adjusted by throttling valves 3.
There is a moving stop 10, which limits
the lowering of the feed assembly, for
the regulation of the protrusion of the
lead. During alignment and adjustment,
the mechanism is disconnected from the
main drive and the parts are moved by
levers 11.
-174-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
The drying furnace is made in a separate housing, in which spiral heaters are
located. The working temperature for the drying is 100 + 10� C. The time that
the furnace is in the operating mode is no more than 30 minutes.
The f urnace for attaching the junctions takes the form of a separate housing,
in which there are four electric heaters. On the outside, the furnace is
covered with a casing which is cooled by water flowing through it. The tempera-
ture in the furnace is adjusted in a range of from 50 to 600� C. The time that
the furnace is in the maximum operating mode is no more than 30 minutes. It is
powered from the AC mains at a voltage of 220 volts. The maximum power of the
furnace is 2.4 KW.
A kinematic schematic of the ring welding, tube swaging and upper ?ead welding
mechanism is shown in Figure 7.20. The main drive for the unit consists of
electric motor 3, the V-belt drive, worm gear reducer 2, cylindrical spur gear
transmission 4 and the distribution shaft 1.
The ring welding mechanism consists of the welding housing, the upper 5 and
lower 11 pneumatic cylinders for moving the electrodes and the lower electrode
clamping assembly 12. The overall travel of the upper electrode is 40 mm. The
working travel of the upper electrode is 10 mm, and that of the lower electrode
is 35 mm.
The tube swaging mechanism includes pneumatic cylinder 6, stationary stop 10
and the guides for holding the strip.
The mechanism for welding the upper lead has a welded frame, on which the
distribution shaft 9 with the control cams and drive gears, as well as the
wire bobbin 7 and electrode assembly 8 are mounted. The entire mechanism is
secured on a separate plate, having four slots. The parts of the mechanism
are rotated by the main shaft 1.
The entire process of assembling the devices is carried out in a controlled
medium, which is assured by the protective suit existing in the unit. It con-
sists of standardized sections of welded structures. Each section is installed
and secured to the upper plate of the frame.
To control all of the mechanisms of the equipment, there is a control panel in
it on which the requisite controls, switches and lights are located, by means
of which the mechanism is turned on and the production process operations are
mon itored.
7.7. Equipment for the Assembly of Power Transistors
The sharp growth in the production of power transisf.ors, the increase in the
currents, voltages and power, and consequently, the increase in the dimensions
of the chips, lead diameters, geometric dimensions as well as the materials
of the package make it necessary to seek out new, more promising methods of
assembly, which would make it possible to boost productivity and quality.
- 175 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540090004-3
FOR OFFICIAL USE ONLY
~ - ,
I
o-_
.v _
b
0
i
~n
u-

- 176 -
FOR OFFICIAL USE ONLY
00
aa
~b
~n w
3
~
0o b
~ co
3 a

X
0)
4
Ai ~
c~
w
o r+
~
.0 3
v~
41
t0 R1
$4
G! 00
aZ
o
b
x a~
3
00 V
~ ~
3
0 00
x ~
fA vi
00
u co
-H 3
Ai w
co
a~ .n
u a~
~
w y
u ~
p m
o �r-+
~ aai c~a
a~ ~ x
~ co u
r1 N
aL ~
0
N
n
Gl
H
~
00
~
G4
~
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090044-3
The existing production of power transis-
tors is characterized by a great divers-
ity of the assembly methods, for exauuple,
the connection of the chips is accomplished
by flux soldering, eutectic bonding, and
bonding in a hydrogen furnace. In one
case, here, the chip is joined directly
to the mounting base, and in the other,
to a gold plated molybdenum disk. The
connection of the leads is also accom-
plished in different ways: flux soldering,
bonding in a hydrogen furnace and ultra-
sonic welding. All of this requires the
development of special equipment to per-
form the indicated operations.
The installation for the ultrasonic bond-
Figure 7.21. The unit for ultrasonic ing of leads (Figure 7.21) to high power
lead bonding. transistor chips is described below. It
provides for the automatic bonding of two
wire leads to the metalized pads of a chip
by means of applying ultrasound and pressure with the subsequent for^.ing of the
leads to assure their orientation relative to the end faces of the transistor
mounting bases. The installation is serviced by a single operator. The produc-
tivity of the installation is 1,200 welds/hr. The devices are fed in a cassette
holder; the capacity of the cassette is 10 devices.
The installation consists of the following assemblies: the control panel 1,
the welding head 2, the mechanism for clamping and feeding the devices 3, the
plate 4, manipulator 5, the electrical equipment cabinet 6, table 7, ultrasonic
generator 8 and the pneumatic assembly 9. The installation can operate in two
modes. In the case of operation in mode I, the leads are welded to a chip in
the unit; in the case of mode II operation, the leads are bonded ultrasonically
to the cross-ties of the transistor mounting bases.
The bonding of a lead to a contact bonding track of a chip is accomplished by
means of the tool which is mounted at the end of the waveguide of a magneto-
strictive ultrasonic transducer. The bonding occurs by virtue of the joint
action of pressure and ultrasonic oscillations on the parts being joined to-
gether.
The working cycle of the process takes place over one revolution of the distri-
bution shafts A, B and C(Figure 7.22). The functional linkage of the operation
of these shafts is accomplished electrically. The f.ollowing mechanisms are
operated by means of the cams arranged on distribution shaft A: the wire feed
meclianism is driven by cam 14; the wire cutting mechanism is driven by cam 18
and the wire hclder mechanism is rotated by cam 20. The shaft is rotated by
electric motor 19. Distribution shaft B controls the operation of the mechanism
for the step feed of the cassette holder by cam 29 and the mechanism for clamp-
ing the devices in the bonding position. The shaft is rotated by electric
-177-
FOR OFFICIAI, USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500094004-3
FOR OFFICIAL USE ONLY
motor 26 through the toothed belt drive 27. The mechanism tior lowering the
riL�oustical system wirh the tool secured to the end of the concentrator is driven
by cam 22, which is positioned on shaft C. The shaft is rotated by e?ectric
motor 21. The clamping of the wire feed jaws 12 is accomplished by an electro-
magnet, built into the stationary jaw 13. The wire, which is cut off by knives
11, is fed to holder 10 near the tool 25. The devices are brought into the
working position in the cassette holder with each step of carriage 8. The final
matching of the bonding pad on a chip to the tool is accomplished by manipulator
30. Shafts A and B, after being turned off, are braked by electromagnetic brakes
4, which are rotated through gear pairs 2 and 3, as well as brake 17.
Figure 7.22. Kinematic schematic of the ultrasonic bonding unit.
Ttie working controls of the installation are controlled by cams, contactless
switches 9, 23 and 16, as well as clamping microswitches, by means of which
the following are accomplished: the device is clamped (cams 1 and 28); the
- 178 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400500090004-3
pneumatic valves for the lead bending and forming cylinders are actuated (micro-
switches 5 and 6); the device feed drive motor is turned on (lobes 7); the feed-
ing of the wire to a specified length (cam 14); the turning-on of the drive
motor for wire preparation (lobe 15); the cutting-off of the wire (cam 18); the
rotation of the wire holder (cam 20); the lowering of the tool (cam 22); the
turning on of the ultrasonic generator and the shutdown of the motors (lobe 24);
and the moving of the cassette holder through one step (cam 29).
In structural terms, the major assemblies and mechanisms of the installation are
made as follows. The welding head is the major actuating mechanism and consists
of the following assemblies: the wire preparation drive, the wire feed and cut
off inechanism, the mechanism for feeding the wire to the tool and the acoustical
head and housing.
The wire preparation drive provides for the requisite motions of the mechanisms
which prepare a wire segment far bonding. It consists of a distribution shaft
with the cams seated on it, by means of which the clamped shutdown of the shaft
is realized after the motor is turned off. In addition, there is an electromag-
netic break.
The wire feed and cutting mechanism provides for feeding the wire from the
spools by means of jaws into the nipple of the holder of the wire transport
mechanism. The cutting of the wire is accomplished by a moving knife, which
turns on its axis relative to the stationary knife. The mechanism has a stop
plate, which is set up relative to the height of the guides, assuring that the
requisite length of the wire is obtained.
The mechanism for feeding the wire to the tool feeds a section of wire, which
is secured in the nipple by means of a spring, which actuates during the motion
of the knife and clamps the wire at the tool. The wire feed is accomplished by
virtue of a reciprocating motion of the lever.
The acoustic head is mounted in a carriage in ball guides on a cantilever bracket
and is rigidly coupled to a shaft, which is supported on two bearings mounted in
the bracket. Because of this coupling, it has the capability of rotating
together with the axis through an angle of + 5�. The horizontal position of
the end �ace of the tool is set by adjusting screws. The rotation of the system
about its axis depends on the height to which the head is lowered after the
tool encounters an obstacle.
The mechanism for clamping and feeding the devices by means of the carriage and
two lugs accomplishes the stepwise feeding of the cassette holder. The cassette
holder moves in the guides on bearings, which extend above the slot. The
clamping of the devices is accomplished on both sides by cams, mounted on a
single shaft. There is an electromagnetic brake, which is coupled to the shaft
through a gear coupling to clamp the shutdown of the shaft at the moment it is
turned off.
The manipulator moves the cassette holder with the devices relative to the tool
in a field of 25 x 25 mm. It is built using ball guides. There are two
- 179 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OF'F'ICIAL USE ONLY
pushbuttans for turning it on and the feed of the devices in the drive control
of the manipulator. The operator controls the operation of the installation from
the control panel and both semiautomatic and adjustment modes are possible.
- 180 -
a
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
CHAPTER EIGHT EQUIPMENT FOR HERMETICALLY SEALING SEMICONDUCTOR DEVICES
One of the decisive factors which influences the stability of semiconductor
device parameters is the composition of the ambient medium around the semicon-
ductor chip with the electron and.hole junctions, which are extremely sensitive
to exposure to all possible kinds of dirt, moisture, various kinds of deforma-
tions, etc. A hermetically sealed package, inside which the junctions are
placed, should reliably isolate them from the environment. Semiconductor device
packages having a leakage of less than 5. 10-6 1- um/sec are considered hermet-
ically sealed. .
The ma3or requirements placed on hermetic sealing operations are: producing
vacuum tight and mechanically strong joints of the package elements; precluding
the possibility of dirt, gaseous emissions and splashes getting into the sealed
volume of the device during assembly; the impermissibility of heating the
junction during the hermetic sealing above 120� for germanium devices and 200�
for silicon devices.
The diversity of existing semiconductor device and IC packages is explained by
the simultaneous use of several methods of hermetic sealing in production and
the types of equipment corresponding to them. The techniques of cold and resis-
tance welding have become the most widespread for hermetically sealing metal-
glass packages, while the methods of soldering with low temperature solders and
roller contact welding are most widely use3 for sealing metal-ceramic packages.
The sealing of semiconductor devices and IC's in monolithic plastic packages by
means uf transfer forming (casting under low pressure) has become widespread.
A large group of semiconductor devices, diodes, is hermetically sealed in all-
glass packages using special equipment; in this case, the alignment, electro-
forming and welding together of the packages is accomplished in the equipment.
The parts being welded are heated by means of a plate or ring type direct incan-
descent heater. Equipment is known in which the heating is accomplished by
radiofrequency currents. In a number of cases, equipment is used to hermetically
seal complex products in multiple lead packages in which the heating of the
parts of the package being ,joined together is accomplished by an electron beam,
focused infrared rays, by a plasma or a laser.
8.1. Equipment for Cold Welding
Cold welding assures good quality of the weld seam: the process takes place
without gas liberation and heating which have a harmful influence on the pro-
perties of an electron-hole junction. To be included among the drawbacks to
this technique are the necessity of increasing the diameter of the package
because of plastic deformation of the components being welded, the necessity
of using ductile metals and the somewhat limited capabilities of welding thin
wall parts.
Cold welding of a semiconductor device mounting base with a piston can be
accomplished using one and two sided compression. In the case of compression on
one side, an annular indentation is formed on one side (Figure 8.1a), and with
- 181 -
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400500090004-3
FOR OFFICIAL USE ONLY
two sided compression, it is formed on both sides (Figure 8.1b). The optimal
variant is one sided deformation in a free volume. In this case, the deformation
needed for seizure of the copper-Fernico pair is 67% as opposed to 72% in the
case of double sided deformation [52].
The optimal degree of deformation of the package parts being welded is assured
through the structural design of the working tool, which makes it possible to
produce tite specified thickness of the metals at the weld site.
The working tool is a punch (Figure 8.2) fabricated of the KhVG or Kh12M alloy
steels and tempered to a hardness of Hrc = 52 to 60; the hard alloy VK20 is also
used. The weld quality is governed by the condition of the surfaces being joined
and the force anplied to the working tool.
~
'
~
a) ~ SI
Figure 8.1. Kinds of cold welding.
For hermetic sealing using cold welding
in semiconductor production, special
hydraulic or pneumatic presses are
used having a force of from 5� 104 up
to 6- 105 N. The parameters of the
most widespread equipment are given
below.
Technical Specifications
Output, welds/hr
Working force, Newtons
Working travel of the
tool, mm
Number of carousel
positions, pieces
I ~
N 2
-~k
I J
4
I
; x; I \ $
: I
Figure 8.2. The wor'.cing tool for
cold welding.
Cold
020.0007 020.0011
600-900 200-600
105 6�104 -
3�105
27 5
12 2
Key: i. Punch from the side of
the part made of softer
material;
2. Cup;
3. Cap;
4. Mounting base;
5. Punch from the side of
the part made of harder
material.
Welding Equipment
2.221.006 2.220.003
1,200 800
105 5�104 -
2�105
10 6
8 6
- 182 -
FOR OFFICIAL USE ONLY
5.333.00.000
500
6�105
10
6
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-04850R000500090004-3
/6 !7
I!
Figure 8.3. Semisutomatic unit for hermetically encapsulating s2mi-
conductor devices.
~3 6 The semiautomatic unit for hermetically
sealing semiconductor devices using a
force of 105 Newtons is shown in Figure
8.3. The semiautomated unit is a 12
position carousel type. All of the
mechanisms, assemblies and parts are
B mounted on a machine tool bed, consisting
Z 71p9 of cast bases 1 and 11 and two plates:
the upper 5 and the lower 14 plates.
!
The carousel 7 with the seats 8 in
. which the assembly of the device package-
Figure 8.4. Kinematic schematic of the p8rts is accomplished prior to welding
semiautomatic heraetic as well as the welding itself are mounted
encapsulating unit. in the upper plate; the unloading
mechanism b for the automatic unloading
. of the welded devices from the carousel
nests; hydraulic press 10, which creates the requisite force and consists of a
hydraulic cylinder with a piston, two columns and a cross-piece. The housing
9 with the upper punch is connected to the cross-piece; there are grabs on the
piston to extract the nest with the lower punch. In case a device jams in the
upper punch, a mechanical device for pushing it out is provided, which is
coupled to the piston.
All of the assemblies which are located on the upper plate of the bed are isolated
from the environment by protective cover 18. The interior volume of the protec-
tive cover is filled with an inert gas or clean dry air during the operation of
the semiautomatic unit. There are two windows each in the front and rear walls
of the protective cover in which locks 16 are inserted (for loading parts and
unloading finished products) as well as seals 19 for the arms of the operator.
A reducer 3, which is coupled to the electric motor by a V-belt drive 4 is
- 183 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500090004-3
FOR OFFICIAL USE ONLY
secured at the bottom of the top plate.
Oil tank 13 with a hydraulic pump and
electric motor 15 are mounted on the
lower plate.
Panels 2 and 12 with the control elements,
as well as the electrical and hydraulic
equipment are mounted on cast bases.
Cabinet 20 with the electrical equipment
is built into the rear portion of the
bed. To fill the lock chambers and
working volume of the protective cover
with the inert gas or cleaned dry air,
a hose with a nipple, which is screwed
into a through-hole in the upper plate.
There are similar nipples in the lock
chambers. A luminescent lamp 17 for
local illumination is placed at the top
in the protective cover.
Figure 8.5. Hydraulic schematic of
semiautomatic hermetic
sealing unit.
The interaction of the actuating mechan-
the isms can be traced using the kinematic
schematic of Figure 8.4. The rotation
from electric motor 1 is transmitted .
through V-belt drive 2 and the worm gear
pair 3 to camshaft 10. The helical groove 7 as well as the cams 4 and 9 which
are rigidly secured to the camshaft drive carousel 6, the unloading mechanism
5 and slide valve 8.
The hydraulic system (Figure 8.5) operates as follows. Industrial oil 20 (GOST
1,07-54) is fed from vane pump 1 through the plate filter 2 into the four-way
valve 5 to working cylinder 4. The oil pressure is monitored by manometer 3.
When the projection of slide valve 5 is covered with oil, the oil fills the
space beneath the piston under pressure and lifts the piston. The oil, which
during this time is above the piston, drains off into the tank. When the valve
is released, the oil executes the return trip and the piston returns to the
initial position. Check valve 6 with the unloading valve provides for a constant
specified pressure and protects the hydraulic system against overloads.
8.2. Equipment for Electrical Contact P.esistance Welding
Resistance welding, in contrast to cold welding, makes it possible to weld
finer parts, does not increase the dimensions of a package, and provides for
a higher productivity. For the purpose of eliminating long term heating of the
devices being sealed together, resistance welding is used which assures local
and brief heat liberation at the weld site.
Par.ameters of capacitor machines for contact welding of semiconductor device
packages are given below [53]:
- 184 -
TOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
Technical Parameters
Perimeter af the package
being welded (the welded
seam), mm
Output, welds/hr
Force on the electrodes,
Newtons
Nominal welding current
(Amplitude value),
amperes
Maximum stored energy,
J � 103
T
vpe of Capacitor M.achine
MRK-4001
IrIItK-10001
MTK-8002
MTK-5-3
22-55
55-100
30-80
3-12
1,200
1,200
900
1,200
1�103 -
2�103 -
1.5�103 -
3�102 -
5�103
1.23�104
1.23�104
3'103
40,000
100,000
80,000
32,000
3.6
16.1
14.8
2.7
The formation of the welded seam in the case of electrical contact resistance
welding occurs by virtue of the heating of the parts being welded by the current
and their plastic deformation with the action of the applied compression force.
The electrical energy is stored in a capacitor bank where the capacitors are
charged from a DC power supply.
The quantity of energy stored is regulated by changing the working voltage and
capacitance of the bank. Changing the working voltage is accomplished by changing
the master voltage, while the capacitance of a bank is changed by means of switch-
ing three sections of a capacitor bank in different combinations by means of
a step switch for the capacitance, P8t.
The battery is discharged through the primary winding of the welding transformer
at the moment the discharge circuit is closed by one of the discharge contactors,
which operate in sequence, changing the direction of current in the transformer
windings in each cycle for the purpose of preventing the magnetization of the
transformer.
The discharge current pulse, and consequently, the welding current pulse are
governed by the parameters of the electrical power section: the working voltage,
the capacitance of the bank of capacitors, and the transformation ratio of the
welding transformer.
An overall view of a capacitive welding machine is shown in Figure 8.6.
The major assemblies of the machine are: the frame 2 with the bracket 8, welding
attachment 11, the pressure drive 10, the pneumatic system 9, protective enclosure
5, gas system 4 with the drier 3, welding transformer 1 with the switch for the
taps 7 as well as choke 6 and the electrical equipment. The machine complement
includes a power supply and control station (not shown in the figure).
In structural terms, the frame takes the form of a welded metal chassis, on the
upper plate of which the bracket and protective enclosure are mounted. The
pressure drive is installed on the bracket; the welding attachment is mounted
inside the protective enclosure. Inside the frame housing are located various
- 185 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
a
3
2
1
6
7
Figure 8.6. General view of the capacitive welding machine.
devices and machine systems; access to the adjustable elements of which is accom-
plished through two doors in the rear wall of the housing. The welding attach-
ment provides for moving the upper electrode during welding and maintaining the
working surfaces of the electrodes parallel, something which is important in
obtaining a high quality seam. The welding attachment housing is fastened to
the base of the protective enclosure.
A slide, with the upper electrode mounted in it, which is connected through the
upper current conductor and flexible buses to the secondary winding of the welding
transformer, travels inside the framework on ball bearings. The lower current
conductor is brought in through a hole in the base of the chassis, where this con-
ductor connects the lower electrode to the secondary winding of the welding trans-
former.
The force is transmitted from the pressure drive to the slide through a set of disk
springs, placed between the slide and a U-shape bracket, which when engaged with
the slide goes into the tailpiece of the pressure drive.
The pressure drive (Figure 8.7), which is intended for producing th_~ force on the
electrodes during welding, consists of pneumatic cylinder 7, piston 6 with the rod
5, spring 4, adjusting nut 3, push rod 1 and the pressure indicator 2. The force
on the electrodes is produced as a result of the spring compression and the trans-
mission of this force through the push rod to the slide of the weldir.g attachment
with the upper electrode secured to it. The structural design of the pressure drive
provides f or a stable specif ied force on the electrodes with considerable fluctuat-
ions in the pressure in the compressed air mains and in the pneumatic cylinder.
- 186 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400500090004-3
FOR OFFICiAL USE ONLY
The air rate of flow through the system is regulated by means of valves, which in
turn make it possible to regulate the rate of travel of the pistons of the pneumatic
cylinders.
The constancq of the gas medium in the welding region is assured by the gas system
and the protective enclosure. The quantity fed into the protective enclosure is
monitored by a direct reading flow meter. The admission of gas into the lock cham-
ber and its extraction from the protective enclosure and the chamber are accompli-
shed by means of vacuum valves.
The weldfng transformer is mounted in the central portion of the machine frame
housing. The primary winding of the transformer is made from two cylindrical
type coils, while the secondary is made from two copper foil packets connected in
parallel. A choke is used to shape the leading edge of the welding pulse, where
the choke winding is inserted in series with the primary winding of the welding
transformer. The core of the choke has a variable sir gap.
The electrical equipment of the machine
consists of the electrical power section
and the control circuits. The power sec-
tion stores the energy from the mains in
the bank of capacitors and feeds it to
the weld site. The control circuits pro-
vide for the sequence and duration of
actuation of the power section components
and the other elements of the machine
during cyclical operation.
s
7 The machine has a broad control range for
the amplitude and width of the welding
pulse as well as the electrode force,
because of which, one can weld devices
with diameters of from 3 to 12 mm. The
welding pulse is adjusted in seven capa-
citance steps of the bank of capacitors
with the voltage varying from 150 to 400
volts, using eight steps for switching
the welding transformer. The force on
the el2ctrodes is adjusted in a range of
3- 10 - 3 � 103 N.
8.3. Equipment for Hermetic Encapsulatiot
With Plastics
Encapsulation using plastics finds wide
Figure 8.7. Structural design of the scale applications for devices used in
^ressure drive. consumer electronics equipment. Various
plastic materials are employed: epoxy
resins with various hardeners and syn:he-
tic elastomers. The most widespread methods of hermetic encapsulation are the
techniques of free potting of the forms and casting at low pressure - transfer
forming.
- 187 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500090004-3
NOR OFFICIAI. USF: ONI,Y
The UGP-50 installation (Figure 8.E) for hermetic sealing of semiconductor de-
v:ces with plastic takes the form of a column type hydraulic press with the
compression plate located at the bottom and the casting cylinder 4 placed at the
top. Control panel 1, cleaner 2, the press 3 and table 6 with the control unit
S located on it are mounted on the welded frame. The frame is enclosed with
sheathing, and the hydraulic equipment is located inside the frame. The servicing
of the installation is accomplished from the instrument side of the panel on
cohich the control buttons are brought out.
~
~
~
~
~
,
~
i
~
i
F
~
~
p
HTTi
~ _ ,
i~
1 f' P
r; ( g ) ~J( a ~

'Pf ~
p, ' qb) dl (e)
i
i
~ : ' ~ 1
_I
c~ (c) ~J( f )
Figure 8.8. The installation for plastic Figure 8.9. Functional schematic si.ot.
encapsulation. ing the hermetic encapsti-
lation nxocess.
Ttie operation of the unit is based on the principle of batch encapsulation in P
stationary compression mold. The sequence for the execution of the hermetic
encapsulation operations is shown tn Figure 8.9. A section of the strip carrier
2 is loaded into open compression mold 1 with the chips bonded to the strip
(Figure 8.9a). w'hen the pump is turned on, there is at first an accelerated
closing of the compression mold, and then a slow closing. The maximum compression
force in the UGP-50 installation is 0.5 MN. A tablet of the compression molding
material 4 is loaded into casti.ng chamber 3 of the compression mold (Figure 8.9b),
then the lowering is speeded up, and thereafter the working stroke of rod S of
the cylinder (Figure 8.9c). The speed of the working stroke is adjusted by means
of the feed mechanism, while the force is adjusCed by means of the stop valves
using the manometers until the requisite pressure is reached in the upper and
lower cavities of the cylinder. During the working travel of the rod, the
material is injected at high pressure (Figure 8.9d). To improve the fluidity,
the tablet of compression molding material is heated beforehand by high frequency
currents. To obtain a high density in the package, it is exposed to the nominal
- 188 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
pressurc for some time (I'igure 8.9e) (the exposure time depends on the compression
molding material and is specified by a time delay relay). Then the rod 5 is lift-
ed, the compression mold is released, the hermetically sealed devices 6 are
removed and the compression mold is cleaned (Figure 8.9f). The installation is
ready for a new cycle. Installations with compression forces of 1.0, 1.5 MN and
more, as well as so-called shuttle presses, which provide for an increase in
output by combining the main and auxiliary process times.
8.4. Equipment for liermetic Sealing by Means of Soldering
Flat metal-glass and metal-ceramic packages are hermetically sealed in a batch
sealing installation by means of soldering using a heated inert gas (Figure 8.10).
One can run the soldering process in the unit both with flux and without it
[54]. The devices being sealed in a ten place cassette holder with the assembled
bases and caps are manually placed on the moving carriage of the installation,
and then the entire sol3ering cycle is accomplished automatically upon instructions
from the programmer. The devices are introduced into the effective zone of a
jet of heated inert gas; there is a separate heater for each package, something
which provides for better observance of the soldering conditions, as well as the
possibility of more precisely maintaining the temperature. A provision is made
for an individual direct reading flow meter to regulate the gas rate of flow
for each heater. The heat flow is on the cap side, and therefore the temperature
of the package base where the semiconductor structure is located is always lower
than the soldering temperature. During the entire soldPring cycle, including
cooling, the cap of the device is clamped with a special device and pressed
against the base with a specified force.
Excessive force leads to splashes, and
weak force pror,otes the shifting of the
cap and the appearance of defective
seals. Following soldering, the her-
metically sealed packages are flushed
with a cold inert gas, which promotes
th.e rate of crystallization and prevents
~ the infiltration of solder inside the
{ integrated circuit package. Upon com-
~ pletion of the cooling, the carriage
; with the cassette holder is automatically
~ returned to the initial position.
8.5. Equipment for Checking the Herr.!etic
Seal of Semiconductor Devices
Figure 8.10. Installation for batch her-
metic sealing by means of
soldering.
The degree of the hermetic seal of the
package of any semiconductor device or
IC is one of the most important para-
meters which influence their operabilit,y
and reliability. The criteria for a
hermetic seal differ depending on the
area of application of a semiconductor
device and IC, as well as the interior
- 189 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500090004-3
FOR OFFiCiAL USE ONLY
volume of the package: thus, for consumer electronics devices, the leakage should
. not exceed 10-~ 1- um/sec, and for especially reliable devices, the permissible
leaks do not exceed 10-8 1- um/sec. This means that over a period of 10 years,
- the package will not admit more trian 1 cm3 of air with a pressure difference of
1 atmosphere.
Several techniqutes exist for checking the hermetic seal of devices:
a) Bubble methods, b2sed on the observation of gas bubbles exiting a device
placed in a liquid. These include the following:
--Fluid method, in which the check of the hermetic seal is accomplished by visual
observation of air bubbles exiting the device package where the device is
placed in silicon oil heated up to a temperature of 200� C;
--The vacuum-liquid method, which is based on observing gas bubbles exiting a
device placed in a liquid, over which a rarefaction is created.
The vacuum-liquid technic,ue has a poor sensitivity of 10-2 1- um/sec (the liguid
method has a sensitivity of 10 1- um/sec), but it can be increased up to 10-j
to 10-4 1- Um/sec, by varying the composition of the liquid, the pressure,
the temperature and the depth of immersion nf the device.
b) The mass spectrometry method, which ia based on reading the amount of helium
exiting through leaks existing in the device package. This is the most widespread
and sensitive method: 10-12 1. um/sec.
c) The halide technique, which is based on reading the concentration of halogens
in the space surro,inding the sensor (the sensitivity runs down to 10-6 1� Um/se0.
d) The radioactive method, which is based on reading the gamma radiation of a
radioactive gas which penetrates inside the package during preliminary pressuriza-
tion of the product being tested (a sensitivity of down to 10-9 1- um/sec).
e) Indirect methods of testing for a hermetic seal, which ar.e based on the
change in the electrical parameters of the product being tested by virtue of thr
intrusion of a liquid inside ttie package (pressurization of the devices in wdter
or acetone, exposure for several days in a heat and moisture chamber at a tempera-
ture of 40 + 5� C and a relative humidity of 95 to 98%).
The use of a particular hermetic seal testing technique is determined based on
the specific structural design and production process features of the products
being tested so as to assure a reliable estimate of product quality. For example,
the utilization of the mass spectrometry method do,~s not preclude the necessity
of checking for the presence of inedium and large leaks, since where they are
present, the helium which was introduced into the device beforehand during hermetic
sealing or pressurization can escape prior to its testing. A significant compli-
cating factor when testing with mass spectrometry can also be the presence of
the flow of helium desorbed by the product package. In certain cases (polymer and
ceramic packages), the desorption is so great that it is commensurate with a leak
in the package. Mass spectrometry testing does not yield an objective estimate
in this case.
- 190 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
2
Figure 8.11. Basic schematic of the automatic unit for checking
a hermetic seal.
The utilization of the vacuum-liquid method of testing products, in which polymer
materials are present, can lead to a loss of seal because of the dissolution of
the ^ompound.
When selecting a method, it is also necessary to take into account technical and
economic indicators, which govern the cost, production process suitability and
equipment productivity for hermetic seal testing.
The most widely used equipment is based on the utilization of bubble (vacuum-
liquid) and mass spectrometry methods. ,
A schematic is shown in Figure 8.11 which illustrates the operational principle
of an automatic unit for testing for hermetic sealing and the interaction of its
mechanisms. The automatic unit consists of a vacuum system and the standard PTI-6
helium leak detector. After starting the leak detector 11 and obc3ining the
requisite vacuum, cases 1 with the devices being tested are loaded into magazine
21, from which they roll down to feed mechanism 20 via thP troughs. When the
slide of the feed mechanfsm goes to the extreme rear position, the case which
is located on its upper surface falls down and appears in front of the slide. When
the slide moves forward, it pushes the case, which is located in the working
position, and takes it place.
-191-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
. _
pump.

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
Mechanism 22, which is controlled by cam 19, presses the case against the seal 18.
Cam 16 opens valve 7, and th.e preliminary vacuum exhaustion of the casing is
accomplished. Then valve 7 is closed and valve 8 is opened, which is driven by
cam 15. The high vacuum exhaustion of the case takes place. The further rotation
of the distribution shaft 17 leads to the closing of valve 8 and the opening of
valve 9 by means of cam 14, where valve 9 connects the vacuum volume of the auto-
mated unit, which is located to the right of valve 9, to the vacuum system of
the leak detector. All the other valves of the automatic unit are closed at this
time. In the case of a seal failure of the device being tested, the helium partial
pressure increases in the mass spectrometer analyzer of the leak detector and a
signal appears which is amplified by an electrical circuit in the automated unit
and causes the actuation of relay 5, which plays the part of an electromagnet,
and lock 4 and barrier 3 are opened in this case.
In order that the relay does not actuate before the case with the device which
caused the increase in the mass spectrometer current is dumped, there is a time
_ delay r21ay in the electrical circuitry which delays the signal to relay 5 by
a few seconds, i.e., by the amount of time from the appearance of the leak detec-
tor signal to the contacting of the case. The case with the unsealed device is
rolled through the open barrier 3 into the collection holder 2 for re3ected
devices. If the device is hermetically sealed, lock 4 remains closed and the
case rolls via barrier 3 into collection holder 6 for good devices. Valve 10,
which is actuated by cam 13, opens immediately after valve 9 closes, and remains
open until valve 8 opens.
Valve 10 serves for additional pumping out of the vacuum system of the automatic
unit, located to the left of valve 9, as well as the vacuum system of the leak
detector. In the case a large quantity of helium from a heavily leaking device
gets into the vacuum portion of the automatic unit and leak detector, this
exhaust line helps to speed up the preparation of the automatic unit for the
execution of the next testing cycle. If it turns out that such a quantity of
helium has gotten into the leak detector that it cannot removed over one prepara-
tion cycle, the large leak blocking circuitry actuates and motor 12 of the auto-
matic unit is cut off. The motor is turned on again only after the system is
completely ready.
The PTI-6 standard mass spectrometric helium leak detector operates in the follow-
ing manner. The molecules of helium which enter the vacuum system of the mass
spectrometric analyzer along with the molecules of other gases and vapors are
ionized by electrons, emitted by an incandescent cathode. The ion beam, which is
subjected to an accelerating voltage, exits through the slot c� a diaphragm stop
- into the mass spectrometer chamber, where the ions are segregated wiih respect
to mass in a homogeneous magnetic field. By choosing the accelerating voltage,
the mass spectrometer is set up in sur_h a fashion that only helium ions impinge
on the ion collector (receiver), which is positioned in the chamber at an angle
of 120� to the source.
Following amplification, the ion current is registered by a voltmeter on a
remote control panel. A change in the voltmeter readings is evidence of the
presence of a leak and of its size. An electronic autorecording potentiometer or
other instrument can be connected to the control cansole.
-192-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
FOR OFF[C[AL USE ONLY
CHAPTER NINE EQUIPMENT FOR TESTING THE ELECTRICAL PARAMETERS OF SEMICONDUCTOR
DEVICES
The electrical parameters of semiconductor devices are measured in practically
all stages of their manufacture, beginning with the processes for producing the
structures right up to reliability tests. The major task of testing parameters
in the initial manufacturing stages is not to pass on to subsequent production
process operations those devices whose parameters are worse than the values
established befnrehand. The parameters of semiconductor devices are measured
at the conclusion of a production process primarily for the purpose of distribut-
_ ing them in vroups in accordance with the limiting values of the parameters esta-
blished for each group by the technical specifications. The third major function
of parameter testing of semiconductor devices is the checking of the capability
of the devices of maintaining their properties when exposed to various media:
temperature, humidity, pressure, vibration, shock, etc. In this case, the
electrical parameters serve as the crite-~�ion of immunity of the semiconductor
devices when exposed to the perturbations enumerated above.
Depending on the level of inechanization and automation of the major and auxiliary
operations, the quality control and measurement equipment can be broken down
into manual, semiautomatic and automated. Included in the manual group is equip-
ment for which the loading and unloading operations of the products being meas-
ured, as well as the reading of the measurement results, are accomplished by an
operator. In automated testing and metering equipment, the loading, contacting,
oriented unloading and sorting of the measured products in accordance with the
measured parameters are realized automatically. Semiautomatic equipment occupies
an intermediate position. In it, only the loading of the products being measured
is accomplished manually by an operator, while the remaining operations are
realized automatically.
In terms�of the information obtained from quality control and measurement equip-
ment, it is broken down into equipment for parameter measurement, which makes it
possible to measure the true value of parameters, and classification equipment,
which sorts the devices being measured into groups depending on the aggregate of
measured parameters. For classification equipment, the sequence for parameter
measurement, the comparison of the measurement results with the specified refer-
ence value and the logic processing of the measurement results of all of the
parameters for the purpose of determining the group are all carried out automatic-
ally. The equipment breakdown given here into measurement and classification
equipment is conditional to a considerable extent, since at times the same devic
can perform both measurement and classification tests.
9.1. Measurement Equipment
Measurement equipment is used when tests are made by the quality control depart-
ment services as well as during various tests, in laboratories during developmen-
tal work, in trial production when placing new devices in production, during
input quality control by consumers of the devices, etc. As a rule, equipment of
this type is designed for testing one or more parameters of the same type and
has a comparatively simple design. Meter type measurement instruments are most
- 193 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFF[CIAL USE ONLY
frequently used as the indicators in it. Digital meters for semiconductor device
parameters have started to become more widespread of late.
A significant quantity of specialized parameter meters intended for measuring one
or more parameters of a certain type of semiconductor devices is in operation at
enterprises fabricating semiconductor devices. As a rule, the measurement con-
ditions are established automatically after the product being measured is connect-
ed, the range of ineasured values is small, the switching from one parameter to
another is accomplished ma.nually and the measured value is read out visually from
the scale of a meter or a. digital display. Such parameter meters are used in
the case of a small production volume, during trial production or in other cases
where the application of complex automated equipment is absent or ineffective.
Moreover, there are universal meters, which are ciistinguished by large ranges of
operating mode settings and measurable values. This equipment is used in research
laboratories for the incoming testing of semiconductor devices and for the
measurement of their parameters when repairing various radioelectronic systems.
Several types of such all purpose parameter meters are being produced by domestic
industry.
;ipe06p830uaxene
uar,puxcmtH
(1)
RC -
4) rexeparop
:1cHnxrenb
L';t.i 1NH01IbTY8T{18
(6)
CXBMtl N3Y@Q8-
uHx nepauexpos
3
NauepNrenbxoe
ycTpaNcreo
rieter
CKCYhI N3YCP8HNfl
napauetpoa
rpnxaAcTOpoa O
3 1
~ 41
1
I
. I
~
NCY041M8 (3~
~Z
nxrauna ~
e4yzpeNnxN
03
1
0 41
1
sJ
02
OCTG4NNH
"1
IINT~I~N BMBWMNP~
03
J
94
Figure 9.1. Block diagram of the L2-23 meter for semiconductor
device parameters.
Key: 1. Voltage converter;
2. Circuits for measuring diode parameters;
3. Internal power supply;
4. RC oscillator;
5. Circuits for measuring transistor parameters;
6. Millivoltmeter amplifier;
7. External power suppiy.
= The L2-23 parameter meter for semiconductor devices is intended for measuring
the major parameters of p-n-p and n-p-n transistors as well as semiconductor
diodes.
- 194 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
A block diagram of the meter is shown in Figure 9.1. The meter is powered from
an internal power supply consisting of two "373" batteries; moreover, the meter
can also operate from an external power source. The high voltage needed for the
measurement of the inverse current of semiconductor diodes is obtained by means
of converting the DC voltage from the internal source to an AC voltage and subse-
quently rectifying it. The 760 Hz alternating current signal needed to measure
h21b and h22b is generating by an RC oscillator. The amplifier of the millivolt-
meter amplifies the AC signal which carries information on the parameter being
measured up to a level sufficient for the deflection of the meter needle. The
parameter measurement circuits for semiconductor diodes and transistors provide
for the connection of the devices being measured and the switching of the measure-
ment resistors and power supplies. We shall analyze the measurement of one of
the most important parameters_of transistors: the common base current gain, h21b,
which is accomplished using the circuit depicted in Figure 9.2. The parameter
h21b is defined as the ratio of the change in the output current of the transistor
to the change in the input current where the output circuit is short circuited
for the alternating current. The short circuit mode is realized by inserting a
capacitor C4 in the collector circuit of the transistor. The alternating current
input signal is generated by the RC oscillator, which is the current generator;
this is assured through the insertion of resistor R2, which considerably exceeds
the input impedance of the transistor. In this case, the alte-nating current of
the emitter-base junction of the transistor being measured is governed by the
- voltage Ugen and the resistance of R2, and in the case of constant values of L'gen
and R2, tFie input current will be the same for all of the transistors being
measured. Thus, the measurement of h21b reduces to the measurement of the tran-
sistor output current. Since the value of h21b is close to unity, the ratio of
the base current to the emitter current
L~ R is usually measured to improve the
 measurement accuracy:
L
C~ Rz ~ _ K _z
S~ N3M (1~
F3 � 2
( 3 ~ ~Z
v~
s
T Rs
M
I C
+EK
Figure 9.2. Circuit for measuring the
current gain, h21b, in a
common base configuration.
Key: 1. Measure;
2. Calibrate;
3. Ugenerator�
=18-'K=1 'K cl-Ih21aI� (9.1)
�~a ~a ~e
The value of the input current ie is
measured as the voltage drop across
the calib.rated precision resistor R4,
while the base current is measured
as the voltage drop across resistor
R5, which should also be a precision
resistor. Resistors R3 and RZ should
either be equal or differ from each
other by a ratio specified beforehand.
The gain is thus:
1-~h:tGl -UGR, (9.2)
Unan Rs
Since Ucal, R4 and RS are constant quantities, the scale of the meter measuring
Ub is graduated directly in the values of h21b�
-195-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USF: ONLY
One can test the following parameters by means of the L2-18 digital low power
transistor parameter meter: the quantity 1/(Bst + 1) (Bst is the static gain);
the collector voltage UCEO lim, at which the onset of the change in the phase of
the base current begins; the voltage between the collector and the emitter in the
saturation mode UCE sat; the voltage between the base and the emitter in the
saturation mode UBE sat; the inverse collector current ICgp; the inverse emitter
current IEgO; and the floating emitter potential UEg fl�
I^ - - - - - - - -
- UpeoEpeaosasens CQttV81`t8r~ Power SuPPl Y
~
moprxposa'reAf
Mmnyaacaoro
roKe
Bcrovare Cse6Maraesop
( 2 -200 B roKa O,I-20 rA
6aos ;
( 3)I nw:exxe ~
I
.
~ "T . ~
I 3aAaMw9
(4) renepa:op
KOYYy?8lOP
(5)
~

-
O
Aunyascueio
I
re0poao0 eonasHe:p7
YoAYansoP ~
ycrnp:ena ( 6 )
(
(
tlegosuR
~ 7)
I CpeaNxseunee I.
ycrpoRciao
9 I
icrnr:ena j
xraxoq
vecto:S 11
~ ne:extoP
I
~
I baox xornaxca- I
I Pyn^tero nanpa-
Aeteszop '
xxx
-
~ I9HMA I
c en
~H848NMH ) ~
I
6110K 8D20Y9TN- I
~
~
I
vecxoro ynpaa-
I
I r:s1ixa
I 6nux we~xxeuxr (15)
~
,

L_-----L

Figure 9.3. Block diagram of the L2-18 digital meter for low power
transistor parameters.
Key: 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
Pulse current driver;
2 to 200 volt supply;
0.1 to 20 mA current regulator;
Master oscillator;
Switcher;
Pulse amplifier;
Peak detector;
Digital voltmeter;
Comparator;
Modulator;
Low frequency amplifier;
Compensating voltage unit;
Mean value detector;
Automatic control unit;
Display unit.
-196-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The procedures for measuring some of these parameters will be treated below.
The meter consists of three units: the converter, a digital voltmeter and a
poweX supp.ly (Figure 9.3).
The converter and the automatic range selector provide for feeding a TiC voltage
to the digital voltmeter input in a range of 0.1 to 1 volt, regardless of the
parameter being measured and its value. The conversion factors for each parameter
and the units of ineasurement are indicated on the front panel of the meter.
The transistors being tested are connected to the measurement circuits through
a switcher. Various connections are made in this case, depending on the para-
meter being measured.
The Measurement of 1/(Bst + 1). The base of the transistor being tested is
connected through the measurement resistance to the common bus. A voltage of
2 to 200 volts is fed to the collector from the supply. The emitter is connected
to the outpvt of the pulse current driver, which converts the voltage pulses of
the master oscillator (a push-pull multivibrator) to current pulses with an
amplitude variable in a range of 1 to 200 mA. Voltage pulses are picked off of
the measurement resistor, the amplitude of which are directly proportional to
the quantity 1/(Bst + 1). These pulses are amplified by a two stage pulse ampli-
fier and converted to a DC voltage by the peak detector. The output voltage of
the peak detector is fed to the input of the digital voltmeter.
The Measurement of U EO lim� The transistor being tested is connected just as for
the measurement of 1~(Bst + 1). The output voltage of the peak detector is not
measured; only its polarity is analyzed, which is indicated on a light display
panel in the form of a> sign when the phase changes and a< sign in the
absence of a phase change. The measurement is performed as follows: the voltage
of the 2 to 200 volt source is changed until the sign < changes to the sign
at this moment, the voltage at the collector of the tested transistor is measured
with an external voltmeter, which yields the quantity UCEO lim�
Current Dieasurement. The transistor is connected in the appropriate measurement
configuration. The current being measured is converted to a voltage proportional
to it, which is first fed to a modulator, and then to a low frequency amplifier,
an average value detector, and finally, to a digital voltmeter.
The digital voltmeter is an automatic compensator with discrete equalization. It
consists of the following assemblies: a comparator, a compensating voltage gene -
ator, an automatic control unit and a display block.
A voltage proportional to the parameter being measured is fed to one of the two
inputs of the comparator in the digital voltmeter. The compensating voltage is
fed to the other input, which is changed discretely in accordance with a program
governed by the operation of the automatic control unit. There is a programmer
in this unit which is triggered by a pulse from the automatic range selector
following the completion of the selection. The program actuates the flip--flops
197 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL L1SE ONLY
of a storage register in a definite sequence, which in turn switch on the requi-
site compensating voltage. The flip-flopa remain turned on if no pulse arrives
from Che output of the comparator, or turn off, if such a pulse arrives.
The presence of a pulse at the output of the comparator means that the compensat-
ing voltage is greater than the voltage at the input. Upon the completion of
the measurement cycle, a tetr,idecimal code (4, 2, 2, 1) is registered in the
flip-flops of the register memory, where this code corresponds to the state of
equality of the compensating and measured voltages, i.e., the digi.tal equivalent
of the parameter being measured. The tetradecimal code is fed to the input of
the display unit, where it is converted to a decimal code. The image of the
decimal numbers is produced by means of IN-1 neon indicators.
BMP!!!N4!!
~ ~ `L~C/170YNflA
(2)
~ 1 ~ oN~msaG
damcvNU~'
Figure 9.4. Block diagram of the PNKhT-1 instrument [transistor
characteristic curve family tracer].
Key: 1. External supply;
2. Horizontal deflection X amplifier;
3. Step function generator;
4. Volts/step;
5. mA/step;
6. Collector supply;
7. Indicator (CRT);
8. Power supply;
9. Vertical sweep Y amplifier.
The comparator takes the form of a DC voltage amplifier with periodtc drift
correction. The amplifier consists of two vacuum-tube and six transistor stages.
The compensating voltage unit consists of a Y configuration potentiometer, de-
signed for a tetradecimal code with weights of 4, 2, 2 and 1, and a group of
relays, the contacts of which switch the potentiometer resistance. The relay
coils are controlled by signals from the flip-flops of the memory register of
the automatic control unit.
- 198 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
A provision is made for the capability of connecting a recorder (printer) to the
neter; a voltage drop with an amplitude of 5 volts is used to trigger it. The
information for the recording is fed out in the 4, 2, 2, 1 tetradecimal code.
Transistors can be rejection sorted with respect to one of the parameters by the
~ meter. The meter takes the fona of a desk top type instrument.
Important information on the electri.cal properties and quality of semiconductor
devices is contained in their volt-ampere characteristics. A whole series of
instruments exists for the visual observation of the volt-ampere characteristics
of semiconductor devices. A block diagram of the PNKhT-1 instrument for observing
transistor characteristics is shown in Figure 9.4, while an external view of this
instrument is shown in Figure 9.5. The major assemblies of the PNKhT-1 instrument
are:
--The collector power supply;
--The step function generator;
--The vertical sweep amplifier (the Y amplifier);
--The horizontal sweep amplifier (the X amplifier);
--The indicator;
--The power supply.
The PNKhT-1 instrument makes it
possible to observe both the families
of transistor characteristics as
well as the volt-ampere characteris-
tics of p-n junctions. To obtain
a family of transistor characteris-
tics on the cathode ray tube screen,
a pulsating voltage is fed to the
collector which is obtained by
rectifying a sine wave and which
serves for the sweep. A step func-
tion changing voltage is fed to the
input of the transistor being tested,
where this voltage is used as the
argcment signal. The voltage across
the junctions of the transistor being
Figure 9.5. The PNKhT-1 scope for observ- studied or the voltages proportio 11
ina transistor characteristics. to the currents through these junc-
served) are fed to the X and
fed to the deflection plates
circuit configuration for a
characteristics of the type
in a common emitter circuit
. tions (depending on which character-
istics of a transistor must be ob-
Y amplifier inputs, and following amplification, are
of the CRT. By way of example, we shall consider the
test transistor to observe the family of output
IC = f(UC) for different values of the base currents
(Figure 9.6).
-199-
FOR OFFICiAL USE UNLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
Ii -1.I
FOR OFFICIAL USE ONLY
In this case, a step voltage is fed to
the base of the transistor, while a half-
wave of the sinusoidal voltage is fed to
the collector. The signal picked off of
resistor R2, which is proportional to
the collector current, is fed to the ver-
-.o ' tical sweep amplifier, while the signal
'(IKl from the collector-emitter junction if
fed to the horizontal sweep amplifier.
. The pulsing and step voltages are syn-
chronized so that during one half-wave,
the current in the base of the transis-
Figure 9.6. Common emitter circtiit con-
figuration for plotting
characteristics of the
type Ic = f (Vc) .
istic is traced on the indicator screen.
of the collector voltage and the next ba
i.stic is traced on the indicator screen:
tor remains constant and one character-
istic is traced during this time. During
the time of the next half-wave of the
collector voltage and the next step of
the base current, yet another character-
During the action of the next half-wave
se current step, the following character-
Ic = f (UC) , etc.
The 13L06I cathode ray tube is used as the indicator; a transparent scale with
a grid is placed in front of its screen. Structurally, the PNKhT-1 instrument
is made in the form of a table top unit. A sloped panel with the contacts for
the connection of the semiconductor devices to be studied is fastened to the
lower portion of the front panel of the unit.
9.2. Classification Equipment
A typical representative of classification equipment is the all-purpose low power
transistor classifier, the KT-2 (Figure 9.7a). This is a semiautomated classi-
fier with a carousel transistor transporter 5, which is structurally separated
from the support frames 1 with the measurement and operating control units, and
the support frame 2 for the logic and computer units. The classifier has pro-
gramming panels 3, test probes 4 and receiving hoppers 6. Universality is
achieved through the capability of changing the classification program, test modes
and limiting values in an operatianally timely manner; the modular structural
design makes it possible to rapidly change the composition of the parameters
being tested by changing the instrutnentation units.
A block diagram of the classifier is shown in Figure 9.7b. The transistors being
tested, with the prestraightened leads.are loaded by the operator in an ariented
manner in the connecting heads 3. The latter are mounted on the carousel 10
and execute a start stop motion together with it, because of which each connecting
head sequentially passes by all of the measurement Fosts. A measurement post
consists of the test probe 2 as well as the measurement and test condition setting
units 9.
The test probes are installed above the carousel. The transistor pins are con-
nected directly to their contacts, which are located in the bottom end face
- 200 -
EOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
,

:;v
Figure 9.7. The KT-2 all-purpose classifier
for low power transistors.
a. External view;
b. Structural schematic.
d during the time the carousel is stopped. The
input circuits and the measurement unit devices
are placed inside the test probes, tihich should
be brought close to the test object to assure
noise immunity and reduce the influence of
parasitic leakage and reactances.
The values of a particular parameter of the
transistor being tested are compared in each
measurement block with its limiting value which
(b) is established beforehand. The comparison re-
sults are transmitted in the form of a binary
code (1 is a value measured greater than the
limit; 0 is a value measured less than the limit) for storage to one of the shi�t
registers 6 of the logic and computing unit (SLU) 8. The storage is needed be-
cause the rejection sorting is based on the measrrement results for all of the
parameters.
Corresponding to each measurement post is its own shift register. In step with
the travel of the tested transistor from one.to measurement post to another, the
- 201 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500090004-3
F'OR OFFICIAL USE ONLY
data on it is accumulated in the registers and is moved simultaneously in them
from the input to the output. The capacity of the registers is not the same;
it is numerically equal to the number of the measurement post to which the regis-
ter belongs (the count starts with the unloading post, to which the number 0 is
assigned). Such a configuration of the shift rzgisters provides for the synchron-
ous output of all data on the transistor being tested to the data processing unit
7 at the moment the transistor arrives at the unloading post. The shift registers
are designed around ferrite-transistor memory cells. The data processing block,
in accordance with the program itered in it and based on the information arriving
at it, generates the instruction for the sorting rejection mechanism 4, which
routes the transistor to recei.ving hopper 5.
The c?assification program is composed on the basis of the technical specifications
and is entered on t;iu program cards in the form of holes punched through at the
requisite points. The cards are placed on the prograIIt panels of the SLU, and
_ special electrical contacts are insei-ted in the rezeptacles for these panels,
which match the holes in the cards.
The operation of all of the measurement blocks is synchronized through the SLU
from the contacts which are closed by cams 1, which rotate synchronously with the
carousel.
SEt-I electromechanical counters are mounted in tr.2 classifier, which count the
number of transistors in each group, the overall number of transistors which have
been sorted, and the number of transistors having a negative test result at any
of the six measurement posts.
In the majority of ineasurement units in the classifier, the compensation measure- ment method is employed, because of whicr:, stringent requirements are not placed
on the stability of the gain and the linearity of the amplifiers.
Voltage and current regulators, regulated pulse current generators as well as low
and high frequency generators are used as the test condition setting units in
the classifier. A provision is made in the test condition setting blocks for
the capability of adjusting the output voltages and currents.
The device for transporting the transistors, a kinematic schematic of which is
shown in Figure 9.8, operates as follows. All of the mechanisms are driven by
the distribution shaft 24. Rotation is coupled to the distribution shaft from
electric motor 1 through V-belt drive 42 and a worm gear reducer (worm 39 and
worm gear wheel 23). A number of auxiliary elements are installed in the trans-
mission from the electric motor to the distribution shaft: safety clutch 43,
free-wheeling clutch 41, which permits only one-way rotation of the drive, and
electromagnetic break 40, which is actuated when the electric motor is turnEd on.
Gnd cam 20, the Geneva mechanism carrier 21, the conical gear of reducer 37 and
cams 38.
- During the rotation of end cam 20, once every revolution of it the rack 11 is
raised and lowered, which rotates two shaft-gears 10 (one directly, and the other
- 202 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE OWLY
2
f
J
y 5 6 7 B
ZJ
x
va x
V/
~i0
q
J!
q K7
!0 11 /2 /3 ~s
f
l2 Z/ 20
I
-
x Zs q217.B7. d2 J3
. Zq
p ZS O~~ O o0 0
~
U J6 JS d9
1.' 37
J8
Figure 9.8. Kinematic schematic of the KT-2 all-purpose classifier.
through rack 9). Eccentrics 6 are mounted on the shaft-gears, which when rotating,
raise and lower disk 7. When the disk is lowered, levers 4 take a position such
that the leads of the transistors being tested, which are installed in the con-
necting heads 12, are pressed against the contacts of test probes 3. When disk
7 of lever 4 is raised, in rotating, the leads of the transistors are disconnected
from the contacts of the test probes.
The levers 4 are instalied on the carousel 5, which executes a start-stop rotating
motion. The carous~.! rotates at that point in time when disk 7 is lifted. The
start-stop motion of the carousel is realized by means of a Geneva mechanism,
the carrier 21 of which rotates along with the distribution shaft 24 and rotates
the Maltese cross 22. The backlash free gear 8, which is engaged to the carousel
gear, is mounted on the same sMaft as the cross.
,
The transistors are off loaded and sorted into groups at a special post. When
lever 4 arrives at this post and its arm with the connecting head is lifted up-
ward, the package of the transistor being tested enters into the fork of unloading
lever 14. By this point in time, the cup on lever 31 is set under tray 13. Lever
14 is driven by shaft 25 through gears 19 and 18, cam 17, lever 16 and pull rod
15. With the rotation of lever 14, the transistor is removed from the connecting
head and by virtue of the weight of the transistor itself, falls through tray 13
into the cup on lever 31. After this, levers 31 and 32, which are coupled through
gea*-s 29 and 28, rake-rod 27 and cam 27 to shaft 25, begin to move and run up to
- 203 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFFICIAL USE ONLY
the turned on electromagnet 30; the lever 32 opens the bottom of the cup and the
transistor falls into one of the hoppers 33.
The number of electromagnet and hoppers corresponds to the number of classifica-
tion groups. The electromagnets are actuated by signa?a from the logic computer
unit.
Shaft 25 is driven by distribution shaft 24 through reduction gear 37, articulated
couplings 36, shaft 35 and reducer 34. With the rotation of distribution shaft
24, cams 38 periodically turn microswitches on and off, because of which electric-
al signals are generated which are used via the logic computer unit for the
synchronization of the operation of all of the classifier devices. An emergency
shutdown lever 22 is provided in the design of the semiautomatic unit.
Main Technical Specifications of the KT-2 Classi_fier
Output, pieces per hour 1,200
Number of classification groups 14
Number of limiting values for the parameters
being tested 29
Frequency of the signal for :he measurement
of high frequency parameters, KHz 20
Power consumption, KW
3.5
While the KT-2 classifier contains the loading and unloading mechanism for the
transistors being measured and a measurement sec:ion as indispensable parts, the
operation of classifying the semiconductor devices into groups can be realized
if any of the general purpose or specialized meters are used in conjunction with
a separate sorter, intended for the loading, contacting and sorting of semiconduc-
tor device into groups.
By way of example, we shall consider the combined operation of the EM-630 para-
meter measurement instrument and the US-5002 sorter. The EM-630 unit is designed
for testing the static parameters of digital integrated circuits having up to
24 pins based on the "reject - good" principle. The measurement of a sequence of
parameters (tests) is accomplished automatically in accordance with a specified
program. The programming is realized by a combination of special pins, inserted
in the appropriate jacks of the programming matrix. The unit can run either an
entire sequence of tests (up to 78 tests) or terminate the meast+rements following
the first rejection. The parameter measurement process is based on the automatic
comparison of ineasured and reference analog signals. A block diagram of the
EM-630 unit is shown in Figure 9.9. We shall analyze the function of its major
assemblies.
The instrument control circuitry 1 provides for the following: the generation of
a pulse train at a frequency of 100 Hz for the counter and the test number indi-
cating circuit 2; control of the 80 bit register for program indication and indi-
cating the result of each test S; triggering the modulator of the test circuit
- 204 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
i
i

~
!8 I
i s�re n+ee i
I
( A) ~
.ai .~e .~rs �::os
~a
II~ OMIOOON
I
"
I
A'cNnroiw~
* I
~ B~ i
!0 11 1t /J I
~
!9 ~
I
~
~
S 6 ~
~C~ ~
~
6p~w
~ * I
D ~
S~. ~.J
I
ewnyrw
! 1 ~
ws~'erplw~
(E)
-------J
Figure 9.9. Block diagram of the EM-630 Figure 9.10. Kinematic schematic of
tester for checking the the US-5002 sorter.
parameters of integrated
circuits 10 in each test; generating the "end of
Key: A. To the integrated circuit `est" signal, which is fed to the exter-
leads; nal monitor 20. The test number indi-
B. Test; cating circuit and pulse counter 2 pro-
C. Re3ect; vide for counting the test pulses, in-
D. Start test; dicating the number of the test and
E. End test. control the operation of the 80 bit
register for indicating the results of
each test 5; they also control the program setting circuit 4. The program setting
circuit serves to specify the program for each test, and store the program for
all tests and feed out Che program for any test to the 80 Uit program register 6.
Program register 6 receives the program for each test from the program setting
- circuit 4, converts it to voltage signals, wb.ich then control the elements pro-
vided by the program of the given test. The integrated circuit lead and test
signal circuit switcher 14 provides for connecting the supply voltages generated
by.block 17, the test signals incoming from block 16, and the loads located in
load unit 15 as well as the reference registers for checking the currents, which
are located in block 13, to the pins of the IC being tested in accordance with
the program. Moreover, the switcher provides for the connection of the voltage
_ being measured to test circuit 10, to which the reference standard signal is also
fed from�code to voltage converter 11, where the measured and reference standard
voltages are compared, the polarity of the difference signal is determined and
the signal that the IM is good in terms of the given test is fed out. Register
5 stores and indicates the test result of each test, feeds this result out to
the test data processi.ng circuitry 3, which in turn generates the "reject" signal,
which is fed to the external monitor. Power supply 18 is intended for powering
all of the units uf the EM-630.
- 205 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLV
The Main Technical Specifications for the EM-630 Unit
Nur~~r of quality control tests 78
Duration of a quality control test, msec 10
Number of leads which can be switched 24
Range of voltages which can be checked, volts 0.01 to 9.99
Range of currents which can be checked, A 0.1 � 10-6 to 99.9 � 10-3
Voltage test error, mV +(1% + 0.5)
Current test error, nA +(2y + 10)
The kinematic configuration of the US-5002 sorter (Figure 9.10) consists of
three sections, which are coupled together pneumatically. The major section,
the travel mechanism 2, receives a reciprocating motion from a pneumatic cylinder,
the rod of which is rigidly coupled to the guides of the travel mechanism. The
pruduct loading mechanism 1 functions cyclically with the latter, where the
operation of this mechanism is likewise based on the reciprocating motion of the
rod of another pneumatic cylinder. The third section is composed of the unload-
ing and sorting mechanisms 4 of a modular design with individual pneumatic cylin-
ders. The pneumatic cylinder rod of each mechanism is rigidly coupled to the
three other unloading rods. The drive for each unloading mechanism is pneumatic-
ally coupled to the main drive (the travel mechanism) in such a way that the
unloading rods begin to move only after the travel mechanism is passed by the
guides two-thirds of the way to the position of the elements. The return travel
of the working elements of a11 three mechanisms is accomplished simultaneously.
The cycle time can be adjusted by two factors: the standstill time of the travel
mechanism in the measurement and unloading positions (electrical control); chok-
ing down the cross-sections of the internal bores (pneumatic control). The
loading mechanism 1 pushes out one product each from the case (or other loading
device) during each cycle, where this produce is the one whose parameters are to
be measured. This product is moved by the travel mechanism to the contact posi-
tion, where the product being measured is connected by means of the contacting
device 3 to the power source setting the conditions as well as to the measurenent
circuits [through switcher 14 (see Figure 9.9), if we are speaking of the opera-
tion of the US-5002 sorter in conjunction with the EM-603 measurement instrument].
During the return motion of the travel mechanism, the product whose parameters
have just been measured, enters the unloading and sorting mechanism, and depend-
ing on the measured parameters and signals from the measurement unit, goes into
the appropriate case. The products being measured are housed in this case in
special satellite carriers with standard dimensions, which makes it possible to
standardize the transport assemblies, loading, unloading and contacting devices.
9.3. Automated Systems Using Computers for Parameter Testing
Quality control and measurement complexes which contain a measurement unit and
computer (EVM) have become widespread recently, where the computer controls the
feed of the mode setting currents and voltages to the product being measured,
provides for switching the pins of this product in accordance with the
-206-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL i
measurement circuit for a particular parameter, as well as the measurement of
the parameters and the processing of the measurement results. In this case, the
entire sequence of parameters is measured automatically in accordance with the
cumputer program. Following the completion of the measurement process, the
computer makes the decision as to the conformity of the parameters of the meas-
ured product to their specified values. Information on the measurement results
can be fed out in various forms: the true value of the parameters, printed out
on paper by a nwneric printer operating as part of the computer hardware; a
"reject - good" light signal on the panel of the measurement unit; control sig-
nals to the sorter for the products being measured, by means of which the pro-
ducts are broken down into groups.
r (3) ,ppo

(1~ Mareprseas 7T-i ~
2 ~4-8-e peaPAlW
.I2 Bi
+12 V
3~5-e psspe7pi
~
~ AcaOpatop
(7)
i
� ~ ~ N
�p~~~e
: 6) ~
0YR lbChl sao -s+so
AlOQUlIOYaMA
'St8~t.11
~
(9)
(21) ~
( 8
BxoAxue
"
"Haaaao harepesre
I
3B1(
~
s
easus
.
et
p
~eyecrso
"K
baox
OHOU " cwxzpo-
I
xarax7terr
eraeu~M
"Nareprreas ro:os
I
Pormarp
~
PerrcrP
AeNtl1Ct
'
(14~
"Peayaa:et rozos"25 (22)
I
1
L)
I
( 1
C~aeop
J(9220PATOp
NeAtrepre
ea~as:azos
^Ctp06
a"
e:o
a
~
~
~
Aexeea
p
ea�ecw~~. qr~
p
p
Qorn
ara
"C
or i
13~
omp
~
.
Gating
~
I
I H862p8:eas
' aoreps
aaana (16 )
baor seAeeNe
ycaosrA areae-
~p" rasop
( 9)
Hornapascp
~
r
; ~
rra  rpsaaa-
om
ara
p
eoro sesaeaxe
~
Permosp
(18)
~
!
peelasrare
I
, I
NomIsyeroe
I
ranene (26
~
~

Figure 9.11. Block diagram of the UT-1 meter.
Key: 1.
UT-1 meter;
9.
Input gates;
2.
Bits 6 to 8;
10.
Buffer register;
3.
"Pass";
11.
Adder;
4.
Bits 3 to 5;
12.
Data register;
5.
Decoder N;
13.
Data decoder;
6.
Input-output pulses;
14.
Instruction exchange unit;
7.
Instruction decoder;
15.
Classification result
8.
Computer;
display;
- 207
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFF'ICIAL USE ONLY
Key to Figure 9.11 [cont.]: 16. Plan number selector;
17. Result register;
18. Unit for specifying the bias and limiting vaZue
conditions;
~ 19. Switcher;
20. Comparator;
21. "Start measurement";
22. Synchronization unit;
23. "End of tests";
24. "Meter ready";
25. "Result ready";
26. Product being tested.
Meter ready
Htono~~b StSrt
Start nycfi N�vano
(2~
?deter N~"'cpu- ncpedava daN-
m~Ae Ndr /-ro iecra
read
B
ynb -
,armres
p,~,.~~d-
x~epe~
~nynQ
llep')dava
111Vs.7-JoI
lfcpedava
Pt~ynaan
5
y aamo
-
(12 ) avanv uimcorNa.v �
~
6N115Nnn.n. ~
.
llpnnye,r
(13._
.
6-N12 6"Iz
6N21 LZ~. .
~
511". .
6N04
~
6NUl
.
(14 ~ Ptaynemam romo4 "
.
6NOt
~ 15 HaNeu rttneimaNu
ti 'J
(16 ) .Cnot'o6 Townapamopa " 26-JOwC /OMC
~ 17 ~.R.YJURQ!(flA PfJf//16/AQ/n? M~16CC//~tid'Ql(~IL
Start of n-th test
No.oeo n-ta ncra
10
Figure 9.12. Time diagram showing the interaction of the UT-1 meter
and the computer.
Key: 1. Result ready;
2. Transmission of the data
of the first test;
3. Waiting for the result;
4. Transmission of the data
of the n-th test;
5. Transmission of the re-
sult to the computer;
6. Result ready;
7. Transmission of the classi-
fication result to the meter;
8. Waiting for the result;
9. Transmission of the result
to the computer;
- 208 -
FOR OFFICIAL USE ONLY
" j .
) neoedava pwy,j-
mama 0Aaau- ieter ready
rarmroe ~u,~.auuu t~ - arc0umena
OacnBa- Me umtwb ~os llcpedcva dox-
xut pe- tpe owa pary n Le Nyz I-ao 11
w~,a e aeM y muma u m. d( ~
10. Start;
11. Transmission of the
data of the first test,
etc.;
12. "Start measurement";
13. "Pass";
14. "Result ready";
15. 6N14 ("end of tests");
16. "Comparator mode";
17. "Classification result
displ.ay".
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500094444-3
FOR OFFIf
The measurement time for a single parameter usually amounts to a few milliseconds,
and thus it is possible by means of a computer to carry out the testing operations
for several tens of parameters of a semiconductor device, for example, an inte-
grated circuit, in fractions of a second. The high productivity of parameter
testing is the major distinctive feature of computer controlled monitor and
meaaurement equipment. The sAcond feature is the measurement pi�ecision, which is
due to the use of the "weighting" measurement procedure and the comparison of
- the quantity being measured with a reference value. A characteristic representa-
tive of the family of computer contro?led measurement aitd claasification equip-
ment is the all-purpose UT-1 meter, which is intended for checking the static
parameters of low and medium power transistors and diodes. The operational
principle of the meter will be clear following an analysis of its structural
- configuration (Figure 9.11) and the time diagram showing the interaction of the
r.eter and the computer (Figure 9.12).
After the system is started, the computer operates by periodically interrogating
the meter readiness starting with the initial program address, by sending the
6N11 instruction to the meter with a period equal to the execution time for two
instructions in the computer. After the arrival of the "Start" signal at the
meter synchronization unit (from the pushbutton on the control panel with manual
loading or from the automatic sorter), the synchronization unit feeds out the
"Meter Reaay" signal to the instruction exchange unit, which with the arrival of
the next 6N11 instruction will feed a"Pass" signal to the computer. After re-
ceiving this signal, the computer leaves the periodic interrogation mode, feeds
out the 6N12 instruction, upon which the meter readiness flip-flop is reset and
the number of the classification plan is transmitted to the computer. Following
this, the initial data for the performance of the first test are transmitted
(1 to 5 words for the 6N21, 6N22, 6N24, 6N31 and 6N32 instructions each) and at
the end, the 6N04 instruction, "Start Test", is fed out. The 1 to 5 words are
written into the data register and are decoded by the data decoder, the signals
from which are fed to the unit for specifying the bias and limit value conditions,
as well as to the switcher. As a result, the requisite circuits are switched for
testing the requisite parameter and the specified currents znd voltages are
applied to the pins of the product being tested.
The instruction "Start Test" is converted in the instruction exchange unit to the
"Start Measurement" signal, which when received by the synchronization unit, the
latter generates the "Comparator Gating" signal 26 to 30 msec following the
"Start Test" instruction. This delay is necessary so that the transient pro-
cesses from the switching are finished bei^re the comparator begins to compare
the actual value of the parameter with the specified ultimate value.
After putting out the "Start Test" instruction, the computer changes over to the
periodic interrogation of the readiness of the measurement result. The interro-
gation is accomplished by means of periodically sending the 6N01 instraction to
the meter. When the test is completed, the synchronization unit will feed out
the "Result Ready" signal to the instruction exchange unit, which upon the
arrival of the next 6N01 instruction, will feed out the "Pass" signal to the
computer. The computer then quits the periodic interrogation mode and feeds out
the 6N02 instruction, upon which tte result readiness flip-flop is reset and the
- 209 -
FOR OFF'[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR OFFICIAL USE ONLY
result is transmitted to the computer. The computer compares the result obtained
with the specified limiti.ng value, and then transmics the initial data for the
performance of the next test. Following the measurement of all of the parameters,
specified by the classification plan, the computer determines the classification
grnup of the product being tested and feeds cut the instruction 014, in accor-
dance with which the classif ication res,ilt (the group number code) is transmitted
to the data register, and after it, to the classification result display. In
accordance uith the 6N14 instruction, the "End of Tests" sig:ial is fed to the
synchronization unit, because of which, the generation of the "Comparator Gating"
signal is inhibited and the "End of Tests" light lights up on the front panel
of the meter.
Rin 1 Ro~r (1) -Uinv
R6.i
- VoQP
+roe (4) ynt-~
+10 VOLtS p_n neptxo
p-n j ung
M
f,
~Rin 2
Figure 9.13. Circuit for measuring the inverse current of a
p-n junction.
Key: 1. Feedback resistor 1;
2. Feedback resistor 2;
3. DC amplifier 2;
4. DC amplif ier 1.
Following the transmission of the classification result, the computer shifts
over to the mode for interrogating the readiness of the meter using the 6N11
instruction, which continues until the next "Start" signal is transmitted. One
computer can control several meters. In this case, each meter is assigned its
own ordinal number and the N decoder in each meter is correspondingly aligned.
The computer periodically interrogates the readiness of the meters and thp meas-
urement results. When any meter feeds out a reply signal to the computer, it
transmits the corresponding data to it and continues to interrogate the other
meters.
By way of example, we shall consider the measurement of the inverse current of a
p-n junction using the UT-1 meter, which is based on the principle of comparing
the measured and reference values. A resistor Rm (Figure 9.13) is connected in
series with the p-n junction being tested. Two voltages are applied to the network
consisting of the p-n junction and the resistance Rm: a voltage Uinv is applied
-210-
FOR OFFICIAL USE ONLY
Comparator
KoMnapamp
9.isoa
Output
10 volts
(2; R~7 d -lOB
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
from the tested junction side, while a reference voltage of opposite polarity
Ee is applied on the side of resistor Rm. When the two currents flowing through
the p-n junction and resistor Rm are equal, the voltage at point 1 is U1 = 0.
If the currents are not equal though, the voltage U1 will be positive or negative.
Thus, the polarity of the voltage at point 1 ie an indicator of the equality of
the measur2d and ultimate values of the currents.
The rollowing can be written for the limiting value of the current:
Ilim Ee/ Rm
(9.3)
The quantity Ilim is specified by setting definite values of Ee and Rm. The vol-
tage Ee is produced by an operational amplifier which takes the form of direct
current amplifier UPT-2 which has feedback. A reference voltage of 10 volts is
fed to the input of UPT-2. The setting of Ee is accomplished by changing the
input resistance of UPT-2, Rin2, since the following equality is justified for
UPT-2:
' Ee Uin(Rfb2/Rin2) ' (9.4)
where Uin = 10 volts is the reference voltage and Rfb2 = KOhm [feedback resistance
2].
Thus, by changing Rin2 in accordance with the computer program, one can produce
various values of the reference voltage Ee at the output of UPT-2. The inverse
voltage Uinv is generated by operational amplifier UPT-1 and is set by changing
the feedback resistance Rfbl�
- The moment the limiting value of the current is equal to the measured value, i.e.,
the point in time when the voltage at point 1 becomes equal to zero, is regis-
tered by th4 comparator, which transmits the appropriate signal to the computer.
In this case, it is not necessary to carry out the testing process until the
point in time when the limiting current value is equal to that flowing through
the p-n junction. It is sufficient in the classification mode to determing
whether the current through the junction is greater or smaller than the specified
limiting value, and to feed the appropriate signal to the computer.
9.4. Contacting Assemblies for Checking the Parameters of Semiconductor Devices
One of the most important mechanisms whose influence is felt on the mean time
between failures, the service life, the productivity of ineasurement and test
equipment as well as the confidence levels of the measured parameters is the
contact making assembly. A contacting assembly (KU) is understood to be that
device which makes it possible to repeatedly connect the devices being tested to
the electrical circuitry, in this case assuring a minimum connection resistance,
Rcong for the electrical current flowing through the plug connection between
the contact and the lead to the product which is connected, as well as minimal
induced currents in the measurement circuits and maximum insulation resistance
between the contacts in a specified temperature range. It is apparent from the
211 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
FOR OFF7CIAL USE ONLY
definition of a contact assembly that its major technical characteristics will
be the follow-Lng:
--The permisaible current, A;
--The permissible voltage, V;
--The capacitance between any contact pairs, pFd;
--The contact inductance, nHy;
--The connection resistance between the contact device contacts and the lead of
the device being tested, ohms;
--The insulation resistance between two contact pairs and between any contact and
the device package, ohms;
--The wear resistance, the number of contact making cycles;
--The operating temperature range, �C.
The nominal values of the technical specifications of contact assemblies for
various semiconductor products are stipulated by the corresponding standards
[55-57J.
Depending on the structural design, contact assemblies ar.�e brolcen down into two
major classes: contact assemblies without a mechanical drive and mechanically
3riven contact assemblies.
Contact assemblies without a mechanical drive are intended for operation in
manual quality control and measurement equipment as well as test stand equipment,
where the operations of loading the products into the contact assemblies and
removing them are accomplished manually. In this case, the leads of the products
fall directly on the contacts of the contact assemblies. Loading is realized in
the majority of cases with lead friction against the contacts, something which
lead.s to the destruction of the coating on the semiconductor device leads and
to rapid wear of the contacts of the contact assembly.

.
:I
I
P
2
3
4
Figure 9.14. The contacting device for
transistors with flexible Figure 9.15. Schematic of the contact-
leads. ing device of ineasurement
equipment which operates
using the start-stop
principle.
- 212 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
Various contact systems are used in contact assemblies without a mechanical
drive deper.ding on the type of semiconductor device packages. For diodes and
transistors with rigid leads, these are usually push-in contacts, where the
insertion and removal of the device are accomplished with friction, while the
contact force is produced by means of the elastic properties of the contact
material and additional springs.
There are seveisl contact system designs for transistors with flexible leads:
bent tubes made of nickel alloy, bronze tubes with a lobe shaped contact bent
back inward, collet chuck type clamps, as well as guillotine type terminals.
Contact assemblies having a collet chuck or guillotine tyr-. contact system make
it possible to insert and remove the devices without frictiin, but are complex
to manufacture and have a low productivity, and for this reason have not found
wide scale applications in the production of semiconductor products.
Contact assemblies for transistors with flexible leads using contacts made of
nickel tubes have become widespread. Such contact assemblies are shown in
Figure 9.14. The contact force in the tubular contacts is created by virtue of
the deformation of each lead of the transistor being tested, which duplicates the
profile of the contact tube. Contact is made at several points ia this case.
Since the insection and removal of the devices are accomplished by overcoming
the frictional forces between the device lead and the contact tube, the protective
coating of the device leads and nonuniform contact wear take place as a consequence
of this. Especially severe wear is observed at points where the tube bends. A
tubular contact provides for a resistance of Rcon 5' 10-2 ohms. The instability
of the contact resistance, which is explained by the differing degree of curvature
and differing cleanliness of the lead surfaces are to be numberPd among the
drawbacks of a tubular contact device. It is necessary that the contact assembly
by mounted in a vertical position, while the lower end should not be clamped or
sealed shuc so as to spontaneously remove from the contact tubes any wear products
caused by friction with the leads of the devices being measured.
Mechanically driven contact assemblies, which are schematically depicted in Figure
9.15, are used in automated and semiautomatic measurement and test equipment
which operates using the start-stop principle. They operate in the following
manner: the leads of the device 3 are automatically inserted in the receptacles
of the support 4 and connected to the electrical circuit by the contacts 2. The
requisite force is produced by elastic element 1. Contact assemblies with a
mechanical drive should be distinguished by a high service life, since they are
intended for operation in high output equipment.
Their serviceability depends on meeting four conditions:
--The simultaneous entry of the leads of a single product during the measurement
time;
--A low resistance and stable contact between the product and the electrical
circuitry during the measurement time;
--Connection without deformation of the product;
- 213 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
--Spontaneous removal o� inadvertent objects from the contact region.
Difficulties arise in meeting the first condition which are due to the ambiguous
arrangement of the leads of the ma3ority of products (diodes and transistors
with flexible leads, all types of series produced IC's). For this reason, it is
necessary to begin meeting the condition with an analysis of the structural
design of the semiconductor product, the quality of the leads and possibl_e devia-
tions from the geometrical shape. ,
The condition for a low resistance stable contact between the product and the
electrical circuitry will be met if Rcon between the contact and the product
lead is no greater than the permissible maximum value for the entire operatiunal
time of the equipment. Greater contact compression forces lead to rapid wear
of the contact and support surfaces. For this reason, in developing contact
assemblies it is necessary to determine the optimal contact pressure depending
on the structural design of the contact, its material and coating as well as
the specif ic nature of the lead of the product being tested (diameter, coating,
etc.).
The most characteristic type of semiconductor devices with an indeterminate
arrangement of the leads are transistors with flexible leads. The transistor
leads having a diameter of 0.2 to 0.3 mm are at times deformed and tangled
together. For this reason, it becomes necessary to straighten the leads of
the transistors prior to measuring the electrical parameters. Special lead
straightening machines have been designed for this purpose. At times, the
lead straightening operation is performed by the contact assembly. One of the
structural designs for a contact assembly kith a comb for straightening out
transistor leads prior to measurement is depicted in Figure 9.16. In this case,
the leads of the transistors being tested 2, which are fed to the measurement
position by means of shuttle 8, pass through the teeth of the comb 7prior to
coming in contact with the contacts 1, where the comb teeth fan the transistor
leads out in a definite manner and the probability of the leads making contact
with the contact areas 1 will be significantly higher than in the case of
unstraightened leads. The comb 7 and the shuttle 8 are not shown in Figure 9.16
(at the left). After the transistor gets into the contact position, contacts
1 are pressed against the transistor leads by a clamping force P, which is
applied to plates 3, which are spring ioaded and secured together with the current
conducting flexible elements 4 on base 5. The conducting elements 4 and the
springs of the plates 3 aLe separated by insulating washers 6.
The vertical i.mpinging of the leads on the contacts is assured by the width of
the contacts, which must always be chosen greater than the possible deviations
of the transistor leads. The task of horizontal alignment of the leads on the
contacts reduces to the condition for the teetYi of the comb getting between the
- transistor leads (see drawing section B-B).
The design of contact assemblies for integrated circuits (IC's) presents con-
siderable difficulty, since IC's have a larger number of leads (usually from
8 to 64) with relatively small package dimensions. In this case, the leads of
- 214 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500090004-3
View A Bud A
d=
2
~n
~
6 + +
. IP _
6 S 4 3 1A
7
B-B
I 6=6
~
Z ~ oa
~6 1
71
Figure 9.16. A contacting device with a comb for splaying out
the leads of transistors.
IC's housed in flat packages with a planar arrangement of the leads, are as a
rule made of inelastic strip materials and are readily deformed. The use of
push-in contacts, similar to the contacts treated earlier for other types of
semiconductor devices, is characteristic of IC's in the 201.12, 301.8, 301.10 and
301.12 packages. For flat pack IC's, the elastic properties of the contact
material are utilized to compensate for dimensional deviations and produce the
contact force.
Contact assemblies without a mechanical drive for integrated circuits in the
- 401.14, 301.12 and 201.14 packages are shown in Figure 9.17. Their operational
principle is clear.
To automate the measurements of the electrical parameters of integrated circuits,
it is expedient to place the latter in special satellite carriers, which make
it possible to use automated vibration load equipment and automatically sort
the measured IC's into groups. The use of satellite carriers with standard
dimensions has made it possible to standardize�the design of contact assemblies
for integrated circuits in various packages. A series of standardized contact
assemblies is being produced by domestic industry for integrated circuits with
J a low level of integration; the technical specifications for these assemblies
are given in Table 9.1, while the structural design of twa types is shown in
Figure 9.18. In the KW-1 and KW-2 contact assemblies (Figure 9.18a), the
satellite 4 with the integrated circuit in a circular package which is placed
in the satellite, where the IC rests on the support surface 5, causes the contacts
3 which are arranged in a circle to move towards the center of the circle. In
this case, the contacts go into the correspunding grooves in the satellite,
where the IC pins are positioned, connecting them to the electrical circuits of
the measurement instrument through contacts 1, and electrical connector 2, which
serves to disconnect the contact assembly during the replacement and repair
of the latter.
The KUU-6 contact assembly (Figure 9.18h) for integrated circuits in flat
packages also functions in a manner similar to that described above. The
difference consists in the fact that the contacts in this case are arranged in
- 215 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500090004-3
FOR OFFICIAL USE ONLY
Figure 9.17. Contact assemblies for integrated circuits.
four rows so that their ends are in one plane. The IC being tested is clamped
in the KW-6 by means of special pins, which are located in the assembly, and
holes in the satellite 4. In this case, the IC leads make contact with the
contacts 3 of the contacting assembly. With the further squeezing of the satel-
lite, by virtue of the elastic properties of the contact material a reliable
contact is made between the IC leads and the electrical circuits of the measure-
ment instrument through electrical connector 2. In all of the standardized
assemblies treated above, contact is made with the IC's being tested using a
two wire Kelvin system, something which makes it possible to segregate the vol-
tage and measurement circuits.
~
2
~
a)
z ,
(b) 6J-' i
Figtire 9.18. Mechanically driven contact assemblies for integrated
circuits.
a. In the 301.8 and 301.12 packages (KUU-1 and KUU-2);
b. In the 401.14 package (KW-6).
- 216 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500094444-3
N
1J
U
'b
~
~
u
~
H
O
w
~
d
~
~
~
Ol
U.
U.
Q
~
c
a
~
c
c
c
w
c
c
c
M
~
c
c
4
i
G
r
I
M
E
~I
rn
a
~
H
w
1.~
00
M'1
00
v1
3
b
a)
d N C!
Of $4 ' w
z m$4 W wu
d (D H ~C 0 C0 O Ln
U Q.' IJ CL W
0 o.~ew :3 az
P. a+ a.+ o U) cn
I 00
~
~
~ r
l �
O
O
z
I O~ 4-i U N
G ai O 0 N
�
�
u
N
tn
~
N O
H rl 3.D U U
v II
v ll
r1
I
C~l O ~
v1
ca
U:3 U aJ H'd
N
~
~ 0 0 0 k'
n II
n II
c~
-W w a
~
~
u c
o u o
G r.
co

�
aaa
�
o m o
0
0 0 Q
U -W U
I
u ai
~
O td q c0 G
U a~ 1-1 +J
~ w
o ~ r
Ln
w
~
Ln
U 1+
O �
r-I O
~
co
~ 0l W O cti
%O
-t
00
N N
Z .0 O U LJ
~-I
N
N
en w w
W G!
I
A ~ ~ ~
w
~
~ ~ ~
a
cO w a N
t-~a oH a
r-i
o
0
0
0 000
cr1
m
N
N rl r-I ~t
N
-~r
1l'1 ~O
a m
W N r-I
H o d .a
- 217 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFF[CIAL USE ONLY
CHAPTCR TCN TEST EQUIPMENT
5emiconductor devices which are used in different equipment function under complex
and diverse conditions of exposure to the environment and mechanical loads. The
structural design of semiconductor devices and the observance of the technological
processes for their manufacture should guarantee their normal operation under
the conditions stipulated in the technical specifications. However, in practice
because of the degradation of the quality of the raw materials, deviations from
the fabrication technology, worker errors, equipment failure and a number of
other reasons, not all of the manufactured semiconductor devices can maintain
their parameters under difficult conditions of exposure to the factors indicated
above. Because of this, all semiconductor devices or sample batches are subjected
to various kinds of tests during their manufacture so as to confirm the capability
of the semiconductor devices of functioning under the stipulated conditions
while retaining the electrical parameters within the stipulated range.
As a rule, devices are subjected to tests in the concluding stages of the pro-
duction process for their manufacture.
The major kinds of tests of semiconductor devices are:
--Mechanical;
--Climatic;
--For immunity to special effects;
--Aging;
--Reliability and service life.
The requirements placed on the immunity of semiconductor devices to various
effects, the immunity principle, the testing procedure and the circuit configura-
tion for the devices being tested are stipulated in the overall technical spe-
cifications (OTU) for the semiconductor device, by the special technical speci-
fications (ChTU) for each particular series or type of semiconductor device as
well as by the state (GOST) or sectoral (OST) standards.
10.1. Equipment for Mechanical Tests
The major kinds of inechanical tests .,f semiconductor devices are the following
[58]:
--Tests for the absence of freely moving particles inside a package which are
capable of disrupting device operation;
--Tests for the absence df short term short circuits and breaks in the circuits
of the semiconductor device leads;
--Tests for resistance to shock and vibration loads;
--Tests for resistance to exposure to linear acceleration.
- 218 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL
Figure 10.1. Kinematic schematic of the
SU-1 shock test stand.
Figure 10.2. Kinematic schematic of
the VS-68 vibration
test stand.
Metallic particles which are capable of causing short circuits of the leads of
semiconductor devices may remain inside a package in the fabrication of such
devices. One of the most widespread reasons for the appearance of inetal particles
is the sparging of the metal during the hermetic sealing of the packages. Poor
quality execution of the operation of lead bonding can be the cause of an unrelia-
ble contact between a chip and the external leads of the semiconductor device.
Permanent breaks and short circuits in the leads of semiconductor devices are
easily detected when measuring the electrical parameters. To ascertain short
term disruptions of this type, the devices being tested should be subjected to
shock and vibration loads, and the presence of short term short circuits and
breaks is registered by special equipment.
Equipment for mechanical tests should incorporate the following devices:
--For generating mechanical loads (vibration, shock or linear loads) with the
requisite parameters;
--For securing the devices being tested;
--For setting the electrical operating conditions (where necessary, stipulated in
the sectoral or special technical specificationa).
Moreover, equipment intended for testing for the absence of short term breaks
and short circuits in lead :ircuits should also contain devices capable of
registering these defects.
Primarily mechanical and electrodynamic shock test stands are used to produce
219 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
shock loads. Piechanical stands have found the greatest application because of
the comparative simplicity of their structural design as well as the absence of
electrical interference arising during their operation.
A kinematic schematic of the mechanical SU-1 shock test stand is shown in Figure
10.1, which provides for the free fall and sharp deceleration on its platform of
the products being tested.
The products being tested, which are placed in special cassette holders, which
provide for reliable fastening and supply the proper electrical conditions where
necessary, are secured to platform 1, which is coupled by means of cam 6 to the
drive mechanism, consisting of electric motor 4, V-belt drive 3, and a reducer
containing two pairs of cylindrical gears 2 and 5. With the rotation of cam 6,
platform 1 first lifts up, and then falls sharply on the supports 7, and a shock
load is thereby imparted to the products being tested. The number of shocks per
unit time is adjusted by varying the r.p.m. of direct current motor 4, while
the size of the shock load is adjusted by changing the adjusting washers on
supports 7. The number of platform impacts is counted by a counter.
The Main Technical Specifications of the SU-1 Test Stand
Number of impacts per minute 10 to 100
Acceleration, g 10 to 150
Maximum weight of testable products, kg 50
Platform dimensions, mm 285 x 452
Overall test stand dimensions, mm 620 x 540 x 750
Overall control console dimensions, mm 632 x 640 x 854
The vibration test stand, a kinematic schematic of which is shown in Figure 10.2,
imparts vibration loads to the products being tested. The tested products are
placed on table 4. A rotation is transmitted from electric motor 1 through V-belt
drive and gears 7 to shafts S. Gear sectors are fastened at the ends of these
shaf.ts: stationary 6 and moving 8 sectors. With the rotation of the shafts 5
at the same angular speed in opposite directions, the horizontal components of
the unbalanced forces mutually cancel out, while the vertical forces are slimmed
and cause the vertical motion of table 4, which is rigidly fastened to shaft 2.
The amplitude of the oscillations is ad,justed by moving the moving gear sectors
relative to the stationary ones, while to adjust the amplitude, there is adjusting
screw 3. The frequency of the vibrations is regulated by changing the rotational
speed of electric motor 1.
Electrodynamic vibration test stands are used to test semiconductor devices and
IC's for vibration resistance at frequencies above 500 to 1000 Hz; in these test
stands, the vibration of the table with the products being tested fastened to it,
is accomplished by means of the motion of a metal core in an alternating magnetic
field, produced by a sine wave or pulsed voltage. The WE-5/1000 electrodynamic
vibration test stand, consisting of the VE-5/1000 vibration test stand, the
- 220 -
FOR OFFICIAL i]SE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
WS-3 amplifier and the control console for the unit are pictured in FiRure
10.3.
Figure 10.3. The UVE-5/1000 electrodynamic vibration test stand.
a. Control console;
b. Amplifier;
c. Vibration test stand.
The Major Technical Specifications for the UVE-5/1000 Vibration
Test Stand .
Nominal load capacity of the vibration test stand, kg 5
Frequency range, Hz 5 to 10,000
Maximum acceleration for a load weigl:ng S kg, g 30
Maximum travel amplitude, mm 7�5
Magnetic field intensity at the level of the
vibration test stand table, no more than, A/m 400
When testing for the absence of short term breaks, shock loads with a specified
acceleration and frequency are imparted to the semiconductor devices. In this
case, the appropriate electrical conditions are created for the devices being
tested and meters which registe= the appearance of a pulse from a short term
break should be connected to each tested product during the entire testing time.
Tests for the absence of short circuits in the leads and for freely moving par-
ticles in the package of semiconductor devices are performed under the same
conditions, with the exceptfon of the fact that the tested products are placed
on the platform of the vibration test stand. Additionally, the electrical con-
ditions during the tests for the absence of breaks and short circuits are differ-
ent, about wh ich something will be said below.
- 221 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
Devices for which no disruptions of the contacts, short circuits and breaks in
the leads were detecteti are considered to have passed the tests. A short circuit
in the lead circuitry of a semiconductor device is characterized by a resistance
RB.C, which is inserted in parallel with this circuit, and by the short circuit
time. A break can correspondingly be characterized by a resistance Rbreakt inaer-
ted in series with the lead circuit, and by the duration of the break process.
Values of the resistances Rs.C, and Rbreak can vary in a range of from 0 to
depending on the factors which caused the short circuit or break. In line with
this, the signal which appears in the circuits of the product being tested and
which characterizes the occurrence of a short circuit or break can have different
values under identical test conditions.
NcmorrrR
A
'
NOo~i~rw0~
~ 1
Monp~Nr~vr,t
qempa0cm~s
9)
eter
NcmOrNr~
w
( 1
~ ~NOauaero0rw
cm
sdemh
~
~ano~w~~i
p
9
Figure 10.4. Block diagram showing
the testing of diodes for
the absence of breaks (a)
and short circuits (b).
Key: 1. Voltage source;
RH = load resistance.
(1)
McnrorNUK �
I
-'f~
~
R"
� Mcmorwu~r
Nan
~~rfNU~
~cnp~wcnu~
N
p
o
V o 1 t a ge
Nn.~u~omap-
source
"0f
ytmpodtmDa
Circuit configurations for diodes are
shown in Figure 10.4a and b, while circuits
for transistors when testing them for the
absence of short circuits and breaks are
shown in Figure 10.5.
Voltage sources are necessary to set the
electrical test conditions, while the
metering devices register the pulses
which appear across the load resistance
RH in the case of the appearance of short
term short circuits or breaks. In this
case, the polarity of the recorded pulses
differs when testing for a break and for
short circuits. The sensitivity of the
metering devices should be such as to
be able to determine the values of the
resistance RS.C, and Rbreak for a definite
value of the voltages at the leads of the
devices being tested.
e er Transistors which operate at low and
medium frequencies (up to 300 to 500 Mz)'
Figure 10.5. Block diagram showing the are tested for the absence of short cir-
testing of transistora cuits and breaks in the active mode, i.e.,
for the absence of short When an inverse voltage is applied to the
circuits and breaks. collector and a forward valtage is applied
to the emitter. The values of the voltage
Key: 1. Voltage source. at the collector UK [Vcc] and the emitter
current IE are stipulated in the special
teciiical specifications for the specific semiconductor devices. Testing high
frequency transistors in such a mode can lead to the appearance of self-excitation
phenomena in the devices being tested, which in turn leads to false actuation of
the metering units. To combat these phenomena, blocking chokes, capacitors, etc.
are usually placed close to the products being tested. In a number of cases,
it is expedient to test high frequency transistors with the junctions blocked,
i.e., it is necessary to apply inverse voltagea to the emitter-base and collector-
base junctions.
- 222 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
6
S
4
3
2
1
.
F-55?-3 >R0~
RT
o ~
A A
~
Besides the shock and vibration test
stands which test for the absence of
short circuits and breaks, which were
treated above, devices are needed to create
7 the electrical conditions as well as
fastening and contacting devices for the
products being tested and indicators. All
of this is combined in the RT-120 installa-
tion for registering short circuits and
breaks in transistors (Figure 10.6). It
consists of rack 7, in which all of the
electrical blocks are housed, and rotating
unit 8, which is mounted on stand 9 and
which makea the contacts with the tran-
sistors being tested and rotates them in
two mutually perpendicular planes. Some
/Z ~ -
�
. .~B

!0 ii
licip
'I
Figure 10.6. The RT-120 tester for registering short circuits and
breaks in transistors.
six indicator units 5 are housed in rack 7, in each of which there are 20 indica-
tor cells. Each cell is connected by its own input circuits to one transistor
under test, while the output circuits are connected to a small light mounted on
the front panel of the metering unit. The lights from the 20 meter indicator
cel.ts are combined together in signaling display 6. The power supply for the
transistors being tested 3, the power supply for the indicator cells 4, a block
of filters 2 and the control unit for the vibration and impact test stand 1 are
located in the lower portion of rack 7. The rotating devire 8, which is connected
to the rack by cable 10, is made in the form of a rigid platform, which rotates
through 90� on its own axis. The cassette holders with the transistors being
tested (20 transistors in each cassette) axe clamped in the rotating device by
means of plate 11 and flywheel 12. The traneistors are connected to the metering
- 223 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
- FOR OFFICIAL USE ONLY
and mode setting units through special contacts, which provide for reliable con-
tacts during the impact and vibration testing time.
The RD-120 unit for registering short circuits and breaks in diodes is structurally
similar to the RT-120 unit treated above.
The resistance of semiconductor devices to linear acceleration is tes[ed in a
centrifuge. The goal of these tests is to check the strength of the bonding of
the chip to the device package, the strength of the bonding of the leads to the
bonding pads of the chip, the hermetic seal quality and the quality of the metal
to glass seals at the sites of leads in the device and to generally check the
capability of the device of performing its functions during the process of
exposure to linear loads. The criterion for a device being good following the
tests is the maintenance of the electrical parameters within the specified range
or the satisfying of other requirements indicated in the special technical speci-
fications.
10.2. Equipment for Climatic Tests
The main function of climatic testing equipment is to check the operability of
semiconductor devices and IC's when exposed to various climatic factors [58].
The operability criterion for tested products is the preservation of their struc-
ture, external appearance and electrical parameters, which are checked either
during the tests or after exposure to the climatic factors.
The following categories of climatic tests have been established by existing
standard setting documents:
--For heat resistance;
--For cold resistance;
--For moisture resistance with both short term and long term exposures;
--For resistance to exposure to a cyclical change in temperature;
--For resistance to exposure to reduced and elevated pressures;
--For resistance to exposure to a sea fog;
I --For fungal resistance.
Semiconductor devices are tested for thermal and cold immunity in heat and cold
chambers respectively, where they are exposed for a definite time, most often
30 minutes, with the electrical conditions maintained. Upon the expiration of
this time, the electrical conditions are removed from the devices under test
and the parameters of the devices being tested are measured until they are re-
moved from the chamber. Cold immunity is tested at a temperature of -60� C,
while heat immunity is tested at a temperature of +70� C for germanium devices
and +125� C for silicon semiconductor devices. Other values of the test tempera-
tures are also possible, which are indicated in the special technical specifica-
tions for the device being tested.
- 224 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
For moisture resistance testing, the devices being teated are placed in a mois-
ture chamber without any electrical power being applied to them and kept there
for 4 to 30 days. In this case, a definite temperature is established in the
chamber in a range of +40 to +60� C at a relative humidity of 98%. At the end
of the tests, the electrical conditions are supplied for 5 minutes for the devices.
Testing for resistance to a cyclical temperature change, which has the purpose of
checking the quality of the seals of the device leads to either glass or ceramic
material as well as the quality of the connection of the chip to the mounting base
of the device, is carried out by means of placing the products being tested in
the heat and cold chambers by turns, the temperatures in which are previously
brought to the ultimate values indicated in the particular technical specifica-
tions. The exposure time for products in the chambers is limited by the time it
takes to reach temperature equilibrium or is established by some standard setting
documents. The devices should be moved from chamber to chamber within the time
specified in the particular technical specifications (no more than 1 to 5 minutes).
In this case, the temperature in the chamber after inserting the products in it
should change by no more than 10� C and be restored after a time of no more than
5 minutes, if other values are not indicated in the particular technical specifi-
cations. The time, which includes the exposure of the products in heat and cold
chambers taking into account the movement of the products from chamber to chamber,
is called the cycle time. The number of test cycles is specified by the particu-
lar technical specifications. The devices are as a rule tested without electrical
power applied.
'rhe immunity of semiconductor devices to reduced and elevated pressure is checked
in a pressure testing chamber. The level of the pressure, the testing time and
the necessity of applying electrical power and the electrical parameters are
stipulated by the relevant standard setting documents.
To test for sea fog exposure, which is carried out for the purpose of determining
the corrosion resistance of semiconductor devices in an atmosphere saturated with
viscous salt solutions, the devices are placed in a chamber in which a fog is
produced f rom sea water by means of an aerosol device, pulverizer or in some
other fashion, or from a salt solution obtained by dissolving sodium chloride in
distilled water. The fog should have a dispersion particle size of 1 to 10 um
(95% droplets), and au absolute moisture content of 2 to 3 g/m3. During the
testing, the devices being tested are placed in the chamber so that the solution
spray and drops from the ceiling, walls and system of supports do not fall on
the products.
When test ing for immunity to fungi, the semiconductor devices are placed in a
special fungi formation chamber, where they are sprayed with a water suspension
of a mixture of fungi spores, prepared in accordance with special instructions.
The testing is carried out at a temperature of about +30� C and a relative
humidity of 95% in the absence of air circulation.
There is a large products list of apecialized and general purpose test equipment
to test semiconductor devices for resistance to exposure to climatic factors.
- 225 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500090004-3
FOR OFF[CIAL USE ONLY
a
w o a~ v
1-4 H r-4 f-I
N
G
-r4
L
N
a~
U
~rl
~
tC
0'
~
U
w
0
N
L+
dl
~
~
~
cd
a
0
A
a~
H
r-1
O
~
W
~
H
.y a w
~
.~e
8
i
rz
a m
.14
�
m o
G
ro
m
m
0
0
m o
r.
N H
O
N
Ol
t!1
O
tA
N E-4
O
U
A
U
A
U
i- uco
oc
a
(1)
a
C)
R
d
a
oe
Y s 1+
c7
n

n
.G
n
4
2
e7
C
~
a
=
'J
H
~
d
L
H
U
OO
4-4
i-,
O
r
;
O
L
r

O
v

r
O
~
A
A
r

O
r
i'
O

a= 7
tA
F-
u
-
:E
~
f
v
:E
r
u
~
~
~
X
+j
~
Fo-
v
~
Fo-
:+C
l-o
!-o
w
s
d C
M
C=3q
~

S
o
0
0

0
r
a
c
I
I
I
I
I

~
~
~
~
~
I
I
v
Y~ M

o
O
O
7
n
G
q q
~ C
~
. = a
�
�
�
~a~ �

+
I
i
i
I
I
I
I
I
I
$
f y ~
~
O~
N
o~
ao
00
N
~
O
N
Ln
O
N
o ~ 3
O
O
~
O
O
O
d'
a'
O
V'
CO
a~ s'~
o
0
0
0
0
-
o
0
0
0
0
0
0
Y
~
o d
~
O
~A
~ ~
~r

ir+
~y
C�
~
+
+
+
o
~
+
~n
q!' CU
~o f-1
4
Od

IX
C
f
1^
M
v
~
C")
lA
~
h

`~x a
~
ctl
"
~

f

^

o
Go
0
0
C'i
i-:
C'
o
C
p
~
+N+
~
~
v
~
H

+
= a
^

f,

~
I
I
^i
N
H
o
%
6
F-
F-
F-
I-
F

:
F
%
E-
C~
C
CC
F
F-

:l
t
:L
:i
v
i

Y
!
X
C
a+ .C
u .G
r
�
4, rl
N tti
G) ~
~
~
Y
:7
r!
t7
G
G
m~
x
v
u
aai
e~i'
e+
ai
s
~
F
-
~
L
L
s
~
'
r
,
r

j
=
e0
W
t0
-
w~
.
A"
1
C
Y
O
s
6i
1-
u
h
eJ
r
ai
1^
a
+
r
cl
O
O.
O
G
O
G.
CO
tC
-
t
e7
e7
~
Gl
A
ra
n
A
u
ca
n
~o
N
~
d
u
U
b
0
cc
0
~a
~a
C
2
0
i
c
i
c
o
S
7
f
Ti
m
n
~
~
~

t
, .

1
~
L

LA l
=
F-
F
tG
~4
F
F
-

.a

cc
co
� w
fJ
U
~
to
C
w

ry4
M-1
4
a
+1
a)
LO)
u u
c~

U ctl (n F+ Cl 1+
�w �w
~ � ~ a W {4 f-I V1 ~ L~
~ aa.W 0) .~-H
WM'b cd N CL u13
0 ~ l r U" U U'b C'b
a) b c~ co
19aa~i~~`O ~
rl :3 i-+ .-I O N N N Cl
O.C c0 Gl U4+ 1+ aJ H
> p N 'G Cl ~ Gl ~
Gl
p r>i ~ 41cO
rl ai ~ Q tA Ol u N
.~c cc r+ ~41 U) a tn a
r-I iC 41 cb Gl p'H 0
3 a~~ x a H~ H
rl N M-It u1 %O 1- 00 ON
~i n o ~ ~ w o 227
. C!
x
- 226 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR (
Climatic testing equipment can be broken down into manually controlled equipment
and automated equipment with respect to the manner of loading the products being
tested into the medium and their removal from it, as well as with respect to
the manner of recording and processing the test data. Manually controlled equip-
ment, as a rule, is utilized in sample tests, where not all of the semiconductor
devices being manufactured are subjected to testing, but only a portion of them.
The productivity of this equipment is determined by the capabilities and skills
of the operator. The test results are recorded manually by the operator. The
major technical specifications of manually controlled climatic equipment are
given in Table 10.1.
In automated teat equipment which is used in the mass testing of semiconductor
devices, the loading of the products into the chambers, their removal following
the ;t:sts and sorting into groups are accomplished automatically.
Chambers with manual control of the heat, cold, moisture, sea fog, fungal growth
as well as pressure test chambers consist of the chamber itself where any of
the climatic conditions are created and the products being tested are placed,
the climatic support system, as well as the devices for regulating the climatic
parameters of the chamber within the range permissible for the testing and an
instrument for recording the values of the climatic parameters. Depending on
the external dimensions and useful volume, the equipment can either be of a
desk top or console type design. As a rule, such equipment is all-purpose and
can be used for testing various electronic hardware. In this case, the enterprises
performing the tests of, for example, semiconductor products, should themselves
design the devices for making the contacts with and switching the products under
test, which are placed in tht chambers, as well as the devices which assure the
proper electrical conditions for the tests.
An elevated temperature is produced in heat chambers by means of electrical heat-
ing, and either the vaporization of compressed gases or compressor cooling is
used to produce a below freezing temperature in cold chambers. We shall consider
the structural design of a console type heat and cold chamber (Figure 10.7). A
two stage compressor 1 is used to obtain a below freezing temperature in the
working volume of the chamber. A positive temperature is produced in the chamber
by means of electric heaters 5. The fans 4 are used to mix the air in the chamber
to abtain a uniform distribution of the temperature field over its entire working
volume. The setting and regulation of the temperature are accomplished by thermal
regulator 2; the current value of the temperature in the chamber is registered
by instrument 3, which can either be a meter, a digital meter or an autorecorder.
r[oreover, there should either be holes in the chamber for the electrical cables,
or special sealed entrances for supplying voltages from the external supplies to
the products being tested, as well as for the measurement of the parameters of
the tested products.
Moisture chambers differ from heat and cold chambers to a minor degree: special
devices are used to create environments with an elevated humidity instead of
devices for heating and cooling. One can use a fan with a vaporizer as such a
device. The vaporizer, which.takes the form of a reservoir with water and a water
- 227 -
FOR OFFICIAI. USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
FOR OFFIE'IAL USE ONLY
heater, humidifies the ambient air by
I
virtue of heating and evapo-Ating the
~
J
y~~~
TEI I '
water. The fan circulate,s the air in the
,
chamber. Semiconductor devices are usual-
,
ly tested for moisture resistance at a
~
`8
-
S
relative humidity of 85 to 98%. iests
s
for moisture resistance are usually per-
~
-
formed at an elevated temperature. Be-
cause of this, moisture chambers are
-
ith heat chambers
th
d
ll

.
er w
y ma
e toge
usua
Figure 10.7. A heat and cold chamber. pressure testing chambers, which are in-
tended for testing semiconductor devices
for exposure to elevated and reduced atmospheric pressure, likewise differ little
in structural terms from the heat and cold chamber treated here. A specific fea-
ture of them is the elevated requirements placed on the strength and hermetic
seal integrity of the walls, seals as well as the electrical leads. Sometimes,
the tests in a pressure testing chamber should be accompanied by heating or cooling
of the tested products. Because of this, there exist combined chambers which
combine tests at elevated and reduced temperatures as well as with elevated and
reduced atmospheric pressiires. A reduced atmospheric pressure is created by means
of vacuum pumps. An elevated atmospheric pressure is produced by means of a
compressor.
Figure 10.i. The set of equipment for testing the electrical parameters
of transistors at temperatures of -60 and +120� C.
In contrast to the all-purpose climatic testing equipment treated above, the
UKT-120 (position 1) and the UKT-60 (position 3) (Figure 10.8) installations for
testing the electrical parameters of transistors take the form of specialized
heat and cold chambers respectively and are designed for heat and cold resistance
testing, as well as for measurementa of the major parameters of transistors
(inverse currents and the gain) at temperatures above and below freezing. Each
of these installations is equipped with a switcher for the sequential connection
of the transistors to the tester to measure the parameters and a device for setting
- 228 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONL'1'
the electrical operating mode of the transistor being tested. The switcher and
the control point setter are housc:d inside the installations. The parameters
are measured by an IKT-2 digital~3ieter-classifier (position 2). There is a
display in it on which the ordinal number of the transistor, the parameter being
measured and its value are indicated. Moreover, all of these quantities are
automatically recorded on a numeri: printer.
Evaporating liquid nitrogen is used as the coolant ;n the UKT-60 installation.
A Dewar flask with liquid nitrogen is located iuside the installation. The
exposure time of the tested transistors in the heat chamber or in the cold
chamber is set by a timing relay located inside the unit.
s
s
~
1
4
,i
2
I
Figure 10.9. Cassette holder for test-
ing transistors with flex-
ible leads.
The heat chamber (in the UKT-120 unit),
in which the temperature is maintained
in a range of +100 to +130� C with a
precision of + 2� C, and the cold cham-
ber (in the UKT-60 unit) having a tem-
perature in the operating mode of -60
+ 2� C, are located in the upper left
corner of the installations. There are
two circular holes on the front side
in the heat and cold chambers, in which
the cassette holders with the transistors
being tested are inserted. Similar open-
ings, positioned symmetrically in the
lower portion of the installations,
serve for holding a spare set of cassette
holders. t4oreover, one can place in
them the cassette holders with the tran-
sistors just extracted from the chambers
to keep them at the ambient temperature.
The electrical connectors are fastened to the external surface of the cassette
holders, and it is through these electrical contacts that the emitter, base and
collector leads of the transistors being tested are connected to the switcher.
The holders are interchangeable for various types of transistors.
A cassette holder for low power transistors with flexible leads is depicted in
Figure 10.9 [591. It is distinguished from similar devices by its increased
contact reliability and simple structural design. This is achieved in that the
contact system of the holder is made in the form of moving flat insulating plater
with contact elements placed them, along one of the edges of which there are
grooves for the positioning of the teeth of the insulating comb. The insulating
plates 3 are put together in packets on cylinderical guides 4 with the flat
contacts 9 secured to the packets. The requisite gap between the plates is
achieved by springs 5 and adjusted by nuts 6. To prevent the transistors touch-
ing each other, they are separated by isolating frame 7, the ribs o� which move
freely in grooves 10 of plates 3. The contacting force is produced by the
rotation of eccentrics 1. There are plug connectors 8 to which the cassette
holder contacts are connected for the connection of the cassettes to the measure-
ment equipment and to the electrical power supplies and instrumentation.
- 229 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
The force produced by the rotation of eccentrics 1 is uniformly transmitted by
stiff plate 2 through the springs to plates 3, which in moving along guides 4,
produce the requisite contact force on each transistor lead.
The automation of the measurement of semiconductor device parameters in a range
of temperatures from below freezing to above freezing has a considerable economic
' impact, since the measurement of the parameters takes only fractions of a second
itself in the overall operational cycle, while the insertion, exposure to the
, medium at a specific temperature, as well as the removal and recording of the
measurement results is more or less easily automated. In this case, automated
testers are used as the measurement equipment, which was the topic in thP preced-
ing chapter, ahile run-through chambers are usually called climatic equipnent,
having in mind the fact that the products being tested in an automated cycle
sequentially pass through all stages from the loading to the sorting into the
appropriate containers following the completion of the tests [60].
Semiconductor devices are usually tested in run-through chambers while placed in
p,roup or individual carriers. The use of satellite carriers in automated equip-
ment is covered in more detail in Chapter 15.
S
S
e
9
IO
Figure 10.10. The PPS-130 semiautomated unit for diode instability
testing.
We� sh:,11 ronsider the structural design and operational principle of the PPS-130
ancl PPS-60 semiautomatic units, which are intended for checking the drift in the
pcir,imwtrrs of alloy diodes at temperatures of +130� C(PPS-130) and -60� C
(i'1'S-(i0). In the units, the diodes being tested are placed in a group carrier
rassette, which is loaded either manually or automatically on an independent
piece of equipment, while the diodes are removed from the cassette upon the com-
p]etion of the tests automatically, being allocated to individual hoppers in
accordance with the value of the parameters measured.
-230--
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
J 2 3 q

APPROVED FOR RELEASE: 2447102/09: CIA-RDP82-44850R444544494444-3
FOR OFFICIAL USE ONLY
The structural design of the units is quite similar, and their difference consists
in the system for producing the appropriate temperatures.
In the PPS-130, this is a coil heater, and in the PPS-60 unit, it is a valve for
delivering liquid nitrogen and a tubular coil.
The unit (Figure 10.10) consists of the following major assemblies: the frame 12,
in which the chamber 1 is mounted and all of the remaining assemblies and mechan-
isms. There are two rectangular openings in the front wall of the chamber for
the cassettQ holders to enter and leave it. Two-level guides 13 are installed
inside the chamber, along which the holders move, one pushing the other. The coil
heaters 2(or the tubular coil with the coolant in the PPS-60 unit) are secured
on special brackets to these same guides.
Four axial fans 4 are installed for the purpose of assuring uniform temperature
distribution in the upper portion of the chamber. The mechanism 8 which feeds
tlie cassette holders into the chamber 1 is mounted on the front of the chamber.
The mechanism 15 which removes the cassettes from the chamber 1 is mounted on the
back of the chamber opposite the lower level of guides. The cassette holders
are transferred from the upper level to the lower by means of inechanism 14.
1~)' 6
Figure 10.11. Kinematic schematic of the PK-5005 run-through
chamber.
The contacting device 6 is positioned in the immediate vicinity of the outpu,:
opening of the chamber, where this device is used to connect the diodes being
- 231 -
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICiAL USE OhLY
tested to the instrumentation. Mechanism 11 moves the contact sockets, and also
opens and closes door 7. Mechanism 9 unloads the cassette holder from the cham-
ber, while mechanism 10 sorts the tested diodes into groups. All of the mechan-
isms operate in a definite sequence, which is assured by the electrical circuitry.
The temperature is automatically maintained at the specified level by means of
an automatic control system, the sensor 3 of which is located in the center of
the chamber volume. There is a reference graphic control chart S on the front
panel for operational convenience, which depicts the operating sequence of the
mechanisms in the unit.
The PK-5005 run-through chamber is an automated unit for measuring the parameters
of transistors and integrated circuits in a temperature range of from -65 to
~ +150� C. In this case, the products being tested can be exposed in the chamber
automatically for from 6 to 30 seconds. The parameters are measured by an
external measuxement instrument [60]. The operating principle of the PI:-5005
run-through chamber is illustrated by the kinematic schematic shown in Figure
10.11.
The products being tested, which are placed in a special satellite (Figure 10.12),
are fed into the loading magazine 9(Figure 10.11) from which they are pushed out
one at a time by pneumatic cylinder 10 into the drum type transport magazine.
The slots in carousel 8 of this mechanism are thus gradually f illed with the
products being tested in the satellites. Carousel 8 has 31 slots, arranged
uniformly about the perimeter. Each of the slots holds 7 satellites, which feed
under their own weight from the loading magazine and are arranged one on top of
the other. Following the loading of seven satellites in one slot, the carousel
rotates clockwise through an angle of 11.6� and the next slot comes up to the
loading magazine, where this slot had been in the measurement and unloading pos-'
tion. Each subsequent rotation of the carousel occurs after feeding seven
satellites sequentially into the measurement position and correspondingly loadin"
seven satellites into the slots of the carousel in the loading position, which
is realized by means of transport mechanism 7. The loading of the products beitik,
tested into the chamber and the rotation of the carousel caupled to it take place
in a definite cycle, set by the control unit, which is nor indicated in the
schematic. The exposure time of the products in the chamber prior to the measurc-
ment of their electrical parameters depends on this cycle, as well as the capaci..y
of the carousel. � -
Figure 10.12. Satellite carriers: for u�nencapsulated transistors (a),
integrated circuits in the 101ST14-1 package (b) and
transistors with flexible leads (c).
232 -
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
The motion of the carousel 8 and the travel mechanism 7 is accomplished by means
of pneumatic drive 3, which is driven by compressed air at a pressure of 2 to 5
atm. The carousel is rotated through each seven operational cycles of the travel
mechanism 7 by means of the ratchet wheel 1 and lever 2, which are located on
the shaft of the travel mechanism 7.
After the travel mechanism has brought carousel 8 to the new position 30 times,
the products being tested which were first placed in the carousel slots come to
the measurement and unloading position. The transport mechanism 7 engages the
satellite with the product and brings it to contacting device 6, where all of
the leads of the tested product are connected to the circuits of the external
measurement instrument. In this case, the control unit feeds the "Start Measure-
ment" signal to the external meter for the parameters. Following the completion
of the measurements, a"End Measurement" signal is fed to the control unit from
~ the metering instrument. The control block generates the signal to turn on the
electropneumatic valve of drive 3. In this case, transport mechanism 7 transports
the satellite to the unloading position, where it falls into unloading magazine
4 by virtue of its own weight, and then into the corresponding hopper of the
unloading device S. Having stopped in the unloading position for about 200 msec,
the transport mechanism 7 returns to the contacting position, extracting behind
itself the next satellite with the product from the carousel slot, located in
the measurement and unloading position. The connecting device is equipped with
special catches, while the satellites have special holes for locking in the
contacting position.
The high temperature is maintained in the chamber by means of the coil heater,
while the environment with the below freezing temperature is produced by the
evaporation of liquid nitrogen.
The Major Technical Specifications for the PK-5005 Run-Through Chamber
Output for a product exposure time in
the chamber of, pieces per hour:
3 minutes
6 minutes
Exposure time of the products in the
chamber, minutes
Temperature range in the chamber, �C
Precision in setting the temperature, �C
Temperature fluctuations around the set
point, �C
Temperature distribution nonuniformity in
the product transport region, �C
4,200
2,100
6 to 30
(-65) to (+150)
+ 1.5
+ 0.5
4
Semiconductor devices can be tested for exposure to a cyclical temperature change
in tlie simplest case by means of conventional heat and cold chambers, which were
treated above. In this case, the transport of the products from one medium to
233 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFF7CIAL USE OnLY
the other, which is realized manually, should not take more than one minute in
accordance with the requirements of the existing standard setting documents. The
chamber temperature following the insertion of the products being tested into it
from another medium should not change by more thdn 10� C. It is difficult to
meet these and a number of other requirements in the case where heat and cold
chambers are used which are not coupled to one another. Several types of special-
ized units exist for testing for exposure to cyclical temperature change (thermal
cycling).
L_J
L_J
r-7 i
L_J
F--i
L_J
6)
9
S
Figure 10.13. The UTTs-60/160 semiautomated temperature cycling unit.
a. General view;
b. Schematic drawing of the climatic testing unit.
The UTTs-60/160 semiautomatic temperature cycling unit (Figure 10.13a) consists
of two parts: the control console (depicted on the left) and the climatic unit,
a schematic drawing of whicti is shown in Figure 10.13b. The heat chamber 1 and
cold chamber 2 are made in the form of hollow toroids, inside which there is an
intense air flow circulation. Cassette holders with the products being tested
are transported by means of carousel 7 and holder transport mechanism 2 clockwise
about the periphery, and simultaneously upward and downward, alternately falling
into each position in the heat and cold chambers. The carousel 7 and mechanisms
2 are driven by electric motor 4. All of the mechanisms and devices in the unit
are mounted in the assembly frame 5. Liquid nitrogen is used as the coolant; the
feed of the coolant is regulated through valve 6, which is controlled by a
thermal regulating system located on the control panel. For repair and preventive
maintenance of the unit, its upper section with the heat chamber can be lifted
by means of an electric hoist, located in the lower portion of the installation.
The UTTs-60/160 semiautomated unit makes it possible to perform thermal cycling
operations in three and five cycle multiples. Upon completing the last half-
cycle, the cassette holder with the products is pushed out into the unloading
position.
- 234 -
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42109: CIA-RDP82-00850R000500090004-3
An external view (Figure 10.14a) and an operational schematic (Figure 10.14b) of
the single chamber TO-5081 semiautomated thermal cycling unit are shown in Figure
10.14. A distinctive feature of the unit is the absence of a mechanism r_o trans-
fer the products being tested from one medium to another, since these producta
are fixed in a stationary position in the chamber, where an elevated and reduced
temperature environment is alternately produced in the chamber [60].
The TO-5081 semiautomated unit consists of the chamber 12, which is connected to
the branch pipes 2 and 15 and air ducts 5 and 11 by means of the channel switches
1 and 9, and along with these units, forms closed loops. There are guides for
the placement of the cassette holders with the tested products in the walls of
chamber 12. The holders for various types of semiconductor devices differ in
their structural design, but the external dimensions of all types of cassette
holders are the same. The products being tested can be placed in the cassette
holders both chaotically and with an ordered row layout. A block of heating
elements 4, consisting of five ESP-01 resistive elements and which serves as a
sensor in the temperature regulating system, is placed in the centex of chamber
12. The upper 1 and lower 9 switches for the channels serve to switch *.he air
flows through the chamber 12. The drive for the channel switches is pneumatic.
One sensor each 14 for the temperature regulation system, which takes the form
of a resistance thermometer, is installed in branch pipes 2 and 15. A thermal
relay 3 is additionally inserted in branch pipe 2 for the emergency disconnection
of the semiautomated unit in the case where the temperature norm is exceeded in
the "heat" channel.
Figure 10.14. The single chamber TO-5081 semisutomated heat cycling unit.
a. External view;
b. Operational schematic.
-235-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR OFFICIAL USE ONLY
A three-phase heater 6 is used to heat the air in the "heat" channel. The evapor-
ation tube of liquid nitrogen feed valve 10 serves to cool the air in the "cold"
channel. A heater 13 is inserted in the air duct 11 of the cold channel to dry
the air duct when repairing the semiautomated unit and after the completion of
operation.
Fans 7 and 8 serve to product an intense flow of heated and cooled air through
chamber 12.
All of the major controls, the digital display for the number of half-cycles and
meters which show the value of the temperature in the channels are placed on the
panel of the control console located in the top left corner of the unit. Chamber
12 is covered with a door with a special device which reliably hermetically seals
the joint between the door and the chamber.
In the case where an intense flow of hot air flows through the chamber with the
devices being tested, as shown in Figure 10.14b, the cold air loop is closed into
itself in the idle mode. The temperature in the working and cold loops is main-
tained in the specified range by a temperature regulation system. After the
specified time for a half-cycle has elapsed, i.e., the time for the exposure of
the products in one particular medium, the channel switches 1 and 9 automatically
connect the cold air channel to chamber 12 with the products under test, while
the hot air channel is switched to the idle mode. The number of cycles in the
TO-5081 semiautomatic unit is specified beforehand by a special unit on the con-
trol panel, and after the time for all cycles has elapsed, the semiautomatic unit
is cut off and signals the conclusion of the tests.
The Major Technical Specifications for the TO-5081 Semiautomatic
Thermal Cycling Uni_
Useful chamber volume, liters
80
Specific mass rate of flow of the liquid nitrogen
per kilogram of tested products, per hour, kg 4
Compressed air flow rate, mm3/hr 0.5
Maximum weight of the products which can be
tested, simultaneously loaded into the chamber, kg 43
Range of working tetaperatures, �C from -65 to +200
Temperature fluctuations at the operating point, �C 2
10.3. Equipment for Aging and Reliability Testing
The reliability of semiconductor devices is characterized by the probability of
tlieir failure free operation for a specified period of time. The failure rate
expressed as a function of their operating time is characterized by the greatest
fciilure rate during the period immediately following the start of device testing.
This is explained by the revealing of hidden manufacturing defects. Then the
failure rate falls off and is practically constant over time. This is the main
operating time of the devices.
- 236 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
Manufacturers of semiconductor devices, in striving to deliver products with a
high reliability to consumers, subject the devices to conditioning, the duration
of which corresponds to the burn-in time, i.e., the revealing of unreliable devices
ig accomplished in the manufactur3ng stage prior to the delivery of the producta
to consumers. The conditioning time falls in a range of from a few hours to
several hundreds of hours and is primarily determined by the level of fabrication
technology for the devices and the requirements placed on their reliability.
To determine a quantitative reliability indicator, the semiconductor devices are
subjected to special tests, which reduce to exposing the tested products to
definite electrical and temperature conditions, and monitoring the electrical
parameters and recording the devices which fail. The test conditions and criteria
for good products are defined by the special technical specifications.
In terms of its functional configuration, equipment for aging and reliability
testing is identical. But a number of specific requirements placed on each of
these types of tests leads to the necessity of designing specialized equipment
for both conditioning semiconductor devices and reliability testing.

6n A' 6K
BP BK
I
6r ~ s~v
BF BTR
I Kr ~
I
59 T
IT .
~ I
I 1
6P ~ ~ 63
BR B
Figure 10.15. Block diagram of the STT-
2000M test stand for the
electrical and thermal
conditioning of integrated
circuits.
Key: BP = Power supply;
BR =
Mode setting units;
K =
Cassette holders;
KT =
Heating chamber;
BK =
Pionitor unit;
BTR =
Thermal regulating
unit;
IT =
Measurement instrument;
BZ =
Protection unit.
- 237 -
The conditioning of discrete semicon-
ductor devices and IC's can be carried
out both at room temperature (elec-
trical aging) and at an elevated tem-
perature (electrical and thermal
aging), both in static and dynamic
modes.
A block diagram of the STT-2000M test
stand, which is designed for the
electrical and thermal conditioning
of integrated circuits, is shown in
Figure 10.15. The electrical con-
ditions are set by the power supplies
BP, while the dynamic mode for
switching the IC's under test is
provided by mode switching units BR.
The electrical circuits being sub-
jected to electrical and thermal
conditioning are placed in special
cassette holders K, made in the form
of printed circuit boards and placed
in the heat chamber KT, the tempera-
ture in which can be 'arought up to
+150� C. The monitor unit BK serves
to monitor the test conditions, while
the thermal control unit BTR sets
and maintains the temperature in tlle
heat chamber KT at the specified
level. The measurement instrument
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR OFFICIAL USE ONLY
Figure 10.16. The STT-2000M test stand
for the electrical and
thermal conditioning of
integrated circuits.
Figure 10.17. The UETT-T test stand
for electrical and
thermal conditioning
. of transistors.
IT registers the value of the temperature in the chamber, while the protective
unit BZ protects a number of assemblies of the test stand against overloads,
such as the fan motors and the heaters. The STT-2000M test stand is structurally
made of two sections: the rack with the power supplies and mode control unit!=
and the table with two chambers, mounted one on the other (Figure 10.16). A
spare set of cassette holders with contacting devices for the tested integrated
circuits is placed on the table.
The Major Technical Specifications for the STT-200011 Test Stand
Test stand capacity, pieces 2000
Triggering pulse repetition rate, Hz 50
Pulse amplitude, volts 5 to 6
Temperature in the heating chamber, �C 40 to 150
The UETT-T test stand for the electrical and thermal conditioning of transistors
(Figure 10.17) consists of a single rack, in the lower portion of which the
power supplies to produce the collector and emitter voltages are placed, while
the heat chamber is placed in the upper section, where five loading units with
contacting devices for the connection of the tranaistors under test are inserted
in this chamber. The test stand is intended for static electrical and thermal
aging of low and medium power transistors with flexible leads, connected in a
-238-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
common base configuration. The conditioning of the transistors is carried out
With the application of inverse bias to the collector and emitter junctions.
The Major Technical Specifications of the UETT-T Test Stand
Test stand capacity, pieces 5,400
Temperature range, �C 40 to 200
Power supply voltage for the transistors
being tested, volts 3 to 100
1
Figure 10.18.
The UNTM/T-2 unit for
reliability testing
of low and medium
power transistors.
In contrast to equipment for conditioning,
test stands for testing semiconductor devices
for service life and reliability should make
it possible to monitor the operating con-
ditions and measure the electrical parameters
for each product under test during the
teating time. Moreover, the products being
tested should not fail because of defects
in the test stand equipment. All of this
does not allow the design of eqiiiument for
reliability testing for a large number of
products to be tested simultaneously and
requires the incorporation of all possible
protective, warning and automatic recording
devices.
The UNTM/T-2 unit for the reliability test-
ing of low power transistors is shown in
Figure 10.18. The heat chamber is located
in the upper portion of the rack in this
installation. The checking of the operating
conditions and electrical parameters in
each product under test is accomplished
through special electrical connectors.
A temperature regulating unit is located below�the heat chamber, where this
unit has a device for signaling when the set temperature conditions are disrupted.
An electrical operating mode monitor unit and the power supplies which provide
for the maintenance of the specified electrical test conditions stipulated in
- the special technical specifications for the specific type of device are installed
in the lower portion of the rack. The products being tested are secured either
by means of special terminals, or by means of soldering to provide a reliable
contact during the tests.
The Major Technical Specifications for the UNTM/T-2 Unit.
The capacity for three sections having different
independent electrical conditions, units
~ -239-
,
FOR OFF'ICIAL USE ONLY
150
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00854R000500090004-3
FOR OFFICIAL USE ONLY
Range of temperatures, �C 40 - 155
Temperature maintenance precision, �C + 1
Setting range for the collector-base voltage, volts 1- 60
Setting range for the collector current, mA 3- 50
- 240 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
CHAPTER ELEVEN PRODUCTION PROCESS EQUIP'MENT FOR THE FINAL OPERATIONS
The final operations in the production of semiconductor devices and integrated
circuits include the application of protective coatings to the finished devices,
their marking and packaging.
For the purpose of protecting semiconductor device packages and IC's against
external effects during operation, thin films of varnish and paint materials or
metals are applied to their surface.
Degreasing and drying, priming and drying as well as painting and drying can be
numbered among the operations of the production process of painting semiconductor
device packages. The painted devices undergo a 100% examination for external
appearance and selective quality control.
The semiconductor devices are marked for the purpose of designating the type of
device, the trademark of the manufacturing plant, the date of manufacture and the
mark of the quality control department which confirms the good condition of the
device, and where necessary, polarity marks. The devices are marked with fast
drying marking paints or nitrocellulose enamels of various colors. The marking
label is applied to integrated circuit and semiconductor device packages primarily
using the so-called offset method.
Packaging is the final operation in the process of manufacturing semiconductor
devices and protects the devices against mechanical damage and other effects
during transportation and storage. There are several methods of packaging:
--In cardboard or plastic boxes, where each device is placed in a separate nest
to prevent its moving;
--In polyethylene packets;
--In polyethylene material in which cells are produced underneath the devices by
means of vacuum forming [61]. 11.1. Equipment for the Protective Coating of Finished Devices
Depending on the type of semiconductor device, various techniques are Lsed to
apply protective coatings: painting, nickel plating and tinning. The most
widespread methods of painting are dipping, flushing with a continuous stream and
spray painting.
The structural design of an automatic painting unit for semiconductor devices in
metal-glass packages, using preliminary straightening of the leads, drying and
feeding of the finished devices to the next operation, is shown in Figure 11.1.
The automatic unit is attended by a single operator and paints from 8,000 to
10,000 devices per hour.
The production process is realized in the following sequence. The devices which
are degreased beforehand are loaded into vibrating hopper 7, from which they
- 241 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
Ftgure 11.1. Automatic unit for painting and drying [semiconductoi]
devices.
are fed piece by piece to the lead straightener 6. The uniform feed is accom-
plished by means of setting the requisite gap in an electromagnet. Having fallen
into the electromagnetic scoop of the lead straightener, the devices are oriented
by an el.ectromagnetic field and are transported to a drum, which in rotating
continuously, catches and feeds them into a slot between upper and lower cams,
where the leads are straightened. (The leads of the devices may not be straight-
ened. This occurs because the devices are not rolled out and untwisted. To
climinate this defect, it is necessary to adjust the pressure of the rubber sur-
Eace of the cams against the device. In that case where the rubber is worn our,
it must be replaced.) The straightened leads are rolled down into the magnetic
holding tray 5, where they are uniformly distributed over the entire length and
are fed into the loading drum by means of a magnetic field, where this drum in
rotating continously catches and feeds them to transfer loading drum 2, between
the disks of which the tube of the painting assembly 3 is located. In passing
through under the tube, the devices are painted and fall onto a comb where the
excess paint is removed, which drains off into a funnel and goes through an
opening into the pump tank. The width of the jet is regulated by a lever located
on the upper part of the painting assembly. The painted devices fall from tray
11 into the radiative heat chamber for drying them, where they are dried in an
ultraviolet spectrum for 7 to 9 minutes at a temperature of 130 + 10� C. The
temperature in the chamber is regulated by a slide valve which is located in the
air duct. When leaving the drying chamber, the devices fall into the unloading
transporter 1, by means of which the devices are loaded into the corresponding
piicking case and forwarded to the next operation.
The major assemblies and mechanisms of tre automated unit are: the vibrating
hopper, the lead straightener, the magnetic holding tray, pump 4, the painting
assembly, drying chamber 13, the unloading transporter, control panel 9 and drive
12. Small table 8, the vibration hopper, lead straightener, magnetic holding tray
with the delay unit for devices with unstraightened leads and the device for
- 242 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL i
N
stopping the feed of the devices to the loading drum as well as the panel with
the transfer drums and painting assembly are mounted in frame 10. A pump is
mounted underneath the small table which feeds the paint into the painting assem-
bly. It is equipped with a coarse cleaning filter, a drain valve and a housing
with an exhaust unit. The vibrating hopper orients the device and feeds them
into the lead straightener. The drying chamber which is shown in Figure 11.2 is
located in the left side of the frame and takes the form of a thermally insulated
enclosure 2 with double doors. The inside door is made in the f orm of panel 3
on which the PRK-2M lamps are mounted with reflectors. Inside the chamber, the
coil heater 1 is secured to a textolite plate. It consists of a stationary brass
coil and a brass rotating disk with cutouts in which the devices are loaded. The
internal elements of the chamber have a light reflecting surface. The panel of
magnets 4, by means of which the devices are held and moved along the groove of
the coil from the periphery to the center is located behind the coil heater.
z
Figure 11.2. Thermal radiation drying
chamber (section through
A-A).
The rotation of the transfer drums and
the coil heater is accomplished by drive
12 (see Figure 11.1), which consists of
an AOL-12-4 electric motor, a worin gear
pair and a number of intermediate gears.
A three step pulley is mounted on the
electric motor shaft to change the r.p.m.
of the disks, the transfer and loading
disks, and correspondingly, the drying
time. The electric motor is secured in
one of three positions. The drum 2 is
turned by tcao RD-09 motors.
11.2. Labeling Equipment
The offset method of applying the markir.g
label has become the most widespread
technique in semiconductor device and IC
production. Additionally, coded markings
are used in the fabrication of micro-
miniature devices. There are more than
20 ways of applying a marking brand, in-
cluding direct, flat application, stencil-
ing, etc.
1)epending o1 the structure of the device package, the marking is applied either
on rlie end face of the package, its side surface, and so on (Figure 11.3) [4].
An :iutomated unit for marking and drying devices is shown in Figure 11.4. It
t�ontiists oE the following major assemblies and units: the marking unit I, the
infrared drying conveyor furnace II and the combination unit III. When attended
by a single operator, the automated unit provides for a kinematic productivity
of 5,600 devices per hour.
- 243 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500490004-3
FOR OFFICIAL USE ONLY
ai
, U
M
7
N
b
co
w
0
v
� co ~
' w 44 :J
~ i-i U
! ^ f~l1 M
N ^
0) U
U ~ ,d
v 41 w
o ~
O Rf
-W
b0
-1 GJ
~ e o m A ~
ctl H
'-1
bQ U
0
44
Sx-i N
m 'd
~ ~ .
p
W
O
. .w
0 0
yH ~
u ~
R1 �rl
U tA
�r1
rl [Q
a $4
ca H
Q) cb
YP091'I. ~
O E-~
cy1
~
Eiq
~
G1
$4
:3
00
-H
w
n
h
b
4.
n
h
-1
0%
r%
1
- 244 -
FOR OFFIC[AL USE ONLY
~
N
Cl
u

>
v
b
~
0
~
u
~
b
C".
O
U
rl
d
~
~
b0
A

>1
H
'b
'd
~
cd
00
0
.rj
rz
1-
O
W
~
.4
G
~
C.1
.r{
~
~
Ei
O
u
Qi
~
~
~
~

:3
00
-H
w
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
The operational principle of the automated unit consists in the following: the
devices which are degreased beforehand and have straightened leads are loaded into
vibration hopper 1 in batches of no more than 700 pieces each. The devices are
fed from the vibration hopper by means of a directed magneric field into each
of the forks of carousel 4 by means of the transfer drum, which takes them from
the loading drum 2, rotating synchronously in this case with both them and the
carousel. The polarity of the device is determined on the forks of the carousel.
In the case of improper orientation of the device relative to the marker, the
device is rotated through 180�. This is accomplished through a belt drive, which
is engaged by means of an electromagnetic assembly. A pair of blocking contacts
which gives the instruction to shut the automated unit down is provided to prevent
improperly oriented devices falling onto the marking drum. 'fhe correctly
oriented device goes onto the marking drum 5, where the marking sign is applied
to the cylindrical surface of the device package by means of rolling using a
rubber roller, on which there is the corresponding protruding marker. The rota-
tion of the device about its axis is accomplished by virtue of friction between
the device package and the rubber insert of the printer mechanism.
The marked devices are transloaded into the carriage of the distribution mechan-
ism 6 and layed out on the chain conveyor 10, on which they are fed into the
furnace chamber for preliminary drying. The final drying of the marking label
is accomplished on the grid of a strip conveyor 11, located in the lower portion
of the chamber, to which the devices are transferred by the transloading drum
from the chain conveyor. The dried devices are unloaded from the strip conveyor
to the receiving hopper. In step with the accumulation of finished devices,
they are periodically unloaded and forwarded to the subsequent operations.
The marking unit I consists of the following major assemblies and components:
the vibration hopper, the marking mechanism, the drive sprocket wheels 7, the
rubbing rollers 16 which-uniformly apply the paint to the printing plate of
roller 3, drive 14 and control panel 17. The unit is driven by an electric motor
through a V-belt drive, a conical worm gear reducer, having two output shafts
for driving the marking mechanism and the master block of sprocket wheels for
the chain conveyor drive. Manual drive is used when aligning the unit for the
drives of the chain conveyor and marking mechanism assemblies.
The marking mechanism is structured from the following major assemblies and
components: the loading drum, the collecting drum, two bushings with built-in
permanent magnets, a carousel, consisting of a chassis on which 10 toothed small
shafts are uniformly mounted in a circle where the shafts have fork brackets
fastened to the square tail stems. The forks are clamped by means of flat sprir-s
in a definite position, because of which they are always set in a plane paralle1
to the axis of rotation of the carousel. The permanent magnets built into the
forks hold the devices and keep them from falling out during the rotation of
the carousel. The marking drum is made in the form of a disk with two rim
flanges, about the outer diameter of which the grooves for the placement of the
devices are uniformly arranged. The printer mechanism consists of the frame,
cover., nut and holder with the rubber insert. The marking paint for the outer
surface of the cylindrical rubber insert is transferred from the marking pattern
to the package of the device, making the necessary marking in this case. Then
-245-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONI.Y
tlie devices are fed by means of a similar drum into slots in the carriage, which
executes a reciprocating and cyclical motion along the axis of the shaft, placing
the devices in two rows in the indentations in the combs of the chain conveyor.
The reciprocating motion of the carriage is realized by means of a cam working
through levers and stems with a stationary fastened fork and tray, which hold the
devices in the slots of the carriage during its motions.
The stereotype block mechanism consists of a disk with two carriages which are
fastened in a stationary manner to a shaft; the templates are fastened to these
carriages; a bracket with rollers and slide blocks, by means of which the period-
ic reciprocating motion of the carriages along the slotted shaft of the stereotype
block is realized; an electromagnetic assembly, which engages the belt drive for
orienting the devices by means of an electromagnet; an armature and brackets;
a Geneva mechanism, by means of which ttie carousel and the printing drum are
rotated cyclically; as well as a brake which suppresses the inertia of the rotating
narts of the marking mechanism.
The infrared conveyor drying furnace consists of the chain conveyor 10, the strip
conveyor 11, the aprocket drive wheels 7, the upper and lower heate:rs 8(the
KI-220-1000 infrared lamps), thermocouple 9, strip conveyor drive 12, air duct 13
and frame 15.
The control panel 17 consists of the panel on which the toggle switches for con-
trolling the following are mounted: the marking mechanism, the plate and strip
conveyors and the vibration hopper. The corresponding signaling lights are also
placed here.
Thc setting and monitoring of the thermal modes oE the drying chamber, as wcll
the adjustment and automatic mainten.ance of the specified temperature 15� C
in the drying region) are accomplisYed by means of the regulating devices and
instruments.
11.3. Packing Equipment
Figure 11.5. The mountirig of devices on a cardboard card.
tJith mass production, the packaging of semiconductor devices, especially miniattire
ones, is extremely labor intensive. The presence of external leads creates
certain difficulties when packii.tg the devices. Finished devices are frequently
packed in cardboard boxes, shaped sheets of polyethylene or polyethylene packets.
A general view of devices mount.ed on a cardboard card is shown in Figure 11.5a.
-246-
:wOR OFFICIAL USE ONC,Y
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42109: CIA-RDP82-00850R000500090004-3
:1AL USE ONLY
Holes are punched in the cardboard card prior to mounting for the installation of
the devices. The cardboard cards and the finished�devices (classified into groups)
are delivered to the work positions, where workers iastall a device in a'hole and
fasten it with a nut, placed an the heat sink (Figure 11.5b). After this, the
devices, along with the accompanying sheets, are placed'in cardboard boxes and
dispatched to the finished product warehouse.
Yet another method is that of packing finished devices in shaped cells. Polyethy-
lene with a thickness of 2 to 2.5 mm is used in this case. The cells are produced
at a temperature of 160� to 250� C corresponding to the configuration and overall
dimensions of the devices. A formed polyethylene sheet for packing integrated
circuits is shown in Figure 11.6. Such sheets are fed to the work positions,
where the devices are loaded into the cells. After this a second sheet is placed
on top and they are sealed together.* Devices packaged in this way are then for-
warded for delivery to consumers.
r
Fig
s in
In the structural design of the instal-
lation described below (Figure 11.7),
the semiconductor devices are packaged
in polyettiylene and cellophane packets. Figure 11.7. A unit for packing semi-
The unit is attended by a single operator conductor devices.
and its output is 2,000 devices per hour.
100 devices are packaged in a single
packet. The automatic fabrication of packets with dimensions of 84 x 60 mm is
carried out in the unit as well as the counting of the devices loaded into the
packages. The batch loading of devices into ready packets is done manually by
the operator. A packet is fabricated from a doubled polymer polyethylene-cello-
phane film.
The following operations are performed in the unit:
--The fabrication of the packet;
--The automatic counting out of the requisite number of devices (50 pieces);
--The loading of the devices into the packet;
--The sealing of the loaded packet.
- 247 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
polyethylene cells.

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The operational principle of the packaging unit consists in the following. The
packet is fabricated from polyethylene-cellophane film in the sealing unit 6,
where this film is unreeled from two bobbins and fed through an upper pair of
guide rollers to the sealing rollers, which are driven. The bobbins with the
stock of tape are mounted on brackets, which are fastened to the chassis of the
sleeve sealing assembly. There are flat springs for clamping the end of a strip
prior to winding a bobbin. The rollers are heated up to a temperature in a range
of 160 to 250� C, necessary to sealing the film. When the strips move between
the rollers, they are sealed along the edges. Here, the cross welding is accom-
plished at definite spacings. In continuing their motion, the sealed strips
(sleeves) are fed into the lower pair of rollers, which move it into the knife
assembly where the sleeve is cut into packets.
The rotational speed of the sealing rollers is regulated by an r.p.m. controller
8, w}iich is located in the table pedestal. The devices are loaded into the
packets from guide 4, which is a unique kind of holder to which the devices are
fed in an oriented position from vibrating hopper 3. With tlie accumulation in
the guide of a definite number of devices (units of 50 pieces each), the ready
zind open packet is brought up to the lower section. After this, by pressing on
a lever, the upper support is opened and the lower one is closed. In this case,
the upper support cuts out the 51st device from the series of devices appearing
one af[er ttte other. The lower 50 devices fall from the guide into the packet.
The iever returns the supports to the initial position (the top one is open and
the lower one is closed), while the guide is filled with a new batch of devices
from the vibrating hopper. The devices are loaded twice in groups of SO pieces
eacli time into a packet. Upon completing the loading into the packet, au accom-
panying sheet is inserted and the loaded packet is sealed shut. This is accom-
plished in sealing unit 5. For this, the edges of the packet which are to be
sealed are placed in the lower heater of the assembly and the actuating foot
pedal 9 is pressed. After the upper heater is lowered until it makes contact
witli the bottom one, the foot is taken off of the pedal. After a few seconds
have elapsed (the duration of the pulse is adjusted in a range of from 1 to 6
seconds), the upper heater automatically returns to the initial position. The
devices which are packaged in the packets are fed to the finished product ware-
house.
The unit has a table 1, on the plate of which all of the ma3or assemblies of the
unit are secured, including the unit control panel 2. The table has a pedestal,
in which the power panel is located as well as the electrical equipment. Located
on the front side of the pedestal are the handle of the RNO-250-0.5 regulator,
by means of which the r.p.m. of the sealing rollers of the sleeve sealing assem-
bly is clianged, as well as the handle for adjusting the voltage of the pulsed
he.lter oC the sleeve sealing assembly and the handle of regulator 7(a RNO-250-
0.5), by means of which the voltage fed to the electromagnets of the vibrating
luoppcr is adjusted. Handle 11 of the automatic unit AST-2 is placed in the left
purtion of the table, where this unit is intended for turning the mains voltage
(in and oEf. There is a recess in the rear part of the table in which the pneuma-
tic control panel 10 is located. The recess is covered with a small removable
door.
- 248 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
PART III LINES AND SYSTEMS FOR THE MASS PRODUCTION OF SE*4ICONDUCTOR DEVICES AND
INTEGRATED CIRCUITS
CHAPTER TWELVE THE THEORETICAL PRINCIPLES OF THE COMPREHENSIVE MECHANIZATION AND
AUTO;IATION OF SEMICONDUCTOR PRODUCTION
12.1. Problems of Comprehensive Automation and Specific Features of Semiconductor
Production
Semiconductor production differs substantially from machine construction in a
number of specific features. The major feature consists in the fact that the
process of producing the major element of the device, which determines all of
its functional capabilities - the chip with the p-n structure - is the result of
a series of successive operations inside the volume of the chip. In contrast to
machine structures which can be taken apart, a chip is practically a single com-
ponent, and a re3ection in only one operation leads to the rejection of the
device as a whole. A second feature is the microscopically small dimensions of
a device and the exceptionally high requirements placed on the overall technical
l.evel for the production, the purity of the materials used and the conditions
under which the process is performed. The third feature is the large number of
different products in the products list with a relatively unstable market situation
as well as the presence of such production process operations as classification
or sorting according to the types of devices with different parameters following
their final fabrication (Figure 12.1). The specific features enumerated here
bring about a high level of production process losses. The level of production
process losses depends in the final analysis on the technical level of the
technology and production which is achieved in the industry. It is rather high,
something which leads to the necessity of planning it for each product, in con-
trast to machine building production, where production process losses are the
result of rejects in the fabrication of individual parts, are insignificant and
are not planned for, with the exception of certain products in special instrument
making, the production technology of which has a number of general features in
common with electronic instrument making.
A'7J72~
KT~/7 t 311 2B KTJ1~6
A'r: ccuq~u,r ~ a KT1/19
Classication KT312V
Figure 12.1. Schematic showing the
separation of finished
KT312 devices into
groups in the classi-
fication operation.
Because of this, all of the design calcu-
lations for lines to be developed, includ-
ing the determination of the quantity of
necessary production process equipment,
the arrangement of the monitor equipment,
etc., must be carried out taking into
account the average static plan norms for
production process losses (the percentage
yield). Correspondingly, the problem of
reducing losses through reducing the level
of production process losses takes up
first place in semiconductor production.
The major ways of solving it are improving
the quality of the materials used,
249 ~
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500090004-3
~OR OFFICIAL USE ONI.Y
creating speciat product ~ conditions, stabilizing the parameters of the pro-
duction process, etc. Wtei creating comprehensively mechanized production, decis-
ive factors are also the ctioice of the production process, which assures minimal
losses, as well as the development and introduction of equipment which assures
stability and reproducibility of the parameters of the production process.
The most objective criterion for evaluating the work of any sector of the national
economy is the growth in the productivity of the labor force. This indicator can
be utilized to analyze and substantiate objective laws which govern technical
progress, quantitatively estimate them and forecast the developmental paths of
new hardware, since in the final analysis, this indicator is related to the pro-
duction cost and the qualitative level of production. The theory of machine and
labor productivity developed by Professor G.A. Shaumyan, in figuring the final
parameters of new hardware and determining its efficiency, works from the condi-
tion of attaining a maximum growth in labor productivity [62, 63]. The starting
postulate of this theory is the concept of an ideal continuous service machine
with an infinite service life"and absolute reliability, the productivity of which
is governed only by the production process (the technological productivity):
eideal - k
Of course, the general trend in automation is towards increasing the production
process productivity, i.e., developing progressive technological processes and
methods, and creating highly productive tools for production based on them. How-
ever, this approach is a far from adequate tool for the creation of high efficiency
production, since the degree of utilization of the capabilities of a production
process in an actual machine or line can differ substantially, but is always less
than the ideal. From the viewpoint of machine and labor productivity theory, any
time during which the production process is not under way is lost. For this
reason, both cyclical losses (idling time for machines or an automated line) aiid
non-cyclical losses, despite their different nature, are treated as losses. The
actual productivity is:
6 = knntechnorg
where n is the productivity coefficient which takes into account the cyclical
losses; ntech is a coeff icient which takes into account the losses due to technical
factors; rlor is a coeff icient which takes into account the losses due to organi-
zational fac~ors.
Losses are broken down into the following six kinds, each of which determines the
corresponding problem of the comprehensive automation of production processes.
Losses of kind I are cyclical, and define the problem of automating the working
cycle, and creating continuous service machines and lines; problems of kind II
are related to the tool (changing, adjusting, truing, etc.), and define the pro-
blem of automating the changing and adjusting of a tool; losses of the III kind
are the adjusting and repair of machine mechanisms and define the problem of
service life and reliability of automated systems; losses of the IV kind pertain
to organizational factors (receiving the material, turnover of the finished parts
- 250 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-40850R000500090004-3
FOR OFFICIAL USE ONLY
and the collection of waste, the absence of semi-finished products, etc.), and
define the problem of automating production control; problems of the V kind are
those of product rejection and def ine the problem of product quality; losses of
the VI kind relate to setting up equipment again because of the transition to the
fabrication of another product and def ine the problem of automated production
flexibility.
Formulated in general form, these problems are fully applicable to semiconductor
production, however, because of its specif ic features, the signif icance of the
individual problems is substantially and qualitatively redistributed as compared
to those of machine building production, and the tasks in the field of comprehen-
sive mechanization are made more precise and specific [64].
The Problem of Automating a Working Cycle and Designing Continuous Operation
Machines and Lines
The following approaches to the solution of this problem as applied to semiconduc-
tor production can be noted:
1. The wt.de scale application of the batch processing technique is here one of
major ways of curtailing losses of the first kind. The features of semiconductor
production noted above provide for a considerably higher efficiency in the applica-
tion of the group method than in the case of other types of production. The
group coefficient, i.e., the number of elements subjected to simultaneous process-
ing, amounts to more than 1,000 in semiconductor production in the operations of
producing junctions on a wafer, and with an increase in wafer diameter up to 100
to 150 mm, the number of elements of a device which can be processed simultaneously
on a single wafer reaches 4,000 and more. If one considers that in a number of
chemical treatment operations, a special batch package is employed which makes it
possible to load up to 100 or more wafers, then it can be asserted that the
difference in the batch methods in semiconductor production is of a qualitative
nature and has a substantial influence on the organization of this production and
the techniques for automating it.
2. The use of through-going production process satellites and group interoperation
containers. In the overwhelming majority of fabrication stages in semiconductor
production, as a rule, the product is not put in hoppers or mechanically trans-
loaded, and its transportation at times results in additional rejects. For
this reason, a universal technical solution in the creation of a continuous auto-
mated flow is the use of new principles for transfer between operations based on
the utilization of through-going production process satellites to assure the
maintenance of product reliability, where such satellites are frequently insepara-
bly coupled to the structure of the device, as well as principles based on the
use of strip carriers and special cassette holders.
3. The design of specialized transloading and collecting holders. Those specific
features of the production process such as the necessity of performing operations
in a controlled gas environment, in a vacuum, or in a dust free volume, as well
as the limitation on the storage time of the process stock require the design of
special loading and unloading as Fe11 as collecting devices to meet these
-251-
FOR OFFICIAG USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/42109: CIA-RDP82-00850R000500090004-3
FUR OFMICIAI. USE ONLY
conditions. Specialized transloading devices are also being designed for brittle
or easily deformed elements of a product. In the init ial production stages, where
wafers are manipulated, the automation of their transloading is most efficiently
realized by means of using individual production process transport cassette
collecting holders.
One of the ways of automating the auxiliary transport and load ing-unload ing
operations is the utilization of industrial robots and automated manipulators,
which are computer controlled [65]. The introduction of such robots is of
exceptionally great importance in the stage of the complete automation of semi-
conductor production for a large products list.
Losses of the second and third kinds and the problem of automating the replacement
and adjustment of a tool which are related to them, as well as the problem of
the service life and reliability of automated machines and lines in semiconductor
production are also of no less importance than in other sectors of precision
instrument making. There are no substantial specif ic differences between the
problems in precision instrument making and those in semiconductor production,
and for this reason, they are not treated here.
The Problem of Control Automation
The problem of automating production control with an elevated level of automation
for the major production processes, and correspondingly, with an increase in
the specific share of organizat ional losses is taking on ever increasing import-
ance. The introduction of elements for the scientif ic organization of labor
based on the simplest organizat ional equipment, standard collecting holders and
transloading devices frequently has a great impact. Of course, the task in the
creation of comprehensively mechanized production is the introduction of an
automated production control system (ASUP) based on computers.
The Problem of Product Quality
Reducing production process losses leads to a direct decrease in the labor
expenditures in the same production where the losses were reduced, something
which can be seen from Table 12.1, where two production variants are cited with
an arbitrary output of 1,000 pieces of f inished devices. The overall percentage
TABLE 12.1. The Influence of an Increase in the Percentage Yield in Individual
Production Stages on the Overall Reduction in the Labor Input
Re uirements for a Device
Production Indicators
for the Various Stages
Percentage yield, %
Production volume,
pieces
[xihor input require-
ment
Stage I
Variant Variant
1 2
50 75
5,000 1,670
450 150
Sta
ge II
Total
Variant
Variant
Variant
Variant
1
2
1
2
40
80
20
60
2,500
1,250
1,000
1,000
250 125 700 275
- 252 -
FOR OFFtCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500490004-3
FOR OFF'if
yield in the second case is three times greater than in the first, the labor
intensity in the first step is three times lower, and two times lower in the
second. The overall labor intensity is reduced by a factor of more L'nan 2.5 times.
For this reason, ttte design of a unit which produces a drop in the production
process losses for semiconductor devices can frequently lead to a more substantial
result as regards the reduction of labor intensity than the mechanization of any
manual operation, without increasing the yield percentage. And even a significant
increase in labor productivity in an individual mechanized operation leads to a
reduction in the overall labor productivity in the production complex given the
condition that this mechanization is accompanied by even a slight reduction in
the percentage yield. For this reason, any engineering decision concerning the
introduction of new processes and the creation of tools for comprehensive mechan-
ization should be made as a result of a technical and economic analysis, based
primaril.y on an evaluation of the change in the level of production process losses.
Recause of the special importance of the problem of quality in semiconductor pro-
duction, one of the most important tasks is the design of automated quality con-
trol and measurement equipment for the monitoring and classification of both
the finished products with respect to their parameters, as well as the technology
and components of a device during the fabrication process, the design of automated
production process equipment, equipped with sensors, monitor instruments, and
built-in microprocessors for monitoring, controlling and optimizing the production
process modes. The final task is the creation of automated control systems for
the production process (ATU TP) based on computers. The production process losses
should be reduced to a minimum, and the possibility of producing the requisite,
previously specified group of devices during the production process should be
realized by means of these systems.
The Problem of Yroduction Flexibility
T}ie problem of flexibility of highly mechanized production is also particularly
acute for semiconductor production. The group method yields almost unlimited
possibilities for increasing productivity in the first production stages. The
same possibilities have been obtained recently in the stage of quality control
and classification operations because of the use of universal testers as part of
a computerized complex. The high productivity and universality of these instal-
lations comes into contradiction with the relatively poor productivity of assembly
lines and the sometimes limited demand for specific types and groups of devices.
This problem can be resolved through the creation of high productivity production
systems for products lists with many products, including high productivity all-
purpose production of chips and specialized assembly operations.
12.2. The Systems Approach to the Planning of Automated Production
The considerations set forth above make it possible to def ine the production of
Semiconductor devices as a complex probabilistic system, in which the production
processes are structured based on the conditions of the combined work of man
and machine. This means, first of all, that in all stages of the design and
development of automated production, it must be rreated as a complex system
wtiich converts expenditures to product as a result of the mutual interaction of
- 253 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLV
all of its components (people, equipment, the technical conditions for production,
production stocks of materials, etc.); secondly, production control must also be
treated as a complex system, which performs the functions of resource distribution,
analysis and quality control for the final purpose of decision making; thirdly,
the two systems - production and control - must also be studied in their interac-
tion, since each production component is tied to each control function [66, 68, 69].
The relatively small amount of research in the f ield of comprehensive automated
production system design was one of the reasons for the fact that in the technical
literature devoted to this topic there is as yet no standard terminology.
At the present time, production, which incorporates several comprehensively
mechanized lines, is called comprehensively mechanized production or a comprehen-
sively mechanized shop, if it is enclosed within the framework of a single shop.
However, it is expedient to use these terms only as applied to a specific produc-
tion structure, while in the developmental stage for a univer.sal project plan,
their application is not.convenient. For this reason, the term "line" is fre-
quently applied without substantiation to several lines, arr:inged requentially
in different production stages. The terms "complex" and "set" which are frequently
used also lead to an ambiguous interpretation, since they are introduced in GOST
2.101=68 [State Standard 2.101-68] as a broad concept of specif ied products, the
main difference between which and the assembly unit is the fact that they are not
put together at the manufacturing enterprise in the assembly operations.
The term "system" is also a very broad and ambiguous concept. From the formal
viewpoint, a system is an aggregate of functional components which interact with
each other to achieve the set goal. A man working with a machine is already a
system. A"man-machine" system is only a component of a flow line system, etc.
In much system researc h, the concept of an "system" applies only to the process.
However, it must be recalled that systems analysis and systems engineering were
created primarily for working with objects in the physical world for the purpose
of creating technical systems. It is expedient to use the term "production sys-
tem", which most precisely reflects the existence of a facility, which basically
takes the form of an aggregate of technical hardware, lines and sections, in
contrast to a"system of production", which takes the form of a process for a
specific purpose, because of which the individual components are transformed into
a useful product. The ma3or terms and definitions adopted by the authors are
given below.
A complex is two or more specific products which are not put together at the
manufacturing enterprise in the assembly operations, but which are intended for
the performance of mutually related operational functions. Each of these specific
products, which is incorporated in the complex, serves to perform one or more
main functions, established for the entire complex (for example, a flow line, an
automatic telephone exchange, etc.).
A set is two or more products which are not put together at the manufacturing in
the assembly operations and which take the form of a set of products having a
general operational function of an suxiliary nature (for example: a set of spare
parts, a set of tools and accessories, etc.).
- 254 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500090004-3
FOR 1
A flow line is a set of main, auxiliary or lifting and transport production pro-
cess equipment, machines and mechanisms (consisting of a minimum of two units of
the main equipment which perform various operations), in which the operations of
reprocessing or assembly, which are carried out with human participation, and
these operations are assigned to definite equipment or definite work positions.
In this case, the sequence for the equipment configuration or the working posi-
tions conforms, as a rule, to the sequence for the performance of the operations.
A comprehensive mechanized flow line (KML) is a line in which all of the main
operations'of the production process for the fabrication of a product are per-
formed by mechanisms, machines or other kinds of equipment with a mutually linked
productivity, and additionally, the processes for transporting the products from
one working position to another are mechanized.
An automated flow line (APL) is a set of main, suxiliary, lift and transport pro-
duction process equipment, machines and mechanisms (consisting of a minimum of
two units of main equipment which carry out different operations), which execute
the operations of a portion of the production process for the fabrication of a
product without direct human participation and in a definite production process
sequence at a definite pace. In this case, there are both overall control and
automated transport devices to move the products from one type of equipment to
another, while man performs only the functions of set-up, observation and control.
The initial loading and final loading operations (or one of them) may be performed
manually in individual cases.
A comprehensively mechanized (automated) production system (KMPS) is a complex
hav ing an overall production program for a specific purpose and which takes the
form of an aggregate of a minimum of two comprehensively mechanized (automated)
lines, coupled by material transport flows and joined together by a cotmnon (auto-
mated) technology and production control system.
The planning of large systems such as production systems for semiconductor pro-
duction is impossible without a preliminary engineering and economic analysis of
the production process, without preliminary work to optimize the structure and
parameters of the system being planned, as well as to tie together and match up
the main indicators for the lines incorporated in the system.
An effective tool for analyzing a system and optimizing its parameters is modeling:
the main tool for checking the theories and design methods being created, as well
as the main tool of the optimal design theory. The model, in being a copy or
abstract representation of the major characteristics of any process, shows the
links which exist between the cause and effect, between the tasks and the capa-
bilities. The creation of a mathematical model for a production process is a
r:ecessary condition and the first step in the work on its automation. However,
the study of production systems using mathematical models would be impossible
without computers. The utilization of mathematical optimization techniques using
computers in the design of production systems opens up further possibilities to
improve the efficiency of production, and along with this, is the basis for the
development of the principles of computer sided design for automated lines and
production systems.
- 255 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
FOR OMFICIAL USF; ONLY
An important problem for the developers of comprehensively mechanized production
systems in the initial stage of planning is becoming the determination of the
optimal parameters of the system and the lines incorporated in it, as well as
the determination of an efficient structure for it, the optimum production volume
and the choice of the quality control system (Figure 12.2). The sequence for
the determination of optimal system parameters shown in the schematic is of a
conditional nature, since it is necessary to take into account their unavoidable
interrelationship during the project planning process.
The task of reliably determining the
~
Rei6ep
f inal parameters of the hardware being
c�mpyKmy17e1
designed, including the production
q~euN o~~,~-
engineering parameters of enterprises
~
~~Rnbi~OtO ob~- � ~
is directly related to the problem
r,.wa npna
Qodtma�

71
of parameter optimization; production
eN o~ �cre.~u
'
~
engineering parameter optimization is
~
ON/IlAO bHbS
an important condition for reliable
o~~~pou~~
M N
K
K
forecasts of scientific and technical
progress. Having calculated reliable
data on the f inal indicators of the
Production facilities being developed,
Figure 12.2. The sequence for the selec-
it is as if we are obtaining informa-
tion of the major parameters
tion from the future. This source of
of a production system.
infarmation makes it possible to depart
Key: 1. The selection of the
from traditional methods of extrapola-
structure;
tion (although only within the limits
2. The choice of the opti-
of the period being examined) and in-
mum production volume;
crease the reliability of predictions
3. The choice of the system
of technical progress for this period.
for positioning the
The economic interpretation of the
quality control opera-
indicators and the achievements of
tions.
scientific and technical progress are
of first rate importance in this case.
It is specifically this moment that should
become a connecting link betwePn the
general economic forecast of sector develo
pment and the particular technical
results anticipated with the creat'ion and
introduction of the new hardware and
technology into production. This is a complex and as yet not fully solved problem
in terms of inethodology.
The special position of the three following tasks follows from the definition of
a production system itself: the creation of the system for assuring the appro-
priate production conditions; the creation of the quality control, production and
technology control systems as well as the creation of organized material transport
f lows.
One of the most important factors which assure the technical level of production
is the design of the system for providing production systems with water and gases
of the appropriate degree of purity, as well as with dust-free environments.
Atthough many devices intended for this purpose are developed for applications
in individual installations and are even built into them, such a system should be
designed as a whole as applied to comprehensively mechanized production.
-256-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FQR OFFI(
The development of a quality control system is likewise necessary for the success-
ful functioning of a production system. The basis on which quality is maintained
and controlled, as well as quantitative monitoring - the feedback - in the final
analysis develops a quality control system into a production and technology control
system.
The performance of the functions which comprise the basis of the production process
depend on the system of material flows. For this reason, the methods of moving
materials and semi-f inished products, related to the use of the latest equipment,
should be developed simultaneously with the resolution of other production prob-
lems. A situation is more often encountered where the inclusion of a new piece
of equipment in a flow line makes the existing procedure for the movement of semi-
finished products ineff icient. J. Riggs [67] underscores the fact that attention
devoted only to one part of the overall flow line leads to the fact that the
solution of one problem generates another, on the solution of which the existence
of the entire flow line will depend. In this case, the author employs an analogy
with a river, where the cleaning of the bottom on any section of the river does
not increase the volume of water flowing through this section.
Chapters 13, 14 and �15.1 of this book are devoted to questions of designing sys-
tems for providing production with pure media, control systems for technology and
production as well as material transport flows using comprehensively mechanized
lines in semiconductor production.
12.3. The Engineering Economic Analysis of a Technological and Production Process
The engineering policy in all stages of the design of comprehensively mechanized
flow lines and production systems should be based on an engineering economic
analysis of the production process, for which this equipment is being developed.
The following goals must be kept in mind in this case:
1) The engineering economic stage by stage production analysis should precede the
formulation of the task and the advanced project plan, so as to determine the
production steps and operations where the greatest labor outlays and materials
are concentrated, and to determine the most "critical" production points as well
as determine the stages and operations in which there is the potential possibility
of obtaining the maximum effect, and thereby, establish the points for the neces-
sary concentration of the efforts of system designers;
2) The technical and economic analysis of the processes in the individual stages
should assist in determining in the preplanning stage through which components
the desired effect may be obtained (savings of the main or auxiliary time, savings
in materials, etc.), and thus, in choosing the direction for the solution of the
problem;
3) And, finally, the analysis should assist in determining the most efficient
path for solving the problem and choosing the optimal technical variant in the
developmental stage.
Some techniques of technical and economic analysis of production as applied to
problems which arise in project planning and design work on the automation of
semiconductor production are presented below.
- 257 -
ROR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
A Model of Material Flows of a Production System Based on Cost Indicators
When determining individual labor productivity at an individual enterprise or
section, only the expenditures for human labor at the given enterprise (or section)
are taken into account. Correspondingly, a savings in human labor, including
savings due to reduction in the labor input requirement, means a growth in labor
productivity.
The special position of the problem of quality in semiconductor device production
is responsible for the fact that the balance sheet for labor productivity is
governed in many respects here by the level of production process losses. This
circumstance does not allow one in the comprehensive analysis of production systems
to limit oneself to the labor input requirement indicator- for an individual sec-
tion, since the use of this indicator without taking production process losses
into account can lead to erroneous conclusions in the choice of the approach to
the design of new equipment. It is necessary to work from the overall labor
productivity, which is defined by all of the work time expenditures per product
unit, i.e., by the expenditures of current and past human labor, embodied in raw
materials, working materials, fuel and labor tools. The total outlays of present
and past human labor are characterized by the "economic input requirement for a
product" indicator which has been developed in recent years. However, the prac-
tical utilization of this indicator involves a number of procedural and practical
difficulties. Therefore, it is expedient in practical calculations to use an
^:,timate of the optimality of a variant based on the "minimum production cost"
criterion, since the production cost indicator sufficiently precisely accounts for
the additional expenditures of materials and semi-f inished products related to
production process losses.
First section 2nd section i-th section n-th
/-u y~atmoK '-d y~atinoK i-d yvccmoK n-u uvac~mK
M~=NpCp I Mt Ni
(t) I
K
Qr Ks Atppr T` At .4,00
L..._
p~
(2) Tflz \rnt~'
/ U
sec t ion
Figure 12.3. A model of the material f lows for a production system.
Key: l. Deff 1[section 1 cost];
2. TPl [section 1 production process losses];
3. Deff n [n-th section effective cost].
A model of material flows in a production system, which tal:es into account expen-
ditures related to production process losses, can be represented in the following
form (Figure 12.3). The system consist5 of n production process sectians or lines;
-258-
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500090004-3
at the end of each of these, the product quality is checked, which, as was noted
above, is one of the conditions for singling out a section as a subsystem. As
a result of the quality control in each section, the rejects and production pro-
cess losses TP are ascertained. Since quality control operations can also be
carried out within a section, we understand Ki ir. this model to be the entire qual-
ity control sum, which is realized in the section, including the final one, while
TPi is the sum of all of the losses ascertained in the section as a result of the
quali.ty control. The semi-finished products from the preceding section are fed
to the i-th section, where these products are acknowledged as good following the
Ki_1 quality control with an overall cost of Deff i-1, as well as the additional
materials for the processing in the i-th section of the entire amount of semi-
finished products which arrive at this point. The cost of these mat'erials (or
semi-f inished products) is expressed by the Mi. The labor expenditures for produc-
tion and quality control of the operations, carried out in the i-th section, com-
prise the quantity Ti.
To carry out an engineering economic analysis of a production system using the
model proposed here, it is necessary to introduce an indicator which evaluates
the cost level of the production proczsE losses.
The Cost Coefficient of Production Process Efficiency
The generally accepted characteristic at the present time for the sifting out
of defective products in semiconductor device production is the good product yield
coefficient (or percentage), K; let the quanzities Ki be the yield coefficient
in the i-th operation (or section) and rlg i is the overall yield coefficient in
the i-th operation from the start of the prccess. This in3icator is rather infor-
mative and convenient in estimating the techni..al level of production, for the
operationally timely analysis of the course of a production process as well as
the analysis of a local production process. It is needed in the calculation of a
whole series of parameters for complex lines, including their effective produc-
tivity, the line pace, the quantity of process stock, etc.
However, the good product yield coeff icient does not suff iciently completely
reflect the cost level of losses in a comprehensive analysis of several production
sections arranged in serj.es, since it does not take into account the differing
volume of losses in the indicated sections, including those related to the arrival
of additional materials and semi-f inished products at the given section. It is
apparent that for production processes producing devices which differ in their
structural design and production technology, where these devices have the same
yield coefficient, the specific share of the production process losses diff ers
in the cost expression. For this reason, in line with the model adopted here, w.:
shall introduce a cost indicator along with the good product yield coefficient,
where this cost indicator takes into account the production process losses [68],
and we shall consider its relationship to the yield coeff icient (Figure 12.4).
Let Ci be the cost of all of the outlays (for materials, processing, product
measurements, etc.) to produce one product in the i-th production section (or
step). Then one can write the total expenditures in the section, Di, in the
following form:
- 259 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
)
D;=,(M,+T,) + (M2+T2) + . . . +i(M;+7'j) _
=No(Co+Ci+TIoiCs+ ++18 t-A),
where Co is the product cost at the
start; ND is the number of products
started in production.
The cost retained in production aftar
the i-th section amounts to:
t
D3(D(b {=ijBjNoC=+locNo k E 0 Ch.
a
N
Y y~
~ , N l
~
5)
R ti~
~
h
R
6)
Figure 12.4. The ccst coef f ic ient of
produr_tion process eff i-
ciency.
a. Material balance sheet
for the production of
semiconductor devices;
b. The relationship of
the yield factor EB
and the cost coeffi-
cicnt of production
process efficiency
Zc.
Key: 1. Deff n feffective cost
of the n-th section];
2. Effective cost of the
(n - 1)th section;
3. TP1 [production pro-
cess losses of section
1] ;
4. Production process
losses of section n;
5. ncn - Ceff n/Dn'
!
where C= Z C; is the aggregate of
k-o
expenditures in the i sections for
one product; ngiNo is the number of
good products following the i sections.
The quantity:
1
Tiet Y. Ch
11C'! k=o
pi c
Qn g Cr,
L
k =0
is the cost coeff icient of technological
production efficiency, which is intru-
duced in a manner similar to the good
product yield coeff icient. It is not
difficult to see from the latter rela-
tionship that the following inequality
is observed:
'9 ai>ct
. w y.J (b � N N � r-i
r-I O'G JJ N~o cC
Nr4 U+4 �~vi N N D
0044 U H ~
u u^'4 ~ 1-4 `o ~ ~ ~
aki ~ N~ oow 3w'1 ~
d Cl q~ f~=+ O'~ a
cna~�~ �~u
r-1 u1 ~ O~ U O rl c0 -I N
N
DG
-281^
FOR OFFiCiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
The degasifier (Figure 13.5) is intended for the removal of carbon dioxide from
the H-cationized water. The cylindrical vinyl plastic housing 8 consists of
three sections, 5 and 6, f illed to a height of 2.6 m with vinyl plastic chordal
baffles. The baffles are made so that the plates 3 of one row overlap the gaps
between plates 4 of the next row. The water exits on top through a branch pipe
and is uniformly distributed by a special plate 7 over the surface of the attach-
ment. There are 48 water d istribution branch pipes on the plate and 8 branch
pipes for the air outlet. The branch pipes for water distribution are distributed
uniformly over the area of the plate. Raschig rings are also used to f ill the
body of the degasifier. The degasifier column is mounted on tank 9, in which the
water is collected following degasif ication. An ESU-1 level indicator is mounted
in the tank. Air is delivered to the lower portion of the column from fan 2,
which is mounted on frame 1 alongside the tank. The air from the fan goes upward
to counter the flow of water and goes out into the atmosphere.
The combined action filter is intended for holding the mixture of KU-2-8 chS
and AV-17-8 chS ion exchange resins. The housing is made from steel pipe, the
diameter of which is computed as function of the output of the unit; the internal
surface of the pipe is rubber coated. The structural design of the combined act-
- ion filter is similar to the structure of the conventional filtration column
described above, with the exception of the center drain unit.
The basic production process scheme for producing ultrapure water at a rate of
2 m3/hr is shown in Figure 13.6. A high degree of purity is assured by virtue
of the use of a multistage processing system, which includes the removal of micro-
particles in the mechanical filter 5(the removal of particles with dimensions of
more than 20 um); the coarse cleaning filters 7 and 11 remove particles larger
than 6 um; the fine cleaning filters 22 and 27 remove partic?.es larger than 2 um
and 0.2 um respectively; the purification employs the reverse osmosis technique
using roller type elements 13, degasification, double ion exchange purificatici:
in combi.ed action filters 21 as well as ultraviolet sterilization. To provide
for continuous system operation, collecting tanks 25 with a capacity of 10 m3
each are provided where necessary for the restoration of the combined action fil-
ter; additionally, the combined action filters are structurally designed so that
they allow rapid replacement with a filter which has been restored in a separatt
special section. Where a greater purified water consumption is required, the
collecting tanks make it possible for two production process finishing "lines"
to operate in parallel, which boosts the system output up to 4 m3/hr.
The unit for water purification using reverse osmosis (Figure 13.7) is intended
for water purification in a semiautomatic operating mode, and when operated in
conjunction with combined action filters, makes it possible to obtain water with
an electrical resistivity of 8 to 10 MOhm � cm. The starting wa~er, which is
cleaned in a mechanical f ilter, is fed into the purification block. All of the
assemblies of the purification block are mounted on frame 1, to the bottom frame
of which the high pressure pump 9 is fastened as well as the tank for the acid
solution 7 and the tank for the washing solution 8. There are three purification
modules S, a filtrate flow rate indicator 4 and a concentrate flow rate indicator
6, as well as filters for cleaning the washing solution 3 and a pH meter 2 all
arranged in vertcal racks in the frame. The installation is controlled from a
control console, on which there are the following: pushbuttons for checking the
- 282 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
Figure 13.7. Installation far water purification using reverse
osmosis.
operability 10; for resetting the alarm signal and turning the unit on again 15;
"Alarm Signal Cutoff" toggle switch 14; small monitor lights for installation
operation; the adjusting control for pH meter 12; ammeter 13; counter 11; and the
main switch with mechanical interlocking 16. The power panel is located inside
the control console and power is brought into the console from below.
The basic production process schematic of the water purification unit using
reverse osmosis is shown in Figure 13.8. Th initial water at a pressure of 3
to 4 kgf/cm2 is fed through valve 1, mechanical f iltar 2, valve 3, and electro-
magnetic valve 6~-o the multistage centrifugal pump 7 via a pipe into which a
33% acid solution is also fed from tank 24. The acid ia supplied to correct the
pH of the initial water, which should fall in a rlnge of 5.2 to 5.6. The water
is delivered by pump 7 at a pressure of 28 kgf/cm through open valve 21 and pipe
15 in part to filtration modules 14 and in part recirculates through bypass valve
23. In flowing through modules 14, the water is collected in the collector 13,
and then passes through the flow rate meter 12 and is fed through valve 9 to
the user or for further purif ication. The pressure of the water fed into the
modules is monitored by means of manometer 20 and contact manometer 19. The
concentrate is drained from the modules 14 through receiving manifold 16, valve
17 and valve 11. The concentrate pressure is checked by a manometer and adjusted
by valve 17. A portion of the water incoming for purification is fed through
valve 22 to sensor 8 to check the pH and is dumped into the drain.
In the case where the purification modules 14 become fouled, they are chemically
flushed with a solution prepared in tank 4. A 2% (by weight) solution of citric
acid is used to wash out iron, with the subsequent addition of ammonia (NH40H)
to obtain a pH = 4. In the case where calcium sulphate is washed out, the same
solution is used, but with a pH of 7 to 8. Calcium carbonate is washed out using
a solution of sulphuric or nitric acid with a pH of 4 while organic substances
are removed by an alkali solution with apH of 13.
The solution is delivered by pump 7 from tank 4 through valve 3, sensor 5 and
valve 6 to the filtration modules 14; the filtrate is then returned through
valve 9 to tank 4. Prior to returning to tank 4, the concentrate is cleaned in
- 283 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500090004-3
FOR OFFICIAL USE ONLY
n n .
is
Concentrat is .
Ma~Y~MmOsm / ~ ~ ~
IO /1
Filtrate
9
mun~mpam '
Z J S 6 d 11 m
ti ZI
Mm ~ - �
t
1q '
xuru- NctoJrd~ �~urd ~ M=SO~ .
R8W Eoda ~ C) c!~
aater ;J
'~B)
i
Figure 13.8. Production proceas scheme for the water purif ication
unit using reverse osmosis.
Key: A. Chemical solution;
B. To the drain;
C. SU = Not further defined.
filters 10. A contact thermometer 18 is ir.stalled at the outlet from the pump 7
on the pressure delivery line, where this thermometer disconnects the unit when
the temperature goes higher than the permissible value.
13.3. Equipment for Finish Water Purification
Following the preliminary purification of water with a resistivity of 1 MOhm - (-i
and more, it fed for thorough desalinization to a finish purif ication installation.
The configuration of a string of finish purification units, designed for the use
of ion exchange in a mixed layer of KU-2-8 chS and AV-17-8 chS ion exchange
rosins, includes type UF-100A, UF-250 and UF400A units. Also incorporated in t'r-
configuration of a string of f inish purification installations are type UFE-100
and UFE-250 units, which employ the technique of electrodialysis with intermem-
brane filling using a mixture of the ion exchange materials indicated above. The
outputs of the units are 100, 250 and 400 liters per hour respectively; the elec-
trical resistance of the water following purification in the units is 15 to 20
MOhm�cm at a temperature of 20 �C. The units should be placed close to the points
where the water is used. In this case, to prevent contamination by the atmosrhere
of the assembly shops during the recovery of the spent resins, this proce'!-~ is
carried out at a central location in an isolated room using special recovery equip-
ment. For this purpose, the filtration column of the finishing units is made so
as to be easily disassembled. The continuous monitoring of the course of the pro-
cess is accomplished using a continuous flow meter and an instrument for measuring
the water resistivity at the filter column inlet and outlet.
The UF-250 finishing water purification unit is shown in Figure 13.9. Desalinated
water with an electrical resistivity of 2 to 3'�'Ohm�cm, obtained from the
- 284 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
centralized unit, is fed through sensor 1, inlet valve 2 and direct reading flow
meter 3 to the filter column 5, and then through outlet valve 6 and sensor 7
to the user.
E:~ R
PA Q PV Z
R J
R al (a) R` r Rs R Q-
~ti ~
~ Q D
^ el
(
(b)61 v
Figure 13.10. Water resistivity
metering circuits.
Figure 13.9. The UF-250 final water
purification unit.
The degree of water purification is moni-
tored by means of instrument 4, which
indicates the resistivity of the water at
the outlet. When the resistivity falls
below the permissible value, a signal light
on the panel of the instrument turns on,
the initial water feed valve into the
filter turns off and the filter is dis-
connected for restoration. The f ilter is
made in the form of a cylindrical housing
of plexiglass. There is a drainage disk
in the lower portion. There is a plug
in the upper cap of the filter for the
release of air from the fil*_er in the
initial operating period of the unit. A
direct reading flow meter serves to meas-
ure the water rate of flow incoming for
purification. Its housing is also made
of plexiglass. There are divisions which
show the water rate of flow on the exter-
ior surface of the housing.
The sensors of the flow-through meter for monitoring the water resistivity,
which are secured to the frame of the unit 8, consist of a cup and a nozzle,
made of plexiglass. Two stainless steel electrodes are screwed into the cup,
where the gap between the electrodes is set duiing the adjustment of the instru-
ment. The sensor is connected to the measurement instrument' by a shielded cable.
'Phe basic electrical resistivity measurement circuits for deionized water are
shown in Figure 13.10.
The operation of the meter, a block diagram of which is shown in Figure 13.10a,
is based on the measurement of the current flowing through a measurement cell,
D. With a constant voltage applied to the circuit, when the resistivity of the
solution in the measurement cell changes, its resistance changes, and
- 285 -
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R040500090004-3
FOR OFFICIAL USE ONLY
Lonsequently also the current flowing through the circuit. Thus, one can judge
the resistivity of the solution in the cell from the current in the circuit, i.e.,
R = f(I).
The scale of the microammeter PA is graduated in values of the resistivity (MOhm �
cm) .
If the additional resistance R in the circuit of Figure 13.10b is chosen two to
three orders of magnitude greater than the resistance of the cell being measured,
D, then a change in the current flowing through the circuit with a change in
the resistance of D from Dmax to Dmin can be disregarded in practice.
Thus, one can judge the resistance of the measurement cell by measuring the voltage
drop across it, i.e., R= f(DU). The scale of the millivoltmeter PV is graduated in values of the resistivity
( NtOhm � cm) .
l
The operation of the meter, the measure-
ment portion af which is depicted in
Figure 13.10c, is based on the measurement
of the imbalance of the bridge which is
due to the change in the resistance of the
measurement cell D, inserted in one cf its
arms.
The thermistor Rt serves to compensate for
the change in the water resistivity wit;t
a change in temperature.
The imbalance signal is fed following
amplification te the meter, the scale of:
which is graduated in MOhm � cm.
The URS-1 installation (Figure 13.11) i
used to segregate and recover the KU-2-8
. chS and AV-17-8 chS ion exchange resins,
which are used up in the UF finish puri-
Figure 13.11. Unit for segregating a.nd fication installations, where this unit
recovering resins. can also be used as a finish water purifi-
er with an output productivity of 800
liters per hour when the two filtration
columns operate in parallel, or as a unit for producing desalinated water having
a resistivity of 1 MOhm � cm and higher from the water mains. In the latter case,
the filter columns which are loaded with the resin mixture, are connected in
series through a degasif ier and the output productivity of the unit will be 400
liters per hour.
- The unit has two columns for the ion exchange resins, a degasifier column for
the :emoval of carbon dioxide which is filled with Raschig rings made of
-286-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
polyethylene tubing, as well as a pump for delivering water from the deg:sifier
tank to the second purif ication stage. Located on the control panel for the tin it
is a meter for measuring the water resistivity, which makes it possible to deter-
mine the quality of resin recovery or the thoroughly desalinated water wtlich is
produced. An ejector system is provided in the unit for the preparation of the
recovery solutions.
The L'FE-250 finish water purifier (Figure 13.12) takes the f orm of a composite
structure, consisting of the support frame, electrodialyzer, direct reading flow
me[ers, cutoff valves, sensors, electrical circuitry and a piping system. The
support frame consists of a frame, base and sheathing; the frame and the base take
the form of a welded structure of angle irons and serve for the housing and
securing of 211 of the assemblies. The direct reading flow meters, electrical
circuitry, valves, solution tanks, water resistivity sensors and piping system
are located within the support frame. The dialyzer is secured to the front wall
of the frame. The instruments for metering the electrical parameters (ammeter
and voltmeter) are mounted oi1 the front panel of the unit as well as the water
resistivity measurement block and the current density adjustment control. The
cutoff valves are fastened to the inclined front panel. The side walls of the
frame are covered with removable bent steel section sheathing. The rear wall of
the frame is covered with a removable door. There are a plug connector for thu
electrical current (power maims),a grounding bolt and outlets for the production
process pipes on the rear wall of the base. A group disconnect switcti is located
on the right side wall of the base. There are holes in which handles are inserted
in the side walls of the frame for moving the unit.
The electrodialyzer takes the form of a
composite structure of two electrode
chambers E, desalinization chambers U
and brine chambers B, the number of
which depends on the output productivity
of the installation. The chambers are
assembled in the following sequence:
E--B--D--B--D...B--E.
Figure 13.12. The installation for the
f inish purif ication of
water using electroioniz-
at ion .
All of the chambers are separated from
each other by�cation and anion exchange
membranes and are joined together through
gaskets using studs, which are insu?ated
with polyvinylchloride tubing. To pro-
vide for a gap between the electrode and
tlie membrane in the electrode chambers
and between the membranes in the brin,
chamber, gaskets are inserted made of
perforated and corrugated vinyl plastic
film. There are holes which are plugged
with plugs in the upper portion of the
desalinizatlon chambers, through wtiicii
the resins are loaded in; there are holes
in the lower portion for water delivery
to the electrode and brine chamhers.
- 287 -
FOR OFFiCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
FOR OFF ~ iAL USE ONLY
,
!
i
There are manifolds for water distr_t,tion to the chambers in the upper and lower
portions of the electro3ialyzer; there are air vents in the upper portion for the
removal of air from the desalinization and electrode chambers. Direct reading
flow meters are installed in the unit for the measurement of the purified water
flow (unit output) and the water flow in the electrode and brine chambers.
The unit operates continuously and requires no chemical regeneration. The recovery
cf 07 resins takes place dLring the desalinization process by virtue of the
pa electrolysis of the water into H+ and OH- with the application of direct
cur. -
The major user of ultrapure water is equipment for the chemical treatment of
wafers. This fact is taken into account in the desion of equipment for water
purification. A schematic of a recirculation systr..m for producing ultrapure
water is shown in Figure 13.13. The major companents of the system are secured
in a dust free box, in the working volume of which the chemical treatment unit
is placed. For reducing the water flow rate and expendi*_ures related to its
purification, a provision is made in the system for the capability of the repeat
use of a portion of the water (with a r_esistivity of more than 1 MOhm � cm) follow-
ing the washing operation. In struc*_ural terms, the system consists of the block
- of filters 1, the water delivery u:iit 2, the water return unit 3 and the fine
~ cleaning filter for the deionized water 4.
flLL' ~
rp
r
Figure 13.13. The recirculation system for producing ultrapure water.
Key: 1. Block of filters;
2. Water deiivery unit;
3. Water return unit;
4. Fine cleaning filter;
5. Cascaded washing bath;
6. Resistivity meter.
- 288 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500090004-3
NOR OFFICIAL USE ONI.Y
The incoming water in the block of filters (Figure 13.14) is sequentially cleaned
of inechanical particles larger than 2 Um in the preliminary cleaning filter 6,
which takes the form of a three layer filtration system: two layers of cardboard,
which play the part of a preliminary filter and a substrate, as well as a layer
of filtering material positioned between them (the extended filtration surface is
achieved by virtue of corrugating the filtering layers). The water is then puri-
fied of organic impurities by their sorption by a layer of macroporous anionic
exchange material or activated charcoal (filter 7), and also cleaned of ionic
impurities in a mixed layer of ion exchangers (combined action f ilter 8). The
units of the filter are changed in step with the degradation of water quality,
which is checked by resistivity meter 9, as well when the hydraulic resistance
increases above a set level by virtue of the clogging of the filters, which is
registered by manometer 4. The block is also equipped with a flow rate meter 2
and a check valve 1, which prohibits the flow of water in the return direction
when the system is disconnected.
Figure 13.14. Basic schematic of the
block of filters.
1 Ncz
In
water
o' Pur if ied
water
Following the fine cleaning filter, the
water is fed to the user (in Figure 13.13,
to the cascade washing bath 5), and then
to the return unit and through the open
valve is dumped into the drain. Upon the
signal from the meter 6 when the drain
water reaches a specified resistivity,
the valve is switched and the water is
fed into the collecting tank of the water
delivery unit.
To retain the deionized water parameters
achieved following purification, the
retaining f ittings are made of technical
plexiglass and teflon, while the distri-
bution system for the water delivery from
the intermediate centralized purification
equipment, is made of high pressure poly-
ethylene pipes, vinyl plastic pipes or
seamless cold drawn pipes of corrosion
resistant steel; high pressure polyethyl-
ene pipes or teflon 4D pipes are used to
deliver the deionized water from the
finish purif ication units. The piping
should be able to be disassembled for
ease of washing and repair. The pipes
are washed no less than once per quarter
with a 3 to 5% solution of hydrogen
peroxi,de or sodium chlorate which is kept
h
Key: 1. Check valve;
2. Flow rate indicator;
3. Filter housing;
4. Manometer;
5. Housing;
6. Preliminary cleaning
f ilter;
7. Filter for removing
organic contaminants;
8. Filter for removing
ionic impurities;
9. Resistivity meter.
in the piping for no less than one hour. After the solution is drained out, t e
internal surface of the pipes are flushed with type V deionized water until a
resistivity of the flushing water of 1 MOhm � cm and an oxidability of no more
than 1.5 mg 02/1 are achieved.
-289-
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3
v.5 6 7 B'! 9

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500090004-3
F'OR OFFI('IAI, IISF. ONI.Y
13.4. L'quipment for Gas Purification and Drying
The installations for the chemical adsorption purification of nitrogen and hydro-
$en used in semiconductor production have an output productivity of from 100 to 1
nm3/hr*; installations for the adsorption purif ication of air have an output of
from E,000 to 30 nm3/hr and diffusion purif ication units for hydrogen have outputs
of 4, 10 and 50 nm3/hr. The ma3or impurities which must be removed from the
gases being cleaned are moisture, oxygen, hydrogen and dust.
The basic production process scheme for the purification of nitrogen in the UOGA-
50 (I), operating as part of a complete set with the BtTV-50 (II) excess hydrogen
removal unit, is shown in Figure 13.15. In the case of operation with simultan-
eous restoration of one of the adsorbers (for example, A1), the isolation fittings
are set in such a position by the switching of the control levers that the nitro-
gen incoming for purification goes into adsorber A1, heated by the heater 6uilt
into it, and extracts the moisture released from the silica gel. The restoration
takes place at a temperature of 180 to 190� C.

kTater
L- TJ J
1
~ o----- AYamNa i
i ovucmKy
1 ~
e IM~ I
~ p 1voDOQoD ~
T
~ I
I XN ~
I ~
I A, At ~
I ~
I ~
I
a t er
~ eoaMrr-"
47ater i
L__
Figure 13.15. Basic production process
schematic of the UOGA-50
nitrogen purification unit.
Key: 1. Nitrogen for purification;
2. Freon;
3. Hydrogen;
IM = Actuating mechanism;
K = Electromagnetic valve;
To avoid the failure of the seals of the
isolation fittings because of exposure
to high temperatures, the nitrogen is
water cooled in a heat exchanger T, fol-
lowing A1 (T2 in the case of restoration
of adsorber A2). Then hydrogen is added
to the nitrogen through electromagnetic
valve K and the automatically controlled
valve of the actuating mechanism IM in
an amount which is 0.5 to 1.5% greater
than the stoichiometric ratio for the
amount of oxygen impurity. The nitrogen
and hydrogen mixture is fed to the puLI-
f ier 0, which is filled with a palladium
catalyst, in the presence of which the
hydrogen bonds the oxygen so that at a
temperature of 90 to 100� C, the residual
volumetric fraction Qf the latter amouiit.,
to less than 1- 10-4%. Then the nitro-
gen being cleaned is fed into the reactor
*At normal atmospheric pressure.
- 290 -
0= Pur if ier [ f illed with
palladium catalyst];
R = Reactor;
A1 = Adsorber 1;
T1 = Heat exchanger 1.
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500090004-3
FOR OFFICIAL USE ONLY
R of the hydrogen removal unit, where the following reaction takes plsce:
350� C
Cu0 + f Iz-+Cu I 1:0.
rhe copper oxide, which is precipi*_ated on alumogel, assures a residual volumetric
oxygen content of less than 1. 10-3Y. Being cooled in the heat exchanger T3 to
a temperature of 30 to 35� C, the nitrogen is fed to heat exchanger I, which is
cooled by freon by the MF-SbM type refrigerator KhM. When cooied down to 3 to 5�
C, a large fraction of the moisture condenses in the evaporator and is drained
from it through an automatic condenser outlet tap; and additionally the sorption
capability of the adsorbent is increased. The nitrogen is fed from the evaporator
to the adsorber A2, where the residual moisture content is reduced down to 10 mg/
m3, and fed into the mains through the FAG-50 dust removing filter F. When the
copper oxide turns to copper in the reactor P, valve K cuts off the hydrogen feed.
The removal of zhe oxygen now no longer takes place in purifier 0, but in the
reactor P in accordance with the reaction:
Cu35o~� C 0,-->2 CuO.
The residual volumetric oxygen content is 1
controlled [771.
Vucmdg ~~oJapod
I (2)
Om,xodbr
B cB~v
0
o �
o �
o �
o �
(3)
0
o �
o � Bodapad
o ( � No ovacnry

Figure 13.16. Schematic showing the opera-
tion of a palladium filter
for hydrogen purification.
Key: 1. Pure hydrogen to the
user;
2. Waste to the flare;
3. Hydrogen for purification.
10-4%. The valve K is automatically
The installation is equipped with
automatic instruments for monitoring
the residual oxygen and hydrogen
content (PKG-1S) as well as the
moisture content (DV-1). The con-
figuration of the UOGA-100 unit is
similar to that described abave.
Zeolite is packed in the adsorbers
in the UOvA-25 unit. The evacuation
of the moisture liberated from the
zeolite during recovery is accom-
plished by means of a nitrogen (or
air) flow which is exhausted into
the atmosphere.
The units for the adsorption and
catalytic purification of hydrogen
differ from the nitrogen purif ica-
tion units in that there are no
blocks in them which involve the
dosing of hydrogen into the flow
being purified or for monitoring
its residual content.
ln tlie hydrogen purification units (UOGV), heat exchanger T3 (Figure 13.15) is
inscrted directly after the purifier 0.
The UOGV units can be employed where it is necessary to remove hydrogen and mois-
ture from oxygen.
- 291 -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500090004-3

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500090004-3
Z G
~ O
914
H -W
r-I
W ~
td
z -w
o u,
H
Q r
H
U
H w.
44 C
H
P~
~
a
~
C.7
I
a
r
>
E
FOR OFFI('IAI. USE ONI.Y
0
~
o
o
�
`O
c
a w
o

~
C)
0

V V
~
d O
O
14
U H
W
N
o
~
~
j
p
j
v1
O 00 M
rl
~
p
N
EI
C
7
'd
a
N
U
~G
~
>
1
I
~
~
N
I
I
~1
G
l
~
?
,--I
I
D
O
I
u1 e'n
H
1
.
0
O
.d
O
u1
O I
O
N ~ N
N
O~
W
O
O
~
~
O
N
N
a
r-I 0
0
0
0
~
O
O ~ 1
H
O
~
O
7
x
O
N N
~
N
m
>
o
~
'
a
~
+
Cl
N
4)
u1
(30
~
~
O
N
C)0
00 ~
C
9pq
'

z
N
N
z
~
~
>
I
~
OO
N
I
u1
H
O
.Y cn
r-1
1
j
1l1 M
1
1
H
I
d`r~
rl
z
V1
o
p
u,
c)
0
-4 -4
o i
i
+
o
~
Q
N r-1 N
'
L+
~
00Ln
O
O
O
+
~
O
41
r-I N
'r
7

N r-1
0
O
~
u cn ~ ~
a
w
ro
a.
3 N
r-'
~
H
O
o
i c
~
N
~
p'
~
r~
~
a a
a
i
a
i
~
G
~
~
u
p
p
i
a
o0
3
0 o
o
p
w..
a,+
u
.
M
a~
p
N
C!
CJ O a+ ;:L
O 00
U
rl
r-I
H
^
?
00 H N
11
.ri
JJ C;
e'-1
M
ld
C)