A COMPARATIVE ASSESSMENT OF SOVIET OPTICAL COMMUNICATIONS CAPABILITIES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP92B00181R000300270029-8
• Release Decision: RIPPUB
• Original Classification: U
• Document Page Count: 63
• Document Creation Date: December 27, 2016
• Document Release Date: April 22, 2013
• Sequence Number: 29
• Publication Date: January 21, 1987
• Content Type: REPORT

V
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
UNCLASSIFI1ED
OM(
A COMPARATIVE
ASSESSMENT OF
SOVIET OPTICAL
COMMUNICATIONS
CAPABILITIES (U)
Draft Report No. M2012 s7
NOVEMBER 1985
REVISED 21 JANUARY 1987
Prepared by:
K. Bower
J. Esposito, GTE-SSD
Contract DAAH01-85-C-A093
Western Division
GTE Government Systems Corporation
100 Ferguson Drive
PO Box 7188
Mountain View, CA 94039
ILLEGIB
UNCLASSIFIED
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
GTE GOVERNMENT SYSTEMS CORPORATION - WESTERN DIVISION
P.O. Box 7188
Mountain View, California 94039
A COMPARATIVE ASSESSMENT OF SOVIET OPTICAL
COMMUNICATIONS CAPABILITIES (U)
Draft Report No. M2012
November 1985
Revised 21 January 1987
K. Bower
J. Esposito, GTE-SSD
Prepared for
U.S. Army Missile Space Intelligence Center
Under Contract DAAH01-85-C-A093
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
(U) TABLE OF CONTENTS
Section
ABSTRACT
1 INTRODUCTION
2 STATE-OF-THE-ART
2.1 Fibers
1
3
6
2.2 Components 8
2.3 Systems 15
3 SOVIET TECHNOLOGY 18
3.1 Fibers 18
3.2 Components 22
3.3 Systems 25
4 PROJECTIONS - WESTERN 34
4.1 Fibers 34
4.2 Components 39
4.3 Systems 42
5 TECHNOLOGY. REQUIREMENTS 45
5.1 Fibers 45
-:5.2 Components 48
5.3 Systems 49
6 LOCAL AREA, LONG DISTANCE F.0 SYSTEMS 51
REFERENCES 53
ii
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
1
ABS TRACT
The purpose of this report is to indicate whether or not
the Soviets have the capability to deploy 5 Mb/s fiber optic
communication systems to support an ABM system.
The evidence provided by the literature search seems to
indicate that the Russians can deploy. such a system without
importing technology from the West. They demonstrated a
short distance system capable of transmitting 114 Mb/s as
long ago as 1977. Since that time, they have demonstrated a
system as long a/.8 Kb; and they have demonstrated a 5 Km long
link utilizing both wavelength division multiplexing and
duplex transmission. They have also done a number of studies
on the radiation hardness of optical fibers and components.
More important than the systems actually exhibited by
the Soviets, are the capabilities of the components that they
have developed. They are working on distributed feedback
lasers and Schottky barrier photodetectors which are capable
of multi-Ghz emission and detection, respectively.
Additionally, they have demonstrated the ability to produce
single mode fibers; it would therefore seem that the Soviets
may already be able to transmit data in the low Gb/s range.
For comparison, the West has demonstrated the following
capabilities:
1. 2 Gb/s transmission for 130 Kb unrepeatered.
2. 400 Mb/s transmission for 250 Km unrepeatered.
3. Laser modulation rates above 20 Gb/s.
M2012
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
4. Schottky barrier detectors with bandwidths near 20
GHz.
Most, if not all, the systems described in the Russian
literature seem to have been assembled from domestically
produced components; and, in fact, several of those
components were described as having been mass produced. It
seems doubtful that Russia can routinely manufacture
components as capable as those described in its literature
since the West cannot do so, but there is a large gap between
the prototypes exhibited and a 5 Mb/s system. This gap is
large enough that it seems likely that a 5 Mb/s system can be
deployed.
2
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
Section 1
(U) INTRODUCTION
The basic question to be answered by this report is
whether or not the Soviet Bloc can produce fiber optic
transmission systems by utilizing, only their domestic
technology. Stated specifically, can the Soviets produce
systems with a capacity of 5 Mbps without importing
technology from the West?
Beyond simply answering whether the Soviets can build
systems of a certain data capacity, this report will address
the broader question of where the Eastern Bloc stands in this
particular technology. To do this, it is necessary to review
the state-of-the-art, review the state of Soviet art, and
estimate what lag , if any, exists between the two. The
method used to accomplish this task is outlined in the
following paragraphs.
To review the state-of-the-art it was only necessary to
update the author's own knowledge of the field. This was
accomplished by reading a number of recent review articles.
If the review articles spoke of areas in which the author's
knowledge was not current, specific articles in those areas
were obtained. In this manner, it was possible to ascertain
both where the industry stands at this time, and where it is
likely to go in the next decade.
To review the state of Soviet art was more difficult
because their review articles heavily reference Western
3
M2012
ILLEGIB
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
sources. Thus, unless a Russian review article has a
preponderance of Russian references, it is difficult to
ascertain the Soviet contribution. Therefore, Soviet review
articles may simply restate for a Russian audience the
content of Western review articles. It was therefore
necessary to read a number of Russian papers in each specific
area of interest. Theoretical papers were ignored because we
must acknowledge that Soviet theoretical knowledge is on the
par with the West's, if for no other reason that that they
continually reference Western literature. Experimental
papers, on the other hand, can often tell one more than the
result of a particular experiment; often, the sources of
components and equipment can be determined. Similarly, if a
paper describes the development of a dispersion measuring
instrument capable of sweeping to a GHz, one can be sure that
its authors have access to lasers and detectors with response
times on the order of 1 ns.
To estimate the lag between Soviet and Western
technology one must correlate a development's first
appearance in Western literature with similar developments in
Soviet literature. This requires that, the history of the
development of fiber optics in both the Eastern and Western
blo-6's be researched and compared. But even such a comparj.son
does not give the entire picture, because the open literature
does not give grie an indication of Soviet manufacturing
capabilities.
4
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
It therefore seems that one is confronted with two
choices when trying to estimate Soviet capabilities in any
particular field. One can assume that the worst case always
prevails; thus, if the Russians demonstrate a prototype
system, one assumes that they can manufacture such systems in
quantity. The other method of estimating Soviet capabilities
depends on knowledge of their historical lag between
prototype and production, but such knowledge is likely only
to be available to the intelligence community. Without this
specialized information, the author reverts to the first
scheme and assumes that the Soviets' ability to deploy rests
at the level of the most advanced prototype demonstrated.
The evidence provided by the literature search indicates
that the Soviets can deploy a fiber optics system without
importing technology from the west. This assessment is based
on the fact that the Soviets have demonstrated prototypes of
systems with the required technology and does not consider
whether they can reliably manufacture all of the components
of a demonstrated system in quantities necessary for
widespread deployment. They demonstrated a short distance
system capable of transmitting 114 Mb/second as long ago as
1977. Since that time they have demonstrated a system as
long as 8 kilometers; and they have demonstrated a 5
kilometer long link utilizing both wavelength division
multiplexing and duplex transmission. They have also done a
number of studies on the radiation hardness of optical fibers
and components.
5
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Section 2
(U) STATE-OF-THE-ART
2.1 Fibers
Since the development of the first 20 dB/Km fiber in
1970, the losses in high silica fibers have dropped to the
point where they rival their theoretical limits. These
losses for single mode fibers are 0.16 dB/Km at 1550 nm (1)
and 0.27 dB/Km at 1300 nm (2). For multimode fibers, the
losses are somewhat higher, 0.31 dB/Km at 1550 nm (3), due to
higher concentrations of dopants and consequently higher
Rayleigh scattering. However, even though not much
improvement in either silica fiber type is expected, there
are a number of materials which potentially have losses in
the range .001-.01 dB/Km. These materials will be discussed
in the section of future projections.
Just as the losses in optical fibers continued to drop
through the decade of the seventies, bandwidths continued to
climb with the evolution in fiber designs. The highest
bandwidth multimode fiber mentioned to date has a bandwidth
of 9.7 Ghz Km at 1310 nm (4). However, it must be mentioned
that the peak bandwidth for this fiber was maintained only in
a small spectral range of about 1310 nm. It must also be
mentioned that production fibers have much lower bandwidths.
For example, a recent mass production run yielded an average
of 1.83 Ghz Km at 1300 nm (5).
M2012
ILLEGIB
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
In contrast to multimode fibers, which show sharply
peaked dispersion vs. wavelength characteristics, single mode
fibers are capable of maintaining low values of dispersion
over wide wavelength ranges. The best fiber so far maintains
less than 2 ps/Km-nm over the range 1280 - 1650 nm (6).
These numbers were achieved with a multiply-clad (depressed-
cladding, W, or index-well) fiber and indicate that: 1. The
practical limits to data transmission over optical fibers may
be as high as 100 Gbn/s over a 100 Km span (7); 2. We
eventually may be able to wavelength division multiplex (WDM)
a large number of channels into this low dispersion band; 3.
Material dispersion is so small in certain fibers that fast
systems can be built even with light emitting diodes (LED).
Because the loss minimum of silica fibers occurs at
about 1550 nm and the dispersion minimum occurs at about 1300
nm, a class of fibers with the dispersion minimum shifted to
the lowest loss region was developed. Best results reported
are 0.24 dB/Km at 1550 nm where the dispersion minimum also
occurs. For the sake of completeness, I will add that the
best values obtained at the shorter wavelength of 850 nm are
approximately 2.5 dB/Km and several hundred MHz Km (9).
Cabling adds about 0.2 dB/Km to the values recorded for raw
fiber.
In addition to their optical properties, fibers have
also been continually improved in terms of their mechanical
properties. The best strength thus far reported is 5 x 105
psi (10), which is comparable to steel strands of the same
7
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
size. Because of these increases in strength, the lengths of
fiber that can be continuously drawn has also been
increasing. A continuous fiber of 100 Km in length has been
drawn (11) .
2.2 Components
Besides the fiber, the other key components of a fiber
optic transmission system are the transmitter and the
receiver. Admittedly, other components such as connectors,
couplers, and multiplexers are important, but the fact
remains that a point-to-point system can be built with just a
transmitter, a fiber, a receiver, and a means of splicing
fibers. Consequently, the concurrent development of devices
compatible with the evolving fibers has been mandatory.
Sources, both LEDs and lasers, that are compatible with
optical fibers have experienced the same progression through
the wavelengths as have the fibers, themselves. The initial
devices were made of AlGaAs and had peak emission
wavelengths in the 800-900 nm range, corresponding well with
the first low loss window of silica fibers. For the 1.3
micron and 1.55 micron windows, devices made of InGaAsP,
which can be made to emit in the 1.1-1.65 micron range
depending on exact composition, are most often used.
Although some work is being done on the development of
high output and/or low threshold, AlGaAs injection laser
diodes (ILDs), the state-of-the-art is generally represented
by InGaAsP devices. Before going on to discuss the longwave
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
devices, however, the best performances from AlGaAs should be
summarized. Shortwave lasers with thresholds as low as 21.5
ma and output powers as
been demonstrated (2).
output of 200 mw CW is
Diode.
high as 50 MW in a
single mode
have
In fact, an AlGaAs laser with an
commercially available from Spectra
Reliability studies (12)
times for AlGaAs LEDs are greater
indicate that mean
than 106 hours while
made from the same material show mean life times ranging
6.7 x 104-5 x 105 hours. Reference 2, however, mentioned
AlGaAs ILDs had projected life times of 106 hours. The
difference between the two values cited may simply depend on
differences in specified operating conditions; generally, the
lower the power output and the operating temperature, the
longer the diodes will last.
The following information about InGaAsp lasers and light
emitting diodes is freely abstracted from references 2 and 5,
which are themselves review articles.
As with their shorter wavelength cousins, longwave LEDs
sometimes offer a cheaper, easier to use, and more reliable
alternative to longwave ILDs. Since LEDs normally couple on
50-100 microwatts into a 50 micron multimode core, they were
initially thought unsuitable for use with single mode fibers.
Recent studies, however, have shown that not necessarily to
be the case. By increasing the dopant density, the carrier
lifetime can be reduced to about 2 ns, thus increasing the
LED's speed; this increase in speed, unfortunately, comes at
the expense of output power. Modulation frequency can thus
life
ILDs
from
that
9
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
be extended to about 1 Ghz, but the useful frequency range is
considerably lower. Modulation frequency and coupled power
are not the only parameters limiting the applicability of
LEDs to single mode systems because spectral width is another
limiter. Surface emitting LEDs at 1300 nm have spectral
widths of 120 nm, while at 1600 nm the width is 140 nm. Used
at the wavelength of zero chromatic dispersion (1300 nm), a
surface emitting diode therefore has a maximum gain-bandwidth
product of 2.4 Ghz-Km. On the other hand, edge emitting LEDs
have spectral widths of only about 70 nm at 1300 nm. This
implies that their maximum gain-bandwidth product is
approximately 7 Ghz-Km; they also couple 2-3 times more power .
into fibers than surface emitters. Projected mean life times
for surface emitters are on the order of 109 hours (2) while
values for edge emitters are expected to be at least several
hundred thousand hours (12)
The best performance to date for an IGaAsP edge emitting
LED used in a single mode systems is 180 Mb/s for a distance
of 35 Km (13). Such performance indicates that there are
likely to be a number of potential applications for single
mode-LED systems in the future. Longwave LEDs are
commercially available from a number of Western vendors.
Multimode InGaAsP/InP lasers with peak wave lengths in
the 1300 nm range have been commercially available for
several years. More recently, lasers emitting at 15509 nm
and lasers emitting in a single mode have become available,
but they are very expensive. These single mode lasers emit a
10
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
single transverse, or spatial, mode but they do not emit in a
single longitudinal mode. This means that although they have
narrow beam widths and can be efficiently coupled to single
mode fibers, they do not emit a single wavelength. The
single frequency ILD is therefore the hottest topic in
emitter research.
As mentioned, single mode lasers for both short and long
wavelengths have been produced, but they both share certain
deficiencies. At high modulation rates, there is a tendency
for their spectral outputs to broaden and to hop around. In
such characteristics were also shared by the single frequency
lasers, they would completely destroy its single frequency
characteristic at high speed. The spectral spreading, also
called chirp, and hopping of the output would increase the
chromatic dispersion experienced by pulses, and limit the
ultimate data rate achievable with these lasers.
Consequently, single frequency lasers must maintain their
single frequency output not only when run in the CW mode but
also when pulsed rapidly. Such lasers
labs and are called dynamic single mode
Many types of dynamic single mode
exist in the research
lasers.
(DSM) lasers are being
developed, but they can be broadly grouped into four families
(14): coupled cavity, frequency selective feedback,
injection locked, and geometry controlled. All these
different lasers are designed to achieve the same goal,
selecting one frequency, and one frequency only, for lasing
action. Coupled cavity lasers work by allowing the light to
11
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
pass through an additional cavity. Thus only the frequency
that is resonant in both cavities is allowed to propagate.
The difficulties in tuning the frequencies of cavity lasers
was cause for the development of the frequency selective
lasers. Frequency selective lasers use either eternal
diffraction gratings or integral Bragg reflectors to 'select
the output wavelength. The internal Bragg reflectors are
formed directly under or above the laser cavity by integrated
optics techniques; this type of laser is also called a
distributed feedback laser (DFB). In injection locked
lasers, the diode laser's output is forced to contain a
single frequency by coupling it to another single frequency
laser such as a HeNe. The principal geometry controlled DSM
laser is the short cavity laser. The geometry of this laser
forces- adjacent modes to be spaced by about 20 angstroms
rather than nearly overlapping each other. High reflectivity
coatings or filters can then be used to enhance the
wavelength of interest. Hybrids of the various families are
also being investigated.
DSM lasers have been demonstrated that produce spectral
lines as narrow as one millionth of an angstrom (14) and that
have maintained narrow linewidths even when pulsed at several
Gb/s. Which type will eventually be most popular will
probably depend on which is most easily manufactured,
although manufacturers are leaning toward the DFB laser.
British Telecom Research Labs, AT&T Bell Labs and NTT Labs in
Japan have traded the world record for data rate-distance
12
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
numerous times as a means of displaying the superiority of
their designs. Commercial realization, however, remains at
least a year or two in the future.
Silicon is an ideal detector material for the short
wavelength fiber optic systems because its bandgap nearly
matches the systems' 800-900 nm operating range. Commercial
silicon PIN (positive intrinsic negative) diodes are
therefore highly developed and can be expected to provide the
following characteristics: quantum efficiencies above 90%,
response times below 1 ns, and dark currents about 1010 amps.
Likewise, silicon APDs (avalanche photodiodes) are
commercially available with the following parameters:
quantum efficiencies near 100% response times about 1 ns,
current amplification factors of about 100, noise factors of
about 5, and primary dark currents around 1011 amps (2).
The first detector used for the 1300 nm wavelength was
germanium, but devices constructed from silicon suffer from
larger dark currents and noise factors than do comparable
silicon devices. Germanium APDs also suffer from decreased
high frequency response at 1550 nm has recently described
(15). These problems with germanium are caused by its
physics and therefore cannot be solved by improving
manufacturing techniques; new materials must be found.
Before detailing the work being done on alternatives to
germanium, the best results reported for germanium should be
detailed. The best report found for a Ge APD described a
device with the following parameters: excess noise factor of
13
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
7 at a gain of 10, a risetime of 150 ps, a frequency response
of 2 GHz, a quantum efficiency of about .8, and a total dark
current of 1 microamp (2, 5). Ge APDs are commercially
available with somewhat poorer specifications.
Another device that has been commercially available for
several years is the InGaAsP/InP PIN diode with integral GaAs
FET (field effect transistor) amplifier. This device can be
used from several megahertz to several hundred megahertz, and
it theoretically will exhibit lower noise than a Ge APD at
all frequencies below 1 GHz. However, what is really needed
is a better APD; and that is the area in which detector
research is currently focussed.
Other than germanium, the most developed technologies
for longwave detection are InGaAsP and InGaAs. PIN diodes
made from these materials have been commercially available
for a number of years; APDs of two different types, planar
and mesa, have been demonstrated. The mesa is the more
developed of the two types, and the best device found
described in the literature had a noise factor of 5 at a gain
of 10 with a primary dark current of 3 nanoamps (2). Its
noise factor if already better than germanium APDs, but
further work will be needed before commercialization.
Other materials have been used to produce long
wavelength detectors, among these are HgCdTe, and AlGaSb or
AlGaAsSb/GaSb systems have already been demonstrated in both
PIN diodes and APDs. The APDs have displayed efficiencies in
the .9-.9 range, gains in the 40-100 range, and bandwidths
14
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
above 1 GHz. Unfortunately, their dark currents are high.
In HgCdTe, a photodiode with dark current of 1 nanoamp,
bandwidth of 850 MHz, and responsitivity of .8 at 1300
been demonstrated. However, commercial development of
nm has
any of
these detector systems seems to lie sometime in the future.
In addition to PIN diodes and avalanche photodiodes, two
other devices are being researched. InGaAs/InP has been used
to produce phototransistors (16), and there is hope that
these may in the future be developed into high speed devices
with the capability to measure very low light
levels. For
ultra speed detection, although only at short wavelengths so
far, the GaAs Schottky barrier
photodiode is being
developed. Thus far, a device with 18 GHz bandwidth has been
demonstrated (17).
2.3 Systems
,The pushers of the
communications have been
military organizations;
state-of-the-art in optical fiber
the world's phone companies, not its
and the phone companies have been
shattering the records so rapidly that it is hard to report
exactly what those records are. As an example, between June
of 1983 and September of 1984, seven records were set. The
players in this competition were: AT&T Bell Labs, Nippon
Telephone and Telegraph, and British Telecom Research Labs.
Although it is possible that the record has since changed
hands, as of early this year, Bell Labs had both ends of the
record (18). The record for bandwidth-distance was achieved
15
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
with a system that transmitted for 130 Km unrepeatered at a
data rate of 2 Gb/s. As for sheer unrepeatered distance, a
420 MHz signal was sent 203 Km. Performance curves of
typical fiber optic systems are shown in Figure 1.
Besides these "hero" experiments, much researdh is being
done on optical multiplexing and heterodyning. The reason
that the hero experiments are receiving so much attention is
that they push the major systems components to their
limitsexciting. To illustrate this point, one need only look
at the components used to produce these records. All seven
systems used 1500 nm light to take advantage of the fiber's
absorption minimum; fiber losses including splices average
only .25 dB/Km for all systems. All seven systems used
single frequency lasers; five used DFB lasers, .one used a
cleaved-coupled-cavity, and one used both injection locking
and distributed feedback. As for receivers, only one system
used a PIN diode while three others used GeAPDs. The other
three systems used a new APD invented by Bell Labs. Made of
InGaAsP, this device is called a separate absorption,
grading, and multiplication region (SAGM) APD., Sensitivity
for the SAGM APD is claimed to be 7 dB greater than a PIN-
FET, and 4 dB greater than a GeAPD (18).
16
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
-n
(D
?-?
8-6ZOOLZOOCOOMI,81,00EIZ6da-V10 CZ/170/C1.0z eseeiej J04 panaiddv Ado paz!l!ueS u! PeWsseloaCI
(sign) 3iyti 118
LI
DISTANCE (km)
1
? I
8-6ZOOLZOOCOOMI,81,00EIZ6da-V10 CZ/170/C1.0z eseeiej J04 panaiddv Ado paz!l!ueS u! PaisseloaCI'
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Section 3
(U) SOVIET TECHNOLOGY
3.1 Fibers
Although Soviet fiber optic technology does seem to lag
the West by several years, the gap is not large, and the
Soviets are thus able to demonstrate very capable systems.
This fact can be illustrated by tracing some of the evolution
of the Soviet fiber optic capability.
As long ago as 1977, the Russians had demonstrated the
ability to produce fibers with losses below 1 dB/Km (19); the
reported values for that fiber were: .7 dB/Km at 1600 nm, 1
dB/Km at 1Km at 1200 nm, and 5 dB/Km at 820 nm.
Additionally, it was clear that the Russians were well aware
of the military potential of optical fibers because the
authors of reference 19 proposed building systems in the 110-
1700 nm range. Their proposal was based not only on the
lower losses that they had achieved at the longer wavelengths
but also on referenced articles both Western and Russian,
indicating that fibers had greater radiation resistance at
wavelengths greater than 1 micron. Simultaneously with the
proposal to build systems at the longer wavelengths, a system
capable of transmitting data for .5 Km at 1540 nm was
demonstrated (20).
By 1978, the Soviets had produced their first
strengthened fiber optic cable (21). Reported losses for ten
fiber cable were 5 dB/Km at 1060 nm and 8 dB/Km at 900 nm.
18
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
By the following year, a load bearing cable was announced
(22). Although designed for oil well logging, the high
hydrostatic pressures and high loads that this cable had to
withstand indicate that with perhaps minor modifications, it
could have been adapted for use under the sea. Excess losses
in this cable were reported as 2 dB/Km above those of the raw
fibers.
In addition to their work on various types of cables,
the Russians have duplicated virtually every type of fiber
being produced by the West. In 1980, they announced their
first single mode fiber (23). It exhibited losses below 5
dB/Km over the range 870-1020 nm, and maintained 99% single
polarization for a kilometer. In that same year the Russians
also demonstrated a fairly low loss, silicone clad fiber
(PCS) with losses of 20 dB/Km at 980 nm (24). The developers
of that 200 micron fiber thought that the observed losses
were due to high losses in the silicone rather than the
silica, so they retested the fiber by exiciting only lower
order modes (modes less likely to encounter the core-cladding
interface). Losses consequently dropped to 8 dB/Km at both
860 and 980 nm. The authors therefore concluded that by
purifying their silicone, production runs of 10 dB/Km could
be achieved. Since one of the main attributes of PCS fiber
is its radiation resistance, it was not surprising to see
these authors also speaking of increased radiation resistance
at the longer wavelengths.
19
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Because the standard two layer, single mode fiber has a
very small core, it is difficult to produce interconnection
devices for this type of fiber. Consequently, much effort
has been expended in trying to produce fiber designs that
propagate a single mode yet have larger core sizes. Among
these fiber types are the already mentioned W fiber and the
ring fiber. Thus in 1981, the Russians introduced a ring
fiber (25) and a W fiber (26). The best results obtained
with the ring structure were 4.6 dB/Km at 1020 nm. On the
other hand, losses for the W fiber were 3.2 dB/Km at 1200 nm,
and single mode propagation was achieved with a 13.5 micron
core as opposed to the 8 micron core that would have
characterized an equivalent two layer fiber.
Another area of interest in both the East and the West
is the application of optical fibers to data communications.
Since this field often involves much shorter fiber runs that
does telecommunications, larger core fibers with lower
bandwidths are acceptable. In turn, larger core fibers mean
less expensive sources and connectors can be used. Thus, it
was in the Russians interest to develop larger core fibers,
And they have done so. A 1982 paper described an all glass
fiber, as opposed to PCS, with a 100 micron core (27). This
fiber had a graded index and showed both larger effective
numerical apperture and higher bandwidth than a comparably
sized PCS fiber. Its reported parameters were an NA of .4,
modal dispersion of 2.5 ns/Km, and a loss of 6 dB/Km at 850
nm. The material dispersion was also measured and found to
20
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
be approximately 100 ps/Km-nm at 850 nm. The dispersion
minima was shifted to the 1350-1400 nm range. Another 1982
paper (28) reviewed work on all polymer fibers, which have
some potential short run applications. Of interest in this
article were the facts that the Russians are paralleling
Western efforts in this area and that "Polymer fiber
waveguides subjected to gamma rays or neutrons in doses of 5
x 105 rad have been found to recover their properties after
50 msec".
The latest document pertaining to silica waveguides
should serve as an indicator of where the Russians stand with
respect to the West. That paper (29), dated 1984, reported
on a graded index, multimode fiber that was obviously
developed for long haul applications. The reported losses
were .25 dB/Km at 1550 nm, .50 dB/Km at 1300 nm, and 2.45
dB/Km at 850 nm. These values can be constrasted with the
theoretical limits predicted for the materials used, these
limits were: .20 dB/Km at 1550 nm, .40 dB/Km at 1300 nm and
2.43 dB/Km at 850 nm. The bandwidth was measured at 870 nm
and was 900 MHz-Km.
As previously mentioned, the Russians are just as
interested in the effects of ionizing radiation on optical
fiber systems as is the West. And although it is likely that
some papers were classified, a number of these articles were
published in the open literature (30-35). The last of these
references is a review article that draws on both Russian and
21
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
Western sources. One reference on the effect of radiation on
light emitting diodes was also found (36).
The final question that one might ask about current
Soviet capabilities in fiber optics is where do they get
their fibers? An auxiliary question is, "Can we deny them
access to optical fibers?". It has become apparent in the
courses of performing this search that the Soviets are
developing their own capabilities in parallel with the West.
It is true that they do seem to lag somewhat, but they also
seem to eventually match our accomplishments. Thus, there
appears to be no way to embargo anything tangible, and the
restriction of research results would most likely slow
development in the West. Besides, there is at least one
fiber company in Yugoslavia, ISKRA>
3.2 Components
In the development of semiconductor lasers, the West has
a lead over the Soviets, but the lead is by no means
enormous. In 1980, the Russians demonstrated InGaAsP/InP
lasers that could run CW at room temperature (37). These
lasers had threshold currents as low as 30 milliamps, and
output powers as high as 10 milliwatts in the 1240-1280 nm
rnage. Additionally, they were claimed to be kink-free and
to emit a single transverse mode when strongly pumped. More
reliable single mode operation at 1300 nm seems to have been
achieved with a design termed a three-layer waveguide laser
(38).
22
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
By 1982, the range of continuous wave (CW) operation had
been extended to the 1500-1600 nm range (39, 40). The output
powers for these lasers were 2-3 nm, and threshold currents
were as low as 23 mA. Some of the lasers were also claimed
to have single frequency outputs, although it is likely that
their outputs were only single frequency in the CW mode.
The Soviets have also been interested in the various
aspects of ILD behavior. Among the areas being investigated
are the following: laser to fiber coupling efficiency (41,
42), threshold current reduction (43, 44), and the
temperature sensitivity of InGaAsP lasers (45). Their
results in terms of coupling efficiency have been quite good
with claimed efficiencies of 90%, or 16 mW, into multimode
fiber and 36% into single mode fiber. In addition to these
studies of laser properties, the Soviets have also been
working on different laser materials. Lasers have been
fabricated from GaSb/GaAlAsSb (46) for use at 1600 nm. While
achievements with GaSb are not impressii.re when compared with
what has been achieved with InGaAsP lasers, the mention of
this research serves to illustrate the breadth of the Soviet
program.
The Russians have also been experimenting with the use
of optical feedback to produce single longitudinal mode
outputs and ultrashort pulses. A diffraction grating has
been coupled to a 1300 nm laser to produce a single frequency
output with up to 4 mW output power tunable over a 24 nm
range (47). However, this laser cannot be termed a dynamic
23
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
ID
M2012
single model laser because its frequencies above 20 MHz.
Optical feedback by an external mirror also has been used to
produce tunability in a short wavelength laser (48), but mode
hopping could not be suppressed with that design. In
addition to its use in producing single frequency outputs,
feedback by diffraction grating has been used to produce
pulse trains of very narrow pulses (49, 50). The best
results reported in these experiments were 2.7 ps pulses at a
repetition rate of 660 MHz.
The first laser displayed by the Soviets that appears to
be true dynamic single mode laser made its appearance in
1984. This laser apparently used a holographic diffraction
grating to select the frequency. It also used a second diode
laser as a modulator and travelling wave amplifier. The
results achieved included both modulation at up to 2 GHz and
reduction in the optical spectral width (51). With this
laser, the Russians can be said to have entered the fourth
gneeration of optoelectronic systems research. However, they
are not yet world class competitors.
Before addressing the Soviet capabilities in the field
of photodetedctors, it should be mentioned that they have
also been doing resesarch into light emitting diodes. As
long ago as 1978, they produced long wavelength LEDs made by
liquid phase epitaxy (52). These InGaAsP/InP LEDs were very
efficient, and yielded output power of 15 mW at 50 mA. They
would thus be suitable replacements for lasers in some
applications.
24
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
In the field of detector research, the Russians also
seem to lag. They seem to have mastered silicon because
their Si research is on such advanced topics as enhancing
spectral range and producing sub-nanosecond response times
(53-55). They are also working with germanium photodiodes
(56) and Schottky barrier diodes (57). In fact, they claim
to be the first to combine an integrated optic waveguide with
a Schottky barrier to produce a detector with enhanced
responsivitity. Several combinations of materials were tried
including gold with GaAsP on a GaP substrate and silver with
GaAsP/GaP. It was mentioned that Schottky barrier detectors
would be operable at microwave frequencies; the authors
estimated time constants of less than .5 ns.
Despite their activity in other areas of detector
development, the Soviets definitely seem to be lagging the
West in the area of longwave APDs constructed from alloys of
groups III and V of the Periodic Table (III-V). This
statement is based on analysis of a 1983 Russian review
article on the subject (58). In that review, Russian
references were used but only when referring to research on
III-V materials. When articles pertaining to the actual
construction of prototype APDs were used, they had all been
written by Western authors. Thus, it appears likely that
Soviet research into the subject dates back, not much more
than two years.
3 . 3 Systems
25
M2012
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
This section contains the essence of the entire study,
since it will attempt to answer whether or not the Russians
can deploy a 5 Mb/s fiber optic system.
The Soviets undoubtedly lag the West in fiber optics;
however, the gap is not as large as one would normally be
inclined to assume. Since they have access to our
literature, it would be naive to assume that they could not
at some point duplicate our achievements. On some
developments, the delay between attainment in the East and
the West seems to be four or five years, while other
developments are duplicated within a year or two. In fact,
the estimation of the exact lag between the two blocs could
in itself be the subject of a study. By correlating, on a
one for one basis, the development in the two blocs of the
principal milestones in the field, it should be possible to
achieve a good estimate of the exact lag between the blocs.
However, that is not the subject of this report; so we will
here concentrate on building evidence to support the
contention that the Soviets long ago achieved the capability
in question.
The Russians cannot necessarily match the West's
technology in elegance, but they can often achieve the same
results through brute force. In a 1977 study designed to
study the dispersive properties of fibers, the Soviets
displayed the ability to send signals modulated beetween 800
and 940 MHz through 25 meters of high loss (600-700 dB/Km)
fiber (59). The system consisted of a HeNe laser beam
26
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
modulated by an external modulator, the fiber was high loss,
and the detector was a photomultiplier coupled to a microwave
resonator; the setup was designed as a measurement system,
not a communications link, but it did demonstrate some
capability. That capability was displayed in the very same
journal issue as Reference 59. In an article about using
fiber optics to communicate between two computers, an eight
fiber link that could transmit 64 Mb/s was demonstrated (60).
Therefore, the actual optical signal on each channel was
transferring data at 8 Mb/s for 20 meters over high loss
fibers. Thus, by 1977 the Soviets had already exceed the
speed in which we are interested, so what remains to be seen
is whether they have improved their distance capability.
Later in 1977, the Russians improved the speed of their
systems by demonstrating two new systems: a 8.5 Mb/s system
designed to transmit 120 phone calls, and a 114 Mb/s system
designed to transmit either digital color TV or 1440
telephone circuits (61). However, the fibers used still had
losses of 500 dB/Km. Several months later the first system
to use low loss fibers was demonstrated (62). The data
transfer rate between two computers was only 11 KB/s, but
notable were the introduction of a mass producible LED and
fibers with 10-20 dB/Km losses, both Russian designs. The
transmission distance was .75 Km.
1978 saw the introduction of 1300 nm systems in Russia.
A 2 Km long telecommunications line was demonstrated, and the
stated system margins indicated that the system could
27
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
actually operate in the 10-20 Km range (63). The data rate
was 10 Mb/s, and specifically stated as an advantage of 1300
nm transmission in addition to the greater transmission
distances was the improved radiation resistance. This is not
normally of concern in civilian telecommunications.
The following year, radiation resistance was again
mentioned as an impetus in the development of a 1300 nm
system demonstrating wavelength division multiplexing (64).
Although only .5 Km long and unidirectional, this system was
notable for several reasons. It demonstrated the use of
fiber optic couplers and claimed single mode laser emission
was made possible by proper control of the pumping current.
With this control of the spectral output, the researchers
were able to multiplex two signals separated by only 10 nm
and still maintain crosstalk to less than -20 dB. However,
the fact that spectral output of standard ILDs cannot always
be controlled by controlling the pump current was illustrated
in a 1982 paper by the same authors, when they extended their
1300 nm WDM system to 5 Km and introduced duplex transmission
(65). Two signals were multiplexed in each direction
(although the system was claimed to be capable of handling 10
channels in each direction) with a separation of 17 ns
between channels. However, the lasers could not produce
single frequency outputs so their spectral widths had to be
narrowed t4o 8 nm by filtering, not by current control. Also
interesting was the fact that the couplers used had rather
high losses, but they were serviceable and they were Russian
28
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
manufactured. The system's pass band was 5 MHz for its 5 Km
length mostly because it was intended as a demonstration of
multiplexing capability not data transfer capability. The
researchers had used step index fiber, but they noted that
the pass band could be greatly increased simply by installing
graded index fiber.
The authors of reference 65 made an interesting point
about the need for single frequency lasers. It has been
mentioned that fibers are being developed that will allow the
wavelength division multiplexing of hundreds or thousand of
channels. Similarly, WDM components have been developed and
are commercially available for this purpose. However,
systems demonstrating WDM have invariably transmitted
relatively few channels. This is because the close spacing
of channels requires sources that are both spectrally narrow
and drift-free; DSM lasers are needed.
The longest system found in the Soviet literature was an
8 Km system with a pass band of 6.5 MHz (66). This must have
been a short wavelength system because the source used was
A1GaAs laser, and 8 Km at that wavelength is not bad. As
with many of the other Russian systems found in the
literature, the purpose of this system was not to display the
system's capabilities but rather to experiment with mode
coupling in step index fibers. All the Russian papers seem
to concentrate on demonstrating one achievement, be it 1300
nm transmission, high speed, WDM, or long distance. Not one
paper followed the Western penchant for pushing a system to
29
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
its limit in bandwidth-distance product. Perhaps the
reporting of research into such areas is prescribed for
military reasons, perhaps the authors simply want to avoid
unfavorable comparisons with Western results, or perhaps the
proper literature was simply not discovered. However, the
factors remain that the Soviets have displayed high speed
short distance systems, low speed long distance systems, low
loss-high bandwidth fibers, and high speed sources and
detectors. It therefore seems fairly reasonable to assume
that there is nOthing preventing the assembly of the various
components into a high speed long distance system. To
further support this statement, it is only necessary to point
out that the Poles, who because of their political situation
do not share the latest in Soviet technology, have
independently demonstrated an 8 Mb/s fiber optic system and
the ability to produce their own fiber (67).
EMP, as a propagated signal, can damage equipment or
cause RF interference. In the case of fiber optics
transmission systems, the concern is more with damaged
equipment. A surface burst generates an EMP due to the rapid
ionizing radiation effects. The EMP can be divided into two
regions: a source region which is associated with the air
volume and ground current returns where the EMP is generated,
and the radiated region. Detonations ranging from about 70
miles altitude to several hundred miles altitude give rise to
high altitude EMP. This is the dominant EMP survivability
concern for most systems. The source of the HEMP is
30
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
ionization at the top of the atmosphere cause by X-Rays and
GAMMA rays moving downward from the nuclear burst points.
Residual GAMMA radiation can travel over 2000 miles.
Within the source region, which extends to 3-5
kilometers from the source, electric fields may exceed 100
kV/meter. The EMP energy coupled to a system must be treated
as a survivability issue because permanent damage may result.
Beyond this 3 to 5 kilometer distance, the radiated fields
are less intense, and less than high altitude EMP fields.
The latter field strength is generally specified at 50
kV/meter as a system survival requirement. While fiber optics
are attractive for use in High Altitude EMP (HEMP) mitigation
programs, any such application must include protection
against radiation damage.
As mentioned before, degradation of fiber optics systems
due to nuclear radiation is of major concern. Nuclear
radiation could cause damage to the transmitters, receivers,
and repeaters of fiber optic systems. Prompt radiation and
fallout could cause darkening of the fiber optic cable, which
might result in loss of transmission. The darkening of the
fiber optic cable is caused by the formation of "charge
traps" that absorb and scatter the incident light. This
produces an induced attenuation in the fibers which
progressively decreases after the source of radiation is
removed. The state of the cable may or may not return to the
original state in the absence of the radiated field.
Radiation experiments have lead to two interesting
31
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
observations. First, a correlation seems to exist between
the level of radiation-induced loss and the intrinsic
material loss. Fibers with higher loss have a higher
radiation sensitivity, which probably is due to the higher
impurity concentration. Second, after a certain dose,
polymer-coated fused silica fibers become extremely
saturated; other fibers tend not to saturate and show a
fairly linear behavior. The magnitude of the induced loss
effect at a given time depends on the nature of the
radiation, its total level and rate of application, the
wavelength of observation, the composition of the glass, the
temperature and, in some cases, the previous history of the
material and the optical power fed to the fiber.
Reference 100 defines the level of increased earth
potential as a function of the magnetic disturbance (which
could be created from EMP) and the type of earth structure.
Typically, a fiber optics link connects two sets of terminal
equipment together via wideband fiber optic cable through
wideband repeater amplifiers. A secondary copper line is run
in parallel with the fiber optic line to provide monitoring
control of the repeaters. There are two possible failure
modes that can occur from this increased potential:
a. The power supplies for the repeaters (or
regenerators) distributed along the line
can fail and/or;
b. The monitor and control units that regulate
the signal amplitude along the link can fail.
32
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Either or both failures noted above would increase the
susceptibility of the fiber optic link to EMP.
33
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Section 4
(U) PROJECTIONS-WESTERN
4.1 Fibers
It has already been mentioned that high silica fibers
have already attained their theoretically predicted minimum
losses. Further improvements, therefore, can only be
expected in entirely different glass systems. A number of
materials are potential candidates for the optical fibers of .
the future. Among these candidates are certain crystalline
materials, such as KRS-5, and several families of glasses:
heavy metal fluorides (HMF), heavy metal oxides, and
chalcogenides. In the past five years, much work in both the
East and the West has been done on these materials because
they promise: transparency from the near UV (.2-.3 microns)
to the mid-IR (7-8 microns), losses a factor of 10 to 100
lower than the best silica, and less susceptibility
ionizing radiation in the mid-IR (68).
Little work, either theoretical or experimental, has
been done on the heavy metal oxide glasses with the exception
of germanium oxide. There also seems to be more research
being performed on glasses than is being performed on
crystals, even though there are indications that certain
crystals, KC1 for example, may ultimately exhibit the lowest
losses of all the materials being considered (69). On the
other hand, the chalcogenide and HMF glasses are being widely
studied.
to
34
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Perhaps one of the reasons that glass research has been
given such impetus was a paper published in 1981 that
predicted ultimate losses of .001 dB/km for certain HMF
glasses and .01 dB/km for chalcogenide glasses (70). If
losses this low can actually be achieved, it may be possible
to span oceans without using repeaters. Another reason for
the interest in mid-IR fibers was another 1981 paper that
reported on the effects of gamma rays on bulk samples of a
HMF glass. The researchers found increased losses at both
the high and low ends of the spectral range, but found no
incremental losses in the range 2.5-4 microns (71). No
effects were recorded even of doses of 45 Mrad, and the
region where there appears to be no radiation damage overlaps
the region of the lowest loss; these glasses therefore show
promise for military applications.
One of the reasons that the authors of reference 70
predicted such low losses for the HMF and chalcogenide
glasses was that Rayleigh scattering was viewed as one of the
most important loss mechanisms in low loss fibers; and
because Rayleigh scattering is predicted to be lower in the
low melting point glasses, such as HMF and chalcogenide.
However, when the initial predictions were made, losses in
the 300-500 dB/km range were the state-of-the-art. Since
that time, losses have dropped considerably; and the
projections, at least for chalcogenide glasses, may have
changed. The latest paper published by some of the same
people who authored reference 70, reports on the development
35
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
of three different chalcogenide glasses (72). The best
results reported in reference 72 were for a glass made of 40%
As and 60% S. Lowest losses for this glass were 35 dB/km at
2.44 microns. While both reporting improved performance and
reiterating that chalcogenides are less prone to
crystallization than HMF glasses, the authors nevertheless
also mentioned a "weak absorption tail". This tail is
evidently caused by a loss mechanism not originally included
in the estimates of the ultimate losses of chalcogenides. It
now appears that chalcogenide glasses may never exhibit
losses below about 10 dB/km. If that is the case, the only
application for these glasses may be the short distance
transmission of energy from a CO2 laser.
As already mentioned, the only heavy metal oxide given
extensive notice appears to be germanium oxide. The latest
results for this glass are 4 dB/km at 2.0 microns and 15
dB/km at 2.4 microns (73). The low losses thus far achieved
are due in part to the ability to apply techniques developed
for silica fibers directly to Ge fibers. Unfortunately, this
similarity with silica extends to the Rayleigh scattering
coefficient of germanium oxide fibers; these fibers are
likely to have ultimate losses very nearly equal to those of
high silicate fibers. If that turns out to be the case, then
Ge based fibers will face two difficulties when compared with
Si based fibers. First, germanium oxide does not transmit
very far into the IR (only to about 2.6 microns), and second,
Ge is expensive when compared to Si. The authors of
36
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
reference 73 therefore suggest that one of the realistic uses
for germanium oxide fibers might be the construction of
optical signal amplifiers for ultra long distance
communications systems.
The glasses with the best chance of exceeding the
performance of silica fibers is therefore the heavy metal
fluorides, and work in this area is proceeding rapidly.
Leaders in this field are the United States, France, and
Japan.
There are two distinct categories of glass in the HMF
family. Most development efforts seem to be concentrating on
one of these categories, the glasses based on either
fluorozirconate or fluorohafnate. An excellent review
article (74) has summarized the achievements to date.
Drawing freely from that paper, the following summary of
properties can be offered.
The best losses reported are 8.5 dB/km at 2.1 micron and
6.8 dB/km at 2.55 micron. Material dispersion zero is in the
range 1.6-1.7 microns while the lowest losses are expected to
be in the range 3-4 microns. However, waveguide dispersion
can be used to offset the material dispersion and shift the
zero toward the region of lowest loss. Doping with chlorine
or bromine can achieve the same goal. Additionally, the
material dispersion is low (several ps/km-nm) over a wide
wavelength range, so even without offset from waveguide
dispersion, the overall dispersion will be low over a wide
37
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
range. Coupled with the wide spectral transparency, this
means that WDM of very many channels may be possible.
The strengths of HMF fibers are currently running about
1/3 that of silica fibers, but their theoretical limits are
closer to those of silica. Strengths reported to date are in
the range 10 E4-10 E5 lb./sq-in., but theoretical limits may
be 5x10 E5 lb./sq-in. The practical limits may be in the
range of 3x10 E5 lb./sq-in.
The ultimate fracture toughness of HMF glasses is
expected to be 1/2 to 2/3 that of silica, but silica's values
are high. Therefore, the attainment of a sizable fraction of
silica's toughness will yield acceptable fiber strengths.
One of the main barriers to attainment of high strengths in
fluoride glasses is their susceptibility to attack by water.
In fact, there has been some talk about the possibility of
having to hermetically seal HMF fibers. This may not be
necessary, however, because Corning has just announced a
newly developed HMF with an order of magnitude less
solubility in water. So far this glass has only been
prepared in bulk form and has not been drawn into fiber, but
future prospects are good.
It therefore seems that within the next ten years some
form of mid-IR fiber will be in use. At this point in time,
it appears that the most likely candidates for the production
of these developed fibers are the HMF glasses. These glasses
have already been developed at a more rapid rate than were
silica fibers. And considering that silica fibers went from
38
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
a breakthrough -20 dB/km to their theoretical limits in ten
years or less, low loss, mid-IR fibers
are likely to appear
by about 1990 unless fundamental problems hinder development.
4.2 Components
Articles written just
maximum rate at which a
a few years ago claimed that the
laser diode could be directly
modulated was about 2 GHz, yet this limit has already been
exceeded. By using such
schemes as combining high drive
currents with low modulation
modulated at greater than 20
depths, lasers are now being
GHz in research laboratories.
In fact, lasers are commercially available with bandwidths of
6 GHz (from Ortel in California). Over the next several
years, we should therefore see higher
become widely available.
and
Similarly,
photodetectors with bandwidths near 20
demonstrated in the laboratory. . These
prove commercially viable.
One thing that was noticed was that the older papers
listed lower values of the theoretical limit for system
bandwidth-distance pi.oduct than did the newer papers, and the
limits have invariably been exceeded. The author would
therefore compound these errors by stating speed limits which
devices will never exceed. The fibers themselves may set the
ultimate limits.
However, to push the fibers' data handling capacity to
the limits, the most sensitive detectors must be used with
higher speed lasers
Schottky barrier
GHz have also been
too may eventually
39
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
lasers that emit pulses that will suffer neither modal nor
chromatic dispersion. Thus, the devices most necessary for
the attainment of fiber limited bandwidth-distance products
are single frequency lasers and long wavelength APDs. And,
as has already been mentioned, both devices are being hastily
pursued in the world's top laboratories. Some shakeout in
terms of device designs should occur. in the next several
years with commercialization of the devices following shortly
afterward. With the widespread availability of such devices,
heterodyne detection may become a reality. Figure 2 shows
performance characteristics of long wavelength and single
frequency devices.
40
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
8-6ZOOLZOOCOOM1-81-00EIZ6dCl-V10 CZ/170/C I-0z eseeiej -104 panaiddv Ado paz!l!ueS u! PeWsseloaCI
(uni) ONIOVdS H3J.V3d311
0
Ul
BIT RATE (Gb/s)
../
....
iv w 4
/
/
---
-
... .0.
...I
....0- ..... .....
1
lk
i
..
a.. ... sm.
- , ?
....
/
I
I
I
I
I
I
11
I
PAS.121
li
4
:
; ;
OO:
:1'1:
II
11
S I
:41G
Will
,
ZIOZW
8-6ZOOLZOOCOOM1-81-00EIZ6dCl-V10 CZ/170/C I-0z eseeiej -104 panaiddv Ado paz!PeS u! PeWsseloaCI
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
4.3 Systems
By combining the components and the fibers previously
discussed with some of the devices mentioned only in passing,
such as WDMs and integrated optics, future systems will have
much more capacity than is currently exhibited. In the near-
term (2-5 years), we can expect to see higher power and
smaller spectral widths from single frequency lasers, lower
noise from long wavelength APDs, the introduction of some
integrated optical devices, and coherent transmission and
detection. In the longer term, ultra low loss mid-IR fibers
may be developed as well as purely optical repeaters;
research on these items is currently in a more preliminary
stage.
Of the topics just mentioned, integrated optics and
coherent transmission show great promise in the near-term.
Therefore, accomplishments to date in those areas will be
briefly summarized in the following paragraphs.
A number of discrete components such as couplers,
modulators, and switches have already been demonstrated in
integrated form. Through the use of integrated circuit
techniques, waveguides and other structures can be formed
from optically active materials. The parameters of the
structures can then be varied by varying some external
parameter, usually an electrical field, thereby controlling
the amplitude, phase, or direction of a light beam traversing
the structure. To date, the main achievement of this
technology is the miniaturization of fiber components; but
42
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2 0 12
eventually whole systems comprised of emitters and/or
detectors, modulators, directional couplers, and optical
switches may be formed on a single substrate.
The other technology al-ready showing considerable
promise is that of coherent transmission. There are two main
advantages to this technology: first, a 10 to 20 dB
improvement over direct detection is possible, and second,
the 50,000 GHz of bandwidth between 1250 nm and 1600 nm can
be effectively utilized. Coherent systems can employ either
homodyne detection or heterodyne detection. In both systems,
the incoming signal is mixed with a local oscillator (i.e.,
laser) thereby improving the signal to noise ratio. In a
heterodyne system, the local oscillator is not at the same
frequency as the signal, while in a homodyne system, the
frequencies of local oscillator and signal are the same. The
homodyne system is harder to implement, but it does enjoy a 3
dB advantage over heterodyning.
The performance of a coherent system depends on the
modulation scheme, be it phase (PSK), amplitude (ASK),
or
frequency shift keying (FSK). PSK provides the best
performance, but imposes the most stringent requirements on
the components. The local oscillator's output must be
polarization matched to the incoming signal, which means that
either the more expensive polarization maintaining fibers
must be used or that the oscillator's polarization must be
adjusted by an integrated optical device. Also, the laser
linewidth must be very narrow. The typical multimode laser
43
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
has an output about 3 nm or 500 GHz wide, far too broad for
coherent transmission. Even distributed feedback lasers with
linewidths of tens of MHz can only be used for ASk and FSK
transmission. Even so, the best coherent results are
comparable with the best results for direct detection. For
PSK, injection locked lasers with linewidths of 10 KHz must
be used to generate the light. The signal must then be phase
modulated by an external modulator, usually an integrated
optical device made from the electro-optically active
material lithium niobate.
The best results in heterodyning to date were just
announced several months ago in Venice (ECOC/I00C '85).
There, Bell Labs talked about two coherent systems, 400 Mb/s
over 150 km and 1 Gb/s over 148 km. Also presented at that
conference were results from Japan. NTT announced that they
had achieved 400 Mb/s over 250 km, while NEC talked of
transmitting 140 Mb/s over 234 km.
44
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Section 5
(U) PROJECTIONS - SOVIET
5.1 Fibers
The Soviets are also pursuing research on mid-IR fibers,
and for the same reasons that the West is pursuing such
research. In fact, one of the Rusqian papers mentioned
something that should have been obvious, but that did not
seem to be covered in the Western literature surveyed.
Another reason for wanting to shift into the mid-IR is that
single mode fibers for that region could have core sizes of
several tens of microns, thereby reducing the difficulty of
coupling fibers together. The fact that this was mentioned
in a Russian article may be significant, because it may
indicate a deficiency in their manufacturing capabilities.
Producing low loss connectors for single mode fibers is
perhaps one of the most demanding manufacturing jobs
currently on the horizon. It is difficult in the West, and
perhaps more difficult in the Eastern Bloc.
As mentioned earlier, the materials presently showing
the most potential for development into ultra low loss
optical fibers are the HMF glasses, and the Russians are not
among the leaders in this field of research. They are,
however, pursuing research in low loss crystals, and have
been doing so since at least 1980. A paper (75) written in
that year reported that losses of 13-21 dB/km had been
45
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
achieved in bulk samples of polycrystalline TIC1, KRS-5, and
KRS-6.
The Russians have tried to extrude fibers from the
crystals mentioned and from T1Br, but have been unable to
maintain the low losses achieved in bulk samples. A 1981
paper (76) reported on the results obtained with
polycrystalline fibers and found that the best losses that
could be attained were about 1300 dB/km at 5-6 microns and
10.6 microns. Since these wavelengths correspond to the
outputs of carbon monoxide and carbon dioxide lasers
respectively, the researchers tested the fibers' power
handling capability and found no damage at 1 kw/sq. cm. They
also calculated the material dispersion for KRS-5 and KRS-6
and report zero dispersion at 6.5 and 5 microns,
respectively. Material dispersion was found to be low (below
10 ps/km-nm) over wide spectral ranges for both crystals,
making the materials attractive for communications if lower
losses can be achieved. For comparison, at the time the
Russians were reporting 1300 dB/km, they were referencing
Western sources that had achieved 430 dB/km 2-3 years
earlier.
In addition to extruding polycrystalline fibers, the
Soviets have also tried to grow single crystal fibers from a
melt (77). Results were obtained for the following
materials: T1C1, T1Br, KRS-5, KRS-6, AgC1, AgBr, KBr, and
CsI. Lowest losses were achieved with T1C1, and were 3000
dB/km at 10.6 microns. This was about an order of magnitude
46
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
above the best Western results at the time and two orders of
magnitude above the values recorded for bulk materials. Both
pieces of information indicate that the researchers are
having trouble with their fiber production process.
The Soviets also are active in research into
chalcogenide glasses (78-81), and, in fact, claim to have
originated the field of study in 1954. The materials that
they are using are mostly sulfides and selenides of arsenic.
They are making progress with these materials, as witnessed
by the fact that during a one and half year period between
1981 and 1983 reported losses dropped from about 8000 dB/km
(78) to 580 dB/km (80). The authors of reference 80 also
calculated the material dispersion from their glass'
measurable properties, and found a zero at 4.89 microns and
values, not exceeding 20 ps/km-nm over a wide spectral range.
In addition to the fact that these materials are
transparent over the range of 1-11 microns, the Soviets have
reported other properties that may be extremely useful.
The Soviets report discovering a photoinduced optical
absorption in the fibers that they were testing (79). The
absorption of the fibers could be increased by exposing them
to light from a He-Ne laser, but they were not affected by
higher power densities of longer wave radiation. The
reseachers therefore suggested several possible applications
for this phenomenon. They suggested that fibers could be
graded with complex profiles simply by exposing them to
properly shaped laser beams; this grading would only be good
47
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
for short distances, however, because 630 nm light does not
propagate very far in these fibers. The technique thus might
be applicable to the grading of integrated optical devices.
Another suggestion from the researchers. was that He-Ne laser
beams could be imaged on the beveled ends of fibers to create
phase diffraction gratings. If the fibers are exposed to 630
nm light long enough, a permanent structural transformation
seems to occur. This change, however, is apparently
permanent only at normal temperatures. The annealing of the .
fibers by heating almost to the softening point apparently
erases the induced changes. Further work on this phenomenon
(81) has led to the modulation of a beam at 1150 nm by a beam
at 630 nm. In this case, the shorter wavelength light was
pulsed into the fibers, and the changes apparently were thus
not permanent. Since the absorption of 630 nm light was a
factor of 100 times higher than the absorption of the 1150 nm
light, the modulating beam did not propagate to the detector.
The Soviets are apparently making good progress in the
development of low loss crystals and chalcogenide glasses,
and even appear to be making some initial discoveries.
Overall, however, their efforts in the development of mid-IR
glasses appear to lag the West's by several years.
5.2 Components
It appears that in the future the Soviets will continue
to duplicate Western achievements in the area of component
development, but with a several year lag. They are already
48
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
working on fourth generation optoelectronic devices such as
single frequency lasers. Also, they have probably also begun
research on type III-V avalanche photodiodes by now; but in
that technology they undoubtedly lag the West by quite a few
years. Without these components, they can forget about
producing such advanced systems as those employing heterodyne
detection or multiple channel multiplexing.
Because of time constraints, the author was unable to
cover a number of topics. The Soviets are actually pursuing
research in such advanced subjects as integrated optics and
optical amplification, and they can be expected to match our
accomplishments in these areas in the future. Therefore,
several references on these subjects are offered for the sake
of completeness.
We have already mentioned optical amplification with
respect to some of the Soviet single frequency laser
research, but for the interested reader, reference 82 is
offered. Integrated optical devices are covered in
references 83 through 87. Multiplexers and couplers are
covered in references 88 through 91. Optically bistable
devices (i.e., the potential building blocks of optical
computers) are covered in references 92, 93 and 94.
5.3 Systems
If the Soviets can eventually duplicate our efforts in
terms of fiber and component development, it seems clear that
they can eventually match our systems capability. Thus, we
49
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
can expect that as our systems become faster and more
efficient, their's will also. Since Western research papers
almost never reference Soviet sources and since Russian
papers heavily reference Western sources, it must be admitted
that they seem to feed off of our technology. However, since
the Soviets are acknowledged to have an extensive
intelligence network in the West, censorship of Western
research might have a greater slowing effect on Western
progress than it does on Eastern progress.
As has been mentioned repeatedly, this report details
the capabilities the Soviets have thus far demonstrated, it
does not say anything about what the Soviets can reliably
manufacture. It may be entirely true that the Soviets
possess a number of talented individuals, who are
concentrated in a very few research institutes; and it may
also be true that the Russian manufacturing capabilities lag
several technological generations behind these scientists.
But it must be emphasized that at least some Russians are
working on fourth generation fiber systems, while a 5 Mb/s
system lies well within the capabilities of first generation
fiber optics. Thus, it is quite likely that they can deploy
a 5 Mb/s system.
5 0
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2 0 12
Section 6
(U) LOCAL AREA, LONG DISTANCE F.O. SYSTEMS
The use of fibers are much more advanced in Local Area
Networks and User Networks than in long distance carrier
networks because of the ease of installation and the
bandwidths they provide. Fibers are being used for building,
campus and public subscriber networks [95,96]. Existing
western technology is sufficient to permit the use of fibers
and related digital modems and multiplexers to be used for
present users' requirements. Even in the bloc countries
fiber is being employed in these user networks, but advanced
networks are limited by the lack of adequate microchips for
the modems and multiplexers.
Whether or not fiber optics is used in long distance
carrier networks is primarily a cost issue. If there is an
installed base of other types of media such as radio or
coaxial cable, then the tendency is to upgrade those first
before installing fibers. Within the United States there is
a major move to install fibers. The developing countries are
using more fiber than radio because they do not have a basic
investment in many major toll systems. Radio systems are
favored in some countries that have remote or mountainous
regions due to the problems of maintaining long cable or
fiber routes. However, the costs of installing fiber is
dropping and by the early 1990s most all new toll
transmission systems will be fiber, with radio and satellite
51
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
systems confined to specialized markets [97,98]. There are
presently few, if any, new long distance cable systems being
installed except in the bloc countries. In the bloc
countries new long distance fiber systems will likely start
appearing by 1990 as they develop the capacity to produce the
high data rate electronics required by the fiber systems.
[99].
52
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
Section 7
(U) REFERENCES
1. M. Schwartz, Optical Fiber Transmission-From Conception
to Prominence in 20 Years, IEEE Communications Magazine, Vol.
22, pp. 38-48, May 1984
2. T. Li, Advances in Optical Fiber Communications: An
Historical Perspective, IEEE J. Sel. Area Comm., VolInvited.
SAC-1, pp. 356-372, April 1983.
3. T. Miya et al, An Ultimately Low-Loss Single Mode Fiber
at 1.55 Microns, Electron. Lett., Vol. 15, pp. 106-108, Feb.
1979.
4. H. Unger, Trends in Optical Communications, Presented at
the International Conference on Communications, Amsterdam,
Netherlands, May 14-178, 1984.
5. Y. Suematsu, Long Wavelength Optical Fiber
Communications, Proc. IEEE, Vol. 71, pp. 692-721, June 1983.
6. L. Cohen, W. Mammel, S. Jang, Low-Loss Quadruple-Clad
Single Mode Lightguides and Dispersion Below 2 ps/Km nm over
the 1.28-1.65 micron Wavelength Range, Electron. Lett.,
Vol., 18, pp. 10-23-1024, November 1982.
7. D. Marcuse and C. Lin, Low Dispersion Single Mode Fiber
Transmission - The Question of Practical vs. Theoretical
Maximum Transmission Bandwidth, IEEE J. Quantum Electron.,
Vol. QE-17, pp. 869-877, June 1981.
8. B. Ainslike et al, Monomode Fibre with Ultra-Low Loss and
Minimum Dispersion on 1.55 microns, Electron. Lett., Vol.
18, pp. 824-844, September 1982.
9. Special Issue on Optical Fiber Communications, Proc.
IEEE< Vol. 68, October 1980.
10. S. Nagel, Recent Advances in the MCVD Process Rate and
Fiber Performance, Invitged Paper, Fourth Int. Conf. on
Integrated Optics and Fiber Optic Communication, Tokyo,
Japan, June 27-30, 1983.
11. M. Kawachi et al, 100 Km Single Mode VAD fibers,
Electron. Lett., 1983.
12. H. Yonezu et al, Reliability of Light Emitters and
Detectors for Optical Fiber Communication Systems, IEEE J.
Sel. Area Comm., Vol. SAC-1, pp. 508-514, April 1983.
53
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
13. AT&T LED Sets Distance/Bit-Rate Record, Photonics
Spectra, Vol. 19, pp. 8, October 1985.
14. T. Bell, Single Frequency Semiconductor Lasers, IEEE
Spectrum, pp. 38-45, December 1983.
15. M. Niwa et al, High Sensitivity Hi-Low Germanium
Avalanche Photodiode for 1.5 micron Wavelength Optical
Communication, Electron Lett., Vol. 20, pp. 552-553, June
1984.
16. J. Campbell et al, Avalanche InP/InGaAs Heterojunction
Phototransitor, IEEE J. Quantum Electron., Vol. QE-19, pp.
1134-1138, June 1983.
17. H. Blauvelt et al, Fabrication and Characterization of
GaAs Schottky Barrier Photodetectors for Microwave Fiber
Optic Links, Apopl, Plys. Lett., Vol. 45, pp. 195-196, 1
August 1984.
18. T. Bell, Technology '85 Communications, IEEE Spectrum,
pp. 56-59, January 1985.
19. A Belov et al, Glass-Fiber Waveguide with Losses Below 1
dB/Km, Soy. J. Quantum Electron., Vol. 7, pp. 1170-1172,
September 1977.
20. E. Godik et al, Investigation of the Possibility of
Using Phosposilicate Fiber Waveguides in the Near Infrared,
Soy. J. Quantum Electron., Vol., 7, pp. 1169-1170, September
1977.
21. A. Belov et al, Low-Loss Fiber-Optical Cable, Soy. J.
Quantum Electron., Vol. 8, pp. 414-415, March 1978.
22. Kh. Alimov et al, Load-Bearing Optical Cable, Soy. J.
Quantum Electron., Vol. 9, pp. 1580-1581, December 1979.
23. A. Gur'yanov et al, Single-Mode Low-Loss Fiber
Waveguide, Soy. J. Quantum Electron., Vol. 10, pp. 1052-1053,
August 1980.
24. A. Boganov et al, Glass Fiber Waveguide Made of
Anhydrous Quarts Glass with a Reflecting Silicone-Rubber
Cladding, Soy. J. Quantum Electron., Vol. 11, pp. 101-102,
January 1981.
25. M. Bubnov et al, Three Layer Optical. Waveguide of the
Ring Type, Soy. J. Quantum Electron., Voil. 11, pp. 204-206,
February 1981.
26. A. Andreev et al, Single Mode Low-Loss W-Type Fiber
Waveguide, Soy. J. Quantum Electron., Vol. 11, pp. 782-783,
June 1981.
54
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
27. V. Grigor'yants et al, Large-Aperature Fiber Waveguides,
Soy. J. Quantum Electron., Vol. 12, pp. 939-941, July 1982.
28. V. Gagulov, Polymer Fiber Waveguides, Soy. J. Quantum
Electron., Vol. 12, pp. 1587-1592, December 1982.
29. A. Belov et al, Graded Fiber Waveguide with Extremely
Low Losses, Soy. J. Quantum Electron., Vol. 14, pp. 440-441,
April 1984.
30. A. Boganov, Hyroxyl-Free Quarts Glass for Low-Loss Fiber
Optical Waveguides and its Comparative Radiation-Optical
Properties, Soy. J. Quantum Electron., Vol. 7, pp. 558-562,
May 1977.
31. A. Andreev et al, Influence of Gamma Irradiation on the
Temperature Dependence of the Optical Losses in Quartz-
Polymer Fiber Waveguides, Soy. J. Quantum Electron., Vol. 11,
pp. 1095-1096, August 1981.
32. Yu. Larin et al, Investigation of the Distribution of
Color Centers Along Gamma Irradiated Fiber Waveguides, Soy.
J. Quantum Electron., Vol. 12, pp. 360-362, March 1982.
33. E. Dianov, Pulsed Optical Bleaching of Fiber-Optic
Waveguides with a Pure Quartz Glass Core, Soy. J. Quantum
Electron., Vol. 12, pp. 500-504, April 1982.
34. G. Kosinov et al, Investigation of the Radiation
Strength of an Optical Cable Irradiated by 8 meV Electrons,
Soy. J. Electron., Vol. 13, pp. 638-639, May 1983.
35. E. Dianov et al, Radiation-Optical Properties of Quartz
Glass Fiber-Optic Waveguides (Review), Soy. J. Quantum
Electron., Vol. 13, pp. 274-289, March 1983.
36. A. Ptashchenko, V. Suskov and V. Irkha, Characteristics
of Radiation-Induced Degradation of Light-Emitting Diodes
Exhibiting Absorption and Reemission of Luminescence Photons,
Soy. Phys, Semicond., Vol. 15, pp. 1338, November 1981.
37. V. Bezotosnyi et al, Buried Mesastripe CW Room-
Temperature GaInPAs-InP Heterojunction Lasers in the 1.24-
1.28 micron Wavelength Range, Soy. J. Quantum Electron., Vol.
10, pp. 1146-1148, Sept. 1980.
38. M. Vasil'ev et al. Three-Layer Waveguide in GaAsP/InP
Injection Lasers, Soy. J. Quantum Electron., Vol. 14, pp.
431-432, March 1984.
39. L. Dolginov et al, Continouous-Wave Injection Lasers
Emitting in the 1.5-1.6 micron Range, Soy. J. Quantum
Electron., Vol. 12, p. 1127, September 1982.
55
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
M2012
40. D. Akhmedov, I. Ismailov, and N. Shokhudzhaev,
Fabrication and Investigation of GaInPAs/InP Heterolasers,
Soy. J. Quantum Electron., Vol. 12, 1568-1570, Dec. 1982.
41. Yu. Ayunts et al. Matching of Single-Mode Optical
Waveguides to Semiconductor Lasers, Soy. J. Quantum
Electron., Vol. 12, pp. 1428-1432, November 1982.
42. V. Duraev et al, Optical Fiber Coupling of 1.2-1.6
micron Radiation Emitted from Buried Mesastripe Injection
Lasers, Soy. J. Quantum Electron., Vol. 13, pp. 382-384,
March 1983.
43. L. Dolginov et al, Injection InGaAsP/InP Lasers with a
Threshold Current Density oif .5 kA/Sq-cm at 300K, Soy. J.
Quantum Electron., Vol. 14, p. 439 April 1984.
44. P. Eliseev, B. Sverdiov, and N. Shokhudzhaev, Reduction
of the Threshold Current of InGaAsP/InP Heterolasers by
Unidirectional Compression, Soy. J. Quantum Electron, Vol.
14, pp. 1120-1121, Aug. 1984.
45. L. Dolginov et al, Temperature Dependences of the
Emission Characteristics of GaInPAs/InP Injection Lasers,
Soy. J. Quantum Electron, Vol. 12, pp. 1237-1238, Sept. 1982.
46. A. Adlivankin et al, Characteristics of Radiation
Emitted by GaSb/GaAlAsSb Injection Heterolasers, Soy. J.
Quantum Electron, Vol. 13, pp. 1532-1534, Nov. 1983.
47. A. Bogatov et al, Tunable CW Emission in the 1.3 micron
Range from a GaInPAs/InP heterolaser with an External
Dispersive Resonator, Soy. J. Quantum Electron, Vol. 12, pp.
963-965, July 1982.
48. I. Goncharov et al, Influence of an External Feedback on
the Tuning of the Emission Frequency of a Semiconductor
Laser, Soy. J. Quantum Electron, Vol. 13, pp. 643-645, May
1983.
49. A. Bogatov et al, Direct Detection of Picosecond Pulses
Emitted by an Injection Laser with Active Mode Locking, Soy.
J. Quantum Electron, Vol. 13, pp. 1303-1304, Oct. 1983.
50. Yu. Bessonov et al, Generation of Picosecond Pulses in
an Injection Laser with an External Selective Resonator, Soy.
J. Quantum Electron, Vol. 12, pp: 1510-1512, Nov. 1982.
51. D. Annenkov et al, Spectrally Matched Modulation, at
Frequencies up to 2 GHz, of Injection Laser Radiation in a
Traveling-Wave Amplifier, Soy J. Quantum Electron, Vol. 14,
pp. 163-164, Feb. 1984.
56
I Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
52. Z. Nesterova et al, Transmission of Microwave-Modulated
Radiation along Fiber Optical Waveguides, Soy. J. Quantum
Electron, Vol. 7, pp. 910-912, July 1977.
53. A. Grudinin et al, Investigation of the Response Time of
a Silicon Photodiode Governed by the Transit Time in the
Space-Charge Region, Soy. Phys. Semicond., Vol. 15, pp. 527-
530, May 1981.
54. A. Vilisov, V. Voronkov, and V. Pozolotin, Sinicon P-I-N
Photodiodes with V-like Reflection Relief, Soy. Phys..
Semicond., Vol. 15, pp. 570-571, May 1981.
55. S. Ryvkin et al, Concept of Fast-Response Photodiodes,
Soy. Phys. Semicond., Vol. 16, pp. 669-672, June 1982.
56. I. Neizvestnyi et al, Detectivity Spectrum of Germanium
Photodetectors in the Fundamental Absorption Band Region,
Soy. Phys. Semicond., Vol. 14, pp. 1087-1089, Sept. 1980.
57. V. Karavanskii et al, Integrated-Optics Photodetector
Utilizing the External Photoelectric Effect in a Schottky
Barrier, Soy. J. Quantum Electron, Vol. 13, pp. 259-261, Feb.
1983.
58. V. Korol'kov and M. Mikhailova, Avalanche Photodiodes
made of Solid Solutions of III-V Semiconductor Compounds
(Review), Soy. Phys. Semicond., Vol. 17, pp. 355-363, April
1983.
59. L. Dolginov et al, High-Efficiency GaInPAs/InP Light-
Emitting Diodes, Soy. J. Quantum Electron, Vol. 8, pp. 1404-
1405, Nov. 1978.
60. B. Alyab'ev et al, Eight-Channel Optical Fiber
Communication Line between Computer Units, Soy. J. Quantum
Electron, Vol. 7, pp. 920-922, July 1977.
61. Yu. Vorob'ev et al, Quantum Electronic Devices in
Prototype Fiber Optical Communications Line, Soy. J. Quantum
Electron, Vol. 7, pp. 1020-1022, Aug. 1977.
62. M. Belovolov et al, Investigation of Optical Fiber
Systems for Communication between Computer Units, Soy. J.
Quantum Electron, Vol. 7, pp. 1404-1406, Nov. 1977.
63. Zh. Alferov et al, Fiber-Optical Long-Distance
Telecommunications Line Operating at the Wavelength of 1.3
microns, Soy. J. Quantum Electron, Vol. 8, pp. 1403-1404,
Nov. 1978.
64. M. Belovolov et al, Prototype Fiber-Optical
Communications Line with Spectral Multiplexing in the 1.3
57
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23 : CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
micron Region, .Sov. J. Quantum Electron, Vol. 9, pp. 1473-
1475, Nov. 1979.
65. Zh. Alferov et al, Multichannel Duplex Fiber-Optic
Communications Line Operating at the Wavelength of 1.3
microns, Soy. J. Quantum Electron, Vol. 12, pp. 1088-1090,
Aug. 1982.
66. M. Belovolov et al, Fiber-Optic Communication Line with
Multimode Waveguides for Data Transfer over Distances up to 8
km, Soy. J. Quantum Electron, Vol. 13, pp. 1619-1622, Dec.
1983.
67. R. Romaniuk, Fiber Optics and their Applications,
presented at the Third National Symposium of the Polish
Academy of Sciences, Jablonna, Feb. 15-17, 1983.
68. M. Drexhage and 0. El-Bayoumi, Heavy Metal Flouride
Glasses for Mid-IR Military Applications, Aerospace America,
pp. 66-69, April 1985.
69. T. Miyashita and T. Manabe, Infrared Optical Fibers,
IEEE J. Quantum Electron, Vol. QE-18, pp. 1432-1450, Oct.
1982.
70. S. Shibata, M. Horiguchi, K. Jinguji, S. Mitachi, T.
Kanamori, and T. Manabe, Prediction of Loss Minima in Infra-
Red Optical Fibres, Electron. Lett., Vol. 17, pp. 775-777,
Oct. 1981;
71. A. Rosiewicz and J. Gannon, Effects of Gamma Irradiation
on Mid-IR Transmitting Glass, Electron. Lett., Vol. 17, pp.
184-185, March 1981.
72. T. Kanamori et al, Chalcogeni de Glass Fibers for Mid-
Infrared Transmission, J. Lightwave Tech., Vol. LT-2, pp.
607-612, Oct. 1984.
73. H. Takahashi and I. Sugimoto, A Germanium-Oxide Glass
Optical Fiber Prepared by a VAD Method, J. Lightwave Tech.,
Vol. LT-2, pp. 613-616, Oct. 1984.
74. D. Tran, G. Sigel, and B. Bendow, Heavy Metal Fluoride
Glasses and Fibers: A Review, J. Lightwave Tech., Vol. LT-2,
pp. 566-586, Oct. 1984.
75. V. Artyushenko et al, Soy. J. Quantum Electron, Vol. 10,
p. 1181, 1980.
76. C. Artyushenko et al, Thallium Halide Fiber Waveguides
for Middle Infrared Range, Soy. J. Quantum Electron, Vol. 11,
pp. 239-240, Feb. 1981.
58
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
77. A. Vasil'ev et al, Single-Crystal Fiber Waveguides for
the Middle Infrared Range, Soy. J. Quantum Electron, Vol. 11,
pp. 834-835, June 1981.
78. N. Vechkanov et al, Infrared Fiber Waveguides Made of
Chalcogenide Glasses, Soy. J. Quantum Electron, Vol. 12, pp.
260-261, Feb. 1982.
79. V. Abashkin, A. Andriesh, and V. Ponomar', Glassy
Chalcogenide Semiconductor Fibers, Soy. J. Quantum Electron,
Vol. 12, pp. 1571-1574, Dec. 1982.
80. A. Bagrov et al, Fiber Waveguides for the Middle
Infrared Range Made from As-S and As-Se Glasses with Optical
Losses Below 1 dB/m, Soy. J. Quantum Electron, Vol. 13, pp.
1264-1266, Sept. 1983.
81. A. Andriesh et al, Characteristics of the Absorption of
Radiation in the 1-8 micron Wavelength Range by Fibers Made
of Glassy Arsenic Sulphide, Soy. J. Quantum Electron, Vol.
14, pp. 855-857, June 1984.
82. I. Goldobin et al, Injection Traveling-Wave Laser
Amplifier Based on a GaAlAs Double Heterostructure, Soy. J.
Quantum Electron, Vol. 14, pp. 255-259, Feb. 1984.
83. R. Dokhikyan et al, Prototype of an Integrated-Optics
Four-Digit Analog-Digital Converter, Soy. J. Quantum
Electron, Vol. 12, pp. 806-808, June 1982.
84. Yu. Gulyaev et al, Corrugated Focusing Grating for
Coupling Radiation in and out of Diffused Lithium Niobate
Waveguides, Soy. J. Quantum Electron, Vol. 13, pp. 520-523,
April 1983.
85. V. Atuchin, K. Ziling, and D. Shipilova, Investigation
of Optical Waveguides Fabricated by Titanium Diffusion in
Lithium Niobate, Soy. J. Quantum Electron, Vol. 14, pp. 671-
674, May 1984.
86. S. Bozhevvol'nyi et al, Investigation of an Electro-
optic Modulator Formed from Coupled Channel Diffused
Waveguides in Lithium Niobate, Soy. J. Quantum Electron, Vol.
12, pp. 1165-1169, Sept. 1982.
87. V. Zolotarev et al, Integrated-Optics Channel Switch
Formed from Electric-Field-Induced Waveguides in Lithium
Niobate, Soy. J. Quantum Electron, Vol. 12, pp. 1241-1243,
Sept. 1982.
88. M. Miler, V. Sychugov, and T. Tulaikova, Multichannel
Demultipleser for Optical Communication Lines, Soy. J.
Quantum Electron, Vol. 14, pp. 404-406, March 1984.
59
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8
M2012
89. M. Belovolov et al, Experimental Fiber-Optic
Communication Line with Spectral Multiplexing of LED
Radiation by a Diffraction Grating, Soy. J. Quantum Electron,
Vol. 12, pp. 242-243, Feb. 1982.
90. V. Karavanskii et al, Investigation of a Frequency-
Division Data Channel Mul.tiplexer, Soy. J. Quantum Electron,
Vol. 10, pp. 783-784, June 1980.
91. M. Belovolov et al, Low-Loss Fiber-Optical Directional
Couplers, Soy. J. Quantum Electron, Vol. 10, pp. 911-912,
July 1980.
92. D. Akhmedov et al, Heterojunction'Phototransistors Made
of InGaAsP Solid Solutions, Soy. Phys. Semicond., Vol. 16,
pp. 235-236, Feb. 1982.
93. T. Murina and N. Pozanov, Operation of Hybrid Optically
Bistable Devices, Soy. J. Quantum Electron, Vol. 11, pp. 711-
714, June 1981.
94. A. Maier, Optical Transistors and Bistable Devices
Utilizing Nonlinear Transmission of Light in Systems with
Unidirectional Coupled Waves, Soy. J. Quantum Electron, Vol.
12, pp. 1490-1494, Nov. 1982.
95. R. Rosenberg, "Next Step for Fiber Optics: The Local
Loop", Electronics, November 27, 1986.
96. J. Gosch, "Philips Local Area Net Employs Fiber Optics",
Electronics, December 23, 1985.
97. Space Net KMART usage discussions with SPRINT
personnel., December 1986.
98. T. Bell, "Communications", IEEE Spectrum, January, 1986.
99. Selected Material from Soviet Communication Exhibits,
Moscow, 1986.
100. Henry Root, GTE-SSD, "Survivability of Communications
Over Cable Transmission Systems During Periods of Magnetic
Disturbance", Electronics For National Security, 27-29
September, 1983.
60
Declassified in Part - Sanitized Copy Approved for Release 2013/04/23: CIA-RDP92B00181R000300270029-8