PHOTOGALVANIC CELL RESEARCH (SANITIZED) FINAL REPORT

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-03424A001200030004-3
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 56
• Document Creation Date: December 22, 2016
• Document Release Date: February 16, 2012
• Sequence Number: 4
• Publication Date: March 15, 1960
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
UUNFIDENTIIAL
Period Covered: 1 May 1959 to 1 January 1960
Report Date: 15 March 1960
C U
Fr%Nfr--~
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
This is the Final Report pertaining to Photogalvanic Cell
Research, It concerns the progress made during the
period 1 May 1959 to 1 January 1960.
This program was aimed at studying and developing electrolytic
batteries which could operate normally without irradiation and be re-
charged by light.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
Page
Preface ........................................................ ii
List of Illustrations .......................................... iv
List of Tables ................................................. v
I. ABSTRACT .............................................. 1
II. INTRODUCTION .......................................... 1
III. PHOTOGALVANIC SILVER HALIDE BATTERIES ................. 3
A. Silver Halide Membrane Type ....................... 3
B. Electrolytic Analogs of Semiconductor Junctions ... 11
C. Silver Halide Suspension Type 17
D. Discussion of Silver Halide Batteries ............. 17
IV. PHOTOGALVANIC SEMICONDUCTOR BATTERIES ................. 19
V. CONCLUSIONS ........................................... 25
Appendix A: CYCLIC PHOTOGALVANIC SILVER HALIDE CELLS 26
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
Figure
Page
1
Schematic Illustrations of an Ag, AgCl, Aqueous FeC12II
FeCl3 (Pt) Photogalvanic Battery with Rectification
Achieved by an Anion-Impermeable Membrane Combined with
Complex Formation ........................................
4
2
Schematic Illustration of an Ag,AgCl, Aqueous FeC12IJ
FeCl3 (Pt) Photogalvanic Battery with Rectification
Achieved by a Cation-Impermeable Membrane with a Three-
Electrode Arrangement ...................................
5
3a
Design of Photogalvanic Cell with AgX Membrane ..........
6
3b
Reaction Steps in AgX Photogalvanic Batteries ...........
7
3c
Reaction Steps in AgX Photogalvanic Batteries ...........
8
4
Comparison of Fixed and Mobile Charges in P-, N-, and I-
type Semiconductors, Ion-Exchange Resins, and Water ......
Conceivable Semiconductor Junction Arrangements and
12
Their Respective Electrolytic Analogs
14
6
Schematic Configurations of a P-N Semiconductor Diode
with Ohmic Contacts, Its Electrolytic Analogs, and the
Electrolytic Analog of a P-I-N Structure with Ohmic
Contact .................................................
Schematic Configuration of an AgX Membrane Type Photo-
galvanic Cell with Rectification Achieved by an Electro-
lytic Analog of a P-I-N Junction ........................
15
8
Combination of a Photovoltaic Diode with an Electro-
lytic Cell or with the Possible Circuit Equivalents
of the Cell .............................................
20
Reversible Concentration Cell with Selectively Permeable
Membrane ................................................
20
10
Electrolytic Resistance Cell ............................
20
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
LIST OF ILLUSTRATIONS (cont'd)
Figure
Title
Page
11
Photogalvanic Semiconductor Cells without Charge
Storage Capacity ........................................
12
Photogalvanic Semiconductor Cell with Charge Storage ....
2!i
LIST OF TABLES
Table Title Page
I Comparison of Photochemical and Photogalvanic Quantum
Yields in AgX Systems after Various Periods of Illumina-
tion by 30-Watt Tungsten Lamps with Reflecting Focusing
Mirrors (Color Temperature 2600?K) ...................... 18
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
A basic chemical requirement for the operation of light-rechargeable
batteries utilizing light-sensitive electrodes or separating membranes,,
is the presence in the electrolyte of an oxidation-reduction couple whose
oxidation potential falls within a range determined by the oxidation-
reduction potentials of the light-sensitive material and its photochemical
products (cf.,Egs. (1) through (8),,Section III). The chemical reaction
steps that determine the performance of photogalvanic cells equipped
with light-sensitive silver halide (AgX) membranes are given.
An improved performance may be expected for cells in which a suitable
electrical space-charge field is created. Such a field could be formed
either by an electrolytic analog of a semiconductor P-K junction (Section
IV. B.) or by the formation of an actual P-K junction created by immersing
a semiconductor in a suitable electrolyte (Section V). The latter method
was shown to be effective in experiments with AgX and Ge. However, photo-
galvanic cells utilizing other light-sensitive semiconductors may, also be
improved by the methods developed during the present study (Section VI).
The light-rechargeable batteries initially proposed operated in
the manner described below.
Given a battery deriving its electromotive force E from the difference
between the potentials El and E2 of the following half cell reactions:
MX+e-- M+X-
El = Elo - 0.059 log (X-)
Ox + e- - Red-
E2 = Ego - 0.059 log (Red-)
(ox)
so that the overall reaction on discharge is
Ox +M+X MX+Red-
E=E2-El
(1)
(2)
(3)
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
where MX is a light-sensitive, rather insoluble salt of a metal M (silver,
copper, etc.) and a non-metal X (I, Br, Cl or 0), Red- and Ox are com-
ponents of a suitable oxidation-reduction couple, and (X-), (Red-), (Ox)
are the concentrations of the respective species in the aqueous electrolyte.
Upon suitable irradiation of the light-sensitive discharge product MX,
photolysis occurs:
MX + hv--i M + 1/2 X2 (4)
which will be accelerated if the X product is removed through the secondary
reaction
1/2 X2 + Red- -* Ox + X-
,_ _ c_o - n ncn i,.,. (X-) (Ox)
(Red-) (X2) 1/2
so that the overall resulting reaction, (4) + (5), is
(5)
MX + h -V + Red--> M + Ox + X- (6)
which is the reverse of reaction (3); i.e., the battery is effectively
recharged by light.
Reaction (5) can occur only if the potential E is positive, which
is tantamount to the condition that the potential E4 ?or the reaction
1/2 X2 + e- X_
E4 = E4? - 0.059
(X
(7)
be higher than E2. Furthermore, because reaction (3) must also yield a
positive potential, it follows that
El < E2 < E4
(8)
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
This requirement imposes a basic limitation on the choice of a suitable
oxidation-reduction couple.
AgX batteries, developed before the present contract
came into effect, are shown in Figures 1 and 2. These, in turn, were
improved by using light-sensitive, semipermeable membranes instead of
light-sensitive electrodes (Figure 3a). The light sensitive membranes
consisted of rolled single crystal or polycrystalline sheets of AgC1
about 0.002" thick which could be partly converted on one side to AgBr
or AgBr + AgI by exposure to a solution of HBr with or without HI.
latter batteries,
The developments presented in the subsequent sections
IV.B.,C.,D.;and V) occurred exclusively during the present contract.
A. Silver Halide Membrane Type
The performance of photogalvanic cells, in which an AgX sheet
acts both as the basic light-sensitive material and as a semipermeable
membrane for the separation and storage of the photochemical reaction
products, is formulated in the following steps and is shown schematically
in Figures 3b and 3c:
Photochemical Charging (Figure 3b)
1. hv---> e- +h+
2. e- + Ag+ (in AgX) --) Ag (in AgX)
3. h+ (in AgX)--.h+ (at AgX-electrolyte) interface)
4. Ag+ (at AgX-electrolyte) interface) --p
Ag+ (at AgX-electrolyte,, interface)
5. h+ + X---),1/2 X2 (at AgX-electrolyte) interface)where B-
and h+ are free electrons and holes, respectively.
6. 1/2 X2 + Red, + Ag+--)Ox1 + AgX
7. Oxi (at AgX-electrolyte) interface) -j
Ox,+ (at Pt, electrode)
25X1
25X1
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Fe+3
Fe +2
MEMBRANE -
PERMEABLE TO
CATIONS BUT NOT
TO ANIONS
Fe (H2PO4)n
Fe+2
CI
H 2 P04
FIGURE I SCHEMATIC ILLUSTRATION OF AN Ag, AgCI, AQUEOUS
FeC 12 II FeCI 3 (Pt) PHOTOGALVANIC BATTERY WITH RECTIFICA-
TION ACHIEVED BY AN ANION- IMPERMEABLE MEMBRANE COM-
BINED WITH COMPLEX FORMATION
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
LOAD
Mh
MEMBRANE PERMEABLE TO ANIONS
BUT NOT TO CATIONS
0.04 m ( FeCI3 + FeCI2)
-0.1 m HCI SOLUTION
CONTAINING IO-4 TO 10-3 m
(FeCI2 + FeCI3 )
FIGURE 2 SCHEMATIC ILLUSTRATION OF AN Ag, AgCI, AQUEOUS FeC 1211 FeC 1 3 (Pt)
PHOTOGALVANIC BATTERY WITH RECTIFICATION ACHIEVED BY A CATION- IMPERMEABLE
MEMBRANE WITH A THREE - ELECTRODE ARRANGEMENT
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
\U/
FIGURE 3a DESIGN OF PHOTOGALVANIC CELL WITH AgX MEMBRANE
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
PtIr
Red a
Ag+
Ag +
e'
(2\) +Ag+
A~-1
ELECTROLYTE I L- AgX
(HF+AgF + Redl +O x=)
ELECTROLYTE IL
( HX+ Reda + 0 xa )
FIGURE 3b REACTION STEPS IN AgX PHOTOGALVANIC BATTERIES
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Pt1
I
Red =
(2') \ \ R / / +X-
ELECTROLYTE I
HF+AgF+Red, +Ox ,)
Z ELECTROLYTE IL
(HX+Reda+Oxa)
FIGURE 3c REACTION STEPS IN AgX PHOTOGALVANIC BATTERIES
by e- + h+ - (22)
ti
PtIr
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
8. Ox11+ (at AgX-electrolytell interface) >
h+ + Red11 (at AgX-electrolytell interface)
9. Red11 (at AgX-electrolytell interface)--> Red11 (at Pt11)
10. h+ (at AgX-electrolytell interface)- ,h+ (at Ag in AgX)
1 1 . h+ + Ag -~ Ag+ (in AgX)
Discharge through a Load (Figure 3b)
12. Oxi + e- -) Red1 (at Pt1 electrode)
13. Red1 (at Pt1)- Red1 (at AgX-electrolyte) interface)
14. Red1-,--> Ox 111+ + e- (at Pt11 electrode)
15. Ox11+ (at Pt11) -,Ox11 (at AgX-electrolytell interface)
16. Ag+ (at AgX-electrolyte11 interface) --,
Ag+ (at AgX-electrolyte) interface)
Undesirable Competing Reactions (Figure 3c)
17. e- + h+ -i heat
18. e- + R -) R-
19. R- + h+ --3 R (where R, R- may be an impurity or lattice
defect acting as a center for the recombination of holes
with electrons)
20. h+ + Ag -* Ag+
21. Ox1+ (at AgX-electrolyte) interface) h+ + Red1
22. h+ + X------> 1/2 X2 (at AgX-electrolytel1 interface)
23. 1/2 X2 (at AgX-electrolyte11 interface) X Red11 --->X- + 02111
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Steps (17) and/or (18) - (19) and/or (20) compete with steps
(2) and (3). Step (21) competes with steps (6) through (10), and steps
(22) - (23) compete with steps (5) through (15). The difference between
steps (20) and (11) is that in the latter step, the positive hole had
originated at the AgX-solution II interface (step (8)).
The charging and useful discharging reaction steps (1) through
(3), and (5) through (16) involve the motion of positive charges from
Pt1I through electrolyte II through the AgX towards electrolyte I and
Pt1,whereas the competing parasitic reactions involve charge motion in
the opposite direction, steps (20) through (22); or in either directiom,
steps (17) through (19).. Step (17) is of minor importance except under
conditions of intense illumination. Hence, in the absence of the recombina-
tion centers R, the main problem would consist in favoring steps (1)
through (16) and obviating steps (20) through (23).
An electric field within the AgX favoring the motion of positive
charges from right to left should tend to obviate the undesirable steps
(20) through (23) and favor all but one of the desired steps (1) through
(16). The one exception is the charging step (4+) which would have to
proceed against the direction of the field. However, when charge is
withdrawn as the cell is illuminated (which occurred in most of the
experiments) the discharging step (16) should cancel step (4+), since these
two steps are equal and opposite in direction. Therefore, under these
conditions the proposed field should lead to improved yields. Two alternate
methods of achieving this field without applying an auxiliary external
voltage are discussed in Sections IV. B. and V.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
B. Electrolytic Analogs of Semiconductor Junctions
It is known that space-charge regions exist in electrolytes
near metal electrodes, precipitated particles (Helmholtz and Gouy type
layers), and selectively permeable ion-exchange membranes. The latter
resemble semiconductors to the extent that they contain immobile species
of one charge sign and mobile species of an opposite sign.
The mobile H 0+ and OH ions in aqueous solutions and ion-3 exchange membranes are analogs of the holes, h+, and electrons a-, while
the immobile anionic and cationic groups in ion-exchange membranes re-
semble the acceptors and donors in semiconductors.
The recombination of holes and electrons giving rise to the
equilibrium condition in semiconductors
W) (h+) = Ks (9)
is analogous to the acid-base neutralization equilibrium
(OH-) (H30+) = KW 10)
and also to the precipitation equilibrium for insoluble salts such as
AgX
+) (X-) = K
(A
(1i)
g
AgX
where the symbols in parentheses are the concentrations of the respective
species. Because the constants Ks, KW and KAgX are of the order of
l0-13-13 (millimoles/cc)2 both in water and in good semiconductors,
Equations(9) and (10)afford a means of reducing the concentration of the
mobile species of any one charge sign to an extremely low value,,thus
providing the basis for good rectification. However, if other highly
mobile ionic species Y+ and Z-, e.g., Na+ and/or NO ions, which are
not subject to recombination, are also present in solution, then the
product of oppositely charged carrier concentrations may become much
larger than Kw. The aqueous solution would then be equivalent to a
semiconductor with a high value of Ks, i.e..,a "low band-gap" semiconductor
with rather poor rectifying ability. Finally, when the concentration of
mobile Y+ and Z_ ions exceeds 0.1 M, the aqueous solution becomes equivalent
to a highly conductive "degenerate" semiconductor with no appreciable
rectifying ability.
When the concentrations of y+ and Z- ions are less than 0.1 M,
Figure 4 shows the analogy between semiconductors containing an excess of
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
0 h h+ O
h+ ' +O h+
U kV ht
h+ 0 h+
e-
e O
O e-
e e O
0
e- 0 e-
H30+ 0 H30+
H30+O H30+
O H3 0+ O H3 0+
H3O+ O H30+
LOH
FIGURE 4 COMPARISON OF FIXED (ENCIRCLED) AND MOBILE CHARGES
IN P-, N - , AND I-TYPE SEMICONDUCTORS (a), (b), AND (c),
RESPECTIVELY, AND ION-EXCHANGE RESINS (d) AND (e ),
AND WATER (f). THE RIGHT-HAND DIAGRAMS ARE ELECTRO-
LYTIC ANALOGS OF THE ADJACENT SEMICONDUCTOR TYPES.
0 0H- O+ 0H-
0ri_ 0H
0H- O OH -(D
U+ OH- O+
0H- DOH-0 .,0
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
holes (p-type), electrons (n-type) or neither (intrinsic or n-type), on
the one hand,,-and aqueous solutions or resins containing an excess of
H30+ ions, OH- ions, or neither (pure water), on the other hand.
Figure 4 demonstrates the possibility of obtaining an elec-
trolytic analog of every conceivable type of semiconductor junction,
simply by arranging the corresponding components in corresponding order,
as shown in Figure 5. Of the eight diagrams of Figure 5, only (c), (d),
(e) and (f) are currently used. The I-P-I and I-N-I arrangements, (a)
and (b), do not seem to offer useful applications in semiconductor tech-
nology; however, their electrolytic analogs (e) and (f), are the only
arrangements in which ion-exchange membranes have been extensively used
thus far. Nevertheless, because the P-I-N and P-N arrangements (c) and
(d) are known to have excellent rectifying characteristics, it is worth-
while to examine their electrolytic analogs (g) and (h) more closely.
Arrangements (g) and (h) of Figure 5 require electrical con-
tacts to the end regions similar to the "ohmic" contacts of the p- and n-
type semiconductor regions required for the utilization of a diode in an
external circuit. Of course, reversible half cell electrodes with fairly
concentrated electrolytes would form the equivalent of an ohmic contact
as long as the concentrated electrolytes remain only on one side of each
membrane, as shown in Figures 6a and 6b. However, the concentrated
electrolytes composed of small ions would rapidly permeate the membranes,
and destroy all the junctions. Fortunately, it is possible to use highly
soluble large sized ions impermeable through the membranes,.e.g., long-
chain organic quaternary ammonium or sulfonate ions. This system is
indicated in Figures 6c and 6d, which represent the practical electrolytic
analogs of semiconductor P-N and P-I-N diodes.
As long as the membranes shown in Figures 6c and 6d remain
impermeable to the long-chain ions, the reversible reduction-oxidation
couples, composed of ions of the type Y+ and Z- should remain separated
for an indefinitely long time. Thus, it is theoretically possible to
construct rechargeable inert-electrode batteries with reactants stored
indefinitely in adjacent aqueous solutions.
An application of arrangements analogous to those of Figures
6c and 6d is shown in Figure 7, which represents an improved model of an
AgX cell of the Figure 3 type. The latter cell yielded the highest
currents with very thin sheets of AgX. However, the thin sheets readily
developed pores and caused the mixing of the electrolytes. Because the
arrangement of Figure 7 provides for a permanent separation of the two
electrolytic solutions, extremely thin AgX sheets may be employed.
-13-
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
H3O EXCHANGE MEMBRANE
H30 + EXCHANGE MEMBRANE
H30+ EXCHANGE MEMBRANE
(g)
OH EXCHANGE MEMBRANE
OH EXCHANGE MEMBRANE
FIGURE 5 CONCEIVABLE SEMICONDUCTOR JUNCTION ARRANGEMENTS (a), (b), (c), AND
( d ) AND THEIR RESPECTIVE ELECTROLYTIC ANALOGS (e) , (f ), (g), AND (h)
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
METALLIC
CONTACT
HEAVILY
P -TYPE
REVERSIBLE
ELECTRODE
httwc :a1 BL E
ELECTRODE
REVERSIBLE
ELECTRODE
H30+ H30+ H30+
+ + +
H30+EXCHANGE
OH- EXCHANGE
+ OH_
Y+ H30t H30t
MEMBRANE
MEMBRANE
OH- OH- Z_ +
Z
+1 - _
H 0
Y mow` 3 +
_ + +
Z
H3O + EXCHANGE
OH- EXCHANGE
H30+ Y+
MEMBRANE
H2 0
MEMBRANE
Z
OH- OH- OH-
H30+ Y +
Z -'-OH I
REVERSIBLE
ELECTRODE
(d)
FIGURE 6 SCHEMATIC CONFIGURATIONS OF A P-N SEMICONDUCTOR DIODE WITH OHMIC CONTACTS (a), ITS
ELECTROLYTIC ANALOGS (b) AND (c), AND THE ELECTROLYTIC ANALOG OF A P-I-N STRUCTURE WITH OHm,,1C
CONTACTS (d). DIAGRAM (b) REPRESENTS A TRANSIENT CONFIGURATION. DIAGRAMS (c ) AND (d) REPRESENT
CONFIGURATIONS STABILIZED BY LONG-CHAIN IONS.
OR
DEGENERATE
SEMICONDUCTOR
REGION
METALLIC
CONTACT
HEAVILY
N-TYPE
OR
DEGENERATE
SEMICONDUCTOR
REGION
REVERSIBLE
ELECTRODE
CONCENTRATED
H3O EXCHANGE
OH- EXCHANGE
CONCENTRATED
SOLUTION OF
SOLUTION OF
Z- AND H30+
MEMBRANE
MEMBRANE
Y+AND OH-
REVERSIBLE
ELECTRODE
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
INERT
ELECTRODE
INERT
/ELECTRODE
X
Y+
Y+
+
X- EXCHANGE
Y+ EXCHANGE
MEMBRANE
MEMBRANE
X +
Ag+ _ Y+
X-
Y Ag+
FIGURE 7 SCHEMATIC CONFIGURATION OF AN AgX MEMBRANE TYPE
PHOTOGALVANIC CELL WITH RECTIFICATION ACHIEVED BY
AN ELECTROLYTIC ANALOG OF A P-I-N JUNCTION
7 AgX
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Reaction between photolytic halogen and the ion-exchange
membranes adjacent to the illuminated AgX had to be prevented in the cell
of Figure 7. This reaction was prevented by an exhaustive bromination
process which, unfortunately, also rendered the initially transparent
membranes opaque to blue light. However, it should be possible to pre-
pare transparent exhaustively chlorinated or fluorinated ion-exchange
membranes suitable for the arrangement of Figure 7-
C. Silver Halide Suspension Type
Two 15 cc compartments separated by an anion exchange membrane,
each containing a Pt electrode, were filled with the same Solution of
0.005 M FeSO)+ plus 0.5 M H2SO4. One compartment also contained 10 gms of
fine AgBr particles and a small glass-encapsulated stirring magnet. The
compartments, once filled and closed; were placed in the dark and allowed
to equilibrate overnight. Subsequently, the solution containing'AgBr,
was continuously illuminated with a 30-watt microscope lamp and stirred.
A flow of electric current through a 50-ohm load began immediately. The
current climbed slowly to a maximum value of 1.5 ma within ,2 hours and
remained in the range of 1 to 1.5 ma for more than 3 hours. The quantum
yield was 20% to 50%.
Two objectionable features of this type of cell were: (a)
stirring was required for maximum current output; (b) the cell was es-
sentially a concentration cell, hence the voltage output was neither
likely to reach nor appreciably exceed the value of 0.1 v; and even such
a low value could not be maintained for a long period of time. Neverthe-
less, the initial current yields with these cells bridge the gap between
the high photochemical quantum yields (5o% to 100%) obtained for periods
of several hours with stirred AgX powders (and for the first five minutes
of illumination with AgX sheets), and the relatively low photogalvanic
quantum yields (1% to 5%) obtained with AgX electrodes and membranes for
periodsof several days. See Table I.
D. Discussion of Silver Halide Batteries
In Table I the photogalvanic quantum yields obtained with the
three types of AgX cells described in Sections III and IV are compared
with measured photochemical quantum yields.
In AgX membrane type cells, photochemical yields decreased
rapidly upon prolonged illumination because Ag accumulated near the AgX
surface. This explanation was confirmed by etching down the surface of
previously illuminated AgX sheets; the etched sheets gave initial photo-
chemical quantum yields comparable to those of freshly illuminated,
-17-
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Comparison of Photochemical and Photogalvanic Quantum Yields in AgX
Systems after Various Periods of Illumination by 30.-Watt Tungsten Lamps
with Reflecting Focusing Mirrors (Color Temperature 2600?K)*
QUANTUM YIELDS AFTER ILLUMINATION TIME OF:
0-5 min.
5 min.
to
3 hrs.
3 hrs.
to
1 day
1 day > 1
to week
1 wk.
a. Photochemical.
Stirred AgCl and AgBr powders
50-100%
50-100%
10-50%
---
AgCl and AgBr sheets
50-100%
negligibly
---
---
b. Photogalvanic Cells
Cells with AgX membranes
---
1-2%
1-2%
1% =
1%
Cells with AgX electrodes
1-5%
1-5%
1-3%
0-2%
0
Cells with AgBr in suspension
---
20-50%
1-10%
0
0
In computing the quantum yields, only the light of wavelengths shorter
than the absorption edge of the AgX was considered.
previously unexposed sheets. The ability of AgX membrane type cells to
deliver appreciable currents for periods of several weeks, suggests that
Ag accumulation is somewhat reduced by a regeneration mechanism of the
kind shown in Figure 3b.
Excessive surface accumulation of Ag was probably the main
cause of deterioration in AgBr suspension type cells and at least a
partial cause of deterioration in the other cell types.
AgX electrode cells, which did not contain solid Ag initially,
deteriorated in less than 1 week partly because of the formation of pores.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Although batteries with solid Ag electrodes delivered appreciable current
for much longer periods of time, the true photogalvanic quantum yields
could not be measured easily (see Appendix A, Section I. A.).
Initial maximum photogalvanic quantum yields were higher for
the poorer cells (AgBr in suspension) and lower for the cells with AgX
membranes. The reverse was true when total photogalvanic currents were
computed over a period of more than 1 week: the AgX membrane cells sup-
plied the larger amounts of total charge upon prolonged illumination.
Most of the cells tested were composed of solid AgCl sheets
covered with AgBr-AgI on the illuminated side, containing 1 M HC1 in
the illuminated compartment,and 0.01 to 0.1 M FeS04 (with or without
0.01 M Ag2S04) in 0.5 M H2S04 (or saturated CuCl in 1M HC1)in the dark
compartment.
The potentialities of the AgX cells are still inadequately
explored. In addition to the large variety of available membranes, long-
chain ions, and oxidation-reduction couples having suitable electrode
potentials, there are reasons to experiment with (a) the AgX thicknesses
and the ratios of AgBr to AgI for maximum light-sensitivity and ionic
mobility, and (b) the inclusion of impurities for extending the wave-
length spectrum of the useful radiation, improving ionic mobilities, and
removing or neutralizing harmful recombination centers.
IV. PHOTOGALVANIC SEMICONDUCTOR BATTERIES
Before examining photogalvanic semiconductor cells, it is con-
venient to consider the circuits of Figure 8. Electrolytic cell (a) of
Figure 8 may be of the form shown in Figures 9 and 10. It is composed
mainly of two inert electrodes immersed in an electrolytic solution con-
taining reversibly oxidizable and reducible ions, Red- and Ox, such that
the reaction
Red Ox + e- (12)
can occur at either one of the electrodes at appreciable current densities.
The electrolyte composition is initially uniform throughout the cell.
When switch 1 is closed, cell (a) at first behaves like a purely resistive
element, as represented by the equivalent circuit (b) with switch 2 closed.
Reaction (1) will then proceed from left to right at one inert electrode
and from right to left at the other. This will produce local changes in
the concentrations of Red- and Ox with continued passage of photovoltaic
current through the cell.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
C(
FIGURE 8 COMBINATION OF A PHOTOVOLTAIC DIODE WITH AN ELECTROLYTIC
CELL (a) OR WITH THE POSSIBLE CIRCUIT EQUIVALENTS OF (a)
,_ ELECTRODES _\
xx
I
e+Ox I Ox+e
-SEMIPERMEABLE
MEMBRANE
FIGURE 9 REVERSIBLE CONCENTRATION CELL WITH SELECTIVELY
PERMEABLE MEMBRANE
INERT i
ELECTRODES
P-N JUNCTION
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
In the cell of Figure 9, these local concentration changes
are tantamount to the charging of a concentration cell. The cell is
therefore equivalent to circuit element (c) or (d) of Figure 8 with
switch 3 or 4 closed. The charging can then continue only until the
counter-voltage built up in the cell is equal to the maximum photo-
voltage of the illuminated diode. However, if the semipermeable mem-
brane is eliminated, as shown in Figure 10, then mixing occurs and the
concentration cell counter-voltage is reduced. A steady state is reached
when the electrode reaction products are carried away as fast as they
are formed, by convection currents, diffusion, and by electric fields.
The cell is then still equivalent to circuit element (d) of Figure 8,
but with a low value of the counter-voltage E. Finally, if the con-
centrations of the Ox and Red- components are high and the electrodes
almost adjacent, the counter-voltage E may become negligibly low and the
cell may be considered nearly equivalent to the purely resistive element
(b).
A simplification of the arrangement of Figure 1 would be
(a) if a semiconductor electrode behaved like an inert
electrode when immersed in the electrolyte of cell
(a) and
(b) if the electrolyte changed the conductivity type of the
region of the semiconductor so as to form a P-N type
junction therein.
It is known that an electrolyte can form a p-type layer
over an n-type semiconductor or an n-type layer over a p-type material. by
forming a surface layer of ions of the same charge sign as the majority
carrier in the semiconductor.1 This layer of ions is formed either by
preferential adsorption of certain ions which have a specific affinity
for the semiconductor surface, or by an electrochemical potential gradient
across the semiconductor-electrolyte interface favoring the displacement
of ions of one charge sign towards the semiconductor surface. ?p3
Since illumination of a semiconductor P-N junction is known
to yield excellent photovoltaic effects, illumination of an electrolyte-
induced P-N junction should also generate marked photogalvanic currents.
1. W. H. Brattain and C. G. B. Garrett, "Electrical Properties of the
Interface Between a Germanium Single Crystal and an Electrolyte,"
PHYSICAL REVIEW, Vol. 94, No. 3, May 1954, p. 750.
2. Ibid.
3. W. H. Brattain and C. G. B. Garrett, U. S. Patent No. 1,870,344,
p. 3, lines 50 to 75.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Cells based on this phenomenon are shown in Figures 11 and 12.
When the above requirements (a) and (b) are met, the cells of
Figures 11 and 12 become equivalent to the arrangement of Figure 8a.
The electrolytic cells of Figure 11 resemble that of Figure 10, while
the cell of Figure 12 is analogous to that of Figure 9, in allowing for
charge storage. In Figure lla part of the light is absorbed by the
electrolyte solution which must therefore be free of appreciable con-
centrations of light-absorbing ions. This limitation is by-passed in
the arrangement of Figure llb where the light may pass through a
transparent ohmic contact. Such a contact will be even more useful in
the cell of Figure 12. The latter cell operates as described below.
Upon illumination, the photovoltaic flow of electrons through
the solution, the separating membranes, electrode A, diode D, and the
semiconductor, results in the charging reactions
Red-f+e- + Ox, (at the semiconductor) (13)
OxII + e--+ Red-II (at the electrode A)
On discharge through a load, the reverse reactions can then occur at
electrodes C and B. respectively. The diode D (which may not be needed
in some systems) prevents discharge through the photovoltaic charging
circuit in the dark.
A variation of Figure llb is shown in Figure 3a where the
transparent ohmic contact consists of a second electrolyte solution
which tends to induce the same conductivity type to the AgX surface as
that of the original bulk material.
Cells of the Figure lla type actually yielded appreciable
photogalvanic currents with:
(a) n-type AgCl or AgBr as semiconductor electrode, Ag or Pt
as the ohmic contact, Fe+2 - Fe+3 as the Red- - Ox couple,
and Cl- or Br as the adsorbed ions inducing p-type con-
ductivity to the AgX surface; and
(b) p-type Ge as semiconductor and Ti+2 - Ti+3 and/or Cr+2 -
Cr+ as the Red- - Ox couples.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
OHMIC CONTACT TO
SEMICONDUCTOR
ELECTRODE
FIGURE II PHOTOGALVANIC SEMICONDUCTOR CELLS WITHOUT CHARGE
STORAGE CAPACITY . (0) LIGHT MUST PASS THROUGH THE
ELECTROLYTE SOLUTION BEFORE REACHING THE SEMI-
CONDUCTOR. (b) LIGHT CAN PASS THROUGH A TRANSPARENT
CONDUCTOR FORMING AN OHMIC CONTACT TO AN EVAPORATED
SEMICONDUCTOR FILM.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
SEMIPERMEABLE MEMBRANE
OR SEPARATING SYSTEM
ZD
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
A typical electrolyte solution in the latter cell type was
composed of 0.0015 M GeCll + 0.05 M NaCl + 0.01 M HC1 + 0.15 M (TiC12 +
TiC13) + 0.5 M (CrC13 + CrC12). Photogalvanic currents of 10-100 ?A have
already been obtained with such cells upon relatively weak illumination,
but much more study will be required to optimize the yields from these
cells.
A basic chemical requirement for the operation of light-recharge-
able batteries utilizing light-sensitive electrodes or separating membranes,
is the presence in the electrolyte of an oxidation-reduction couple whose
oxidation potential falls within a range determined by the oxidation-
reduction potentials of the light-sensitive material and its photochemical
products (cf. Eqs. (1) through (8),Section III.A.).
Although the light-responsive substances studied were mainly AgX
and Ge, the designs shown in Figures 1 to 3, 7, 11, and 12, as well as
the relevant basic theory, should be applicable to other photogalvanic
systems. Furthermore, the possibi'ity of photogalvanic effects arising
from a combination of a semiconductor P-N junction photovoltaic effect
And regular electrode behavior has been at least partly confirmed by the
experiments with Ge and AgX types of cells.
Since the operation of the AgX type photogalvanic cells also appears
to be related to a P-N junction photovoltaic diode effect, the established
theory of this effect appears applicable. Now, the optimum theoretical
efficiency for the conversion of solar radiation into electricity is ob-
tainable with semiconductors having a 1.5 e.v. intrinsic band-gap width,
which is less than half the width of the silver halides (3 to 4 e.v.).
Therefore, it appears desirable to investigate photogalvanic systems
utilizing semiconductors with an optimum band-gap width.
This program has augmented the theoretical and technical knowledge
of photogalvanic phenomena; however, there were many barriers to develop-
ment of an efficient photogalvanic cell. For example, in the AgX membrane
type cells, Ag accumulated near the AgX surface and photochemical yields
decreased substantially upon prolonged illumination. In the AgX suspension
type cells, excessive accumulation of Ag caused deterioration of the cells;
voltage output was low and could not be maintained long. Cell deteriora-
tion was also encountered in the AgX electrode type cells. These deteriorated,
partially because of the formation of pores.
Because of these difficulties and the additional experimentation
yet required in associated areas, continued research activity with an
efficient photogalvanic cell as its end product would not be practical at
this time.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Cyclic Photogalvanic Silver Halide Cells
ABSTRACT
In order to distinguish between photoconductive or photocatalytic
and "truly photogalvanic" effects, continuous photocurrent measurements
were performed on cyclic Pt, Ag - AgC1, aqueous FeCl2 - FeCl3, Pt and Pt,
Ag - AgBr, aqqeous FeBr2 - FeBr3, Pt batteries in a state of nearly com-
plete discharge. These batteries could b% partly recharged by light of
wavelengths shorter than 4300 A and 4800 A,respectively, with maximum
photogalvanic quantum yields of 2 to 5%. These quantum yields vary main-
ly with method of electrode preparation, electrolyte composition, and
light wavelength, but not with light intensity for intensities equivalent
to usual solar radiation levels. Direct recombination of photochemical
reaction products, corrosion of Ag by an oxidant in the electrolyte, and
electrode impedance and polarization effects are the main current limit-
ing factors.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
A. Photogalvanic Quantum Yield Studies
Since Becquerel1 first discovered the ph.otogalvan:ic effect hundreds
of papers on the subject appeared in the scientific literature. ;13
However most of these have dealt with open.-circuit photopotentials
rather than photocurrents. Although the former can offer interesting
clues as to the type of photochemical products formed, if any, i.t is
still essential for any practical photogalvanic cells and also for any
thorough understanding of the kinetics of the photogalvanic process to
obtain vr.lues of (a) the quantum yields of the individual photochemical
reaction steps and (b) the net yields of the current that can be deliver-
ed for a given light input. Whereas the latter yields have been easily
determined for dry phot.ovoltaic cells such as the Si solar cells, there
exist difficulties in measuring them in most photogalvanic systems.
To appreciate these difficulties it is necessary to distinguish at
this point between those systems which in the dark: a) are in or near
a thermodynamic equilibrium and in which the light either gives rise
directly to products dischargeable through a battery circuit or else
generates an e.m.f. capable of producing chemical changes by electro-
.lys.is; and those which. b) are not in true thermodynamic equilibrium and
in which the light catalyzes an otherwise extremely sluggish. reaction or
else accelerates the discharge of a battery by a photoconductive and/or
depolarizing effect. Unfortunately, the systems for which p~otogurrent
measurements have been reported seem to fall into group (b). ,5, Fur-
themore, the term photocurrent (defined as the increase in current de-
livered by a cell when the potential of the illuminated electrode is
returnee to the value in the dark by applying or adjusting an. external.
voltage 15) does not distinguish between. current generated by a photo-
voltage and that due to photoconductivity.
Let us consider for example a battery having an open circuit e.m.f.
Ed in the dark and which. comprises one electrode covered 1., a photosensi-
tive material which also forms a current-limiting layer :riving a very
high resistance Rd. When, this electrode is illuminated, the current may
increase from the value in. the dark
Id ... Ed/Rd
to a much higher value
11 Ed/R1
(1)
(2)
in light, due solely to a photoconductive effect. Now, It is l nown
that photoconductors can. have gains higher than one, i.e. one photon
of light may allow passage of hundreds of electrons when an &J ectri.c
field is applied across the photoconductor. Thus one may obtain
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
deceptively high values for the photogalvanic quantum yields with
batteries which would normally deliver a current in the dark, no matter
how small. (either spontaneously or with the aid of an external polarizing
voltage), having the same polarity as the observed photocurrent.
Equations (1) and (2) describe essentially a conversion of chemical
into electrical energy accelerated by light. If the term catalysis is
generalized to include every type of acceleration of electrode tlischar.ge
reactions, then equations (1) and (2) can. be considered as describing a
special. form of the phot.ocatal.ytic effect, and the current Il can be
considered as a "photocatalytic" rather than "photogalvanic" current.
This objection. seems to apply ito the few reported photocurrent measure-
ments in electrolytic systems.6
On the other hand, in systems of group (a) the dark current Id may
again be described by equation (1.); however, its low value must be
associated with a low value of Ed rather than with a high Rd. The
current could then increase to a higher value
Il = El/Rd
(3)
due to a photo-gen.errated e.m.f. Since the latter implies actual con-
version. of light into electrical energy, equation (3) describes a truly
photogalvanic current.
In frequent cases, however, both the e.m.f, and the resistance may
change with illumination so that
11 ` E:1/R1
(4)
The distinction between the photoconductive and the truly photogalvanic
effects may then pose a problem. Of course, when El is of opposite
polarity than Ed, then Il must represent a truly photogal.vani.c current
regardless of the value of R1. However, even In batteries having an
e.m.f. in the dark of the same sign as in. light it should be possible,
in principle, to measure truly photogalvani.c currents by a method of
exhaustion., i.e. by starting with a very limited quantity of reducible
and/or oxidizable components and. drawing much more current over a long
period of time than could have been delivered by complete discharge with
100% current efficiency. Obviously, the method of exhaustion will be
most sensitive when a battery can be completely discharged at the start
of the i.1.l.umi.n.at.Ion. or, what is tantamount. but simpler, when the battery
is assembled with the discharge products replacing at least one essential.
but missing reactant.
If a completely discharged battery can deliver appreciable current
upon illumination, then the light apparently effects at least a partial
regeneration of the battery reactants. These regenerated reactants thus
undergo cyclic chemical changes.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
In the present work, the photogalvanic quantum yield Yp is defined
as the ratio of the number of electrons Ne caused to go around an extern-
al circuit per number of absorbed photons Nph in a cyclic photogalvanic
cell.
Yp = (Ne/Nph) cyclic (5)
For the most unambiguous measurements, the cell should have an initial
open circuit voltage in the dark close to zero or possibly even of a
sign opposite to that observed upon illumination. It is not necessary
that the cell voltage reverse or disappear when illumination is dis-
continued. In fact, it is preferable for the cell to have a certain
ability to store the charge and e.m.f. generated by the light. However,
one of the criteria for a truly light-induced current is a marked increase
of e.m.f. from a negligibly low value or complete reversal of polarity
upon illumination of the initially assembled cell.
A review of previously reported truly photogalvanic systems suggests
that the photogalvanic currents are usually both transient and extremely
low (i.e., less than lM) because of the absence of provisions for
effective cyclic utilization and regeneration of reaction products. For
instance, Becquerel's photogalvanic cellsl consisted essentially of one
bare Pt electrode and one Pt electrode covered with AgX (X = I, Br or Cl)
both immersed in the same electrolyte solution. However, the halogen
formed in the photochemical reaction
AgX+h o)---iAg+1/2X2 (6)
cannot diffuse to the bare Pt electrode at a sufficiently high rate to
give rise to appreciable electrode discharge reactions, because appreciable
diffusion of the X2 would require a rather high concentration of X2 near
the AgX. This would be difficult (if not impossible) to achieve by
illumination on account of the tendency for the X2 to recombine with the
photolytically formed Ag.
The present studies are aimed at the development of cyclic photo-
galvanic cells and of methods of maximizing the quantum yield Yp. A
preliminary survey8 led to the choice of photogalvanic cells based on
reaction (6).
B. Cyclic Silver Halide Photogalvanic Cells
a) Photochemical Charging and Discharging Reactions (Figures 1
and 2)
In the present work (see Figure 1), the cells differ from the
Becquerel cell in that they contain in addition the components Red- and
A-3
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
Ox of a suitable reduction-oxidation couple which can remove the X2 at
the AgX-electrolyte interface via the reaction
1/2 X2 + Red- ---o Ox + X- E7 = E,?7 - 0.059 Log (A-) (Ox) 1/2 (7)
(Red-) (X2)
where the symbols in parentheses represent the concentrations of the
respective components.
The net result of reactions (6) and (7) is then
by + AgX + Red- --.p Ag + X- + Ox
(8)
which can be considered as a photochemical recharging of an Ag-Ox battery
completely discharged via the reactions
AgX + e-~ Ag + X- E9 = E9 - 0.059 Log (X
Ox + e-" -7 Red- E10 = E10 + 0.059 Log (Ox)
(Red-)
yielding an overall discharge reaction
Ox + Ag + X---..* Red- + AgX Eli = E10 E9
which is the reverse of reaction (8).
(9)
(10)
In order for the cell to be cyclic, the discharge reaction
should occur spontaneously; i.e., Ell should be positive. Furthermore,
in order for reaction (7) to occur, it is necessary that the potential
E12 for the reaction
1/2 X2 + e---) X- E12 = E12 - 0.059 Log (x)/(x2)'2 (12)
be higher than E10. Hence
E9 < E10 < E12
which is a basic limitation on the choice of the Red - Ox couple.
b) Competing Wasteful Reactions (Figure 3)
Reaction (6) can be considered to be composed of the basic
electronic process
(13)
h1)> e- + h+ (14)
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
where e- and h+ are electrons and holes, respectively, followed by
e- + Ag+> Ag
h+ + X ---)' 1/2 X2
as shown in Figure 2
(15)
(16)
This suggests the possibility of several wasteful recombina-
tion reactions shown in Figure 3
or, what is more likely,
followed by
(17)
(18)
(19/
where R, R may be an impurity or lattice defect acting as a center for
the recombination of holes with electrons.
On the other hand, the useful discharge reaction step (9) can
also be regarded as the sum of
AgX F- Ag + h+ + X-
(20)
e- - h+ (at Pt-AgX interface) (21)
where the h+ originates at the Pt electrode as shown in Figure 2. The
last two steps must be differentiated on the one hand from the following
wasteful reactions shown in Figure 3
h+ + Ag--.t Ag+ (22)
Ox (at AgX-electrolyte interface)---Red- + h+ (23)
followed by reaction (22), where the h+ originates from reaction (14+) or
(23), and on the other hand from the sequence of reactions (10), (20), and
(21) all occuring at the Pt-AgCl electrode in case of porosity of the AgX
film, as shown at the bottom of Figure 3. All these wasteful reactions
represent short-circuiting paths for the recombination of e- with h+ without
passage of the e- through an external circuit.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
c) Other Current-Limiting Factors
Even in the absence of the above parasitic reaction mechanisms,
the current-density is limited by the resistance across the AgX film and
by concentration and electrode polarization effects.
Attempts to reduce these limitations introduces other difficulties.
E.g., thinner films of AgX will have lower resistance, but they also form
the short-circuiting pores considered in the preceeding section. Further-
more, an appreciable fraction of the useful radiation of wavelengths
approximating the absorption edge of the AgX may be lost by the thinner
films but absorbed and converted into useful currents by the thicker ones..
Similarly, concentration polarization effects are reduced with high (Ox)
concentrations, but these accelerate the parasitic reaction (23)-
C. Experimental Procedure
a) Optical System
The optical arrangement is shown schematically in Figure 4.
Light from a 30-watt microscope lamp having a color temperature of 2400
to 2700?K was focused onto an area of 1 to 2 cm2 of the Pt-AgX electrode
g by means of the hemispherical mirror m and the condensing lens C. Most
of the infrared radiation was absorbed by the water filter W. Filters B
and D could be of the narrow-band interference type for experiments with
monochromatic radiation, neutral density filters for studies with varying
light intensities, or Wratten filters for cutting out some of the longer or
shorter, wavelength components in special orienting experiments.
Relative light intensity measurements were performed with the
photocell P located behind the transparent cell E. Absolute light intensity
measurements were obtained by focusing the light source on a previously
calibrated photovoltaic Si cell about 1 x 2 cm2 in area (not shown in
Figure 4).
b) Test Cell Assembly
The test cell E was a square transparent Lucite box with an
outer jacket containing circulating water from a thermostat bath at 25?C.
Electrode f was a 52 mesh Pt gauze made of 0.004-inch diameter wire,
transmitting about. 60 to 70% of the impinging radiation. Electrode g
was prepared by abrading the surface of a 0.002-inch thick Pt foil with
grade 320 emery paper, etching in aqua regia, rinsing, immersing in
molten AgX contained in a -porcelain crucible at about 700?C and slowly
withdrawing the Pt while allowing the edge of the Pt to touch the crucible.
The resulting films of AgX over the Pt surface were well-adhering,
about 5-20 microns thick, and impervious to the electrolyte. The latter
consisted of solutions of 0.001 to 0.1 (usually 0.01) M FeC12 plus 0.01 to
A-6
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
1. (usually 0,.2) M HCl in experiments with Pt-AgCl electrodes, and of 0.01
M NaBr plus 0.005 M FeS04 with Pt-AgBr. The ratio of Fe+2 to Fe+3 ions
was maintained nearly constant in some experiments by addition of FeC13
complexed with an excess of HF.
All reagents were of analytical grade. The AgC1 was Baker
Analyzed reagent, whereas the AgBr was prepared by precipitation of
0.2 M AgNO3 solutions with either 0.2 M NaBr or 0.2 M HBr.
c) Electrical Measurements
A model 153 X 12V-X-6 25-mv Brown potentiometer recorded the
voltage drop across one of two resistance boxes connected in series with
the test cell electrodes f and g of Figure 4. Current-sensitivity and
total external resistance could be adjusted at will by means of the two
resistance boxes.
Internal cell resistances were not measured by means of the usual
alternating current bridge method, in order to remove the possibility of
any net charging. effects resulting from the application of an external
alternating voltage,. Instead, the internal cell resistances were deduced
from a series of current measurements with varying external resistances
in the high external resistance range, where the current output was low
enough to eliminate polarization effects. The same measurements also
yielded the values of open circuit cell voltage, which agreed with
readings of a 10=.megohm vacuum-type voltmeter. On the other hand, the
quantum yield measurements usually involved relatively low external
resistances
d) Photochemical Quantum Yield Measurements
The yields of reaction (3) were also measured by purely chemical
methods. A beaker containing a stirred mixture of 4 Gm powdered AgX and
100 cc of 0.001M FeX2 plus 0.02M HX was illuminated with the above mention-
ed 30-watt microscope lamp for 1-to-10 minute intervals. The powder was
then allowed to settle and 10-cc samples of the solution were withdrawn
and analyzed for Fe+2 ions by a permanganimetric method.9 Control titra-
tions were also performed on samples from mixtures stirred in the dark
and from illuminated mixtures of 4 Gm. Cr20 with the same solution,
to correct for any air-oxidation of Fe+2 ions. The rates of conversion
of Fe+2 to Fe+3 ions in the control experiments were negligibly low in
comparison with those observed upon illumination of the AgX.
The discharge characteristics of a typical Pt-AgCl, FeCl2, Pt cell
are shown in Figure 5. Upon illumination of the Pt-AgCl electrodes in
initially assembled test cells, the open-circuit e.m.f. rose from values
of less than 50 my to a value of about 0.4 v corresponding to the e.m.f.
A-7
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
of an Ag-AgCI-FeC13-Pt battery, This satisfies,, then.,, one basic. criter-
ion. enounced near the end of Section I.A. Furthermore, the slow con-
tinuous rise in current and voltage over a period of more than 10 minutes
indicates a gradual. build-up of photochemical reaction products. Finally,
a most convincing proof of true photochemical charging was obtained by
illuminating the cells for varying time intervals while on open-circuit
and integrating the total. current that could be withdrawn. in the dark
immediately following the illumination, as shown in Figure 6. The plot
of charge withdrawn versus time of illumination is linear (Figure 7),
in agreement with the charging reaction of Section I.B. (a).
The electrodes used for the results of Figure 7 were among the
earliest and least efficient ones, yielding an average current of less
than 10?A during continuous illumination. In order to optimize the
current yields, the thickness of the AgCl-layers and composition of the
solution were varied. Using a 1/100 M solution of FeC12 which was
approximately 1/1000 molar in. FeC13, charging currents of up to 80?A
for a short period of time and up to 4011A for several. hours were obtained,
when the AgCl-layers were 5-20 microns thick.
With thinner layers, higher porosity and lower light absorption led
to a lower efficiency whereas with thick layers (> 20?) the high impedance
of the Pt-AgCl electrode limits the current.
The best results obtained correspond to quantum yields of 2 to 5%
for light of wavelengths shorter than 4,000 A. The relative sensitivity
of two Pt-AgCl electrodes to various wavelengths is shown in Figure 8.
By varying the intensity of illumination with neutral density
filters, it was found that the current yield increased linearly with
intensity (Figure 9), showing no sign of any intensity saturation effect
with the light source used.
With Pt-AgBr electrodes the voltage and cur-rent changes upon
illumination were small. and usually in a direction opposite to that
expected by the reaction scheme of Section I.B. (a), when the AgBr was
prepared from NaBr and AgNO3 solutions. However, with AgBr prepared from
HBr and AgN03 solutions, both the polarity of the photo-induced voltage
and the photogalvanic quantum yields were comparable to those obtained
with AgCl. Furthermore, since the fraction of light from tungsten at
2600?K absorbed by AgBr is about four times larger than the fraction
absorbed by AgCl., the actual. photogalvanic currents with the same illumina-
tion were correspondingly increased by a factor of three to four. With
monochromatic light of various wavelengths, the relative photogalvanic
quantum yields varied as shown in Figure 10.
When exposed to the white light of a microscope lamp the AgBr
electrodes showed a change of color and a simultaneous decrease in photo-
galvanic efficiency within a few hours. It is knownl0 that light of wave-
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
lengths longer than the absorption. edge may disperse the photolytically
formed Ag forming an Ag colloid with an absorption band extending into
the intrinsic absorption region of the AgX. This colloid may decrease
the amount of light available for photolytically active absorption.
Furthermore, the colloidal Ag may accelerate hole-electron recombina-
tion via the reaction steps (15) and (22). (Cf. Figure 3 section I.B.
(b).)
This possibility was partly confirmed with filtered light in the
range of 3300-?+x+00 AO
was used for illumination of the AgBr. Almost no
color change could be observed under steady state conditions and the
photolytic sensiti.vity of the electrodes decreased only after a few
days, this decrease being probably due to the formation of pores.
It is noteworthy that the spectral sensitivity curves for both
AgCI. and AgBr (Figures 8 and 10) show maximum quantum yields for
wavelengths just below the absorption edges of the respective halides.
To explain the decrease in quantum yield with decreasing wavelengths
one may assume that the latter, being highly absorbed at the very
surface of the AgX, result in the formation of Ag near the AgX-electrolyte
interface where re-oxidation by Fe+3 ions is relatively fast. Evidence
for this kind of recombination is given in. Figures 11 and 12. In
Figure 11, 80% of the charge formed by illumination of a Pt-AgCl, FeCl2,
Pt cell for 8.5 minutes is shown to disappear within 20 minutes. In
Figure 12, the photogalvanic current is found to be inversely propor-
tional to the Fe+3 ion concentration, for concentrations above 10-4 M
FeCl3.
Whereas the photogalvanic quantum yields were usually less than 5%
with both Pt-AgCI. and Pt-Agar electrodes the photochemical quantum
yields for the conversion of Fe+2 to Fe+i ions by illumination of both
AgC1 and AgBr powders (cf. section. I. C. (d) ) were 50-100%. These
high yields may be attributed partly to the absence of the stringent
electrochemical requirements for the delivery of useful photogalvanic
currents and partly to the large effective surface area of the powdered
AgX with a corresponding low surface density of photochemically formed
Ag taking part in wasteful recombination reactions.
Partial photochemical charging of a Pt-Ag-AgX, aqueous FeX3, Pt
cells was achieved with maximum photogalvanic quantum yields of 2 to 5%
in cells remaining In. a state of nearly complete discharge.
Tn partly charged batteries, i.e. containing more than l0-4 M FeX3
or art appreciable surface concentration of Ag, direct attack of the Ag
by Fei.j ions appreciably reduces the photogalvanic quantum yields in
direct proportion to the Fe+3 ion concentration. When the latter
reaches or exceeds 0.001M, the quantum yield may be reduced by a factor
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
of ten or more. On the other hand, for Fe+3 ion concentrations below
10-4M the photogalvanic current is severely limited by concentration
polarization at the Pt-FeX3 electrode. The optimum Fe+3 ion concentra-
tion is therefore limited to the fairly narrow range of 10-3 to 10-4M.
However, the most serious drawback of these cells is the deteriora-
tion observed after 1 to 5 days of illumination, attributable mainly to
the development of pores in the AgX film (cf. Figure 3, Section I. B.
(b) ).
Whereas the last difficulty applies only to AgX-covered Pt electrodes,
the problem of diffusion-limited current with low photochemical product
concentrations and of recombination with higher concentrations is likely
to be encountered in most photogalvanic systems. Some approaches towards
a solution of this problem are presented in a following publication.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
1. E. Becquerel, Compt. rend. 9, 561 (1839)
2. A. W. Copeland, 0. D. Black, and A. B. Garrett, Chem. Rev. 31,
177, (1942)
3. K. M. Sancier, Trans. Conf. Use of Solar Energy (University of
Arizona Press, 1958) V. pp. 43-56
4. H. Luggin. Z. Physik. Chem., 23, 577 (1897)
5. V. I. Veselovsky, J. Phys. Chem. (U.S.S.R.) 22, 1302 (1948)
6. A. Goldman and J. Brodsky, Ann. Physik, 44, 849 (1914); C. G.
Fink and D. K. Adler, Trans. Am. Electrochem. Soc. 58,175 (1930)
7. A. Rose, Proc. I. R. E. 43, 1850 (1955)
9. I. M. Kolthoff and E. B. Sandell, Textbook of Quantitative
Inorganic Analysis (Macmillan, New York, 1946) pp. 592
10. C. E. K. Mees, The Theory of the Photographic Process (Macmillan,
New York, 1946) pp. 285 ff.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
Pt
SCREEN
2 Op
2Op
Ag X
AQUEOUS SOLUTION OF RED, Ox, AND X
FIGURE 1 Schematic Drawing of a Cyclic Silver Halide Photogalvanic
Cell
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
Pt
SCREEN
Red--
e + Ox ?
R
% ed- + 2X2
Lg+ rAg
+ e
AgX
rIjUZ',Z 2 Dischar?ins and Photochemical Chargin?
,, ,_ Reactions at the
Electrodes of the Cell of Figure 1
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
AgX
(14) '
-- ht + e-
U9)
(18)
ht, t e-
I 'A
Ag A9t
Ox
(23)
h
l
Ag
Agt' Ag
(20) r
ELECTROLYTE
SOLUTION
Red Ox + e
001
PORE IN AgX
AgX
F'I(;U1(E 3 tTasteful J ecombi.nati.on P&:chanisms at the 1't:-A)X Electrode
of Figure I
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
rIGLi3 4 Optical Systems for Photogalvanic Quantum Yield Measurements.
Light from the 30-watt microscope lamp L is focused by means
of the adjustable reflecting mirror m and condensing lens C
onto the Pt-Agl electrode g. Intensity and composition of the
liryht are adjusted by means of the water filter W and inter-
changeable filters B and D. Phototube P located behing the
test cell E measures relative light intensities when the elec-
trodes f and g are temporarily removed.
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
E
0 8
LIGHTS ON
I
L
0 10 20 30 40 50
TIME ( minutes)
Fr--'U2 ,Z ?ypical Changes of Current Delivered by a Pt-AgCl, FeC12,
Ft Cell with and without Illumination
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
N
CL
E
a
Q,
lv~
Z
Ui
U
LIGHT
OFF
LIGHT
ON
0 40 so
TIME (minutes)
120
160
FIGURE 6 Current Withdrawn in tha Dark from a Pt-A,Cl, FeC12, Pt Cell
Fnllnwinv Tliff-ranfP aririd,. of Tllnminn tinn
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
8 16 24 32
PERIOD OF ILLUMINATION ( minutes)
FIGURE 7 Charge Withdrawn from Two Different Pt-Ac;cl Electrodes
After Various Periods of Illumination
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424A001200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
3,000 3,500 4,000 4,500
WAVELENGTH (R)
0
-0-- AgCI LAYER 20.u THICK
I -,A- AgCI LAYER < 51u THICK
I
I t
I
A/
5,000
5,500
8 Relative Dependence of the Photogalvani.c Quantum Yield
from i't-A-r,C1 Electrodes on Wavelength
A - L
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
102
10-3
10-2 10''
LOG INTENSITY ( relative)
FIGURE 9(a) Logarithmic Plot
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
0 40
0.4 0.6
RELATIVE INTENSITY
FIGURE 9(b) Linear Plot
A-21
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
w 0.4
r
d
~
Q
2,500
3,000 3,500 4,000
LIGHT WAVELENGTH (~ )
4,500
5,000
FIGURE 10 Relative Photo;alvanic Quantum Yields from Pt-Ao"Br Electrodes
with Light of Various Wavelengths
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
4 8 12 16
TIME OF STORAGE ON OPEN CIRCUIT FOLLOWING
ILLUMINATION ( minutes)
FIGURE 11 Loss of Charge During Open Circuit Storage
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
 Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3
0
10
3 4 5
1 / (FeC 13) (liter / millimole )
FIGURE 12 Inverse Proportionality of Photogalvanic Current to the
FeC13 Concentration Observed with Two Pt-AgC1 Electrodes
Declassified in Part - Sanitized Copy Approved for Release 2012/02/16: CIA-RDP78-03424AO01200030004-3