An electrical power source can be made by exposing a p‐n junction to radioactivity or light, so that the junction field separates electron‐hole pairs produced by the radiation. Expressions for maximum power, optimum load resistance, and efficiency are derived from an equivalent circuit and rectification theory. Power and efficiency increase with source current Ig of separated charges and zero‐bias junction resistance. Ig increases with energy and intensity of radiation, but is limited by self‐absorption in the radioactive isotopes. Estimates of attainable power and efficiency for silicon cells are 3·10−3 watt cm−2 and 15 percent for solar radiation, averaged, allowing for night, weather, and varying angle of incidence; and 3·10−4 watt cm−2 and 8 percent, for beta radiation from Sr90–Y90 of activity 32 Curie/g. However, lattice defects produced by Sr90–Y90 beta radiation impair cell performance by increasing electron‐hole recombination. A theoretical estimate of threshold energy for radiation damage in silicon is about 0.3 Mev, about half the experimental value reported for germanium. Avoiding radiation damage by annealing, by absorbers, and by use of less energetic isotopes is discussed. The Y90 beta spectrum is given; it is used in estimating damage rates in germanium, which are high, and efficiencies obtainable with absorbers, which are low. Theory and experiment are compared for Sr90–Y90 cells of silicon and of germanium.

1.
H. G. J.
Mosely
,
Proc. Roy. Soc. (London)
A88
,
471
(
1913
).
2.
E. G.
Linder
and
S. M.
Christian
,
J. Appl. Phys.
23
,
1213
(
1952
).
3.
P.
Rappaport
and
E. G.
Linder
,
J. Appl. Phys.
24
,
1110
(
1953
).
4.
J. B.
Kramer
,
The Electrician
93
,
497
(
1924
).
5.
P. E.
Ohmart
,
J. Appl. Phys.
22
,
1504
(
1951
).
6.
Chapin
,
Fuller
, and
Pearson
,
J. Appl. Phys.
25
,
676
(
1954
).
7.
Ehrenberg
,
Lang
, and
West
,
Proc. Phys. Soc. (London)
A64
,
424
(
1951
).
8.
P.
Rappaport
,
Phys. Rev.
93
,
246
(
1954
).
9.
Shockley
,
Sparks
, and
Teal
,
Phys. Rev.
83
,
151
(
1951
);
R. L.
Wallace
, Jr.
, and
W. J.
Pietenpol
,
Bell System Tech. J.
30
,
530
(
1951
).
10.
A preliminary estimate of $2.00 per Curie for Sr90Y90 at a specific activity of 33 Curies per gram appears in “Industrial Uses of Radioactive Fission Products,” SRI Project 361, Stanford Research Institute, Stanford, California (1951).
11.
It is a convenient mnemonic to recall that this charging effect is the origin of the terms “p‐type” and “n‐type”:
E. F.
Kingsbury
and
R. S.
Ohl
,
Bell System Tech. J.
31
(
4
),
802
(
1952
).
12.
W.
Shockley
,
Bell System Tech. J.
28
,
435
(
1949
);
C.
Wagner
,
Physik. Z.
32
,
641
(
1931
).
13.
Detailed consideration of these matters is beyond the intended scope of the present discussion.
14.
K. G. McKay (private communication).
15.
H. B.
Briggs
,
Phys. Rev.
77
,
287
(
1950
);
H. B.
Briggs
,
Phys. Rev.
77
,
727
(
1950
);
H. B.
Briggs
,
J. Opt. Soc. Am.
42
,
686
(
1952
).
16.
Smithsonian Meteorological Tables (The Smithsonian Institution, Washington, 1951) Sixth revised edition, Table 131.
17.
As observed at Washington: W. E. Forsyth (Editor), Measurement of Radiant Energy (McGraw‐Hill Book Co., Inc. New York and London, 1937), p. 77.
18.
M. Becker, “The Infra‐Red Optical Properties of Silicon and Germanium,” thesis, Purdue University, June, 1951.
A selenium film over germanium or silicon surfaces has been found to reduce reflection loss:
H. B.
Briggs
,
Phys. Rev.
77
,
287
(
1950
).
19.
See reference 6, also
G. L.
Pearson
and
C. S.
Fuller
,
Proc. Inst. Radio Engrs.
42
,
760
(
1954
).
Measured values of I and Wm are 0.018 amp cm−2 and 6⋅10−3watt cm−2 (D. M. Chapin, unpublished).
20.
See reference 10, also R. I. Powers and A. F. Voight, AECU‐840, May 1, 1950.
21.
H. H.
Seliger
,
Phys. Rev.
88
,
408
(
1952
).
22.
J. W.
Weymouth
,
Phys. Rev.
84
,
766
(
1951
).
23.
The transport mean free path κ−1 employed as length unit is evaluated as 4.8⋅10−2cm, and use is made of the dependence of energy E and dE/ds on range s. See Sections 8.3, 8.4, and 8.6 of the Appendix.
24.
The empirical treatment of beta intensity in terms of an absorption coefficient also gives essentially the same electron flux. If energy is not too small, electrons emerge equal in number to those in about 0.2 of a beta range:
R. G.
Baker
and
L.
Katz
,
Nucleonics
11
(
2
),
14
(
1953
).
25.
Brown
,
Fletcher
, and
Wright
,
Phys. Rev.
92
,
591
(
1953
).
26.
E.
Klonte
and
K.
Lark‐Horovitz
,
Phys. Rev.
82
,
763
(
1951
);
E.
Klonte
and
K.
Lark‐Horovitz
,
86
,
643
(
1952
).,
Phys. Rev.
27.
If needed, the Sr90 spectrum may be obtained by a procedure similar to that given.
28.
W. T. Read (private communication).
There is evidence that this cross section may be even larger:
P.
Rappaport
,
Phys. Rev.
94
,
1409
(
1954
).
29.
Theory similar to that of Weymouth (reference 22) can furnish reasonably accurate estimates.
30.
W. Kohn (private communication).
31.
W. Kohn (unpublished). For copper, EL is about 25 ev, and there is evidence that there is not much variation of EL with different materials.
32.
L. Pauling, The Nature of the Chemical Bond (Cornell University Press, Ithaca, New York, 1940), p. 53.
33.
Conyers Herring (private communication).
34.
E. M.
Conwell
,
Proc. Inst. Radio Engrs.
40
,
1327
(
1952
).
35.
C. S.
Fuller
and
J. A.
Ditzenberger
,
Phys. Rev.
91
,
193
(
1953
).
36.
Teal
,
Sparks
, and
Buehler
,
Proc. Inst. Radio Engrs.
40
,
906
(
1952
).
37.
See: reference 16; Smithsonian Physical Tables (The Smithsonian Institution, Washington, 1933), eighth revised edition, pp. 610ff;
and The President’s Materials Policy Commission, Resources for Freedom, A Report to the President, June 1952, Vol. IV, Chap. 15, “The Possibilities of Solar Energy.” Note that a minor effect of load mismatch has been neglected, and that there are no readily available data from which the error that may be involved in the assumption of no change in energy spectrum can be estimated.
38.
W.
van Roosbroeck
,
Phys. Rev.
91
,
282
(
1953
).
39.
W. van Roosbroeck, J. Appl. Phys. (to be published).
40.
H. A.
Bethe
and
R. F.
Bacher
,
Revs. Modern Phys.
8
,
82
(
1936
), Eq. (215b).
41.
E. N.
Jensen
and
L. J.
Laslett
,
Phys. Rev.
75
,
1949
(
1949
);
Braden
,
Slack
, and
Shull
,
Phys. Rev.
75
,
1964
(
1949
);
H. C.
Price
, Jr.
, and
L. M.
Langer
,
Phys. Rev.
76
,
454
(
1949
);
and
L. M.
Langer
and
H. C.
Price
, Jr.
,
Phys. Rev.
76
,
641
(
1949
).
42.
W. Kohn (unpublished).
43.
L.
Katz
and
A. S.
Penfold
,
Revs. Modern Phys.
24
,
28
(
1952
).
44.
Note that the distribution in E0, of total emergent number is given by 12(1−x/s0)B(E0)dE0 per source electron.
45.
F.
Rohrlich
and
B. C.
Carlson
,
Phys. Rev.
93
,
38
(
1954
).
46.
Bethe
,
Rose
, and
Smith
,
Proc. Am. Phil. Soc.
78
,
573
(
1938
). Direct application of diffusion theory is unsatisfactory because it assumes small energy loss in a transport mean free path, which in aluminum is of the order of the residual range after straight‐path transport.
47.
H. W.
Lewis
,
Phys. Rev.
78
,
526
(
1950
).
This content is only available via PDF.
You do not currently have access to this content.