GERMANIUM INFRARED POINT SOURCES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-03300A001600010027-3
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 35
• Document Creation Date: December 22, 2016
• Document Release Date: December 28, 2011
• Sequence Number: 27
• Publication Date: March 20, 1956
• Content Type: REPORT

1
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3 ,(f7
GERI&NIUU INPRARED POINT SOURCES
MARCH 20. 1956
STAT
STAT
STAT
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
INTRODUCTION
It has been previously reported that if an electric
current is brought through a pin point in a semiconductor
material such as germanium, the electrons are attracted to-
wards the point, thereby leaving holes in the semiconductor
material. The recombination of electrons from the negative
electrode with these holes causes the emission of infrared
radiation at about 1.8 to 3 microns.13 This phenomenon is
evident in most semiconductors under the same conditions; i.e.,
the injection of positive holes in a semiconductor of the
M-type or the injection of electrons in a semiconductor of the
P-type will cause this effect.
It has also been previously reported that
germanium is transparent to infrared radiations beyond 1.3
microns, but its transparency can be modulated by an electric
current.15 The modulated transparency of the semiconductor
material is probably caused by the formation of holes, as
described in the paragraph above.
Germanium infrared point sources14 have been develop-
ed that utilize the above described phenomena in devices for
the modulation, transmission and reception of infrared radiant
energy and for the establishment of radiation barriers for the
scanning of an area to sense the presence of objects. The
ease of modulation of these sources coupled with their
efficient infrared output allows their use in devices for trans-
mitting and receiving infrared radiation over relatively long
distances. Because of the very low time constants of the
sources and detectors, the devices will perform even more
satisfactorily than normal telephonic communication equipment.
The narrowness of the infrared energy beam emitted
from the source makes it preferable to transmit the signals
between two fixed points or to utilize special supplementary
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
equipment for aiming the transmitter at the receiver, or
vice versa. This paper will review the historical and
theoretical aspects of the phenomenon and discuss methods
of utilizing improved germanium infrared point sources in
communication devices.
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
HISTORICAL
Lossew 1,2 first observed the injected light
emission from crystals of silicon carbide in 1924. The early
literature3'4 related the light emission to the presence of
an "active layer" on the surface of the crystal. Later the
light emission was attributed to the Bremstrahlung of
electrons.5'6 This theory did not explain the fact that
spectral distribution was nearly independent of the applied
field strength.
Two types of emission were reported (a) a bluish
light, and (b) a yellow light, the type emitted depending
on the direction of current flow. The parts of the crystal
emitting the yellow light did not coincide, in general, with
those emitting blue light or current reversals; nor did the
same crystal necessarily emit both types. It was concluded,
therefore, that the mechanism of exitation differed for the
two cases.
To disprove the "active layer" theory of light
emission,3'4 Lossow removed the layer by grinding and measur-
ed the resistance between two probes placed on the surface at
various intermediate stages of the grinding process.? The
resistance was found to increase by several orders of
magnitude with progressive removal of material and then to
drop sharply after twelve microns had been removed. Probing
the potential distribution while current was passing from the
surface to the crystal showed the presence of a rectifying
junction at the boundary of the active layer and the bulk
silicon carbide. Yellow light emission occured when the
current passed through the boundary in the forward direction.
In a later paper it was mentioned that the active layer con-
sisted of an n-type layer, whereas the bulk consisted of p-
type material.
-3-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
Tetzner investigated the luminescence phenomena
observed during the passage of currents through impure silicon
carbide crystals.8 He noted that their apliearance depended
on the polarity of the applied voltages. When silicon carbide
was the cathode, luminescence occurred over a small area
surrounding the metal electrode and apjUeared to be located just
below the glassy surface. Probe measurements showed this
region to be well-conducting. When the crystal was made the
anode, blue or green luminescence appeared in the form of ir-
regularly distributed pinheads. Anodic luminescence required
higher voltages. Tetzner also investigated the use of
electrolytes as an alternative to metal electrodes.
It was considered that the mechanism of.cathodic
luminescence was similar to that of normal phosphors except
that the luminescence centers were replaced by larger
"luminescent domains". The process was thus expected to depend
on impurity content since its absence in chemically pure silicon
carbide had been confirmed. Both forms of emission had very
short decay periods. .
When a crystal was irradiated-with ultra violet light,
the areas which fluoresced concided with the regions of cathodic
luminescence. The regions of anodic luminescence could not be
similarly excited. Anodic luminescence .was thought to be a
thermal effect due to highly localized power dissipation in
internal spreading resistances.
In 1951 Lehovec, at al investigated the iiijected
light emission of silicon carbide crystals.9 Previously
published data on the intensity and spectral distribution of
the light had been of a qualitative nature. Lehovec, at al
investigated the emission as a function of temperature and of
current through the crystal. They found that the emission
spectrum extended from 4500k to 65008 at room temperature and
was nearly independent of current from 0.1 ma. to 50 ma. The
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
light intensity increased approximately proportionally to
current (efficiency about 106 quanta per electron at room
temperature for a particular crystal.) (See Appendix for a
discussion of the experiments).
New photoelectric devices utilizing carrier in-
jection were disclosed by Lehovec in late 1952.10 The
detection of injected carriers by their absorption and by
emission due to their recombination with majority carriers
was discussed. Two new components were proposed, (1) The
"photo modulator" which permitted modulation of a light beam
by the change in absorption due to injection carriers, and
(2) the "graded seal junction" which permitted light to be
transmitted to and from .a p-n junction without substantial
absorption in the bulk semiconductor.
In 1953 Lehovec published a paper on the light
emission produced by current injected into a green silicon
carbide crystal.11 He discussed the light emission which
arises from the passage of a current in the forward direction
over a p-n barrier in a transparent (pure) silicon carbide
crystal. The results differed from previous results obtained
on a dark blue ("impure") silicon carbide crystal in several
respects. The spectral distribution-of light emission was
found to be independent of current and temperature. Both
the efficiency of light emission and the decay constant of
light emission increased exponentially with decreasing
temperatures. This was explained by a non-radiative transition
of excited electrons. In a dark blue; silicon carbide crystal
the resistances in s eries with the p-n barrier were largely
eliminated, and light emission at a voltage of the same
magnitude as by/e, was observed.
Newman reported on his optical studies of. injected
carriers in 1953.12 He investigated the radiation produced
by the direct recombination of electrons and holes in germanium.
-5-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
Briggs and Haynee13 received a patent on methods and
devices for generating infrared energy. This was accomplished
by creating a high rate of recombination of electron-hole pairs
in semiconductors by injecting foreign charge carriers into the
semiconductor body. (See discussion in Appendix).
-6-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
INFRARED POINT SOURCES AND THEIR APPLICATION
Fig.l - ARRANGE:..-NT FO PRODUCTION OF INFRARED ENERGY FROM
A PIECE OF SEMICONDUCTOR MATERIAL
The Briggs and Haynes infrared emitting device,13
Figure 1, comprised a single crystal of n-type germanium 1,
a point contact to the germanium 2, a base 3, all oo:ineoted
in series with a do current source 4. The positive point
contact electrode caused a migration of electrons in the
germanium towards the positive electrode 2 with a consequent
formation of holes in the germanium 1. Electrons from the
negative eleotrodi 3 then entered the germanium and combined
with the holes. This combination of electrons with the
holes occurred with an emission of infrared energy at ap-
proximately 1.8 to 3 microns,
-7 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
Fig.2 - MODULATING DEVICE FOR SINGL+. POINT CONTACT SWI-
CONDUCTOR SOURCE.
Modulating the infrared energy emitted from the
germanium source was accomplished as shown in Fig.2.
Speaking into the microphone 5 c-used a fluctuation of
current from the do source 6 to pass through the primary
coil of the transofrmer 7. The secondary coil of the
transformer 7 then modulated the output of the do source 8
and the current flowing from the base 9 into the germanium
pellet 10 modulated the electron flow to point contact 11,
thus causing a modulated infrared emission from the germanium
pellet. It was important that the polarity of the device be
sufficient so that the direction of electron flow within the
germanium 10 did not change. It was possible by this method
to modulate the infrared output up to very high frequencies
(in the order of megacycles per second).
- 8 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The radiation emitted from the germanium pellet was
observed to be only in the infrared region. Because there was
no emission of visible light, an infrared filter was not
necessary as in the usual infrared projection devices. The
germanium, being transparent to the infrared radiation formed
by the recombination of holes and electrons, permitted the
escape of the radiation for projection purposes.
Recently semiconductor infrared sources have been
constructed in such a manner so that they have an optimum
geometric shape determined by the general principles of optics
for the projection of fine beams of infrared radiation. The
exit surface of the semiconductor has been shaped to constitute
a diopter approximately stigmatic in relation to the point of
emission to provide for optimum focusing. The infrared
radiation was emitted very close to the point contact but the
combination of the point contact and the stigmatic diopter
shape of the germanium crystal furnished an apparent pin point
source which, when placed at the focus of an optical instrument
such as a mirror, yielded very fins beams of radiation without
a.e:praciable aperture. For some applications, how,ver, it was
advantageous to form the germanium pellet so that the emitted
radiation was not focused.
The optimum operating current for the above described
units was up to 500 milliamperes at 0.5 to 160 volts.
Fig.3 - MODULATING DEVICE FOR TIZIULTIi'L z i'OINT CONTACT SEMI-
CONDUCTOR IR SOURCE
- 9 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The germanium crystal 17 in Fig.3 above, was employed
as an infrared emitter utilizing several point contacts 16 for
the positive electrodes. Several points were used instead of a
single point to increase the infrared radiation output of the
germanium crystal. The points were placed close together so
that the geometric form of the germanium crystal could focus the
radiation emitted from all the points. An oscillator 12 emitted
a signal which passed through transformer 13 and modulated the
current input from the do source 14. Electrons within the
germanium migrated towards the positive point electrodes 16 and
new electrons entered the germanium crystal through the negative
electrodes 15. The combination of the electrons entering the
crystal through 15 with the holes produced by the migration of
electrons already ,in the crystal towards points 16 produced the
infrared energy which was emitted as shown by dotted lines 18.
Fig. 4 Fig. 5
DEVICE FOR PRODUCING A FINE NON-MODULATED BEAM OF INFRARED
ENERGY B TWEEN TWO AREAS OF MODULAT::D INFRARED ENERGY.
- 10 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
In Figure 4 two germanium pellets 19 and 20 were
placed at the focal point of a mirror 23. A signal of 400
cycles per second was emitted by the oscillator 21, thus
modulating the current entering the germanium crystal 19.
Another oscillator 22 emitted a signal of 1,000 cycles per
second, thus modulating the current into the crystal 20. The
modulated infrared energy output of the crystals 19 and 20
impinged on mirror 23. Two modulated infrared beams were
reflected, one at 400 and the other at 1,000 cycles per second.
The intersection of these two beams comprised a narrow beam 24
which was sensed as a non-modulated continuous signal.
Figure 5 shows diagrammatically the areas covered by the
infrared energy reflected by mirror 23 in Figure 4. Area 25
represents the 400 cycle beam. Area 26 represents the 1,000
cycle beam. The continuous non-modulated beam is shown by
line 24.
It was thus possible by the use of the device
represented by Figure 4 to provide an extremely fine beam of
infrared energy that can be used for the alighment of signal
devices.
Fig.6 Fig-7
DEVICE FOR PRODUCING A FINE B!;AM OF NON-MODULATED INFRAR3D
ENERGY B:.:T'WEEN TWO AREAS OF MODULATED INFRARJD ENERGY EMPLOYING
A SINGLE SEMICONDUCTOR UNIT.
- 11 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
In the figure above a single crystal of germanium
29 was placed at the focal point of the mirror 23 instead
of the two crystals of germanium used in the apparatus shown
in Figure 4. The oscillator 30 emitted a signal of 400
cycles per second through point 27 into the germanium 29. The
oscillator 31 emitted a signal of 1,000 cycles per second
through the point 28 into the germanium 29. The modulated
infrared energy emitted from the points 27 and 28 impinged
on mirror 23 and was reflected in the pattern shown in
Figure 7. The device represented by Figure 6 could also be
operated in another manner. The output of oscillators 30
and 31 could be modulated at the same frequency, for example,
400 cycles per second. The output of oscillator 30, however,
would be modulated so that the signal would be transmitted as
short pulses, or dots. The output of oscillator 31 would be
modulated in such a manner so that the signal would be trans-
mitted as long pulses, or dashes.
In Figure 7 area 33 represents the infrared energy
emitted from crystal 29 at point 27 which has been reflected
by the mirror 23. Area 35 represents the infrared energy
emitted by orystdl 29 from point 28 and reflected by mirror 23.
In the case where two different frequencies were
produced, a detector placed in area 33 received a modulated
signal of 400 cycles per second. The detector in area 35
received a 1,000 cycles per second sign&, and the detector in
the area represented by line 34 received a continuous signal.
In the case where the output of each oscillator was the same
but the signal was emitted in dots and clashes, a detector
placed at area 33 would receive dote at 400 cycles per second.
A detector placed in area 35 would receive dashes at 400 cycles
per second, but the detector placed in the area represented by
line 34 would receive a continuous signal.
- 12 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
Fig. 8 - RADIATION PATTERN PROJECT::D BY 3"1ICONDUCTOR IR
SOURCE WITH FOUR POINT CONTACTS
In figure 8 a germanium crystal 37 was placed at the
focal point of the parabolic mirror 36. The crystal had four
infrared emitter points, 38, 39, 40 and 41. By employing the
procedures described in this report, the output of each point
could be modulated at a different frequency or at the same
frequency but with different pulse durations. For example,
the output at points 39 and 41 could be 400 cycles per second
with one point emitting dots and the other dashes. The output
of points 38 and 40 could be kodulated at 2,000 cycles per
second with one point emitting dots and the other dashes. At
the receiving station a high frequency signal (2,000 c.p.s.)
and a low frequency signal (400 c.p.a.) could be detected around
an axis which is a straight line in space at approximately the
axis of the parabolic mirror 36. The signal reflected at the
axis of the mirror would be received as a continuous signal by
a photosensitive cell.
- 13 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
Fig. 9 - CO! UNICATION SYSTEM F2wIPLOYING IR POINT SOURCE
In Figure 9 the infrared output of germanium source
43 was modulated by microphone 44. The emitted infrared
radiation impined on mirror 42 and was reflected on to
mirror 45, then to the photoconductive surface 48 of photo-
cell 47. The resistance of the photoconductive surface 48
changed with the variations in impinging infrared radiation
and the modulated output of the cell was amplified by
amplifier 49. The electrical output of amplifier 49 was then
received by speaker 50.
- 14 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The germanium 43 has an infrared emission spectrum
with a peak at 2.5 microns. There is no visible light
emitted. Lead sulfide cells are sensitive up to 3.5 microns
at room temperature and up to 4 microns at the temperature of
liquid air (-180?C). Lead telluride cells are sensitive to
6 microns when cooled to the temperature of liquid air (-180?C).
Either cell can be used in the device represented by Figure 9
or in any of the devices described in this report employing
infrared responsive photoconductive cells.
It is important to note that the efficiency of the
device represented by Figure 9 is high because no infrared
filter or lenses are necessary for the transmission or
reception of the infrared radiation. In devices previously
reported, an incandescent infrared source such as a tungsten
filament lamp was employed. Filters had to be used in these
devices so that only the infrared energy would be transmitted.
A germanium source, optically shaped, will project a beam
without the filter or lenses necessary in other types of
infrared projection apparatus.
574
Fig.10 - Indium UNDER GE POINT CONTACT
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
It has been previously reported that the deposition
of indium in pellet or layer form on an n-type germanium crystal
directly under the point electrode contact will increase the
hole density in the area adjoining said contact. In Figure 10
an n-type germanium crystal 53 is employed as an infrared
emitter. An indium pellet 51 is placed in or on the surface of
the crystal directly under the point contact electrode by
nechanical means or by electrodeposition. A modulated signal
from oscillator 55 causes the metal at 51 to become more positive,
attracting electrons from the germanium. This will increase the
hole population throughout the germanium, but particularly in area
52. In effect, then area 52 becomes a p-type germanium section.
The inclusion of the pellet Gr layer of indium or other metal with
similar properties under the point contact is advantageous because
the infrared output of the crystal is increased, the units are
easier to manufacture, and the resulting impedance values of the
device are more satisfactory than that of pure germanium for in-
strument-ition purposes
6C
Pig.ll - IR POINT SQURUPLBYIa 3 INDIUM PELLETS
Multiple electrode structures on a single germanium
crystal blank have also been produced. (Figure 11). `Indium
pellets 56, 57 and 58 were not into the hemisphere of the n-
type.germanium crystal 60 directly under the positive point
contact electrodes and the negative electrode 59 was positioned
on the side of the hemisphere. The pellets were placed
sufficiently close to the center of the shaped crystal so that
- 16 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
the infrared radiation was emitted approximately normal to the
crystal surface. The infrared source, composed of the indium
or similar metal between the point contact and the semiconductor
material was employed in some of the devices described in this
report.
63
12 - IR 7 DINT SOURCE - CkkT SIAN OVAL
It may be assumed that the emission area in the
germanium source is almost a pin point. It is then possible
to determine the best procedure for projection of the infrared
energy in a narrow or broad beam. One method would be to use
a hemisphere shaped germanium source in combination with a
mirror which would gather and reflect the infrared energy with-
in a wide angle. The infrared radiation emitted from the
germanium source, could also be focused by a proper shaping of
the germanium crystal.. For example, if emission takes place
at 61 in Figure 12, the portion of the germanium which allows
the exit of a parallel beam from point 61 is a section 62 of
the Jartesian oval 63. With the Cartesian oval section it is
- 17 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
not necessary to use a mirror or lens to project a narrow
parallel beam of infrared energy from the crystal.
Fig.13 - IR POINT SOURCE - SPHERE SECTION
The index of refraction of germanium in the infrared
region is approximately four. Calculations indicate that
with an index of refraction of four, the Cartesian oval des-
cribed above is almost a sphere. Thus the practical solution
in the production of the infrared source was to see a cross-
section 65 (Fi&ure 13 ) of a germanium sphere cut slightly
beyond its center at 67. The point electrode was placed at
66. The plane at which the sphere of germanium was out to
get the desired focus of emitted infrared energy was determined
mat matically or by infrared measurements on the emission from
the sphere as it was being shaped to conform with the optimum
geometric configuration for projection of parallel beams of
infrared radiation.
- 18 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
Fig. 14 - IR POINT 3OJRCE-CONICAL CROSS SECTION OF A CARTESIAN
OVAL
Figure 14 shows the use of a conical cross-section 68
of a Cartesian oval. The point contact 69 with an indium or
similar metal underlay is at, the apex of the cone, permitting
the emission of parallel beams .f infrared energy as represented
by do`-tdd lines 70. This shape allows optimum utilization of
the germanium with a minimum amount of material. The impedance
of the unit is also higher than in the germanium hemisphere,
because of the conic ii shape of the source. The device will
project parallel beams of infrared radiation quite satisfactorily,
even though there may be some loss of lateral radiation at the
pdnt contact.
lie
Fig-15 - MODULATION OF IR POINT SOURCE
- 19 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The germanium infrared source requires a current
of about 600 to 800 milliamperes per point for satisfactory
operation. When an indium layer or pellet is placed between
the point and the germanium, the current requirements are
substantially higher. The germanium source thus requires a
do power supply of the same order of magnitude as that required
for the operation of a standard carbon microphone. Figure 15
shows an arrangement whereby the modulation of current from
the direct current source by the carbon microphone 128 will
produce a modulated infrared signal from the germanium source
129. It is thus possible to construct a very simple germanium
infrared source for short distance communication without
complex electronic, optical or mechanical systems.
- 20 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
A P P E N D I X
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
U.S.Patent 2,683,794 - INFRARED SOURCES
Howard R.Briggsr and James R.Haynes
Assigned to Bell Telephone Laboratories
Patented July 13, 1954.
The invention discloses devices and methods for
generating infrared energy with semiconductor point source
devices. The radiation can be generated in narrow frequency
band widths and in controlled intensities. The infrared
energy is generated by recombination of electron-hole pairs
in the semiconductor.
Only extrensie semi-conductors are considered in
the invention. An a xtrensic somiconduotor has the
characteristic that the conductivity at room temperature is
largely due to the presence of impurity atoms. If these are
of a type which furnish excess electrons, the impurity atoms
are called donors and the semiconductor is known as n-type.
Conversely, if the impurity atoms provide electron deficits
or holes, the atoms are called acceptors and the semiconductor
is known as p-type. If both acceptor and donor impurities
are present, the semiconductor will be either n or p type
depending upon the a xaess of the one impurity over the other
and the conductivity so produced is due either to excess
electrons or holes, but not both. The concentration of
excess electrons or holes due to impurities is a constant at
any given t empertiture , depending only on the composition of
the semiconductor.
This invention requires the presence of both free
electrons and free holes simultaneously in the material so
that direct electron-hole recombination will occur to produce
radiation. High concentrations of free electrons and free
holes are necessary for the production of radiation and, there-
fore,:it is advantageous to use material having an excess
- 21 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
concentration of either donor or acceptor atoms so that the
concentration of either holes or excess electrons is high
prior to the injection of charge carriers of the opposite type
into the material.
Silicon and germanium are described as preferred
materials for the infrared generator in accordance with the
invention.
The crystalline form of both germanium and silicon
is the diamond lattice in which each atom forms four covalent
(electron pair) bonds with neighbouring atoms. The electrons
in these covalent bonds are unable to move in electric fields
of ordinary intensity and thus do not contribute to the
electrical conductivity of the crystal. It was found however
that if visible light was allowed to fall on the germanium or
silicon crystal, some of the valence electrons are ejected from
the bonds leaving behind electron deficits, or positive holes.
Both the ejected electrons and holes are capable of moving in
an electric field so that the conductivity is increased.,
As the wavelength of the light used to irradiate the
crystal is oontinuousl,,r increased, the incident quanta have
progressively less energy. A point is finally reached at
which the incident quanta have just sufficient energy to break
a covalent bond and so produce an electron-hole pair. The
corresponding wavelength is known as the internal longwave
limit of crystal. Careful measurements of the increase in
conductivity and incident light intensity showed that one
electron-hole pair was produced for each photon of light
absorbed for light with quanta having sufficient energy. The
minimum energy required to produce an electron-hole pair and
thereby increase the number of free charge carriers was that
which would transfer an electron having maximum energy in the
valence band to a condition of minimum energy in the conduction
band. This minimum energy was found to be 1.1 ev. for silicon
and 0.7 ev. for germanium. Thus, the internal longwave limits
-22-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
for these materials are about 1.2 and 1.8 microns respec-
tively.
The conductivity which is an index of the number
of free charge carriers present can be increased in a semi-
conductor by the application of forms of external energy
other than light energy. The increase in the conductivity
of intrinsic semiconductors with temperature is usually due
to the breaking of large numbers of valence bonds by thermal
energy.
The conductivity of semiconductors can also be in-
creased by carrier injections. This may be done by causing
a current to flow from a metal point into an n-type semi-
conductor or from a p-type semiconductor into a metal point.
Large concentrations of injected carriers can also be obtained
by using a p-n junction with current flowing from the p to the
n-type material so that electrons are injected into the p-type
semiconductor and holes into the n-type. Since neutralization
of space charges requires that each.ninority carrier be
accompanied by an additional carrier of the opposite side,
extremely high densities of excess electrons and holes can be
produced in the immediate vicinity of the junction. whenever
the conductivity of the semiconductor is increased by one of
the ajove means. The crystal reverts to its original
conductivity when the energy source is removed. Than, if
the voltage applied across the p-n junction is suddenly removed,
the carrier concentration in the neighbourhood of the junction
decays to its original value. This restoration of the
original conductivity implies some process or processes of
recombination of the injected charge carriers and carriers of
the opposite sign present in the material.
The Conservation of Energy Law demands that the re-
combination of an eleotroriole pair be accompanied by the
release of an amount of energy equal to that required to
produce a pair. Conceivably this energy may appear either in
the form of quanta of thermal lattice vibrations (phonons), or
- 23 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
it may be radiated in the form of light quanta (photons),
or both. With carrier concentrations ordinarily obtained
in semiconductor devices of the prior art, only an extremely
small fraction of recombination energy could have been of a
nature which would appear as characteristic radiation. Only
with extremely high concentrations of free carriers can the
chance of direct electron-hole capture be made at all
comparable to that of carrier capture through recombination
centers so that usable amounts of characteristic radiation
result. Electron-hole concentration rates sufficient to
produce usable quantities of substantially single frequency
electromagnetic energy can be obtained by injecting charge
carriers of one type at high rates into properly prepared
and shaped bodies of semiconductor material which normally
contain an excess of charge carriers of the opposite type.
Thus electrons can be injected into p-type material, holes
can be injected into n-type material or an n-p junction can
be utilized by drawing a forward current across it to inject
electrons from the n-section into the p-section and holes
from the p section into the n section.
The narrow range of radiation frequencies associated
with the energy evolved in the dropping of an electron across
the entire forbidden band of the material is referred to as
characteristic radiation as opposed to that form of radiation
having a broad emission spectrum called black body radiation,
resulting from thermal agitation at high temperature.
Various means of injecting charge carriers to a
semiconductor are known and may be employed to generate
characteristic infrared energy. For example, by the applica-
tion of visible light to a semiconductor, or by bombardment
of then semiconductor with high energy particles. In the
following discussion, the term injection is intended to in-
clude excitation of minority charge carriers by incident
-24-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
light or bombardment as well as emission of carriers from
physical contact. Minority charge carrier injection
rates equivalent to that obtained from an emitter at a
current density of at least 10 amp. per square centimeter
is sufficient to produce usable infrared energy.
A crystal of germanium or silicon, having
significant impurity in its various portions, so that these
portions are highly conductive, (i.e. of the order of 0.01
to 0.1 ohm lom-l)is provided with one portion which is of
strongly n-type material and another portion which is of
strongly p-type material. Such an n-p junction can be
employed as an excellent emitter of foreign carriers into
both the n and p type portions of the semiconductor body
biased by a current in the forward direction of conduction,
i.e. where the p-type material is biased positive relative
to the n-type material. The resulting recombination
produces energy at least a portion of which is infrared
radiant energy of a narrow and width. By changing the
intensity of current at the n-p junction, the infrared out-
put is changed, thus an on-off ei nal, or a signal of
modulated intensity is produced.
From the theory of electro-magnetic wave generation
not forth above it is to be expected that the characteristic
recombination in a semiconductor would occur across a fixed
energy gap and a sharp and definite wavelength of energy
would result. The wavelengths of the energy obtained from
characteristic recombination in actual eemiocm duator samples,
extend over a range which may be as wide as a few tenths of
a micron. This apparent anomaly resulting from a lack of
sharp characteristic wavelengths may be partly due to kinetic
energy at the holes and electrons, and partly to variations
in the energy gap of the material actually produced by
.r25-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
thermal vibration of the crystal lattice. It has been
observed that the radiation intensity does not have the
broad wavelength distribution characteristic of a black body,
but is sharply peaked. The peak value is quite close to
the best estimates of the energy required to break a covalent
bond and so produce an electron-hole pair in germanium.
It was found that the radiation intensity increased
as the sample was cooled so that about five times as many
quanta were radiated at the temperature of liquid air as at
room temperature. It is therefore evident that a much great-
er proportion of the recombination energy is radiated at these
low temperatures than at room temperature and that therefore
the device is much more efficient when cooled.
The proceeding discussion indicates that the
infrared output is dependent on the rate of direct recombina-
tion which is proportional to the product of hole density and
electron density. High densities of majority charge carriers
can b_~ provided for in the semiconduotive material by the
addition of significant impurities to the material by known
techniques. Large yaantities of minority charge carriers can
be injected into the semiconductor by employing point contacts
for junction emitters and by passing current from 10 amp. per
square centimeter to at least 1000 amp. per square centimeter.
The physical arrangement of the units should be each
that the temperature rise due to these lh current densities
in kept to a minimum, for example by cooling or by .punting
the unit in an intimate heat transfer relationship with a body
having a large thermal capacity. In this regard one suitable
construction is to form a semiconduotive body containing an
n-p junction with massive ends with which ohmic contacts are
made and providing it with a thin intermediate section having
a large ratio of radiating surface to total volume. The p-n
- 26 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
junction in the section should b e centrally located and the
lengths of section from the junction to the masses should
be comparable to the drift length of the carrier, i.e. through
which half the injected carriers recombine to provide effective
heat conduction to the heat sinks containing the masses.
SUMMARY
It has been discovered that usable infrared energy
of a narrow frequency band can be produced by electron-hole
recombination across the forbidden band in the semiconductors.
This energy can be induced by passing suitable electrical
currents through the properly prepared materials. It is
advantageous in the pperation of semiconductors as infrared
sources that carrier injection concentration be high, that
recombination occur near the surface of the semioonducor and
that it be across the entire energy gap. These characteristics
are obtained by employing good emitters, employing semi-
conducive material having high concentrations of majority
carriers, and employing semiconductive bodies formed so that
a large portion of the recombination occurs near the surface.
Devices of this nature have exhibited high efficiencies based
on the radiated quanta which succeed in getting out of the
sample and the number of electron-hole pairs. They operate
at usable outputs at frequencies having plus lengths of the
order of ten microseconds and they have an essentially
infinite life.
-27-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
ABSORPTION OF INFRARED LIGHT BY FREE CARRIERS IN GERMANIUM
H.B.Brigge and R.C.Fletcher
Physical Review 21, 1342 (1953)
The absorption of infrared light associated with
the presence of free carriers in germanium was measured by
injecting these carriers across a p-n junction at room
temperature. The absorption was found to be proportional
to the concentration of carriers. The absorption as a
function of wavelength showed the same sharp maxima previous-
ly observed in p-type germanium. These bands were found to
change with temperature. An explanation of this absorption
was offered in terms of a degenerate energy band scheme.
The presence of donor or acceptor impurities in
germanium was found to cause additional absorption of infra-
red light of a wavelength greater than the fundamental
absorption limit of 2 microns. 16917,18 Some such absorp-
tion can be expected according to the classical Drude-Zener
theory19 of the optical properties of free carriers which, in
this case, are the holes and electrons released by the ionized
impurities. However, the observed absorption is two to three
orders of :magnitude higher than the Drude-Zener theory would
predict. In addition, for p-type germanium, the absorption
has structure instead of increasing uniformly with the square
of the wavelength as the Drude-Zener theory predicts for this
wavelength range. These observations have lead to the
performance of two additional types of experiments. First,
in order to show that the additional absorption, particularly
in p-type germanium, is really caused by the free carriers, the
excess absorption caused by free carriers which have been
injected into the germanium have been measured. Second, the
absorption by p-type germanium in thermal equilibrium has been
measured as a function of temperature.
- 28 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-03300A001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The first experiment depends upon the injection
of free carriers across a p-n junction when the junction is
biased in the forward, or low resistance, direction.20 if
the light, whose absorption is to be measured, is allowed to
pass through the sample in a narrow beam parallel and near
to the junction, the injected carriers will cause an increase
in absorption which can be measured. 10
A single crystal of germanium containing a junction
was finished to a rectangular block 1.5 1 1.0 X 0.4 em., the
plane of the junction being perpendicular to the long
dimension. The sides exposed to light were finished plane
parallel and polished. The germanium was fitted in a groove
in a bakelite mount. The increased absorption caused by the
injected carriers was determined by modulating the voltage
applied across the ;;unction with a 10 cps rectangular wave
obtained from a motor-driven com.iiutater and detecting the
rectangular wave of transmitted light with either a lead
sulphide photocell or a thermocouple followed by an amplifier
and rectifier. This signal was compared with the signal
obtained by eliminf:,tfng the modulation voltage across the
junction and inserting a light chopper, driven by the same
motor as the oomnutater at the same frequency, and using the
same detection scheme.
Measurements were also made on p-type and n-types
samples in thermal equilibrium (normal). Extinction co-
efficients for these samples were obtained by the conventional
means of measuring the ratio of incident to transmitted light,
and using a value for the reflectivity obtained from a separate
experiment. The carrier densities were calculated from
resistivity measurements using the drift mobilities measured
at room temperature for lattice scattering and corrected for
impurities scattering according to the Conwell-`Neieskopf
formula.21
- 29 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
The temperature dependence of the absorption of
the several p-type eamjOles was measured by putting the sample
in contact with a copper block, which was cooled by an
appropriate refrigerant. Higher temperatures were obtained
by passing heavy currents through the samples. Temperatures
were measured with an iron constantan thermocouple attached
by pressure contacts to the samples.
The simplest way to explain the direct proportional-
ity between the absorption and the concentration of the
carriers is to assume that the extra absorption observed in
the infrared is caused by the free carriers, whether they be
introduced by ionization, chemical impurities, or whether they
are injected across junctions.
Two other possibilities were considered to explain
the observed proportionality. First, the absorption might
be associated with the normal ionized acceptor and donor
impurities. This absorption would be proportional to the
number of such impurities present and would thus explain the
observed proportionalities in normal samples. However, the
fact that injected carriers have the same absorption eliminates
this as a possibility. In addition, experiments were carried
out with compensated samples in which the carrier density was
not proportional to the impurities. In these the absorption
was found to be proportional to the carrier concentration and
not to the impurity concentration.
A second possibility is that the absorption is
a 3soeiated with impurities other than the impurities govern-
ing the electrical resistivity or with crystal imperfections.
These are assumed to introduce an energy level into the for-
bidden band from which either electrons or bodies can be
excited by absorbing light. This possibility is suggested
particularly by the absorption bands observed in the p-type
germanium. In order that this absorption be proportional
-0p-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
to the carrier concentration it must be assumed that the
carriers change the population of these levels. For
exampl% consider the absorption caused by ionizing a hole
trapped on an impurity level. In this case the absorption
will be proportional to the number of empty impurity levels.
If the Fermi level is above the impurity level, the relative
population of levels in the impurity level and in the valence
band will be governed by a Boltzmann distribution. That is,
the number of empty impurity levels in thermal equilibrium
will be proportional to the number of free holes. This would
explain the observed proportionality between the absorption
constant and free holes.
When carriers are introduced by injection, however,
the holes are not in thermal equilibrium, and we would expect
that the ratio of empty impurity levels to free holes would
be considerably different. The fact that the arose section
for injected carriers is the same as that ivr carriers in
thermal equilibrium strongly suggests that the absorption is
not caused by impurities or imperfections. Moreover it should
be pointed out that in order for thb cross section to remain
constant from sample to sam:1e, the concentration of impurity
levels has to remain constant. ,`very sample tried had
essentially the same cross section. This same cross section
was also measured on samples prepared in other laboratories
with different doping materials.
It should be observed that, in order for the absorption
to be ascribed to empty energy levels in the forbidden band, the
cross section for capture of protons by these empty levels is
required to be extremely large.
The authors did not believe that the extra absorption
beyond the fundamental absorption edG?e could be associated with
the normal acceptors or donors because the cross section measured
- 31 -
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
in normal samples was the same for (1) injected carriers,
and (2) compensated samples. They did not believe that it
was associated with impurities or imperfections because (1)
the cross section was the same for injected carriers, (2) the
concentration of such uncontrolled impurities was required to
be the same for different samples and (3) too large a cross
section was required. They believed the most reasonable
explanation for the extra absorption was that the big carriers
themselves were absorbing the light by some mechanism other
than that proposed by the simple Drude-Zener theory.
-32-
I Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3
REFERENCES
1. Loseew, 0., Wireless World and Radio 2719 93 (1924)
2. Loseew, O.,Z.Fernmeldeteohnik, .11 97 (1926)
3. Lossew, O.,Philoeophical Magazine, 6, 1028 (1928)
4. Loesew, O.,Physik Z., 30, 920 (1929)
5. Lossew, 0., Physik Z.,, 695 (1931)
6. Claue, B., Ann.Physik, 11, 331 (1931)
7. Loesew, 0., Compt.Rend.Acad.Soi.,U.R.S.S.,2Q, 363 (1940)
8. Tetzner, H., Zeitechrift ftur Angewandte Physik, 1, 153(Sept.194e'
9. Lehoveo, K., et al, Physical Review, ?j, 603 (1951)
10. Lehoveo, K., Proceedings of the I.R.E. AO, 1407 (1952)
11. Lehovec, K., Physical Review, 84, 20 (1953)
12. Newman, R., Physical Review 21, 1313 (1953)
13. U.S.P. 2, 683, 794, July 13, 1954, H.R.Briggs and J.R.Haynee
14. Rocard, Yves, Ecole Normal* Superieure, Paris, France,
Psrsoi:al Communication
15. Briggs, H., and Fletcher, R., Physical Review, 21, 1342 (1953)
16. Fan, H.Y. and Backer, M., Proceedings of the Reading Conference,
132-147, Butterworthe Scientific Publications,London (1951)
17. Briggs, H.B., Physical Review, U. 727, (1950)
18. Briggs, H.H. and Fletcher, R.C., Physical Review 8 , 1130 (1952)
19. Seitz, C.F., ""Modern Theory of Solids", p.39, McGraw Hill Book
Co.Inc., New York (1940)
20. Shockley, W.P., Bell ''ye-,ems Technical Journal, 28, 435-489
(1949)
21. Conwell, E.M., Proceedings of the IRE;, 40, 1330 (1952)
- 33-
Declassified in Part - Sanitized Copy Approved for Release 2011/12/28: CIA-RDP78-0330OA001600010027-3