RCA_PhototubesAndPhotocells_PT-60_1963

Multiplier Pho#otubes 49
►oo
~~—~`
a ~~
i\
6
\ ~~ / ~ MULTIALKALI AND
a \~ --- ~~ QUARTZ

az ( / \ \
3 O \
S-5~~` `
E ~,,,~ \\\
1
F \~ 5-13 \
10 \
`y
ws
m I `1, ' r\\
J
K6 • 1 1
F- yy
• I~
awW
•e + ~ y`1
4 •
~~► 1y
e ~~
Iy
• ` y
~
e ,
~`
Te+QUARTZ '' y
2
~

~ Cs—Te+LiF 1

0.

( 2000 3000 4000 5000 6000 7000
1000

WAVELENGTH—ANGSTROMS

Fig. 41. Spectra] response characteristics of various photocathode-window combinations nsefut

for the ultraviolet: S-5 (Cs-Sb with 9741 glass) ; S-13 (transmission type Cs-Sb with
quartz) ; Cs-Te with quartz ; Cs-Te with LiF window ; multialkali with quartz.

greatest. In the blue region, which the angle of incidence of light on
is most important in scintillation the cathode. The effect is compli-
counting, the maximum effect is of cated, but to a first approximation it
the order of 2.5 per cent (assuming is possible to explain the increased
10 per cent transmission, identical red sensitivity with angle of inci-
cathode and first-dynode sensitivity, dence in tubes having 5-11 response
and a secondary-emission ratio at (shown in Fig. 42) by the increased
the first dynode of 4:1). The effect
absorption of red light with greater
would be negligible for dynodes
angle of incidence.
made of silver-magnesium or copper-
beryllium. The effect of temperature on
spectral response is minor and oc-
The spectral response of a curs primarily in the red region of
phototube is somewhat sensitive to the spectrum,l' as shown in Fig. 43.

50 RCA Phototube Manual

At higher temperatures, the distri- current is amplified in the same way
bution of electrons in the semi- as the photo-originated electrons.
conductor photocathode is shifted to
higher energy levels arid, conse- Usually, only the dark current in
quently, more electrons ale emitted the anode circuit of the tube is of
by the near-threshold energy radia- importance to the observer. This
tion. anode dark current is described in

Dark Current this section as the "dark current" of
the multiplier phototube.
Even though a multiplier photo-
tube is in complete darkness, elec- The dark current and the re-
tron currents may be observed to sulting noise of the multiplier photo-
flow from dynode to dynode. These tube is of particular concern because
currents increase in magnitude to- it is usually the critical factor in
ward the anode end of the tube since limiting the lower level of light
any electron component of the dark detection. It is important to under-
stand the variation of dark current
in the multiplier phototube as a
function of various parameters, in

io I
`~8♦ 1
8 - TYPE 8058
S.SUPPLY VOLTS=1000 l~~ i
\
4
~,.~— B - 85'
2 \

103 '. ~_____-~~. ~
e
1
6
B=o"
~_
4 ~
~
F-

N
Z
W 2

N

F2
W 10

0e

a

6

W
4

!-
a

J

W 2
K

10

e
6

4

2

I 4500 5000 5500 6000 6500 T000 7500
4000

WAVELENGTH —ANGSTROMS

Fib. 42. Spectral-response characteristics for a Cs-Sb transmission-type photocathode (S-11) for
two different incident angles of radiation. The increased red response for the large

angle of incidence may be explained in part by the increase in absorption due to the
longer path ; compare the absorption characteristic shown in Fig. 3S.

IYuJtiplier Photo#ubes 51

i.o

o.s

v 06 n z vz i
w

0.4

N 0.2

a0 e

N
W
O:

F

W 0.2

V

0.4

0.6

—0.8 3000 4000 5000 6000 7000 8000
2000
WAVELENGTH —ANGSTROMS

FSg. 43. Temperature cceffieient of Cs-Sb cathodes as a function of wavelength; data taken at
20 degrees centigrade. Note the Iarge positive effect near the threshold.

order to realize the ultimate in low- may also prevent external arc-over
light-level detection. Although all of resulting from high voltage.
the peculiarities of dark-current
production in the multiplier photo- Ohmic leakage is the predomi-
tube are not well understood, the nant source of dark current at low-
possible sources of dark current are voltage operating condition. It can
be identified by its proportionality
described below: with applied voltage. At higher volt-
ages, ohmic leakage is obscured by
Dark current in a multiplier other sources of dark current.

phototube may be categorized by Fig. 44 shows the typical varia-

origin into three types: ohmic leak- tion of dark current of a multiplier
phototube as a function of applied
age, dark or "thermionic" emission voltage. Note that in the mid-range
of voltage, the dark current follows
of electrons from the cathode and the gain characteristic of the tube.
The source of the gain-proportional
other elements of the tube, and re- dark current is the dark or thermi-
onic emission of electrons from the
generative effects. photocathode and the first-dynode
stage. Because each electron emitted
Ohmic leakage, which results
from the imperfect insulating prop- from the photocathode is multiplied
erties of the glass stem, the support-
ing members, or the plastic base, is by the secondary-emission gain of
always present. This type of leakage the tube, the result is a unipotential
is usually negligible, but in some output pulse having a magnitude
tubes it may become excessive be- equal to the charge of one electron
cause of the presence of residual multiplied by the gain of the tube.
(There are statistical amplitude
metals used in the processing of the variations which will be discussed
photocathade or the dynodes. Con- later.) Because the emission of
densation of water vapor, dirt, or thermionic electrons is random in
grease on the outside of the tube time, the output dark current con-
may increase ohmic leakage beyond sists of random unidirectional pulses.
reasonable limits. Simple precau- The time average of these pulses,
tions are usually sufficient to elimi- which may be measured on a do
nate this sort of leakage. In
unfavorable environmental condi-
tions, however, it may be necessary
to coat the base of the tube with
moisture-resisting materials, which

52 RCA Phototube Manual

Los UNSTABLE REGION to operate the multiplier phototube
in the range where the thermionic
a REGENERATIVE component is dominant. In this
range, the relationship between sen-
IONIZATION EFFECTS ~. sitivity and noise is fairly constant
as the voltage is increased because
z ~~~ both the photoelectric emissicn and
the thermionic emission are ampli-
o.la SUM OF OHMIC-LEAK- fied by the same amount.
- AGE AND AMPLIFIED
sB -_ THERMIONIC-EMISSION At higher dynode voltages, a
0 a CURVES regenerative type of dark current
develops, as shown in Fig. 44. The
f ~~ OHMIC dark current becomes very erratic,
LEAKAGE= and may at times increase to the
F2 practical limitations of the circuit.
Continued flow of large dark cur-
~Z o.ol rents may cause damage to the
sensitized surfaces. Some possible
~s causes of the regenerative behavior
will be discussed in more detail later.
s All multiplier phototubes eventually
become unstable as the gain is in-
Uq creased.
Y
The best operating range can
~0a 2 AMPL FIED generally be predicted from a con-
sideration of the ratio of the dark
w 000a1 THERMIONIC current to the output sensitivity of
EMISSION the tube. This ratio, known as the
o equivalent anode-dark-current in-
put (EADCI), is shown as a func-
~6 tion of luminous sensitivity in Fig.
45. The EADCI is equivalent to the
aa light flux on the photocathode which
would result in an output-current
00001 change equal to the dark current
observed. If thermionic emission
0 20 40 60 80 100 120 140
VOLTS PER STAGE

Fig. 44. Typical variation of dark current
with voltage fora multiplier phototube.

meter, is usually the principal do
component of the dark current at
normal operating voltages. The limi-
tation to the measurement of very
low light levels is the variable char-

acter of the thermionic dark-current

component. It is not possible to bal-
ance out this wide-band noise com-
ponent of the multiplier phototube,
as it might be to balance out a
steady ohmic-leakage current. Nev-
ertheless, it is usually advantageous

a 10-I 10'2 10-I LUMINOUS SENSITIVITY—AMPERES/LUMEN 10q 105
10- ~ 102 I 10~ 102 .,,103 108
z
f-

rc

'c,i 10-u 10-8

aY. N

p~ W

Q~

O3 J

a 10_18 10 9
'- s c

wq

J

22

10_I 103 2 4 6 8105 106 107
W X 10
10q

AMPERES/WATT

Fig. 45. "Equivalent anode dark current input" as a function of the luminous sensitivity for
carious multiplier phototubes. The Equivalent Anode Dark Current Input represents the
light flux which would result in an output current change just eyual to the dark current.
Optimum operating range is usually where this function is near a minimum.

Multiplier Phototubes 53

were the sole source of dark current, were compiled, the thermionic emis-
the EADCI would be a horizontal sion shown is probably representa-
line on the graph. In some tubes a
regenerative condition sets in before tive of the best achievable at the
the really flat operating region has
been attained. present state of the art.

As may be expected, the thermi- The variation of dark current or
onic component of the dark current noise with temperature is most im-
is very much a function of the portant for ultimate low-light-level
temperature. Fig. 46 shows the sensitivity. Various cryostats have
temperature variation of the equiv- been designed to take advantage of
the reduced noise at low tempera-
alent of the dark current at the ture:~ An important practical con-
sideration at low temperatures is the
photocathode (anode dark current
prevention of condensation of mois-
divided by gain) per unit cathode ture on the window. In a Dewar-type
arrangement, condensation is not a
area. The data for this figure were problem; in simpler set-ups mois-
obtained from a number of tubes ture condensation may be prevented
having well defined (flat) minima of by a controlled low-humidity atmos-
the EADCI curves. In some cases, phere at the external window.
the photocathode equivalent of the
In most multiplier Phototubes,
output dark current was derived the electrostatic potential of the
from noise measurements because of walls surrounding the photocathode
the difficulty of separating the leak- and dynode-cage region is important
age from the electronic component particularly with respect to the onset
of dark current at low temperatures. of the regenerative dark-current
(The relation between thermionic
emission and noise is discussed component. The bulb wall can be
later.) The data shown in Fig. 46
represent the actual dark emission maintained near photocathode po-
per unit area associated with par- tential if the bulb is wrapped or
painted with a metallic coating
ticular photocathode types rather maintained at cathode potential; the
than with particular tubes. Because connection of the metallic coating to
of the manner in which the data the cathode is usually made through
a high impedance to avoid shock
W z 0~ o hazard. A positive potential on the
W 0
Flole bulb wall can cause noisy operation.
~6 ~W ~
Even though the bulb is not con-
W sl 06 aW ~~- nected to a positive potential, the
proximity of a shield or container
Wz SII at positive potential may lead to the
aio-le S20
2Z development of a positive charge.
~6
Fig. 47 shows the effect of various
N ~W
bulb-shield potentials on the dark
W _z OS ci~W noise. This effect may not be ob-
served for all tubes and all types,
N10 IB w~ but should be recognized as a pos-
J sible source of increased noise.
6 W Q
Excess noise or dark current is
K 9 IO often accompanied by fluorescent
W effects on the inner surface of the
Oy bulb. When the potential of the
f2
IO, ya bulb is positive, stray electrons at-
alole
W tracted to the bulb cause the emis-
OB Y sion of light on impact, depending on
~, 4
~3 O the nature of the glass surface and
~2
the presence of contamination. Sec-
WIC 1a6 ondary electrons resulting from the
Y impact of the stray electrons on the
-6
0a 4
3.0 3.4 3.8 4.2 4 6 50
RECIPROCAL TEMP. —(103/T)—• K

60 20 0 -20 -60
TEMP.—•C

FiP. 46. Temperature variation of dark cur-
rent for multiplier Phototubes in the
manner of a Richardson plot

54 RCA Phototube Manual

TU6E TEMPERATURE=25° C

NW N Iy (SIGNAL)
2 ZO~W
2000 100
H ~J
F

10-12 a ~ V

~~ s _> a 1500 75 w
~z w a
o- N
6W 01000
Z r

-1000 -800 -600 -400 -200 4 500
EXTERNAL SHIELD VOLTS 0

RELATIVE TO ANODE VOLTS 50 wi
0
Fig. 47. Effect of external-shield potential NOISE 0
on the noise of a 1P21 multiplier photo- z
tube. Note the desirability of maintain-
25 a
ing anegative bulb potential.

glass surface are collected by the 0
most positive elements in the tube
and help maintain the positive po- 0 50 100 150 200 250 300
tential of the inner surface of the FOCUS-SHIELD POTENTIAL
glass. Under these circumstances, it
(RELATIVE TO CATHODE)—VOLTS
is possible to observe the formation
Fig. 48. Typical variation of noise output
of glowing spots on the inside of the of a 6342A multiplier phototube as a
function of focus-shield potential. Also
glass bulb, provided the eye is dark- indicated is the output signal current
adapted and the applied voltage is showing how only a minimum loss in
sufficiently high. Some of this emit- signal sensitivity results from a choice
ted light may reflect back to the of shield potential which minimizes
cathode and cause regenerative dark noise.

current. is= (27)
Other surfaces within the multi-
l+bt +i°
plier phototube are also important
in the control of regenerative dark where is is dark emission as a func-
current. In the 6342A, for example, tion of time, a and b are constants,
the focus-shield potential shows a t is time, and i° is the constant level
point of distinct noise minimum,1° as of dark emission. However, the sig-
shown in Fig. 48. The mechanism of nificance of this relationship is not
this behavior is not understood at apparent at present. It is possible
present.
that the heavy initial exposure to
Another phenomenon which de-
serves mention in connection with N 106 O-13 N
dark current is the effect of previous X10
WU
exposure to light. If a photocathode o0 lo
O 14 oHZ~W
is exposed to strong light, with or U _
without an applied voltage, a meas- ~ 103
urement made immediately after- 0pl102 ID 15 U~W
ward shows a higher-than-normal
dark current which decreases rapidly ~ to 1 10 102 103 1016 Z ~Y
with time. The effect is more marked TIME —MINUTES
when the exposure is richer in short- W to17>Qo~~y
wavelength radiation. Very long wI
periods are required to attain the s
low dark current associated with 0.1
equilibrium, as shown in Fig. 49. 10 18 w w
0
An analysis of the rate of decay
of the dark emission suggests the 104
following reciprocal relationship:
Fig. 49. Variation of dark current following
exposure of cathode to cool-white fluo-
reseent-lamp radiation. The various
cathodes are identified by their apectral-
response symbols.

Multiplier Phototubes 55

light alters the Fermi level of the For an excess impurity semicon-
photocathode as a result of the intro- ductor where thermionic emission
duction of excited states and that originates from the impurity centers,
the equation for thermionic emission
the photocathode then decays to its is written as follows:

initial state. For best operation and 4~ emk2T$ 1 s h°/°
low noise, therefore, it is recom-
(E.+R/2) a (31)
mended that the photocathode be e kT

kept dark at all times, or at least where n° is the impurity concentra-
for many hours before making low- tion and R is the depth below the
conduction band to the impurity
level measurements. level.l~ Photoemission still originates
Origins of Dark Current. In a primarily from the valence band.

metal, the electrons which escape as In Fig. 46, the slope of the
thermionic emission are generally
from the top of the conduction band curve of cathode dark current as a
(see Fig. 1). Thus, the work func- function of reciprocal temperature
tion for photoemission and thermi- has approximately the magnitude
onic emission is the same. Thermi- predicted, but it is difficult to explain
onic emission as a function of work
function ~ in volts and temperature the change in slope. Perhaps there
T in degrees Kelvin is given by the
familiar Richardson equation: are "islands" of different impurity
level or concentration, or it may be
4,r emk9T' —~ e/kT that the impurity concentration is a
j= h° a function of temperature. Exposure
(28) to temperatures above the normal
operating range sometimes results
where j is the thermionic current in permanent reduction in dark
density; e, the electron charge; m,
the electron mass; k, Boltzman's current.
constant; and h, Planck's constant.
In MKS units, the equation becomes Another source of dark emission
from the photocathode results from
— ~ e/kT
radioactive elements in the tube or
For semiconductor photocath-
odes, the work functions of photo- surroundings which cause scintilla-
emission and thermionic emission tions in the glass envelope of the
may be quite different. The work tube. For example, it is very difficult

function for photoemission (see Fig. to obtain glass that does not contain

6) is the potential height from the potassium, which has a natural com-
ponent of radioactive KSO. Some
top of the valence band to the glasses, such as fused silica, have
vacuum level, or E. (the electron comparatively low background radi-
affinity) plus Eg (the forbidden gap, ation.
i.e., the separation of valence and
conduction bands). For an intrinsic Noise
semiconductor, thermionic emission
The amplification of a multiplier
originates from the valence band, as
phototube is usually high enough so
does photoemission but the "work
function" is not the same as for that the noise in the dark current
photoemission. In the case of an completely dominates coupling-re-
intrinsic semiconductor, thermionic- sistor noise except for very small
emission density can be expressed as resistances. Under normal operating
conditions, the noise which limits
4,~ emk4T' (E. +E2 / e detectability results from amplified
kT (30) thermionic emission. Thermionic
7= h' e emission may originate from both
the photocathode and the dynode
surfaces. The latter emission is
usually negligible as a result of the

56 RCA Phototube Manual

difference in gain through the tube. the operation of multiplier photo-
If the multiplication of the second- tubes is the ratio of the signal to
ary emission is assumed to be noise- the noise. The output signal current
free, the expression for the rms noise Is may be expressed as follows:

current is similar to that for a IB=µFR (35)
vacuum phototube, Eq. (22), as fol-
lows: where F is the flux in lumens on

I' of = µ L2 a is ~f] "' (32) the photocathode and R is the sensi-
tivity in amperes per lumen. (F and
where µ is the amplification factor R can also be expressed in terms of
of the multiplier phototube and the watts instead of lumens.) The fol-
other symbols are as defined pre- lowing simplified expression may be
viously. As in the case of vacuum written for the ratio of the do signal
phototubes, it is possible to calculate current to the rms noise current in a
the minimum value of coupling re- bandwidth of:
sistance Rwhich may be used. If it
is assumed that the tube-generated IB FR (36)
[2 a is nf] '~
noise is just equal to the coupling ~
I2 ~f
Johnson noise, as in Eq. (24), the
In the detection of low light
minimum value of R is given by levels, it is often advantageous to
modulate the light by means of a
— 2kT (33) "chopper" and to couple the multi-
R µ e is plier phototube to an amplifier hav-
ing a narrow bandpass at the
For example, if the equivalent cath- chopping frequency. In this way the
ode dark current is is 10-'a ampere do component of the dark current is
and the gain of the tube is 108, the eliminated, and the inherent signal-
value of R for which the two noise to-noise ratio of the multiplier
sources are equivalent is 50 ohms. phototube is more readily realized.
Thus, the multiplier phototube is far
superior to the vacuum phototube For example, when the modula-
for the amplification of very small tion of the light is sinusoidal, a
light signals. The possibility of very modulation factor M can be defined
low values of coupling resistance as the peak-to-peak cathode photo-
permits the observation of high- current amplitude divided by the
speed phenomena not possible with average cathode photocurrent. The
vacuum phototubes having large rms output-signal current can be
coupling resistances.
expressed as
The above discussion does not
consider the increase in noise result- itMµ
ing from the secondary-emission
2~2
amplification mechanism. If this
where it is the average photocathode
source is included, a more refined
expression may be written:18 current. If the average cathode cur-

g '.`~ rent is small compared with the
= N[2eia(1~"m-1)Of,
I' of~ equivalent cathode dark emission id,

(34) the following expression may be

written for the signal-to-noise ratio

S/N:

S/N — rms modulated signal current
rms noise current
where m is the secondary-emission
ratio per stage (usually of the order Miti
of 4) and B is a statistical factor
(found by measurement on a 5819 to 4[eia( 1-~ B 1'~.
be 1.54.18) m-1
/ of J
An important consideration in
(37)

Multiplier Phototubes 57

It is frequently advantageous to not clearly understood. In the case of
rate a multiplier phototube by its tubes such as the 931A and 6655A,
equivalent noise input or ENI. This which have CsSb dynodes, thermi-
figure is the amount of light in onic emission may originate from the
lumens (or other radiation units) dynodes as well as from the cathode.
which produces an rms signal cur- This explanation could well account
for the many small pulses. On the
rent just equal to the noise current other hand, it has been speculated
that the distribution of secondary
in a bandwidth of one cycle. For ex- emission itself may be exponential.
ample, if a square-wave modulation This point of view is proposed by
is assumed for which the peak-to- J. A. Baicker ~ who also suggests
peak amplitude is just equal to the that the two-slope characteristic
unmodulated current with on time shown in Fig. 50 is the result of
equal to off time, a modulation fac- single and multiple emission of elec-
tor M of 8/~r can be assumed in Eq. trons from the cathode. The multiple
(37). If F is the unmodulated flux, emission may be the result of impact
the average cathode current i=, is by positive ions. However, the evi-
dence for this assumption is not as
then given by yet clearly established. The statis-
tics of secondary emission are dis-
it = FR/2 (38) cussed further below.

For the case where S/N is equal to Io5
unity, ENI is equal to the unmodu-
lated light flux F. The value of ENI lo"
is then determined as follows:
Io3
S =1
N Z

(ENI)R 0
c~ 102
B ~ =~z
to
La eis (1 -♦- m-1 / 1
(39) 20 40 60 80 100 120
BIAS VOLTS
or
Fig. 50. Pulae-height distribution o1 dari~
g 1 ~. current pulses in two 7264 multiplier
Phototubes. The data were obtained 1>9
L~ e is (1-I-m-1/ 1 integral-bias counting.

ENI — Noise in the Signal. When the
R
(40) photocurrent is well in excess of the

Note that this equation may be used thermionic emission, measurement
to determine the equivalent photo- precision is limited by the random-
cathode dark current. For example, ness of photoemission and secondary
when B equals 1.54 and m equals 4, emission. This type of limitation is
the value of is is given by most prevalent in applications such
as the detection of a star against the
is - 0.4 X 1078 R2 (ENI)' (41) background of the sky, where the
modulated signal is produced by
When the dark current is ob- scanning .back and forth across the
served on a wide-bandpass oscillo-

scope, it consists of unidirectional

pulses of variable amplitude. It is
presumed that these pulses represent
thermionic electrons from the cath-
ode amplified by secondary emission.
The distribution of the heights of
these pulses is quite closely expo-
nential° with a trend to a double-
slope characteristic

The cause of this distribution is

58 RCA Phototube Manual

star and in the detection of small S/N. _ M It E 1/a
marks on scanned paper. The expres-
sion for the rms noise current output 4 B
is identical to Eq. (32) or to Eq.
(34) (when secondary-emission sta- e l+ m-1 Of
tistics are included), except that the J(44)
average cathode photocurrent it is
substituted for ia. When the modu- This equation may be used to pro-
lation of the light is sinusoidal and vide an approximate measure of the
the magnitude of the modulation is
small compared with the background, collection efficiency, or at least a
the expression for the rms noise cur- relative comparison between tubes'
rent in the signal, N., may be deter-
mined by a development parallel to The noise in the signal is simi-
that of Eq. (37), as follows: lar to the dark noise in that it also
consists of random pulses of variable
S/N. =rms modulated signal current amplitude. The random spacing of
rms noise current in signal
the pulses corresponds to the basic
it ~z
randomness of the emission of
M e (1 B l of photoelectrons. In this case pulses
4 + m-1 / originating from the dynodes are
negligible. The distribution of anode
(zv) pulse heights has been measured at
the Lawrence Radiation Labora-
In an application of this type, which tory" for single-photoelectron in-
requires maximum sensitivity in the puts; data obtained are shown in
presence of a light background, the Fig. 51. The curve passes through a
multiplier phototube used should maximum for small pulses instead of
have high cathode sensitivity. Gain increasing indefinitely near zero
is unimportant except at the first pulse height, as the dark-current
stage, where high dynode-No. 1-to- distribution apparently does. The
cathode voltage is required to mini- distribution is approximately that
mize noise from the statistical vari- calculated from Poisson statistics.
ation of secondary emission (through
the factor m). ---90% CONFIDENCE INTERVALS

In most practical multiplier ~ zo ze-
phototubes, some of the photoelec-
J_ -1
trons fail to enter the secondary- m

emission section of the tube as a m 15
result of imperfect design or mis- 0
alignment of tube components. The rc
collection efficiency for the photo- a
electrons is usually near unity, but
in some tubes may be of the order of ~ to
0.5 or less. If this consideration is
included in Eq. (42), the cathode a
current must be the collected current
rather than the emitted current. The J
collected current may be defined as
follows: ~S

it (collected) — it (emitted) X E (43) ~~ -_~_

where a is the collection efficiency. 5 10 5 20 25 30
Eq. (42) then becomes
RELATIVE PULSE AMPLITUDE

Fig. 51. Measured amplitude distribution of
anode pulses due to single phot-oelectron
inputs for a 2-inch diameter, 14-stage
multiplier phototube having C4RSb dy-

nodes. Gain per stage is approximately 3.

Scintillation Counting~'-°~ An-
other type of application in which

Multiplier Phototubes 59

the statistics of operation of the The main peak of the curve at the

multiplier phototube are important right is associated with mono-
is scintillation counting. In a typical energetic gamma rays which lose
application, nuclear disintegrations their entire energy by photoelectric
produce gamma rays which cause conversion in the crystal. Pulse-
scintillations in a crystal such as height resolution is measured as the
NaI(Tl). A multiplier phototube width of this distribution peak at its
coupled closely to the face of the half height divided by the pulse
crystal converts the scintillations to height at maximum.
electrical output pulses. Because the
energy of a light flash is closely Pulse-height resolutions meas-
proportional to the gamma-ray en- ured in this manner vary fx•om 6
ergy and because multiplier photo- to 20 per cent. Multiplier phototubes
vary considerably in their ability to
tubes are linear in operation, the resolve scintillation pulses of differ-
electrical pulse height can be used ent heights. Good optical coupling
as a direct measure of the gamma- is required to use all the light from
ray energy. However, the number of the scintillation effectively. This re-
photoelectrons per scintillation is quirement makes it necessary to
relatively small (of the order of sev- provide the tube with a semitrans-
parent photocathode on the window-
eral hundred). The output pulses faceplate and a scintillating crystal
vary in height because of the sta- coupled directly to the faceplate.
tistics of the small numbers and
because the scintillations themselves High and uniform photocathode sen-
vary. An important requirement in
nuclear spectrometry is the ability sitivity is essential, especially in the
to discriminate between pulses of
various heights; hence, the im- spectral region corresponding to the
importance of Pulse-Height Reso-
lution. blue emission from the crystal. Fig.
53 compares the distribution of light
Measurement of the pulse- from the NaI(Tl) source and the
height resolution of a multiplier
phototube has not been standardized; 5-11 spectral response commonly
however, a NaI(T1) crystal and a
Cs~" source of gamma rays are gen- used in the coupling multiplier. It
erally used as a reference combina-
tion. Atypical pulse-height distri- to /~ 0
bution curve is shown in Fig. 52.
~e /~ e~a
/~ Ni
s 0
/ 11
q2vWi 4 6 a~
0 _RESOLUTION• 98685 5 2100% i- I~r~~ 2W
JW 2
4000 SCOO 6000 4W
0 Q
3000
2W

0

_ •12.1 WAVELENGTH—ANGSfROM3
O

03D Fig. 53. Distribution of light from a scin-
wV tillating NaI (TI) crystal compared
w with the typical (5-11) spectral re-

FULL WIDTH AT HALF sponse of multiplier phototubes most
F 20 ~ ~ MAX.=92-61.5=10.5 UNITS commonly used in scintillation counting.

0 PEAK MIDPOINT
U POSITION

10 •86.5 UNITS -

20 40 60 RO loo is also important that electrons
PULSE HEIGHT emitted from the cathode be effi-
ciently used by the first dynode.
Fig. 52. Distribution of pulse heights ob-
Pulse-height-resolution measure-
served in a scintillation counting ex-
periment using gamma rays from Cs~,, ments are a reliable guide to efficient
t;o excite a NaI (Tl) crystaL operation in scintillation counting.
It should be noted, however, that

60 RCA Phototube Manual

pulse-height resolution is not solely photoelectric conversion.
determined by the characteristics of
the multiplier phototube; the prop- It is generally desirable to have
erties of the scintillating crystal, its a relatively flat plateau which ex-
.housing, and the coupling to the tends for several hundred volts. A
phototube and the location of the good plateau characteristic is par-
gamma-ray source are also im- tially determined by such properties
portant. as low dark current and high photo-

Another characteristic used to cathode sensitivity; it is also de-
termined by the particular amplifi-
describe the effectiveness of multi- cation-voltage characteristic. Thus,
plier phototubes for scintillation a rapid variation of gain with volt-
counting is the so-called Plateau age results in a short and steep pla-
Characteristic. Although the term teau; this effect is not inherently
is widely used, there is no ac- undesirable, but merely corresponds
to a scale change. Plateau charac-
cepted definition for plateau; adop- teristic should not be used for indis-
criminate comparison of different
tion of this concept may be traced types of multiplier phototubes. Pulse-
to the parallel use of scintillation height-resolution data are a more
counters and Geiger counters. Pulse- fundamental guide to the choice of
height resolution is a preferred multiplier phototubes than plateau
measure of scintillation counting effi- characteristics. Similarly, a pulse-

ciency; however, because of the in- height distribution provides greater
terest in plateau characteristics for
certain applications, a brief descrip- insight to the source of the scintil-
tion of the general implication of
the term is given below. lations than the integral-bias-type

The plateau characteristic is ob- plateau curve.
tained in the same manner as pulse-
height-resolution data, except that When a multiplier phototube is
the pulses are recorded by integral- used to observe very short light
bias rather than differential-bias.
The number of pulses larger than flashes, such as those which occur
a particular value is plotted as a in scintillation counting, After-
function of the voltage applied to Pulses~° (i.e., minor secondary
the multiplier phototube. The plateau pulses following the main anode-
current pulse) are sometimes ob-
which develops (Fig. 54) corre- served. Two general classes of after-
pulses are characterized by the time
sponds not to the valley to the left between the main pulse and the
of the photopeak in Fig. 52, as might
be expected, but to the region of after-pulse: (1) delays of the order
the curve at the extreme left in of nanoseconds; (2) delays of the
Fig. 52 just before the sharp up- order of microseconds. The former
turu.~' Operation on the plateau have been shown to be the result of
corresponds to the counting of scin- feedback of light to the photo-
tillation pulses originating from cathode. The fact that light is gen-
Compton-scattering, as well as from erated in the output stages of the
tube has been verified by many ob-
03 servers, especially in tubes in which
the construction is open enough to
2 PLATEAU permit observation of the dynode
areas. The light output follows the
N (COUNT INCREASE current level in the tube; at reason-
< 10%/100 VOLTS)
fl ably high levels of current, the dy-

=ocoo coo loon Izoo laoo 1600 nodes at the end of the tube seem
to be covered by a blue-green light.
APPLIED VOLTS
Light generated by luminescent
Fig. G4. Typical plateau characteristic.
effects associated with the high-

density pulse current in the output
stages of the tube is reflected and
transmitted back to the photocathode.

Multiplier Phototubes 61

Delay time is a combination of the velops. Post" was able in this man-
transit time of the secondary and ner to operate 931A tubes at volt-
primary electrons through the mul- ages of 4 to 5 kilovolts with very
tiplier phototube and the transit high gains, output-current peaks to
time of the light itself. This type 0.7 ampere, and pulse rise times of
of feedback has been minimized by the order of 0.8 X 10-° second.

baffles designed into the structure Transit-Time Effects
of the tube. After-pulses in the nano-
One of the principal advantages
second range are usually observed of multiplier phototubes compared
with simple photodiodes is the high
only under conditions of very-high- amplification achieved without ap-
gain operation, of the order of 10°. preciable loss in frequency response.
Calculated power spectrum of the
The second type of after-pulse, noise for type 931 is shown in Fig.
of microsecond delay, has been ob- 56~' (The 931 was an earlier form
served in many different types of of the present 931A multiplier photo-
tube. No consistent explanation of tube. The electrode structures are
their cause has been proposed, al- identical; the principal differences
though they are clearly some type are in the supporting members and
of regenerative effect. Fig. 55 shows test specifications, particularly for
some exaggerated after-pulses from dark current and sensitivity.) The
an experimental tube. Z'Vhen the gain figure shows the power spectrum of
and voltage are sufficiently high, the noise pulses arising from the am-
even "after-after" pulses may be plification of single electrons from
the photocathode. The loss of fre-
observed. quency response above 100 mega-
cycles results partly from the
30 A=DECAY OF INITIAL PULSE shaping of the output-current pulse
B=FIRST AFTER -PULSE by the passage of the amplified elec-
FNw C=SECOND AFTER-PULSE tron cloud from the next-to-the-last
ww 20 dynode through the anode structure
A to the last dynode. The electron
toe:n~. cloud is then further amplified at
e the last dynode and passes again to
Ua_ the anode, even oscillating in the
c to anode space before final collection.
w -J~ 10
ozase
aZ g
0 TIME AFTER PEAK OF
INITIAL PULSE —MICROSECONDS

Fig. 55. Multiple after-pulses observed in an 0
experimental multiplier phototube fol-
lowing a high current primary pulse. fa
The initial pulse is generated by the
light from a pulsed CRT; time con- Za
stant of the Phosphor and circuit com-
bined is approximately 0.8 microsecond. w~ a
~o:w~
Studies of the microsecond after- ~~aZ e
pulses under various conditions in- N ~} CURVE IS RELATIVE TO LOW-
dicate that they are of several dif-
ferent origins. In one class, the am- FREOUENCY VALUE
plitude of the after-pulse increases s i 12
at least as rapidly as the square of
the amplification factor; this rela- ~ww~
tionship suggests a regeneration ef-
fect dependent on feedback from the ~w 16
anode end of the tube.
Nli INTER-STAGE VOLTAGE=100
Feedback effects which have
time delays of the order of micro- aZ 20
seconds can be circumvented in spe-
cial applications by pulsing the ~' 2a 4 6 8100 2 a 6 81000 2
voltage so that high gain is achieved
10 2
before the regenerative pulse de-
FREQUENCY —MEGACYCLES

Fig. 56. Power spectrum of the noise from
the RCA 931 multiplier phototube op-
erated at 100 volts per stage, as cal-

culated by R. D. Sard.

62 RCA Phototube Manual

The electron cloud is also spread out Because of the very short time
by variations in transit time through response of the multiplier phototubes,
the multiplier section. These transit- it has been difficult to find a test
light source which has a sufficiently
time variations arise from the dif- short time of flash. A spark source,'

ferent energies and directions of which produces a good delta-function
light impluse having a main pulse
secondary electrons. This calculated width less than 10-~ second, is suit-
able for the measurement of the time
frequency response has been sup- delay in multiplier phototubes. Fig.
57 shows the transit-time delay for
ported by pulse measurements which an 8053 as a function of supply volt-
ages; the curve closely approximates
show that a type 931A multiplier

phototube has atharinse10t_ime esec(o10ndt.o31 90
per cent) less

The actual delay time between

the arrival of a photon and the re-

cording of an electrical pulse in the

anode circuit of the multiplier photo- too0 I x10-2 2vx-1i0-Z 3210-p 4210-2
tube may be much longer than the
pulse rise time. The pulse delay

time is the result of the accumulated

transit times for the several stages m 75
of the tube. Fora 931A, the transit-
delay time is approximately 16.7 X Z
10-" second when the tube is oper- W
ated at 100 volts per stage.
N
In tubes having large photo- i'
cathodes suitable for scintillation Z 50 r'

counting, one of the principal transit- WI
i= 2s

time effects occurs in the space be-

tween the cathode and the first "100 10 3 2 I.3 085 0.70

dynode. The large photocathode 0 zo s z.s I.s I.I o.ao

necessitates a fairly long path to SUPPLY KILOVOLTS BETWEEN
ANODE AND CATHODE
the first dynode to provide good
photoelectron collection from the FiS. 57. Transit time as a function of the
entire photocathode. In addition, for square root of reciprocal applied volt-
some tubes, all areas on the photo- age for a type 8053 multiplier photo-
cathode are not equally distant in
transit time of the photoelectrons to tube.

the first dynode. In tubes designed the inverse half power of the applied
for short-time resolution, the photo- voltage. This effect is to be expected,
cathode areas are curved and the provided the effects of initial veloci-
ties and secondary-emission delay are
first few dynodes are oriented to negligible. The intercept of the line
minimize transit-time spread.l' on the time axis is at 5.2 X 10-'

In a new type of multiplier second. This effect has been pre-

phototube presently under develop- viously observed;'itis probably due
to the delay induced by the circuit
ment, transit times are further re- within the tube. Fig. 58 shows the
duced by the use of accelerator grids time delays for a number of multi-
between dynodes (See Fig. 35)11 plier phototubes over a range of
operating voltages.
Secondary electrons are thus sub-
jected to a higher accelex•ating field In most applications, the delay
near the emitting surface and to a time of a multiplier phototube is not
decelerating field near the collector. as ilnportant as the pulse rise time
The effect is almost an order-of- or the pulse width for adelta- func-
magnitude improvement over tubes tion input. Fig. 59 shows the output
pulse width at half maximum am-
having comparable cathode areas. plitude for an 8053 as a function of
In some high-speed tubes, the last-
the reciprocal square root of the
dynode and anode leads are construc-
ted to form atwin-lead transmission voltage. The light source used was
line to the elements themselves.

Multiplier Photot:~bes 63

0 rise time, which is closely related to
the pulse width, has been measured
U for a number of multiplier photo-
W8 tubes as the time required for the

0 pulse to rise from 10 to 90 per cent

6 of the maximum value. These values
G are plotted for a range of typical
operating voltages in Fig. 60. No
I correction has been made for the
finite rise time of the light pulse
W4 itself, which is estimated to be in the
f order of 0.6 X 10-° second.

H

~2

to

500 1000 1500 2000 2500 3000
,~ ANODE-TO-CATHODE APPLIED VOLTS

Fig. 58. Transit time as a function of sup- N2
ply voltage (log scale) for a number 0z
of multiplier phototubes. vw

the spark light source described N
above. The pulse width observed is
210
spread by the width of the light
pulse itself, approximately 0.7 X 10 9 le

second. The output pulse width also %6
follows the inverse half power of the
applied voltage, as does the transit- a
time delay. This relationship in-
oa
dicates that the transit-time spread 0

W

2
N

is determined more by the tube

structure (e.g. unequal path lengths) 1 1000 1500 2000 25003000
than by initial velocities. These data
also suggest a finite intercept on the ANODE-TO-CATHODE APPLIED VOLTS

time axis of 4 X 10-° second. Fig. 60. Rise time (10 to 90 percent) as a
function of voltage for s number of
t multiplier phototubes.

-2 V2 3x10_2 4 xIO-2
FULL WIDTH
40 0 1x10 2x102
TRANSIT TIME SPREAD

m OF OUTPUT PULSE AT HALF If the inherently wide passband
MAXIMUM. of the multiplier phototube is to be
0z fully used, it is important not to
W 30 limit the response in the anode out-
put circuit. Ina 931A tube, for ex-
0 ample, the capacitance of the anode
to all other electrodes is 6.5 pico-
az farads. If the capacitances of the
20 leads and the input of a first-stage
amplifier tube are included, the total
J ~.
shunt capacitance can be 20 pico-
:~ 3 2.0 1.3 0.95 0.T0
farads. If it is desirable to maintain
~ 10 25 1.6 I.I o.eo a time constant of 10~ second, the
coupling resistor must be less than
KILOVOLTS 500 ohms. At the very shortest
times, the tube elements become part
Fig. 59. Transit time spread of an 8053 as of a transmission line and "ringing"
measured by the full width of the oub (transient oscillation) occurs if the
tube and circuit are improperly de-
put pulse at half maximum as a func- signed and matched.
tion of the inverse half power of the
applied voltage.

Rise Time. For the same delta-
function input mentioned above, the

64 RCA Phototube Manual

Linearity Of sitivity of the tube and cause an
Output Current apparent non-linearity.

The output current of a multi- Xrl ~_.
plier phototube has been shown to be
proportional to the light input over TYPE 931A SATURATION
a wide range of values40 The limit to _VOLTS PER STAGE=100 AT 45 MA
linearity occurs when space charge mW 10 3
begins to form. The first limitation W
of space charge is not necessarily in - MAXIMUM /
the space between the last dynode a 105 DEVIATION FROM

and anode, but more frequently in LINEARITY— 3:G
the space between the last two w 16T
dynodes. The voltage gradient be-
tween anode and last dynode is v 109

usually much higher than for other 0
dynodes, and therefore results in the 10 11
limitation at the previous stage, even
though the current is less. The max- ro-l3 I I 1
imum current, at the onset of space
charge, is proportional to the 3/2 I~H p-12 10- p N)-B 10-8 104 N)2
power of the voltage gradient in the LIGHT FLUX —LUMENS
critical dynode region. By use of an
unbalanced dynode voltage distribu- Fig. 61. Range of linearity, current as a
function of light flux, for a type 931A
tion and increasing the interstage multiplier phototube.
voltages near the end of the tube, it
is possible to increase the output Temperature Effects i

current in a given tube. In the case The gain of a multiplier photo-
of the limiting currents, it is neces- tube is fairly independent of temper-
sary to restrict the operation to ature over the normal operating
pulsed light to avoid damage to the ranges. Careful measurement with
tube. Fig. 61 shows the range of low current (to avoid fatigue effects)
linear current which can be obtained usually indicates a positive variation
from the 931A, and the region in of gain with temperature. For a
which the space charge limits the 1P21 (having Cs,Sb dynodes) Eng-
linearity. Table VIII shows the max- strom20 has reported an increase in
imum current which can be drawn gain of approximately 0.3 per cent
from various multiplier phototubes. per degree centigrade rise in tem-

It should be pointed out that a perature.
linear behavior is not always ob- Photocathodes usually show a
tained from multiplier phototubes.
For example, if the test light spot on slight increase in sensitivity with
a 931A is not directed close to the increasing temperature at the long-

center of the active area of the wavelength threshold of the spectral
photocathode, disturbing effects may response. Otherwise, however, they
arise from the proximity of the are quite stable with temperature
ceramic end plates. Near the end except for permanent changes, which
plates, the fields are not uniform probably result from redistribution
of cesium at higher temperatures.
and are affected by charge patterns
A most important characteristic
on the insulator spacers, which of multiplier phototubes is the rapid
change with the current level. An increase of dark current with tem-
exterior negative shield placed perature, especially as a result of
around the bulb wall can also im-
prove tube linearity. Aging effects thermionic emission, as discussed
resulting from the passage of exces- previously. Similarly, background
sive current may change the sen- noise increases rapidly with tempera-
ture because it is dependent on the
dark-current origins. Fig. 62 shows

the noise background fora 1P21 as
a function of temperature.

Multiplier Phototubes 65

TABLE VIII

Maximum (Space-Charge-Limited) Output Current for
Various Multiplier Phototubes

Tabs Type Mas. Saturated Over-all Voltage Distribntionx'
Current (Amperes) Voltage 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
931A, 1P21,
1P22, 1P28, 0.045 1000 1, 1.46, 0.83, 1, 1.2, 3.3
'6328, 6472
2648 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
2059 1.7b+ 1200 2, 1, 1, 1, 1, 1, 1.9, 2.7, 4, 1.6
2b00
5819, 6217, 0.04b++ 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1.25, 1.7G, LO
6342A, 6656A, 0.37 2500 2, 1, 1, 1, 1, 1, 1.3, l.b, 1.9, 2.3
6903 3800 2.8, 3.8, 4.4, 5, 6
0.23+++ 4, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1.25, 1.7F, 4.0
6810A, 7264, 1.2# 2750 2, 1.4, 1, 1, 1, 1, 1, 1, 1, 1.25, 1.6,
7265 2b00 1.75, 2.0
2, 1.4, 1, 1, 1, 1, 1, 1.26, 1.6, 1.75, 8.0
7046 0.23## 2200 2, 1, 1, 1, 1, 1, 1, 4, 3.6, 4, 4.8
0.30 2500 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
7850 1950
4459 0,30
0.70
7746 0.23

8053

•Note that types with identical geometry have been grouped together ; it has been as-
sumed that all would have the same space-charge characteristic, although in most cases the
"satellite" types were not separately tested.
►tThe numbers represent the relative divider stage voltages: cathode-to-first dynode, IIrsb
dynode-t-osecond-dynode, etc. In most cases the stage voltages have been increased tmvard
the output end of the tube to increase the space-charge-limited output current.
+Data from D. L. Lasher and D. L. Redhead, RSI, 34, 115 (1963) ; the 2069 is essentially a
five-stage 931A.
++Based on data for 931A which has identical geometry after the first stage_
+++Lawrence Radiation Laboratory Counting Handbook, UCRL-3307, December 2, 1968.
#W. Widmaier, R. W. Engstrom, R. G. Stoudenheimer, "IRE Transactions on Nuclear Science,"
NS-3, November 1956.

##Based on 6810A data because of identical dynode and anode design.

100 VOLTS PER STAGE.

BANDWIDTH: ICPS
LIGHT SOURCE; TUNGSTEN, AT 2870° K INTERRUPTED AT 90 CPS TO

PRO0.UCE PULSES ALTERNATING BETWEEN ZERO ANO FLUX

VALUE SHOWN FOR ANY GIVEN TUBE TEMPERATURE; °ON"
PERIOD OF PULSE EQUAL TO "OFF" PERIOD; RMS SIGNAL
CURRENT =RMS NOISE CURRENT

EXT10RNZL SHIELD VOLTS RELATIVE TO ANODE VOLTS= -1000

a

Iq

F-

a2

Z 10 13
W6
OO
6

Ow 4
Z~

F- O 2

J J 10 IB

~6
4

W2

10 IS

-150 -100 -50 O +50

TUBE TEMPERATURE-DEGREES CENTIGRADE

Fig. 62. Background noise variation in multiplier phototubes as a function temperature.

6S RCA Phototube Manual

Considerable advantage in low- w n
light-level operation may be achieved a 10
by cooling of the tubes. However, not s~s~
all multiplier phototubes can be 0 ~
cooled without compensating for the N ~pG?i
temperature variation of the resis-
tivity of the photocathode layer. For aW z 4jGAT 9~,5,92~
tubes such as the 1P21, in which the 4~4 ~O
photocathode is overlaid on a solid E ID B
conductor, there is no problem. How- 26 y0ro~ '~
ever, in the semitransparent type of O
photocathode, since the photocathode
is a semiconductor, the conductivity I
at low temperatures may become so
poor that the emission of photo- W~
electrons from the center results in
a positive charge pattern which ef- a IOg 4Tti
fectively blocks the normal operation y6
of the tube. No fixed lower-tempera- ~'°°F
ture limit can be given for proper W°
operation because the minimum tem- ~~S~
perature depends upon the photo-
current and the dark-emission cur- \22
I LIQUID NZ DRY ICE
rent, as well as on the particular 106 -200 -160 -120 -80 -40 O
photocathode. Fig. 63 shows the re- 40

sistivity per square for semitrans- TEMPERATURE —DEGREES C

parent photocathodes as a function Fig. 63. Resistance per square as a function

of temperature. of temperature for the Cs-Sb and the
multialkali semitransparent photocath-
Effect Of odes These data were obtained with
Magnetic Fields special tubes having connections to
parallel conducting lines on the photo-
To some degree, all multiplier cathode.
phototubes are sensitive to the pres-
deflection of electrons from their
ence of magnetic fields. Typical loss normal path between stages. Tubes
of sensitivity in the presence of a for scintillation counting are gen-
magnetic field is shown in Fig. 64. erally quite sensitive to magnetic
The loss of gain results from the fields because of the relatively long
path from the cathode to the first

dynode.

If multiplier phototubes are to
be used in the presence of magnetic
fields, as is often the case, it is
essential to provide magnetic shield-
ing around the tube. High-mu-ma-
terial shields are generally available

Fz H PARALLEL TO MAJOR AXIS OF TUBE (POSITIVE VALUES FOR LINES
W OF FORCE TOWARD TUBE BASE)

~ 100

aW

F. 80
W
C
7 60

W
O

z 40 j 7029
a
W

F 20 1 `\
a c6342A
J
W
K
0 1 I I `~ I

-30 -20 -!0 Oy 10 20

MAGNETIC FIELD INTENSITY (H)—GAUSSES

Fig. 64. Variation of output cu.•r:.nc ui several multiplier phototubes as a function of mag-
netic field strength.

Multiplier Phototubes 67

commercially. In some experiments, heavy current, it frequently is re-
even the earth's magnetic field may covered when the tube is removed
be critical, especially if the tube is
moved about. from operation. This recovery is ac-

It is possible to take advantage celerated by an increase in tempera-
of magnetic fields to modulate the
output current of the multiplier ture within the permitted range.
phototube. Under the application of (Too high a temperature may cause
normal fields, no permanent damage permanent loss in sensitivity.) Re-
results. However, it is possible to covery is probably the result of
cause a slight magnetic polarization cesium returning to the dynode sur-
of some of the internal structure of faces.
the tube. If this condition should
occur, the performance of the tube Because recovery occurs even
may be somewhat degraded by loss during operation, sensitivity loss is
in collection efficiency; however, it is not determined by the product of
a simple matter to "degauss" (de- current and time at very low cur-
magnetize) the tube by placing it in rents. Fig. 65 shows the short-time
an alternating magnetic field and fatigue and recovery characteristic
then gradually withdrawing it. A of a typical 1P21 having an initial
maximum field of 100 gausses at anode current of 100 microamperes;
the center of a coil operated on a the recovery is considerably slower
60-cycle alternating current is usu- than the fatigue. At currents of 10
ally sufficient to degauss a tube. microamperes or less, this situation
may be reversed. For a tube with
Fatigue And Cs-Sb dynodes, therefore, little im-
Life Characteristics provement in stability is achieved by
use of an anode current smaller than
It is difficult to predict the pre- 10 microamperes.
cise changes in sensitivity of a mul-
tiplier phototube which may occur loo
during the course of operation. The
fatigue characteristic is a function RELATIVE SENSITIVITY ~~eo
of output current, previous history,
and type of dynode material. Because 60 LIOGNT NO LIGHT ON TUBE
fatigue is quite variable from tube TUBE
to tube, only typical patterns can be
described. ao

Generally, sensitivity changes 20
become more rapid as the output
current increases. In fact, for an in- 0 100 200 300 400 500
dividual tube, the sensitivity change
is approximately a function of the TIME —MINUTES
product of time and output current,
especially for rather large currents. Fig. 65. Short-time fatigue and recovery
It may be that changes in dynode characteristics of a typical 1P21 op-
characteristics result from an altera- erating at 100 volts per stage and with
a light source adjusted to give 100
tion of the surface caused by the im- microampere initial anode current. Al
pact of the primary electrons. One the end of 100 minutes the light ie
turned o$ and the tube allowed to re-
possibility is that cesium is released cover sensitivity. 1~bea recover ap-
proximately as shown, whether the
from the surface and then recom- voltage is on or o$.

bines elsewhere in the tube. The re- Over a longer period, the rate of
sensitivity change decreases, as
lease of cesium may be expected to shown in Fig. 66, but the change
be more or less proportional to total tends to become more permanent and
charge impact, as indicated above. recovery is only partial.

When sensitivity of the tube is
lost as a result of the passage of a

ss RCA Phototube Manual i

loo 160 i

60 140

J J
F Q
> E-
F-
~?60 i 120
w
NW
WO 0
NF
z 100
j~40 o:
as a
I 60
~ ~ 20 rr

0 100 200 300 400 500 rwzn 60
TIME—HOURS
a 40
Fig. 66. Typical sensitivity loss for a 1P21
operating at 100 volts per stage for a J
period of 500 hours. Initial anode cur- W
rent is 100 microamperes and is
readjusted to this operating value at 48, 20
168, and 360 hours.

Multiplier phototubes having 0 100 200 300 400 500
silver-magnesium or copper-beryl- TIME—HOURS
lium dynodes are much more stable
at high operating currents than Fig. 67. Typical sensitivity variation on life
those having cesium-antimony fora 6342A multiplier phototube (silver
dynodes. There is no significant dif- magnesium dynodes) operating with
ference in stability between silver- 1250-volt anode supply voltage for a
magnesium or copper-beryllium period of 600 hours Initial anode cur-
dynodes. A typical characteristic of rent is 2 milliamperes and is readjusted
sensitivity change on life is shown to this operating value at 48, 168 and
in Fig. 67 for a type 6342A multi- 360 hours.
plier phototube (silver-magnesium
This particular phenomenon, ob-
dynodes). The sensitivity tends to served in a developmental type, was
increase at first, then levels off and probably the result of the charging
of the supporting insulator for the
decreases very slowly. The operating dynodes. The effect was observed to
current in this case is 2 milliamperes,
as compared to the 100-microampere 600
current used for the Cs-Sb dynodes a

typified by the characteristic shown g 400 LIGHT
I OFF LIGHT ON
in Fig. 66.
In addition to the life character- w
~ 200
istics, which are probably the result
of changes in the dynode layer itself, U
other changes of a temporary nature
also occur. Not all these changes are a0 0 24 6
well understood; some are charging ~ -2 TIME —SECONDS

of insulators in the tube. 0
Fig. 68 illustrates one of the
Fig. 68. Sudden shift in anode current prob-
peculiar instabilities which are some-
times observed in multiplier photo- ably as the result of insulator spacer
tubes. When the light is first turned charging. Observation was made using
on,the current apparently overshoots an experimental multiplier phototube
in which the effect was unusually large.
and then decays to a steady value.

Multiplier Phototubes 69

occur more rapidly at higher cur- pulsed cathode-ray tube was used as
a light source. Two pulse rates were
rents, presumably because of the studied: 100 and 10,000 pulses per sec-
ond. Pulse duration was one micro-
greater charging current. Observa- second; decay time to 0.1 maximum
tion of the phenomenon at different was 0.1 microsecond. The experiment
stages in the tube showed that it was devised to study the rate at which
originated between the first and sec- the multiplier-phototube output re-
ond stages. Electrons striking the
sponse changed when the pulse rate
insulator probably resulted in sec-
was suddenly switched between the
ondary emission and a resultant posi- two rates. Tubes such as the 6342A,
tive charge. The change is potential
affected the electron optics in the and especially the 8053 and 8054,
space between dynodes. The effect
was observed as an increase in some showed practically no effect (of the
order of one per cent or less). Fig. 70
tubes and as a decrease in others. shows the pulses during the switch-
This particular development tube ing procedure for a competitor's mul-
type was modified to minimize the tiplier phototube. The adjustment
effect by use of a metal shield to
cover part of the spacer at each end z3 i

of the first dynode space. WN
A related phenomenon is the
~~ ~W
variation of pulse height with pulse-
count rate in scintillation-counting ~a2 ~I
applications. Thus, when a radio- wa~ _
active source is brought closer to i
J_, IOOpps 100 pps
a scintillating crystal a greater rate ; 10000 pps
of scintillations should be produced, n.~ I
all having the same magnitude. In ~I
a particular multiplier phototube a QI ~I
few per cent change in amplitude
may result and cause problems in ao ~ ~~ IC I ~ I
measurement. Fig. 69 shows the
typically minor variation of pulse -10 0 10 20 30 40
height with pulse-count rate for a TIME —SECONDS
type 6342A multiplier phototube
Fig. 70. Variation of output pulse height
3 as the rate of pulsing is changed in a
poorly designed experimental tube. Light
LLZ2 pulses were provided from a cathode-
ray tube. At the left of the graph,
=U which shows the pulse-amplitude en-
N velope with time for the output of the
multiplier phototube, the pulses are at
ZW I 100 per second. The pulse rate is in-
ac~s creased suddenly to 10,000 per second
and again reduced as indicated. Changes
0 0.2x106 04 O6 08 10 1.2 14 in amplitude are probably the result
COUNTING RATE PER MINUTE
of insulator charging.

1.6

Fig• 69. Typical variation of pulse height decay curves are approximately ex-
ponential. The phenomena were com-
with pulse-count rate for a 6342A. pletely reversible and were observed
(Cs~ source with a NaI (TI) source).
on many different tube types (to a
In order to investigate the phe- lesser extent). The time-decay period
nomena of pulse-height variation with of several seconds suggests the
pulse-count rate, a purposely exag-
changing of an insulator spacer to a
gerated experiment was devised. In- new potential as the result of the
stead of a scintillating crystal, a increased charge flow and the sub-
sequent modification of interdynode
potential fields.

~o RCA Phototube Manual

References of Multiplier Phototubee", Nucleonics, 10, i

1. McI{ay, K. G., "Secondary Electron Emis- No. 4, Apr. (19b2).
sion", Advances in Electronics, 1, (1948).
19. Morton, G. A. and Mitchell, J. A., "Per-
2. Bruning, N., Physics and Application of formance of a 931-A Type Multiplier sa
Secondary Electron Emission, McGraw- a Scintillation Counter", Nucleonics, 4,
No. 1, (1949).
Hill Book Co., New York (1954).
8. Kollath, R, Hsnbneh der Physik, 21, 20. Engstrom, R W., "Multiplier Phototube
Characteristics; Application to Low Light
232 (1966). Levels;' J. Opt. 3oc. Am., 37, (1947).

4. Massey, H .S. W. and Burhop, E. H. S., 21. Baicker, J. A., "Dark Current Photo-
Electronic and Ionic Impact Phenomena, multiplier" IRE Transactions on Nuclear
Oxford Univ. Press, New York (1952). Science", NS-7, (1960).

6. Zworykin, V. K., Ruedy, J. E. and Pike, 22. Trusting, R. F., Kerns, W. A. and Knud-
E. W., "Silver Magnesium Alloy as a sen, FI. K., "Photomultiplier Single-Elec-
Secondary Electron Emitting Material", tron Statistics", IRE Transactions on Na-
J. Appl. Phyn., 12 (1941). clear Science, NS-9, (1962).

6. Rappaport, P., "Methods of Processing 23. Birka, J. B., Scintillation Counters, Mc-
Silver-Magnesium Secondary Emitters for Graw-Hill Book Co., New York (1953).
Electron Tubes", J. Appl. Phys., 25 (1954).
24. Morton, G. A_, "Photomultipliers for
7. Wargo, P., Iiaxby, B. V., and Sheperd, Scintillation Counting;' RCA Review, 10,
W. G., "Preparation and Properties of
Thin-Film Mg0 Secondary Emitters", J. (1949).

Appl. Phys., 27 (1956). 26. Crauthamel, C. E., Applfed Gamma-Ray
8. Allen, J. S., "An Improved Electron Multi- Spectrometry, Pergamon Press, New York

plier Particle Counter", Rev. $ci. Instr., (1960).
26. Bell, P. R., `Beta- and Gamma-Ray
18 (1947).
Spectrometry", Interaeience, New York
8. Rajchman, J. and Snyder, R. L., "An (196b).
Electrostatically Focused Multiplier", 27. Wall, N. 3., and Albvrger, D. E., Naclear
Spectroscopy (Fay Ajzenberg-Solove, Ed.)
Electronics, 13, Dec. (1940). Academic Press, New York (1960).
28. Proc. Fifth, Sixth, Seventh, and Eighth
30. Rajchman, J., "Lea Cuzanta Residuels Scintillation Counter Symposia ; IRE
daps Lea Multiplicateurs d'Electrona Elec- Transactions on Nuclear Science; NS-3,
trostatiques", (Thesis), Kundig, Geneva No. 4 (1956) ; NS-5, No. 3 (1968) ; NS-7,
Nos. 2, 3 (1960) ;and NS-9, No. 3 (1962).
(1938).
11. Morton, G. A., Matheson, R W. and 29. Engstrom, R. W. and Weaver, J. L., "Are
Plateaus Significant in Scintillation
Greenblatt, M. H., "Design of Photo- Counting?", Nacleonica, 10 (19b9).
multipliers for the Sub-Millisecond Re-
80. Mueller, D. W., Best, G., Jackson, J. and
gion", IRE Transactions on Naclear Singletary, J., "After-Pulsing In Photo-
multipliers", Nucleonics, 10, No. 6 (1952)
Science, NS-5 Dec. (1958).
12. Engstrom, R. W. and Matheson, R. W., 31. Post, R. F., "Performance of Pulsed
Photomultipliers", Nacleonica, 10, No. 6
"Multiplier Phototube Development Pro- (1962)

gram at RCA Lancaster", IRE Transac- 82. Sard, R. D., "Calculated Frequency Spec-
tions on Naclear Seienee, NS-7, June- trum of the Shot Noise from a Photo-
multiplier Tube", J. Appl. Phys, 17,
Sept. (1960). (1946).

13. Sommer, A. IL, "Use of Lithium in Photo- 83. Kerns, 8. A., "Improved Time Response
emissive Cathodes", Rev. Sci. Inetr. 20 in Scintillation Counting", IRE Transac-
tions on Naclear Science, NS-3, (19b6).
(1966), 28 (1957).
14. Lontie-Bailliez, M. and Mesaen, A., L'In- 84. Smith, R. V., "Photomultiplier Transit-
Time Measurements", IRE Transactions
fluence de la Temperature sur lea Photo-
mnitiplicatears, Centre de Physique Nu- on Nuclear Science, NS-3, (1956).
clesire, Universite de Louvain.
15. Wiggins, S. C. and Earley, K., "Photo- 86. Covell, D. F. and Euler, B. A., "Gain
multiplier Refrigerator", Rev. Sci. Inatr. Shift Versus Counting Rate in Certain
Multiplier Phototubes", USNRDL-TR-521,
23, No. 10> Oct. (1962). U.S Naval Radiological Defense Labora-
tory, San Francisco (1961).
1&. Pagano, R, Damerell, C. J. 3., Cherry,
R. D., "Effect of Photocathode-to-First-
Dynode Voltage on Photomultiplier Noise
Pulses", Rev Sci. Instr. 33, No. 9, Sept.

(1962).

17. Wright, D. A., Semi-Conductors, Methuen
and Co., New York (19b5).

18. Engstrom, R W., Stoudenheimer, R. G.,
and Glover, A. M., "Production Testing

Photocells

P HOTOSENSITIVE devices in In a circuit, a cadmium sulfide
cell having ohmic contacts acts as an
which electron flow occurs in a solid ohmic impedance. One of the im-
photoconductive material are called portant parameters of such a cell is
photocells. In a photoconductive its conductance as a function of
material, electrical conductivity is a illumination. Fig. 72 shows a char-
acteristic of this type in which the
function of the intensity of incident
t-o s
electromagnetic radiation. Although
many materials are photoconductive -a
to some degree, this section is limited to
to the three types which are most
useful commercially: cadium sulfide,
germanium, and silicon.

Cadmium-Sulfide -6
Photoconductive Cells
0rn 10

The basic elements of a cadmium- w -a DARK CONDUCTANCE p4
sulfide photoconductive cell include to
a ceramic substrate, a layer of photo- Iazu- -lo 104 10 z 10 10
conducting cadmium sulfide, metallic ~ to ILLUMINATION — FOOTCANDLES
00zu _Iz
electrodes, and a protective en-
closure. The photoconductive layer is w
prepared from cadmium sulfide
which has been treated with various -l4
activating materials (such as a to
chloride and copper). The electrodes
are formed by evaporation through -16
a mask of a metal such as tin, in- 10 -6
dium, or gold. The finished cell is
protected from moisture by a glass 10
or glass-metal envelope of the type

shown in Fig. 71.

GLASS ELECTRODE Fig. 72. Conductance as a function of illumi-
METAL nation for acadmium-sulfide photocell

WINDOW~~ CHSE

BASE slope of the curve is nearly constant
PIN around a given operating point. The
conductance G may be expressed as
CERAMIC follows:
SUBSTRATE
PHOTOCONDUCTIVE G — G~ Ly (45)
CADMIUM
where G~ is the conductance for unit
SULFIDE illumination, L is the illumination,

Fig. 71. Typical RCA cadmium-sulfide photo- and y is the slope of the character-
cell.

~~

72 RCA Phototube Manual

istic. The performance of the cell at more quickly at high light levels and
a given operating point is described the rise time is usually longer than
by specifying G~ (expressed in terms the decay time. Typical photocurrent
of the current drawn through the cell rise curves are shown in Figs. 73
at a given applied voltage) and y. and 74.

For a typical cell, the 7163, G~ and y to

are 53 micro-ohms and 1, respec- TEMPERATURE =25' C
tively; the photocurrent measured at I VOLTAGE=22.5 VDC
1 foot-candle and 50 volts (ac) is
approximately 2 milliamperes. ~I
PREVIOUS DARK STORAGE
The capacitance of these cells
AT LEAST 16 HOURS
does not respond instantaneously to 10

changes in incident illumination be- ~0 2

cause of the presence of electron ~W to ~~, l04
a
traps within the forbidden gap of I ~vz2~v 106
z loB
the cadmium sulfide. Although the o°
u
build-up and decay of conductance °g t-olo 4
on the application or removal of illu- a
mination is only approximately ex- toi ~
ponential and depends on the magni- -12
tude of the illumination, the term J~
time constant is frequently used to ro
describe the time required for the
~- -DARK CURRENT
conductance to rise to 63.2 per cent
io4 10'2 10- I I 10 102 103 w4
of the maximum value or to fall
TIME AFTER EXCITATION IS APPLIED —MINUTES
from the peak to 36.8 per cent of
Fig. 73. Photocurrent rise characteristics
the maximum value. For example, if for a cell selected for low dark current.
a cell has been in the dark for a
long time and is then illuminated
with 10 footcandles, the time con-
stant is approximately 70 millisec-
onds. In general, the cell responds

CURVES ARE INDEPENDENT OF VOLTAGE.
AMBIENT TEMPERATURE=25' C

CURVE ILLUMINATION—FOOTCANDLES*

A 10
BI
C 0.1

D D.DI

COLOR TEMPERATURE 2870'K
5 SECOND STORAGE IN DARK PRIOR TO EXCITATION.
----S-MINUTE STORAGE IN DARK PRIOR TO EXCITATION.

j 10 a ~' a
s i
I
Q6 i~

~ 4A i B~ C i i
♦- ~ 2 A~ -r D~
ZQ C
WW B D
tr F 10

~~ e
UQ 6
O
O F- a
=Z
dU 2

aW

2 4 6H 2 4 6e 2 4 68 0 2 4 68
100
0.01 O.I
TIME AFTER EXCITATION IS APPLIED —SECONDS

Fig. 74. Typical rise characteristics of a cadmium-sulfide cell.

Photocells 73

In addition to the short-term tions is described in terms of dark
time effects just described, other phe- current and decay current. Dark cur-
nomena resulting from previous light rent, the current passed under speci-
exposure proceed more slowly. In fied conditions of voltage and tem-
general, long exposure to high levels perature when the cell has been in
of light makes the cell slightly less the dark a long time, is usually ex-
sensitive and somewhat faster in re- tremely low. Because of time effects
sponse. These changes are reversible; it is more convenient to specify the
the cell reverts to its former condi- decay current, which is observed at
tion during storage in the dark. Be- a given interval after removal of the
light used fora sensitivity deter-
eause of the long-term time effects, mination. For a typical photocell
such as type 7163, the decay current
cells should be preconditioned to is below 40 microamperes at a volt-
age of 60 volts, 10 seconds after re-
light before measurement of sensi- moval of 1-footcandle illumination.
tivity. Acommonly used production- Typical photocurrent decay curves
testing preconditioning schedule pro- are shown in Figs. 75 and 76.

vides for exposure of the cells to a In the typical curve of photo-
current as a function of applied
b00-footcandle fluorescent light for voltage at various levels of illumina-
16 to 24 hours. Voltage is not applied
to the cell during the preconditioning tion shown in Fig. ?7, linearity ex-
schedule.
tends over six orders of magnitude
Certain time effects are also re-
lated to the application of voltage, of voltage. Peak-to-valley response

and as a result cells are often as a function of frequency of square-

slightly less sensitive under ac than wave light input is shown in Fig. ?8.
under do operating conditions.
The curve shows that the frequency
For most applications, the con- response improves as the level of

ductance of the photocell must be illumination increases.

substantially lower in the dark than
when the cell is illuminated. Cell per-
formance under unilluminated condi-

CURVES ARE INDEPENDENT OF VOLTAGE.
AMBIENT TEMPERATURE=25' C

CURVE ILLUMINATION—FOOTCANDLES~

A !0
BI
C 0•I
D 0.01
* GOLOR TEMPERATURE 2870'K.

~.zo roos `~
z~
W~ 6
U_
4
~u

aW.WX BC D

FO z
2
w j 10

e
~~ 6

~U)W 4

~J

Oj 2
2
a.~
0

2 4 68 2 4 68 2 4 6B 24 6B
10
0.001 0.01 0.1

TIME AFTER EXCITATION IS REMOVED—SECONDS

Fig. 75. Typical decay characteristics of acadmium-sulfide cell.

74 RSA Phototube Manual

0 lo° POWER DISSIPATION= Ip ~~
0.3 WATT—►~ ~p0i ~
to
p~ TEMPERATURE=25° C ~h ii
_p B VOLTAGE=22.5 VDC ~O~' ~
10 C
CURVE ILLUMINATION 102 00 GP `co
_4 D (F00T C AIJDLE S~
WIO
w A 102 ~l ~o
a B
C 10 N ~O ~O
106 D 2 w O~
H ~E
zw 10 104 ~~~P
j10 vJ
V 10 4 \
O 10 6
H I
O F
_ -w z
a10
~ 106 o°
7
U

LIGHT SOURCE=2870° K

—STEADY—STATE VOLTAGE
108 --VOLTAGE APPLIEDIN PUI v5

-12 ~ ~~
10

E

-la 101 102 101 1 10 102 103 10
103
10 -4 -2 O 2 4 6
10
10 10 10 10 10 DC VOLTS BETWEEN TERMINALS

TIME AFTER EXCITATION IS REMOVED—MINUTES

Fig. 76. Photocurrent decay charaeteristica Fig. 77. Curve showing photocurrent as a
for a cell selected for low dark currents.
function of anPlied voltage at various
levels of illumination.

CURVES ARE INDEPENDENT OF VOLTAGE.
AMBIENT TEMPERATURE=25' C

CURVE ILLUMINATION — FOOTCANDLES ~

A 0.01
B O.1
CI
D 10
E 100
M COLOR TEMPERATURE 2870 K~

r t~~ ~~~ ~_
z A
~ 1008 ~~ Dt
~~ 'W' a
UJ 4 C
OJ B

p> 2

=a FO 108

wQ 6

~W 4
a
J2

aI

2 4 6 8 2 4 69 2 4 6a 2 4 6 8 2 4 6 B

0.01 01 0 100 1000

LIGHT PULSES PER SECOND—~ON~TIME EQUALS OFF~TIME

Fig. 78. Peak-t-ovalley response as a function of frequency of apuare-wave light input,

Photocells 75

The sensitivity of a cadmium- fide photocells varies as a function of
sulfide cell tends to decrease as the the wavelength of incident illumina-
ambient temperature rises. The tion, as shown in Fig. 80. The re-
typical curves shown in Fig. 79 in- sponse curve is centered within the
dicate that the effect is marked at
the lower level of illumination, but visible range, and has a peak of sen-
becomes negligible at 10 footcandles sitivity near 5800 angstroms. Be-
and higher. cause the spectral response of cad-
mium sulfide closely matches that of
The sensitivity of cadmium-sul- the human eye, cadmium-sulfide cells

CURVES ARE INDEPENDENT
OF VOLTAGE.

CURVE ILLUMINATION—
FOOTCANDLES~

2 A 10

loon BI
C 0.1
D 0.01

* COLOR TEMPERATURE
2870' K.

D

6

3 C` \
\\
}
F ~.~B \~ 25~C

F- A ~. ~ r.J
N t
Z ~'~. ..~•~,` ~ ~
w
rn
w l0

Q
J
K

1

0
e

-80 -60 -40 -20 0 20 40 60 80

AMBIENT TEMPERATURE—`C

Fia. 79. Sensitivity ae a function of temperature for a Cadmium-sulfide ce1L

76 RCA Phototube Manual

usually used as either a photoconduc-

~loo FOR EQUAL VALUES OF tive or photovoltaic device. In photo-
~ so RADIANT FLUX AT conductive applications, the cell is
ALL WAVELENGTHS. biased in the reverse direction, and
the output voltage is developed across

om: a series load resistor. In photo-
voltaic applications, the cell is used
~ 60 to convert radiant power directly
into electrical power.
N

ao A photojunction device operated
in the photoconductive mode has a
W characteristic similar to that of Fig.

N

> 20 81, but rotated 180 degrees, as shown

a 5000 '7000 in Fig. 82. The circuit analyzed by
WJ o
Fig. 82 is shown in Fig. 83; as the
~ 3000 9Doo illumination on the cell is increased,

WAVELENGTH —ANGSTROMS a change in voltage is developed

across the resistor.

Fig. 80. Spectral response of a cadmium-
sulfide cell as a function of the wave-
length of incident illumination.

can be used for control applications LOAD LINE
in which human vision is a factor, ~ ILLUMINATION>0
such as street-light control and auto-
matic iris control for cameras.

Junction Photocells ~~

In some applications, photocon- Fig. 82. Analysis of a photojunction device
ductive materials such as germanium in the photoconductive mode of opera-
and silicon are used in junction de- tion.
vices; a p-n junction formed of such
material has a nonohmic character- LIGHT
istic, as shown in Fig. 81. The solid
curve applies when the device is in R~
the dark; when light is applied to
the cell, the curve shifts downward
as shown. The junction photocell is

I ~ o/ Ebb

,o, z~ Fig. 83. Photoiunction device connected {n
the photoconductive mode.
O
Germanium P-N Junction
=Q Q Photocells

Jr` ~'~// A germanium photocell, such as
~~ . / the 4420, which has a quantum
efficiency of approximately 0.45, is
PHOTOCONDUCTIVE PHOTOVOLTAIC intermediate in sensitivity between
a phototube and atypical cadium-
QUADRANT QUADRANT sulfide photoconductive cell. The 4420
has a dark current of less than 35
Fig. 81. Current-voltage characteristic for a microamperes, and the current
photojunction device.

Photocells 77

through the cell increases by about POWER DELIVERED
TO THE LOAD
0.7 microampere for each increase in
illumination of 1 footcandle. The LOAD LINE
increase in photocurrent is linear
with the increase of illumination. V- - >

The response of germanium junc- Fig. S4. Current-voltage characteristic oP a
tion photocells to sudden changes in photojunetion device connected in the
illumination is fairly rapid. The 4420, photovoltaic mode
for example, has a time constant
(photocurrent-decay characteristic) the load line consistent with the load
resistor used in series with the cell.
of approximately 7 microseconds. Be- The power delivered to the lead is
cause of this relatively fast response, determined by the area of the rect-
the germanium cell is useful for angle constructed as shown in Fig.
optical excitation frequencies well 84. The value of the load resistor
may be adjusted to provide maxi-
above the audio range. mum power from the cell for a
The germanium junction devices specific condition of input radiation.

contribute relatively little noise to a The performance of a silicon solar
circuit. The noise is 1/f in character; cell is frequently described in terms
a typical value of the equivalent of conversion efficiency, which is de-
noise input at 1000 cycles per second fined as the ratio of the electrical
(1-cycle-per-second bandwidth) is 60 output power to the incident radiant
microfootcandles. power, when the load resistance is
adjusted to provide an output volt-
Silicon Photovoltaic age of 0.46 volt, which is near the
Cells maximum-transfer point. The spec-
tral content and the intensity of the
Silicon solar cells, such as the illumination used must also be speci-
fied. Several typical illumination-
SL2205 and the SL2206 are junction source specifications are listed below.
devices used to convert the radiant
power of the sun to electrical power Tungsten efficiency radiation
for space applications. The cell con- from a bank of tungsten-filament
sists of a thin slice of single-crystal lamps operated at a color tempera-
p-type silicon up to two centimeters ture of 2800 degrees Kelvin is passed
square into which a layer (about 0.5 through a filter of 3 centimeters of
micron) of n-type material is dif- water, which absorbs unwanted in-
fused. The bottom contact of the cell frared radiation. The intensity is ad-
is usually a continuous layer of justed to provide a calibrated photo-
solder, and the top contact consists voltaic cell with ashort-circuit cur-
rent equal to the current measured
of a series of grid lines (electrodes). when the calibrated cell is illumi-
A non-reflective coating of silicon nated by the sun at air-mass one at
monoxide is usually applied to the an intensity of 100 milliwatts per
top surface to minimize reflection of square centimeter (extrapolated).
usable radiant energy from the sili- The load resistor is adjusted to pro-
con surface. vide 0.46 volt.

N-on-p type cells are formed by
diffusing phosphorous into a p-type
base. The advantage of the n-on-p
cell over the p-on-n cell is that it is
far more resistant to degradation
from the high-energy particles (pro-
tons and electrons) encountered in
space applications.

When the electrical performance
of a silicon solar cell is analyzed, the
characteristic curve of Fig. 81 is in-
verted and appears as shown in Fig.
84. The analysis consists of drawing

78 RCA Phototube Manual

Table Mountain. The source is are processed so that the peak of
the natural sunlight on a clear day response is shifted toward the
on Table Mountain in California. shorter wavelengths. The shift is ob-
(Table Mountain is the former site tained by making the n-layer (n-on-p
of the Smithsonian Astrophysical cells) thinner and taking advantage
Observatory.) of effects resulting from the absorp-
tion of light in silicon. A typical re-
Air Mass Zero. The source is sponse curve for an n-on-p cell is
shown in Fig. 86.
natural sunlight at a distance suffi-
Wioo
ciently above the surface of the earth
to eliminate atmospheric effects. U

Air Mass One. The source is nat- ~ 80
ural sunlight on a clear day at the
surface of the earth (sea level) when U
the sun is directly overhead.
U
Solar simulators. The source is 1
a combination of xenon and tungsten
lamps adjusted to approximate the ¢0x 60
sunlight above the earth's atmos-
phere. F 40

As a result of the spectral re- z
N 20
sponse of the silicon cell and the
spectral distribution of the light a0
sources, it is usually found that the
air-mass-zero efficiency is less than 3000 5000 7000 9000 Ii000
the tungsten efficiency. The Table WAVIELENGfH —ANGSTROMS
Mountain efficiency, on the other
II
hand, is higher than the air-mass-
zero efficiency. a

The spectral response of a silicon »~~>~JmK c~ rO rWcL?LW¢
solar cell has its peak in the near- Flp.. 86. Spectral response characteristic for
infrared region. The radiation from
sunlight, however, shows a peak of a silicon solar cell.
intensity near 4750 angstroms, as
shown in Fig 85. For maximum con- Other materials now in the de-
velopmental stage also show promise
w l in the field of solar-energy power
conversion. Gallium arsenide, for ex-
a~ 2 oz oa os os I ample, has a narrower band-gap
than silicon. Consequently, because
~~0~010e3 WAVELENGTH—MICRONS its peak spectral response is farther
in the blue region of the spectrum,
WU 6 the potential conversion efficiency of
the material exceeds that of silicon.
a~
Data-Processing
F~ a Cells

~Qwo_ Data-processing (read-out) cells
3¢ z are multiple-unit silicon photovoltaic
devices used for sensing light in such
~H applications as reading punched
cards, and in axial position indi-
~~.,'~ioa2 cators. Atypical cell consists of a
thin piece of silicon on which several
w 6 n-on-p photovoltaic elements have
been formed. Although these types
z 4

0 L2

Fig.85. Radiation from the sun (outside the
earth's atmosphere) ; Johnson solar
spectral irradiance curve—Journal of
Meteorology, 11, 431 (1964).

version efficiency, the solar cell
should respond to more blue radia-
tion than is characteristic of the
band-gap of silicon. Therefore, cells

Photocells 79

have been designed for specific ap- tends to about 9 microns. Because
plications in data processing, the the active gold level is relatively
technique for preparing such devices near the filled band, the cell must be
is very flexible and permits wide cooled to suppress noise resulting
variations in the location, size, and from thermal excitation and recom-
number of the sensing elements. bination. Furthermore, because the
Individual mechanical packages, of radiation results in a relatively small
several cells per package, can be
placed in various spatial arrays to change in the resistance of the
form sensing strips over large areas.
germanium element, the input signal
Germanium And
Germanium/Silicon should be chopped, and the output
Infrared Detectors amplified by a low-noise amplifier. A
typical cell consists of the germa-
The atmosphere is not uniformly nium element within an integrating
transparent to electromagnetic radi- chamber, a Dewar envelope contain-
ation. It contains various "atmos- ing liquid nitrogen for cooling, and
pheric windows" through which a germanium window treated with
radiation of specific wavelengths can an antireflection coating. The elec-
readily pass. As shown in Fig. 87, trical circuit consists of the cell, a
the p-type gold-doped germanium load resistor, a voltage source, and a
cell and the zinc- or gold-doped low-noise amplifier.
germanium-silicon-alloy cells have
response ranges which take advan- Germanium-silicon alloy cells
tage of such windows. are similar to the p-type gold-doped
germanium device. The alloy, how-
The sensitive material of the p- ever, makes it possible to use an
type gold-doped germanium cell is activator level which is closer to the
germanium activated with gold and valence band. The spectral sensitiv-
compensated with arsenic or anti- ity of the alloy cell extends to 14
mony. Incident radiation excites elec- microns. As a result of this extended
trons from the filled band to a low- response, the device must be cooled
lying impurity level contributed by to a lower temperature. Pumped
the gold. The spectral response ex- liquid nitrogen at a reduced pressure

(50 degrees Kelvin) is used as a

coolant; liquid neon and liquid

hydrogen are also sometimes useful.

--- p -TYPE GOLD-DOPED GERMANIUM
F 100 p TYPE GOLD-DOPED GERMANIUM-SILICON ALLOY
p-TYPE ZINC -DOPED GERMANIUM-SILICON ALLOY

U ^\ /r \~

W SO /\

I /1 /~ `1

ZW j ~, ~

a so 1 ~'

a ~ `\
1.
w
> 40 f ~~_ ~ ~
r ~~
w
i
20
4I `~' ~~ ` 18
0 4
2 \ '\
\
14 16
6 B 10 12
WAVELENGTH —MICRONS

Fig. 87. Spectral response of infrared-sensitive detectors.

General Application
Considerations

MANY criteria must be considered range of each class of photosensitive
device. Multiplier phototubes are in-
before choosing a photosensitive de- dicated for the lowest levels of radia-
vicefor aparticular application. Some tion because of their inherent ad-
application requirements can be filled vantage of secondary-emission gain.
Higher levels of radiation should be
by only one type of device; however, avoided in the case of multiplier
for many applications, any one of
phototubes because of fatigue, sat-
several possible devices may be suit- uration, and life problems. However,
able. This section provides a general the range can be increased upward
guide to the selection of photosensi- for a multiplier phototude by de-
tive devices for most typical applica- creasing the applied voltage.
tions.

Level of Light or Radiation Because the gain of a gas-filled
phototube is much less than that of
a multiplier phototube, it cannot be

All the devices described in this used at as low a level of light; how-

manual can be used over wide ranges ever, such tubes can operate satis-

of light or radiational level. Of course factorily at higher ranges. Vacuum

optical, environmental, or circuit ad- phototubes can be used at higher

justment may be necessary to ac- levels of radiation than either gas-

comodate the particular device. Fig. filled phototubes or multiplier photo-

88 shows the approximate useful tubes. Cadmium sulfide photocells
are usually most useful in the higher

MULTIPLIER 1. ranges of light intensity. At the
PHOTOTUBES lower ranges, they are somewhat
GAS —FILLED I limited by response-time problems.
PHOTOTUBES I Solar cells (photovoltaic cells, en-
ergy-conversion typed) are of use
VACUUM II only at the upper range of radiation
PHOTOTUBES levels. At low levels, the power de-
iiFENERGY CONVERSION TYPE veloped is too small to be of value
CdS PHOTOCELLS
for most applications.
PHOTOJUNCTION The far-infrared detectors are
CELLS

PHOTOVOLTAIC
CELLS

IO 12I0 1010 8 10 GI0 4 10 z IOG generally useful only for the very low
WATTS (2870' K) TUNGSTEN levels of radiation. If a large amount
of infrared radiation is available, the
10 I 10 I Is 10 isI0I4 K) I2 I complication of special cooling can
be avoided by other simpler devices
12 IGIO IOG such as the bolometer or thermo-
couple.
LUMENS (2870' K)
Spectral Energy Distribution
Fig. SS. The range of power or luminous
flux for which the various detector types Although many of the different
are useful. The indicated ranges are detectors of radiation have a broad
only for guidance, and may at times spectral sensitivity range, there are
be exceeded. The overlapping scales of
luminous and power fluxes are scaled
in reference to a tungsten lamp op-
erating at 2570 degrees Kelvin color
temperature. For a different reference
source, the luminous equivalent of the
power flux would be different.

80

General Application Considerations 81

times when the spectral energy dis- of the eye. In this case, the spectral
tribution of the source is of special
response may be modified by means
importance in selecting the proper
device. For the far-infrared region, of special color filters to provide a
combination detector and filter which
only the solid-state photoconductors does match the eye response. The
filter characteristic required may be
are suitable. Similarly, in the ultra- obtained by dividing the spectral sen-
violet region, only photoemissive de- sitivity characteristic of the eye by
vices are practical; however, in the the spectral response of the device to
visible range, many devices can be be corrected. Fig. 89 shows the close
used. The spectral response char- approximations which can be ob-
acteristics of the various types of tained by using commercial color
devices are included in the pertinent filters.
sections of this manual.
Fig. 90 shows the transmission
Some applications require a de-
vice having a specific spectral sensi- characteristics of a number of other
tivity to match a particular response.
However, for many applications, no useful filters and radiation trans-
such detector exists. For example, mitting substances. Color filters are
measurement of light levels requires
that the receiver have the same spec- frequently used to isolate various
tral sensitivity as the eye. Although
the devices described in this manual regions of the spectrum to improve
detection or matching characteristics
have spectral response covering the of photodetectors. The transmission
of various glasses and the atmos-
range of visible radiation, none of phere are also provided because

the responses exactly matches that these factors have an important in-
fluence on practical optical systems.

I.o

___STANDARD-EYE ~
~\
S-4 WITH WRATTEN FILTER /
No. 06 \

-- S-II WITH WRATTEN FILTER ~ ~~ ,
0.8 No.106
~/

w ~! /
1
N
Z

N 0.6

it ~,
w // \~
a 0.4
~/ \
o.
yp ~~,
4000
,~~ ~~~_®

5000 6000

WAVELENGTH —ANGSTROMS

Fig. 89. Spectral responses S-4 and 5-11 modified by the Wratten filter 106 closer match
spectral sensitivity characteristic of the eye.

82 RCA Phototube Manual

(a)

~" ~ _-

NZw PYREX (0-53) ' V~ ~2 8 (2-60)
a
5113 (5'58) lr.'2408 (2-60) ~ 15000
60 /~ `

w ~ ,1 2540 (7-56)~j
40
% ~\ I7
a 5000
20 1 1~~

o

WAVELENGTH-ANGSTROMS

(b) r-
'n
I
~ TRANSMISSION OF 2000 YARDS
OF ATMOSPHERE (17 MM
PRECIPITABLE WATER).

Z~W8 )

U
K
a 6~

1 /0y41I

N
2
¢ 21 i
f

2 4 68 .! LJ J B
IQQQQ 2 4 `6 IQDDQQ
1QOO

WAVELENGTH —ANGSTROMS

Fig. 90. Transmission characteristics of various useful color filters, glasses, and the
atmosphere. ( (A) 2540, 6113, 2408, Pyrex (B) 2000 yds of air.)

Frequency Response MULTIPLIER 7) I
PHOTOTUBES 31 I
The approximate useful range 7)
of frequency response for the various GAS- FILLED 7i
types of photosensitive devices is PHOTOTUBES
shown in Fig. 91. All the detectors 32
VACUUM
are useful from do operation to a PHOTOTUBES TEENERGY CONVERGION TYPE
DC IOo IOZ 10° 106 106 IOIo
specific cutoff frequency. However, CdS PHOTOCELLS
FREQUENCY-CPS
for photocells of the germanium PHOTOJUNCTION
type, in which the dark current is CELLS
quite substantial, the detection of
unmodulated signals is not practical. PHOTOVOLTAIC
A modulated signal can be auto- CELLS
matically discriminated from the
dark-current component. A compari- Fig. 91. Range of frequency response of
son of ranges in Fig. 91 shows that various photodetectors. Recommendation
the multiplier phototube is especially of operation for de (unmodulated light)
useful in detecting very-high-fre- sources is indicated by the symbol do
quency modulations. at the left end of the frequency range.

General Application Considerations 83

In some cases the light to be is moved across a slit, a sine wave
detected is modulated by the nature modulation can be achieved (see Fig.
of the application, or is in the form 92a). A circle passing over a square
of light pulses. For cases in which at 45 degrees (Fig. 92b) and having
the light is unmodulated, it is often a diameter equal to the diagonal
desirable to provide discrimination produces an approximate sine wave
between signal current and dark cur- with about 10-per-cent first-har-
rent by one of several means, such monic distortion (Uniform illumina-
tion of the apertures is assumed.)
as alight-chopper wheel.
The light chopper wheel is usu- Typical Applications

ally about 8 inches in diameter and As discussed in the previous
made of a material such as 0.030- sections, photosensitive devices are
inch Duralumin. Holes or alternate extremely versatile devices which
segments are so placed that when are being used in an ever-increasing
the wheel is rotated in the beam of variety of applications. Shown be-
light, a modulated beam falls on the low is a brief listing of several typi-
light detector. If a 1800 rpm syn- cal applications for each category
of photosensitive devices. Choice of
chronous motor is used to drive the a particular type for a specific ap-
plication should be based on the
disk, and three equally spaced 60-

degree cutouts are made in the disk,

the resultant modulation (as the

cutouts nass over ~ radial slit) ap-

0 40 80 120 160 200 240 280 320 360

DEGREES

Fig. 92a. Illustration o4 a alit passing over asine-wave type aperture to produce sine-wave
modulation.

proximates a 90-cycle square wave, ratings and characteristics shown in
the Technical Data Section.
which is a convenient frequency be-
cause it is not a multiple or sub- Vacuum Phototubes
multiple of the 60-cycle current.
Almost any type of modulation can Photometry
be accomplished with properly Spectrophotometry
shaped apertures. For example, if Industrial controls
the moving aperture is shaped like Facsimile
Colorimetry
the area under a sine curve and it

~ ~~ RCA Phototube Manual

- /~

- /-

-
-
\\

- is ~ -

_ ~f \ \ SINE WAVE _

/ ~ MUTUAL AREA —
/ \ OF CIRCLE
-/ AND SQUARE -
/
/ \

-~ \
/
\ -
/ t _
_ /~ /
\ \\
i~
-- 80 ~ ~ ~ i i ~n
160 200 240 260 320 360
l'EGRF,ES

Fig. 92b. Production of sine-wave approximation by nse of circular holes and a square aper-
ture of 45 degrees.

Typical Applications (con't) Photoconductive Cells
Industrial controls
Gas Phototubes Camera iris control
Industrial controls Street light control
Sound reproduction
Photojunction Cells
Multiplier Phototubes Sound reproduction
Scintillation counting Data processing
Photometry
Spectrophotometry Photovoltaic Cells
Flying-spot generator Solar power conversions
Star tracking Industrial controls
Cerenkov radiation measurement Photometry
Laser detection
Industrial controls
Colorimetry
Timing measurement

Interpretation
of Data

T HE data given in the TECH- ment, load variations and environ-
mental conditions.
NICAL DATA SECTION include
For the most part, electrode
ratings, characteristics, minimum voltage and current ratings for

circuit values, and characteristic phototubes are self-explanatory and
require little discussion. However, it
curves for RCA vacuum and gas should be noted that the maximum
photodiodes and multiplier photo- average cathode current (for gas and
vacuum phototubes) and the maxi-
tubes. Data for photoconductive cells mum .average anode current (for
and photojunction cells are also in- multiplier phototubes) are averaged
cluded. This section discusses the over an interval no longer than 30
parameters given in the data and in- seconds.
dicates briefly the method of meas-
urement used for the more important
parameters.

Maximum Ratings Characteristics

Ratings are established on The characteristics given in the
TECHNICAL DATA SECTION are
phototube and photocell types to typical values which indicate the
performance of the device under cer-
help equipment designers utilize the tain operating conditions. Charac-
teristic curves represent the char-
performance and service capability
of each device to best advantage. acteristics of an average tube;
These ratings are based on careful however, individual tubes (like any
study and extensive testing by the manufactured product) may have
tube manufacturer, and indicate characteristics that range above or
limits within which the operating below the values given in the char-
acteristic curves. The more im-
conditions must be maintained to en-
sure satisfactory performance. The portant of these characteristics for
phototubes are discussed below.
maximum ratings given for the
The spectral-sensitivity charac-
photosensitive devices included in
teristic represents a response curve
this manual are based on the Abso- which is typical of the spectral re-
sponse obtained with a given photo-
lute Maximum System. This system tube. Such a curve also indicates the
range of maximum response. The
has been defined by the Joint Elec- short-wavelength cutoff of the spec-
tron Device Engineering Council tral response is fairly well fixed by
the ultraviolet absorption properties
(JEDEC) and standardized by the of the tube envelope. The long-wave-
National Electrical Manufacturers length cutoff is determined by such
Association (NEMA) and the Elec-
tronic Industries Association (EIA). factors as the thickness of the photo-
cathode layer and the particular ac-
Absolute-maximum ratings are
limiting values of operating and en- tivation of the photosurface.
vironmental conditions which should In some critical applications, an
not be exceeded by any device of a
specified type under any condition of exact knowledge of the spectral re-
operation. Effective use of these
ratings requires close control of sup-
ply-voltage variations, component

variations, equipment-control adjust-

85

86 RCA Phototube Manual

sponse of a phototube may be re- LIGHT DC
SOURCE SUPPLY
quired. Asimplified system for the
measurement of spectral response VOLTAGE

as a function of wavelength is shown I MEG
in Fig. 93. For this measurement,
one or more monochromators are FI¢. 94. Typical circuit for measuring
used to select narrow bands of radia- sensitivity of photodiodea
tion from a given source of radiant
energy. The output radiant flux in cal circuit for the measurement of
luminous sensitivity of photodiodes;
the narrow bands is then directed to Fig. 95 shows a similar circuit used
a calibrated radiation thermocouple for multiplier phototubes. For these
and measured in watts. This same measurements, the phototube is
output flux is also directed to the placed in a light-tight shielded en-
photocathode and measured in am- closure to prevent extraneous radia-
peres by a current-reading device. tion from affecting test results. The
light source normally used is an
VOLT- aged 50-candlepower tungsten-fila-
METER ment lamp operated at a color tem-
perature of 2870°K and having a
l lime-glass envelope. The lamp is
calibrated for color temperature and
MONO- SHUTTER-SOURCE candlepower against a secondary
~ CHROMAl9R standard lamp supplied by the U.S.
National Bureau of Standards.
AM-
METER For the measurement of lumi-
nous sensitivity of vacuum photo-
Fi¢. 93. Typical system for determining tubes, avoltage in the order of 250
spectra]-sensitivity characteristics of volts is generally applied; for gas
phototubes. photodiodes, 90 volts is used. As
shown in Fig. 94, a microammeter
Luminous sensitivity is defined is inserted in series with the photo-
as the output current divided by the diode. Aone-megohm resistor is also
incident luminous flux at constant
electrode voltages. This parameter is
expressed in terms of amperes per
lumen (a/lm). Fig. 94 shows a typi-

APERTURE.
A

CALIBRATED LIGHT DY DYz DY3
DYN
LAMP ATTENUATOR PHOTO—
CATHODE

ANODE

Q

Fig. 9b. Typical circuit for measuring sensitivity of multiplier phototubes.

S

Interpretation of Data 87

used in series with the phototube at constant electrode voltages. Cath-
because it represents a typical load
ode radiant sensitivity is the amount
and, in addition, provides a measure
of safety for the meter in case of a of current leaving the photocathode
short circuit. With multiplier photo-
tubes, the tube is connected across a divided by the incident radiant power
voltage divider, as shown in Fig. 95,
of a given wavelength. These pa-
which provides voltages as specified
rameters are generally expressed in
for the individual type. The current
passing through the divider should terms of amperes per watt (a/w).
have a value at least ten times that
of the maximum anode current to be Although these characteristics can
measured. A luminous flux in the
range of 10~ to 10~ lumen (0.01 to be measured by use of the circuits
10 microlumens) is directed to the
described above with the addition
photocathode.
of a calibrated radiation thermo-
(As shown in the data for some
multiplier phototubes, the luminous couple, they can be more readily
sensitivity can also be given with the
computed from the measured cath-
last dynode stage used as the out-
ode and anode luminous sensitivity
put electrode. With this arrange- values. This calculations• Z uses the

ment, an output current of opposite conversion factors shown in Table
polarity to that obtained at the anode
is provided. Under this condition, the IX, and is dependent on established

load is connected in the last dynode typical spectral-response character-

circuit and the anode serves only as istics. For this reason, the result

the collector.) may differ somewhat from the value

Cathode luminous sensitivity is directly measured.
the photocurrent emitted per lumen
of incident light flux at constant A suitable light source for use in
electrode voltage and is expressed
in terms of microamperes (~a or this type of measurement is a spark
10-° ampere) per lumen. For photo-
diodes, this characteristic is meas- discharge in a mercury switch cap-
ured by means of the circuit shown
in Fig. 94 with a light flux of ap- sule. Although this source does not

proximately 0.1 lumen applied. From provide an ideal delta-function pulse,
a practical standpoint, there is no
distinction made between the meas- the resulting rise time of approxi;
urement of anode and cathode lumi-
nous sensitivity for photodiodes. In mately 0.4 nanosecond is useful for
the case of multiplier phototubes, a
measuring circuit similar to that many purposes.
shown in Fig. 95 is used with a do
voltage of 100 to 250 volts (specific Current amplification for multi-
value given in data for individual
type) applied between the cathode plier phototubes is the ratio of the

and all other electrodes connected as anode luminous' sensitivity to the
anode. Light-limiting apertures are
used and a light flux in the order photocathode luminous sensitivity at
of 0.01 lumen is generally applied.
The measured photocurrent (minus constant electrode voltages, and is

the dark current) is then divided by a computed parameter. (This char-
the specified light level to determine
cathode luminous sensitivity. acteristic can also be given as a ratio

Radiant sensitivity is the out- of radiant sensitivity.) Because of
put current divided by the incident
radiant power of a given wavelength its magnitude, the amplification fac-

tor for multiplier phototubes is gen-

erally expressed in terms of millions,

and is plotted logarithmically.

For a gas photodiode, the am-

plification factor is a ratio of photo-

currents at two different anode volt-

ages: the operating voltage (usually

90 volts) and 25 volts.

Equivalent anode-dark-current

input is the anode dark current of a

phototube divided by the radiant or

luminous sensitivity. This parameter

serves as a useful figure of merit for

the comparison of dark current in

multiplier phototubes, and ocran10b_ve
expressed in pico~vatts (pw

watt) or nanolumens (nlm or 10-°

lumen). The measurement method

for this characteristic is similar to

that for luminous sensitivity except

88 RCA Phototube Manual

TABLE IX
Spectral Responses fer Devices with Related Photocathode

Characteristic Values

Typical Typical
Photocathode
Luminous 1Saz Typical Dark Emiasion2
Aadfant at 25~C
6ensitivSty Luminous 6ensitivity gypical (amperes/~m9
guanturo
Device Conversion (8evn) 8ensitivity~ (Orvn) Y 10- I6
8-Number Factor (k) (2780oH:) (p,a/lumen) ~mn (ma/w) Efficieucys
(lm/w) (µa/lumen) gm.z (angstroms) (per cent)
S-1
S-3 93.9 25 60 8000 236 0.36 800
S-4 286 6.5 20 4200 1.86 0.55
S-5t 977 40 110 4000 39.1 0.2
S-8 1252 40 3400 b0.lt 12 0.3
S-9 755 80 3650 2.26 18t 0.13
5-10 683 3 20
S-11 608 30 110 4800 20.6 0.77 70
5-13 804 40 20.3 6.3 3
5-17 795 60 100 4500 4
5-19 664 60 110 4400 48.2 5.6
S-20 —'~ 126 14 1.2
S-21 428 40 80 4400 47.7 0.3
779 150 13 Q3
160 4900 83 21 —
30 70 ~ 22• 11•
64.2 18
260 4200
23.4 6.6
60 4400

r Care must be used in converting eIDeZ to a am:: figure. Photocathodes having maximum
Inmen sensitivity frequently have more red sensitivity than normal, and the formula cannot be
applied without revaluation of the spectral response for the particular maximum sensitivity
device.

a 100 per cent quantum efficiency implies one photoelectron per incident quantum, or
e/hv = x/12,396, where ~ is expressed in angstrom units. Quantum efficiency at ~,,,nx ~s
computed by comparing the radiant sensitivity at ~maY with the 100 per cent quantum effi-
ciency expression above.

s Most of these data are obtained from multiplier phototube characteristics. For tubes

capable of operating at very high gain factors, the dark emission at the photocathode is taken

as the output dark current divided by the gain (or the equivalent minimum anode dark cur-
rent input multiplied by cathode sensitivity). On tubes where other do dark-current sources
are predominant, the dark noise figure may be used. In this ease, if all the noise originates
from the photocathode emission, it may be shown that the photocathode dark emission in
amperes is approximately 0.4 X 10~s X (equivalent noise input in lumens times cathode aenai-
tivity in amperes per lumen)a. The data shown are all given per unit area of the photocathode.

s No value for k or ~mn: is given because the spectral response data are in question. The
values quoted for o and typical quantum efficiency are only typical of measurements made at
the specific wavelength 2637 angstroms and not at the wavelength of peak sensitivity as for
the other data.

~ The S-6 spectral response is suspected to be in error. The data tabulated conform to
the published curve, which is maximum at 3400 angstroms. Present indications are that the
peak value should agree with that of the S-4 curve (4000 angstroms). Typical radiant eenai-
tivity and quantum efficiency would then agree with those for S-4 response.

that the applied voltage is adjusted teristie can be expressed in terms
until a specified anode luminous sen- of picolumens (plm or 10-'~ lumens).
sitivity is obtained. The light flux is The test conditions for this measure-
then removed and the anode dark
ment are similar to those used for
current is measured. If the meas-
luminous sensitivity measurements
urement is made in terms of radiant except that the incident radiation is
flux, the wavelength of radiant en- modulated by a mechanical "light-
ergy is also specified and the applied chopper" which produces asquare-
voltage is adjusted to a specified wave signal at the phototube anode.

anode current. Transit time of multiplier photo-
tubes is defined as the time interval
Equivalent noise input is that between the arrival of a delta-func-
value of incident luminous flux which, tion light pulse (a pulse having finite
when modulated in a stated man- integrated light flux and infinitesi-

ner, produces an rms output current mal width) at the entrance window
equal to the rms noise current within
of the tube and the time at which
a specified bandwidth. This charac-

Interpretation of Data 89

the output pulse at the anode termi- measure rise time. For this measure-
ment, the incident light usually il-
nal reaches peak amplitude. Fig. 96 luminates the entire photocathode.
The rise times of the light source
shows a typical circuit for the meas- and the oscilloscope (or other dis-
play devices used), however, must
urement of transit time. As shown, be considered.

a small pulsed light spot of specified Transit-time spread is the time
interval between the half-amplitude
diameter is directed on a central area points of the output pulse at the
anode terminal, which results from a
of the photocathode. Because transit delta function of light incident on the
entrance window of the tube. This
time is dependent upon the magni- parameter is seldom measured di-
rectly because it is difficult to dif-
tude of the applied voltage, the
ferentiate between the decay of the
voltages are specified for each type.
light source and the decay of the
For this measurement, the phototube. Transit-time variations

length of the delay cable is adjusted between anode pulses as a function

until the time difference (T,) be- of light-spot position on the photo-

tween the phototube output pulses cathode serve as a satisfactory indi-
cation of transit-time spread. As
and the marker pulses from the light given in the data section, this para-
meter is expressed as greatest delay
source can be observed on a sampling between anode pulses. For measure-
ment of this parameter, a small-
oscilloscope. Transit time (T:) is diameter light spot is initially cen-
tered on the photocathode; the tran-
then equal to
sit time of the phototube is then
Tc=Tai-Ta—T.—T.
determined and used as a reference
where Ta is the electrical transit time point. The same light spot is then

of the delay cable; T, is equal to directed to another specified point

the time it takes the light pulse to on the cathode and this transit time
is determined. The difference be-
travel the distance between the mer- tween these transit times indicates
the transit-time spread.
cury-switch light-pulse generator

and the photocathode; T~is equal to

the electrical transit time of cable
«A„

Anode-pulse rise time indicates

the time required for the instan-

taneous amplitude of the output

pulse at the anode terminal to go

from 10 per cent to 90 per cent of

the peak value. This parameter is

normally e1x0p-8ressesecdondin). nanoseconds
(nsec or The circuit

shown in Fig. 96 can also be used to

APERTURE OUTPUT
PULSE

LIGHT MULTIPLIER I CABA~L, E OSCILLD—
LIGHT ~ BEAM ~ ^ i{ SCOPE
PULSER DELAY=TaI PHOTOTUBE ~c`7l DELAY=T

DELAY=T1

DELAY CABLE ELECTRICAL MARKER
—PULSE FROM LIGHT

PULSER

DELAY=Td

TRIGGER CABLE

Fig. 96. Typical circuit for determining time response of phutotubea.

References Devices, RCA Review, 21, No. 2, dune
1960.
1. Engstrom, R. W., "Calculation of Radiant 3. Kerns, Q. A., Kirsten, F. A., and Cox,
Photoelectric Sensitivity from Luminous G. C., "Generator of Nanosecond Light
Sensitivity", RCA Review, 16, No. 1, Pulse for Phototube Testing", Rey. Sci.
March 19b5.
Instr., 30, pp. 31-36, Jan. 1959.
2. Engstrom, R. W., "Absolute Spectral

Response Characteristics of Photosensitive

QUICK SELECTIOF~ GUIDE

FOR RCA MULTIPLIER PHOTOTUBES

Spectral Numher RCA Luminous Anode Maximum Wavelength
Response of Stages type Sensitivity Supply Diameter
Maximum of Maximum
(a/Im) (volts) (in.) Length Response

(in.) (angstroms)

S-1 10 7102 4.5 1250 1.56 4.57 8000
1000 1.31 3.69 4000
S-4 9 1P21 80 1000 1.31 3.69 4000
4000
S-4 9 931A 24 1000 1.19 2.75
3.12 4000
S-4 9 6472 35 1000 1.31

S-4 9 6328 35

S-4 9 71.1.7 35 1000 1.31 3.12 4000
1000 1.31 3.69 3400
S-5 9 1P28 50 1000 1.31 3.69 3650
1000 2.31 5.81 4500
S-8 9 1P22 1.0
1250 2.31 5.81 4400
5-10 10 6217 24

5-11 10 2020 6

5-11 10 2067 15 1000 1.56 2.80 4400
5-11 10 4438 27 1000 1.56 3.91 4400
5-11 10 4439 27 1000 1.56 3.91 4400
5-11 10 4440 27 1000 1.56 4.12 4400
5-11 10 4441 27 1000 1.56 3.18 4400

5-11 10 5819 25 1000 2.31 5.81 4400
5-11 1000 1.56 4.57 4400
10 6199 27
5-11 1250 2.31 5.81 4400
5-11 10 6342A 18 1000 2.31 5.81 4400
5-11 2000 2.38 4400
10 6655A 50 7.5

14 6810A 3050

5-11* 14 7046 180 2800 5.25 11.12 4200
5-11 7264 2000 2.38
5-11 14 7746 875 2000 2.31 7.5 4400
5-11 10 7764 1200 1200 0.78 6.12 4400
2.75 4400
5-11 6 7767 0.3 1250 0.78
4 4400
10 16

5-11 12 7850 6000 2300 2.06 6.31 4400
1500 5.81
5-11 10 8053 19 2.31 6.31 4400
1500 3.65 7.69 4400
5-11 10 8054 19 1500 5.31 6.56 4400
1000 2.31 4400
5-11 10 8055 19

5-13 10 6903 24

5-17 10 7029 40 1000 1.56 3.75 4900
1000 1.31 5.69 3300
5-19 9 7200 40
2400 2.38 7.5 4200
5-20 14 7265 3000 1800 2.38 6.78
4200
5-20 10 7326 22.5

•Extended 5-11 response; response from approximately 2600 to 6b00 angstroms.

90

~eclznical
Data

T HIS section contains maximum ratings, characteristics, and curves for

RCA gas, vacuum, and multiplier phototubes. Also included is a chart de-
scribing RCA's line of photocells. All ratings given in this section are based
on the Absolute Maximum System.

Tube types are listed according to the numerical-alphabetical-numerical
sequence of their type designations. For socket and shield data for all photo-
tubes, see page 182. For Key to Terminal-Connection Diagrams, see inside

back cover.

A number of abbreviations have been used to
simplify the tabulation of parameters. Some of
the less familiar are shown below.

a/lm —ampere per lumen

a/w —ampere per owr a1t0t-e lumen
nlm — nanolumen
nsec — nanosecond or 10_s second

pf — picofarad or 10-'~ farad

plm — picolumen or 10-" lumen

pw — picowatt or 10-'~ 1w0_ae ttampere
,~a —microampere or

NINE-STAGE, side-on type having S-4 response. Wave- 1P21
length of maximum spectral response is 4000 ± 500
MULTIPLIER
angstroms. This type makes use of cesium-antimony PHOTOTUBE
dynodes and acesium-antimony, opaque photocathode.
Window material is Corning No. 0080 lime glass or
equivalent. Tube weighs approximately 1.6 ounces
and has anon-hygroscopic base.

DIRECT INTERELECTRODE CAPACITANCES (Approx.): 4.4 Df
Anode to Dynode No. 9 8 Dt
Anode to All Other Electrodes
1000 1250 max volts
MAXIMUM RATINGS (Absolute-Maximum Values): 78000 250 max volts
Supply Voltage (DC or Peak AC) 250 max volts
0.04 260 max volts
Between anode and cathode• 0.1 maa
Between anode and dynode No. 9 80 76 maz ma
Between consecutive dynodes °C
Between dynode No. 1 and cathode 750
Average Anode Current 12000 volts
Ambient Temperature a/w
0.04 sw
TYPICAL CHARACTERISTICS: 12 am
DC Supply Voltage'•
Radiant Sensitivity (at 4000 angstroms)
Cathode Radiant Sensitivity (at 4000 angstroms)
Luminous Sensitivity'

91

92 RCA Phototube Manual

Cathode Luminous Sensitivity*` 40 40 µa/Im
Current Amplification (in millions) 2 0.3
Equivalent Anode-Dark-(~'DrT.'nt Input at 25°C nlm
0.5 max —
at a luminous sensitivity of 20 a/lm plm
Equivalent Noise Inputj' 0.5 —

'~OPeration with a supply voltage of less than 600 volts do is not recommended ; if used,
illumination must be limited so that cathode photocurrent does not exceed approximately 0.01

microampere.

'~wDC supply voltage (E) is connected across a voltage divider which provides 1/10 of E be-
tween cathode and dynode No. 1, 1/10 of E for each succeeding dynode stage, and 1/10 of E
between dynode No. 9 and anode.
'With light input of 10 microlumens from a tungsten-filament lamp operated at s color
temperature of 2870° K. Range of luminous sensitivity is 40 to 800 a/lm at 1000 volts; 5.4
to 100 a/]m at 750 volts.

'With light input of 0.01 lumen from atungsten-filament lamp operated at a color tempera-

ture of 2870° K ; 100 volts applied between photocathode and all other electrodes connected as

anode.

tAt tube temperature of 25°C and with external shield connected to cathode; bandwidth is
equal to 1 cycle per second. A tungsten-filament light source at a color temperature of
2870°K is interrupted at a low audio frequency to produce incident radiation pulses alter-
nating between zero and the value stated. The "on" period of the pulse is equal to the "off"
period.

TYPICAL EFFECT OF MAGNETIC
FIELD ON ANODE CURRENT

TYPE IP21
INCIDENT 0.167' DIA. LIGHT SPOT NORMAL TO AND
CENTERED ON PHOTOCATHODE GRILL.
SENSITIVITY AND CURRENT UNIFORM MAGNETIC FIELD PARALLEL TO MAJOR
AMPLIFICATION CHARACTERISTICS _OF TUBE. POSITIVE VALUES OF MAGNETIC FLUX A%IS
F- - FOR LINES OF FORCE TOWARD TUBE BASE. ARE_

TYPE IP21 Z
VOLTS /STAGE • I / IO E
ANODE-TO -DYNODE No.9 VOLTS = I/IOE ~ 1006

e i~ ' SUPPLY VOLTAGE
y BETWEEN ANODE
U 4 p~i ~ .,~`

4~—~~~~~~■■~■~~■ 11111 ~~~ AND CATHODE=500 V
2~~~~~11111Y 10a~~~~~~~~~~ O2

4~~~~■11116—__--■■■■■ 106

N W 6
J

4

Q

2

KI

-60 -40 -20 0 20

w103-_ -_ -■_ 1'111 MAGNETIC FIELD INTENSITY—GAUSSES
92C5-766472
J~'~{~~~~%
8~~~

~~ 5~~i ~~~~ 111 1 3/16 MAX.
er R ,►
+~~= v~~d'N"' DIA.

T9 BULBS — 5/ly MIN.

—A'~ PHOTO-
n m~ CATHODE

nnm~
IMF
~~ ~Q15~~y~~~Iillll 15/16 MIN. 3I
I/ ~~JQ.GP'~5~~~~11111 I/I6
SMALL-SHELL
J~~~~~~~~~ SUBMAGNAL MAX.
II-PIN BASE
ti ~.~~a`~~~~~1■1■■1■1■1 JEDEC N!
~_II~~■11111 BII-88

~~~~~11111 1 5/ly MA%,
DIA.
--■„1„
DYy

500 6 7 es
1000 1250
ANODE-TO-CATHODE SUPPLY VOLTS (E)

92CM-645474

v

DYI ~ K
WCIDENT RADI4TION

Technical Data 93

VTYOPLETSI/PS2TAGE =100 TYPICAL ANODE CHARACTERISTICS
2
25

Na 20
W LIGHT FLUX—MICROLUMENS=I5
K
10
1.5 5
a

J
J_

WI

O
O

aZ

0.5

0 50 100 150 200 250 92CM-6956T4
ANODE-TO-DYNODE-No.9 VOLTS

NINE-STAGE, side-on type having S-8 response. Wave- 1 P22
length of maximum response is 3650 ± 500 ang-
stroms. This type makes use of cesium-antimony MULTIPLIER
PHOTOTUBE
dynodes and acesium-bismuth, opaque photocathode.
Window material is Corning No. 0080 lime glass or
equivalent. Tube weighs approximately 1.6 ounces

and has anon-hygroscopic base. For outline and
terminal-connection diagram, refer to type 1P21.

DIRECT INTERELECTRODE CAPACITANCES (Approx.): 4.4 pf
6 pf
Anode to Dynode No. 9
Anode to All Other Electrodes

MAXIMUM RATINGS (Absolute-Maximum Values): 1250 max volts
250 max
Supply Voltage (DC or Peak AC) 250 max volts
Between anode and cathode volts
Between anode and dynode No. 9 250 max volts
Between consecutive dynodes 1 max
Between dynode No. 1 and cathode b0 max ma

Average Anode Current °C
Ambient Temperature

TYPICAL CHARACTERISTICS: 1000 750 volts
750 110 a/w
DC Supply Voltage* 0.0023 a/w
Radiant Sensitivity (at 3650 angstroms) 0.0023 0.145 a/lm
Cathode Radiant Sensitivity (at 3650 angstroms) 1.0 µa/Im
Luminous Sensitivity*" 3 3
Cathode Luminous Sensitivity+ nlm
Current Amplification 330,000 45000 plm
Equivalent Anode-Dark-Current Input at 25°C
0.375 max —
at a luminous sensitivity of 0.4 a/lm 7.6 —
Equivalent Noise Input**

"DC supply voltage (E) is connected across a voltage divider which provides 1/10 of E
between cathode and dynode No. 1, 1/10 of E for each succeeding dynode stage, and 1/10 of E
between dynode Na 9 and anode.
**With light input of 10 microlumens from a tungsten-filament lamp operated at color
temperature of 2570°K. Range of luminous sensitivity is 0.116 to 16 a/lm at 1000 volts;
0.016 at 1.85 a/lm at 750 volts.
+With light input of 0.01 lumen from atungsten-filament lamp operated at a color temperature
of 2870°K ; 100 volts applied between photocathode and all other electrodes connected as anode.
++At tube temperature of 25°C and with external shield connected to cathode; bandwidth is
equal to 1 cycle per second. Atungsten-filament light source at a color temperature of 28-0°K
is interrupted at a low audio frequency to produce incident radiation pulses alternating be-
tween zero and the value stated. The "on" period of the pulse is equal to the "off" period.

94 RCA Phototube Manual

2.5 TYPICAL ANODE CHARACTERISTICS
TYPE IP22 2000
_VOLTS/STAGE=100

2

ANODE MILLIAMPERES 1500
I.5

LIGHT FLUX—MICROLUMENS=1000

500
0.5

0 50 100 150 200 250 s2crn-6sesn
ANODE-TO-DYNODE-No.9 VOLTS

SENSITIVITY AND CURRENT
AMPLIFICATION CHARACTERISTICS
TYPE IP22
VOLTS /STAGE = I/10 E
ANODE-TO-DYNODE-No.9 VOLTS=1/10 E

pp —___-~~~~, 4
a mm0 ro a mm0 ro a~~~~~111112
~~~~~IIIII, DT
~~~~~~~~~~ e TYPICAL ANODE-DARK-CURRENT
Y—AMPERES/LUMEN (COLOR TEMP. 2870° K) ~~~~~~~~~~ CHARACTERISTIC
6

___--~~~,~4 TYPE IP22
~~~~■11.112
~~~ , .1111.,06 I 6 TUBE TEMPERATURE=25° C

YZ

~~~ \4 ~~~~G~ e Gf 4
0~
6 IJ 2 UNSTABLE

0 o~ loe

9F' ~~' / / ~' as 6
~y~\ ,
~ P+~~.~~F~°P`~~//~111,~ Fz 4

~ mmQ n~ a mm— ~ !t' ~ ~. ~~~~~~ g a wl- 2
~ ~ tAQ y /~t~~~ a~ J2

—~ ~. p ~~~~ -~~~~~4 ¢ 108

1Q~I~ 1 I~~~~~6 ~j 6 466x1 2 966 2 4 6610 2 46i~

wc~ 4 LUMINOUS SENSITIVITY—
AMPERES/LUMEN 92cs-96607
2

16

2
Q01

2 /- 5~~y'~~~111112

W as~ _`~`~,~_~_,Itll pa

/~~~~111112
ooa %~~~~IIIIIID2
e e9
5p0 ~ 1000 1250

ANODE-TO-CATHODE SUPPLY VOLTS (E
92CM-9674T1

Technical Data 95

NINE-STAGE, side-on type having S-5 response. Wave- 1P28

length of maximum response is 3400 ± 500 ang- MULTIPLIER
PHOTOTUBE
stroms. This type makes use of cesium-antimany
dynodes and acesium-bismuth, opaque photocathode.
Window material is Corning No. 9741 ultraviolet-
transmitting glass or equivalent. Tube weighs ap-
proximately 1.6 ounces and has a non-hygroscopic
base. For outline and terminal-connection diagram,
refer to type 1P21.

DIRECT INTERELECTRODE CAPACITANCES (Approx.): 4.4 pf
6 pf
Anode to Dynode Na. 9
Anode to All Other Electrodes 1250 msa volts
volts
MAXIMUM RATINGS (Absolute-Maximum Values): 250 max volts
volts
Supply Voltage (DC or Peak AC) 250 msa
Between anode and cathode 250 max ma
Between anode and dynode No. 9 0.5 max
Between consecutive dynodes °C
Between dynode No. 1 and cathode 75 msa

Average Anode Current
Ambient Temperature

ANODE MILLIAMPERES ~ TYPICAL ANODE CHARACTERISTICS
100 TYPEIP28
4 90 VOLTS/STAGE=100
80
3
/ 70

2 SIN 60
~!
50
I~ _LIGHT FLUX—MICROLUMENS=40

30
20

10

50 100 150 200 250 92CM-663273
ANODE-TO-DYNODE-No.9 VOLTS

TYPICAL CHARACTERISTICS: 1000 750 volts
61800 7900 a/w
DC Supply Voltage* 0.05 a/w
Radiant Sensitivity (at 3400 angstroms) 0.05
Cathode Radiant Sensitivity (at 3400 angstroms) 6.4 a/lm
Luminous Sensitivity 50 a/lm
30 40 µa/lm
At 0 cps** 40 0.16
With dynode No. 9 as output electrode 1.26 nlm
Cathode Luminous Sensitivity* nlm
Current Amplification (in millions) 1.25 max —
2 max — plm
Equivalent Anode-Dark-Current Input at 25°C pw
0.75 —
at a luminous sensitivity of 20 a/lm: 0 OGOa5 —
With anode as output electrode
With dynode No. 9 as output electrode

Equivalent Noise Input:
Luminous"
Ultraviolet3

*DC supply voltage (E) is connected across a voltage divider which provides 1/10 of E bo-
tween cathode and dynode No. 1, 1/10 of E for each succeeding dynode stage, and 1/10 of E
between dynode No. 9 and anode.

96 RCA Phototube Manual

•`With light input of 10 microlumens from a tungsten-filament lamp operated at color
temperature of 2870°K. Range of luminous sensitivity is 17.5 to 300 a/]m at 1000 volts.
*With light input of 0.01 lumen from a tungsten-filament lamp operated at a color tem-
perature of 2Si0°K ; 100 volts applied between photocathode and all other electrodes connected

as anode. temperature of 25°C and with external shield connected to cathode; bandwidth is
cycle per second. Atungsten-filament light source at a color temperature of 2870°,r
++At tube
equal to 1
is interrupted at a low audio frequency to produce incident radiation pulses alternating be-
tween zero and the value stated. The "on" period of the pulse is equal to the "off" period.
tUnder the same conditions as (**) except that a monochromatic source radiating at 2537

angstroms is used.

EQUIVALENT—NOISE—INPUT SENSITIVITY AND CURRENT
CHARACTERISTICS AMPLIFICATION CHARACTERISTICS

TYPE IP28 TYPE IP28
VOLTS/STAGE=100 VOLTS/STAGE= I/10 E
BANDWIDTH~I CPS ANODE-TO-DYNODE-No.9 VOLTS=I/IOE
EXTERNAL SHIELD-TO-ANODE VOLTS-1004 4
~~~~~1111
2 ~~~~~1111

~I2 —~~~ y ~~~
s ~~~ ,~~,~a~ ~~1
6 ~~_ ~y~,:,~,~.P, ~...11.!.1

4Z 1MEN (COLD ~ p+~\Jp~ P~ie9~`~,~1~ .111
W
nm0 N J -~ -LQ fv~I~~~~~~ ~ r~ CURRENT AMP
2J A 01 Cod
~ra~~%~~11~~1
Ib13 ~~ ~ A
e V, 1~' .~~~illli ~mO N
6N
—I~~~~~~~~t~
Z
//~~~~11111
4 Z~ ~~~~11111
JW
2j

WO
1014
9
6

4

2 ~~~~11111

10i5 ---■,1„1
s 7 &s
-12p -80 -40 0 40 500 1000 1250
ANODE-TO-CATHODE SUPPLY VOLTS (E
TUBE TEMPERATURE—°C
92CM-654774
92CM-750371

1P29 SIDE-ON type having S-3 response. ' Wavelength of
maximum spectral response is 4200 ± 1000 ang-
GAS stroms. This type makes use of a semicylindrical
PHOTODIODE photocathode, and has a direct interelectrode capaci-
tance of 3 picofarads. It weighs approximately 1.1
ounces.

MAXIMUM RATINGS (Absolute-Maximum Values): Rating 1 Rating 2 molts
µa/eq in
Anode-Supply Voltage (DC or Peak AC) SO max 100 max
Average Cathode-Current Density 50 max 25 max µa
Average Cathode Current 10 mas 5 max
Ambient Temperature 100 max °C
100 max

Technical Data 97

TYPICAL CHARACTERISTICS: 90 volts
0.011 a/w
Anode-Supply Voltage µa/lm
Radiant Sensitivity (at 4200 angstroms) 40
Luminous Sensitivity* µa
Gas Amplification Factor** 9 max
Anode Dark Current (at 25°C) 0.10 max

MINIMUM CIRCUIT VALUES: 80 or less 100 min volts

Anode-Supply Voltage 0.1 min — megohms
DC Load Resistance: megohms
0 min 2.5 min megohms
For average currents above 6 µa — 0.1 min megohms
For average currents below 5 µa
For average currents above 3 µa —
For average currents below 3 µa

•With light input of 0.1 lumen from atungsten-filament lamp operated at a color temperature
of 2870°K. Range of luminous sensitivity is 20 to 70 µa/Im.
•*Ratio of luminous sensitivities at 90 and 25 volts.

TYPICAL ANODE CHARACTERISTICS

ANODE MICROAMPERES n N
TYPE IP29 y
~~
6
J} O
4
v
2 ~~~

v

00h

0 20 40 60 80 100

ANODE VOLTS 92CM-647272

PHOTO- I X16
CATHODE MAX.

MIN. MAX.

NCV Va TB BULB 2 %B 4 /e
t 3j32 MAX.
DWARF-SMELL
NCIDENT~NADIATgN SMALL 4 -PIN

BASE
JEDEC NeA4-26

MAX.

98 RCA Phototube Manual

1P37 SIDE-ON type having S-4 response. Wavelength of
maximum spectral response is 4000 ± 500 angstroms.
Cs:~S This type makes use of a semicylindrical photo-
PHOTODIODE
cathode, and has a direct interelectrode capacitance
of 3 picofarads. It weighs approximately 1.1 ounces.
For curves of typical anode characteristics, refer to
type 5581. For outline and terminal connection
diagram, refer to type 1P29.

MAXIMUM RATINGS (Absolute-Maximum Values): Rating 1 Rating 2 volts
80 max 100 max µajsq in
Anode-Supply Voltage (DC or Peak AC) 50 max
Average Cathode-Current Density 10 max 25 max µa
Average Cathode Current 5 max
Ambient Temperature 75 max °C
75 max

TYPICAL CHARACTERISTICS: 90 volts
a/w
Anode-Supply Voltage 0.13
Radiant Sensitivity (at 4000 angstroms)
Luminous Sensitivity _ 135 µa/lm
Gas Amplification Factor*+
Anode Lark Current Iat 25°C) 5.6 maa

0.05 maz µa

MINIMUM CIRCUIT VALUES: 80 or less 100 volts

AnodeSupply Voltage 0.1 min — megohma
DC Load Resistance: 0 min — megohms
megohms
For average currents above 6µa — 2.6 min megohms
For average currents below 5µa — 0.1 min
For average currents abuve 3µa
For average currents below 3µa

With light input of 0.1 lumen from atungsten-filament lamp operated at a color temperature
of 2870°K. Range of luminous sensitivity is 75 to 206 µa/lm.

t'~Ratio of luminous sensitivities at 90 and 25 volts

1P39 SIDE-ON type having S-4 response. This type is
similiar to type 929, but has a maximum anode
VACUUM
PHOTODIODE dark current at 2b°C of O.00b microampere at 250

volts, and makes use of anon-hygroscopic base for
increased resistance between anode and photocath-
ode pins under adverse conditions of high humidity.

T9 Bta.B I ~16MAx. WCIDENT' RLDI4TION
5~8.
PNOTOCATNOOE MIN. vi

13/ly MIN. 2 iz
MA%.
INTERMEDIATE-
SNELL OCTAL ! ~D2 j 116
M X.
S-PIN BASE
JEDEC Ns
BS -10

IAL
7

1P40 SIDE-ON type having S-1 response. This type is
similar to type 930, but has a maximum anode dark
GAS current at 25°C of 0.005 microampere at 90 volts,
PNOTODIODE and makes use of anon-hygroscopic base for in-

creased resistance between anode and photocathode
pins under adverse conditions of high humidity.