S Photomultiplier _ Manual -= Photomultiplier Manual
The purpose of this Manual is to help designers and users of electro optical equipment using modern photomultiplier tubes achieve a better understanding of these devices. The Manual includes information on the construction and theory of operation of photomultipliers together with discussions of photomultiplier characteristics and the methods used to measure these characteristics. Information on typical applications is given along with basic data dealing with the measurement of light and radiant energy, noise statistics, spectral matching factors, and other topics as shown in the Table of Contents. Finally, a data section is included that gives typical characteristics of RCA photomultipliers.
RCA I Electronic Components I Harrison, N.J. n7~9
Copyright 1970 by RCA Corporation (All rights reserved under Pan-American Copyright Convention) Printed in U.S.A. 9/70 2
Contents
Introduction ...... 3 Photoemission and Secondary Emission ...... 5 Principles of Photomultiplier Design ...... 19 Basic Performance Characteristics ...... 30 Statistical Fluctuation and Noise ...... 56 Application of Photomultipliers ...... 77 Voltage-Divider Considerations ...... 103 Photometric Units and Photometric-to-Radiant Conversion .. 112 Radiant Energy and Sources...... 119 Spectral Response and Source-Detector Matching ...... 127 Technical Data ...... 134 Outlines ...... 161 Basing Diagrams ...... 174 Index ...... 183
Information furnished by RCA ,is bel ieved to be accurate and reliable. However, no responsibility is assumed by RCA for its use; nor for any infr ingements of patents or other rights of third parties which may result from its use. No license is granted by implication or otherwise under any patent or patent rights of RCA. 3
Introduction
HOTOMULTIPLIERS are ex Raman spectroscopy, laser ranging, Ptremely versatile photosensitive precision measurements, and commu devices that are widely used to de nications. Photomultipliers are used tect and measure radiant energy in in video playback equipment for the ultraviolet, visible, and near home entertainment, and in the infrared regions of the electromag audio-visual educational field; optical netic spectrum. In most of this range, character-recognition equipment util photomultipliers are the most "sensi izes photomultipliers for computer tive" detectors of radiant energy processing of financial statements. available. The basic reason for this The history of the development of superior sensitivity is the use of sec the photomultiplier is an interesting ondary-emission amplification which one. The use of secondary emission makes it possible for photomultipliers as a means for signal amplification to approach "ideal" device perform was proposed as early as 1919.1 The ance limited only by photoemission active development of photomulti statistics. Other outstanding charac pliers was begun approximately 35 teristics of photomultipliers include years ago by combining a photo a choice of photosensitive areas from emitter and one or more stages of a fraction of a square centimeter to secondary-electron-emission amplifi hundreds of square centimeters, in cation in one vacuum envelope. In ternal amplifications ranging from 1935, lams and Salzberg 2 of RCA 10 to as much as 109 in some types, described a single-stage photomulti extremely fast time response with plier which had a secondary-emission rise times as short as a fraction of amplification of 6; in 1940, Rajch a nanosecond, and output signal man and Snyder 3 of RCA described levels that are compatible with auxili a 9-stage electrostatically focused ary electronic equipment withou: photomultiplier that served as a pro need for additional signal amplifica totype design for a family of types tion. that are still used throughout the Photomultipliers are used with world today. Other forms of photo scintillators and Cerenkov radiators multipliers were developed in the for the detection and measurement late 1930's, and with the sharp in of X-rays, gamma rays, and energetic crease in research in atomic energy particles, and find wide usage in nu and nuclear physics, much progress clear research, industrial controls, in photomultiplier development was and space exploration. The availabil made in the 1940's and 1950's. The ity of many different types of lasers rapid growth of electro-optics in the has opened up many new fields of 1960's further advanced the develop application for photomultipliers in ment of photomultipliers. The ad- 4 RCA Photomultiplier Manual vent of the bialkali photocathode 6. G. Weiss, "On Secondary-Elec and more recently the Extended Red tron Multipliers," Z. t. Techn. Multi-Alkali (ERMA) photocathode Physik, Vol. 17, No. 12, pp. provided further improvements. Very 623-629 (1936). recently, advances in solid-state 7. W. Kluge, O. Beyer, and H. theory and semiconductor-materials Steyskal, "Photocells with Sec research have opened new vistas in ondary-Emission Amplification," the development of photomultipliers Z. t. Techn. Physik, Vol. 18, No. having performance capabilities that 8, pp. 219-228 (1937). . were literally impossible only a few 8. J. R. Pierce, "Electron-Multiplier years ago. Design," Bell Lab. Record, Vol. 16, pp. 305-309 (1938). REFERENCES 9. V. K. Zworykin and J. A. Rajch 1. J. Slepian, U. S. Patent 1,450.- man, "The Electrostatic Electron 265, April 3, 1923 (Filed 1919). Multiplier," Proc. IRE, Vol. 27, 2. H. lams and B. Salzberg, "The pp. 558-566 (1939). Secondary-Emission Phototube," 10. C. C. Larson and H. Salinger, Proc. [RE, Vol. 23, pp. 55-64 "Photocell Multiplier Tubes," (1935). Rev. Sci. Instr., Vol. 11, pp. 3. J. A. Rajchman and R. L. 226-229 (1940). Snyder, "An Electrostatically 11. R. C. Winans and J. R. Pierce, Focused Multiplier Phototube," "Operation of Electrostatic Electronics, Vol. 13, p. 20 Photomultipliers," Rev. Sci. (1940). [nstr., VO'I. 12, pp. 269-277 4. P. T. Farnsworth, "An Electron (1941). Multiplier," Electronics, VO'I. 7, 12. R. W. Engstrom, "Multiplier pp. 242-243 (1934). Phototube Characteristics; Ap 5. V. Zworykin, G. Morton, and plication to Low Light Levels," L. Malter, "The Secondary I. O. S. A., Vol. 37, No.6, pp. Emission Multiplier-A New 421-431 (1947). Electronic Device," Proc. IRE, Vol. 24, pp. 351-375 (1936). 5
Photoemission and Secondary Emission
HIS section covers the means by photoemission and then amplify the Twhich radiant energy is con signals by means of secondary emis verted into electrical energy in a sion. photomultiplier tube. A basic component arrangement of Photomultipliers convert incident a typical modern photomultiplier is radiation in the visible, infrared, and shown in Fig. 1. Radiant energy en ultraviolet regions into electrical sig ters the evacuated enclosure through nals by use of the phenomenon of a "window" which has a semi trans-
INCIDENT RADIATION
SEMITRANSPARENT r PHOTOCATHODE r-~~~==~~~
TYPICAL PHOTOCATHODE PHOTOELECTRON TO DYNODE No.1 ELECTRON - TRAJECTORIES OPTICS
VACUUM ENCLOSURE
ELECTRON MULTIPLIER
1-12: DYNODES 14: FOCUSING ELECTRODES 13:ANODE IS: PHOTOCATHODE
Fig. 1 - Schematic of typical photomultiplier showing some electron trajectories. 6 RCA Photomultiplier Manual parent photocathode deposited upon E - hv (1) its inner surface. The photocathode p - =~A emits photoelectrons by the process of photoemission, i.e., by the inter where v is the frequency of the in action of the incident radiant energy cident radiation, A is the wavelength with electrons in the photocathode of the incident radiation, h is Planck's material. Photoelectrons from all quantum of action, and c is the parts of the photocathode are ac velocity of light. Substitution of the celerated by an electric field so as to numerical values of these constants strike a small area on the first dy from Appendix A provides the fol node. Secondary electrons resulting lowing relation: from the process of secondary emis sion (i.e., from the impact of the photoelectrons on the first dynode) Ep = 123~.5 (2) are accelerated toward the second dynode by an electric field between where Ep is in electron-volts and A dynodes No. 1 and No. 2 and im is in nanometers. Thus, photons of pinge on the second dynode. The visible light which are in the wave impact of the secondary electrons on length range from 400 to 700 nano dynode No. 2 results in the release meters have energies ranging from of more secondary electrons which 3.1 to 1.8 electron-volts. are then accelerated toward dynode No.3. This process is repeated un Fundamentals of Photoemission til the electrons leaving the last dy An ideal photocathode has a node are collected by the anode and quantum efficiency of 100 per cent; leave the photomultiplier as the out i.e., every incident photon releases put signal. If, on the average, 4 sec one photoelectron from the material ondary electrons are liberated at each into the vacuum. All practical photo dynode for each electron striking that emitters have quantum efficiences be dynode, the current amplification for low 100 per cent. To obtain a a 12-stage photomultiplier is 412, or qualitative understanding of the vari approximately 17 million. Thus, the ations in quantum efficiency for liberation of a single photoelectron different materials and for different at the photocathode results in 17 wavelengths or photon energies, it is million electrons being collected at useful to consider photoemission as the anode. Because this pulse has a a process involving three steps: (1) time duration of approximately 5 absorption of a photon resulting in nanoseconds, the anode current is the transfer of energy from photon approximately 1 milliampere at the to electron, (2) motion of the elec peak of the anode-current pulse. tron toward the material-vacuum interface, and (3) escape of the elec PHOTO EMISSION AND tron over the potential barrier at PHOTOCATHODES the surface into the vacuum. Energy losses occur in each of Photoemission is a process in these steps. In the first step, only which electrons are liberated from the absorbed portion of the incident the surface of a material by the inter light is effective and thus losses by action of photons of radiant energy transmission and reflection reduce with the materiaJ.1 The energy of a the quantum efficiency. In the sec photon Ep is given by ond step, the photoelectrons may Photoemission and Secondary Emission 7 lose energy by collision with other models that describe semiconductor electrons (electron scattering) or photoemitters is illustrated in its with the lattice (phonon scattering). simplest form in Fig. 2. Electrons Finally, the potential barrier at the can have energy values only within surface (called the work function in well defined energy bands which are metals) prevents the escape of some separated by forbidden-band gaps. electrons. EA - ELECTRON AFFINITY The energy losses described vary EG-BAND GAP T ,..-----r--- from material to material, but a CONDUCTION BAND EA I major difference between metallic and semiconducting materials makes f II EA + EG separate consideration of each of FORBIDDEN BAND EG --+- these two groups useful. In metals, a ------~~-~-~i----LJ-- large fraction of the incident visible VALENCE BAND I FERMI light is reflected and thus lost to I LEVEL the photoemission process. Further - SEMICONDUCTOR I VACUUM _ losses occur as the photoelectrons rapidly lose energy in collisions with Fig. 2 - Simplified semiconductor the large number of free electrons energy-band model. in the metal through electron-electron At O·K, the electrons of highest en scattering. As a result, the escape ergy are in the so-called valence depth, the distance from the surface band and are separated from the from which electrons can reach the empty conduction band by the band surface with sufficient energy to over gap energy EG • The probability that come the surface barrier, is small, a given energy level may be occu and is typically a few nanometers. pied by an electron is described by Finally, the work function of most Fermi-Dirac statistics and depends metals is greater than three electron primarily on the difference in energy volts, so that visible photons which between the level under consideration have energies less than three electron and a reference level called the volts are prevented from producing "Fermi level." As a first approxima electron emission. Only a few metals, tion, it may be said that any energy particularly the alkali ones, have levels which are below the Fermi work-function values low enough to level will be filled with electrons, and make them sensitive to visible light. any levels which are above the Fermi Because of the large energy losses level will be empty. At temperatures in absorption of the photon and in higher than WK, some electrons in the motion of the photoelectron the valence band have sufficient en toward the vacuum (the first and ergy to be raised to the conduction second steps described above), even band, and these electrons, as well as the alkali metals exhibit very low the holes in the valence band created quantum efficiency in the visible re by the loss of electrons, produce gion, usually below 0.1 per cent (one electrical conductivity. Because the electron per 1000 incident photons). number of electrons raised to the As expected, higher quantum effi conduction band increases with tem ciencies are obtained with higher perature, the conductivity of semi photon energies. For 12-electron-volt conductors also increases with tem photons, quantum yields as high as perature. Light can be absorbed by 10 per cent have been reported. valence-band electrons only if the The concept of the energy-band photon energy of the light is at least 8 RCA Photomultiplier Manual equal to the band-gap energy E(;. fraction of the incident light, photo If, as a result of light absorption, electrons can escape from a greater electrons are raised from the valence distance from the vacuum interface, band into the conduction band, and the threshold wavelengths can photoconductivity is achieved. For be made longer than those of a metal. photoemission, an electron in the Thus, it is not surprising that all conduction band must have energy photoemitters of practical importance greater than the electron affinity EA' are semiconducting materials. The additional energy EA is needed In recent years, remarkable im to overcome the forces that bind the provements in the photoemission electron to the solid, or, in other from semiconductors have been ob words, to convert a "free" electron tained through deliberate modifica within the material into a free elec tion of the energy-band structure. tron in the vacuum. Thus, in terms The approach has been to reduce of the model of Fig. 2, radiant energy the electron affinity, EA. and thus to can convert an electron into an in permit the escape of electrons which ternal photoelectron (photoconduc have been excited into the conduc tivity) if the photon energy exceeds tion band at greater depths within EG and into an external photoelectron the material. Indeed, if the electron (photoemission) if the photon energy affinity is made less than zero (the exceeds (EG + E A ). As a result, vacuum level lower than the bottom photons with total energies Ep less of the conduction band, a condition than (EG + EA ) cannot produce described as "negative electron affin photoemission. ity" and illustrated in Fig. 3), the The following statements can escape depth may be as much as therefore be made concerning photo 100 times greater than for the emission in semiconductors. First, normal material. The escape depth light absorption is efficient if the of a photoelectron is limited by the photon energy exceeds EG• Second, energy loss suffered in phonon scat energy loss by electron-electron scat tering. Within a certain period of tering is low because very few free time, of the order of 10-12 second, electrons are present; thus, energy the electron energy drops from a loss by phonon scattering is the pre level above the vacuum level to the dominant loss mechanism. The es bottom of the conduction band from cape depth in semiconductors is which it is not able to escape into therefore much greater than in the vacuum. On the other hand, the metals, typically of the order of tens electron can stay in the conduction of nanometers. Third, the threshold wavelength, which is determined by the work function in metals, is given VACUUM LEVEL by the value of (EG + EA) in semi conductors. Synthesis of materials with values of (EG + EA) below 2 electron-volts has demonstrated that threshold wavelengths longer than SURFACE those of any metal can be obtained REGION in a semiconductor. Semiconductors, therefore, are superior to metals in Fig. 3 - Semiconductor energy-band all three steps of the photoemissive model showing negative electron process: they absorb a much higher affinity. Photoemission and Secondary Emission 9 band in the order of 10-10 second The changes which result in the without further loss of energy, i.e., energy-band structure are two-fold. without dropping into the valence In the first place, the acceptance of band. If the vacuum level is below electrons by the p-type impurity the bottom of the conduction band, levels is accompanied by a down the electron will be in an energy state ward bending of the energy bands. from which it can escape into the This bending can be understood by vacuum for a period of time that is observing that a filled state must be, approximately 100 times longer than in general, below the Fermi level; if an energy above the bottom of the whole structure near the surface the conduction band is required for is bent downward to accomplish this escape, as in the materials repre result. In the second place, the sented by Fig. 2. Therefore, a ma potential difference between the terial conforming to the conditions charged electropositive layer (cesium) of Fig. .3 has a greatly increased and the body charge (filled zinc escape depth. Under such circum levels) results in a further depres stances, the photosensitivity is sig sion of the vacuum level as a result nificantly enhanced. Substantial re of a dipole moment right at the sponse is observed even for photons surface. with energies close to that of the Another way to describe the re band gap where the absorption is duction of the electron affinity is to weak. Efficient photoemission in this consider the surface of the semicon case results only because of the ductor as a capacitor. The charge on greater escape depth. one side of the capacitor is repre The reduction of the electron affin sented by the surface layer of cesium ity is accomplished through two ions; the other charge is represented steps. First, the semiconductor is by the region of filled acceptor levels. made strongly p-type by the addition The reduction in the electron affinity of the proper "doping" agent. For is exactly equal to the potential dif example, if gallium arsenide is the ference developed across the ca host material, zinc may be incor pacitor. porated into the crystal lattice to a In a more rigorous analysis, the concentration of perhaps 1,000 parts amount by which the energy bands per million. The zinc produces iso are bent is found to be approximately lated energy states within the for equal to the band-gap, and the bidden gap, near the top of the vacuum level is lowered until the valence band, which are normally absorption level of the electroposi empty, but which will accept elec tive material is essentially at the top trons under the proper circumstances. of the valence band. The p-doped material has its Fermi level near the top of the valence Measurements of band. The second step is to apply Photocathode Sensitivity to a semiconductor a surface film of an electropositive material such There are various ways of specify as cesium. Each cesium atom be ing the sensitivity of a photocathode. comes ionized through loss of an For many purposes, a knowledge of electron to a p-type energy level the response of the photocathode to near the surface of the semiconduc particular wavelengths or equivalent tor, and is held to the surface by energies of radiation is necessary. electrostatic attraction. Sometimes it is convenient to have 10 RCA Photomultiplier Manual a single number to describe the re It should be noted that the spec sponse of the photocathode to some tral response of photoemitters is standard broad-band source, such as modified by the window material. a tungsten lamp. Therufore, the short-wavelength limit Variation in photocathode sensi of the spectral response is generally tivity over the range of wavelengths a characteristic of the window rather to which the photocathode material than of the photocathode. is sensitive is usually described by A common practice is to specify means of a spectral-response curve the response of the device to the in such as the one shown in Fig. 4. cident flux from a tungsten lamp with The abscissa may represent the wave a lime-glass window operating at a length in nanometers or angstroms, color temperature of 2854°K. The or the photon energy in electron units of measurement are most fre volts. The values can be converted quently amperes per lumen. Some from one scale to the other by use times sensitivity is expressed in of Eq. 1. terms of photocurrent and the total The ordinate can also be expressed incident radiant power from the in many ways. The most common lamp on the photocathode window. scale is amperes per incident watt, This sensitivity is commonly stated as shown; a second common scale is in amperes per watt. It is the general that of quantum efficiency, the av practice to give sensitivity measure erage numbers of photoelectrons ments in terms of incident flux. For emitted per incident photon. some purposes, such as the study of photo emission phenomena, it may 100 be useful to specify the sensitivity in 80 terms of absorbed flux. 60 Opaque and Semitransparent ... 40 !i' Photocathodes ...... 3< E The photocathode consists of a I 20 material that emits electrons when >- I- ;; exposed to radiant energy. There ;:: are two types of photocathodes. In in 10 ....z one, the opaque photocathode, the U) 8 .... light is incident on a thick photo I- 6 ::> emissive material and the electrons -' 0 U) are emitted from the same side as m.. 4 that struck by the radiant energy. In the second type, the semitranspar ent photocathode, the photoemissive 2 material is deposited on a trans parent medium so that the electrons are emitted from the side of the 200 400 600 800 photocathode opposite the incident WAVELENGTH - NANOMETERS I I I I radiation. 8 6 4 2 PHOTOELECTRON ENERGY -eV Because of the limited escape depth of photoelectrons, the thick Fig. 4 - Typical spectral-response curve for a bialkali photocathode with ness of the semitransparent photo a 008 lime-glass window. cathode film is critical. If the film Photoemission and Secondary Emission 11 is too thick, much of the incident and bialkali (K2CsSb). Typical spec radiant energy is absorbed at a dis tral-response curves for these ma tance from the vacuum interface terials are shown in Fig. 4 and Fig. greater than the escape depth; if the 6. Additional information about film is too thin, much of the incident these and other photocathodes of radiant energy is lost by transmis practical importance is shown In sion. Because radiant-energy absorp Table I. tion varies with wavelength, as shown The long-wavelength response of in Fig. 5, the spectral response of a the multialkali photocathode has semitransparent photocathode can be been extended recently by process controlled to some extent by control ing changes including the use of an of its thickness. With increasing increased photocathode-film thick thickness, for example for alkali ness at the expense of the short antimonides, blue response decreases wavelength response. A typical and red response increases because of spectral-response curve for one of the difference in absorption coeffi these Extended Red Multi-Alkali cients. (ERMA) photocathodes is shown in Fig. 7 in comparison with a con ventionally processed multi alkali photocathode.
I- Z ...u a: 60 ...0.. I l- z ;::0 i 0.. 40 " a: <0: 0 (f) E Photocathodes of Practical Importance The photocathodes most com monly used in photomultipliers WAVELENGTH - NANOMETERS are silver-oxygen-cesium (Ag-O-Cs), cesium-antimony (Cs3Sb), multi Fig. 6 - Typical spectral-response alkali or trialkali [( Cs) N a2KSb] * , curves, with 008 lime-glass window, for (a) silver-oxygen-cesium (Ag-O-Cs), * The parenthetical expression (Cs) in (b) cesium-antimony (Cs,Sb), (c) multi- dicates trace quantities of the element. alkali or trialkali [(Cs)Na.KSb]. 12 RCA Photomultiplier Manual Table I - Nominal Composition and Characteristics of Various Photocathodes. WIYlltl&lh Conversion of Quantum DIrk Nominal Response Typ. of Envelope Factorb LuminDIIS Maximum Sensitivity Elliciency Emillion Composition DesltJlltion Photo· Material" (lumen/wall Sensitivity Response, at >.max at >'mll at25°C cathod. at >.max) 0.max (mA/wall) (percent) Ax 10-1I/em 2 (nm) Ag-O-Cs S-1 0 0080 92.7 25 800 2.3 0.36 900 Ag-O-Rb S-3 0 0080 285 6_5 420 1.8 0_55 CS3Sb S-19 0 SiOz 1603 40 330 64 24 0.3 CS3Sb S-4 0 0080 1044 40 400 42 13 0.2 Ca3Sb S-5 0 9741 1262 40 340 50 18 0.3 CS3Bi S-8 0 0080 757 3 365 2.3 0.77 0.13 Ag-Bi-O-Cs S-10 S 0080 509 40 450 20 5.6 70 CS3Sb S-13 S SiOZ 799 60 440 48 14 4 CS3Sb S-9 S 0080 683 30 480 20 5.3 CS3Sb S-l1 S 0080 808 60 440 48 14 3 CS3Sb S-21 S 9741 783 30 440 23 6.7 CsgSb S-17 O· 0080 667 125 490 83 21 1.2 NazKSb S-24 S 7056 1505 32 380 64 23 0.0003 K-Cs-Sb S 7740 1117 80 400 89 28 0.02 (Ca)NazKSb S SiOZ 429 150 420 64 18.9 0.4 (Cs)NazKSb S-20 S 0080 428 150 420 64 19 0.3 (Cs)NazKSb S-25 S 0080 276 160 420 44 13 (Cs)NazKSb ERMA- S 7056 169 265 575 45 10 1 Ga-As Oh 9741 148 250 450 37 10 0.1 Ga-As-P Oi Sapphire 310 200 450 61 17 0.01 InGaAs-CsOd Oh 0080 266 260 400 71 22 I d CszTe S liF 120 12.6 13 Csi S liF 120 24 20 Cui S liF 150 13 10.7 • Numbers refer to the following glasses: Princeton, New Jersey. The dark emission indi 008()-Corning Lime Glass cated is a calculated value based on a bandgap 9741-Corning Ultraviolet Transmitting Glass of 1.1 eV for the particular composition studied. 7056--Corning Borosilicate Glass See also B. F. Williams, "lnGaAs-CsO, A Low 774()-Corning Pyrex Glass Work Function (Less Than 1.0 eV) Photoemitter," SiQ,,-Fused Silica (Suprasil-Trademark of Engle Appl. Phys. Letters 14, No.9, p. 273 (1969). hard Industries, Inc., Hillside, New Jersey) • Not relevant. b These conversion factors are the ratio of the radiant sensitivity at the peak of the spectral f Data unavailable; expected to be very low. response characteristic in amperes per watt to • Reflecting substrate. the luminous sensitivity in amperes per lumen for a tungsten test lamp operated at a color tempera h Single crystal. ture of 2854°K. I Polycrystalilne. _ An RCA designation for "Extended-Red MultialkalL" 0= Opaque d An experimental photocathode, private communi cation from B. F. Williams, RCA Laboratories, S = Semitransparent Photoemission and Secondary Emission 13 1.00/0 100 (3) where Ns is the average number of secondary electrons emitted for Ne .... primary electrons incident upon the 1- surface. ~ " 0~~ :> ,ooloo.~ ~ 10 6 z w (f) f- 4 w 6 f- .... - ~ ~ (f)z ,of. w 8 (f) w 6 f- ::l -' 0 4 (/) Fig. 7 - Typical spectral-response New Photocathode Materials 200 400 600 800 1000 The negative-electron-affinity ma WAVELENGTH - NANOMETERS 'erials described earlier are used in Fig. 8 - Typical spectral-response Ipaque photocathodes. Typical spec curves for GaAs(Cs) and GaAs.P1_.(Cs) l ·aI-response curves for GaAs(Cs) opaque photocathodes. a Id GaAsxP1_x(Cs) opaque photo Cl thodes are shown in Fig. 8. De Fundamentals of velopment of negative-electron-affin Secondary Emission ity materials is proceeding rapidly. Therefore, improved sensitivities and extended spectral ranges may be an The physical processes involved ticipated. in secondary emission are in many respects similar to those already de SECONDARY EMISSION scribed under Fundamentals of Photoemission. The main difference When electrons strike the surface is that the impact of primary elec of a material with sufficient kinetic trons rather than incident photons energy, secondary electrons are causes the emission of electrons. The emitted. The secondary-emission steps involved in secondary emission ratio or yield, 8, is defined as follows: can be stated briefly as follows: 14 RCA Photomultiplier Manual 1. The incident electrons interact to 2200 electron-volts. The total with electrons in the material and number of excited electrons pro excite them to higher energy states. duced by a primary is indicated by 2. Some of these excited electrons the area of the individual rectangles move toward the vacuum-solid inter in the figure. face. As an excited electron in the bulk 3. Those electrons which arrive at of the material moves toward the the surface with energy greater than vacuum-solid interface, it loses en that represented by the surface bar ergy as a result of collisions with rier are emitted into the vacuum. other electrons and optical phonons. When a primary beam of electrons The energy of the electron is very impacts a secondary-emitting ma rapidly dissipated as a result of terial, the primary-beam energy is these collisions, and it is estimated dissipated within the material and a that the energy of such an electron number of excited electrons are pro will decay to within a few times the duced within the material. Based on mean thermal energy above the bot experimental data, approximations tom of the conduction band within 12 have been made 2 based on the as 10- second. If the electron ar sumption that the range of primary rives at the vacuum-solid interface electrons varies as the 1.35 power of with energy below that required to the primary energy and that the num traverse the potential barrier, it can ber of electrons excited is uniform not escape as a secondary electron. throughout the primary range. The Therefore, only those electrons ex numbers of excited electrons pro cited near the surface of the ma duced are indicated in Fig. 9 for terial are likely to escape as second primary energies varying from 400 ary electrons. The probability of .5 r--,...-,---,---.,----r----,---,---,----,--r--r----.IO .., 9 8 T E c I 7 en 0 400 1200 1600 2200 z >- 0 '= Eo a: -' .3 6 t- m 0 0 0 40 60 eo 100 120 DEPTH - NANOMETERS Fig. 9 - An illustration showing the processes of secondary emission. See text for explanation. Photoemission and Secondary Emission 15 escape for an excited electron is Experimental secondary-emission assumed to vary exponentially with yield values are shown as a func the excitation depth, as indicated in tion of primary-electron energy in Fig. 9. If the product of the escape Fig. 10 for MgO, a traditional sec function and the number of excited ondary-emission material, and for electrons (which is a function of GaP(Cs), a recently developed nega primary energy and depth) is inte tive-electron-affinity material. Al grated, a secondary-emission-yield though both of these materials display function may be obtained, as indi the general characteristics of sec cated in the insert at the top of Fig. ondary emission as a function of 9. The model which has been as primary energy, as expected from sumed thus explains the general the model illustrated in Fig. 8, the characteristics of secondary emis secondary-emission yield for GaP(Cs) sion as a function of primary energy. increases with voltage to much Secondary-emission yield increases higher values. Even though electrons with primary energy, provided the are excited rather deep in the nega excited electrons are produced near tive-electron-affinity material and the surface where the escape prob lose most of their excess energy as ability is high. As the primary a result of collisions, many still es electron energy increases, the number cape into the vacuum because of of excited electrons also increases, the nature of the surface barrier. but the excitation occurs at greater When secondary electrons are depths in the material where escape emitted into the vacuum, the spread is much less probable. Consequently, of emission energies may be quite the secondary-emission yield even large, as illustrated in the curve of tually reaches a maximum and then Fig. 11 for a positive-affinity emit decreases with primary energy. ter. The peak at the right of the '"I o .J ~ 200 >- z Q en i'" w > « o'" is 100 u w '" M9 0 (MEASURED) 2 4 6 8 16 18 20 PRIMARY ENERGY - keY Fig. 10 - Typical curve of secondary-emission yield as a function of primary electron energy in Gap(Cs) and MgO. 16 RCA Photomultiplier Manual (/) z 0 100,....---. 0:: f-- U W ...J W >- 80 0:: 0 '"z 0 u w (/) 0 W f-- ~ ::;: w 40 "- 0 0:: W al ::;: ::> 20 z w > i= ...J '"W 0:: o 60 100 120 140 160 ELECTRON'ENERGY - ELECTRON VOLTS Fig. 11 - Typical secondary-electron energy distribution; peak at right is caused by reflected primary electrons. curve does not represent a true sec grade the signal-to-noise ratio in a ondary, but rather a reflected pri photomultiplier; the greatest contri mary. Data are not available for the bution to the noise occurs at the first emission energies from a negative dynode. If the secondary-emission electron-affinity material, but they ratio of the first dynode is improved are expected to be considerably less through the use of a negative-elec than for positive-affinity materials. tron-affinity material, such as GaP, Negative electron affinity has not the' emission statistics of the first dy only expanded the field of photo node are markedly improved and the emission, but has proved very im performance of the multiplier more portant in the development of greatly closely resembles that of a perfect improved secondary-emitting ma amplifier. Photomultiplier statistics terials for use as dynodes in photo are explained in greater detail in the multipliers. With negative-electron section on Statistical Fluctuation and affinity materials, such as GaP, the Noise. secondary-emission ratio increases The incorporation of high-gain sec almost linearly up to very high pri ondary-emission materials through mary energies; factors up to 130 out the dynode chain makes it have been measured. This factor possible to design photomultipliers compares with typical values of less with fewer stages for given amplifi than 10 for previously used dynode cations. The use of fewer stages also materials. reduces the variation of gain with voltage changes. Effects of Secondary-Emission Statistics Time Lag in Photoemission and Secondary Emission The statistical distribution of sec Because both photoemission and ondary-emission yields tends to de- secondary emission can be described Photoemission and Secondary Emission 17 in terms of the excitation of elec in photoemission, however, is not trons within the volume of the solid well established. One investigation 8 and the subsequent diffusion of these determined a limit of 3 X 10-9 sec electrons to the surface, a finite time ond. From careful measurements 9 interval occurs between the instant of the time performance of fast that a primary (photon or electron) photomultipliers it can be inferred strikes a surface and the emergence that the limit must be at least an of electrons from the surface. order of magnitude smaller. Furthermore, in the case of second While these limits are a useful ary emission, the secondaries can be guide to the type of time perform expected to reach the surface over a ance to be expected in present period of time. Within the limita photomultipliers, they will probably tions of a mechanistic approach to have less significance as new photo a quantum phenomenon, time inter multipliers using semiconducting vals for metals or insulators of the photoemitters and secondary emit order of 10-13 to 10-14 second may ters are developed. Semiconductors be estimated from the known energy having minority-carrier lifetimes of of the primaries, their approximately the order of microseconds are now known range, and the approximately available. Probably, by combination known diffusion velocities of the in of this characteristic with negative ternal electrons. In negative-electron electron affinity, higher gains and affinity semiconductors, it is known quantum efficiencies can be achieved, that the lifetime of internal "free" but at a sacrifice of time response or electrons having quasi-thermal ener band width. However, the first gen gies (i.e. electrons near the bottom eration of negative-electron-affinity of the conduction band) can be of emitters (e.g., GaP) has actually re the order of 10-10 second. sulted in photomultipliers having Thus far experiments have pro better time performance because a vided only upper limits for the time smaller number of stages operating lag of emission. In the case of sec at higher voltage can be used. At ondary emission, a variety of ex this time it can only be concluded periments have established limits. that in the future photomultipliers Several investigators 3, 4, 6 have de will probably be designed to match duced limits from the measured per in more detail the requirements of formance of electron tubes using a particular use. secondary emitters. Others, making direct measurements of these limits, determined the time dispersion of REFERENCES secondary emission by letting short electron bunches strike a target and 1. A. H. Sommer, Photoemissive comparing the duration of the result Materials, John Wiley, New York ing secondary bunches with the (1968) measured duration of the primary 2. R. E. Simon and B. F. Williams, bunch. By this means an upper limit "Secondary Electron Emission," of 6 X 10-12 second was determined IEEE Trans. Nucl. Sci., Vol. NS- for platinum 7 and an upper limit 15, No.3, p. 167 (1968) of 7 X 10-11 second for an MgO 3. G. Diemer and J. L. H. Jonker, layer 5 formed on the surface of an "On the Time Delay of Second AgMg alloy. ary Emission," Philips Research The upper limit for the time lag Repts., Vol. 5, p. 161 (1950) 18 RCA Photomultiplier Manual Foerster, 4. M. H. Greenblatt and P. A. 7. E. W. Ernst and H. Von Secondary Miller, Jr., "A Microwave Second "Time Dispersion of Appl. ary-Electron Multiplier," Phys. Electron Emission," 1. (1955) Rev .. Vol. 72, p. 160 (1947) Phys., Vol. 26, p. 781 J. W. Beams, 5. M. H. Greenblatt, "On the Mea 8. E. O. Lawrence and surement of the Average Time "Instantansity of the Photoelectric Vol. 29, p. Delay in Secondary Emission," Effect," Phys. Rev., RCA Rev., Vol. 16, p. 52 (1955) 903 (1927) Kerns, and R. F. 6. C. G. Wang, "Reflex Oscillators 9. M. Birk, Q. A. of the Using Secondary-Emission Cur Tusting, "Evaluation Phys. Rev., Vol. 68, p. 284 C-70045A High-Speed Photomul rent," Sci. (1945) tiplier," IEEE Trans. Nucl. Vol. NS-ll, p. 129 (1964) 19 Principles of Photomultiplier Design HIS section describes the physi high-speed electron multiplier pro Tcal and electrical parameters of vides a photomultiplier having a interest in the design of photomul very fast time response. Some tube tipliers. The effects of dynode and types utilize a spherical-section anode structures on such parameters photocathode on a plano-concave as time response, sensitivity, and dark faceplate to facilitate scintillator current are discussed. A section cov coupling. This design is usually ering ruggedized photomultipliers de limited to faceplate diameters of two scribes their structure and method of inches or less because of the exces testing. sive thickness of the glass at the edge. The plano-concave faceplate may PHOTOCATHODE-TO-FIRST also contribute to some loss in uni DYNODE REGION formity of sensitivity because of in ternal reflection effects near the thick As mentioned in the preceding edge of the photocathode. section, two classes of photocathodes ar.e used in photomultipliers: opaque Design Analysis and semitransparent. Because opaque photocathodes are an integral part Photocathode-to-first-dynode struc of the electron-multiplier structure, tures are usually axially symmetric, they are discussed later in connec with the symmetry axis passing tion with the electron multiplier. through the center of the photo Semitransparent photocathodes are cathode. This symmetry simplifies formed on planar or spherical-section the electron-optical analysis of the windows which are part of the photo photocathode-to-first-dynode region cathode-to-first dynode region. Semi because the problem is reduced from transparent photocathodes have large, three dimensions to two. The first readily accessible light-collection step in the analysis is the solution of areas which may be directly coupled the Laplace equation, V2 V = o. to scintillation crystals. Because a solution of this equation A planar-photocathode design, in closed form is impossible for most such as that shown in Fig. 12(a), geometries, numerical solutions are provides excellent coupling to a scin obtained;1 one solution 2 results in tillation crystal, but its time re the photocathode-to-first-dynode re sponse is not as good as that of a gion shown in Fig. 12(b). Perform spherical-section photocathode. The ance is determined by use of a spherical-section photocathode shown computer to trace equipotential lines in Fig. 12(b) when coupled to a and electron trajectories which can 20 RCA Photomultiplier Manual FOCUSING ELECTRODE F FACEPLATE- INCIDENT RADIATION SEMI- I TRANSPAR~ PHOTOCATHODE I I INTERNAL CONDUCTIVE I LCOATING I FRONT - END REG ION (a) /FACEPLATE FOCUSING ELECTRODE INCIDENT RADIATION FRONT-END REGION (b) Fig. 12 - Front-end region of a photomultiplier with (a) a planar photocathode, and (b) a spherical-section photocathode. then be superimposed on a schematic velocities,3 some photoelectrons be diagram of the tube structure as gin their trajectories with unfavorable shown in Fig. 13. angles of launch and are not col Collection efficiency and time re lected on the useful area of the first sponse may be predicted from an dynode. Although collection effi analysis of the electron trajectory. ciency is difficult to measure, it can The collection efficiency is defined be calculated by computer by the as the ratio of the number of photo use of Monte Carlo simulations.4 electrons which land upon a useful In modern photomultiplier struc area of the first dynode to the num tures, first-dynode collection efficien ber of photoelectrons emitted by the cies vary from 85 to 98 per cent. photocathode. If all the photoelec Ideally, the emitted photoelectrons trons began their trajectories at the should converge to a very small area surface of the photocathode with on the first dynode; in practice this zero velocity, a collection efficiency electron-spot diameter is approxi of 100 per cent would be possible; mately 1/ 8 to 1/ 20 of the cathode however, because of the finite initial diameter, depending upon tube type. Principles of Photomultiplier Design 21 TYPICAL ELECTRON TRAJECTORIES TYPICAL EQUIPOTENTIAL LINES Fig. 13 - Cross section of a photomultiplier showing equipotential lines and electron trajectories that were plotted by computer. first dynode allows simple photoelec ELECTRON MULTIPLIERS tron collector systems to be used. This structure is used only when In a photomultiplier having an time response is not of prime consid opaque photocathode, the photo eration. The relatively slow time re cathode is part of the electron sponse is a result of the weak elec multiplier or dynode-chain structure. tric field at the surfaces of the dynode The circular cage structure, shown vanes. This structure is the most in Fig. 14(a) for a "side-on" type of flexible as to the number of stages. photomultiplier, employs an opaque The box-and-grid dynode struc photocathode and is one of the earli ture, shown in Fig. 16, also has a est mUltiplier configurations. Its rela large first-dynode collection area. tively fast time response is a result The time response of this structure of its small size and high interdynode is similar to that of a venetian-blind field strengths. This same structure multiplier (slow) because of the weak is also used in "head-on" type photo electric field within the box. multipliers, as shown in Fig. 14(b). A linear-multiplier structure, such Because of the type of symmetry in as that shown in Fig. 17, offers good the design, the number of dynodes time response. Because the dynode in these structures can be reduced by stages (with the exception of the units of two. first three and the last two) are itera The "venetian-blind" dynode struc tive, the number of stages in this ture shown in Fig. 15 has a compact, structure may be varied without elec rugged geometry and is generally tron-optic redesign. used with planar photocathodes be The continuous-channel multiplier cause the relatively large area of the structure:; shown in Fig. 18 is very 22 RCA Photomultiplier Manual INCIDENT +-__++- ____ -I4..;.R_ADIATION o = OPAQUE PHOTOCATHODE I - 9 = DYNODE' ELECTRON MULTIPLIER 10 = ANODE (a) INTERNAL CONDUCTIVE COATING I GRILL INCIDENT RADIATION I SEMI- I TRANSPARENT;:----,.! PHOTOCATHODE I FACEPLATE SHIELD FOCUSING I ELECTRODE I h I 1- 10' DYNODES= ELECTRON MULTIPLIER II ANODE (b) Fig. 14 -- The circular-cage multiplier structure (a) in a "side-on" photomultiplier, (b) in a "head-on" photomultiplier. ELECTRON MULTIPLIER r------~I------, FACEPLATE INTERNAL CONDUCTIVE COATING Fig. 15 - The venetian-blind multiplier structure. Principles of Photomultiplier Design 23 SEMITRANSPARENT PHOTOCATHODE INCIDENT RADIATION INCIDENT RADIATION I-S-DVNODES - ELECTRON MULTIPLIER IO-ANODE Fig. 16 - The box-and-grid multiplier structure. ~ ~ SEMI- ~ TRANSPARENT \ ~PHOTOCATHODE~\ FOCUS RING \ INCIDENT I RADIATION FACEPLATE INTERNAL FOCUSING CONDUCTIVE ELECTRODE COATING 1-10- DVNODES =ELECTRON MULT1PLIER II-ANODE Fig. 17 - The linear-multiplier structure. compact and utilizes a resistive emit- primary electrons impinge to cause ter on the inside surface of a cylinder secondary emission from the oppo rather than a discrete number of dy- site side. There are, however, diffi nodes. The lack of discrete dynodes culties in this construction: ( 1 ) in causes the electron-multiplication sta- obtaining rugged self-supporting dy tistics to be poor because of the vari- nodes of reasonable size, and (2) in able path lengths and the variable preventing transmission of primary associated voltages. The gain of a electrons which tend to degrade the continuous-channel multiplier is de- statistics of electron transit time. termined by the ratio of length, 1, to inside diameter, d; a typical value of this ratio is 50 but it may range from 30 to 100. An interesting multiplier structure which has potentially good time re sponse is a series of close-spaced parallel-plane structures that use transmission dynodes. The dynodes Fig 18 - The continuous-channel consist of thin membranes on which multiplier structure. 24 RCA Photomultiplier Manual Multiplier Analysis The analysis of the multiplier structure requires solution of the Laplace equation. This solution can be obtained with an analog model, such as the rubber membrane,6 or Fig. 19 - A dynode whose over-all in digital form by use of a com rounded shape and field-forming ridges converge the electron cascade. puter. 1 Initial analysis requires that electrons from a large area of one stages of tube designs requiring large dynode be converged onto a smaller anode pulse currents. Such tubes area on the next dynode. Converg usually require higher interstage po ence must take place in two dimen tential differences near the anode to sions to keep the electron cascade overcome space-charge saturation from spilling off the ends of the dy effects. nodes. The over-all rounded shape of the dynode shown in Fig. 19 con ANODES verges the electrons in one dimen sion, while the field-forming ridges The primary function of the anode at or near the dynode ends converge is to collect secondary electrons the electrons in the second dimen from the last dynode. The anode sion. should exhibit a constant-current In an analysis of multiplier struc characteristic of the type shown in tures, space-charge effects may be Fig. 20. The simplest anode struc ignored in the early stages because ture, shown in Fig. 21, is a grid-like of the low current densities. How collector used in some venetian-blind ever, space-charge effects must be structures. The secondary electrons taken into account in the output from the next-to-Iast. dynode pass 400 (/) W 0:: W " ;;; g 300 a: u ::!' I t- r5 200 a: a: :J u --- W o o :i 100 - - o 50 100 200 POTENTIAL BETWEEN LAST DYNODE AND ANODE - VOLTS Fig. 20 - Constant-current characteristic required of an anode. Principles of Photomultiplier Design 25 the next-to-Iast dynode pass through the anode grid. In more sophisticated photomultiplier designs this phe GRID-LIKE nomenon may be suppressed by ANODE auxiliary grids that shield the anode from the effects of the impinging electron cloud. Fig. 21 - The simplest anode structure, a grid-like collector. TRANSIT-TIME CONSIDERATIONS through the grid to the last dynode. Secondary electrons leaving the last Photocathode Transit-Time dynode are then collected on the Difference grid-like anode. Photomultipliers de signed to have fast time responses A time parameter of interest is require anodes with matched-imped the photocathode transit-time differ ance transmission lines. Most high ence, the time difference between the speed circuits are designed to utilize peak current outputs for simultane a 50-ohm impedance, which requires ous small-spot illumination of differ a 50-ohm transmission line within the ent parts of the photocathode. (In tube and a suitable connector or practice, of course, one small area lead geometry outside the tube to at a time is illuminated and the permit proper interface. A more de spread in illumination-peak output tailed description of these require intervals is measured.) In a planar ments is given in the section on cathode design, the transit time is Voltage-Divider Considerations. longer for edge illumination than for Anode configurations and internal center illumination because of the transmission lines may be analyzed longer edge trajectories and the by use of the standard methods of weaker electric field near the edge cavity and transmission-line analy of the photocathode. The center-to sis. These methods yield approximate edge transit-time difference may be design parameters,7 which are op as much as 10 nanoseconds. The timized experimentally by means of spherical-section photocathode af time-domain reflectometry, TDR 8, 9. fords more uniform time response TDR provides information about the than the planar photocathode be discontinuities in the characteristic cause all the electron paths are impedance of a system as a function nearly equal in length; however, the of electrical length and is an ex transit time is slightly longer for tremely useful approach in the de edge trajectories than for axial tra sign of voltage-divider circuits and jectories because of the weaker elec mating sockets which do not readily tric field at the edge. lend themselves to mathematical The photocathode transit-time dif analysis. ference is ultimately limited by Certain anode structures in fast the initial-velocity distribution of rise-time photomultipliers exhibit a the photoelectrons; this distribution small-amplitude pulse (prepulse) that causes time-broadening of the elec can be observed a nanosecond or tron packet during its flight from two before the true signal pulse. In the photocathode to the first dynode. grid-like anode structures this pre The broadening effect can be mini pulse is induced when electrons from mized by increasing the strength of 26 RCA Photomultiplier Manual the electric field at the surface of the multipliers yield nearly isochronous photocathode. electron trajectories because of their compensation characteristics; the Electron-Multiplier higher-velocity electrons have longer Time Response path lengths between stages, and thus arrive at the same time as elec Because the energy spread of sec trons which began with lower initial ondary electrons is even larger than velocities. Because this type of multi that of photoelectrons, initial-velocity plier must be operated at a unique effects are the major limitation on voltage for a given magnetic field the time response of the electron strength, gain is not adjustable un multiplier. Multiplier time response less electromagnets are used. is usually improved by the use of Certain materials have advantages high electric-field strengths at the dy over other materials in providing node surfaces and compensated de good multiplier time performance. sign geometries. In a compensated For example, a standard dynode ma design, such as that shown in Fig. terial, copper beryllium (CuBe), has 22, longer electron paths and weaker a maximum gain per stage of 8 at fields alternate with shorter electron a 600-volt interstage potential dif paths and stronger fields from dy ference. In contrast, gallium phos node to dynode to produce nearly phide (GaP) exhibits a gain of 50 equal total transit time. or more at a 1000-volt interstage Excellent time response is achieved potential difference.1o The advantage with crossed-field multipliers which in time performance of the GaP use both magnetic and electro dynode over one of the CuBe type static focusing fields. Crossed-field is that the number of stages, and thus the total transit time, may be re duced, and that the energy spread of the secondary electrons is less. RUGGEDIZED PHOTOMULTI PLIERS New and increased demands on the capabilities of photomultipliers to survive severe vibrational environ ments have resulted in the develop ment of a series of ruggedized photo multiplier types. The ruggedization_ of these photomultipliers has been achieved without degradation of elec trical characteristics. Ruggedization of glass-envelope photomultipliers has been accom plished by moving the dynode cage extremely close to the stem (thereby drastically shortening the lead lengths and raising their mechanical resonant Fig. 22 - A compensated-design frequency), and by using heavier multiplier. leads, extra spacers to hold the dy- Principles of Photomultiplier Design 27 node cage in place, and a special pulse and the time duration of the heavy-duty welding process on the pUlse. The tubes may receive more metal-to-metal joints. A new type of than one shock pulse in each of the tube, the stacked-ceramic photomul orthogonal axes. tiplier, has been developed for use x- in the most severe mechanical en vironments. In this type, each dynode is mounted on the inside of a flat metal ring which is separated from y- y+ x- the next dynode ring by a ceramic ring spacer. The metal rings and ceramic spacers are then brazed to x+ y- gether to form the vacuum envelope. z+ Ruggedized tubes are tested for their ability to withstand sinusoidal and random vibrations and for their x+ resistance to shock. Test-parameter SECTION AA' values for RCA ruggedized photo multipliers are given in the Data A A' Section. The "sinusoidal vibration test" is performed on apparatus that applies a variable sinusoidal vibration to the tube. The sinusoidal frequency is varied logarithmically with time ~- from a minimum to a maximum to a Fig. 23 - Test axes used in random minimum value. Each tube is vi vibration and shock testing of stacked brated in each of the three ortho ceramic photomultipliers. gonal axes shown in Fig. 23. The BASING CONSIDERATIONS total time for the vibration of all three axes for ruggedized photomul Photomultiplier tubes may have tipliers is given in the Data Section. either a temporary or a permanently The peak acceleration is expressed attached base. Dimensional outline in units of the acceleration of gravity diagrams such as those shown in at the earth's axis (32.17 ft/s2 ) and Fig. 24 are provided in the Techni is denoted by the symbol g. The cal Data Section of this manual and distance over which the tube is vi in individual tube bulletins. Indicated brated is referred to as the "double on the diagrams are 1he type of base amplitude." Double amplitude is the employed, maximum mechanical di distance between the limits of travel mensions, radii of curvature where of the tube in both directions of the applicable, pin/ lead details, location sinusoidal vibration. and dimensions of magnetic parts The shock test is performed on used (in tubes utilizing minimum apparatus that applies a half-wave number of magnetic materials), and sinusoidal shock pulse to the photo notes regarding restricted mounting multiplier tube. The tube is sub areas, again where applicable. jected to the shock in each of three Photomultiplier tubes intended to orthogonal axes shown in Fig. 23. be soldered directly to circuit boards The shock pulse is expressed in or housings are supplied with semi terms of the peak acceleration of the flexible or "flying" leads and a tem- 28 RCA Photomultiplier Manual FACEPLATE ...... ~~~ .685 *.035 --'llR. r B PROTECTIVE '~A PLASTICSHIELD WD~ 14 SEMIFLEXIBLE DUMET ....1-* R F LEADS .016:!:.OO4 DIA. 1.5 MIN LENGTH TEMPORARY BASE ~ JEDECNo B20-102 ~ Dimensions Inches mm A 194 max. 100.0 max. 8 3.50 ~:~~ 88.9 ~k5 C .5 min. dia. 12-7 min. dia D .78 max. dia. 19.8 max. dia. E .755 max. dia. 19.18 max. dia. F .38 max. 9.7 max. H .75 min. 19.0 min. P .30 max. 7.6 max. R 1.0 max. 25 max. The dimensions in millimeters are derived from the basic inch dimensions (1 inch = 25-4 mm) Fig. 24 - Typical dimensional-outline drawings showing the type of base supplied with each tube and pertinent notes. porary base, intended for testing a lead indexing reference. A lead purposes only, that should be re connection diagram, such as the one moved prior to permanent installa shown in Fig. 26, relates terminal to tion. PIN I - DYNODE No. r PIN 2 - DYNODE No 3 A lead-terminal diagram that PIN 3-DYNODENO.5 PIN 4 - DYNODE NO.7 shows photomultiplier-tube lead PIN 5 - DYNODE NO.9 PIN 6 - ANODE orientation with the temporary base PI N 7 - DYNODE No.IO PIN B - DYNODE No.8 PIN 9 - DYNODE NO.6 removed, shown in Fig. 25, provides P1NIO - DYNODE NO.4 PINII -DYNODE No.2: 12 SEMIFLEXIBLE PIN 12 - PHOTOCATHODE LEADS DY, .016:!: .004 DIA. DIRECTION OF LIGHT INTO END OF BULB OYIO LEAD r - DYNODE No. I LEAD 2 - DYNODE NO.3 LEAD 3 - DYNODE No.5 LEAD 4 - DYNODE NO.7 LEAD 5 - DYNODE NO.9 LEAD 6 - ANODE 1'2 LEAD 7 - DYNODE No. 10 "1 LEAD 8- DYNODE No.8 LEAO 9- DYNODE No.6 LEADIO - DYNODE No.4 LEAD II - DYNODE No.2 I LEAD 12- PHOTOCATHODE (NOTEINDEX 21 ~ (NOTE 2) DY, l' .47±.01 DIA. Fig. 26 - Lead-connection diagram: (a. with base connected, (b) with tempora. Fig. 25 - Lead-orientation diagram. base removed. Principles of Photomultiplier Design 29 electrode. Care must be exercised in 5. K. C. Schmidt and C. F. Hendee, interpreting basing and lead-terminal "Continuous-Channel Electron diagrams to insure against possible Multiplier Operated in the Pulse damage to the photomultiplier re Saturated Mode," IEEE Trans. sulting from incorrect connections. Nucl. Sci., Vol. NS-I3, No.3, p. 100 (1966) REFERENCES 6. A. D. White and D. L. Perry, "Notes on the Use of a Rubber 1. H. E. Kulsrud, "A Programming Membrane Model for Plotting System for Electron Optical Electron Trajectories," Rev. Sci. Simulation," RCA Rev., Vol. Instr., Vol. 32, No.6, p. 730 28, No.2, p. 351 (1967) (1961) 2. R. W. Engstrom and R. M. 7. I. A. D. Lewis and F. H. Wells, Matheson, "Multiplier-Phototube Milli-Microsecond Pulse Tech Development Program at RCA niques, Pergamon Press (1959) Lancaster," IRE Trans. Nucl. 8. "Time-Domain Reflectometry," Sci., Vol. NS-7, No. 2-3, p. 52 Hewlett Packard Application (1960) Note 62 3. E. A. Taft and H. R. Philipp, 9. "Time-Domain Reflectometry," "Structure in the Energy Distri Tektronix Publication 062- bution of Photoelectrons from 0703-00 K3Sb and Cs3Sb," Phys. Rev., 10. G. A. Morton, H. M. Smith, Jr., Vol. II5, No.6, p. 1583 (1959) and H. R. Krall, "The Perform 4. D. E. Persyk, "Computer Simu ance of High-Gain First-Dynode lation of Photomultiplier-Tube Photomultipliers," IEEE Trans. Operation," RCA Reprint PE- Nucl. Sci., Vol. NS-I6, No.1, 473 (1970) p. 92 (1969) 30 Basic Performance Characteristics HIS section discusses the opera Equipment Design Parameters Ttional characteristics of photo multipliers that are listed in the Characteristic Range Values for Technical Data section of this Man Equipment Design are included in ual and in the technical data bulle. this Manual and in individual data tins that describe each type of bulletins in addition to the maximum photomultiplier. The index at the ratings described above. Minimum, back of the Manual gives the loca maximum, and typical values of se tion of particular characteristics of lected parameters for a given tube interest within the pages of this sec of a particular type are listed under tion. this heading to assist an equipment designer in selecting the proper de CLASSES OF DATA vice for his application. Knowledge of parameter values allows the de Ratings signer to specify the proper latitude of voltage and gain controls, addi Photomultiplier ratings are estab tional amplification, etc., which will lished to help equipment designers permit optimum utilization of the utilize the performance and service device he has chosen. Of the para capability of each device to best ad meters listed, some are measured vantage. These ratings are based on and some calculated from the mea careful study and extensive testing sured quantities. Not all parameters and indicate limits within which the are described for every device be operating conditions must be main cause many parameters are im tained to insure satisfactory perform portant only in specific infrequent ance. The maximum ratings given applications. are based on the Absolute Maximum System, a system initiated by the RATINGS AND DESIGN Joint Electron Device Engineering PARAMETERS Council (JEDEC) and subscribed to by the National Electrical Manufac Voltage turers Association (NEMA) and Electronic Industries Association The maximum voltage ratings for (EIA). photomultipliers are usually given as Basic Performance Characteristics 31 the maximum voltage to be applied possible application in which special between the anode and cathode, be or "tapered" voltage distribution may tween the anode and the last dynode, be desirable is in equipment in which between consecutive dynodes, and large anode pulse currents may be between the cathode and the first encountered. To reduce the possi dynode. A maximum anode-to-cath bility of space-charge limiting, it is ode voltage is specified so that the often desirable to increase the volt tube does not become regenerative. age between the last few successive At excessive values of voltage the dynodes and between the last dynode output current will become very and the anode. It may also be de noisy and may become so large as sirable to increase the normal cath to exceed the current rating and ode-to-first-dynode potential when cause damage to the secondary optimum pulse-height resolution is emitting surfaces of the dynodes desired. which could result in eventual tube failure. Maximum-voltage ratings are Cathode Current also specified between the last dy node and the anode. This rating is The magnitUde of the maximum usually made somewhat lower in permissible cathode current depends value than the ratings between con upon the type of photocathode and secutive dynodes to reduce the pos the type of substrate or conductive sibility of breakdown between the layer upon which the photocathode last dynode and the anode as a re is placed. In a tube having a semi sult of the close proximity of the transparent bialkali photocathode up two. Maximum voltages between to approximately one inch in di consecutive dynodes are specified to ameter, a peak cathode current of prevent the possibility of breakdown 1 X 10-8 ampere at a tube tem between these elements and also to perature of 22·C or 1 X 10-10 prevent excessive ohmic leakage in ampere at -100·C should not be the region of the stem leads. The exceeded. Tubes having larger photo inter-dynode maximum voltage rat cathodes should not be subjected to ing maintains operation below the cathode currents exceeding 1 X 10-9 point of maximum secondary emis ampere at a tube temperature of sion above- which the gain actually 22·C or 1 X 10-11 ampere at decreases. The specification of a -100·C. Because of the resistivity maximum cathode-to-first-dynode of the photocathode, the voltage drop voltage rating also reduces the pos caused by higher peak cathode cur sibility of leakage, breakdown, and rents could produce radial electric reduction in secondary emission. fields across the photocathode, and The sum of the individual maxi result in poor photoelectron collec mum voltage ratings may exceed the tion at the first dynode. Photo maximum anode-to-cathode voltage surface resistivity also increases with rating. This statement does not mean decreasing temperature and thus re that the over-all voltage rating can duces the maximum permissible cur be exceeded, but that, depending rent still further. upon the application, special volt Opaque cesium-antimony photo age distributions may be made. A cathodes can be operated at peak detailed discussion of voltage distri cathode currents as high as 100 bution is given in the section on microamperes per square centi Voltage-Divider Considerations. One meter. Semitransparent cesium-anti- 32 RCA Photomultiplier Manual mony photocathodes (CsSb) should Temperature be operated at currents less than 1 microampere per square centimeter A maximum ambient temperature, at room temperature, and propor and in some instances a minimum tionally less as the ambient tempera temperature, is specified for all ture is reduced. Semitransparent photomultipliers. The specification of multialkali and Ag-O-Cs (S-l) photo maximum ambient temperatures re cathodes have lower resistivities duces the possibility of heat dam than the CsSb types and can be oper age to the tube. Because dark current ated at currents as high as 100 micro (noise) increases with temperature, it amperes per square centimeter at is possible that a point of instability room temperature; again, derating is may be reached at high temperatures required as temperature is reduced. at which the tube will break down and the resulting large currents permanently damage the secondary emitting surfaces of the dynodes. Anode Current Even if the dark current does not increase to a level that causes re All photomultipliers have a maxi generation, the tube might be dam mum anode current rating. The pri aged as a result of changes in the mary reason for such a rating is to composition of the photocathode and limit the anode power dissipation dynode materials. Cesium, for ex to approximately one-half watt or ample, is very volatile and may be less. Consequently, the magnitude of driven from the photocathode and the maximum anode current is re dynode surfaces to condense on the stricted to a few milliamperes when cooler portions of the tube and the tube is operated at 100 to 200 thereby reduce gain and cathode volts between the last dynode and sensitivity or both. the anode. Space-charge effects must It is recommended that photomul also be considered in computation tipliers be operated at or below of the magnitude of the maximum room "temperature so that the effects recommended anode current. As the of dark current are minimized. The anode current increases, operation variation of dark current, or noise, becomes nonlinear because of the is most important because of its ef space-charge buildup between the fect on ultimate low-light-level sensi last few dynodes and the last dynode tivity. Various cryostats and soIid and the anode. Even though a photo state thermionic coolers have been multiplier acts essentially as a con designed that reduce dark current stant-current device, the signal volt at low temperatures in low-light-level age developed across the load is in applications. An important consider series with the last-dynode-to-anode ation in the use of these devices is voltage and consequently opposes it. to prevent condensation of moisture Nonlinear operation may occur as on the photomultiplier window. A the load resistance and load current controlled low-humidity atmosphere become large. or special equipment configuration Another possible adverse effect of may be necessary to prevent such operating a given photomultiplier at condensation. an excessively high anode current is When the temperature is varied, the increased fatigue that occurs as small changes in the spectral re the average anode current increases. sponse characteristics of the photo- IBasic Performance Characteristics 33 I 1.0.-----r---r--..----r--r--~----r-_.-_r____,-_,____, O.B ~ 0.6 ..... w ~ 0.4 :I: ""U ~ 0.2 ~ ::l 0 a:: t- ~ -0.2 u a:: ~ -0.4 -0.6 -O.B '------''----'_~_--'-_ __'_ _ __'__-'-_ __'_ _ _'__ _'__ _'_____J 200 300 400 500 600 700 BOO WAVELENGTH - NANOMETERS Fig. 27 - Temperature coefficient of cesium-antimony cathodes as a function of wavelength at 20·C. Note the large positive effect near the threshold. cathode may be observed. Fig. 27 transparent photocathodes of the shows a temperature coefficient of bialkali and cesium antimony types sensitivity of a CsSb photocathode as have particularly high resistivity. a function of wavelength.1 The in Consequently, the operating tem crease of red sensitivity with tem peratures of these photocathodes perature is typical for photoemission. Alhough it is recommended that photomultipliers be operated at .or below room temperature, many de vices have minimum operating tem peratures. The materials used in producing photocathodes are semi conductors. Their conductivity de creases with decreasing temperature w until the photocathode becomes so u resistive that a sizable voltage drop ~ 10~ Iii 6 may occur across the cathode sur in 4 W face when cathode current flows. a:: 2 Such a voltage drop may result in LIQUID N2 DRY ICE 10' ~~~=o=~:=-~===::I:==---!-~'::-J loss of linearity of the output cur -200 -160 -120 -80 -40 0 40 rent as a function of light level. Fig. TEMPERATURE - DEGREES C 28 shows the resistance per square Fig. 28 - Resistance per square as a of common semitransparent photo function of temperature for the cesium cathode materials; measurements antimony and the multialkali semi were made in special tubes having transparent photocathodes. These data were obtained with special tubes hav conections to parallel conducting ing connections to parallel conducting lines on the photocathode. Semi- lines on the photocathode. 34 RCA Photomultiplier Manual should be kept above -lOO·C, de Tubes designed for scintillation pending somewhat upon the photo counting are generally very sensitive cathode diameter, the light-spot to magnetic fields because of the diameter, and the cathode current. relatively long path from the cathode The larger the photocathode diameter to the first dynode; consequently, and the smaller the light-spot di such tubes ordinarily require electro ameter, the more severe the effect. static and magnetic shielding. Mag Opaque photocathodes and photo netic fields may easily reduce the cathodes having conductive sub anode current by as much as 50 per strates can be operated at cryogenic cent or more of the "no-field" value. temperatures. The three curves in Fig. 29 show Another reason for avoiding the the effect on anode current of mag operation of photomultipliers at ex netic fields parallel and perpendicular tremely low temperatures is the pos to the main tube axis and parallel to sible phase change that this type of the dynodes. The curves are usually operation may cause in some of the provided for one or more values of metal parts. These changes are par over-all applied voltage and indi ticularly probable when Kovar is cate the relative anode current in used in metal-to-glass seals. Tubes per cent as a function of magnetic utilizing Kovar in their construction field intensity in oersteds. Fig. 30 should not be operated at tempera shows the variation of output current tures below that of liquid nitrogen of several photomultiplier tubes as (-196·C). a function of magnetic-field intensity In some tubes, particularly those directly parallel to the major axis of with multialkali photocathodes, it is the tube. The magnitUde of the effect sometimes observed that the noise depends to a great extent upon the actually increases as the temperature structure of the tube, the orientation of the photocathode is reduced be of the field, and the operating volt low about -40·C. The reason for age. In general, the higher the operat this noise increase is not understood. ing voltage, the less the effect of However, most of the dark-current these fields. reduction has already been achieved High-mu material in the form of at temperatures above -40·C. foils or preformed shields is avail In general, it is recommended that able commercially for most photo all wires and connections to the tube multipliers. When such a shield is be encapsulated for refrigerated used, it must be at cathode potential. operation. Encapsulation minimizes The use of an external shield may breakdown of insulation, especially present a safety hazard because in that caused by moisture condensa many applications the photomulti tion. plier is operated with the anode at ground potential and the cathode at Exterior Magnetic and a high negative potential. Adequate Electrostatic Fields safeguards are therefore required to prevent personnel from coming in All photomultipliers are to some contact with the high potential of extent sensitive to the presence of the shield. The application of high magnetic and electrostatic fields. voltage, with respect to the cathode, These fields may deflect electrons to insulating or other materials sup from their normal path between porting or shielding the photomulti stages and cause a loss of gain. plier at the photocathode end of the Basic Performance Characteristics 35 current and noise output because of SUPPLY VOLTAGE E IS ACROSS A VOLTAGE DIVIDER PROVIDING 116 OF E BETWEEN voltage gradients developed across CATHODE AND DYNODE NQ I; 1112 OF E FOR the envelope wall, the application of EACH SUCCEEDING DYNODE -STAGE; AND 1112 OF E BETWEEN DYNODE No.IOAND ANODE. high voltage may produce leakage current to the cathode through the PHOTOCATHODE IS FULLY ILLUMINATED. tube envelope and insulating ma TUBE IS ORIENTED IN MAGNETIC FIELD AS SHOWN BELOW' terials; this current can permanently damage the tube. It is possible to modulate the out ~ I~./:Y>~~ 11 put current of a photomultiplier with a magnetic field. The application of a magnetic field generally causes no permanent damage to a photo 100 multiplier although it may magne (I) tize those internal parts of a tube eo that contain ferromagnetic materials (tubes are available which contain 60 practically no ferromagnetic ma ~40 terials). If tube parts do become mag w netized, the performance of the tube ~ 20 w may be degraded somewhat; how a. ever, the condition is easily corrected ~I~ z (2) by degaussing, a process in which a w tube is placed in and then gradually &:80 :J withdrawn from the center of a coil u w 60 operated at an alternating current 0 0 of 60 Hz with a maximum fiield ~40 strength of 100 oersteds. w 2:20 ... encountered in space to several at may be expected for a given light mospheres as perhaps would be pres input. Photocathode and anode sensi ent in an underwater application. tivities are stated in terms of lumi Although photomultipliers them nous or radiant sensitivity. selves may operate at very low pres Current amplification is stated in sures, precautions must be taken to terms of the ratio of anode output reduce the possibility of breakdown current to photocathode current and between external tube and circuit is measured at specified values of elements in partial-pressure environ electrode voltage. Typical photomul ments as predicted by Paschen's Law. tiplier current-amplification values The gas contributing to the partial range from 104 to 109 when the pressure could be the result of out tubes are operated at the voltages gassing of encapsulating and other recommended for the number of materials used external to the photo stages and dynode materials used. multiplier. Fig. 31 describes the anode sensi High-pressure operation of photo tivity in amperes per lumen and the multipliers depends primarily upon typical amplification characteristics the physical size and strength of the of a photomultiplier as a function of materials employed in their construc the applied voltage. These curves tion. Although a two-inch-diameter tube employing stacked ceramic-to SUPPLY VOLTAGE (E) "CROSS VOLTAGE DIVIDER PROVIDING IIIOOFE BETWEEN CATHODE AND metal brazed construction can easily DYNODE No.1; 1110 OF E FOR EACH SUCCEEDING withstand 4 or 5 atmospheres, other DYNODE; AND 1/10 OF E BETWEEN DYNODE No.9 8 ANODE. all-glass two-inch-diameter tubes B with very thin entrance windows are 6 6 limited to only slight additional load 4 4 ing above atmospheric pressure. ~'" on 2 IX) Helium Penetration. Because he N lium permeates through glass, the Q;IO' 107 15 B 8 operation or storage of photomulti I- 6 6 Z pliers in environments where helium a: 4 4 0 9 i= 0 - !::IO 10~ unusable. Continued operation will ~ 8 8 result in the deterioration of the 6 photocathode response and lead to 4 an inoperative tube. PHOTOMULTIPLIER DATA 500 1000 1200 1500 SUPPLY VOLTS (E) BETWEEN ANODE AND CATHODE Sensitivity Fig. 31 - Typical anode sensitivity and amplification characteristics of a photo Sensitivity provides an indication multiplier tube as a function of of the magnitude of the signal that applied voltage. 38 RCA Photomultiplier Manual illustrate the wide range of ampli Radiant sensitivity is the output fication of which a photomultiplier anode current divided by the inci is capable, provided a well regulated dent radiant power at a given wave voltage power supply is available. length, measured at a constant It is customary to present these electrode voltage. Cathode radiant sensitivity, amplification, and voltage sensitivity is the amount of current data on logarithmic scales that cover leaving the photocathode divided by the normal range of operation; the the incident radiant power at a given resultant curves are then closely ap wavelength, usually the wavelength proximated by straight lines. Curves of peak response. Values at other of minimum and maximum sensitiv wavelengths can be computed from ity are shown, as well as those of the relative-spectral-response curve. the typical sensitivity and amplifi Alhough radiant sensitivity may be cation. measured directly with suitable Luminous sensitivity is defined as monochromatic sources and radia the output anode current divided by tion-measuring equipment, it is the luminous flux incident on the usually computed from the measured photocathode, measured at a spe anode and cathode luminous-sensi cified electrode voltage; it is ex tivity values and expressed in terms pressed in amperes per lumen, Allm. of amperes per watt, A/W. This Luminous sensitivity is most fre computation uses conversion factors quently stated in terms of a tungsten and is dependent upon established filament light source in a lime-glass typical spectral-response character envelope operated at a color tem istics. perature of 2854 oK. A typical lumi Some users prefer to use integral nous flux for test purposes is in the radiant sensitivity instead of lumi range of 10-8 to 10-5 lumen. nous sensitivity because the latter Cathode luminous sensitivity is expression, although very commonly the photocurrent emitted per lumen used, does have an element of in of incident light flux on the photo appropriateness especially when used cathode at constant electrode volt to express the sensitivity of devices age, and is also expressed in terms which have sensitivity beyond the of amperes per lumen, A/1m. In the visible range of radiation. measurement of cathode luminous Integral radiant sensitivity is the sensitivity in a photomultiplier, a dc ratio of the anode or photocathode voltage of 100 to 250 volts is applied current to the total radiation in watts between the cathode and all other impinging on the photocathode from electrodes connected together as an a tungsten test lamp operating at a anode. Light flux in the order of color temperature of 2854°K. Be 10-2 to 10-4 lumen is then applied, cause it is very difficult to calibrate and the measured photocurrent is a test lamp in total radiant watts, it divided by the specified light level is customary to use a luminous cali to yield the required amperes per bration and a conversion factor for the lamp of 20 lumens per watt. lumen. Another sensitivity parameter It is customary in sensitivity often specified for tubes, especially measurements to subtract any dark those intended for scintillation de current which may be present so tectors, is blue sensitivity. Because that only the effect of the radiation most scintillators produce radiation is recorded. in the blue region of the spectrum, Basic Performance Characteristics 39 tubes intended for these applications fluxes range between 10-5 and 10-7 are processed for maximum sensi lumen, depending upon the tube tivity in this region. It has been sensitivity, when the anode blue found, during processing of such sensitivity is measured, and between photomultipliers, that the maximum 10-2 and 10-4 lumen when the cath sensitivity for blue and for white ode blue sensitivity is measured. light often does not occur at the Blue sensitivity is expressed in am same time. The anode blue sensitivity peres per "blue" lumen. is measured with rated voltages ap A spectral response curve pro plied to the electrodes and a source vides the most complete specifica of blue light incident upon the cath tion of the sensitivity of a photo ode. The blue-light source has a cathode. Fig. 33 shows a set of spectral output as indicated in Fig. typical spectral-response character 32 and consists of the flux of tung istics. Spectral-response characteris sten radiati 100 e 6 WAVELENGTH--NANOMETERS Fig. 32 - Spectral output of a blue 2 light source used in measuring anode blue sensitivity. of 2854 OK which is transmitted IIOLO-..l..--L----l--,l--...JLI-..J through a blue filter. Sensitivities are 7OO termed "blue" but are expressed in WAVELENGTH- NANOMETERS terms of the light flux which is in Fig. 33 - A set of typical spectral cident upon the filter. Incident light response characteristics. 40 RCA Photomultiplier Manual peak sensitivity at each wavelength spectral sensitivity for the desired over the useful spectral range of the wavelength. device. The center curve provides the absolute radiant sensitivity in Dark Current units of milliamperes per watt. The third curve shows the per-cent quan Current flows in the anode circuit tum efficiency at each wavelength. of a photomultiplier even when the The wavelength of peak response tube is operated in complete dark and the long-wavelength cutoff are ness. The dc component of this cur primarily functions of the photo rent is called the anode dark current, cathode material. The short-wave or simply the dark current. This cur length cutoff is primarily a function rent and its resulting noise com of the window material. Each win ponent usually limit the lower level dow material has its characteristic of photomultiplier light detection; cutoff region that varies from about as a result, the anode dark-current 300 nanometers for commonly used value is nearly always given as part lime glass to 105 nanometers for of the data for any tube. lithium fluoride. It is important to upderstand the Cathode quantum efficiency may variation of dark current in the be defined as the average number of photomultiplier as a function of vari photoelectrons emitted per incident ous parameters so that the ultimate photon. The cathode quantum effi in low-light-Ievel detection can be ciency, QE, in per cent at any given achieved. Although not all the wavelength can be calculated from mechanisms by which dark current the following formula: is generated within a photomultiplier are understood, they can be cate QE = S 12~9.5 (100) gorized by origin into three types: ohmic leakage, emission of electrons from the cathode and other elements where S is the cathode radiant sensi of the tube, and regenerative effects. tivity at the wavelength A in amperes Interior ohmic leakage, the pre per watt, and A is the wavelength in dominant source of dark current at nanometers. From this relation and low operating voltage that can be the absolute (radiant) spectral-re identified by its proportionality with sponse characteristic it is possible to applied voltage, results from the im calculate the typical quantum effi perfect insulating properties of the ciency at any wavelength within the glass stem, the supporting members, spectral range of the photomultiplier or the base, and is always present under consideration. but usually negligible. Contamina The relative spectral response can tion consisting of dirt, water vapor, also be used to calculate cathode or grease on the outside of the tube quantum efficiency if the radiant may also contribute to ohmic leak sensitivity at the wavelength of peak age. A source of leakage inside the response is known. This calculation tube may be residual alkali metals is made by multiplying the peak used in the photomultiplier process radiant -sensitivity by the percentage ing. of the peak response (obtained from As the operating voltage of the the relative-response curves) for the photomultiplier increases to mid wavelength in question. The resultant range values, the dark current begins product is the absolute (radiant) to follow the gain characteristic of Basic' Performance Characteristics 41 the tube. The source of this gain proportional dark current is the dark or thermionic emission of electrons from the photocathode or the dy nodes. Because each thermionically generated electron emitted from the photocathode is multiplied by the secondary-emission gain of the tube, the output is a pulse having a mag nitude equal to the charge of one ~~~~~~~~--~~~4 electron multiplied by the gain of 0-1000 -BOO -600 -400 -200 0 EXTERNAL SHIELD VOLTS the tube. This process, involving sta RELATIVE TO ANODE VOLTS tistical variations, is discussed in de Fig. 34 - Effect of external-shield po tail in the section on Statistical tential on the noise of a IP21 photo Fluctuation and l'IJoise. Because the multiplier. Note the desirability 01. emission of thermionic electrons is maintaining a negative bulb potentia. random in time, the output dark cur potential with respect to the cathode rent consists of random unidirec on the envelope wall can cause noisy tional pulses. The time average of operation by contributing to the gen these pulses may be measured on eration of minute ionic currents a dc meter and is usually the main which flow through the glass and component of the dc dark current at produce light at the cathode as a normal operating voltages. The vari result of fluorescence within the able nature of the thermionic dark glass. This noise need not arise from current limits the sensitivity of a direct connection of the envelope tube at very low light levels. It is wall to a positive potential; close not possible to balance out the ef proximity of the tube to such a po fect of this current in the same man tential will yield the same effect. In ner as the effect of a steady dc addition to the increased dark cur current resulting from ohmic leak rent and noise, the life of the photo age. It is usually advantageous to cathode is usually shortened when operate a photomultiplier in the regenerative effects are present. To range where the thermionic com prevent this deterioration, the en ponent is dominant because in this velope wall should be maintained range the signal-to-noise ratio of the near photocathode potential by tube is maximum. wrapping or painting it with a con Regenerative effects occur in all ductive material and connecting this photomultipliers at higher dynode material to cathode potential. The voltages and cause the dark current connection is usually made through to increase and become very er a high impedance to reduce the shock ratic. At times the dark current may hazard. Insulating materials must be increase to the practical limitations selected to limit the leakage currents of the tube and circuit. Permanent in the vicinity of the photocathode damage may be caused to the sensi to less than 1 X 10-12 ampere. tized surfaces. Excess noise or dark current can Regenerative effects may be trig also result from a number ot other gered by the electrostatic potential sources: fluorescent effects produced of the walls surrounding the photo on the inner surfaces of the bulb cathode and dynode-structure re during high-current operation, light gion. As shown in Fig. 34, a positive generated within the tube at the 42 RCA Photomultiplier Manual various dynodes and at the anode as related to the actual application of a result of electrons striking these the photomultiplier. elements, field emission occurring at The best operating range for a sharp points within the tube, ionic given photomultiplier can usually be bombardment of the photocathode predicted from the quotient of the as the result of minute quantities of anode dark current and the luminous residual gases, and scintillations in sensitivity at which the dark current the glass envelope of the tube caused is measured. This quotient is iden by radioactive elements within glass tified as the Equivalent Anode Dark (most glas:;es contain some radio Current Input (EADCI) in the active K40). Technical Data section of this Man Photomultipliers with bialkali ual and in individual tube bulletins, photocathodes, and to a lesser ex and is the value of radiant flux inci tent those with cesium antimony dent on the photocathode required photocathodes, exhibit a temporary to produce an anode current equal increase in dark current and noise to the dark current observed. The by as much as three orders of mag units used in specifying EADCI are nitude when the tube is exposed mo either lumens or watts. mentarily to blue light or ultraviolet The curves in Fig. 36 show both radiation from sources such as typical anode dark current and fluorescent room lighting, as shown equivalent anode dark current input in Fig. 35. This increase in dark (EADCI) as functions of luminous current occurs even though voltage is not applied to the tube, and may LUMINOUS SENSITI VITY IS VARIED BY ADJUST ING THE SUPPLY VOLTAGE (E) ACROSS VOLT persist for a period of from 6 to 24 AGE DIVIDER WHICH PROVIDES 1110 OF E hours after such irradiation. PER STAGE. LIGHT SOURCE IS A TUNGSTEN-FILAMENT LAMP OPERATED AT A COLOR TEMP. OF2854°K TUBE TEMPERATURE-22°C Fig. 35 - Variation of dark current following exposure of photocathode to cool white fluorescent-lamp radiation. The various photocathodes are identified by their spectral-response symbols. ,2468102468,0' Dark-Current Specification. Dark LUMINOUS SENSITIVITY-AMPERES/LUMEN I I I I I current values are often specified at /100 1300 1500 1700 1800 a particular value of anode sensitiv SUPPLY VOLTAGE(El-VOLTS ity rather than at a fixed operating Fig. 36 - Typical and equivalent voltage. Specifications of dark cur anode-dark-current input (EADCI) m rent in this manner are more closely a function of luminous sensitivity. Basic Performance Characteristics 43 sensitivity. These curves can be used cified operating conditions. If the to predict the best operating range type of modulation and bandwidth for a tube by considering the ratio used in the measurement is known, of the dark current to the output an equivalent noise input, ENI, can sensitivity. The optimum operating be calculated from the signal-to range occurs in the region of the noise ratio. Equivalent noise input minimum on the EADCI curve, the is defined as the value of incident region in which the signal-to-noise luminous or radiant flux which, ratio is also near its maximum. The when modulated in a stated man increase in the EADCI curve at ner, produces an rms output cur higher values of sensitivity indicates rent equal to the rms noise current the onset of a region of unstable and within a specified bandwidth, usually erratic operation. Many curves of I Hz. this type also include a scale of To determine the ENI for any anode-to-cathode supply voltage cor other bandwidth, it is necessary only responding to the sensitivity scale. to multiply the published ENI value Dark-Current Reduction. Dark by the square root of that band current may be reduced by cooling width. The conditions under which the photomultiplier in a refrigerant, a value of ENI is obtained are simi such as dry ice. This procedure is lar to those used to measure lumi recommended in those applications nous sensitivity, except that the light in which maximum current amplifi incident upon the photocathode is cation with minimum dark current is modulated by a light chopper so that required. It should be noted, how a square-wave signal with equal on ever, that at reduced temperature and off periods is produced at the the resistance of the cathode in photomultiplier output. creases and may reduce the maxi It is of interest to consider what mum peak anode current. ENI may be expected for a photo If care is taken to avoid damage multiplier if it is assumed that the to the photomultiplier by operation only source of noise is the photo with excessive current, the dark cur cathode dark emission, it. If the rent can often be reduced by a photocathode sensitivity is given by process of operating the photomulti S in amperes per lumen aI)d the un plier in the dark at or near the maxi modulated light flux by F in lumens, mum operating voltage. This process, the unmodulated output signal is called dark aging, may require sev equal to SFJL, where p, is the gain eral hours to several days. After such of the multiplier. When the light is a process of aging, it is recommended ohopped in an equal on and off that a photomultiplier be operated square wave with its peak amplitude for several minutes at the reduced the same as that of the unmodulated voltage before measurements are at light flux, the rms signal output at tempted. the fundamental frequency (with a Equivalent Noise Input. Dark cur narrow-bandpass measuring device) rent is the average or dc value of is SFp,(2)1/2/ 1T• The rms noise cur discrete pulses occurring at random rent from the dark emission in a intervals. The fluctuations or noise band pass of af is given by the ex associated with these pulses limit the pression JL(2eit af)1/2, where e is the measurement precision. Noise is charge of an electron. This expression evaluated in terms of a signal-to neglects the noise increase from dy noise-ratio measurement under spe- node amplification statistics which 44 RCA Photomultiplier Manual may increase this noise current by emission current, the noise in the a few per cent. (See the section on signal current is negligible. Statistical Fluctuation and Noise.) Because the dark emission is re In an evaluation of the expression duced as the temperature of the for ENI, the signal must be equated photocathode is reduced, ENI may to the dark noise, as follows: be reduced by cooling the photomul tiplier. Fig. 37 shows the variation of ENI for a IP21 (opaque CsSb photocathode) over a wide range of temperature. The value of ENI is then approxi The differential dark-noise spec mately the value of F for the condi trum shown in Fig. 38 displays the tion of this equation: number of pulses counted per unit of time as a function of their height. Such a spectrum indicates the num ENI = F bers of single and multiple electron events originating at the photo For M = I, it 10-15 ampere (a cathode. Submultiple electron emis reasonably low value of photocathode sion results from electron emission dark-emission current), and S = 100 at one of the dynode surfaces. microamperes per lumen, the ENI A differential dark-noise spectrum is equal to 4 X 10-13 lumen. This is obtained with a multichannel value is equivalent to a photocathode pulse-height analyzer. The calibra peak signal current of 4 X 10-17 tion of the single-photoelectron ampere. Because this signal current pulse height is determined by il is almost two orders of magnitude luminating the photocathode with a less than the photocathode dark- light level so low that there is a 100 VOLTS PER STAGE. BANDWIDTH: I Hz LIGHT SOURCE: TUNGSTEN. AT 2B54°K INTERRUPTED AT 90 Hz TO PRODUCE PULSES ALTERNATING BETWEEN ZERO AND FLUX VALUE SHOWN FOR ANY GIVEN TUBE TEMPERATURE; "ON" PERIOD OF PULSE EQUAL TO "OFF" PERIOD; RMS SIGNAL CURRENT =RMS NOISE CURRENT EXTERNAL SHIELD VOLTS RELATIVE TO ANODE VOLTS = -1000 TUBE TEMPERATURE - DEGREES CENTIGRADE Fig. 37 - Equivalent noise input in lumens for a 1 P21 photomultiplier as a function of temperature. Basic Performance Characteristics 45 very low probability of coincident extrapolation to show the location photoelectron emission. The dark of the single photoelectron peak; the pulse distribution is then subtracted solid portion shows the actual dif from the subsequent combination of ferential dark-noise spectrum. The dark pulses and single-photoelectron rapid change in slope of this curve pulses, so that. the remainder repre in the pulse-height region of less sents only that distribution resulting than one-half photoelectron is as from single-photoelectron events. By sumed to be the result of electron adjusting the gain of the pulse-height emission from the first- and second analyzer, the single-electron photo dynode surfaces. peak can be placed in the desired The slope of the curve for the channel to provide a normalized pulse-height region between 1 and distribution. 4 photoelectrons is as expected for The dark-pulse spectrum of Fig. single-electron emission, when the 38 is characteristic of photomulti statistical nature of secondary-emis pliers intended for use in scintilla sion mUltiplication is considered. The tion counting and other low-light number of pulses in this region may level pulse applications. The dashed be reduced by cooling the photo portion at the left of the curve is an multiplier. The slope of the curve for the pulse-height region greater than 4 photoelectrons is presumed to be caused by multiple-electron-emission events. These multiple pulses are caused by processes such as ionic bombardment of the photocathode. SOLID-LINE PORTION INDICATES DARK-PULSE SPECTRUM. Other mechanisms contributing to TUBE TEMPERATURE: 22·C ONE - PHOTOELECTRON PULSE HEIGHT=4 the noise spectrum include cosmic COUNTING CHANNELS. INTEGRATING TIME CONSTANT = lOps rays, field emission, and radioactive (R=IOKll C= contaminants that produce scintilla tions within the glass envelope. Cool ing has little effect upon reduction of the number of these multiple electron pulses, but extended opera tion of the tube _may improve performance. Operation of the tube may result in erosion of sharp points and reduce the possible contribution of field effects. In addition, improve ment occurs because residual gases are absorbed within the tube, ion bombardment of the photocathode is reduced, and the resulting multiple secondary-electron emission is les sened. For many applications it is useful 2 4 12 to have a summation of the total DARK- PULSE HEIGHT -PHOTOELECTRONS number of dark pulses. In Fig. 38, Fig. 38 - Typical dark-noise pulse for example, the sum of dark-pulse spectrum. counts from one electron equivalent 46 RCA Photomultiplier Manual height to 32 electron equivalent PULSE HEIGHT RESOLUTION IN PER CENT heights is 3 X 104 counts per min IS DEFINED AS 100TIMES THE RATIO OF THE WIDTH OF THE PHOTOPEAK AT HALF ute. The value of 32 is chosen to THE MAXIMUM COUNT RATE IN THE be essentially equivalent to infinity. PHOTOPEAK HEIGHT (AI TO THE PULSE HEIGHT AT MAXIMUM PHOTOPEAK Another useful figure is the summa COUNT RATE (81. tion of pulses in the range from 4 ..J ~~------~ to 32 equivalent electron heights. a: This sum for this particular distri ~ PHOTO bution is 2.2 X loa counts per min ~~ P~K ute and corresponds to the rate of ~~ 6~ A multiple-electron events. U'" ",,,,1:1: tt~ Pulse Counting a:::I ... n. za: Pulse-Height Resolution. An im ::I'"on. portant application of photomulti ...... u'" 0::1 pliers is in scintillation counting. "'~ (For a more detailed account of 9:0 PULSE HEIGHT B ffi ~ ~ scintillation counting, see the section n.L-______on Photomultiplier Applications.) In Fig. 39 - Typical pulse-height distribu many scintillation-counting applica tion curve. tions, a photomultiplier is coupled to a thallium-activated sodium iodide width of the photopeak at half the NaI(Tl) crystal in which scintilla maximum count rate divided by the tions are produced by gamma rays pulse height at maximum count rate. resulting from nuclear disintegra Consequently, the lower the pulse tions. Because the output of a photo height resolution, the greater the multiplier is linear with light input ability of the photomultiplier to dis and because the light energy of scin criminate between pulses of nearly tillations is directly proportional to equal height. the gamma-ray energy over a certain The pulse-height resolution para range, an electrical pulse is obtained meter is often measured by using which is a direct measure of the the 662-keV photon from an iso gamma-ray energy. Consequently, an tope of cesium, CS137, and a cylin important requirement of photomul drical thallium-activated sodium tipliers used in nuclear spectrometry iodide scintillator. The CS137 source is the ability to discriminate between is placed in direct contact with the pulses of various heights. The para metal end of the scintillator pack meter indicating the ability of a age and the faceplate end of the tube to perform this discrimination crystal is coupled to the tube by is called pulse-height resolution. means of an optical coupling fluid. A typical pulse-height distribution The load across which the pulse curve obtained with a CS137 source height is measured may consist of a and an NaI(Tl) crystal is shown in 100,000-ohm resistor and a capaci Fig. 39. The main peak of the curves tance of 100 picofarads in parallel. at the right is associated with mono This combination provides a 10- energetic gamma rays which lose microsecond time constant. their entire energy by photoelectric Peak-to-Valley Ratio. In many conversion in the crystal. Pulse low-energy scintillation-counting ap height resolution is defined as the plications, noise as well as resolution Basic Performance Characteristics 47 becomes important. A parameter The peak-to-valley ratio of the which provides an indication of measured spectrum is a measure of photomultiplier usefulness in resolv the extent to which tube noise ob ing low-energy X-rays is the pulse scures the observation of low-energy height spectrum obtained for an gamma rays. However, at the larger Fe55 radioactive isotope and shown values of peak-to-valley ratio (30 in Fig. 40. This parameter is mea to 50), the properties of the crystal sured by using an NaI(Tl) scintilla become dominant. tor with a thin beryllium window Afterpulsing. Afterpulses, which and an Fe55 source. The decay of may be observed when photomulti this isotope of iron results in a 5.9- pliers are used to detect very short keY X-ray. From the resulting spec light flashes as in scintillation count trum a peak-to-valley ratio and a ing or in detecting short laser pulses, pulse-height resolution can be cal are identified as minor secondary culated. pulses that follow a main anode current pUlse. There are two general types of afterpulses; both are char acterized by their time of occurrence Fe 55 SOURCE, ACTIVITY I"CURIE. .. SCINTILLATOR: HARSHAW TYPE H~ 0.005 in relation to the main pulse. The BERYL\.JUM WINDOW, No i {Ttl, 7IB' DIA., 0.040' THICK. first type results from light feedback CATHODE-TO-DYNODE No.1 VOLTS: 660 DYNODE No.I-TO-DYNODE No.2 VOLTS: 108 from the area of the anode, or pos DYNODE NO.2-TO-DYNODE NO.3 VOLTS:151 EACH SUCCEEDING DYNODE-STAGE VOLTS: 108 sibly certain dynodes, to the photo ANODE-TO-CATHODE VOLTS: 2000 cathode; the intensity of the light is FOCUSING ELECTRODE IS CONNECTED TO OYNODE No.1 POTENTIAL. proportional to the tube currents. ELECTRON MULTIPLIER SHIELD IS CONNECTED TO DYNODE No.5 POTENTIAL. When this light feedback reaches the cathode, the afterpulse is produced. Afterpulses of this type, character ized by a delay in the order of 40 w 10· to 50 nanoseconds, may be a prob I- :J lem in many older photomultipliers ~ :< having open dynode structures. The a: w 0- time delay experienced with this type 00 I- of afterpulse is equal to the total Z W transit time of the signal through the w> photomultiplier plus the transit time I 00 of the light that is fed back. W I- been shown to be the result of ioniza z 10 tion of gas in the region between ~ Z :J the cathode and first dynode. The 0 PEAK .. u VALLEY PEAK VALLEY 50 time of occurrence of the afterpulse depends upon the type of residual gas involved and the mass of the gas ion, but usually ranges from 200 nanoseconds to well over 1 micro 12 second after the main pulse. When the ion strikes the photocathode, sev eral secondary electrons may be Fig. 40 - Typical pulse-height spec trum obtained for an Fe" radioactive emitted; thus, the resulting after isotope. pulse has an amplitude equal to sev- 48 RCA Photomultiplier Manual eral electron pulse-height equiva Time Response lents. These pulses appear to be identical to the larger dark-current The commonly used methods of pulses, and it is suspected that many specifying time response involve of the dark-current pulses are the frequency-domain and time-domain result of photocathode bombardment parameters. Frequency response is by gas ions. usually characterized in terms of the Several gases, including N2+ and frequency at which the ac com H2 +, are known to produce after ponent of the output current de pUlses. Each gas produces its own creases to 3 dB below the low characteristic delay following the frequency output level, and may be main pulse. The most troublesome, measured with an electro-optic modu perhaps, is the afterpulse caused by lator and a variable-frequency driving the H2+ ion; this afterpulse occurs oscillator. Time-response measure approximately 300 nanoseconds after ments are made with small-spark the main pulse. One source of hy sources,2 light-emitting diodes, or drogen in the tube is water vapor mode-locked lasers 3 that approxi absorbed by the multiplier section mate delta-function light sources. In before it is sealed to the exhaust sys most measurements the entire photo tem. Other gases which may cause cathode is illuminated. afterpulsing may be present as a The frequency response of a result of outgassing of the photomul photomultiplier is related to the tiplier parts during processing or anode-pulse rise time, a parameter operation. Present photomultiplier usually provided for each photo processing techniques are designed multiplier type. The anode-pulse rise to eliminate or at least to minimize time of a tube, illustrated in Fig. 41, the problem of afterpulsing. is affected primarily by the mechani- 20 10 2 8 Z 8 6 !oJ '"~ 4 .8 .6 A~--~~~~~~~~~~~~~ __~~~~~~-L~~ 1000 1200 1400 1600 1800 2000 2500 3000 3500 4000450°5000 APPLIED VOLTAGE - VOLTS Fig. 41 - Anode-pulse rise times as a function of anode-to-cathode applied voltage for a number of photomultipliers. Basic Performance Characteristics 49 cal configuration and the operating STABILITY values of a device and is specified as the time for the current to increase The operating stability of a photo from 10 to 90 per cent of maximum multiplier depends on the magnitude anode-pulse height when the tube is of the average anode current; when illuminated with a delta-function stability is of prime importance, the light source. use of average anode currents of 1 Frequency response is also limited microampere or less is recommended. by the transit-time spread. This Additional considerations affecting spread is the time interval between stability are discussed in the subse the half-amplitude points of the out quent paragraphs. put pulse at the anode terminal, as Fatigue shown in Fig. 42. Transit-time spread is seldom measured directly because Tube fatigue or loss of anode sensi it is difficult to determine whether ti"ity is a function of output-current the fall time of the pulse results from level, dynode materials, and previous the fall time of the light source or operating history. The amount of the tube. Figs. 41 and 42 indicate average current that a given photo that transit and rise times are re multiplier can withstand varies ducl~d as over-all applied voltage is widely, even among tubes of the increased. Tubes utilizing focus same type; consequently, only typi structures of either the linear or the cal patterns of fatigue may be cited. circular-cage type exhibit better The sensitivity changes that are time-resolution characteristics than a direct function of large currents tubes of the venetian-blind type; in imposed for great lengths of time are many high-speed applications this thought to be the result of erosion of difference is of major importance. the cesium from the dynode surfaces Fig. 42 - Transit-time spreads as a function of supply voltage for a number of photomultipliers. 50 RCA Photomultiplier Manual during periods of heavy electron 100 r---y---.,---r----,.-__ bombardment, and the subsequent deposition of the cesium on other 80 areas within the photomultiplier. ~ Sensitivity losses of this type, illus- > trated in Fig. 43 for a IP21 operated ~ at an output current of 100 micro- ~ amperes, may be reversed during ~ periods of non-operation when the ~ cesium may again return to the dy- ~ node surfaces. This process of return 0:: 20 o 400 500 > t- > Fig. 44 - Typical sensitivity loss for E a IP21 operating at ZOO volts per stage 60 LIGHT for a period of 500 '"z ON hours. Initial anode w TUBE current is 100 microamperes and is re 1---+1+-- NO LIGHT ON TUBE '"4J adjusted to this operating value at 48, > 40 168, and 360 hours. 5..J '"0:: 20 IP21 having CsSb dynodes, operat ing at an output current of 100 microamperes. Tubes operated at lower current levels, of the order of o 100 200 300 400 500 TIME - MINUTES 10 microamperes or less, experience less fatigue than those operated at Fig. 43 - Short-time fatigue and re higher currents, covery characteristics and, in fact, may of a typical IP21 actually operating at 11 volts per stage and with recover from high-current a light source adjusted to give 100 operation during periods of low-cur- microamperes initial anode current. At rent operation. • the end of 100 minutes the light is Fatigue rates are turned aD and also affected by the tube allowed to re the type of dynode cover sensitivity. Tubes recover ap materials used in proximately as shown, whether the a tube. Copper beryllium or silver voltage is on or aD. magnesium dynodes are generally more stable at high operating may cur be accelerated by heating the rents than the cesium antimony photomultiplier during periods of types. The sensitivity for tubes non-operation utiliz to a temperature ing these dynodes very often within in the maximum temperature creases during initial hours of rating of the tube; heating above the operation, after which a very gradual maximum rating may cause a perma decrease takes place, as illustrated nent loss of sensitivity. in Fig. 45. Sensitivity losses for a given oper ating current normally occur rather Photocathode Fatigue rapidly during initial operation and at a much slower rate after the tube A photomultiplier with a multi has been in use for some time. Fig. alkali photocathode (S-20 spectral 44 shows this type of behavior for a response) tends to decrease in sensi- Basic Performance Characteristics 51 Sensitivity may decrease quite rapidly upon application of light and volt age during the first few minutes of operation. With a few hours of con tinued operation, the sensitivity usu ally increases until it has regained its initial value. If the source of the high radiant energy is removed from 100 the tube, the sensitivity also tends to return to its initial value. >- 80 "Hysteresis" :;;~ ~ iii Many photomultipliers exhibit a z 60 CI) tt:mporary instability in anode cur '" rent and change in anode sensitivity '"> ~ 40 for several seconds after voltage and ..--' light are applied. This instability, '"0: sometimes called hysteresis because 20 of cyclic behavior, may be caused by electrons striking and charging the dynode support spacers and thus 0 100 500 slightly changing the electron optics within the tube. Sensitivity may Fig. 45 - Typical sensitivity variation overshoot or undershoot a few per on life for a photomultiplier having cent before reaching a stable value. silver-magnesium dynodes. Initial anode current was 2 milliamperes and was re The time to reach a stable value is adjusted to this operating value at 48, related to the resistance of the in 168, and 360 hours. sulator, its surface capacitance, and the local photomultiplier current. tivity upon extended exposure to This instability may be a problem in high ambient room lighting. The applications such as photometry mechanism by which this sensitivity where a photomultiplier is used in loss occurs is not clearly understood. a constant-anode-current mode by The decreased cathode sensitivity oc varying the photomultiplier voltage curs primarily in the red portion of as the light input changes. the spectr,:um and is usually perma Hysteresis has been eliminated in nent. It ~is interesting to note that many tubes by coating the dynode photocathodes of the extended red spacers with a conductive material multi-alkali (ERMA) type appar and maintaining them at cathode po ently do not exhibit this loss. tential. Tubes treated by this method The Ag-O-Cs photocathode (S-1 assume final sensitivity values almost spectral response) also suffers a de immediately upon application of crease in sensitivity, particularly dur light and voltage. ing operation, when exposed to high radiant-energy levels normally not Counting Stability harmful to other types of photo cathode materials. The decreased In scintillation counting it is par sensitivity occurs primarily in the ticularly important that the photo infrared portion of the spectrum. multiplier have very good stability. 52 RCA Photomultiplier Manual There are two types of gain stability 10,000 counts per second. The photo tests which have been used to evalu peak counting rate is then decreased ate photomultipliers for this applica to 1000 counts per second by increas tion: (1) a test of long-term drift in ing the source-to-crystal distance. pulse-height amplitude measured at The photopeak position is measured a constant counting rate; and (2) a and compared with the last measure measure of short-term pulse-height ment made at a counting rate of amplitude shift with change in count 10,000 per second. The count-rate ing rate. stability is expressed as the percent In the time stability test, a pulse age gain shift for the count-rate height analyzer, a CS137 source, and change. It should be noted that a NaI(Tl) crystal are employed to count-rate stability is related to the measure the pulse height. The CS137 hysteresis effect discussed above. source is located along the major Photomultipliers designed for count axis of the tube and crystal so that ing stability may be expected to have a count-rate of 1000 counts per sec a value of no greater than 1 per cent ond is obtained. The entire system is gain shift as measured by this count allowed to warm up under operating rate stability test. conditions for a period of one-half to one hour before readings are re corded. Following this period of Life Expedancy stabilization, the pulse height is re corded at one-hour intervals for a The life expectancy of a photo period of 16 hours. The drift rate in multiplier, although related to fa per cent is then calculated as the tigue, is very difficult to predict. mean gain deviation (MGD) of the Most photomultipliers will function series of pulse-height measurements, satisfactorily through several thou as follows: sand hours of conservative operation and proportionally less as the severity of operation increases. Although i-n photomultipliers do not have ele ~ ip - pi! ments which "burn out" as in the MGD = o=i-::.:.l ___ 100 n p case of a filament in a vacuum tube, photomultiplier dynodes do lose sensitivity with operation; however, where p is the mean pulse height, as pointed out earlier, this loss oc Pi is the pulse height at the ith read curs at an extremely low rate and. ing, and n is the total number of even tends to recover during con readings. Typical maximum mean servative operation. gain-deviation values for photomul Factors which are known to af tipliers with high-stability CuBe fect life adversely are high-current dynodes are usually less than 1 per operation, excessive-voltage opera cent when measured under the con tion, high photocathode illumination, ditions specified above. Gain stability and high temperature. becomes particularly important when Operation of photomultipliers in photopeaks produced by nuclear dis regions of intense nuclear radiation integrations of nearly equal energy or X-rays may result in an increase are being differentiated. in noise and dark current as a re In the count-rate stability test, the sult of fluorescence and scintillation photomultiplier is first operated at within the glass portions of the tubes. Basic Performance Characteristics 53 1 Continued exposure may cause dark 10. r---r--'--'--"--'-7r= ening of the glass and a resultant TYPE 931A ,,}- VOLTS PER STAGE -100 reduction in transmission capability. 10' T7 SATURATION Photomultipliers having glass en AT 45 mA 105 MAXIMUM / / velopes should not be operated In DEVIATION FROM helium environments because helium can permeate the tube envelope and lead to eventual tube destruction. 2" Extremely high illumination fo cused onto a spot on the photo cathode actually destroys a portion 1613~/_~_-':----L-:----'-:~--'::::--~-' of the photocathode. It is usually 10-14 10-12 10-10 10-8 10-6 10-4 10-2 better to defocus such a light source LIGHT FLUX-LUMENS and utilize a larger area of the photo Fig. 46 - Range of anode-current line cathode. arity as a function of light flux for a Loss of photocathode sensitivity 931 A photomultiplier. may sometimes result from the prox charge-limiting effects usually occur imity of a high-potential gradient in in the space between the last two the region of the photocathode. Such dynodes. The voltage gradient be a gradient may result from applica tween anode and last dynode is usu tions in which the photomultiplier ally much higher than between dy is operated with the anode near nodes and, therefore, results in a ground potential. Metal fixtures, limitation at the previous stage, even clamps, or shields at ground poten though the current is less. The maxi tial may provide such a gradient mum output current, at the onset of through the glass near the photo space charge, is proportional to the cathode either by proximity or direct 3/2 power of the voltage gradient in contact. The result may be increased the critical dynode region. By use noise output and a deterioration of of an unbalanced dynode-voltage dis the photocathode through an elec tribution increasing the interstage trolysis process through the glass. voltages near the output end of the The destructive process may take a multiplier, it is possible to increase few hours or several hundred hours the linear range of output current. depending upon the material in con Some photomultipliers used in ap tact with the envelope and the plications requiring high output pulse amount of potential difference. The current use an accelerating grid be deterioration process is greatly ac tween the last dynode and the anode celerated as the temperature is in to reduce the effects of space-charge creased above normal ambients. limiting. The potential of such a grid is usually between that of the last LINEARITY and the next-to-Iast dynode and is adjusted by observing and maximiz The anode currents produced by ing the value of the anode output photomultipliers are proportional to current. the level of incident radiation over Another factor that may limit a wide range of values. A linearity anode-output-current linearity is plot over a wide range of light level cathode resistivity; cathode resistiv is shown in Fig. 46 for a type 931A. ity is a problem only in tubes with The limit of linearity occurs when semitransparent photocathodes, par space charge begins to form. Space- ticularly of the CsSb or bialkali type. 54 RCA Photomultiplier Manual Individual tube bulletins and data TYPE 8054 listed in the Technical Data section SUPPLY VOLTS = 1000 of this manual give values of the peak linear and the peak saturated pulse currents. The limits of linear operation are determined, of course, by the particular geometry of the output stages and by the applied voltage. Peak linear and saturation cur rents are measured by both dc and pulsed methods. One common method makes use of a cathode-ray tube with a PI5 or PI6 phosphor. The grid is double-pulsed with pulses of unequal amplitudes but fixed amplitude ratio. As the amplitude of the two pulses is increased, a point is observed at which the amplitude of the larger of the anode pulses does not increase in the same proportion as the smaller pUlse. At this point the tube is assumed to become non linear. The current value at this point is then measured by means of an oscilloscope and load resistor. 400 500 'SOD 700 The maximum saturation current is WAVELENGTH - NANOMETERS found when a further increase in Fig. 47 - Spectral-response character radiation level yields no further in istics for a CsSb transmission-type crease in output. photocathode for two different incident angles of radiation. The increased red response for the large angle of incidence USE OF REFLECTED LIGHT may be explained in part by the increase AT THE PHOTOCATHODE in absorption caused by the longer path. Because, in general, photocathodes total internal reflection occurs at absorb less red light than blue, the both the photocathode-to-vacuum spectral response of the photocathode interface and the glass-to-air inter is somewhat sensitive to the angle of face. Several passes of the light incidence of light on the photo through the photocathode become cathode. There is an increase in red possible and thus increase the prob sensitivity at larger angles of inci ability of producing photoemission. dence, as shown in Fig. 47. It is For total internal reflection to oc possible to take advantage of this cur, the minimum angle at which characteristic of photocathodes to light may impinge upon the face enhance the photoemission signifi plate is given by arcsine (lIn), cantly. where n is the refractive index of If a prism is optically coupled to the photomultiplier window material. the faceplate of a transparent photo The greatest enhancement pro cathode, the angle of the incident duced by the total internal-reflection radiation may become such that method occurs in the red region of ! Basic Performance Characteristics 55 the photomultiplier spectral re FOCUSING-ELECTRODE VOLTAGE IS VARIED sponse; gains greater than 5 have BY ADJUSTMENT OF POTENTIOMETER been achieved at 700 nanometers in CONNECTED BETWEEN DYNODE No.1 AND Cj\THOpE. tubes having an 8-20 spectral re 100 sponse. 80 OUTPUT-CURRENT CONTROL .... AND SPATIAL UNIFORMITY Z LU 0: 0: 60 ::l Many photomultipliers are equipped U LU with a focusing electrode, or grid, 0 between the photocathode and the 0z 40 collection of the photoelectrons ~ ...J emitted from the photocathode. The LU focusing-grid voltage is usually set 0: at the point at which maximum 100 anode output current is obtained. In FOCUSING ELECTRODE VOLTAGE- some applications, spatial uniformity, 'Y. OF DYNODE NO.I-TO-CATHODE VOLTS i.e., the variation of anode current Fig. 48 - A typical focusing-electrode with position of photocathode il characteristic. lumination, may be more important than maximizing output current. In such cases, however, the final adjust general, large tubes having relatively ment of the focusing-grid potential large distances between the activat should not differ significantly from ing elements and the photocathode the adjustment that provides opti exhibit the best cathode uniformity. mized collection efficiency. Fig. 48 In these tubes, the activating ele I shows a typical focusing-electrode ments, of which there may be sev characteristic. eral, approach point sources. Spatial uniformity is determined i by photocathode uniformity and by the uniformity of collection effi ciency, i.e., the proportion of the REFERENCES emitted photoelectrons which strike a useful area on the first dynode. 1. M. Lontie-Bailliez and A. Mes Photomultipliers with venetian sen, "L'Influence de la Tempera blind structures usually exhibit the ture sur les Photomultiplicateurs," best spatial uniformity because they 'Annales de la Societe Scienti{ique have a large aperture area, and thus de Bruxelles, Vol. 73, Series 1, a relatively large first-dynode area. p. 390 (1959) Modern high-speed tubes with fo 2. A. Kerns, A. Kirsten, and C. Cox, cused structures designed specifically "Generator of Nanosecond Light for scintillation-counting applications Pulses for Phototube Testing," also provide excellent spatial uni Rev. Sci. Insfr., Vol. 30, No.1, formity; the collection efficiency in p. 31 (1959) these tubes approaches unity. 3. A. J. DeMaria, W. H. Glenn, Jr., Cathode uniformity is determined M. J. Brienza, and M. E. Mack, by the uniformity of the deposition "Picosecond Laser Pulses," Proc. of the photocathode material. In IEEE, Vol. 57, No.1, p. 2 (1969) 56 Statistical Fluctuation and Noise N this section, some of the sources is at this lower limit of photomulti I of noise and output fluctuations plier operation that noise sources in photomultipliers are examined. become most important. Therefore, These sources may be roughly di most of the discussion in this sec vided into two classes: (1) those re tion centers upon the effect of the sulting from the processes by which noise sources on tube output in this the photomultiplier operates, and mode. (2) those resulting from the way in which a tube is designed and manu PHOTON NOISE factured and the conditions under which it is operated. The first class If the photon flux impinging upon is more fundamental because it com the photocathode originates from a prises the statistical processes of thermal broad-band source, the pho photoelectric emission and electron tons can be assumed to be emitted multiplication through secondary at random time intervals. Under emission. Most of the discussion of such conditions, the probability that class 1 sources deals with these n photons will strike the photo processes and their effect on photo cathode in a time interval T is given multiplier output. Class 2 sources by a Poisson distribution, as follows: include thermionic emission from the photocathode, spurious gas ionization (IpT)n within the tube, light feedback from P (n, T) = --, exp (-IpT) (4) the output stages to the photocathode, n. and random emission caused by the natural radioactivity of the various where Ip is the average photon-arrival materials within the tube. rate. The average value for n and the 2 A photomultiplier is so sensitive variance in n, u p , can be obtained a radiation detector that it can oper directly from P(n, T): ate at the lower limit of detectability where output pulses are initiated by n = IpT (5) single photoelectrons from the photo cathode. A count of the number of 0'; = IpT = ii (6) output pulses in a given time inter val provides a measure of the rate The signal-to-noise ratio iii Up is at which photons strike the photo given by cathode. This mode of operation is SNR = (IpT)l (7) called the photon-counting mode; it Statistical Fluctuation and Noise 57 Thus the signal-to-noise ratio of the From the expressions developed in photon flux increases as vr;:;. Appendix A for the average and vari There are a few cases in which ance, the average output D and the the fluctuation in the photon flux variance (T2 are given by has an additional term.1 In broad band thermal sources, however, Eq. ii = 71 (9) (6) yields the proper expression for (1"2 = 71 (1-71) (10) the noise in the input photon flux. The discussion of these expressions NOISE CONTRIBUTIONS FROM is best understood by imagining a THE PHOTOCATHODE steady stream of photons impinging upon the photocathode. The photons The physics of photoemission and are equally spaced in time and ar photocathodes is discussed in the rive at a rate of Ip photons per sec section on Photoemission and Sec ond. In an interval 7' (where 7' is ondary Emission. In the following much larger than lIIp), the number discussion it is assumed that the time of photons N striking the photo between the absorption of a photon cathode is equal to IpT. Because by and the subsequent emission of an hypothesis Ip is constant, the number electron, when emission occurs, is N does not fluctuate. In a time 7', the short '(about 10-12 second). In ad average number of photoelectrons dition, it is assumed that all the n emitted from the photocathode is statistical processes of absorption, equal to TJN. The fluctuations in electron transport within the photo photoelectron number can be com cathode, and photoemission can be puted by use of Eq. (84) in Appen described and characterized by one dix A for the total variance for two number, the quantum efficiency 'YJ. statistical processes in series. Be For a given photocathode, 71 is a cause the photon flux does not fluc function of the photon wavelength, tuate, the total variance in the as explained in the section on Photo photoelectron number is given by emission and Secondary Emission. A simple model of photoemission (1"!.e. = N (1-71) 71 (11) will aid in understanding the noise contributions of the photocathode. Eq. ,(11) can be derived from Eq. For each photon that strikes the (84) in Appendix A with the follow photocathode, an electron is emitted ing substitutions: 2 with probability 71. Optical-reflection DA = N, DB = 71, U'A = 0, and effects at the various cathode inter U'B2 = 71(1-71)' The signal-to-noise faces are accounted for in the value ratio of the photon flux is infinite of 71. The chance that no electron is because it does not fluctuate; how emitted is 1-71' ever, the photoelection number has The statistics for this process can a signal-to-noise ratio given by be found through the use of the gen erating function Qp.c.(s) derived in 71N p e Ji (12) Appendix A at the end of this sec SNR . . = [ (1-71) tion. In the simplified model of the photocathode this function is given by Thus, for quantum efficiencies less than unity, the statistics of the photo Qp.c.(8r = (1-71) + 871 (8) cathode result in a finite signal-to- 58 RCA Photomultiplier Manual noise ratio for the photoelectron (13) number. As shown below, this state ment is valid even when the input and photon beam fluctuates. Fig. 49 shows a curve of the signal to-noise ratio of the photoelectron where lip is the average number of number produced by the hypothetical photons arriving in a time interval noiseless photon flux. When 71 = 1, T. The average photoelectron num the photocathode is a perfect con ber and its variance are calculated verter and no noise is impressed on from Eqs. ( 81) and ( 84), respec the photoelectron number. When 71 tively, in Appendix A. The average is less than 1, the signal-to-noise number of photoelectrons and the ratio decreases with decreasing 71 variance in the photoelectron number until, for values much smaller than are stated, respectively, as 1, the signal-to-noise ratio becomes proportional to the square root of 71. lip.e. = 11 iip (15) 2 and cr~ .•. = iip 11 (1-.11) + 1/ iip 10..--.--.--.--.----..---, or (16) It should be noted that the statisti 8 cal conversion process within the photocathode has not altered the a: functional form of the variance; it z Ul 6 is still proportional to the average z number of particles. The photoelec o 1= trons appear to emanate from a ran o dom source of electrons with average 4 w~ o value 'Y1iip. It is as though the aver b age photon number were reduced :I: Do by a factor 71 in the conversion 2 process, a result that depends on the variance of the photon flux be ing equal to Dp. The photoelectron signal-to-noise ratio is Fig. 49 - The photoelectron signal-to noise ratio resulting from a noiseless photon flux impinging upon the photocathode. A quantum efficiency of 40 per cent reduces the signal-to-noise ratio of the photoelectron flux to about 63 In practice, the signal-to-noise per cent of that of the photon flux. ratio of the input photon flux is It is important to realize that the never infinite; the flux always con degradation in signal-to-noise ratio tains some noise. In most sources, as a result of 71 being less than unity the signal-to-noise ratio of the photon is irreversible in that no amount of flux is given by Eq. (7) and the noise-free amplification can improve variance of the photon beam is given the photoelectron signal-to-noise by ratio. Statistical Fluctuation and Noise S9 In some applications, multi photon emit [(1_TJ)m-r], and the binomial pulses form the input signal. In these applications, integral numbers of coefficient [ m! ] which de- photoelectrons are emitted from the r! (m-r)! photocathode within a time that is scribes the number of ways such an short with respect to the resolution output can occur among the m capabilities of the photomultiplier as photon inputs. a result of a multiphoton input Figs. 50 and 51 show the probabil pUlse. Examples of this type of in ity distribution for photoemission put can be found in radioactive from the photocathode for a train of tracer scintillations, such as those 4-photon input pulses and for two observed in tritium and carbon spec different values of "1' The spectrum troscopy. in Fig. 50 is computed with If, with "1 equal to 1, the input signal consists of a steady train of photon pulses widely spaced in time, each pulse consisting of m photons, a steady train of photoelectron pulses would result, each consisting of m 0.3 '1 = 0 4 electrons. For values of "1 less than 1, the number of electrons in each pulse, r, can vary from 0 to m (0 ~ r ~ m). The probability that 0.1 r electrons are emitted is obtained through the use of the multiple-input particle generation function de OL--...L...... L.----L_-'---....Lt_ o I 2 3 4 scribed by Eq. (86) in Appendix A. No. OF PHOTOELECTRONS/PULSE-r For m-photon input pulses, the gen erating function Qm(s) is given by [Q(s)]m, where Q(s) is the single Fig. 50 - Photoelectron output pulse particle generating function. Thus, spectra resulting from a flux of 4-photon the function Qm(s) is given by input pulses computed with '1/ = 0.4. "1 = 0.4 and that of Fig. 51 with Qm(S) = [(1-1/) + 1/sjm (18) "1 = 0.01. The difference is strik The probability that r photoelectrons ing. With 'YJ = 004, there is no photo will be emitted is given by emission only 13 per cent of the time; most of the photoemission is di vided between 1- and (2-, 3-, or 4-) P r (1/) = (l),(f~mr(S)1 photoelectron pulses in the ratio of r . S S = 0 (19) 13 to 20. In sharp contrast, when or "1 = 0.01, photoemission does not occur 96 per cent of the time. Of the times when photoelectrons are emitted, single-electron pulses occur 70 times more often than any others; Eq. (20) is just the coefficient of sr 3- and 4-photoelectron pulses almost in the expression for Qm(s). Pr(TJ) never occur. It should be clear that is composed of the chance of r suc in multiphoton-pulse spectroscopy it cessful photoemissions (TJr), the is important that "1 be as high as chance of (m-r) failures to photo- possible. 60 RCA Photomultiplier Manual be purely random. Such effects have been observed but are small in or 1.0 dinary thermal sources.! Another source of photocathode noise is the thermionic emission of '7 ;0.01 single electrons. The strength of the emission varies with photocathode type. In the bialkali cathode, for ex ample, thermionic emission is vir tually absent; electron emission exists in the absence of a photo signal, but it is not temperature-dependent as is thermionic dark emission, and it probably arises from a different source. S-I photocathodes, on the other hand, exhibit relatively large 1" (g. 51 - Photoelectron outpUt pulse dark currents as a result of thermionic resulting from a flux of 4-photon input pulses computed with '1 = 0.01. emission; these currents can be elimi nated to some extent" however, by As in the case of a single-photon cooling the tube. pulse train, the average photoelectron Randomly emitted thermionic elec number and the variance in that trons add a term to the fluctuation number can be found from Qm(s) in the photoelectron current pro by use of Eqs. (87) and (88) of Ap portional to their number. Because pendix A. If ex is the average rate of of their independence with respect to I andomly emitted m-photon pulses, any usual signal current, the term then the average number of photo adds to that of the signal noise in dectrons emitted in a time T (where quadrature. That is, 7 is much greater than II ex) is given by U~.e. = T/ up + nth (24) np .e . = T/ m(ar) (21) where nth is the average number of and the variance is given by thermionic electrons emitted in a u~.e. = T/m(ar) (22) time 7, and YJi'ip is the average num ber of photoelectrons emitted in the The signal-to-noise ratio is given by same time. SNRp . . = [T/ m(ar)j! (23) e NOISE CONTRIBUTIONS FROM The input signal-to-noise ratio THE MULTIPLIER CHAIN (mexT)l/2 is degraded by a factor of 0.63 for YJ = 0.4 and by a fac- Although the amplification (multi tor of 0.1 for YJ = 0.01. plication) of the photoelectron cur The effect of the imperfect conver rent in the dynode chain of a photo sion of photons from a purely ran multiplier is often referred to as dom source into photoelectrons noise-free, careful examination of the within the cathode is a reduction in statistics of the gain mechanisms in the input signal-to-noise ratio by a volved shows that this statement is factor equal to the square root of not entirely correct. Approximately YJ. Deviations from this result indi no:se-free operation can be attained, cate that the photon source may not however, with the use of proper elec- Statistical Fluctuation and Noise 61 tron optics and newly developed dy Statistics for a node materials. In the following dis Series of Dynodes cussion the statistical gain processes in the individual dynodes are ex Using the formulas for the gain amined and then combined to yield and fluctuation for a single dynode the statistical properties of the entire developed in Appendix A, the ex multiplier chain. Much of the work pressions of Appendix B may be used presented was accomplished at a to find the combined gain and vari very early stage in photomultiplier ances for the string of dynodes that history.2 make up a multiplier. It is assumed that one primary electron 'impinging Secondary Emission on the first dynode releases, on the average, (h secondaries with variance Photomultiplier dynodes provide 0"12. The output of the first dynode gain through the process of second striking the second dynode produces ary emission, which was discussed in an average gain at the second stage detail in the section on Photoemission of mz and a variance of 0"2 m2' From and Secondary Emission. Most prac Eq. (84) in Appendix A, m2 and 0" im tical dynodes consist of semiconduc can be related to the individual dy tors or of conducting substrates node statistics as follows: surfaced with a thin insulating film. An energetic primary electron pene trates the surface and, ideally, ex cites electrons from the valence band and to the conduction band within the dynode material. These excited elec trons drift and diffuse toward the 2 where 82 and 0"2 are the average dynode-vacuum interface where they secondary emission and variance for escape with a certain probability into the second dynode for a single in the vacuum as secondary electrons. put electron. Continuing in this man For a given primary energy, it is pos ner, the gain and fluctuation from sible to obtain any number of sec the third stage are given by ondaries ns from zero to a maximum n.(max)' The maximum number is given by the quotient of the primary energy Ep and the energy required and to produce a hole-electron pair = 15~ [15; u~ 151 151 15 2 (29) within the dynode Ep' as follows: U!'3 + uil + ui Eq. (29) can be rearranged to read Over many repeated measurements using primary electrons of the same energy, a truncated distribution is obtained for n.. A model that de scribes the observed distributions The extension of Eqs. (29) and (30) from most practical dynodes is de to k stages is accomplished by add veloped in Appendix B at the end ing terms to Eqs. (26) and (27) as of this section. follows: 62 RCA Photomultiplier Manual The noise added to the input signal is very small. It is in this sense that the multiplication chain is said to U!.k = (iih)2\ ~ + Ch~~~ + provide noise-free gain. Multiple-Particle Inputs •••• (Ch'02"~~Ok_l) a,} (32) The output-pulse distribution for mUltiple-particle inputs is obtained from the generating function for the Eq. (31) states the expected results: multiplier chain. A multiplier chain that the total average gain for a consisting of k dynodes has a gen series of k dynodes is the product erating function Qk(S) given by of the secondary-emission yields of the individual dynodes in the series. Qk(S) = Ql {Qd···· • Qk(S)ll (35) Eq. (32) shows that the relative con tribution of any stage to the total The probability Pk(n) of observing fluctuation decreases with the prox n electrons (for a one-electron in imity of the dynodes to the output put) at the output of a k-stage chain end of the chain. The first stage con is derived from Eq. (74) in Appendix tributes most to the total variance. A, as follows: The higher the first-stage gain, the less each subsequent stage contributes to the total variance. This property is an important feature of the new high-gain GaP first-dynode photo The generating function for a k-stage multipliers. multiplier chain for multiple-particle The signal-to-noise ratio for the inputs is given by Eq. (86) of Ap multiplier chain is given by pendix A, as follows: The probability of observing n output electrons from m input particles is given by For large first-stage gains, the mul tiplier signal-to-noise ratio is high. Most of the noise contribution is When identical dynodes are used, the from the first stage. If, in addition output distribution for single-elec to a large gain, the first stage exhibits tron input pulses evolves toward a Poisson statistics, as explained in steady-state distribution after four or Appendix B, the signal-to-noise ratio five stages, and exhibits little change becomes thereafter. Fig. 52 shows some single-electron distributions3 com puted by use of the Polya statistics for each stage in the chain as ex plained in Appendix B. As b ap- Statistical Fluctuation and Noise 63 STAGE GAIN (1'1':3 (f) z '"o >- S 0.4 iii ~ m o 0:: 11. MEAN ! 0L------+------~2------~~:3~~~=--~~4 PULSE HEIGHT Fig. 52 - Computed single-electron distribution for a range of values of parameter b. Parameter b is defined in Appendix B. pro aches 1, the distribution becomes structure are shown in Fig. 53.4 The more sharply peaked. two curves relate to two different Computer values of multiple structures that have the same dynode particle outputs for a two-stage as a first stage. The solid-line curve 1.6...---'r--r--r--~-~-.--"T""---'--"T""-"T""--'---""-'" CURVE A (i) .... 1.4 STAGE I: 1'=5.0, b=0.20 z ::J STAGE 2: 1'=500, boO.Of 8 Lt.. 1.2 ~ o A 0:: STAGE I: JL=5.0, b-0.20 III STAGE2; b=1.00 ~ 1.0 '1"500, ::J ~ ~ 0.8 :J iii ~ 0.6 o 0:: 0.. >'" ~ ~ 0.2 a: (] I 2 3 4 5 6 7 8 9 10 II 13 NUMBER OF SECONDARY ELECTRONS fN A PULSE FROM DYNODE SURFACE Fig. 53 - Theoretical pulse-height distribution. 64 RCA Photomultiplier Manual shows the output when the first stage gains for the new RCA gallium phos is followed by a high-gain (p. = 500) phide dynodes are 30 to 45, and second stage having nearly pure their statistics are nearly Poisson. Poisson statistics (b = 0.01). The Fig. 54 shows the multiple-particle output peaks are sharply defined, and pulse-height distribution for the pulses up to a ten-electron input tube, and Fig. 55 shows the pulse pulse are clearly resolvable. The height curves for a tritium scintilla dashed line describes the final out tion input for a conventional tube put distribution when the first stage using the standard copper-beryllium is followed by a high-gain (p. = 500) first dynode along with, that for a dynode having an exponential output tube with a gallium phosphide first distribution. Individual peaks are no dynode. The increased resolution of longer discernible; the large variance the gallium phosphide dynode is associated with the exponential sta clearly shown. tistics of the second stage eliminates all the structure in the output of the first stage. RCA has recently developed a I- high-gain gallium phosphide (GaP) '"Z ::> dynode which, when used as the first o stage in a conventional copper beryl u ,; lium (CuBe) multiplier chain, greatly z increases the pulse-height resolution of the photomultiplier. The high-gain first stage in a photomultiplier hav CHANNEL No. ing multiple photoelectron events (COMPRESSED SCALE) originating from the photocathode is Fig. 55 - A comparison of the tritium similar to the case illustrated in Fig. scintillation pulse-height spectra ob 53 for a multiplier where the high tained using a conventional photomul gain second dynode amplifies the tiplier having all CuBe dynodes and a multiple pulses originating from the photomultiplier having a GaP first stage. first dynode which in turn are ini Note: The first photoelectron peak of tiated by single electrons. Typical the 8850 spectra includes dark noise from the photomultiplier, chemiluminescence and phos phorescence from the vial and cocktail, and well as Ii" disinte gration. FLUCTUATIONS IN THE TUBE AS A WHOLE In the previous sections the noise contributions from the photocathode and the multiplier chain were consid 01234567 ered; these results can be combined PULSE HEIGHT -PHOTOELECTRON,EQUI to obtain the signal-to-noise ratio for VALENTS the photomultiplier as a whole. The average number of photoelec Fig. 54 - Typical photoelectron pulse height spectrum for a photomultiplier trons from the photocathode in a having a GaP first dynode. time T is given by Statistical Fluctuation and Noise 65 Dp.e. = 1) 0 iip (39) High dynode gains imply that 1/ (S-1) is much less than 1 and The variance is given by hence that U~.e. = 1) 0 iip (40) U! = 1)iip(02k) (47) 2 It is assumed that CTp = nil or that the input flux of photons displays The signal-to-noise ratio at the anode Poisson statistics. If the form for is given by CT p 2 is not the result of a Poisson process, Eq. (84) in Appendix A must 2 be used to obtain CTpe • Using these expressions to describe the input to the photomultiplier For high-gain dynodes exhibiting chain, the average number of elec Poisson statistics, therefore, SNR" is trons collected at the anode can be essentially that of the photoelectrons, stated as follows: SNRp .•. , as given in Eq. (17). In a photomultiplier in which the n.. = 1/ o np o iih (41) dynode gain is not high but still ex hibits Poisson statistics, SNR" is where mk ·is given in Eq. (31) and is the average gain of a k-stage multi given by plier. The variance for the output electron stream is given by u; = mt°1) oiip + 1/oiipoU~k (42) where CT mk2 is given by Eq. (32) and , is the variance in the average gain As an example, with S = 4 the SNR" . of a k-stage multiplier chain. is decreased by a factor of 0.87 from Eq. (42) can be rearranged as its value for very large S. Doubling follows: S to a value of 8 changes the deg radation factor to 0.94. Further in u! = 1)np \m~+ U!k) (43) creases in 8 do not improve the For equal-gain stages described by SNR" very much. Poisson statistics in the multiplier In the case of fully exponential chain, Eq. (42) becomes dynode statistics, the variance for each dynode in a chain of identical dynodes is given by Eq. (94) in Ap (44) pendix B; i.e., = 152 15 (50) Neglecting the 1/ Sk term and sub U: + stituting the result and Sk for mk where 8 is the average gain per stage. in Eq. (43), The total anode fluctuation then be comes u; = 1/ iip [c~2k (1 + ~~ 1) ] (45) In this case For large dynode gains, 2k u; .". 2(1) 1'i.p ) c5 (52) 66 RCA Photomultiplier Manual Even for large dynode gains, ex related with and (2) those that are ponential statistics increase u a2 by a correlated with the signal pulse. factor of 2 with the result that SNRa decreases by 1/-y 2. This drop in GROUP 1 NOISE SOURCES SNR is accompanied by a severe loss in single- and multiple-electr~n Radioactivity within the pulse-height resolution, as shown 10 Photomultiplier Fig. 53. To resolve single-photoelec The materials used to fabricate tron pulses, the multiplier chain must the internal structure and the glass exhibit both high gain and good envelope of a photomulti~lier m.ay (i.e., Poisson) statistics. . contain amounts of certam radiO A significant improvement 10 active elements that decay and give SNRa results when the photocathode off gamma rays or other high-en~rgy quantum efficiency 7/ is increased. At particles. If one of these emitted present the peak value of 'YJ. for the particles strikes the photocathode or S-20 cathode is 0.35; doublIng 1] to one of the first few dynodes, it will 0.7 would improve SNRa by a factor produce an anode dark pulse. The of 1.4. The ideal photomultiplier, for size of the anode pulse may be which 1] = 1, would have an SNRa equivalent to one or more photoelec equal to the square root of npo The trons emitted at the photocathode. present value of 1] = 0.35 degrades These pulses are randomly emitted. the ideal SNRa by a factor of 0.59. In tube manufacture this type of Considerable improvement can be ex emission is minimized through the pected with the development of careful selection of materials. photocathode materials of increased sensitivity. . Dark Pulses of Unknown Origin The application of these equations to present photomultipliers indicates Dark pulses of unknown origin in that the available photocathode quan a photomultiplier result primarily tum efficiency is the principal degrad from single-electron pulses; a small ing influence on SNRa• Values of 1] fraction consists of 2- and 3-electron less than 1 also decrease the mul pUlses. The rate of occurrence of tiple-photon input-pulse resolution .of these dark pulses varies among mul a photomultiplier. High dynode gams tipliers; the lowest rate is about 150 (8 greater than 6) do not significantly counts per minute for the new RCA- degrade the input signal-to-noise 8855 series. These dark pulses are ratio of photoelectrons provided the randomly emitted, but may be ac dynode statistics are nearly Poisson. companied by one or m?re of t~e noise effects correlated With the sig nal pulse. The problem of the origin OTHER SOURCES OF NOISE IN of unknown pulses is at present a PHOTOMULTIPLIER TUBES very active field of research. Within a photomultiplier there are GROUP 2 CORRELATED sources of noise that are not asso NOISE SOURCES ciated directly with the processes of photoelectric conversion and elec tron multiplication. These sources Gas Ionization can, in general, be separated into two A gas atom or molecule within the groups: (1) those that are not cor- photomultiplier may be ionized by a Statistical Fluctuation and Noise 67 photoelectron pulse. This ionization erally above the recommended maxi may occur at the first-dynode sur mum operating voltage. face or in the vacuum between the Not every signal pulse initiates an photocathode and the first dynode. afterpulse. Therefore, coincidence The positive ion thus created travels techniques, using more than one backward to the cathode where it photomultiplier, can be employed in may release one or more electrons some instances to eliminate this from the photocathode. Because source of noise as well as the un there is a time delay in the ion correlated sources discussed above. emitted electron pulse equal to the Analytically, the fluctuations re time of flight of the ion to the photo sulting from Group 1 non-correlated cathode, the resulting pulse, usually random sources add in quadrature referred to as an afterpulse, occurs to those of the signal. Because they after the true signal pulse at the are random, the dark-pulse variances anode. Afterpulses are caused mainly CT dp2 are proportional to the aver by hydrogen ions and their occur age dark-pulse rate iidp, as follows: rence can be minimized in tube processing. Light Feedback The total anode variances are given by: Primary electrons produce photons as well as secondary electrons within the dynodes of the multiplier chain. Despite the low efficiency of this process, some of the emitted photons may eventually reach the photo- where 7Jiip and DdP are the average , cathode and release additional elec single-photoelectron and the average trons. A time delay is observed single-dark-emitted electron numbers corresponding to the transit time for observed in a time T. The factor the regenerated electron pulse to ~p sets a real limit on the anode reach the point of origin of the light. signal-to-noise ratio. Again, in those Depending upon the type of dynode instances where the incoming radia multiplier cage, this time may be of tion signal comprises more than one the order of 20 nanoseconds. Most photon, coincidence techniques can of the photons comprising the light be employed to reduce the effect of feedback originate in the region of randomly emitted electrons originat the last few dynodes or of the anode. ing at the cathode. If the voltage across the tube is Electrons may originate at dynodes increased, the dark-pulse rate also well along in the multiplier chain. increases and usually produces some At the anode, these electrons appear observable light near the anode re as fractional-photoelectron pulses. gion, a fraction of which is fed back Such pulses also result from inter to the photocathode. The result of stage skipping, generally near the be this positive feedback is that, at a ginning of the chain. With good certain voltage, the photomultiplier statistics in the chain, the fractional becomes unstable and allows the out pulses may be discriminated against put dark-pulse rate to increase to an because the single-electron peak in intolerably high level. The voltage the pulse-height spectrum stands out at which this increase occurs is gen- sharply. 68 RCA Photomultiplier Manual NOISE AND THE BANDWIDTH Af = 1/4RC (58) OF THE OBSERVATION From Eq. (7), the SNR for a ran At high counting rates, noise cal dom photon flux is given by culations are performed with the av erage and variance of the photo SNRp = VIp T (59) electron current rather than with where Ip is the average photon ar individual photoelectron pulses. The rival rate and T is the time interval expressions which have been de of the count. When T = 11 (2E,f) is veloped for signal-to-noise ratio by substituted, consideration of the average number of events in a time T and the variance SNRp = VIp/2M (60) from average may be converted to expressions of signal-to-noise ratio In the case of the photocathode elec involving currents and bandwith, E,f, tron current, by considering the reciprocal nature of the observation time and the band width. 1 - '1p_I ~-.1 p e (61) Noise equivalent bandwidth E,f SNR . . - ~2Af - 2eM may be defined 5 as follows: where i is the photocathode emission 1 current in amperes and e is the M = A2 foo IH(jw) 12 df (55) charge of the electron. If the signal m 0 current is considered as i, the noise where H(jw) is the complex fre in the bandwidth E,f for the photo quency transfer response of the cir current is the familiar shot noise cuit, and Am is the maximum abso formula (2eiE,f)1/2. lute value of H(jw). The "circuit" The equation for the photon noise in this case counts pulses in a time squared is given by T. The network transfer function is the Laplace transform of the im pulse response. The impulse response is a rectangular pulse of width T, The photocurrent noise squared is like a camera shutter. For this case, given by H(jw) = !- (1 - e-irw ) (56) JW Both these equations involve the photon "current" Ip- and Am is found to be equal to T. If the photons are not randomly The noise equivalent bandwidth is emitted, Eq. (62) must be modified. then readily found to be In the case in which the variance in the photon current is given by = 2 Af 1/2T. (57) (J"p , the equation becomes It is of interest to compare this re lation with the equivalent noise bandwidth for exponential impulse response as in an RC circuit; in this (64) case Statistical Fluctuation and Noise 69 For a randomly emitted photon, the With 1J = 0.3 and ak = 106, noise current squared at the anode is given by Ip = 10 i:/(2e2 1J a2k df) (69) or d I~ = 2 e2 1J a2k Ip df (65) Ip ~ 2 X 105 photons per second where the multiplier chain is assumed to be composed of k high-gain dy The value of Ip shown in Eq. (69) nodes exhibiting Poisson statistics, is the lower limit for the average each with an average gain 8. photon current and makes the At the anode, the input resistance squared anode-current fluctuation and capacitance of a preamplifier greater than the squared noise cur generate a noise current squared rent in the photomultiplier preampli given by fier input by a factor of ten. The resulting value of the average photon current corresponds to a photoelec = 4kTg M i~ tric current of about 10-14 ampere, or about 6 x 104 photoelectrons [1 + Rng + 4;2. ~n df2C2J (66) per second. In cases in which the dark current is effectively higher than 6 x 10 4 photoelectrons per sec where g = lIR is the shunt con ond, the photomultiplier sensitivity ductance in the anode lead, C is the limit is set by the dark current and shunt capacitance, 6f is the band not by the preamplifier noise. It is width, k is Boltzmann's constant, T the noise-free gain of the multiplier is the absolute temperature, and Rn chain which increases the rms photo is the equivalent noise resistance of current shot noise by a factor of the preamplifier input. The total 8k and permits the low-input-Ievel noise current squared through R is operation. 6i2aN + in2 , and SNRa is given by SUMMARY SNRa = 7]a k l pe I [2 e2 7]a2klp M The more important expressions + 4k Tg M + Rng relating to signal and noise dis (1 cussed in this section are summarized below. Beginning at the input to the + 4;2 ~n M2C2) ] 1/2 (67) photomultiplier, the signal is followed through the tube, and the variance and the SNR associated with the signal are noted. To maintain the input SNR. 6i2aN First, a randomly emitted photon must be greater than i 2• For ex n flux exhibiting Poisson statistics has ample, with 6J2aN = 10 i2 and an average value ii given by R = 100 ohms, C·= 100 picof~rads, o Rn = 100 ohms, T = 300 K, and ii = IpT 6f = 106 Hz, at room temperature, for the number of photons present i~ = 3.2 10-16 amperes2 (68) in a time interval T when the aver age rate of emission is Ip photons 70 RCA Photomultiplier Manual 2 per second. The variance u is given where 8j and ul are the gain and by variance of the jth stage, respectively, and ink and u mk are the total gain and the total deviation, respectively, for the entire multiplier chain. and the signal-to-noise ratio is given For the photomultiplier tube as a by whole, the anode SNR is given by SNR = iip//T p = VI;:;. = ~ SNRa = ~/[I+/TV8~+/Ti/(8l·8i) The number of photoelectrons + .. ·+/TV emitted from a photocathode of «(81· 82 ••• 8k-l)8~)]l quantum efficiency 'TJ is given by If each stage is identical and has Poisson statistics, and the variance is given by and for large gain (Le., 8 > > 1), The signal-to-noise ratio of the photoelectron ftux is given by In this case the multiplier chain has SNRp .•. = ~ not reduced the photoelectron SNR. If the photocathode contributes If each stage in the multiplier randomly emitted thermionic elec chain is identical and has exponen trons to the signal of photoelectrons, tial statistics, then then the variance of the electron ftux is increased, as follows: and for large gain, where nth is the average number of thermionic electrons emitted in a time T. The SNR is decreased and The SNR for exponential statistics becomes is decreased by 1r{"2 below that of a high-gain multiplier chain with Poisson statistics. When the photomultiplier is oper The signal-to-noise ratio of the ated at high photon counting rates: mUltiplier chain with k dynodes alone is given by SNRa = 'I18 klpe I [2 e2 'I18 2kIp .1£ + 4k Tg .1r(1 + Rng + 4;2 ~n .1f2C2) ] 1/2· Statistical Fluctuation and Noise 71 where Ip is the average photon cur The expression fil CT is the signal-to rent in photons per second, e is the noise ratio for the device in ques electronic charge, l1f is the fre tion (CTln is the relative deviation). quency bandwidth, T is the absolute Obviously, the smaller the value of temperature CK), k is the Boltz CT, the more precisely determined is man constant, Rn is the equivalent the output. noise resistance of the input of the An alternative and more powerful preamplifier that processes the anode method of obtaining the results signal, g is the shunt conductance above, as well as others of a more in the anode lead, and C is the shunt complex nature, involves the use of capacitance in the anode lead. the generating function of the dis 6 tribution [Pn1. The generating func APPENDIX A tion Q(s) in terms of the auxiliary STATISTICAL PROCESSES AND variable s is defined as follows: THE GENERATING FUNCTION N Statistical processes are character Q(s) = ~ sn Pn (73) ized by randomly distributed outputs n-O for a given input. The relative fre and has the following properties: quency of occurrence of any particu lar output is given by the probability that this output occurs. As an ex ample, for a single-particle input, the output of a particle-multiplication device such as a photomultiplier may Q(s)!._I=l (75) consist of from zero to N particles. A particular output n occurs with probability P n' The distribution of a Q(s) I = ii (76) P n sums to unity as follows: as 8-1 N ~ Pn = 1 (70) and n...o The average number of output par ticles is given by N Ii = ~ n P n (71) On any given single measurement, an output consisting of n particles The statistical properties of a is not expected. In fact, n may not number of individual devices in be integral, even though all possible series (the multiplier chain of a outputs must be. There will be fluc photomultiplier is an example) and tuations in the output which are the average output and fluctuation given by the variance CT2, or by the for a single device when the input rms or standard deviation CT, such consists of more than one particle that can now be obtained through the use of the generating function Q(s) (J = [f (n_ii)2.Pn]i (72) of the single-particle-input probabil n-O ity distribution [P n1. 72 RCA Photomultiplier Manual Fig. 56 shows the series combina or tion of two statistical devices. The first has an output distribution [P nlA ~B(S) = I!P1P~ + S2(P1P~ + P2P~P~) which yields an average gain DA and + s3(2 P2P~P;) + s'P2P;P; variance U A2 . The generating func (79) tion for [P nlA is QA(S). Similarly, for the second device the output distri The coefficients of sl (j = 1, 2, 3, 4) bution is [P OlB' the average gain is give the probability of a j-particle DB, the variance is UB2, and the gen output for the entire series combina erating function is QB(S). The rela tion. Each coefficient comprises the tionships between these individual sum of all possible ways to obtain device functions and the total average that particular number of particles output nAB and the variance U AB2 at the output. for the entire serial combination are From Eq. (78), the output pro~ required. The first step is to state ability distribution [PnlAB' the tO,tal the generating function QAB(S) for average output DAB, and the total the serial combination shown in Fig. variance U AB2 for the suies combina 56. tion can be ob~rough the use --I Fig. 56 - A schematic representation of the serial combination of two statistical devices A and B. For a single-particle input the gain and variance of the signal at each stage in the device are shown. The total gain and total variance for the entire combination in terms of each of the individual device characteristics for a single-particle input is given by Eqs. (81) and (84) in the text. The generating function for the of Eqs. (74), (76), and (77). Of most serial combination is given by the interest are the total average out following relation: put and the total variance. From Eq. (76), (80) This relation can be understood more clearly if the two devices are con sidered to have only two possible or outputs, either one particle or two particles. Then the generating func tions are QA(S) = sP1 + S2P2 and QB(S) = sP1' + S2P2'. When these functions are inserted, Eq. (78) be comes = [OQA. OQB] = fiA •fiB (81) OS Os 0_1 As expected, the average output for QAB(S) = SP1P~ + S2P1P; + S2P2P~P~ the series combination is the product + 2Sap2p~p; + s'P2P;P; of the average outputs of the indi- Statistical Fluctuation and Noise 73 vidual devices. The variance (TAB2 QABC o 0 0 0 z = QA {QB [0000 oQz(s)1l is given by (85) Eqs. (81) and (84) provide a recur sion relation between any two de vices in the series so that the total After the differentiation is per gain and variance from any number formed, of devices can be obtained. These recursion relations are used in dis cussions of photocathode and mul criB = Q~ [Q~)2 + Q~ Q~ + Q~ Q~ tiplier-chain statistics. - [Q~12 [Q~12 (83) When the input to a particular de vice consists of multiple, simultane where the prime indicates differentia ous, independent events, the gener tion with respect to s and evaluation ating function for the combination of the derivative at s = 1. By addi is obtained by taking the product of tion and subtraction of a term [QA'] the separate generating functions. [QB']2, Eq. (83) can be rearranged Thus, for two such parallel inputs as follows: to a statistical device, the correspond ing generating functions are: criB = [QB]2 [Q~ + Q~ - [Q~]2] + Q~ [Q~ + QB - [Q~121 N M QA = ~ sm PAm and QB = ~ sm PBm or, in terms of the individual-device m-o m-G average outputs and variances, as follows: When the product is formed, the individual coefficients of Sj represent interpret the bulk of the observed secondary-emission statistics. The distribution has the follow ing form: Fig. 57 - A schematic representation P(n,b) = Jl~ (1 +bJl)-n-ljb nil (1 +jb) of a statistical, multiplicative device n. i-I with an m-particle input. The total gain and variance for the output in terms (89) of the single-particle input character where Pen, b) is the probability of istic are given by Eqs. (87) and (88) observing n secondaries, /L is the in the text. mean value of the distribution, and By use of Eqs. (76) and (77) b is the parameter controlling the llm = m'D (87) shape of the distribution. With b and = 0, the distribution is Poisson, as 2 follows: C1~ = m' 0- (88) The output probability distribution Jln Pen 0) = - e-I' (90) [P n]m is obtained by using the m , nl particle generating function [Eq. (86)] in Eq. (74). With b = 1, an exponential distribu Eqs. (87) and (88) can be obtained tion results, as given by: from Eqs. (81) and (84) by assuming P(n,I) = Jln (1+Jl)-(n+1) (91) in that case (see Fig. 56) that the first serial device is a perfect multi Fig. 58 shows a family of distribu plier with no variance (O"A2 = 0). tions for various values of b. Under this restriction, the input to The generating function for device B is a constant m, and the P(n, b) is given by gain and variance found from Eqs. Q(s) = [1 +bJl (I_S)]-l/b (92) (81) and (84) are identical to those given in (87) and (88). From Eqs. (76) and (77) of Appen dix A, APPENDIX B D=Jl (93) MODEL FOR DYNODE and STATISTICS 2 0- = bJl2 + I' (94) The observed number distributions 2 for secondary electrons vary among The fluctuations increase as /L for the different types of dynodes used an exponential distribution (b = 1), commercially. Nearly all the distri but increase only as /L for the Pois butions fall within the class limited son distribution. The signal-to-noise ratio is given by by a Poisson distribution 7 at one ex treme and an exponential distribution D at the other.4 To describe this wide - = Jl/(bJl2+Jl)t (95) variety of distributions the Polya, or 0- compound Poisson, distribution is Fig. 59 shows a log-log plot of the employed.3 Through the adjustment signal-to-noise ratio as a function of of one parameter, the distribution /L with b as a parameter. The signal runs from purely Poisson to expo to-noise ratio improves with increas nential. Therefore, this one distribu ing /L for the Poisson distribution, tion can be used to describe and to but approaches unity with large /L in Statistical Fluctuation and Noise 75 described by Eq. (89) can be shown to be composed of a number of dif ferent Poisson processes, each with ,.= 25 a different mean value. The mean values are, in turn, distributed ac cording to the Laplace distribution. ~1(j4 go -' Z o ~ 10-5 m ii: I- i5'" .... ~ 10-· '" -' oz iii 4 ~ 4 ~ , g: 1(j7 -' ~ 2 II.. <.!) o in ::E :z: 0.1 L---:l-..L-!-I-~!:u--+--'-+y..~ I- I 2 4 6 8 100 ~ l(j' <.!) 9 Fig. 59 - A comparison of the SNR as a function of the mean value I' for 1(j90 20 40 60 80 100 120 a number of Polya distributions. For No. OF OUTPUT PARTICLES, n Poisson statistics (b = 0), the SNR in creases as the square root of 1'. With o < b ~ 1, the SNR approaches the Fig. 58 - Single-particle output dis square root of b-1 for large mean gains. tri~ution for a dynode displaying Polya statistics. A value of b = 1 gives a probably exponential distribution; b = When b equals 0, the distribution of o gives a Poisson distribution; b = 0.2 the mean values collapses to a delta is intermediate between the two extreme values. function at the value JL, leading to a purely Poisson distrib?tion. For b the exponential distribution. In fact, equals 1, the distribu,ion of tile mean for any non-zero value of b, the sig values is exponential. The physical nal-to-noise ratio approaches b-1/ 2 interpretation for a dynode display for large JL. As shown in Fig. 59, ing non-Poisson statistics is that even small departures from Poisson physical non-uniformities on the dy statistics significantly reduce the sig node surface cause each element of nal-to-noise ratio at moderately high the surface to have a different mean gains of 10 to 20. It can be antici value for emission. Although each pated that departures from Poisson small element exhibits Poisson statis statistics degrade the single-electron tics with respect to emission, the pulse-height resolution. total emission from the entire dy The Polya distribution has an in node is non-Poisson because it com teresting interpretation with respect prises a distribution of Poisson dis to secondary-emission statistics.3 For tributions. It is possible that the basic non-zero values of b, the distributions emission process from a given dy- 76 RCA Photomultiplier Manual node is not a Poisson process. How tribution shown in Fig. 60, it would ever, high-gain GaP dynodes exhibit be difficult to distinguish among one-, nearly Poisson statistics 7 and at three-, and five-particle input pulses, present it is believed that departures whereas the problem virtually dis from this norm are caused, to a large appears for a Poisson distribution. extent, by dynode non-uniformities. The departure from Poisson statis REFERENCES tics affects the single-particle pulse height resolution. Fig. 60 shows the 1. R. H. Brown and R. Q. Twiss, output-pulse distribution from a sin "Interferometry of the Intensity gle dynode for a number of multiple Fluctuations in Light", Proc. particle inputs. The resolution is Royal Soc. London, Vol. 242A, clearly degraded in passing from a p. 300 (1957), Vol. 243A, p. 291 Poisson distribution to an exponen (1958). tial one. With the exponential dis- 2. W. Shockley and J. R. Pierce, "A Theory of Noise for Electron Multipliers", Proc. IRE, Vol. 26, p. 321 (1938); V. K. Zworykin, SINGLE PARTICLE GAIN: 100 G. A. Morton, L. Malter, "The m = INPUT PARTICLE NUMBER Secondary Emission Multiplier A New Electronic Device", Proc. _ 10-1 POISSON DIST., b=O IRE, Vol. 24, p. 351 (1936). ..c m=J 3. J. R. Prescott, Nuc. Instr. Meth c: n. ods, A Statistical Model for photo multiplier Single-Electron Statis tics", Vol. 39, p. 173 (1966). z - o 4. L. A. Dietz, L. R. Hanrahan and i= :::> A. B. Hance, "Single-Electron Re m sponse of a Porous KCI Trans ~ 10-3 en mission Dynode and Application ;:; of Polya Statistics to Particle > I ::; Counting in an Electron Multi a; plier", Rev. Sci. Insf., Vol. 38, ~ 10-" o p. 176 (1967). II:n. 5. M. Schwartz, Information Trans ...o mission, Modulation, and Noise, ::!: :I: McGraw-Hill, p. 207 ((959). ~ 10-5 6. T. Jorgensen, "On Probability Application of Photomultipliers HIS section discusses the major is called the scintillator. A photomul Tapplications of photomultipliers tiplier mounted in contact with the and describes some of the special scintillator provides the means for considerations for each application. detecting and measuring the scintil The applications covered include lation. Fig. 61 is a diagram of a scintillation counting, Cerenkov radi basic scintillation counter. The three ation detection, time spectroscopy, most probable ways in which incident laser range finding, flying-spot scan gamma radiation can cause a scin ning, detection of low light levels, tillation are by the photoelectric ef photometry, and spectrometry. This fect, Compton scattering, or pair listing is not complete even for pres ently known applications, and new ones are being continually devised. r-REFLECTOR The applications discussed, however, (AI,O. POWDER) are some of the major ones and the SCINTILLATOR information given can be readily PROTECTIVE ~COVER (AI) adapted to other applications or to OPTICAL ~~~~~~~GLASS WINDOW new ones. COUPLING - PHOTOCATHODE SCINTILLATION COUNTING AND CERENKOV RADIATION ~ DETECTION PHQTOMULTIPLIER A scintillation counter is a de Fig. 61 - Diagram of a scintillation vice used to detect and register in counter. dividuallight flashes caused by ioniz production. The reaction probabili ing radiation, usually in the form of ties associated with each of these an alpha particle, beta particle, types of interaction are a function of gamma ray, or neutron, whose en the energy of the incident radiation ergy may be in the range from just as well as the physical size and atomic a few thousand electron-volts to number of the scintillator material. many million electron-volts. The In general, for a given scintillator, most common use of scintillation the photoelectric effect predominates counters is in gamma-ray detection at small quantum energies, the and spectroscopy. Compton effect at medium energies, The gas, liquid, or solid in which and pair production at energies a scintillation or light flash occurs above 1.02 MeV. 78 RCA Phetolntdtiplier Manual Scintillation Processes In the photoelectric effect, a gamma-ray photon collides with a bound electron in the scintillator and imparts virtually all its energy to the electron. In the Compton effect, a gamma-ray photon with energy E = Fig. 63 - Pair-production mechanism. hv interacts with a free electron in the scintillator and transmits only part of its energy to the electron, as energy is transferred as kinetic en shown in Fig. 62. A scattered photon ergy. When the positron is anni of lower energy also results. To sat hilated, two photons are produced isfy the conditions of conservation 180 degrees apart, each with an en ergy of 0.51 MeV. The photons are then subject to the normal probabili ties of interaction with the scintil lator. In neutron detection, unlike alpha or beta-particle or gamma-ray detec tion, the primary interaction is with the nuclei of the scintillator rather Fig. 62- Compton-effect mechanism. than its atomic electrons. The inter action may consist of scattering or of energy and momentum, there is a absorption; in either case, some or maximum energy that can be trans all of the energy of the neutron is ferred to the electron. This maximum transferred to the recoil nucleus energy, known as the Compton edge, which then behaves similarly to an occurs when e in Fig. 62 is 180 de alpha particle. grees and cf> is 0 degrees; it is given . In each interaction between a form by of ionizing radiation and a scintil lator, an electron having some kinetic E energy is produced. A secondary T eM = 1 + (1/2a) (96) process follows which is independent of both the kind of ionizing radiation incident on the scintillator and the where E is the photon energy, Of. = type of interaction which occurred. EI m.,c2, m., is the rest mass of an In this secondary process, the kinetic electron, and c is the speed of light. energy of the excited electron is dis The resultant energy imparted to the sipated by exciting other electrons electron can then range from zero to from the valence band in the scin a maximum of TCM• tillator material into the conduction In pair production, as shown in band. When these excited electrons Fig. 63, the energy of a gamma ray return to the valence band, some of is converted to an electron-positron them generate light or scintillation pair in the field of a nucleus. The photons. The number of photons pro gamma ray must have energy at least duced is essentially proportional to equal to two times the rest-mass the energy of the incident radiation. energy equivalent of an electron, 2 In the photoelectric interaction de 2 (moc ), or 1.02 MeV; any additional scribed above, all of the incident Application of Photomultipliers 79 photon energy is transferred to the ters in which the excited electron excited electrons; therefore, the num returns to the valence band with the ber of photons produced in this sec dissipation of heat without emission ondary process, and hence the of light; and phosphorescence centers brightness of the scintillation, is pro in which the excited electron can be portional to the energy of the inci trapped in a metastable state until dent photon. it can absorb some additional energy and return to the valence band with the emission of a photon. An im Scintillation Mechanism portant process in the transfer of energy to the fluorescence, or activa The exact mechanism of the scin tor, centers is the generation of ex tillation or light-producing process is citons or bound hole-electron pairs. not completely understood in all These pairs behave much like hydro types of materials; however, in an gen atoms and, being electrically inorganic scintillator, the phenome neutral, can wander freely through non is known to be caused by the the crystal until captured by fluores absorption of energy by a valence cence centers. The emission of a electron in the crystal lattice and its photon from a phosphorescence cen subsequent return to the valence ter is a relatively slow or delayed band. Fig. 64 shows a simplified en process. Of the three types of cen ergy-band diagram of a scintillation ters through which an excited elec crystal. The presence of energy levels tron can return to the valence band, or centers between the valence and the first, fluorescence, is that sought conduction bands is the result of im for in the preparation of scintillators. perfections or impurities in the crys The second type, quenching, tends to tal lattice. Three types are important: lessen the efficiency of the scintilla fluorescence centers in which an tor because it does not cause the electron, after excitaton, quickly re emission of photons, and the third, turns to the valence band with the phosphorescence, produces an unde emission of a photon; quenching cen- sirable background glow. ELECTRONS I IMPURITY LEVELS - I T EXCITED STATES OR TRAPS I EXCITON I ACTIVATOR CENTER I I -1 GROUND STATE I Fig. 64 - A simplified energy-band diagram of a scintillation crystal: (a) electrons in phosphorescence center; (b) electrons in fluorescence center; (c) electrons in quenching center. 80 RCA Photomultiplier Manual Scintillator Materials photons arriving at the faceplate of a photomultiplier as a result of a The most popular scintillator ma single ionizing event E is given by terials for gamma-ray energy spec trometry is thallium-activated sodium relative standard deviation "" iodide, NaI(Tl). This material is (N )1/2jN (97) particularly good because its re p p sponse spectrum contains a well defined photoelectric peak; i.e., the were Np is the average number of material has a high efficiency or photons arriving at the faceplate of probability of photoelectric inter the photomultiplier per incident action. In addition, the light emitted ionizing event. Np is equal to c E g, by the material covers a spectral where c is the average number of range from approximatel~ 350 to photons per unit energy for the scin 500 nanometers with a maximum at tillation, E is the energy of the inci about 410 nanometers, a range par dent radiation, and g is the fraction ticularly well matched to the spectral of the total number of photons pro pesponse of conventional photomulti duced which arrive at the faceplate pliers, NaI(Tl) does not, however, of the photomultiplier. have a fast decay time in comparison Eq. (97) implies that it is highly to other scintillators, and, therefore, desirable that all emitted photons is generally not used for fast-time be collected at the photomultiplier resolution. As a comparison, the de faceplate. This collection problem cay time constant for NaI(Tl) is can be simplified by careful selec approximately 250 nanoseconds, tion of the shape and dimensions of while for a fast plastic scintillator the scintillator to match the photo the dominant decay time constant is multiplier photocathode dimensions. approximately 10 nanoseconds. The coating of all sides of the scin The decay time of a scintillator in tillator except that which is to be volves the time required for all the exposed to the photomultiplier face light-emitting luminescence centers to plate with a material which is highly return excited electrons to the valence reflective for the wavelengths of the band. In some of the better scintilla photons emitted by the scintillator tors the decay is essentially exponen also proves helpful. Because NaI tial, with one dominant decay time (Tl) is damaged by exposure to air, constant. Unfortunately, most scin it is usually pa£kaged in an alumi tillators have a number of com num case lined with highly reflective ponents each with different decay Al20 s powder; the NaI(Tl) scin time. tillator is provided with an exit win dow of glass or quartz. To avoid Collection Considerations total internal reflection, it is im portant that the indices of refrac Because scintillations can occur tion of the scintillator material, its anywhere in the bulk of the scintil window, any light guide, and the lator material and emit photons in photomultiplier faceplate match as all directions, there exists the prob closely as possible. lem of collecting as many of these If it is not convenient for the photons as possible on the faceplate photomultiplier to be directly cou of the photomultiplier. The relative pled to the scintillator, as when the standard deviation in the number of photomultiplier entrance window is Application of Photomultipliers 81 not flat, light guides can be used. tion, and when the scintillator is Again, care should be taken in de thick in comparison to its diameter, signing the light guide to assure as shown in Fig. 65(a), the photo maximum light transmission. The cathode is approximately uniformly outer side or surface of the light illuminated during each scintillation. guide should be polished and coated However, if the scintillator is thin with a highly reflective material. In in comparison to its diameter, as some cases, a flexible fiber-optics shown in Fig. 65(b), or if the scin bundle can be used. An optically tillation is very weak, the illumination transparent silicone-oil coupling fluid of the photocathode as a function should be applied at the scintillator of position is closely related to the (-light guide, if used)-photomultilier position of origin of the scintillation; interface regardless of the light-con therefore, photocathode uniformity is duction method used. much more important when a thin The next important consideration scintillator is used.' Photocathode in scintillation counting is the con uniformity also becomes more im version of the photons to photoelec portant as the energy of the inci trons in the photocathode. The dent radiation becomes less and the photocathode should have the great number of photons per disintegra est quantum efficiency possible over tion is reduced. the spectral range defined by the In some photomultipliers, the col spectral emissivity curve of the scin lection efficiency for photoelectrons tillator. The method of determining decreases near the outer edges of the the quantum efficiency of a photo photocathode; therefore, best results cathode as a function of wavelength ar~ obtained when the scintillator is is explained in the section on Basic slightly smaller in diameter than the Performance Characteristics. The bi photocathode. alkali photocathode K2CsSb provides the highest quantum efficiency (ap proximately 30 per cent at 385 nano meters) of all photocathodes presently available for matching to most scin tillators. The uniformity of the photo cathode, i.e., the variation in quan tum efficiency at a given wavelength as a function of position on the photocathode, is also important. Be (a) cause the number of photoelectrons emitted for a constant number of photons incident on the photocathode is proportional to the quantum effi ~w=tfcI1VE ciency, any variation in quantum efficiency as a function of position coupUNG results in an undesirable variation in ,- . FLUID the number of photoelectrons emitted r(ff+ as a function of position. L PHOTOCATHODE (b) When the scintillation is fairly Fig. 65 - Scintillator geometries: (0) bright, i.e., when a large number of thick in comparison to diameter; (b) photons are produced per scintilla- thin in comparison to diameter. 82 RCA Photomultiplier Manual Multiplier Considerations Significant Photomultiplier Characteristics The two multiplier configurations most commonly used in photomulti Photomultiplier dark noise is of pliers in scintillation-counter appli particular importance in scintillation cations are the venetian-blind and counting applications when the en focused (either in-line or circular ergy of the ionizing radiation is small, cage) structures. The advantage of or when very little energy is trans the venetian-blind structure is that ferred to the scintillation medium; in the useful portion of the first dynode short, when the flash per event rep area is quite large in comparison to resents only a few photons. If a the photocathode area. On the other photomultiplier is coupled to a scin hand, the focused structure has a tillator and voltage is applied, the small first-dynode area and, there composite of all noise pulses coming fore, requires a more complex front from the photomultiplier is referred end design to provide good collection to as the background of the system. efficiency. However, a well designed The plot of a frequency distribution focused structure, although more of these pulses as a function of en complex, can provide a better collec ergy is shown in Fig. 66. The well tion efficiency than the venetian defined peaks seen in the figure are blind structure. In addition, the fo caused by external radiations and cused structure is several times can be reduced by placing the photo faster than the venetian-blind type, as multiplier and scintillator in a lead discussed in the section on Principles or steel vault to reduce background of Photomultiplier Design. radiation . ...J z '"z C :J: U I~ 01- z~ ::;)0 Ou 0: eI .... lI:o U co: CD", CD :. ::;) z 662 KeV 1.46 MeV 1.76 MeV PULSE HEIGHT Fig. 66 - Distribution of radiation background pulses as a function of en ergy. A calibration spectrum showing the Cs"l1 photopeak is included for reference. Application of Photomultipliers 83 It is desirable that the background liquid scintillation counting has de of the scintillation counter be as low veloped rapidly. The beta-ray ener as possible. Photomultipliers have gies from these radioisotopes may been developed in which the back vary over a fairly broad range; for ground is kept low by the use of example, H3 emits betas with ener radioactively clean metal cans in- gies varying from zero to a maximum ""'-unnnnul. J. lIe: LlUce: IIlUSl com- U] Lin; .LVI.IVnllJ!> ~'"lu,u.JUlJ. monly used radioisotopes in liquid scintillation counting are tritium, H3; carbon, Cl4; and phosphorus, P32. (98) Because these elements are the most suitable radioactive tracers used in where C is the chance coincidence biological experiments, the field of rate per minute, Nl is the dark-noise * Registered Trade Name for General count rate in counts per minute from Electric Co. material. tube No. I, N2 is the dark-noise count 84 RCA Photomultiplier Manual 105 ..J III Z- Z 104 c :z: u ~ Z ::I 105 "- ." I- Z :. 0 u 10 2 .,.1 U ",- :z: 10 ENERGY (COMPRESSED SCALE) Fig. 67 - Scintillation count showing distribution of beta-ray energies. Fig. 68 - Block diagram of a liquid-scintillation spectrometer. Application of Photomultipliers 85 rate in counts per minute from tube lower than for a single tube having No.2, and 'T is the resolving time a very low background count rate. of the coincidence circuit in seconds. When a vial filled only with a scin In a liquid-scintillation spec tillator is placed between the photo trometer employing two tubes with multipliers, and the output from the dark-noise ratings of 30,000 counts coincidence circuit is examined by per minute each and a coincidence use of a multichannel analyzer, a circuit having a resolving time of 10 pulse-height distribution such as that nanoseconds, the number of acci shown in Fig. 69 is obtained. Clearly, dental coincidences is approximately not many of the background pulses 0.3 count per minute. Scintillations shown are caused by the accidental resulting in at least one photoelectron coincidences of dark-noise pulses being emitted from the photocathode from the photomultipliers, but are of each tube should produce a sig caused by cosmic rays or scintilla nal pulse from the coincidence cir tions in the material of the vial and cuit. Because a scintillation of at photomultiplier envelope resulting least two photons is required to ini from the presence of radioisotopes in tiate a pulse from both photomulti the materials of which they are con pliers, the over-all detection efficiency structed. Light originated internally for both tubes together is slightly in either photomultiplier and trans- ~ 1&1 Z Z et :z: ,0 t~ a:lE Z' 10 jl/) al- Z Cl)ja: ¥o 00 eta.. mo a: 1&1m :lE j 0.1 z 0.01 PULSE HEIGHT (COMPR£SSED SCALE) Fig. 69 - Background distribution obtained from a liquid scintillator using a coincidence system. 86 RCA Photomultiplier Manual mitted to the other also causes back less important than the efficiency in ground pulses; this phenomenon is determining the merit of the system. sometimes referred to as crosstalk. In selecting a photomultiplier tube The counting efficiency E for a for a liquid-scintillation application, liquid scintillation counter is de the following items are of major im fined in per cent as follows: portance: high photocathode quan tum efficiency, low dark-noise count (counting rate from coinci- rate, minimum internal-light genera dence circuit)-(background) x 100 tion, low radioactive-background en E disintegration rate of sample velope, low scintillation-efficiency (99) envelope, fast time response, and good energy resolution. The RCA- The best counting efficiency for 4501 V3 photomultiplier has been a given radioisotope is obtained when specifically designed to meet these a highly efficient scintillator and a requirements. photomultiplier with a high quantum efficiency in the spectral emissivity Cerenkov Radiation Detedion range of the scintillator are used. The optical system containing the Cerenkov radiation is generated two photomultipliers and the count when a charged particle passes ing vial is also of major importance; through a dielectric with a velocity it should be designed so that, as far greater than the velocity of light in as possible, the photons produced in the dielectric (i.e., v greater than a scintillation are equally divided be c/ n, where n is the index of refrac tween the photomultipliers. This di tion of the dielectric). Polarization of vision assures that a coincidence the dielectric by the particle results pulse results from as many scintilla in the development of an electromag tions as possible. In some cases, two netic wave as the dielectric relaxes. or more different isotopes may be If the velocity of the particle is counted simultaneously; therefore, it greater than the velocity of light, is desirable that the photomultipliers constructive interference occurs and have matched gains and good energy a conical wavefront develops, as il (pulse-height) resolution capability to lustrated in Fig. 70. Cerenkov radia provide best isotope separation. tion is the electromagnetic counter A figure of merit E2/b has been part of the shock wave produced in developed to aid in tlte eval\.¥ltion of a gas by an object traveling faster systems used in liquid scintillation than sound. It is highly directional, counting; E is the counting effi and occurs mostly in the near-ultra ciency of the system for a given iso violet part of the electromagnetic tope, and b is the background of spectrum. Because the radiation is the system in an energy range in propagated in the forward direction cluding the energy range of pulses of motion of the charged particle, as from the radioisotope. The figure of shown in Fig. 70, Cerenkov detec merit is, however, not necessarily the tors can be made to detect only those best way of evaluating all systems. particles that enter the system from It is valid at very low counting rates a restricted solid angle. where the background is the domi Cerenkov radiation produced in nant factor, but is not of great help an aqueous solution by beta emitters at high counting rates where the can be useful in radioassay tech background of the system becomes niques because it is unaffected by Application of Photomultipliers 87 , measure time differences such as be , , tween a pair of gamma rays or a , combination of gamma J:ays and par , , ticles in cascade de-exciting some , , level in a nucleus. In time spectros , copy, some special considerations must be made in selecting the scin tillator, photomultipliers, and tech nique of analyzing the signals from the photomultipliers. In a photomultiplier, time resolu tion is proportional to (ii)-1/2, where ii is the average number of photo electrons per event. It is therefore important to choose a scintillator material that provides a high light yield for a given energy of detected radiation. It is also important that DIRECTION OF MOTION OF the variation of the time of interac CHARGED PARTICLE tion of the radiation with the scintil lator be as small as possible. This Fig. 70 - Diagram of the Cerenkov minimum variation is assured by at radiation mechanism. tention to scintillator thickness and source-to-detector geometry. The de chemical quenching and because it cay time constant of the light-emit offers the advantages, over liquid ting states in the scintillator should scintillation counting techniques, of be as short as possible, and the simplified sample preparation and geometry and reflective coatings of the ability to accommodate large the scintillator should be selected so volume samples. Because a fast par that variations in path lengths of ticle is required to produce Cerenkov photons from the scintillator to the radiation, rather high-energy beta photocathode of the photomultiplier rays are required; e.g., the threshold are minimized. The photocathode of for Cerenkov radiation is 261 keY the photomultiplier selected should for electrons in water. Because the have a high quantum efficiency. In photon yield for Cerenkov light is addition, the transit-time dispersion usually very low, the same considera or jitter (variations in the time re tions concerning photomultiplier se quired for electrons leaving the lection apply for the Cerenkov de photocathode to arrive at the anode tection as for liquid-scintillation of the tube) should be small over counting. The tube selected should the entire photocathode area. also be equipped with a faceplate The major contribution to transit capable of good ultraviolet trans time spread occurs in the photo mission. cathode-to-first-dynode region and may be a result of the initial kinetic Time Spectroscopy energies of the emitted photoelec trons and their angle of emission. In addition to the energy spectros Focusing aberrations, and the single copy described above, there are oc electron response or rise time, i.e., casions when it is of advantage to the output-pulse shape at the anode 88 RCA Photomultiplier Manual for a single photoelectron impinging where Vt is the discriminator thresh on the first dynode, may also be of old, and Va is the peak amplitude of some importance. Although the sin the anode-current pulse. The frac gle-electron response theoretically tional pulse height has a consider does not have much effect on time able effect on the time resolution ob resolution, it does change the trigger tained; best results are usually ing threshold at which the best time obtained with F equal to 0.2. resolution can be obtained. In fast zero-crossover timing, the The most commonly used time anode pulse is differentiated twice. spectroscopy techniques include lead This double-differentiation produces ing-edge timing, zero-crossover tim a bipolar output pulse that triggers ing, . and constant-fraction-of-pulse the timing discriminator at the zero height-trigger timing; the technique crossover, the time required to col used depends on the time resolution lect approximately 50 per cent of and counting efficiency required and the total change in the photomulti the range of the pulse heights en plier pUlse. Zero-crossover timing is countered. A block diagram of a second to leading-edge timing for basic time spectrometer is shown in time-resolution work with narrow Fig. 71. pulse-height ranges, but is better Leading-edge timing makes use of than the leading-edge method for a fixed threshold on the anode-cur large pulse-height ranges. rent pulse and provides good time In constant-fraction triggering, the resolution over a narrow range of point on the leading edge of the pulse heights. The fractional pulse anode-current pulse at which lead height F at which the triggering ing-edge timing data indicate that the threshold is set is defined as follows: best time resolution can be obtained is used regardless of the pulse height. (100) For this reason, constant-fraction-of pulse-height timing is the best method TIMING PULSE TIME DETECTOR I PICKOFF --v- DEVICE NUMBER TIME r---- I START :;;.:; ~~ "T" J\cr~, HEIGHT IS PROPOR- SOURCE OF I TIME TO TIONAL TO TIME. TIME COINCIDENT I PULSE HEIGHT I-----,-'lr RADIATION I CONVERTER I STOP * I ___L-_--,-T~;;;-GII r" PULSE TIME "U"DELAY MULTICHANNEL DETECTOR 2 PICKOFF (PULSE HEIGHT) DEVICE ~I ANALYZER - I Fig. 71 - Generalized block diagram of a time spectrometer. Application of Photomultipliers 89 for obtaining optimum time resolu LASER RANGE FINDING tion no matter what the pulse-height range. A simplified block diagram of a laser range finder is shown in Fig. Interpretation of 73. The interference filter is used to Counting Data pass the wavelength of radiant en ergy of the laser with a minimum of A number of techniques can be background radiation. Photomulti used to interpret the data obtained pliers used in laser range finding may from a scintillation counter; one of have a relatively small photocathode, the most popular employs a multi but they must exhibit high quantum channel analyzer. The analyzer sorts efficiency, low dark noise, and ·fast the pulses according to their ampli rise time or an equivalent large tude and records the number of bandwidth. Most photomultipliers pulses in a "channel" whose position can provide bandwidths exceeding is proportional to the amplitude of 100 MHz and at the same time main the pUlse. Many pulses are analyzed tain relatively large output signals. in this manner and a frequency dis The bandwidth of a photomultiplier tribution of pulse heights is obtained. can be limited by th~ RC time con Fig. 72 shows a pulse-height distri stant of the anode cirCuit. bution for a scintillation counter The range of a laser-range-finding counting pulses from a sample of system depends on system parameters CS137. and operational environment. The maximum range of a given system may be signal-photon limited or background limited. The photon limited case exists when the back ground and detector noise can be considered negligible. The maximum range in this case is determined by the signal-to-noise ratio in the photo electron pulse corresponding to the scattered laser return beam. As noted in the section on Statisdall F1uetua. tioD and Noise, the signal-to-noise ratio of such a pulse is proportional to the square root of the product of the number of incident photons on Q. 32 keV Barium x-ray line. b. Backscatter peak. the photomultiplier and the quan c. Compton edge. tum efficiency of the photoelectric d. Photopeak, cesium 137, 0.662 MeV. conversion. If the atmospheric at e. 749 keV photopeak of cesium 134 tenuation is neglected and it is as impurity. f. Continuum due to accidental self sumed that the laser spot falls coincidences producing partial sum entirely within the target, the maxi pulses. mum range in this case would be in g. Accident sum coincidence peak. creased as the square root of the h. Detail due to background radiation. quantum efficiency of the photo cathode. This increase follows be Fig. 72 - Pulse-height distribution for cause the number of photons col Q scintillation counter counting pulses from Q sample of Cs'''. lected by the aperture of the receiving 90 RCA Photomultiplier Manual TRIGGER MODULATOR GENERATOR INTERFERENCE FILTER RANGE, TIME AND COUNTING PHOTOMULTIPLIER CIRCUITS WITH INDICATOR POWER SUPPLY Fig. 73 - Simplified block diagram of a laser range finder. system varies inversely as the square approximation the range is increased of the distance from the target. Re in the background-limited case only cent improvements in the multialkali by the fourth root of the quantum photocathode, resulting in the efficiency. Because the photomulti ERMA cathode, described in the plier current caused by the back section on Photoemission and Sec· ground radiation is proportional to ondary Emission, have increased the the solid angle of the scene from quantum efficiency at 860 nano which the photomultiplier collects meters from 0.01 per cent to 1.5 per radiation, the background current in cent. In terms of laser range in the the photomultiplier may be mini photon-limited case, this quantum mized by the use of a photocathode efficiency improvement implies more having a small area or by the use of than a 10-to-l range increase. The a limiting aperture on the faceplate use of a GaAs photocathode is even of the photomultiplier. No loss in more promising at this wavelength. collected laser light need result be In cases where the laser pulse must cause the return beam may generally be detected against the background be considered as originating from a of a daylight scene, it is said to be point source. The system aperture, background limited. If atmospheric of course, must be large enough to attenuation again is neglected, the avoid optical alignment problems. signal is proportional to the number of incident photons times the quan FLYING·SPOT SCANNING tum efficiency of the photocathode. The number of incident photons An important application for from the return beam is inversely which photomultipliers are particu proportional to the square of the larly well suited is flying-spot scan range, again assuming that the tar ning, a system for generating video get is larger than the laser spot. The signals for television display from a noise, however, is independent of photographic transparency. The ele the range and is determined by the ments of a flying-spot scanning sys square root of the product of the in tem are shown in Fig. 74. A cathode cident background radiation and the ray tube, in conjunction with its quantum efficiency. Thus, to a first power supplies and deflection cir- Application of Photomultipliers 91 cuits, provides a small rapidly mov by the light-absorbing filters before ing light source which forms a raster being focused upon each of three on its face. This raster is focused by photomultipliers, one for 'each color the objective lens in the optical sys channel. The output of each photo tem onto the object being scanned, a mUltiplier is then fed to a separate slide transparency or a motion video amplifier. picture film. The amount of light Flying-spot video-signal genera passing through the film varies with tors are used in the television indus the film density. This modulated light try primarily for viewing slides, test signal is focused upon the photomu patterns, motion-picture film, and tiplier by means of the condensing other fixed images. Systems have lens system; the photomultiplier con been developed for the home-enter verts the radiant-energy signal into tainment industry that allow slides an electrical video signal. The ampli and motion-picture film to be shown fier and its associated equalization on the picture tube of any type of circuits increase the amplitude of the commercial television receiver. video signal as required. The flying-spot scanning system is Cathode-Ray Tube capable of providing high-resolution monochromatic performance.' With Several important considerations the addition of (a) appropriate di must be taken into account if the chroic mirrors which selectively re cathode-ray tube in the system is flect and transmit the red, blue, and to produce a light spot capable of green wavelengths, (b) two addi providing good resolution. The cath tional photomultipliers with video ode-ray tube should be operated with amplifiers, and (c) appropriate fil as small a light spot as possible. The ters, color operation is possible. In cathode-ray-tube faceplate should be color operation, the primary wave as blemish-free as possible and the lengths are filtered after separation tube should employ a fine-grain phos- VERTICAL II HORIZONTAL HORIZONTAL SAWTOOTH GENERATORS OUTPUT II AMPLIFIER VERTICAL OUTPUT AMPLIFIER VERTICAL HORIZONTAL SCANNING SCANNING HIGH VIDEO AMPIFIER WITH -~--~~i~:a:: EQUALIZATION -OBJ;CTIVE LENS 1/ SLIDE '------. CONDENSER LENSES SYNCHRONIZING TO LINE SIGNAL ....~,=="'=.-.! AMPLIFIER GENERATOR Fig. 74 - Elements of a flying-spot scanning system. 92 RCA Photomultiplier Manual poor; blemishes adversely affect sig Obiedive Lens nal-to-noise performance and con tribute to a loss of resolution. The objective lens used in a fly The spectral output of the cath ing-spot generator should be of a ode-ray-tube phosphor should match high-quality enlarger type designed the spectral characteristic of the for low magnification and, depend photomultiplier. This match can be' ing upon the cathode-ray-tube light rather loose in a monochromatic fly output, should be corrected for ultra ing-spot generator; however, the violet radiation. The diameter of the spectral output of the phosphor of objective lens should be adequate to the cathodc;-ray tube used in a three cover the slide to be scanned. An color version must include most of enlarging ff 4.5 lens with a focal the visible spectrum. Phosphors length of 100 millimeters is suit used in monochromatic systems may able for use with 35-millimeter slides. provide outputs in the ultraviolet re gion of the spectrum and still per Photomultiplier Tube form satisfactorily. Phosphors such as P16 and P15, when used with an The spectral characteristics of the appropriate ultraviolet filter, display photomultiplier (or photomultipliers the necessary short persistence re in the case of the three-color sys quired in a monochromatic system. tem) and the cathode-ray-tube phos The visible portion of the P15 phor should match. Usually an S-4 or P24 phosphor is used in color or S-11 spectral response is suitable systems. The P15 and P14 phos for use in a monochromatic system phors are, however, much slower or as the detector for the blue and than the P 16 phosphor and cause a green channels; an S-20 response is lag in buildup and decay of output very often utilized for the red chan from the screen. nel. The bialkaIi cathode (K-Cs-Sb) The phosphor lag results in trail is also well suited for the blue chan ing, a condition in which the per nel. The speed of the detector must sistence of energy output from the be sufficient to provide the desired cathode-ray tube results in a con video bandwidth. Most requirements tinued and spurious input to the do not exceed 6 to 8 MHz, a figure well within photomultiplier capabili photomultiplier as the flying spot ties. moves across the picture being The anode dark current of the scanned. The result is that a light photomultiplier should be small com area may trail into the dark area in pared to the useful signal current. the reproduced picture. The signal-to-noise ratio will be Similarly, the lag in buildup of maximized by operation of the photo screen output causes a dark area to mUltiplier at the highest light levels trail over into the light area. The re possible. If necessary, the over-all sult of these effects on the repro photomultiplier gain should be re duced picture is an appearance duced to prevent excessive anode current and fatigue. similar to that produced by a video The amplitude of the light input signal deficient in high frequencies; is usually a compromise between an consequently, high-frequency equali optimum signal-to-noise ratio and zation is necessary in the video maximum cathode-ray-tube life, amplifier. which may be reduced as a result Application of Photomultipliers 93 of loss of phosphor efficiency at high anode pulses which ongmate from beam-current levels. The signal-to the individual photoelectrons, and noise ratio can also be improved by the digital technique in which indi the selection of photomultipliers hav vidual pulses are counted. ing the highest photocathode sensi tivities possible. However, because Charge-Integration Method the spread of photocathode sensitivi ties is seldom greater than two or In the charge-integration method, three to one, the improvement af either the transit-time spread of the forded by such selection is limited photomultiplier or the time charac to two or three dB, an improvement teristics of the anode circuit cause difficult to detect while observing a the anode pulses to overlap and pro television display but desirable and duce a continuous, though perhaps necessary in some critical applica noisy, anode current. The current is tions. modulated by turning the light off and on by means of some mechanical Video Amplifier device such as a shutter or light "chopper". The signal becomes the Photomultiplier gain need only be difference between the current in the sufficient to provide a signal of the light-on and light-off conditions. required level to the succeeding Detection is limited by noise in video-amplifier stages. These stages, the anode current. At low light levels in addition to providing the necessary the noise is caused primarily by the amplification and bandwidth to as fluctuations in dark current of the sure good picture quality, incorporate photomultiplier (as discussed in the equalization circuits composed of sections on Basic Performance Char networks with different time con acteristics and Statistical Fluctuation stants. The relatively long decay and Noise). The noise may be mini time of these circuits generally re mized by reducing the bandwidth of sults in appreciable reduction of the the measuring system. For example, useful signal-to-noise ratio. There a dc system may be used with a fore, the use of short-persistence bandwidth of only a few hertz if an phosphors is recommended to re appropriate low-level current mctt'r duce the required amount of equali is selected. Bandwidth can also be zation. reduced by some technique of av In addition to the video amplifier, eraging the current fluctuations over a gamma-correction amplifier is re a period of time. quired in each channel. The gamma Another technique of charge inte correction amplifier assures maxi gration is to chop the light signal mum color fidelity by making the with a motor-driven chopper disk linearity or gamma of the system having uniformly spaced holes or unity. slots. The output current is then fed through an amplifier having a narrow LOW-LlGHT-LEVEL DETECTION bandwidth tuned to the frequency of chop. Bandwidths of the order of 1 Systems for the detection of low Hz are typical. light levels make use of two basic The earlier discussions on Equiva techniques: charge integration, in lent Noise Input and on Noise which the output photocurrent is Formulas Relative to Photomulti considered as an integration of the pliers contain the information neces- .. RCA Photomultiplier Manual sary to evaluate the performance of where Np is the number of photons photomultipliers in the charge incident on the photocathode win integration method. dow, and TJA is the quantum effi ciency of the photocathode at the Digital Method photon wavelength including a fac tor for the loss of light by reflection In the digital method for the de and absorption, and a factor for the tection of low light levels a series loss resulting from imperfect elec of output pulses, each corresponding tron-collection efficiency of the front to a photoelectron leaving the photo end of the photomultiplier. cathode of a photomultiplier, appears The dark-noise pulses present in at the anode. All of the output pulses addition to the signal pulses originate from the tube ar~ shaped by a pre mainly from single electrons and amplifier before they enter a pulse have a pulse-height distribution as amplitude discrimination circuit. shown in the simplified dark-noise Only those pulses having amplitudes pulse-height-distribution spectrum of greater than some predetermined Fig. 76. Region A of Fig. 76 in value and having the proper rise-time cludes circuit-originated noise, some characteristics pass through to the signal-processing circuits. The digital (echnique is superior to charge inte gration at very low light levels be cause it eliminates the dc leakage component of the dark current as well as dark-current components originating at places other than the photocathode. Fig. 75 shows a digi tal system in block form. C In the special case in which the digital technique is used to count single photons incident upon the I ELECTRON photocathode of a photomultiplier, LOWER LUPPER DISCRIMI NATOR DISCRIMINATOR signal pulses appear at the anode LEVEL LEVEL with an average pulse amplitude PH PULSE HEIGHT equal to em, where e is the electron charge and ffi is the photomultiplier Fig. 76 - Dark-noise pulse-height-dis- gain. The number of signal pulses tribution spectrum. Na arriving at the anode is given by single-electron pulses, and pulses caused by electrons originating at the dynodes in the multiplier section. PREAMPLIFIER ADJUSTABLE TO lPHOTOMULTIPLIER I PULSE AND I- PULSE AMPLITUDE PROCESSING SHAPER DISCRIMINATOR - CIRCUITS Fig. 75 - Block diagram illustrating a digital system for detection of low light levels. Application of Photomultipliers 95 Pulses originating at the dynodes ([he section on Statistical Fluctua a further exhibit less gain than the single-elec tion and Noise contains of photon tron pulses from the cathode. Region discussion of the statistics B represents the single-electron counting.) may be pulse-height distribution and is the The signal-to-noise ratio of the region in which the single-photon improved by the square root the time signal pulses appear. Region C is time of count. Increasing decreasing caused by cosmic-ray muons, after of count is analogous to pulsing, and radioactive contami the bandwidth in the charge-integra nants in the tube materials and in the tion method. The signal-to-noise by a vicinity of the tube. To maximize ratio can also be improved decreasing the the ratio of signal pulses to noise square-root factor by pulses. Be pulses in single-photon counting, number of dark-noise at the lower- and upper-level discriminators cause these pulses originate the number should be located as shown in the photocathode surface, the area figure. can be reduced by reducing by reducing If the rate of photon arrival is Ip, of the photocathode or by the net quantum efficiency is 'Y/, and the effective photocathode area to image only the counting interval is T, the num using electron optics photocathode on ber of signal pulses is Ip'Y/T. The rate a small part of the also be de of single-electron dark pulses is re the first dynode. It may photomultiplier ferred to as I . At the limit of detec sirable to cool the d thermionic tion the number of dark pulses will and thereby reduce the photocathode (as greatly exceed the pulses which are emission from the section on Basic photon originated. Two counts must discussed in the be made, one with the light on and Performance Characteristics). one with it off. The difference in the number of counts is the signal, Very-Low-Light-Level Ip'Y/T, in numbers of electrons for Photon-Counting Technique the count interval. The variance for very-low-light this measurement is given by Before beginning level photon counting, the following special precautions must be taken: 1. The power supply and intercon necting circuits must have low-noise characteristics. The reason for the double entry of 2. The optical system must be the variance for the dark counts IdT carefully designed to minimize pho is that the determination involves the ton loss and to prevent movement subtraction of the light-on and light of the image of the object on the off counts. If it is assumed that the photocathode, a possible cause of number of dark pulses greatly ex error if the cathode is non-uniform. ceeds the light-originated pulses, the It is generally good practice to de signal-to-noise ratio is given by the focus the image on the photocathode, following expression: especially in the case of a point source, to minimize problems which may result from a non-uniform Ip 1/ Tl photocathode. SNR = [2I li (103) must be d 3. The photomultiplier allowed to stabilize before photon 96 RCA Photomultiplier Manual counting is begun. The tube should of considerable significance in pho not be exposed to ultraviolet radia ton counting. These developments tion before use and should, if pos fall into two major categories, sec sible, be operated for 24 hours at ondary-emission materials and photo the desired voltage before the data cathodes. Because of the superior are taken. statistics of gallium phosphide (one 4. The photomultiplier should be of the more recently developed sec operated with the cathode at ground ondary-emission materials discussed potential if possible. If the tube is in the section on Photoemission and operated with the photocathode at Secondary Emission), a tight single negative high voltage, care must be electron distribution curve can be taken to prevent the glass envelope obtained, as shown in Fig. 77. The of the tube from coming into con tight distribution permits easy loca tact with conductors at ground po tion of the pulse-height discriminator, tential or noisy insulators such as as is particularly evident when some bakelite or felt. Without this pre caution, a very high dark noise may result as well as permanent damage to the photocathode. 5. Large thermal gradients must not be permitted across the tube as it is cooling. In addition, care must be taken to avoid excessive condensa tion across the leads of the tube or on the faceplate. Photomultiplier Seledion 01234567 for Photon Counting PULSE HEIGHT -PHOTOELECTRON EQUI VALENTS In the selection of a photomulti Fig. 77 - Single-electron distribution plier for use in photon counting, sev curve taken with a photomultiplier hav eral important parameters must be ing a gallium phosphide first dynode. considered. First, and most impor tant, the quantum efficiency of the of the signal pulses are originated photocathode should be as high as by two or more photoelectrons leav possible at the desired wavelength. ing the cathode simultaneously. If To minimize the thermionic dark the average number of photoelectrons noise emission, the photocathode area leaving the photocathode per pulse should be no larger than necessary were three, a pulse-height distribu for signal collection; the multiplier tion similar to that shown in Fig. structure should utilize as large a 78 would be obtained. Gallium phos fraction as possible of the electrons phide provides a higher signal-to from the photocathode. The over-all noise ratio. than would conventional tube should have as low a dark noise secondary-emitting materials such as as possible. In some applications, the BeO or CsaSb when used in the same rise time and time-resolution capa tube. bilities of the tube may also be im Photocathode developments in portant. clude ERMA (Extended Red Multi A number of more recent develop Alkali), a semitransparent photo ments in photomultiplier design are cathode having a response to 900 Application of Photomultipliers 97 tion of wavelength and require pho tomultipliers having a broad spectral range. The results of measurements of the absorption characteristics of substances are frequently expressed in terms of optical density. This logarithmic method correlates with the way the human eye discriminates differences in brightness. The transmission density D is de fined by I 2 PULSE HEIGHT / No. OF ELECTRONS Po 1 () D = IOglO P = IOglO T 104 Fig. 78 - Pulse-height distribution ob t tained with a photomultiplier having a gallium phosphide first dynode. Light level is such that three photoelectrons where Po is the radiant flux incident per pulse time is the most probable number. upon a sample, P t is the radiant flux transmitted by a sample, and T is nanometers, and several negative the transmission figure or P tl Po. electron-affinity materials. Perhaps Density measurements are useful in the most significant of these ma various applications to films and terials is indium gallium arsenide, other transparencies in addition to whose threshold wavelength in chemical analysis where concentra creases with increasing indium con tion of a solution is studied as a tent. Spectral-response curves for function of wavelength. some of the more recent photo cathodes are shown in the section on Color-Balancing Photometry Basic Performance Characteristics. A color-balancing photometer is PHOTOMETRIC APPLICATIONS used to determine color balance and exposure times necessary to produce Photometry is concerned with the photographic color prints from color measurement of luminous intensity negatives. Such a device is shown in and other parameters related to the block diagram form in Fig. 79; it radiant energy which produces visual allows the matching of the relative sensation. Photometric units and con proportions of red, blue, and green version from radiant power are dis light transmitted by a production cussed in a later section of this negative to those of a master color Manual. Photometric measurements negative. As the first step in the may be made either by visual com matching process, the master nega parison or by more accurate photo tive is used to make an acceptable electric methods. This section dis print. This first print is produced cusses some of the applications of through trial and error by measure photomultipliers in photometry. ment of the relative proportions of red, blue, and green light transmitted Spectrophotometry through a key area of the master negative; the key area usually con Spectrophotometers measure the sists of a flesh tone or a gray area. optical density of materials as a func- The amount of light transmitted is 98 RCA Photomultiplier Manual DENSITY ITIME t---+-o+ REGULATED HIGH VOLTAGE POWER SUPPLY CALIBRATION NEGATIVE " COLOR 7 0 FILTERS"''; "6J "" E>=== Fig. 79 - Block diagram of a color-balancing photometer. a measure of the density. The ex as those obtained with the master posure time using white light and color negative. the lens opening used to obtain the Most color-balancing photometers satisfactory print are noted. employ steady light sources and Next, an area of the production handle small-amplitude signals. Con negative similar to that on the mas sequently, the photomultiplier used ter negative is chosen and the rela must have low values of dark current tive proportions of the red, blue, and and good stability. The life expect green light transmitted are measured ancy of tubes used in this application again. By use of magenta and yel is long because the small signal levels low correcting filters, the color trans reduce the effects of fatigue which mitted by the production negative might otherwise adversely affect can be balanced with that of the measurement accuracy and repeat master negative. When the produc ability. Low-current operation also tion negative is used, the lens open provides for linear operation where ing is adjusted so that the exposure the anode current is proportional to time with white light is the same as the input flux over the range of trans it was for exposure of the master mission values measured. The in negative. When the values of color tensity range of a color-balancing correcting filters and exposure times photometer, given in terms of the thus determined are used, prints from ratio of the radiant flux incident upon the production negative can be ob a sample to the radiant flux trans tained which are very nearly as good mitted by a sample, is usually of the Application of Photomultipliers 99 order of 1000 to 1 or more. It must Logarithmic Photometry be remembered that as in most photomultiplier applications, the In the measurement of absorption power supplies used must be capable characteristics, the changes in bright of providing voltages sufficiently ness levels vary over such a large regulated and free from ripple to range that it is advantageous to use assure minimum variation in sensi a photometer whose response is ap tivity with possible line-voltage varia proximately logarithmic. This re tion. sponse enables the photometer to be equipped with a meter or' read-out Densitometry scale that is linear and provides pre cise readings even at high optical A second example of the use of densities. the equivalent photoelectric method A simplified circuit of a logarith in photometry measurements is in mic photometer capable of measur densitometry, or the measurement of ing film density with high sensitivity diffuse transmission density. and stability and of providing an Although techniques such as those appropriate logarithmic electrical re employed in· the color-balancing sponse and linear meter indication of photometer may be used successfully density over three or four density to measure density over an intensity ranges is shown in Fig. 80. range of 10 or 100 to 1 (density 1 The circuit of Fig. 80, in which to density 2), their use becomes in the photomultiplier operates at a creasingly difficult as the range is constant current, minimizes photo increased to 1000 to 10,000 to 1 multiplier fatigue and eliminates the (density 3 to density 4) or more. The need for a regulated high-voltage large dynamic ranges encountered in supply. The feedback circuit illus color-film processing place severe re trated develops a signal across Rl quirements upon the photomultiplier proportional to the anode current. because it must be operated in a This signal controls the bias applied constant-voltage mode. Problems de to the control tube and automatically velop in this mode at high-density adjusts the current in the voltage values at which the dark current may divider network. By this means, the become a significant proportion of dynode voltage is maintained at a the signal current. The random na level such that the anode current is ture of this dark current, which pre held constant at a value selected to cludes its being "zeroed out", may minimize the effects of fatigue: At lead to output-signal instability. As optical densities of 3, the dynode a result of the type of operation supply voltage may be 1000 volts; needed to produce the dynamic range at optical densities of less than 1, required in density measurements, the dynode voltage may be as low the photomultiplier anode current is as 300 volts. Dynode voltage is trans high at low density values. These lated into density by means of the high currents may result in excessive scale on voltmeter VI. fatigue, and, depending upon the Because there is an approxi operating point, perhaps non-linear mately exponential variation of sensi operation. As with the color-balanc tivity of a photomultiplier with ap ing photometer, a well regulated low plied voltage (as discussed in the ripple power supply is needed to section on Basic Performance Char assure accurate measurements. acteristics) in a constant-current 100 RCA Photomultiplier Manual RI + - REGULATED + r------, HIGH VOLTAGE I I POWER I I SUPPLY I I 1 I L..7---. I REG. LOW- : I VSOuL;~~i I I I I IL ______.J COMPENSATING CIRCUIT "~""~ COCOR Flm,,, ~ Fig. 80 - Block diagram of a logarithmic photometer. mode, the voltage becomes a logarith Association standards for the de mic function of the input light flux termination of diffuse density. and thus a linear measure of density. In photometric equipment, the re SPECTROMETRIC lationship between dynode voltage APPLICATIONS and the logarithm of the incident radiant flux Po is not quite linear Spectrometry, the science of spec and, consequently, some compensa trum analysis, applies the methods of tion is required. This compensation physics and physical chemistry to is provided in the circuit of Fig. 80 chemical analysis. Spectrometric ap by the incorporation of a variable plications include absorption, emis and automatic shunt across the volt sion, Raman, solar, and vacuum meter VI. As the density values in spectrometry, and fluorometry. The crease, the effective value of the uses made of photomultipliers in shunt resistance is reduced. This cir these applications are described in cuit not only compensates for photo the following paragraphs. multiplier non-linearity, but also for the non-linearity of the optical sys Absorption Spedrometry tem employed so that the measure ments obtained through its use will Absorption spectrometry, used to conform to American Standards detect radiant energy in the visible, Application of Photomultipliers 101 ultraviolet, and infrared ranges, is ing a gallium arsenide photoemitter one of the most important of the in can now be used to detect radiant strumental methods of chemical energy from 940 nanometers to the analysis. It has gained this import cut-off point of the photomultiplier ance largely as a result of the de window in the near-ultraviolet. velopment of equipment employing photomultipliers as detectors. The Raman Spectroscopy principle underlying absorption spec trometry is the spectrally selective There are two types of molecular absorption of radiant energy by a scattering of light, Rayleigh and substance. The measurement of the Raman. Rayleigh scattering is the amount of absorption aids the scien elastic collision of photons with tist in determining the amount of the molecules of a homogeneous various substances contained in a medium. Because the scattered pho sample. tons do not gain energy from or lose The essential components of an energy to the molecule, they have absorption spectrometer are a source the same energy hvo as the incident of radiant energy, a monochromator photons. A classic example of Ray for isolating the desired spectral leigh scattering of light from gas band, a sample chamber, a detector molecules is the scattering of the for converting the radiant energy to sun's light rays as they pass through electrical energy, and a meter to the earth's atmosphere; the scattering measure the electrical energy. The accounts for the brightness and the spectral ranges of the source and de blueness of the sky. tector must be appropriate to the Raman scattering is the inelastic range in which measurements are to collision of photons with molecules be made; in some cases this range that produces scattered photons of may include one wavelength; in higher or lower energy than the ini others, it may scan all wavelengths tial photons. During the collision between the near-ultraviolet and the there is a quantized exchange of en near -infrared. ergy which, depending on the state The side-on photomultiplier has, of the molecules, determines whether in general, been the most widely used the initial photon gains or loses en in spectrometric applications. The ergy. The differences in energy levels side-on tube is relatively small and are characteristic of the molecule. has a rectangularly shaped photo If the excitation frequency is 111 and cathode that is compatible with the the photon emitted after the inter shape of the light beam from an exit action has a frequency of 112' the slit. Spectrometric applications re interaction results in a change in the quire tubes with good stability, high energy of an initial photon by hllo, anode sensitivity, low dark current, where hllo = Ihll2 - hill I· When hll2 and broad spectral sensitivity. is greater than hill> the initial pho Before the development of the new ton has gained energy from the mole negative-electron-affinity type photo cule that was in the excited state; emitters, such as gallium arsenide, in the reverse case, the initial photon it was necessary to use more than has given up energy to the molecule one detector in the measurement of in the unexcited state. radiant energy from the near-ultra Early Raman instruments had a violet to the near-infrared part of number of disadvantages and were the spectrum. A photomultiplier hav- difficult to use; it was difficult to 102 RCA Photomultiplier Manual find a stable high-intensity light The scattered photons from a source and to discriminate against Raman interaction are so few in Rayleigh scattering of the exciting number that only the highest-quality line. With the recent development of photomultipliers can be used as de high-quality monochromators and tectors. The tube should have high the advent of the laser light source, collection efficiency, high gain, good a renewed interest in the Raman ef multiplication statistics, low noise, fect as an analytical method 'of and high quantum efficiency over the chemical analysis has taken place. spectral range of interest. Fig. 81 Raman spectrophotometers are gen shows a Raman spectrum. To re erally used to investigate the struc duce the effects of Rayleigh scatter ture of molecules and to supplement ing, a source of noise in Raman other methods of chemical analysis, spectroscopy, the photomultiplier is particularly infrared-absorption spec placed at right angles to the light trometry. beam, whose wavelength has been chosen as long as possible within the range of interest. PENTENE-2-RAMAN SPECTRUM WAVElEN~TH(em)= WAVENUMBER (em-I) 80 I- 60 ~ Q. I- o~ w ~ 40 20 500 Fig. 81 - Typical Raman spectrum. 103 Voltage-Divider Considerations HE interstage voltage gradients RESISTANCE VOLTAGE DIVIDER for the photomultiplier elements T photo may be supplied by individual volt The response of a 931A volt age sources; however, the usual multiplier using a conventional Fig. 82, source is a resistive voltage divider age divider, as shown in is shown placed across a high-voltage source, with equal voltage per stage the light as shown in Fig. 82. in Fig. 83 as a function of flux input.1 (The value of RL was essentially zero for this measure ment.) The anode current is shown + --_--'V\f\r.., current at zero p relative to the divider P -Anode Rn+1 RL -AnodaL.... TYPE 931A Rlliston 1.6 SUPPLY VOLTS -1000 DY n DY n - Dynodo S.... Rn - Vol.... Divider 1.4 A Rn Rlliston / n - T 0,"1 Number DYn-1 of DynodoS_ K - CodIode Rn-I DYn-2 Fig. 83 - The relative resporue of a 931A photomultiplier as a function of the light flux using the circuit of Fig. 82 with equal stage voltage (at zero light level). Curve is taken from Fig. 2 of Ref. 1. K light level. The superlinearity region is explained by the fact that the in crease in dynode voltage resulting Fig. 82 - Schematic diagram of a re from the redistribution of the voltage listjve voltage divider. loss across the anode-to-last-dynode 104 RCA Photomultiplier Manual divider resistor tends to reduce anode The following formula may be collection efficiency. The decrease in used to calculate the voltage be sensitivity which occurs beyond re tween stages: gion A of Fig. 83 results from the extension of voltage losses to the last two or three dynode resistors caus ing defocusing and skippi,ng in the associated dynode stages. In order n+l to prevent this loss and assure a high where rt I rio and Vj is the degree of linearity, the current 1=1 through the voltage-divider network should be at least ten times the maxi voltage between the elements K - mum average anode current required DYb DYI - DY2' and DYn - Anode. to prevent the dynode voltages from varying. In calculating the voltage VOLT AGE-DIVIDER RESISTORS divider current, the average anode current must first be estimated; this The voltage-divider resistor values estimate requires knowledge of the required for each stage can be de value of the input (light) signal and termined from the value of the total the required output (electrical) signal. resistance required of the voltage divider and the voltage-divider ratios VOLTAGE RATIO of the particular tube type. The inter stage resistance values are in propor A resistance voltage divider must tion to the voltage-divider ratios as be designed to divide the applied follows: voltage equally or unequally among the various stages as required by the electrostatic system of the tube. The most common voltage between stages is usually referred to as the stage voltage and the voltage between other where Rj is the resistance between stages as relative to the most com elements DYj and DYj_l. The rec mon stage voltage. ommended resistance values for The voltage distribution is nor a photomultiplier voltage divider mally specified in the following way: range from 20,000 ohms per stage to 5 megohms per stage; the exact values are usually the result of a Relative compromise. If low values of resist Between Voltage ance per stage are utilized, the power K-DYl r O ) drawn from the regulated power sup DYl - DY2 r(2) ply may be excessively large. The DYn_l - DYn r(n) resistor power rating should be at DYn - Anode r(n+l) least twice the calculated rating to provide a safety margin and to pre where n represents the number of vent a shift in resistance values as the dynode stage and r (J) G = 1 to a result of overheating. n+ 1) represents the relative inter Photomutliplier noise or a shift in stage voltage. There is always one gain may result from heat emanating more interstage voltage than there from the voltage-divider resistors; are dynode stages. therefore, the divider network and Voltage-Divider Considerations 105 have the other heat-producing components rangement, however, may the gain should be located so that they will disadvantage of reducing disad not increase the tube temperature. in the critical first stage. This using a Resistance values in excess of five vantage can be avoided by volt megohms should be avoided because zener diode to hold this critical current leakage between the photo age constant. multiplier terminals could cause a variation of the interstage voltage. INTERMEDIATE STAGES The type of resistor used in a di it is de vider depends on the dynode struc In applications in which anode sensi ture with which the divider will be sirable to control the the over-all used. Close tolerance resistors, such tivity without changing a single dy as the wire-wound types, are nor voltage, the voltage of 84 shows mally required with the focused or in node may be varied. Fig. current for line structures. On the other hand, the variation of anode voltages in the venetian a 931A photomultiplier when one of interdynode while blind structures can vary widely with the dynode voltages is varied is held con but little effect on the photomulti the total supply voltage be selected plier. For this reason, the resistors stant. The dynode should dynode string used with a venetian-blind structure from the middle of the dynode poten need only be of a less stable variety, because a variation of or anode such as carbon. tials near the cathode would have a detrimental effect on CATHODE-TO-FIRST -DYNODE photomultiplier operation. REGION The over-all performance of a tube can be improved by maintain ing a high electric field in the cath ode-to-first-dynode region to reduce of photoelec the transit-time spread I Z trons arriving at the first dynode and W Ir the effect of magnetic fields. Ir minimize ~ first-dynode gain, which im U A high W o plies a high cathode-to-first-dynode o voltage, is particularly important in z ..w the detection of extremely low light > levels where the statistics of photo ~ .J become important. W emission Ir A high cathode-to-first-dynode voltage of approximately 600 to 800 volts is recommended in photomul tipliers having a gallium phosphide first dynode. Otherwise, the higher secondary-emission capability of this O~O~O~~~~60~O~~-L7~O-O~~~OCO material and its improved statistics DYNODE NO.6-TO·CATHODE VOLTS are not utilized. variation to control Fig. 84 - The output-current It is frequently useful of a 931 A when the voltage on one the gain of the photomutiplier by dynode (No.6) is varied while the total varying the over-all voltage. This ar- anode voltage remains fixed. 106 RCA Photomultiplier Manual VOLTAGE DIVIDERS FOR at which the peak dynode current is PULSED OPERATION less than 1110 of the average cur rent through the voltage-divider net In applications in which the input work. signal is in the form of pulses, the average anode current can be de termined from the average pulse am plitude and the duty factor. The MEDIUM-SPEED total resistance of the voltage-divider PULSE APPLICATIONS network is calculated for the average anode current for dc operation. In applications in which the out In cases in which the average put current consists of pulses of short anode current is orders of magnitude duration, the capacitance CL of the less than a peak pulse current, dy anode circuit to ground becomes node potentials can be maintained very important. The capacitance CL at a nearly constant value during is the sum of all capacitances from pulse durations of 100 nanoseconds the anode to ground: photomulti or less by use of charge-storage ca plier-anode capacitance, cable ca pacitors at the tube socket. The pacitance, and the input capacitance voltage-divider current need only be of the measuring device. For pulses sufficient to provide the average with a duration much shorter than anode current for the photomulti the anode time constant RLCL, the plier; the high peak currents required output voltage is equal to the product during the large-amplitude light of the charge and 11 CL because the pulses may be supplied by the ca anode current is simply charging a pacitors. capacitor. The capacitor charge then I The capacitor values depend upon decays exponentially through the the value of the output charge asso anode load resistor with a time con ciated with the pulse or train of stant of RLCL. To prevent pulses pUlses. The value of the final-dynode from piling up on each other, the to-anode capacitor C is given by maximum value of RLCL should be much less than the reciprocal of the repetition rate. C 100 ~ An important example of this type = V of operation is scintillation counting, for example with NaI(Tl). The time constant of the scintillations is 0.25 where C is in farads, q is the total microsecond and, because the inte anode charge per pulse in coulombs, gral of the output current pulse is and V is the voltage across the ca a measure of the energy of the inci pacitor. The factor 100 is used to dent radiation, the current pulse is limit the voltage change across the integrated on the anode-circuit ca capacitor to a 1-per-cent maximum pacitance for a period of about 1 during a pulse. Capacitor values for microsecond. To prevent pulse pile preceding stages should take into ac up, the anode-circuit time constant count the smaller values of dynode is selected to be about 10 micro currents in these stages. Conserva seconds. Because the scintillations oc tively, a factor of approximately 2 cur at random, the maximum average per stage is used. Capacitors are not counting rate is limited to about required across those dynode stages 10 kHz. Voltage-Divider Considerations 107 FAST-PULSE APPLICATIONS on a 50-ohm source and a 50-ohm load impedance; careful wiring tech In fast-pulse light applications, it nique is required to assure the best is recommended that the photomulti impedance match. A good impedance plier be operated at negative high match to the transmission line and voltage with the anode at ground po anode of the tube is provided by tential. A typical voltage-divider cir capacitively coupling the last, and cuit with series-connected capacitors sometimes the next to last, dynode used with a positive ground is shown to the ground braid of the 50-ohm in Fig. 85. The parallel configura coaxial signal cable. tion of capacitors may also be used, DYI DY2 DYn-3 DY n-2 DYn-1 DYn ~2THODE~-HIGH VOLTAGE 3R R __ _ Fig. 85 - Series-connected capacitors in voltage-divider circuit using positive ground for pulse-light applications. as shown in Fig. 86. This circuit per- Fig. 87 illustrates the use of by mits lower-inductance connections to pass capacitors (CBP) on the anode the ground plane. The parallel ar- signal line. Because of the bypass rangement requires capacitors of capacitors, the transmission-line ele higher voltage ratings. Regardless of ments in the figure are viewed by a the configuration, the capacitors must fast pulse as a single line between be located at the socket. The capaci- two ground planes. tor arrangements just described may also be applied to negative-ground WIRING TECHNIQUES applications. The wiring of the anode is very Good high-frequency wiring tech critical in pulse applications if line niques must be employed in wiring arity and pulse shape are to be pre photomultiplier sockets and asso served. Most pulse circuitry is based ciated voltage dividers if loss of time DYI DY2 DYn-3 DYn-2 DYn-1 JRTHODE~-HIGH VOLTAGE 3R R __ _ R R R R ~'V\/~~'VV~~~~~ = l·005I'~OO5JlFl·oO'I'F --= -=- ~ Fig. 86 - Parallel-connected capacitors in voltage-divider circuit for pulsed-light application. 108 RCA Photomultiplier Manual I TIME-n. S TIME-n. o 5 Fig. 87 - Bypass capacitors used to make the last two dynodes appear as a ground plane to a fast-pulsed signal. I Z III It: It: resolution in pulse operation is to ~ be minimized. Fig. 88(a) illustrates the pulse shape obtained from an 8575 photomultiplier excited by a 0.5-nanosecond rise-time light pulse. The pulse shape is most easily seen Fig. 88 - (a) Pulse shape obtained from with the aid of a repetitive light photomultiplier excited by a O.5-nano pulser and a sampling oscilloscope. second-rise-time light pulse, and (b) Fig. 88(b) indicates the general type distortion encountered with improper charge-storage capacitors and ringing of distortion encountered with the in an improperly wired socket. use of improper charge-storage ca pacitors, capacitors either too small in value or too inductive. The output pulse appears clipped in comparison the bypass capacitors and coaxial to that of Fig. 88(a) and illustrates cable. The resistors for the voltage the ringing that may occur in an im divider are shown mounted on the properly wired socket, a condition socket. In applications requmng that worsens as pulse amplitude in minimum dark current, the resistors creases. should be remote from the photo Fig. 89 shows a socket wired for multiplier to minimize heating ef a negative high-voltage application. fects. In negative-high-voltage ap The disk-type bypass capacitors are plications, the socket should be mounted in series with minimum lead mounted in a copper or brass co length because the self-inductance of axial cylinder to provide a suitable a few millimeters of wire becomes ground plane to which the capacitors critical in nanosecond-pulse work; can be connected; this construction care should also be taken in dressing is shown in Fig. 90. Voltage-Divider Considerations 109 II Ii Ii pulse amplitudes. Output-pulse am II II II plitude may be increased by increas ill! II ing the tube voltage and may be DYn--O measured with a multichannel ana lyzer or an oscilloscope. The effectiveness of the charge storage capacitors can be verified with a simple linearity test. The out put of a photomultiplier varies as SHORT LENGTH RG!S8V LEADING TO BNC OR Va, where V is the applied voltage GR847 CONNECTOR and alpha (a) is a constant; thus a Fig. 89 - Anode detail of socket wired plot of pulse amplitude as a function for a negative high-voltage application. of the logarithm of the voltage value should yield a straight line. A two-pulse technique may be em PHOTOMULTIPLIER SOCKET MACHINED ployed to test linearity at constant TO FIT TUBING operating voltage. This technique re quires a means for attenuating the light signals by a factor of 0.5 to 0.1; a filter in the light beam suf COPPER fices. In applying the technique, the OR BRASS ratio of the amplitude of the two SHIELD TUBING pulses is established at low pulse amplitudes where linear operation is assured. Next, the pulse amplitude CHARGE STORAGE is increased, either by increasing the CAPACITORS CONNECTED tube voltage or by increasing the TO GROUND CYLINDER light intensity; a constant pulse-am Fig. 90 - Tube socket mounted in plitude ratio indicates linear opera copper or brass coaxial cylinder for tion. negative-high-voltage applications. TAPERED DIVIDERS CHECKING SOCKET AND Some applications require that TUBE PERFORMANCE photomultipliers sustain high signal currents for short time intervals, tens Some photomultipliers typically of nanoseconds or less. In general, display a reflected-pulse rise time of photomultipliers are capable of sup the order of 1.5 nanoseconds for an plying 0.2 ampere or more into a initiating pulse of the order of 50 50-ohm load for short durations. picoseconds when the tube and socket However, the voltage divider must are tested with a time-domain re be tailored to the application to al flectometer, the instrument used to low a photomultiplier to deliver test the anode-pin region of the these high currents. socket. The socket should simulate The principal limitation on current an open circuit. If the output pulse output (into a 50-ohm system) is can be viewed with a sampling oscil space charge at the last few stages. loscope, the pulse shape can be in This space charge can be overcome spected for signs of clipping or ring if the potential difference across the ing, effects more prominent at high last few stages is increased by use 110 RCA Photomultiplier Manual of a tapered divider rather than an ing). If the potential difference is equal-volts-per-stage divider. The reduced from, for example, 100 volts tapered divider places 3 to 4 times to 10 volts, several orders of magni the normal interstage potential dif tude can be compressed into a sin ference across the last stage. The gle decade. The exact relationship progression leading to the 4-times between input-light intensity and the potential difference should be grad compressed output voltage developed ual to maintain proper electrostatic across a load resistor varies among focus between stages; a progression tube types and voltage dividers, and of 1, 1 ... 1, 1, 1.5, 2.0, 2.5, 3 is must be determined empirically. recommended. Anode signals 400 to 500 volts in CURRENT PROTECTION amplitude can be obtained from OF PHOTOMULTIPLIER photomultipliers. These large voltage excursions are often useful, for ex If a photomultiplier is accidentally ample, in driving electro-optical exposed to an excessive amount of modulators. Because the photomulti light, it may be permanently dam plier is nearly an ideal constant aged by the resultant high currents. current source, its output voltage To reduce this possibility, the re into a high-impedance load is limited sistive voltage-divider network may only by the potential difference be be designed to limit the anode cur tween the last dynode and the anode. rent. The average anode current of If the potential difference is 100 a photomultiplier cannot much ex volts, the anode cannot swing ceed the voltage-divider current; through more than a l00-volt excur therefore, the voltage-divider net sion. By impressing a much higher work serves as an overload protection potential difference between the for the tube. If overexposure is ex anode and last dynode by means of pected frequently, interdynode cur a tapered divider, greater voltage rents which can be quite excessive, swings can be obtained. may cause loss of gain. In some ap plications it may be worth while to DYNAMIC COMPRESSION OF protect against dynode damage by OUTPUT SIGNAL using resistors in series with each dynode lead. 1 Most photomultipliers operate linearly over a dynamic range of HIGH-VOLTAGE SUPPLY six or seven orders of magnitude, a range few monitor devices can ac The recommended polarity of the commodate without requiring range photomultiplier power-supply volt changes. When compression of the age with respect to ground depends dynamic range is desired, a logarith largely on the application intended. mic amplifier is sometimes used. The Of course, the cathode is always photomultiplier may be operated in negative with respect to the anode; a compressed-output mode, how however, in some pulsed applications, ever, without the need for additional such as scintillation counting, the compression circuitry. cathode should be grounded and the The voltage excursion of the anode anode operated at a high positive is limited by the potential difference potential with a capacitance-coupled between the last dynode and the output. The scintillator and any mag anode (with no signal current flow- netic or electrostatic shields should Voltage-Divider Considerations 111 also be connected to ground poten tial to eliminate leakage currents TUBE TEMPERATURE =22·C from ground to the photocathode and to prevent the shock hazard caused by the scintillator being at high volt age. In applications in which the signal cannot be passed through a coupling capacitor, the positive side of the ~L-~~~-L-L-L~~~4 power supply should be grounded. -1000 -800 -600, -400 -200 0 The cathode is then at a high nega EXTERNAL SHIELD VOLTS tive potential with respect to ground. RELATIVE TO ANODE VOLTS When this arrangement is used, extra Fig. 91 - Curve of effect of external precautions must be taken in the shield potential on noise in a mounting and shielding of the photo photomultiplier. multiplier. If a potential gradient POWER-SUPPLY STABILITY exists across the tube wall, scintilla tions occurring in the glass will in The output signal of a photomulti crease dark noise. If this condition plier is extremely sensitive to varia exists for a sufficiently long period tions in the supply voltage. There of time, the photocathode will be fore, the power supply should be very permanently damaged by the ionic stable and exhibit minimum ripple. conduction through the glass. To As an approximate "rule of thumb", prevent this situation when a shield power-supply stability (in per-cent is used, the shield is connected to variation) should be at least ten photocathode potential. Light-shield times less than the maximum output ing or supporting materials used in current variation (in per cent) that photomultipliers must limit leakage can be tolerated. currents to 10-12 ampere or less. Fig. 91 shows a curve of the effect REFERENCES of external-shield potential on photo multiplier noise. 1. R. W. Engstrom and E. Fischer, To reduce the shock hazard to "Effects of Voltage-Divider Char personnel, a very high resistance acteristics on Multiplier Photo should be connected between the tube Response", Rev. Sci. Instr., shield and the negative high voltage. Vol. 28, p. 525 (1957). 112 Photometric Units and Photometric-to-Radiant Conversion HOTOMETRY is concerned with the faint object. The response of the Pthe measurement of light; the light-adapted eye (cone vision) is re origins of the photoelectric industry ferred to as the Photopic eye re were associated with the visible spec sponse; the response of the dark trum. It was appropriate, therefore, adapted eye (rod vision) is referred that the units for evaluating photo to as Scotopic eye response. sensitive devices be photometric. To Although characteristics of the day, however, although many of the huma~ eye vary from person to per applications of photosensitive de son, standard luminosity coefficients vices are for radiation outside the for the eye were defined by the Com visible spectrum, for many purposes mission Internationale d'Eclairage the photometric units are still re (International Commission on Il tained. Because these units are based lumination) in 1931. These standard on the characteristics of the eye, it C.LE. luminosity coefficients, listed is appropriate that the introduction in Table II, refer to the Photopic to this discussion begin with a con eye response. They represent the sideration of some of the character relative luminous equivalents of an istics of the eye. equal-energy spectrum for each wavelength in the visible range, as CHARACTERISTICS OF suming foveal vision. An absolute THE EYE "sensitivity" figure established for the standard eye relates photometric The sensors in the retina of the units and radiant power units. At 555 human eye are of two kinds: "cones" nanometers, the wavelength of maxi which predominate the central (or mum sensitivity of the eye, one watt foveal) vision and "rods" which pro of radiant power corresponds to 680 vide peripheral vision. The cones are lumens. responsible for our color vision; rods For the dark-adapted eye, the peak provide no color information but in sensitivity increases and is shifted the dark-adapted state are more sen toward the violet end of the spectrum. sitive than the cones and thus pro A tabulation of the relative scotopic vide the basis of dark-adapted vision. vision is given in Table III. The peak Because there are no rods in the luminosity for scotopic vision occurs foveal region, faint objects can more at 511 nanometers and is the equiva readily be observed at night when lent of 1746 lumens/watt. Fig. 92 the eye is not exactly directed toward shows the comparison of the absolute Photometric Units and Photometric-to-Radiant Conversion 113 Table II - Standard Luminosity Coefficient Values. C.i.E. WlnllRllh C.I.E. WlnllRcth c.1.E. Wmllnllh •••telS) VII .. (1 ••lmetllS) VII .. (.... meters) Vilul (111 0.175 380 0.00004 510 0.503 640 0.00012 520 0.710 650 0.107 390 0.061 400 0.0004 530 0.862 660 410 0.0012 540 0.954 670 0.032 0.0040 550 0.995 680 0.017 420 0.0082 430 0.0116 560 0.995 690 440 0.023 570 0.952 700 0.0041 0.038 580 0.870 710 0.0021 450 0.00105 460 0.060 590 0.757 720 0.091 600 0.631 730 0.00052 470 0.00025 480 0.139 610 0.503 740 0.208 620 0.381 750 0.00012 490 0.00006 500 0.323 630 0.265 760 measure luminosity curves for scotopic and tion from a light source. The funda photopic vision as a function of of luminous intensity is the all other wavelength. mental standard from which The sensitivity of the eye outside the wavelength limits shown in Tables II and III is very low, but not actually zero. Studies with in 2 tense infrared sources have shown that the eye is sensitive to radiation of wavelength at least as long as 1050 nanometers. Fig. 93 shows a com posite curve given by Griffin, Hub bard, and Wald 1 for the sensitivity of the eye for both foveal and peripheral ., vision from 360 to 1050 nanometers. ~ £ According to Goodeve 2 the ultra violet sensitivity of the eye extends >- to between 302.3 and 312.5 nano '= UNITS 1 PHOTOMETRIC 10- L....JW--"----!_"--'---J..---JL--...... 300 400 500 600 700 Luminous intensity (or candle WAVELENGTH - NANOMETERS power) describes luminous flux per Fig. 92 - Absolute luminosity curves unit solid angle in a particular direc- for scotopic and photopic eye response. 114 RCA Photomultiplier Manual Table III - Relative Luminosity Values jor Photopic and Scotopic Vision. WAVELENGTH PHOTOPIC V). SCOTOPIC V). (nil) (B > 3 cd m-') (B < 3 110-1 cd m--') 350 0.0003 360 0.0008 370 0.0022 380 0.00004 0.0055 390 0.00012 0.0127 400 0.0004 0.0270 410 0.0012 0.0530 420 0.0040 0.0950 430 0.0116 0.157 440 0.023 0.239 450 0.038 0.339 460 0.060 0.456 470 0.091 '0.576 480 0.139 0.713 490 0.208 0.842 500 0.323 0.948 510 0.503 0.999 520 0.710 0.953 530 0.862 0.849 540 0.954 0.697 550 0.995 0.531 560 0.995 0.365 570 0.952 0.243 580 0.870 0.155 590 0.757 0.0942 600 0.631 0.0561 610 0.503 0.0324 620 0.381 0.0188 630 0.265 0.0105 640 0.175 0.0058 650 0.107 0.0032 660 0.061 0.0017 670 0.032 0.0009 680 0.017 0.0005 690 0.0082 0.0002 700 0.0041 0.0001 710 0.0021 720 0.00105 730 0.00052 740 0.00025 750 0.00012 760 0.00006 770 0.00003 Photometric Units and Photometric-to-Radiant Conversion 115 photometric units are derived. The 2870 0 K has served as the basic test standard of luminous intensity is the standard for photoelectric measure candela; the older term candle is ments in this country for about 30 sometimes still used, but refers to the years. In order to provide a more new candle or candela. universal standard, most industries The candela is defined by the and government laboratories are radiation from a black body at the changing to 2854 OK color tempera temperature of solidification of plati ture. This latter temperature is in num. A candela is one-sixtieth of the common use in Europe. Illuminant luminous intensity of one square A as designated by the C.I.E. is a centimeter of such a radiator. tungsten lamp operated at 2854 OK color temperature. The text of this 3~------~ Manual utilizes the 2854 OK figure, 2 although tube data are still based on 2870 ° K. The difference, however, is 3 o generally negligible. Test lamps may )- be obtained with either temperature ~ -I calibration. ) ~ -2 Luminous ftux is the rate of flow ~ -3 CI) of light energy, that characteristic of w -4 > radiant energy which produces visual !;i -5 sensation. The unit of luminous flux ...J l:! -6 is the lumen, which is the flux emitted g -7 per unit solid angle by a uniform ...J -8 point source of one candela. Such a -9 source produces a total luminous flux of 4,.. lumens. -10 PERIPt£RY A radiant source may be evaluated 300 500 700 900 1100 WAVELENGTH- NANOMETERS in terms of luminous flux if the radiant-energy distribution of the Fig. 93 - Relative spectral sensitivity source is known. If W(A) is the total of the dark-adapted foveal ana radiant power in watts per unit wave peripheral retina. length, total radiant power over all A suitable substandard for practi wavelengths is f 0 cx>W(,\.) d,\., and the cal photoelectric measurements is the total luminous flux L in lumens can developmental-type calibrated lamp, be expressed as follows: RCA Dev. No. C70048, which oper ates at a current of about 4.5 amperes L = 680 fo "'W(,\) y(,\) d,\ and a voltage of 7 to 10 volts. A typical lamp calibrated at a color where y(A) represents the luminos temperature of 2854 OK provides a ity coefficient as a function of wave luminous intensity of 55 candelas. length. At a wavelength of 555 nano The luminous intensity of a tungsten meters, one watt of radiant power lamp measured in cande1as is usually corresponds to 680 lumens. The numerically somewhat greater than lumen is the most widely used unit the power delivered to the lamp in in the rating of photoemissive de watts. (The photoelectric industry is vices; for photomultipliers the typi in the process of changing the color cal test levels of luminous flux range temperature standard for the tungsten from 10-7 to 10-5 lumen (O.l to 10 test lamp. A color temperature of microlumens). 116 RCA Photomultiplier Manual Table IV - Luminance Values for Some Common Sources. Source Luminance (FooUamberts) Sun, as observed from Earth's surface at meridian...... 4.7 X 108 Moon, bright spot, as observed from Earth's surface ...... , .. " ...... , ...... 730 Clear blue sky...... 2300 lightning flash ...... , .. " ...... , ...... " . 2 X 1010 Atomic fission bomb, 0.1 millisecond after firing, SO·feet diameter ball...... 6 X 1011 Tungsten filament lamp, gas·filled, 16 lumen/watt ...... ' ...... 2.6 X 106 Piain carbon are, positive crater ...... , ...... , .. " ...... , ...... 4.7 X 106 Fluorescent lamp, T·12 bulb, cool white, 430 rnA, medium loading...... 2000 muminance (or illumination) is the density of luminous flux incident on a surface. A common unit of illumi nation is the footcandle, the illumi 103 nation produced by one lumen Vl UJ ...J uniformly distributed over an area 0 102 «2 of one square foot. It follows that a u ~ source of one candela produces an 0 10 0 illumination of one footcandle at a IL distance of one foot. 2 Table IV lists some common values 0 !:i 10-1 of illumination encountered in photo 2 electric applications. Further infor ~ 10-2 ...J mation concerning natural radiation ...J is shown in Fig. 94, which indicates 10-3 the change in natural illumination at ground level during, before, and 10-4 after sunset for a condition of clear sky and no moon.4 10-5 -4 -3 -2 -I 0 3 Photometric luminance (or bright HOURS BEFORE AND AFTER SUNSET ness) is a measure of the luminous flux per unit solid angle leaving a Fig. 94 - Natural illuminance on the surface at a given point in a given earth for the hours immediately before direction, per unit of projected area. and after sunset with a clear sky and The term photometric luminance is no moon. used to distinguish a physically describes the light emission from a measured luminance from a subjec surface, whether the surface is self tive luminance. The latter varies luminous or receives its light from with illuminance because of the shift some external luminous body. in spectral response of the eye For a surface which is uniformly toward the blue region at lower levels ditI&ing, luminance is the same re of illuminance. The term luminance gardless of the angle from which the Photometric Units and Photometric-to-Radiant Conversion 117 surface is viewed. This condition re sinO dO. At an angle of 0, the projec sults from the fact that a uniformly tion of the elementary area is equal diffusing surface obeys Lambert's law to A cos O. Because the luminous (the cosine law) of emission. Thus, flux in a particular direction is equal both the emission per unit solid angle to the product of the source strength and the projected area are propor in candelas and the solid angle, the tional to the cosine of the angle be total luminous flux L in lumens from tween the direction of observation the area A may be obtained by inte and the surface normal. gration over the hemisphere as A logical unit of luminance based follows: on the definition given above is a candela per unit area. When the unit of area is the square meter, this unit L 17 A(1l7r) lumens, or at an angle (J Manual are presented in units of with respect to the normal line the candelas, lumens, footcandies, and flux would be A(l/ 7T) (cos (J) lumens footlamberts; Table V is a conver per steradian. sion table for various photometric All photometric data in this units. Table V - Conversion Table for Various Photometric Units. luminous Intensity 1 1 candela (cd) = 1 lumen/steradian (lm sn- ) luminous Flux 4'17 lumens = total flux from uniform point source of 1 candela. Illuminance 1 footcandle (fc) = 1 lumen/ft" = 10.764 lux (meter candle) (lumen/meter') = 0.001076 phot (lumen/em') luminance 1 footlambert (fl) = 1/'17 candela/ft" = 0.0010764 lambert (1/'17 candela/em') = 1.0764 millilamberts = 3.426 nits (candela/meter') = 0.0003426 stilbs (candela/em') = 0.3183 candela/ft" = 10.764 apostilbs (l/'IT candela/meter') REFERENCES and Luminous Sensitivities," RCA Review, Vol. 28 (1967). 1. D. R. Griffin, R. Hubbard, and G. 4. I.E.S. Lighting Handbook, Illumi Wald, "The Sensitivity of the Hu nating Engineering Society, New man Eye to Infrared Radiation," York, New York (1959). JOSA, Vol. 37, No.7 (1947). 5. R. Kingslake, Applied Optics and 2. C. F. Goodeve, "Vision in the Optical Engineering, Vol. I, Chap Ultraviolet," Nature (1934). ter 1, Table II, Academic Press, 3. R. W. Engstrom and A. L. More New York and London (1965). head, "Standard Test Lamp Tem 6. J. W. Walsh, Photometry, Con perature for Photosensitive De stable and Co., Ltd., London vices-Relationship of Absolute (1953). 119 Radiant Energy and Sources ADIANT energy is energy travel radiation from the hole approaches R ing in the form of electromag that from a theoretical black radia netic waves; it is measured in joules, tor if the cross-sectional area of the ergs, or calories. The rate of flow cavity is large compared with the of radiant energy is called radiant area of the exit hole. The character flux; is it expressed in calories per istic of IOO-per-cent absorption is second, in ergs per second, or, prefer achieved because any radiation enter ably, in watts 'Goules per second). ing the hole is reflected many times inside the cavity. BLACK·BODY RADIATION The radiation distribution for a source which is not black may be As a body is raised in tempera calculated from the black-body radia ture, it first emits radiation primarily tion laws provided the emissivity as in the invisible infrared region. Then, a function of wavelength is known. as the temperature is increased, the Spectral emissivity is defined as the radiation shifts toward the shorter ratio of the output of a radiator at wavelengths. A certain type of radia a specific wavelength to that of a tion called black-body radiation is black body at the same temperature. used as a standard for the infrared Tungsten sources, for which tables region; other sources may be de of emissivity data are available,l are scribed in terms of the black body. widely used as practical standards, A black body is one which absorbs particularly for the visible range. all incident radiation; none is trans Tungsten radiation standards for the mitted and none is reflected. Because, visible range are frequently given in thermal equilibrium, thermal radia in terms of color temperature, in tion balances absorption, it follows stead of true temperature. The color that a black body is the most effi temperature of a selective radiator is cient thermal radiator possible. A determined by comparison with a black body radiates more total power black body: when the outputs of the and more power at a particular selective radiator and a black body wavelength than any other thermally are the closest possible approxima radiating source at the same tempera tion to a perfect color match in the ture. Although no material is ideally range of visual sensitivity, the color black, the equivalent of a theoretical temperature of the selective radiator black body can be achieved in the is numerically the same as the black laboratory by providing a hollow body true temperature. For a tung radiator with a small exit hole. The sten source, the relati ve distribution 120 RCA Photomultiplier Manual of radiant energy in the visible spec FLUORESCENT tral range is very close to that of a i LAMP-DAYLIGHT black body although the absolute temperatures differ. However, the match of energy distribution becomes progressively worse in the ultraviolet i~~300 II! 500 600 700 and infrared spectral regions. WAVELENGTH'" NANOMETERS Fig. 97 - Typical spectral-emission SOURCE TYPES curve for a daylight-type fluorescent lamp. Tungsten lamps are probably the A very useful point source 2 is most important type of radiation the zirconium concentrated-arc lamp. source because of their availability, Concentrated-arc lamps are available reliability, and constancy of oper with ratings from 2 to 300 watts, and ating characteristics. Commercial in point diameters from 0.003 to photomultiplier design has been con 0.116 inch. Operation of these lamps siderably influenced by the character requires one special circuit to pro istics of the tungsten lamp. A relative vide a high starting voltage and an spectral-emission characteristic for a other well filtered and ballasted cir tungsten lamp at 2854 oK color tem cuit for operation. perature is shown in Fig. 96. Although many types of electrical discharge have been used as radiation TUNGSTEN LAMP 2854 eK (STANDARD TEST) sources, probably the most important J are the mercury arc and the carbon arc. The character of the light emitted from the mercury arc varies .... with pressure and operating condi 2:0.4 tions. At increasing pressures, the ~ spectral-energy ..J distribution from :l!Q.2 the arc changes from the typical mer cury-line spectral characteristic to o 3200 an almost continuous spectrum of high intensity in the near-infrared, Fig. 96 - Relative spectral-emission visible, and ultraviolet regions. Fig. characteristic for a tungsten lamp at a 98 shows the spectral-energy distri color temperature of 2854° K. bution from a water-cooled mercury The common fluorescent lamp, a MERCURY LAMP very efficient light 40 source, consists of I I I r.! .- an argon-mercury glow discharge in - a bulb internally coated with a phos - phor that converts ultraviolet radia - tion from the discharge into useful - light output. There are numerous types of fluorescent lamps, each with - a different output spectral distribu - tion depending upon the phosphor r..'lr V. t...... - - and gas filling. The spectral response o 400 500 600 700 shown in Fig. 97 is a typical curve WAVELENGTH--NANOMETERS for a fluorescent lamp of the day Fig. 98 - Typical- spectral-emission light type. curve for a water-cooled mercury-arc lamp at a pressure of 130 atmospheres. Radiant Energy and Sources 121 arc at a pressure of 130 atmospheres. ARGON- ARC LAMP The carbon arc is a source of great intensity and high color temperature. A typical energy-distribution spec trum of a dc high-intensity arc is shown in Fig. 99. Figs. 100 and 101 , show relative spectral-emission char acteristic curves for xenon and argon arcs; Table VI specifies typical parameters for all sources described. Fig. 101- Typical spectral-emission curve for an argon-arc lamp. to detect. Although many of these interesting devices have their prin cipal wavelengths of emission in the infrared beyond the sensitivity range of photomultiplier tubes, some do not. Because of the growing impor ~~~400"-oI-5~00~L-60~O-L.-700"'-'" tance of laser applications and the use WAVElENGTH- NANOMETERS of photomultipliers for detecting their radiation, Tables VII through Fig. 99 - Typical spectral-emission XI are provided as reference data curve for a de high-intensity carbon on crystalline, gas,· and liquid lasers, arc lamp. and on p-n junction light-emitting diodes. In recent years the development of various types of lasers and p-n light LIGHT SOURCES FOR TESTING emitting diodes with very high modu lation frequencies and short rise times Monochromatic sources of many has increased the types of sources. wavelengths may be produced by that photomultipliers are called upon narrow-band filters or monochroma tors. Narrow-band filters are more XENON-ARC LAMP practical for production testing, but, ~'~~~~~r-r-~ at best, such tests are time-consuming and subject to error. Monochromatic sources are not used in general purpose testing because most appli cations involve broader-band light sources; a monochromatic test might grossly misrepresent the situation be cause of spectral-response variations. A broad-band source is probably 600 800 1000 1200 more useful as a single test because WAVELENGTH - NANOMETERS it tends to integrate out irregularities Fig. 100 - Typical spectral-emission in the spectral-response characteris curve faT a xenon-ar~ lamp. tic and more nearly represents the 122 RCA Photomultiplier Manual Table VI-Typical Parameters for the Most Commonly Used Radiant Energy Sources. DC ARC LUMINOUS LUMINOUS AVERABE LAMP nPE INPUT DIMENSIONS FLUX EFFICIENCY LUMINAIltCE POWER (1111) (III) (IIIW-I) (a1I1I1-I) (WATTS) Mercury Short Arc 200 2.5X 1.8 9500 47.5 250 (high pressure) Xenon Short Arc 150 1.3 X 1.0 3200 21 300 Xenon Short Arc 20,000 12.5X 6 1,150,000 57 3000 (in 3 mm X 6mm) Zirconium Arc 100 1.5 250 2.5 100 (diam) Vortex-Stabilized Argon Arc 24,800 3X 10 422,000 17 1400 Tungsten 79 7.9} 10 Light { 10010 1630 16.3 to Bulbs 1,000 21,500 21.5 25 Fluorescent Lamp Standard Warm White 40 2,560 64 Carbon Arc Non-Rotating 2,000 ",,5X 5 36,800 1a.4} 175 to Rotating 15,aOO ""aX a 350,000 22.2 aoo Sun 1600 Table VII - Typical Characteristics of a Number of Useful Crystal Laser Systems. WAVELENGTH MODE AND HIGHEST HOST DOPANT OF LASER TEMPERATURE OF ("m) OPERATION (oK) AI203 0.05% 0.6934 CW, P350 Cr3+ 0.6929 P 300 AI203 O.~ 0.7009 P77 Cr + 0.7041 P77 0.7670 P 300 MgF2 1% 1.6220 P77 Ni8+ MgF2 1% 1.7500 P77 CO~+ 1.a030 P77 ZnF2 2.6113 P77 C1~ + CaW04 1% 1.0580 CW 300 Nd3+ 0.9145 P77 1.3392 P 300 CaF2 1% 1.0460 P77 Ndll.J+ CaMo04 1.a~ 1.0610 CW 300 Nd + Y3AIs012 Nd3+ 1.0648 CW360 P 440 La Fa 1% 1.0633 P300 Nd3+ Radiant Energy and Sources 123 Table VII - Typical Characteristics of a Number of Useful Crystal Laser Systems. (cont'd.) WAVELENGTH MODE AND HIGHEST HOST DOPANT OF LASER TEMPERATURE OF ("II) OPERATION (OK) laF3 1% 0.5985 P 77 PrR+ CaW04 0.5% 1.0468 P77 Pr3+ Y203 5% 0.6113 P 220 Eui+ CaF2 H03+ 0.5512 P 77 CaW04 0.5% 2.0460 P 77 H03+ Y3AIs012 H03+ 2.0975 CW 77 P 300 CaW04 1% 1.6120 P 77 Er3+ Ca(Nb03)2 Er3+ 1.6100 P77 Y3AIs012 Er3+ 1.6602 P 77 CaW04 Tm 3+ 1.9110 P 77 Y3AIs012 Tm 3+ 2.0132 CW 77 P 300 Er203 Tm 3+ 1.9340 CW 77 Y3AIs012 Yb3+ 1.0296 P 77 CaF2 0.05% 2.6130 P 300 U3+ CW 77 SrF2 UH 2.4070 P 90 CaF2 .01~ 0.7083 P 20 Sm + SrF2 .01% 0.6969 P 4.2 Sm 2+ CaF2 .01~ 2.3588 CW 77 Dy + P 145 CaF2 0.01% 1.1160 P 27 Tm 2+ CW4.2 Table VIII - Comparison of Characteristics of Continuous Crystalline Lasers. MATERIAL UNSITIZER OPTICAL PUMP L~mH EFF. (%) POWER OPERATING (WATTS) TEMP. (OK) ACTIVE SYSTEM (I'm) Dy2+CaF2 W 2.36 0.06 1.2 77 Cr3+A1203 Hg 0.69 0.1 1.0 300 Nd H Y3AI5012 W 1.06 0.2 2 300 1.06 0.6 15 300 Nd 3+Y3Als012 Plasma Arc 1.06 0.2 200 300 Nd3+Y3A15012 Na Doped Hg 1.06 0.2 0.5 300 Nd H Y3AI5012 Cr3+ Hg 1.06 0.4 10 300 H03+Y3A15012 [Er3+, Yb3+, Tm3+] W 2.12 5.0 15 77 124 RCA Photomultiplier Manual typical application. The tungsten Sources such as arcs and glow dis lamp has been used for many years charges are difficult to calibrate and because it is relatively simple, stable, show serious time variations. and inexpensive, and maintains its Filters are frequently used to nar calibration. The tungsten lamp emits row the spectral range for specific a broad band of energy with rela purposes; however, they sometimes tively smooth transitions from one contribute to errors because of sig end of the spectrum to the other. nificant transmission outside the Its principal disadvantage as a gen band of interest. Filters are also sub eral source is its lack of ultraviolet ject to change in transmission with output and relatively low blue out time and are very difficult to repro put. duce with identical characteristics. Table IX - Typical Characteristics of p-n Junction Light-emitting Diodes. WAVELENGTH WER CRYSTAL (Pll) ACTION PbSe 8.5 Yes PbTe 6.5 Yes InSb 5.2 Yes PbS 4.3 Yes InAs 3.15 Yes (In x Gal_x) As 0.85-3.15 Yes In (P x As l_x) 0.91-3.15 Yes GaSb 1.6 No InP 0.91 Yes GaAs 0.90 Yes Ga (As1- x Px) 0.55-0.90 Yes CdTe 0.855 No (Zn x Cdl_x) Te 0.59-0.83 No CdTe-ZnTe 0.56-0.66 No BP 0.64 No CU2Se-ZnSe 0.40-0.63 No Zn (Sex Te I-x) 0.627 No ZnTe 0.62 No GaP 0.565 No GaP 0.68 No SiC 0.456 Radiant Energy and Sources 125 I Table X - Typical Characteristics of a Number of Useful Gas Lasers. PRINCIPAL OUTPUT POWER GAS WAVREMaTHS PULSED OR (pm) TYPICAL MAXIMUM CONnNUOUS Ne 0.3324 10mW Pulsed (ionized) 10mW CW Ne 0.5401 1 kW Pulsed (unionized) He-Ne 0.5944-0.6143 10 I'W Pulsed (unionized) 0.6328 1 mW 150 mW CW 1.1523 1-5mW 25mW CW 3.3913 H2O 27.9 1.2 W Pulsed (molecular 118 1 mW Pulsed eXCitation) 118 IOI'W CW CN 337 50mW Pulsed (molecular excitation) Table XI - Typical Characteristics for Two Liquid Lasers. PRINCIPAL LlQIIID WAYELENITH PULSE ENERI' ~) (PULSEWIDTH) Eu(o-Cl BTFA)4DMA* 0.61175 O.lJ Nd+3 :SeOCI2** 1.056 10J (150 I£S) * dimethylammonium salt of tetrakis •• trivalent neodymium in selenium europium - ortho - chloro - benzoyltriflu- oxychloride oracetonate 126 RCA Photomultiplier Manual As a result of the work of indus tested by means of an Nal(Tl) crys trial committees, virtually the entire tal and a CS137 source or a simulated photosensitive-device industry in the NaI light source utilizing a tungsten United States uses the tungsten lamp lamp whose light passes through a at 2870 0 K color temperature· as a one-half stock-thickness Corning general test source. The lamp is cali 5113 glass-type CS-5-58 filter. Inter brated in lumens and is utilized in ference-type filters are becoming in the infrared spectrum as well as the creasingly important in isolating spe visible. Typical as well as maximum cific wavelengths for testing photo and minimum photosensitivities are multiplier tubes for laser applications. quoted in microamperes per lumen. The principal disadvantages of us TESTING FOR SCIENTIFIC AND ing the tungsten lamp as an industry MULTIPURPOSE APPLICATIONS standard test are that it does not pro vide a direct measure of radiant When a photomultiplier is manu sensitivity as a function of wave factured for a variety of purposes, length and that it is a somewhat mis including scientific applications, it leading term when the response of would be highly desirable if sensi the photomultiplier lies outside the tivity were specified by a complete visible range. To assist the scientist spectral response in terms of quan in using photomultipliers, technical tum efficiency or radiant sensitivity. specifications for RCA photomulti This information could then be uti plier types include photocathode lized in conjunction with the known spectral-response curves which give spectral emission of any source to the sensitivity in absolute terms such compute the response of the photo as amperes per watt and quantum multiplier to that source. Complete efficiency as a function of wave spectral sensitivity data, however, length. Methods of computing the re are rarely provided because it is un sponse of a given photodetector to necessary for most practical situa a particular radiation source are out tions and would considerably increase lined in the following section on device costs. Spectral Response and Source-Detec tor Matching, REFERENCES FUNCTIONAL TESTING 1. J. C. DeVos, Physica, Vol. 20 (1954). In many applications it is appro 2. W. D. Buckingham and C. R. priate to test photomUltipliers in the Deihert, "Characteristics and Ap same manner in which they are to be plications of Concentrated Arc utilized in the final application. For Lamps," Journal of the Society example, photomultipliers to be of Motion Picture Engineers, Vol. used in scintillation counting may be 47 (1946). >I< Now being changed to 2854 OK color temperature to agree with international standards. 127 Spectral Response and Source-Detector Matching HIS section covers the signifi includes an indication of the range Tcance of the spectral response of over which the peak of the curve may photomultiplier tubes, describes some be expected to vary, as well as an of the methods used for measuring indication of the range over which this response; and discusses in some the cut-off at either end of the spec detail the calculations and other con tral response may vary. These cut siderations useful for matching the off points are usually expressed at radiation source and the photomul 10 per cent of the maximum value. tiplier tube type for a specific appli The spectral-response characteris cation. tic, and particularly the long-wave length cut-off, are dependent upon SPECTRAL·RESPONSE the chemical composition and the CHARACTERISTICS processing of the particular photo cathode. On the other hand, the A spectral-response characteristic ultraviolet cut-off characteristic is is a display of the response of a determined primarily by the charac photosensitive device as a function of teristic cut-off transmission of the the wavelength of the exciting radia window of the tube. tion. Such curves may be on an ab solute or a relative basis. In the lat MEASUREMENT TECHNIQUES ter case the curves are usually nor malized to unity at the peak of the In order to determine a spectral spectral-response curve. For a photo response characteristic, a source of cathode the absolute radiant sensitiv essentially monochromatic radiation, ity is expressed in amperes per watt. a current-sensitive instrument to Curves of absolute spectral response measure the output of the photo may also be expressed in terms of cathode, and a method of calibrating the quantum efficiency at the particu the monochromatic radiation for its lar wavelength. If the curve is pre magnitude in units of power are re sented in terms of amperes per watt, quired. lines of equal quantum efficiency may The source of monochromatic be indicated for convenient reference. radiation is often a prism or a grat Typical curves are usually included ing type of monochromator. Inter in tube data bulletins. Because there ference filters may also be used to are variations in spectral response isolate narrow spectral bands. Al from tube to tube, the typical spec though interference filters do not tral-response characteristic usually provide the flexibility of a mono- 128 RCA Photomultiplier Manual chromator, they may be indicated in color filters which exclude the wave situations in which repeated measure length of the measurement and in ments are required in a particular re clude the suspected spectral leakage gion of the spectrum. region, or vice versa, are used in The width of the spectral transmis checking the magnitude of the pos sion band in these measurements sible leakage spectrum. must be narrow enough to delineate the spectral-response characteristic in the required detail. However, for ENERGY SOURCES the most part, spectral-response characteristics do not require fine Various radiant-energy sources are detail and generally have broad peaks used to advantage in spectral-re with exponential cut-off characteris sponse measurements. A tungsten tics at the long-wavelength limit and lamp is useful from about 450 nano rather sharp cut-offs at the short meters to beyond 1200 nanometers wavelength end. For spectral mea because of its uniform and stable surements, therefore, a reasonably spectral-emission characteristic. The wide band is used. Such a band has linear type of filament also provides the following important advantages: good optical coupling to the slits (1) because the level of radiation is used in monochromators. A mercury higher, measurements are easier and vapor discharge lamp provides a high more precise; and (2) spectral leak concentration in specific radiation age in other parts of the spectrum is lines and thus minimizes the back relatively less important. Spectral ground scattered-radiation problem. leakage is a problem in any mono The mercury lamp is particularly chromator because of scattered radia useful in the ultraviolet end of the tion, and in any filter because there spectrum where the tungsten lamp is some transmission outside the de fails. sired pass band. A double mono chromator may be used and will MEASUREMENT OF greatly reduce the spectral leakage RADIANT POWER OUTPUT outside the pass band. The double monochromator is at a disadvantage A radiation thermocouple or in cost and complexity. If a pass band thermopile with a black absorbing of 10 nanometers is used, spectral surface is commonly used to measure leakage can be insignificant for most the radiation power output at a spe of the spectral measurements. At the cific wavelength. Although these de same time, this pass band is narrow vices are relatively low in sensitivity, enough to avoid distortion in the they do provide a reasonably reliable measured spectral-response charac means of measuring radiation inde teristic. It is often advisable to vary pendent of the wavelength. The limi the pass band depending upon what tation to their accuracy is the flatness part of the curve is being measured. of the spectral absorption character For example, at the long-wavelength istic of the black coating on the de cut-off where the response of the tector. Throughout the visible and photocathode may be very small, the near-infrared regions there is usually leakage spectrum may play a rather no problem. There is some question, important part; thus it is advisable however, as to the flatness of the to increase the spectral bandpass of response in the ultraviolet part of the measurement. Wide pass-band the spectrum. Spectral Response and Source-Detector Matching 129 The output of the thermocouple is very easy to measure. It is neces or thermopile is a voltage propor sary, however, to be careful to avoid tional to the input radiation power. fatigue effects which could distort This voltage is converted by means the spectral-response measurement. of a suitable sensitive voltmeter to It should be noted that the spectral a calibrated measure of power in response of a photomultiplier may watts. The calibration may be ac be somewhat different than that of complished by means of standard the photocathode alone because of radiation lamps obtained from the the effect of initial velocities and National Bureau of Standards. It is their effect on collection efficiency at theoretically possible to calculate the the first dynode and because of the monochromatic power from a knowl possibility of a photoeffect on the edge of the emission characteristic first dynode by light transmitted of the source, the dispersion charac through the photocathode, especially teristic of the monochromator or the if the first dynode is a photosensitive transmission characteristic of the fil material such as cesium-antimony. ter, and from the transmission char A diagram of a typical arrange acteristic of the various lenses. This ment for measurement of spectral re procedure is difficult, subject to sponse is shown in Fig. 102. In this error, and is not recommended ex arrangement a mirror is used in two cept perhaps in the case of a tung angular orientations to reflect the sten lamp source combined with a radiation alternately to the thermo narrow-band filter system. couple and to the photocathode. Other methods are also used to split MEASUREMENT OF the monochromatic beam so that PHOTOCATHODE OUTPUT both thermocouple and photocathode can be sampled alternately or simul For measuring the spectral char taneously. acteristic of a photocathode, a very sensitive ammeter is required. When SOURCE AND DETECTOR the output of the photocathode of MATCHING a photomultiplier is measured, the tube is usually operated as a photo One of the most important para diode by connecting all elements meters to be considered in the selec other than the photocathode together tion of a photomultiplier type for a to serve as the anode. When the specific application is the photo photomultiplier is operated as a con cathode spectral response. The spec ventional photomultiplier, the output tral response of RCA photomultiplier Fig. 102 - Diagram of typical arrangement for spectral-response measurements. 130 RCA Photomultiplier Manual tubes covers the spectrum from the photocathode spectral response as a ultraviolet to the near-infrared re function of wavelength normalized gion. In this range there are a large to unity at the peak. When Eq. (106) variety of spectral responses to is solved for the peak power per choose from. Some cover narrow unit wavelength, Po, and this solu ranges of the spectrum while others tion is substituted into Eq. (107), the cover a very broad range. The data cathode current is expressed as fol sheet for each photomultiplier type lows: shows the relative and absolute spec 00 tral-response curves for a typical tube of that particular type. The 1 W(}') R(}') d}' It = 0' P ~o ______relative typical spectral-response (107) curves published may be used for 00 matching the detector to a light 1 W(}') d}' source for all but the most exacting o applications. The matching of detec The ratio of the dimensionless inte tor to source consists of choosing the grals can be defined as the matching photomultiplier tube type that has a factor, M. The matching factor is the spectral response providing maximum ratio of the area under the curve de overlap of the spectral distributions fined by the product of the relative of detector and light source. source and detector spectral curves to the area under the relative spec MATCHING CALCULATIONS tral source curve. 00 The average power radiating from 1 W(}') R(}') d}' a light source may be expressed as M = -,-0______(108) follows: 00 W(}') d}' p = Po fo OOW(A) dA (105) 1o where Po is the incident power in Fig. 103 shows an example of the watts per unit wavelength at the peak data involved in the evaluation of of the relative spectral radiation the matching factor, M, as given in characteristic, WeAl, which is normal Eq. (108). ized to unity. If the input light distribution in If the absolute spectral distribu cident on the detector is modified tion for the light source and the ab with a filter or any other optical de solute spectral response of the photo vice, the matching-factor formulas multiplier tube are known, the must be changed accordingly. If the resulting photocathode current It transmission of the filter or optical when the light is incident on the de device is f().), the matching factor tector can be expressed as follows: will be 00 OO W(}') R(}') f}' d}. It 0' Po W (}.) R (}.) d}' (106) 1 = i M = _0______00 where (T is the radiant sensitivity of 1 W(}') f(}') d}. the photocathode in amperes per o watt at the peak of the relative curve, Table XII shows a number of match and R{A) represents the relative ing factors for various light sources Spectral Response and Source-Detector Matching 131 Table XII - Spectral Matching Factors.** OTHER OETECTORS PHOTOCATHODES Pho· Sco· totlk topic st 54 510 511 511 520 525 IY' IY. Nolts " k k k k k k I m Phosphors PI a 0.278 0.498 0.807 0.687 0.892 0.700 0.853 0.768 0.743 P4 a,b 0.310 0.549 0.767 0.661 0.734 0.724 0.861 0.402 0.452 P7 a 0.312 0.611 0.805 0.709 0.773 0.771 0.882 0.411 0.388 P11 a 0.217 0.816 0.949 0.914 0.954 0.877 0.953 0.201 0.601 PIS a 0.385 0.701 0.855 0.787 0.871 0.802 0.904 0.376 0.495 P16 a 0.830 0.970 0.853 0.880 0.855 0.902 0.922 0.003 0.042 P20 a 0.395 0.284 0.612 0.427 0.563 0.583 0.782 0.707 0.354 P22B c 0.217 0.893 0.974 0.960 0948 0.927 0.979 0.808 0.477 P22G c 0.278 0.495 0.807 0.686 0.896 0.699 0.855 0.784 0.747 P22R c 0.632 0.036 0.264 0.055 0.077 0.368 0.623 0.225 0.008 P24 a 0.279 0.545 0.806 0.696 0.827 0.725 0.869 0.540 0.621 P31 a,d 0.276 0.533 0.811 0.698 0.853 0.722 0.868 0.626 0.651 Nal e 0.534 0.923 0.885 0.889 0.889 0.900 0.933 0.046 0.224 Lamps 2870/2854 std 0.516- 0.046* 0.095· 0.060"' 0.072· 0.112· 0.227· 0.071 .of 0.040* Fluorescent g 0.395 0.390 0.650 0.496 0.575 0.635 0.805 0.502 0.314 Sun In space h 0.535- 0.308* 0.388· 0.328* 0.380· 0.4OS* 0.547* 0.179- 0.172- +2 air masses h,i 0.536- 0.236* 0.348· 0.277· 0.315- 0.360* 0.513· 0.197· 0.175- Day sky 0.537· 0.520* 0.556· 0.508* 0.589* 0.581* 0.700* 0.170· 0.218* Blackbodies 60000 K 0.533* 0.308* 0.376· 0.320· 0.375* 0.397* 0.521- 0.167- 0.159- 30000 K 0.512- 0.053* 0.10Z- 0.067· 0.080* 0.120* 0.232- 0.075- 0.044- 28lQOK 0.504* 0.044· 0.090* 0.057* 0.OS9- O.1OS- 0.Z16· 0.OS7· 0.038· 2854°K 0.500· 0.042· 0.088· 0.055· 0.068- 0.103· 0.211· O.OSS· 0.037- 28100 K 0.493- 0.039* 0.081* 0.051· 0.062* 0.097- 0.150* 0.061· 0.034· 2042°K 0.401* 0.008* 0.023* 0.011* 0.014* 0.033* 0.090- 0.018* 0.007* • Entry valid only for JOO·1200-nm wavelength in- Approximately noon sealevel flux at 60· terval. latitude. From Gates between 300 nm and 530 nm. t For the total wavelength spectrum this entry would A 12,OOO"K bla.ckbody spectral distribu- be 0.0294. tlon was assumed between 530 nm and Notes: a Registered spectral distribution. Data 1200 nm. extrapolated as required. k Registered spectral distribution. Data extrapolated as required. b Sulfide type. I Standard tabulated photopic visibility dis- RCA data. tribution. Low brightness type. I! Standard tabulated scotopic visibility dis- e Harshaw Chemical Co. data. tribution. f Standard test lamp distribution. •• Data from E. H. Eberhardt, "Source-Detector I General Electric Co. data. Spectral Matching Factors", Applied Optics, b From Haldbalk af Geo,\lJsics. 7, 2037, (1968) 132 RCA Photomultiplier Manual tor having a photocathode spectral 100 response that will maximize the 8 photocathode current, Ik , for a given 6 light source. Maximizing the cathode current is important to maximize the 4 signal-to-noise ratio. From Eq. (109) it can be seen that the product of the matching factor M and the peak absolute photocathode sensitivity U" 2 must be maximized to maximize the cathode current. It is advantageous (/) then to define the product of f the peak Z absolute cathode sensitivity U" and the OJ 10 matching factor M as N, a value to ~ 8 be selected as a maximum for a given f= light ~ 6 source: w tr 4 N = q M I (110) I Therefore, 2 I h=PN (111) The importance of taking into ac count the absolute photocathode sensitivity, as well as the matching 300 500 700 900 WAVELENGTH - NANOMETERS factor, is illustrated by a comparison of the S-1 and S-20 photocathodes Fig. 103 - Graphic example of factors used in evaluation of matching factor with a tungsten light source operat M. ing at a color temperature of 2854 0 K. The S-1 and the and spectral response characteris S-20 match ing factors are 0.516 and tics. When the spectral range 0.112, re of a spectively. From the matching source exceeded 1200 nanometers, factors alone it appears that the S-1 is the integration was terminated at the best choice of photocathode spectral this wavelength. Because none of the photoresponses response. The S-l has a peak abso exceed 1200 nano lute meters, conclusions sensitivity of 2.5 milliamperes as to the rela per watt tive merit of and the S-20 has a peak various combinations absolute are still valid. sensitivity of 64 milli amperes When M is substituted per watt. Calculation of the for the in variable N for tegral ratio in Eq. (107), the two photocathodes the photo yields the cathode current becomes following results: h= qPM (109) N (S-1) = 0.0025 x 0.516 = 0.00129 N (S-20) = 0.064 x 0.112 = 0.00716 Matching Factor These calculations show that the S-20 photocathode will provide a substan In any photomultiplier applica tially larger response to the tungsten tion it is desirable to choose a detec- lamp than the S-l. Spectral Response and Source-Detector Matching 133 REFERENCES 1. E. H. Eberhardt, "Source-Detec Applied Optics, Vol. 7, p. 2037 tor Spectral Matching Factors," (1968). 134 Technical Data THE tabulated data on the follow- or RCA field sales representatives be ing pages are believed to be ac- contacted for current information on curate and reliable for the listed types the individual types. at the date of printing. However the The data shown have been mea- data, especially for developmental sured using a tungsten-filament lamp types, are subject to change. It is operated at a color temperature of recommended that RCA distributors 2870°K. It is intended that future 'RELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Spectral Response MAXIMUM RATINGS Nominal Spmral TUH Number Viewlnr Ca.1 RCA Outline, Suptlly Averl,1 Responsea Dlamlter 01 Conliru· Slruc· Type Basinr VoI\af1 Anodl Incbes StallS ratlonb lurl' No .• Oialram (E) Currlnl V mA 3/4 10 H C70102B 4 1500 0.01 1-1/8 9 S C C31004A 11 1500 0.01 101 (S-l) 1-1/2 10 H C 7102 18 1500 0.01 1-1/2 10 H C C70114D 19 1500 0.01 2 12 H I C70007A 25 2000 0.01 1/2 9 S C 8571 1 1250 0.02 1-1/8 9 S C 1P21 11 1250 0.1 1-1/8 9 S C 931A 11 1250 1.0 1-1/8 9 S C 4471 11 1250 1.0 1-1/8 9 S C 4472 11 1250 1.0 102 (S-4) 1-1/8 9 S C 4473 11 1250 0.1 1-1/8 9 S C 6328 12 1250 0.1 1-1/8 9 S C 6472 13 1250 0.1 1-1/8 9 S C 7117 12 1250 0.1 1-1/8 9 S C C7075J 11 1250 0.1 1-1/8 9 S C IP28/Vl 11 1250 0.5 103 1-1/8 9 S C lP28A/V1 11 1250 0.5 Technical Data 135 data for RCA photomultipliers be Photomultiplier variants of each of measured using the international the listed types are also available color temperature of 2854 oK. with permanently attached bases, Variants of the listed photomulti with temporary bases attached to pliers are also available having elec semiflexible leads, or with flexible trostatic and magnetic shielding and! leads only. or integral voltage-divider networks. TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22° C Sensitivity anldl Dirk Rldia.t Lullli_ Su"" Current CurrentnA@ $JIcIr1l YoIlIlI Clthode Anode Clthllde Anode Artplilicaliln lIede LUlllinOlS R..... Y IIIA/W A/W uA/l1II A/lm <"'I'll.) SlIIIIdrity A/l_ 1250 2.8 310 30 3.3 1.1 X 105 800@4 1250 1.9 235 20 2.5 1.25 X 105 300@2 1250 2.8 660 30 7 2.3 X 105 1900@4 101 (S·l) 1250 2.8 420 30 4.5 1.5 X 105 1700@4 1250 2.8 940 30 10 3.3 X 105 400@4 1000 34 73,000 35 75 2.1 X 1(J6 2@20 1000 40 120,000 40 120 3 X 1(J6 1@20 1000 40 83,000 40 80 2 X 1(J6 5@20 1000 40 100,000 40 100 2.5 X 106 5@20 1000 40 100,000 40 100 2.5 X 1(J6 5@20 1000 40 160,000 40 160 4 X 1(J6 1@20 102 (S·4) 1000 35,000 35 1000 32,500 35 1000 35,000 35 1000 40 83,000 40 80 2 X 106 5@20 1000 48 160,000 60 200 3.3 X 1(J6 2@40 103 1000 48 160,000 60 200 3.3 X 1()1 2@40 136 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Spectral Response (cont'd) MAXIMUM RATINGS Nominal S,.w., Tube Number Viewlne Call RCA Outline, Supply Averlp Rn,ans.- Dlameier of Conficu· Slruc· Type Basine Voliae' Anodl Inchn SIaCII r.tionb lUll' No •• Dlazram (E) Currenl V mA 1/2 9 S C C70l29H 2 1250 0.02 104 (S-5) 1-1/8 9 S C IP28 11 1250 0.5 1-1/8 9 S C IP28A 11 1250 0.5 105 ($08) 1-1/8 9 S C IP22 11 1250 1.0 106 (S-10) 2 10 H C 6217 26 1250 0.75 3/4 6 H 7764 5 1500 0.5 3/4 10 H 4460 4 1500 0.5 3/4 10 H 7767 4 1500 0.5 3/4 10 H C70l02E 4 1500 0.5 3/4 12 H I C31005B 6 2000 0.5 1-1/2 10 H C 2060 18 1250 0.75 1-1/2 10 H C 2067 20 1250 0.75 1-1/2 10 H C 4438 21 1250 0.75 1-1/2 10 H C 4439 21 1250 0.75 1-1/2 10 H C 4440 21 1250 0.75 1-1/2 10 H C 4441 19 1250 0.75 1-1/2 10 H C 4441A 19 1250 0.75 1-1/2 10 H C 4461 19 1500 1.0 107 (S-l1) 1-1/2 10 H C 6199 18 1250 0.75 1-1/2 10 H C C7151N 22 1600 0.5 1-1/2 10 H C C70132B 20 1600 0.5 2 10 H C 2020 27 1500 2.0 2 10 H C 2061 28 1500 2.0 2 10 H C 2062 28 1250 0.75 2 10 H V 2063 29 2000 2.0 2 10 H C 5819 26 1250 0.75 2 10 H C 6342A 30 1500 2.0 2 10 H C 6655A 30 1250 0.75 2 10 H 7746 31 2500 2.0 2 10 H V 8053 29 2000 2.0 2 12 H I 7850 25 2600 2.0 2 14 H 7264 32 2400 2.0 2 14 H 6810A 32 2400 2.0 Technical Data 137 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22" C Sensitivity Anode Dlr. Radiant Luminous Currenl CurrentnA@ Spectral SUJIIIIJ Amplification Anodl Lullinous RH~ Voltl,e Clthade Anode Calhode Anodl (Approl.) Sensitivity A/lm V mA/W A/W uA/lm A/lm 6 1000 44 44,000 35 35 1 X 10 20@20 6 104 (S-5) 1000 50 120,000 40 100 2.5 X 10 5@20 6 1000 50 250,000 40 200 5 X 10 5@20 6 105 (S·S) 1000 2.3 7500 3 10 3.3 X 10 [email protected] 6 106 (S·lO) 1000 20 50,000 40 100 2.5 X 10 200@20 1200 48 480 60 0.6 1 X 10· [email protected] 5 1250 48 6000 60 7.5 1.25 X 10 [email protected] 5 7.5 1250 48 12,800 60 16 2.7 X 10 4@ 5 1250 56 8800 70 11 1.6 X 10 1.4@ 7.5 6 100 1500 56 120,000 70 155 2.2 X 10 100@ 6 1000 36 36,000 45 45 1 X 10 4.5@20 5 1000 60 16,000 74 20 2.7 X 10 2.6@20 5 1000 36 22,000 45 27 6 X 10 16@20 5 1000 36 22,000 45 27 6 X 10 16@20 5 1000 36 22,000 45 27 6 X 10 16@20 5 1000 36 22,000 45 27 6 X 10 16@20 5 1000 36 22,000 45 27 6 X 10 16@20 5 5@10 1250 48 8000 60 10 1.7 X 10 6 107 (S·Il) 1000 36 36,000 45 45 1 X 10 4.5@20 5 1500 70 56,000 85 70 8.1 X 10 0.8@20 5 1500 70 56,000 85 70 8.1 X 10 0.8@20 5 4@20 1250 40 4800 50 6 1.2 X 10 1250 64 80 6 6@20 1000 61 96,000 76 120 1.6 X 10 1500 56 70 6 i'@20 1000 40 80,000 50 100 2 X 10 5 4@20 1250 64 25,000 80 31 3.9 X 10 6 6@20 1000 61 96,000 76 120 1.6 X 10 1500 56 100,000 70 130 1.8 X 1Q6 16@20 5 4@9 1500 56 34,000 70 42 6 X 10 6 1800 56 510,000 70 640 9.1 X 10 64@160 7 2000 56 3,400,000 70 4200 6.1 X 10 1000@2000 7 2000 56 3,000,000 70 3800 5.4 X 10 1000@2000 138 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Spectral Response (cont'd) I MAXIMUM RATINGS Nominal S_II Tube Numblr Yiewinl Cap RCA Outline, Su ... I, AYWlIII RlSpoIISl' Dia1IIlIIr of ConfiIY· Strue· T'III "slnt VoIlIlI ARodI InchlS SIIIIS rltion b turl c No .• Diallim (E) Currllll V mA 3 10 H V 2064B 38 2000 2.0 107 (S-l1) 3 10 H V 8054 38 2000 2.0 (Cont'd) 5 10 H V 2065 41 2000 2.0 5 10 H V 8055 41 2000 2.0 108 (S-13) 2 10 H C 6903 '30 1250 0.75 109 (S-19) 1·1/8 9 S C 7200 14 1250 0.5 3/4 10 H 8644 4 2100 0.5 3/4 10 H 8645 8 1800 0.1 1·1/2 10 H C C70114C 19 1800 1.0 2 10 H V 4463 29 2500 1.0 110 (S-20) 2 10 H 7326 33 2400 1.0 2 12 H 4459 25 2800 1.0 2 14 H 7265 32 3000 1.0 3 10 H V 4464 38 2500 1.0 5 10 H V 4465 41 2500 1.0 111 1·1/2 10 OW C 4526 23 2000 0.1 1·1/2 10 H C C70114E 19 1800 1.0 112 3 14 S C7oo45C 39 6000 1.0 2 12 H C31OO0A 34 3000 1.0 113 2 12 H C31000B 34 3000 1.0 3/4 10 H C700420 4 2100 0.5 114 2 10 H V C70109E 29 2500 1.0 2 14 H C7268 32 3000 1.0 3/4 10 H 4516 4 1800 0.5 3/4 10 H I C70l02M 4 1800 0.5 1 10 H C C31016A 9 1500 0.02 1 10 H C C31016B 9 1500 0.02 1·1/2 10 H C 4517 18 1800 0.5 115 1·1/2 10 H C C7151Q 22 1800 0.5 1·1/2 10 H C C70114F 19 1800 0.5 1·1/2 10 H C C70132A 20 1800 0.5 2 10 H C 4518 30 2000 0.5 2 10 H V 4523 29 2500 0.5 Technical Data 139 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22' C Sensitivity Anodl Dark Luminous SUPIIIJ Radiant Current CurrlntnA@ Spectrll Voltall Cathode Anode Cathode Anode Amplideation Anodl Luminous lesp ••se a V mA/W A/W uA/lm A/lm (APlltox.) SlnsiliYfty A/1m 1500 64 80 1500 64 35,000 80 43 5.4 X 105 4@9 107 (S-l1) 1500 88 110 (Cont'd) 1500 88 35,000 110 44 4 X 105 4@9 1000 48 72,000 60 90 1.5 X 105 1O@20 1118 (S-13) 1000 65 65,000 40 40 1 X 10' 4@20 109 (S-19) 1500 64 5100 150 12 8 X 10' 1.2@30 1500 64 5100 150 12 8 X 10' 1.2@~ 1500 77 11,000 180 25 1.4 X 105 4@10 2000 68 11,000 160 25 1.6 X 105 4.8@12 1800 64 38,000 150 88 5.9 X 105 3@20 110 (S-20) 2300 64 430,000 150 1000 6.6 X 10' 3O@300 2400 64 3,000,000 150 7200 4:8 X 107 50@looo 2000 68 11,000 160 25 1.6 X 105 4.8@ 12 2000 68 11,000 . 160 25 1.6 X lOS 4.8@12 1250 89 4400 300 15 5 X 10' 2@20 III 1500 77 11,000 180 25 1.4 X 105 4@1O 112 5000 60 140 5 X 106 500@1000 2000 77 270,000 200 700 3.5 X 106 5@200 113. 2000 77 270,000 200 700 3.5 X 106 5@200 1500 60 4300 140 10 7.1 X 10' 6@30 1500 68 3400 160 8 5 X 10' 0.2@6 114 2400 64 2,200,000 150 5200 3.4 X 107 50@lOoo 1500 71 56,000 60 47 8 X 105 0.2@7 1500 79 32,000 67 27 4 X 105 0.2@7 1250 71 12,000 60 10 6 X 105 0.5@7 1250 79 15,000 67 13 1.9 X 105 0.5@7 1500 79 56,000 67 47 7 X 105 0.2@7 115 1500 79 39,000 67 33 5 X lOS 0.3@7 1500 79 39,000 67 33 5 X lOS 0.3@7 1500 79 65,000 67 55 8.2 X 105 [email protected] 1500 79 39,000 67 33 5 X 105 0.24@7 1500 71 32,000 60 27 4.5 X 105 0.5@13 140 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Spectral Response (cont'd) MAXIMUM RATINGS Nominal Splctral Tube Number Viewing Call RCA OUtline, Supply Averall RHponse a Diameter of Configu- Strue- NoType __ Basinl Voltall Anode Inches StalH ration' ture e Dialram (E) Current V mA 3 10 H V 4524 38 2500 0_5 115 5 10 H V 4525 41 2500 0_5 (Coni'd) 5 10 H V C31027 42 2000 0.5 5 12 H V C31029 43 2500 0.5 2 12 H 8575 34 3000 0.2 116 2 12 H 8850 34 3000 0.2 2 12 H I 8851 34 3000 0.2 1-1/2 10 H C C70114J 19 1800 0.5 117 3 14 S C70045D 39 6000 1.0 5 14 H 4522 44 3000 0.5 118 5 14 H C70l33B 44 3000 0_5 2 12 H C31000E 34 2500 1.0 119 2 12 H C31000F 34 2500 1.0 2 10 H V 8664 35 2000 2.0 120 2 10 H V 8664/V1 35 2000 2.0 121 3/4 12 H C31005 6 2500 0_5 122 3 10 H V 4521 40 2000 0.5 1/2 9 S C C70l29E 3 1250 0.02 123 1-1/8 9 S C C31022 15 1250 0.1 124 3/4 10 H C70102N 4 1500 0.5 125 3/4 12 H C70l28 7 1800 0.5 126 2 5 H I C31024 36 6000 0.5 127 1-1/2 10 H C C7l5lU 24 1250 0.75 128 1-1/8 9 S C C31025C 16 1500 0.1 1-1/8 9 S C C31025B 16 1800 0.1 129 1-1/8 9 S C C31025G 16 1800 0.1 130 2 10 H C C7164R 30 1500 0.5 131 3/4 10 H C70042K 4 2100 0.5 Technical Data 141 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22° C Sensitivity Anode Dirk Supply Radiant lumin'1!!! Current CurrentnA@ Spectral Voltage Ctlt!GdIl A~!!t!; Cathode Anode Amplification Anode Luminous Response' Ji mA/W A/W uA/lm A/1m (Approx.) Sensitivity A/1m 1500 71 32,000 60 27 4.5 X 105 1 @ 13 1500 80 32,000 67 27 4 X 105 1.5 @ 13 1500 88 13,000 77 11.5 1.5 X 105 [email protected] 115 A/incident Imd (Cont'd) 1750 88 130,000 77 115 1.5 X 106 20@9 A/incident Imd 2000 97 970,000 85 850 1 X 107 1@2oo 2000 97 710,000 85 620 7.3 X 106 0.6@200 116 2000 97 710,000 85 620 7.3 X 106 0.6@200 2500 79 39,000 67 33 5 X 105 0.3@ 7 117 5000 72 1 X 107 1000@ 1000 2000 88 2,600,000 77 2300 3 X 107 60@2000 118 2000 88 2,600,000 77 2300 3 X 107 60@2000 1500 45 18,000 250 100 4 X 105 10@30 119 1500 45 18,000 250 100 4 X 105 10@30 1500 69 18,000 67 17 2.6 X 105 1 @ 7.5 120 1500 69 18,000 67 17 2.6 X 105 1 @ 7.5 2100 9.2min.@ 9200min.@ 1 X 106 0.1 @3000 121 253.7 nm 253.7 nm A/W 1500 87 19,000 83 18 2.2 X 105 2@ 7.5 122 1000 31 21,000 30 20 6.7 X 105 8@15 123 1000 48 160,000 60 200 3.3 X 106 2@40 1250 72 4800 90 6 6.7 X 104 [email protected] 124 1500 1C@ 3000@ 3 X 105 0.1 @ 3000 125 253.7 nm 253.7 nm A/W 3000 82 110,000 72 93 1.3 X 106 24@85 126 1000 25 8200 60 20 3.3 X 105 10@40 127 1250 33 1300 250 10 4 X 104 0.5@ 10 128 1250 45 2000 160 7 4.4 X 104 0.5@10 129 1250 28 1400 100 5 5 X 104 0.3@5 1250 40 20,000 200 100 5 X 105 30@150 130 1500 45 3600 250 20 8 X 104 6@30 131 I 142 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Spectral Response (cont'd) MAXIMUM RATINGS Nomlnll Specll'll Tube Number Vlmnr Care RCA OuHlne, SUllllly Average Ruponse' Diameter 01 Conn,u· Strue· TiP' Basinl Ve/tqe Anode Inchu Stiles ratianb tUII' No .• illlZiiiln Currlll!l (~ mA 3/4 10 H C70042R 4 2100 0.5 132 1 12 H I C31026 10 2200 0.5 1-1/2 10 H C C7151W 18 1500 0.5 133 3/4 10 H C70042J 4 1800 0.5 136 1-1/8 9 S C C31028 17 1250 0.5 137 2 11 H I C31000K 37 2000 0.1 138 1-1/2 10 H C C7151Y 24 1500 0.5 • See bar chart of Spectral response on page 156. • Type numbers with prefix C are developmental b Viewing configurations: H, head on; S, side on; types. Each of these C numbers identifies a par- and OW, dormer window. tlcular laboratory tube design but the number and 'Cage structures: C, circular cage; I, in line; and the identifying data are subject to chan~e. No V, venetian blind. obligations are assumed as to future manu acture • With blue light source. Tungsten light source at unless otherwise arranged . color temperature of 2870oK. Corning C.S. No.5-58, 1/2 stock thickness. PRELIMINARY SELECTION GUIDES For Photomultiplier, Tubes by Diameter MAXIMUM RATINGS Namlnll Tube SjllClrll Number Vlewlnr Cap RCA OuUlne, Su"" AYenp Diameter RuponseA 01 CanUp· Strue· Type Basinl Veltap Anode Inchu Stales rlHanb tur.' No .• Diacram Current (~ mA 102 (S-4) 9 S C 8571 1250 0.02 1/2 104 (:s-~) 9 S C (;/0l29H 2 1250 0.02 123 9 S C C70129G 3 1250 0.02 101 (S-l) 10 H C70102B 4 1500 0.01 6 H 7764 5 1500 0.5 3/4 10 H 4460 4 1500 0.5 107 (S-l1) 10 H 7767 4 1500 0.5 10 H C70102£ 4 1500 0.5 12 H C31005B 6 2000 0.5 Technical Data 143 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22 0 C SenslUvily Anode Dark Su"I, Radiant Luminous Currenl CurrentnA@ S,.eIral VoIla,1 Calhode Anodl Calhodl Anodl Amplification Anodl Luminous RII(IOIIII" V mA/W A/W oA/11I A/1m (A ....ox.) Sinsitivtly A/1m 1500 44 5500 200 25 1.25 X 105 2@30 1800 43 26,000 250 150 6 X 105 40@50 132 1250 40 10,000 200 50 2.5 X 105 1@20 1800 71 56,000 60 47 8 X 105 0.2@ 7 133 1000 54 175,000 65 200 3.1 X 10' 0.8@20 136 1500 35 21,000 330 200 6 X 105 6@loo 137 1250 40 2400 200 12 6 X 1()4 3@10 138 TYPICAL CHARACTERlnlCS AT THE SPECIFIED SUPPLY YOLU8E AND 22 0 C Sensitivity Anod.Dark Nllllul RatIIaII LIIII_ c.rtnI CIIRIlllIIA@ 'hili VoIlaps.. A.adeLIIII .... IIIIIIIIIr CaIlMldl ...... Cal ...... Y IIA/W A/W IA/I_ A/I_ -=~ S.... 1hI1y A/1m .... 1000 34 73,000 35 75 2.1 X 10' 2@20 1000 44 44,000 35 35 1 X 10' 20@20 1/2 1000 31 21,000 30 20 6.7 X 105 8@15 1250 2.8 310 30 3.3 1.1 X 105 800@4 1200 48 480 60 0.6 1 X 1()4 [email protected] 1250 48 6000 60 7.5 1.25 X 105 [email protected] 3/4 1250 48 12,800 60 16 2.7 X 105 [email protected] 1250 56 8800 70 11 1.6 X 105 1.4@ 7.5 1500 56 120,000 70 155 2.2 X 1()6 100@ 100 144 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Diameter (cont/d) MAXIMUM RATINGS Nominal Tube Spectral Number Vil'ing Cage RCA OUtlinl, Supply Averagl Diameler RI$PODSI' 01 Confiru· Strne· TYPI Basing Vollarl Anodl Inchl$ Sta,ls rationb turl' No .• Diagram (E) Currlnt V mA 10 H 8644 4 2100 O.S 110 (5-20) 10 H 864S 8 1800 0.1 114 10 H C70042D 4 2100 O.S 10 H 4S16 4 1800 O.S 115 10 H C70102M 4 1800 O.S 3/4 121 12 H C3lO0S 6 2S00 O.S (Cont'd) 124 10 H C70102N 4 lS00 O.S 12S 12 H C70128 7 1800 O.S 131 10 H C70042K 4 2100 O.S 132 10 H C70042R 4 2100 O.S 133 10 H I C70042J 4 1800 O.S 10 H C C31016A 9 lS00 0.02 115 1 10 H C C31016B 9 lS00 0.02 132 12 H I C31026 10 2200 O.S 101 (5-1) 9 5 C C31004A 11 lS00 0.01 9 5 C 1P21 11 12S0 0.1 9 5 C 931A 11 12S0 1.0 9 5 C 4471 11 12S0 1.0 9 5 C 4472 11 12S0 1.0 102 (5-4) 9 5 C 4473 11 12S0 0.1 9 5 C 6328 12 12S0 0.1 9 5 C 6472 13 12S0 0.1 1-1/8 9 5 C 7117 12 12S0 0.1 9 5 C C707SJ 11 12S0 0.1 9 5 C 1P28jV1 11 12S0 O.S 103 9 5 C 1P28A/V1 11 12S0 O.S 9 5 C 1P28 11 12S0 O.S 104 (5-S) 9 5 C 1P28A 11 12S0 O.S lOS (5-8) 9 5 C 1P22 11 12S0 1.0 109 (5-19) 9 5 C 7200 14 12S0 O.S 123 9 5 C C31022 IS 12S0 0.1 Technical Data 145 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22° C Sensitivity Anode Dark Nominll SUIIII'Y Radiant Luminous Current CurrentnA@ Tube Volta,e Cathode Anode Cathode Anode Amplification Anode Luminous Diameter V mA/W A/W uA/lm A/1m (Approl.) Sensitivity A/1m Inches 1500 64 5100 150 12 8 X 104 3@30 1500 64 5100 150 12 8 X 104 3@30 1500 60 4300 140 10 7.1 X 104 6@30 1500 71 56,000 60 47 8 X 105 0.2@7 1500 79 32,000 67 27 4 X 105 0.2@ 7 2100 9.2min.@ 9200 min. @ 1 X 106 0.1 @ 3000 253.7 nm 253.7 nm min. AjW 3/4 1250 72 4800 90 6 6.7 X 104 2@ 7.5 (Cont'd) 1500 10@ 3000@ 3 X 105 0.1 @ 3000 253.7 nm 253.7 nm AjW 1500 45 3600 250 20 8 X 104 6@30 1500 44 5500 200 25 1.25 X 105 2@30 1800 71 56,000 60 47 8 X 105 0.2 @ 7 1250 71 12,000 60 10 6 X 105 0.5@ 7 1250 79 15,000 67 13 1.9 X 105 0.5@ 7 1 1800 43 26,000 250 150 6 X 105 40@50 1250 1.9 235 20 2.5 1.25 X 105 300@2 1000 40 120,000 40 120 3 X 106 1@20 1000 40 83,000 40 80 2 X 106 5@20 1000 40 100,000 40 100 2.5 X 106 5@20 1000 40 100,000 40 100 2.5 X 106 5@20 1000 40 160,000 40 160 4 X 106 1@20 1000 35,000 35 1000 32,500 35 1000 35,000 35 1-1/8 1000 40 83,000 40 80 2 X 106 [email protected] 1000 48 160,000 60 200 3.3 X 106 2@40 1000 48 160,000 60 200 3.3 X 106 2@40 1000 50 120,000 40 100 2.5 X 106 5@20 1000 50 250,000 40 200 5 X 106 5@20 1000 2.3 7500 3 10 3.3 X lOS [email protected] 1000 65 65,000 40 40 1 X 106 4@20 1000 48 160,000 60 200 3.3 X 10' 2@40 146 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Diameter (cont'd) MAXIMUM RATINGS N8IIII1I1 T.... S,.eIr,1 NulUer YlIWiIlf RCA 0II1II11, s.... , A' ...... DllIIIIIr R..,... ., CIIIIIIIU· ;::. TJ" BIll.. Vol ...... I.... SlIps rlbb lur.· No .• DllII"'. Curr'". (~ .a 128 9 S C C31025C 16 1500 0.1 1-1/8 9 S C C31025B 16 1800 0.1 (Cont'd) 129 9 S C C31025G 16 1800 0.1 136 9 S C C31028 17 1250 0.5 10 H C 7102 18 1500 0.01 101 (S·l) 10 H C C701140 19 1500 0.01 10 H C 2060 18 1250 0.75 10 H C 2OS7 20 1250 0.75 10 H C 4438 21 1250 0.75 10 H C 4439 21 1250 0.75 10 H C 4440 21 1250 0.75 107 (S-11) 10 H C 4441 19 1250 0.75 10 H C 4441A 19 1250 0.75 10 H C 4461 19 1500 1.0 1-1/2 10 H C 6199 18 1250 0.75 10 H C C7151N 22 1600 0.5 10 H C C70132B 20 1600 0.5 110 (S·20) 10 H C C70114C 19 1800 1.0 111 10 OW C 4526 23 2000 0.1 112 10 H C C701l4E 19 1800 1.0 10 H C 4517 18 1800 0.5 10 H C C701l4F 19 1800 0.5 115 10 H C C7151Q 22 1800 0.5 10 H C C70132A 20 1800 0.5 117 10 H C C70114J 19 1800 0.5 127 10 H C C7151U 24 1250 0.75 132 10 H C C7151W 18 1500 0.5 138 10 H C C7151Y 24 1500 0.5 101 (S-l) 12 H I C70007A 25 2000 0.01 lOS (S-10) 10 H C 6217 26 1250 0.75 2 10 H C 2020 27 1500 2.0 107 (S-11) 10 H C 2OS1 28 1500 2.0 10 H C 2OS2 28 1250 0.75 Technical Data 147 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22· C SHsiti'iIJ AnodeDeIt NOIIInel Su., Rldllnt unRinous Current Currtat nA@ T_ Val .... Clthod. Anode Clthod. Anlde Amt.:tlClflon Anod. Lalll.otIS DlllItIIr V mA/W A/W uA/lm A/1m ( pprOL) SlIIiU1I\JA/lm IICbts 1250 33 1300 250 10 4 X 1()4 0.5@10 1250 45 2000 ... 160 7 4.4 X 1()4 0.5@10 1-1/8 1250 28 1400 100 5 ~ X l()4 0.3@5 (Cont'd) 1000 54 175,000 65 200 3.1 X 1()8 0.8@20 1250 2.8 660 30 7 2.3 X 105 1900@4 1250 2.8 420 30 4.5 1.5 X 105 17oo@4 1000 36 36,000 45 45 1 X 106 4.5@20 1000 60 16,000 74 20 2.7 X 105 2.6@20 1000 36 22,000 45 27 6 X 105 16@20 1000 36 22,000 45 27 6 X 105 16@20 1000 36 22,000 45 27 6 X 105 16@20 1000 36 22,000 45 27 6 X 1Q5 16@20 1000 36 22,000 45 27 6 X 105 16@20 1250 48 8000 60 10 1.7 X lOS 5@10 1000 36 36,000 45 45 1 X 1()8 4.5@20 1-1/2 1500 70 56,000 85 70 8.1 X 1()5 0.8@20 1500 70 56,000 85 70 8.1 X lOS 0.8@20 1500 77 11,000 180 25 1.4 X lOS 4@10 1250 89 4400 300 15 5 X 1()4 2@20 1500 77 11,000 180 25 1.4 X 105 4@10 1500 79 56,000 67 47 7 X 105 0.2@7 1500 79 39,000 67 33 5 X 105 0.3@7 1500 79 39,000 67 33 5 X 105 0.3@7 1500 79 65,000 67 55 8.2 X 105 [email protected] 1500 79 39,00) 67 33 5 X 105 0.3@7 1000 24 8200 60 20 3.3 X 105 10@4O 1250 40 10,000 200 50 2.5 X 105 1@20 1250 40 2400 200 12 6 X 1()4 3@10 1250 2.8 940 30 10 3.3 X 1Q5 4oo@4 1000 20 50,000 40 100 2.5 X lOS 2oo@20 1250 40 4800 50 6 1.2 X 105 4@20 2 1250 64 80 1000 61 96,000 76 120 1.6 X lOS 6@20 148 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Diameter (cont'd) MAXIMUM RATINGS Nominal Tube Spedral Number Viewing Cage RCA Outline, Supply Average Diameter Response' of Configu- Struc- Type Basing Voltage Anode Inches Siages rationb ture c No .• Diagram (E) Current V mA I)'" 10 H V 2063 29 2000 <..U 10 H C 5819 26 1250 0.75 10 H C 6342A 30 1500 2.0 ~Hn (5:11) 10 H C 6655A 30 1250 0.75 (Cont'd) 10 H 7746 31 2500 2.0 10 H V 8053 29 2000 2.0 12 H 7850 25 2600 2.0 14 H 6810A 32 2400 2.0 14 H 7264 32 2400 2.0 108 (S-13) 10 H C 6903 30 1250 0.75 10 H V 4463 29 2500 1.0 10 H 7326 33 2400 1.0 110 (S-20) 12 H 4459 25 2800 1.0 14 H 7265 32 3000 1.0 2 12 H C31000A 34 3000 1.0 (Cont'd) 113 12 H C31000B 34 3000 1.0 10 H V C70109E 29 2500 1.0 114 14 H I C7268 32 3000 1.0 10 H C 4518 30 2000 0.5 115 10 H V 4523 29 2500 0.5 12 H 8575 34 3000 0.2 116 12 H 8850 34 3000 0.2 12 H 8851 34 3000 0.2 12 H C31000E 34 2500 1.0 119 12 H C31000F 34 2500 1.0 10 H V 8664 35 2000 2.0 120 10 H V 8664jVl 35 2000 2.0 126 5 H C31024 36 6000 0.5 130 10 H C C7164R 30 1500 0.5 137 11 H I C31000K 37 2000 0.1 10 H V 2064B 38 2000 2.0 3 107 (S-I1) 10 H V 8054 38 2000 2.0 110 (S-20) 10 H V 4464 38 2500 1.0 Technical Data 149 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22° C Sensitivity Anode Dark Nominal Supply Radiant Luminous CUllent Current nA@ Tube Voltace Cathode Anode Cathode Anode Amplification Anode Luminous lIiameter V mA/W A/W uA/lm A/1m (Appro •. ) Sensitivity A/1m Inches 1500 56 70 1000 40 80,000 50 100 2 X 106 6@20 1250 64 25,000 80 31 3.9 X 105 4@20 1000 61 96,000 76 120 1.6 X 106 6@20 1500 56 100,000 70 130 1.8 X 106 16@20 1500 56 34,000 70 42 6 X 105 4@9 1800 56 510,000 70 640 9.1 X 106 64@ 160 2000 56 3,000,000 70 3800 5.4 X 107 1000@2000 2000 56 3,400,000 70 4200 6.1 X 107 1000@2000 1000 48 72,000 60 90 1.5 X 106 10@20 2000 68 11,000 160 25 1.6 X 105 4.8@ 12 1800 64 38,000 150 88 5.9 X 105 3@20 2300 64 430,000 150 1000 6.6 X 106 30@300 2400 64 3,000,000 150 7200 4.8 X 107 50@ 1000 2000 77 270,000 200 700 3.5 X 106 5@200 2 2000 77 270,000 200 700 3.5 X 106 5@200 (Cont'd) 1500 68 3400 160 8 5 X 1()4 0.2@6 2400 64 2,200,000 150 5200 3.4 X 107 50@ 1000 1500 79 39,000 67 33 5 X 105 0.24@ 7 1500 71 32,000 60 27 4.5 X 105 0.5@ 13 2000 97 970,000 85 850 1 X 107 1@200 2000 97 710,000 85 620 7.3 X 106 0.6@2oo 2000 97 710,000 85 620 7.3 X 106 0.6@200 1500 45 18,000 250 100 4 X 105 10@30 1500 45 18,000 250 100 4 X 105 10@30 1500 69 18,000 67 17 2.6 X 105 1 @ 7.5 1500 69 18,000 67 17 2.6 X 105 1 @ 7.5 3000 82 110,000 72 93 1.3 X 106 24@85 1250 40 20,000 200 100 5 X 105 30@ 150 1500 35 21,000 330 200 6 X 105 6@100 1500 64 80 1500 64 35,000 80 43 5.4 X 105 4@9 3 2000 68 11,000 160 25 1.6 X 105 4.8@12 150 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES For Photomultiplier Tubes by Diameter (cont'd) MAXIMUM RATINGS Nominal Tube SlIIefr.1 Number Viewinl C.gI RCA OUtline, Supply AVlral1 Diamat., R"ponSf'> 01 Confilu, Struc· TYIII Baling Voltall AnCHIe Incbas Sta,ls ration b turl' No .• Di'lI'am (E) Currlnt V mA 112 14 S C70045C 39 6000 1.0 3 115 10 H V 4524 38 2500 0.5 (Cont'd) 117 14 S C70045D 39 6000 1.0 122 10 H V 4521 40 2000 0.5 10 H V 2065 41 2000 2.0 107 (S-l1) 10 H V 8055 41 2000 2.0 110 (S-20) 10 H V 4465 41 2500 1.0 5 10 H V 4525 41 2500 0.5 115 10 H V C31027 42 2000 0.5 12 H V C31029 43 2500 0.5 118 14 H 4522 44 3000 0.5 14 H C70133B 44 3000 0.5 • See bar chart of Spectral response on page 156. • Type numbers with prefix C are developmental b Viewing configurations: H, head on; S, side on; types. Each of these C numbers identifies a par· and DW, dormer window. ticular laboratory tube design but the number and • Cage structures: C, circular cage; I, in line; and the identifying data are subject to change. No V, venetian blind. obligations are assumed as to future manufacture d With blue light source. Tungsten light source at unless otherwise arranged. color temperature of 2870oK. Corning C.S. No.5-58, 1/2 stock thickness. PRELIMINARY SELECTION GUIDES Ruggedized Photomultiplier Tubes by Size Numiul Viewing Tube SfllClrll Conn,u· Call RCA Mllillry DilI11at1r RIIfIOlISI& rationb Structur.' Type SlIICIftcationd IIICbts Number S-4 S C 8571 1/2 (102) S·5 S C C70129H (104) S-l H C70102B MIL-E-5272C 3/4 (101) Technical Data 151 TYPICAL CHARACTERISTICS AT THE SPECIFIED SUPPLY VOLTAGE AND 22° C Sensltiwtlyl ARodl Dirk NOIIiIllI Su"., Radiant luminous Current ClllrlftlaA@ TIIII Vllla.e Cathodl Anode Cathodl Anode A=catiOll Anodl Luminous Dllllllltar V mA/W A/W uA/lm A/1m ( 01.) Sinsitiwtly A/1m IIICbIs 5000 60 140 S X 106 SOO@1OOO lS00 71 32,000 60 27 4.5 X 105 1 @13 3 SOOO 72 1 X 107 1000@ 10,000 (Cont'd) 1500 87 19,000 83 18 2.2 X 105 [email protected] 1500 88 110 1500 88 35,000 110 44 4 X 105 4@9 2000 68 11,000 160 2S 1.6 X 105 4.8@ 12 lS00 80 32,000 67 27 4 X 105 1.5@ 13 5 lS00 88 13,000 77 11.S 1.S X 105 [email protected] A/incident Imd 17S0 88 130,000 77 11.5 1.5 X 106 20@9 A/incident Imd 2000 88 2,600,000 77 2300 3 X 1Q1 60@2ooo 2000 88 2,600,000 77 2300 3 X 107 60@2oo0 ENVIRONMENTAl TESTING" Nominal QuaHIy Tube Coa'ormance Shock Vibration. Acceleratioa Temperatura Diamatar Inspection f Cydinl Inches Design 30 g 20 g 15 g -45+7SoC 11 ms S-2OO0 Hz 1-1/2 hrs 5 min 8 hrs 1/2 Design 30 g 20 g lS g -4S+7SoC 11 ms 5-2000 Hz 1-1/2 hrs 5 min 8 hrs 100% 30 g 20 g 11 ms 20-2000 Hz 15 min 3/4 Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 152 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES Ruggedized Photomultiplier Tubes by Size (cont'd) Nominal Tube Spectral Vlewinc Cap RCA MlRlary Diameter Respoll$la COnnIU Structure C Ty" Spaciftcationd Incbn rationb Number S-l1 H 4460 MIL-E-5272C (107) S-l1 H C70102E MIL-E-5272C (107) 3/4 (Cont'd) 115 H C70l02M MIL-E-5272C 124 H C70l02N MIL-E-5272C 115 H C C31016A MIL-STD 810 B 1 115 H C C31016B MIL-STD 810 B S-l1 H C 4441A MIL-E-5272C (107) S-l1 H C 4461 MIL-E-5272C (107) S-l1 H C C7151N MIL-E-5272C 1·1/2 (107) 115 H C C7151Q MIL-E-5272C S-20 H C C70114C MIL-E-5272C (110) S-1 H C C70114D MIL-E-5272C (101) Technical Data 153 ENVIRONMENTAL TESTING" Nominal Quality Tube Conformanet Shack V1~radon& Aceellr.don Tlmperaturl Dlamlter Inspectlonf CYellnl Inches 100% 30 g 20 g 11 ms 20-2000 Hz 15 min Design 20 g 100 g 20-2000 Hz 6 hrs 1 min 100% 30 g 20 g 11 ms 20-2000 Hz 15 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 3/4 100% 30 g 20 g (Cont'd) 11 ms 20-2000 Hz 15 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 100% 30 g 20 g 11 ms 20-2000 Hz 15 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min Design 75 g 20.7 g (rms) 100 g 11 ms 50-2000 Hz 1-1/2 hrs Design 75 g 20.7 g (rms) 100 g 1 11 ms 50-2000 Hz 1-1/2 hrs 100% 30 g 20 g 11 ms 20-2000 Hz 15 min Design 20 g 100 g 20-2000 Hz 6 hrs 1 min 100% 30 g 20 g • 11 ms 20-2000 Hz 15 min Design 20 g 100g 20-2000 Hz 6 hrs 1 min 100% 20 g 20-2000 Hz 15 min 1-1/2 Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min Design 30 g 20 g 100g 11 ms 20-2000 Hz 6 hrs 1 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 100% 30 g 20 g 11 ms 20-2000 Hz 15 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 154 RCA Photomultiplier Manual PRELIMINARY SELECTION GUIDES Ruggedized Photomultiplier Tubes by Size (cont'd) NDndul VltwtRI TuH SIIICIrII CDnfilU. Cala RCA Mifitlry d Olllllllr Responsa" ratilllb Structure· T", $pedftcatilll Inches NumNr 112 H C C70114E Mll·E·5272C 115 H C C70114F Mll·E·5272C 1-1/2 115 H C C70132A Mll·E·5272C (Cont'd) 117 H C C70114J Mll·E·5272C S·I1 H C C70132B MIl·E·5272C (107) 120 H V 8664 2 120 H V 8664/V1 • See bar chart of spectral response on page 156. e For detailed information on environmental testing, b Viewing configurations: H, head on; and S, side on. request a technical data sheet on the specific type. e Cage structures: C, circular cage; I, in line; and t Quality Conformance Inspection: 100%, every tube V, venetian blind. tested; sample, some tubes tested from each lot; d Military Specifications: and Design initial tubes only are te-sted. MIL·E·5272C, 13 April 1959, Amendment 1, 5 g Vibration: Cycling ranges from minimum; to maxi· January 1960; mum to minimum; time is total time for vibration MIL·STD-SIOB, 15 June 1967 in three axes (equal time for each axis). Technical Data 155 ENVIRONMENTAL TESTING" Nominal QuaUty Tulll CGnlormanCe Shock VllJrlHon g Acceleration Temperature Diameter Inspection! Cyetinl inches DesIgn 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min Design 30 g 20 g 100 g 1-1/2 11 ms 20-2000 Hz 6 hrs 1 min (Cont'd) Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min 100% 20 g 20-2000 Hz 15 min Design 30 g 20 g 100 g 11 ms 20-2000 Hz 6 hrs 1 min Sample 150 g 60 g 11 ms 48-3000 Hz 15 min 1500 g 50 ms Design 60 g 150 g 2 48-3000 Hz 6 hr 6 min Sample 150 g 60 g 11 ms 48-3000 Hz 15 min 1500 g 50 ms Design 60 g 150 g 48-3000 Hz 6 hr 6 hr 156 RCA Photomultiplier Manual SPECTRAL RESPONSE AND RELATED PHOTOCATHODE CHARACTERISTICS SPECTRAL RESPONSE YlgEYL ~PECTRAL RANGE TYPICAL OATA e- .. ·.· .. ·.~· .. :·.i •••••• , h ••••• W·<~··!~···~ ... RCA EIA Cathode Window Range to Rangl to Typical Ranga to Ranll to J Approx. Coda Desig· Matarial Materiai Approxl% Approxl0% Maximum Approxl0% APprOX1~ k S u QE No. natio~ Point Point Response Point Point Im/W "A/1m mA/W % Lime or 101 5-1 Ag·O·Cs l· • .. '-' ... ••.... 1(.. • 94 30 2.8 0.4 Borosil. I Lime or f:I.•• ] 1[" • ~. ;:;::::] to 1180 102 5-4 Cs·Sb Borosil. .. . 1036 40 40 12.4 103 - Cs·Sb UV Glass n·.' • ••J 11"'" ..... 1'£::::::::: 800 60 48 13.2 104 5-5 Cs·Sb UV Glass tl' • ...... t ••= . .. .~. :::::a 1253 40 50 18.2 Lime or ::::::::::::) 105 5-8 Cs-Bi Borosil. 101 ~~ ••• ••• .1:':·:::: 754 3 2.3 0.8 Ag·Bi· Lime or Ill·... r.l L·, ,'.'.', 508 40 20 5.5 106 5-10 O·Cs Borosil. ••• ••-.t: lime or 107 5-11 Cs-Sb Borosil. 1Jr.". [l r. ... ~l::] 803 70 56 15.7 Fused 13.5 108 5-13 Cs·Sb Silica :':':1 794 60 48 5-19 Fused 1593 24.4 109 Cs-Sb Silica • 10·.- .... .-.. ij:·:·:1 40 65 110 5-20 Na·K· Lime or .'1:::::::: ::1 429 150 18.8 Cs·Sb Borosil. ::'1·' 10'• ••• ••• 64 Na·K· Lime or ...:::::::: :::1 295 300 89 20.8 m - Cs·Sb Borosil. It-· • ... 'I to' ... Na·K· 112 UV Glass L·· '.'::::::::::: 428 180 77 22.7 - Cs·Sb ...... 113 Na·K- Pyrex ••J ••,:! 385 200 77 23.9 - Cs·Sb tt· r.· • .,',' " 1',' ::iil 114 Na·K· Fused 428 140 60 17.7 - Cs-Sb Silica r...... ,•...... 'I""" .",:1 115 lime or ::,1 1190 67 79 24.5 - K·Cs·Sb Borosil. '.".' 1,-•• ..... 116 - K·Cs·Sb Pyrex I,,·...... loI:i:1 1140 85 97 31.2 117 - K·Cs·Sb UV Glass tir:. '::01 1r •• ... !'fiil 1190 67 79 24.4 ' - 118 K·Cs·Sb UV Glass iiI· • It ~l Ir::· ...... rlml to 9~0 1140 77 88 27.2 Na·K· Pyrex ::::::::::10' ·.1 [ •• t ...... 180 250 45 9.7 119 - Cs·Sb ... . UV·grade It-•• 1030 60 71 22 120 - K·Cs·Sb Sapphire ... • .1 II••• .',', loIf~t Fused - 11.6 7.2 121 - Cs·Te Silica [ ,-'l-. :[.11 - 200 400 600 800 WAVELENGTH - NANOMETERS Tee h meal . Data 15 7 SPEClKAL ~ESPONSE USEFUL SPECTRAL RANGE TYPICAL DATA ...... CA EIA Cathode Window Range to Range to Typical Rang. to Rang. to ApPlox. ode Desig· Material Mat9rial Approxl% ApPloxl0% Maximum APPloxl0% Approxl% k S q QE No. nation Point Point Response Point Point Im/W IlA/lm mA/W % Aluminum • .-4 ~Jj;!] L22 - K-Cs-Sb Oxide Lr...... 1040 83 87 26.9 UV-grade Cs-Sb ~lj;;! 123 - SapphIre r.· ... '.".' ,-.-. 800 60 48 13.2 124 - Cs-Sb UV Glass .. ... 11 r. r,".·... 'fl"J) 803 90 72 21.2 Lithium Cs-Te Mr. 01,1 125 - Fluoride ... - - 11.5 6.1 Lime or K-Cs-Sb ot::::::1 1140 72 27.4 126 - Borosil. '~4 [...... 82 Ag-Bi- 127 UV Glass ... •• !I r·-... ~;; ,,!,;!il'to 910 410 60 25 6.2 O-Cs 10=· • - .- . -- 128 -- Sa-'As UV Glass ••• ...... •••••••}i 133 250 33 7 129 Ga-As-P UV Glass •• ., • ••• ...... 280 200 56 17 130 Na-K- Lime or ...... 10""1,1,' 203 200 40 9 - Cs-Sb Borosil...... ••• ••• Na-K- Lime or .I. 180 250 45 9.7 131 - Cs-Sb Borosil. ,,' k = Conversion Factor (J = Radiant Sensitivity at Wavelength of Maximum S = luminous Sensitivity Response QE = Quantum Efficiency at Wavelength of Maximum Response 158 RCA Photomultiplier Manual ELECTRON MULTIPLIER STRUCTURES lectron multiplier structures are • For Particle and Radiation Detec E identical to those used in photo tion multiplier tubes. Electron mUltipliers • For Use in a Vacuum System of are intended for use in vacuum sys 10-5 Torr, or Lower tems in the detection and measure • Broad Selection of Mechanical ment of electrons, ions and other and Electrical Characteristics charged particles, as well as X-radia • Various Types Feature an Integral tion and vacuum ultraviolet radiation. Voltage Divider The maximum average anode current • High Stability Copper-Beryllium (30 second average) for all struc Dynodes tures listed in the chart below is 10 microamperes. ,!!!;,1 TYPICAL VALUES ....t •••I- Velllle, RCA tlplier Clle N_lir Radiatie. laode t. DeY. Tulle Slncture If Outer Dp •• II, D,lade rype Wlt_ (CuBe D,I.dls Structure Ilciei NI.l, NI. SI.lIar D,,"es) v.ita Optde (equal ,alts Structure per sllle) Flexible leads, 2000 C1115D 931A Circular 10 sealed in .31 X .94 to 1 X 10" Cage bulb 3000 Glass stem, (20) flexible .375 ± .010 X 3000 C11I1J 6810A In-Line 14 leads (820-102 .375 ± .010 to 1 to 10 X HI" base supplied) 4000 Flange, glass stem, (20) 3000 Cll81K 6810A In-Line 14 flexible leads .375 ± .010 X to 1 to 10 X 10" (820-102 .375 ± .010 4000 base suppl i ed) Stacked ceramic and Kovar Venetian construction, 3000 C31011 8664 14 ring terminals, .900 X .730 to 1 to 10 X 10" Blind Flange mounting, 4000 shipped with sealed flange 2500 C3l 011 A 8664 Venetian 10 Same as .900 X .730 to 1 to 10 X H)" Blind C31017 above 3500 Stacked ceramic and non-magnetic material construc 2500 C3l0118 8664 Venetian 10 .900 X .730 to ItolOxHf Blind tion, ring termi 3500 nals, flange mounting Technical Data 159 RCA TYPICAL VALUES P,.t•• II- I RCA V.llII', t~lilr CII' N•• hr Rldi.ti •• ' ••~e t. C.rr ••t DIY. ..1 Structlr. .f Oller O,.lil, '.,.liflC. TJ,e Wit' (C.Be D,.odes StrlCtlre IIC'IS D~"~'•. 1, t •• It N•. SIMiI.r o, •• des) V.lts TJlic.1 D,.. ~e (1111.1 fins C•• itl ••s StrlCt.r. ,.r StIli) Stacked ceramic Venetian and Kovar con- 2500 &31117& 8664 Blind 10 struction, ring .900 X .730 to 1 to 10 X HI" termina Is, flange 3500 mounting Kovar flange, glass stem, 3900 &31111 4460 In-line 14 (16) flexible .250 ± .005 D. to 5 X 10" leads, shipped 4500 in plastic bag Glass stem, (16) flexible 3900 C31110 4460 In-Line 14 leads, unsealed .250 ± .005 D. to 5 X 10" in bulb 4500 Support brackets, integral divider, 3900 C311118 4460 In-Line 14 (3) flexible .250 ± .OOS D. to 5 X HI" leads, shipped 4500 in plastic bag Integral divider, 3000 1&31111C 4460 In-line 14 shipped in .250 ± .005 D. to 1 X HI" plastic bag 3500 Kovar flange, glass stem, 3400 &31121 4460 In-line 12 (14) flexible .250 ± .005 D. to 2.5 X 10" leads, shipped 4000 in plastic bag Glass stem, (14) flexible leads, 3400 C31121A 4460 In-line 12 shipped unsealed .250 ± .005 D. to 2.5 X 10" in bulb 4000 Glass stem, (14) flexible 3000 &71112F 4460 In-Line 10 .250 ± .005 D. to 1 X l(f leads, shipped 3500 sealed in bulb Kovar flange, glass stem, 3000 &111121 4460 In-line 10 (12) flexible .250 ± .005 D. to 1 X lQ-' leads, shipped 3500 in plastic bag 160 RCA Photomultiplier Manua ELECTRON MULTIPLIER STRUCTURES (cont'd) RCA TYPICAL VALUES Phutomul· voltage, RCA tiplier Cage Number Radiation Anode to CurreRt Dey. Tube Structure of outer Opening Dynode Amplifie. Type With (CuBe Dynodes Structure Inches No.1, tion at No. Similar Dynodes) Volts T:liCiI Dynode (eq~al YDlts Co itions Structure per stage) Glass stem, 3000 CI0102K 4460 In·line (14) flexible 10 leads, shipped .250 ± .005 O. to 1 to 5 X 10' unsealed in bulb 3500 Venetian (16) rod termi· 3000 C70120E 8053 Blind 14 nals, shipped in .800 ± .010 O. to 1 to 10 X 10" plastic bag 4000 Glass stem, 2000 C701290 8571 Circular 10 (12) flexible Cage leads, shipped .06 X .375 to 1 X 10' unsealed in bulb 3000 (13) tab termi- 3000 C70131 7850 In·Line 12 nals, shipped .375 ± .005 o. to 2.5 X 10" in plastic bag 4000 161 Outlines HIS section shows the dimen lection Guides in the Technical Data Tsional outlines for RCA photo Section for outlines and basing ref multipliers. See the Preliminary Se- erence numbers. MAX,DlA:.52~ 20 ~:;j.~VIOLET~ f!.., r-- JTl. SAPl'HIRE .... I I WINDCM -~ ~ ·1 - .r: L_lQr ;f. -.05 11 SEIIIFUXI8L£ LEAOS /- J)20DIA.I,'LONG .25 A. 1.46 -3- .3"OlIN. MAX. PHOTO- / CATHODE INSULATING _--,--r-""T1 PR~~~~IVE , 1.5 L GLASS ENV[l.OP£ MIN. L-..- It GOLD-PLATED 1_1 ~~F~~A~EAOS -z- 162 RCA Photomultiplier Manual T6 BULB f 1.0 MAX. PERMANENTJEDEC No. [12-72 BASE AND PROTECTiVE SHELL =tJ ~ "A~~OI~ _~_c iJ t- 12 SEMI FLEXiBLE LEADS C31005 C31005B 4460 4516 8644 A-3.7 Max. 3.94 Max. C70102B C70042D 7767 8-326 +.06 C701D2E C70042J C70042K . -.12 3.50 !:~~ C7Dl02M C70042R C-0.87 Max. Dia. 0.755 Max. Dia. C701D2N D-0.5 Max. A-3.38 Max. 194 Max. 3.8 Max. E- 0.38 Max. 8-2.94 !:~~ 3.50 !:~~ 3.26 ± 0.15 ·6· C-0.30 Max. 0.30 Max. 0.28 Max. D-0.38 Max. 0.38 Max. 0.30 Max. ·4· .685 R. T6 BULB SMALL-BUTTON NJNAR C)-PIN BASE JEOE:C N2 £:9-37 .40 MAX. TL 14 SEMIFLEXteLE LEADS ~ - 1.5 MIN. LENGTH ·5· ·7· Outlines 163 l-- 1.00 •.06 I DIA. 1 75 MIN '15["-" iii I .-PHeTOCATHODE DlA. ~c----l ~ I I PHOTOCATHODE: PH0TCCATHODE - J I ! i I 4.50:1:.05 MAGNETIC SHIELD '-rr--,-..,---'.--t .6~AX. ~ i4PLATED SEM!FLEXI8lE LEADS .02810 004 DIA. MIN. i,.Ef'oGTH a 1.5 ' -10 - 24 MIN. .OS I OIA·'--!it-"m-- -8- II~t:.4X r- BULB TO I=- .31 MIN.. PHOTO CATHODE ..... ~~ :t.os 01A.1.00~ l ... rDlA.~ __ ...... 'ACEPLATE rr= BASE I ....• iIi !.._?-__ i I -~.'--~ PHOTOCATHODE· .../. - I I J1.31MA~ OIA~Xl= I.MIIA)(. 1'l1 1P22 1P21 931A 4411 4412 1P21/V1 .20 .... X) 4413 1'l1/V1 C311104 C1D15J FOCUSING 13 GOLD-PLATED -ELECTRODE SEMiFLEXIBLE LEADS 1'leA LEAD .016 .t.OO4 OIA. ~ LEAD LENGTH 1.5 MIN. A- 1.94 ± 0.09 1.99 ± 0.09 - 9- -II - 164 RCA Photomultiplier Manual -12 - -15 - BULB T9 PHOTO CATHODE .200 II fLEX!BLE LEADSMAX • .020~:gg; OIA. -13 - ~~mB--~------C310258 C31025G C31025C -- A-0.31 Min. 8-0.94 Min 0.60.2 MinMin. -16 - 5.12 T9 BUI..B ~ID -14 - -17- Outlines 165 1.50.t.06 I 56D~~X'~ 01A'1 .81 ± .06 O. R. 124 MIN , 1.24 MIN OIA. ~I [[rOIl. ;:\_ -FACEPlATE i-I - FACEPlATE -~Ir--Tl V I I 2.44 %.06 ET---ll 280 TI2 BULB 3.35 Tl2 BULB 3." 457 7' METAL tI=t=~==~ FLANGE ~FFJ.l=dd 1.56 MAX. OIA.---j I _, __ ~_JJ 'U 12 SEMI FLEXIBLE !-! l--.30 MAX. olA LEAOS ~ JIJ ,020.:6:.00~ OIA. MIN. LENGTH -us BASE -20- 12 FLEXIBLE- JEDEC No. 812 .. 43 LEADS 1.5 ~ mOO.OOODIA. Lij -18 - k1;52;r:~:.1l FACEPLATE_~_:~'__ J n PHOTOCATHODEJ BASE 12 FLEXIBLE- 2.BI JEDEC No. 812-186 !-tADS 1!5 n .0200.00001A. Lij TI2 BULB J±'06 3.18 MAX. - 21 - ~ SEMI FLEXI8LE 1.5 r-II-M~t - 12 LEADS MIN. .020 t.OO5 'Ui _ DIA. 1.56 MAX. DIA. -19 - -22- 166 RCA Photomultiplier Manual ·23· TIZ BULB _t ~EIl£C .0.812-43 ·24· ·25· Outlines 167 200DIA. ~.o6 ~ 1.68 MIN. FACEPLAlE- -r-= DlA. I ~~- "F ----" 1 4.38 :1:.25 TIS BUlB BASE .EDEC No.814-38 '---or--T+r--'IJ 13 FLEXIBLE LEAOS 1.25 MAX:• .020:J::.0050IA. i -26- - 28- - 27- - 29- 168 RCA Photomultiplier Manual PH010CATHOD£ 6342A, &&S5A, C71I4R 4518 6903 A-5.2 Max. 5.81 Max. 6·9/16 Max. 8-4.25 ± 0.19 4.87 ± 0.19 5·5/8 ± 3/16 .. MUST BE ADEQUATELY INSULATED. C-2.00 ± 0.06 2.00 ± 0.06 2 ± 5/32 Dia. Dia. Dia. ·32· D-1.68 Min. 1.68 Min. 1·5/8 Min. Dia. Dia. Dia. E-Curved Curved Flat ·30· 2.00'DIA. ±.06'~ [ C~~~~~~~ I • PHOTOCATHOD£- 1.1? ,r t.oe 111 R. 15.19" II J±.'" 6.12' ME:DIUM-~HELL DIHEPTAL 14-PIN BASE "EDEC GROUP 5, No. 814-38 , ~J E2.31'MAX.~DIA. ·31 • ·33· utlines 169 2.10 MAX. 01A. FACEPLATE PHOTOCATHODE 4.6S IIAX. - 34- - 3&- 2.06 ---l rFACEPLATt MAx.OIA. 1/ !TAINLESS ,TEEl r :RVSTAL PHOTOCATHOot- ~ lOLotR t o.I• 4.00 .6~ MAX. "'.'" ....1 MAX. 7 lO.O2 "'.,'" CERAMIC GUAAO RING :lRCU1T ItOARD - 37- 180 MAX. CIA. 2.00 CIA. - 35- 170 RCA Photomultiplier Manua !.ooDIA. t.06 =;1 ~~2.~ MIN. ---;! UI ~rtfR MAX. J24 BULB ON 4464 VONLY ----- 13 FLEXIBLE LEADS .020 '.OO!I 014. 2.5 MIN. 1 -38- ,.75 10.25 COAXIAL FITTING !ANOD£l MATES WITH BNe ua -eec/u OR EQUIVAL.ENT. - 39- Outlines 171 -40- UU.oe DIA.,------11 \-----4.31 MIN.DIA. 1_ -PHO'rnCAI'IIODE 7:11 MAX. 8.8T "'.211 e.s T - 41 - 172 RCA Photomultiplier Manual 5.30~.06DIA. 5,00 .:1:..06 014. r ~ 4.75 MIN. 01A.----j "\ ,60 t--'1.0 L.- 2EXHAUST ~ymBULOTIONS 120"OPART) PHOTOC'~ ~ ,:500 DV, 6.5 MAX. or. L.060 I DY> ~O .1 OYO ~ DyO or. ~,020 3.00 OV7 ova or. t.ISO 0'1'10 0'1'11 I.SO MAX.DIA, DYl2 OOARORING A",OO£ ~2.000lA.--l - 42- ~ 5.30'.06010. I" ~.OOJ:.06 OIA. .--- _ I' Mt~/~A . '1'1 'f-- 1.0 L- /2EXHOUSTTUBULATIONS ~120.APART) PHOTO CATHOOE_ --;t=:tf .300 OY, -< 6.0 SMAX. or. ~O L.060 t 1 oroor. ~ OY' DYG ~.020 2.61 DY7 Dn OY. t.l80 0'1'10 -r l.sO MAX.OIA. GUARD RING ANODE 1-200010.----1 -43- Outlines 173 ·44· 174 Basing Diagrams This section shows the basic dia grams for RCA photomultipliers. See FLEXIBLE ENVELOPE the Preliminary Selection Guides in TERMINAL D SMAll PIN the Technical Data Section for out ENVELOPE RIGID ENVELOPE TERMINAL line and basing reference numbers. ORIENTATION SYMBOL LARGE PIN OTHEFt THAN KEY Temporary bases are given at end KEY of this section. Key to Terminal Connections on Basing Diagrams. 8571 C70129G* C70129H* 4460 4516 7767 C70042J 0 C70102Eo C70102Mo C70102N° ·1· ·2· ·3· ·4· 8&44° C781420° C70042Ko C711128· C718421° ·4· ·4· * Supplied with temporary base A. o Supplied with temporary base B. Basing Diagrams 175 C31016Ae -5- C31005B+ -9- C31016B- -G- C31005+ C70128+ -9- -G- -7- -10 - CATHO E LEAD ~C" 0 A~ZgF;.AM PHOTO."(BLACK) L ANODE RETURN LEAD"!" (RED) -8- + Supplied with temporary base C. e Supplied with temporary base D. 176 RCA Photomultiplier Manual 20611° I I' DIRECTION OF RADIATION ·13· ·18· 4441 4441A 4461A I K DIRECTION Of RADIATION: ·14· ·15- ·16· INTO END OF BULB ·19· C70114CO C70114E" K DIREC710N OF• INCIDENT RADIATiON -17 . ·19· C70114D" C711114Fo C711114Jo 4517 6199 7102 C7151W C70132Bo ·18 - • Supplied with temporary base B. ·19· ·20· Basing Diagrams 177 2067" C70132A 0 C7151QO ·20· ·22· 4438 4439 0 4526 0 ·21· ·23· 4440 ·21 • ·24· C7151N° ·22· • Supplied with temporary base B. ·25· 178 RCA Photomultiplier Manual 4463 4523 8053 C70109E DY7 DYe DY6 "7 6 DY9 DYs 5 b 9 :I~YIO - 26- - 29- - 27- -30- 20m 2062t - 28- - 31 - 6810A 7264 NC K DIRECTION OF RADIATION INTO (NO OF BULB - 29- -32- *Supplied with temporary base E. Basing Diagrams 179 7265 C7268 ·36· ·32· ·33· ·37· 44&4 4524 8054 DIRECTION OF RADIATION: INTO END OF lUll ·34· ·38· Flalps ,mide electride c••• cti.s ·35· - 38· t Supplied with temporary base E. 180 RCA Photomultiplier Manual 2065t Request technical data sheet for electrode connection information - 39- - 41 - 4521t Flanges provide electrode connections -42 -- 43- -40- 4525 8055 - 41 - -44- 4465 DY-Dynode G -Grid IC - Internal connection (do not use) NC - No connection METAL P -Anode COLLAR Ie (00 NOTUSEl K - Photocathode - 41 - :I Supplied with temporary base E. Basing Diagrams 181 TEMPORARY BASES -A- - 0- - B- - E- DY-Dynode G -Grid IC - Internal connection (do not use) NC - No connection P -Anode K - Photocathode -c - 182 RCA Photomultiplier Manual 183 Index Data, classes of ...... 30 Absorption spectroscopy ...... 100 Densiometry ...... 99 Acceptor level...... 8 Detection Afterpulsing ...... 47 digital method ...... 94 Anodes ...... 24, 25 low-light level...... 93 Atmosphere ...... 36 photon counting ...... 95 Doping ...... 9 Dynamic compression ...... 110 Band gap 7 Dynode ...... 5, 6 Bandwidth, noise ...... 68 box-and-grid ...... 21, 23 Basing ...... 27 Bialkali photocathode ...... 10, 11 Black-body radiation ...... 119 EADCI ...... 43 Box-and-grid multiplier structure ...... 23 Electric field, exterior ...... 34 Electron affinity...... 7 Electron multipliers ...... 21, 26 Cathodes time response ...... 26 bialkali ...... 10, 11 Energy band ...... 7, 8 ERMA ...... 10, 11 Energy sources ...... '" 128 multialkali ...... 11 ENI ...... 43 opaque ...... 10, 19 Environment ...... 35 semitransparent ...... 10, 19 atmospheric ...... 36 trialkali ...... 11 shock ...... ;...... 35 Cerenkov radiation ...... 77, 86 temperature ...... 32 Charge integration ...... 93 vibration ...... 35 Collection efficiency...... 20 Equivalent Noise Input (ENI) ...... 43 Compression, signal ...... 110 Extended Red Multi-Alkali Conduction band ...... 7 (ERMA) photocathode ...... II, 12 Conversion Eye characteristics ...... 112 of units ...... 112 Counting ...... 46, 51 interpretation ...... 89 Fatigue ...... 49, 50 photon ...... 95 Fermi level ...... 7 Current control ...... 55 Fluctuations ...... 56, 64 Current protection ...... 110 Flying leads ...... 27 Current rating Flying-spot scanning ...... 90 anode current ...... 32 CRO ...... 91 cathode current ...... 31 objective lens ...... 92 photomultiplier ...... 92 Forbidden band ...... 7 Dark current ...... 19, 40 equivalent anode dark current input Gas ionization ...... 66 (EADCI) ...... 42 reduction ...... 43 Dark noise ...... 44, 45, 66 Hysteresis ...... 51 184 RCA Photomultiplier Manual applications ...... 77 Integration, charge ...... 93 characteristics ...... 82 current protection ...... , ...... ,...... 11 0 Lmps ...... 120 data ...... 37 Laplace equation ...... 19, 24 design ...... 19 -for photometry ...... 97 Lasers ...... 121 -for photon counting ...... 96 -range finder ...... 89 -for Raman spectroscopy...... 102 Life expectancy ...... 52 -for scanner ...... 92 Light detection ...... 93 sensitivity ...... 37 photon counting ...... 95 Light feedback ...... •...... 67 testing ...... 126 Photon noise ...... 56 Light sources ...... 121 Linearity ...... 53 Power supply ...... 110 high-voltage ...... 110 Logarithmic photometry...... 99 stability ...... 111 Pulse counting ...... 46, 106 Magnetic field ...... 34, 36 Pulse height Matching distribution ...... ,...... 63 calculations ...... 130 resolution ...... 46 -factor ...... 132 source-detector ...... 127 Quantum efficiency...... 40 Multiplier analysis ...... 24 Multipliers contribution to noise ...... 61 Radiant energy ...... 119 for scintillators ...... 82 sources ...... 120 structures ...... 22, 23 Radiation black-body ...... 119 sources ...... 120 Negative electron affinity...... 8 Raman spectroscopy ...... 101 Noise Ratings ...... 30 -and bandwidth ...... 68 anode current ...... 32 dark ...... 66 cathode current ...... 31 gas ionization ...... 66 temperature ...... 32 group I and II sources ...... 66 voltage ...... 30 multiplier contributions ...... 60 Reflected light ...... 54 photocathode contributions ...... 57 Regenerative effects ...... 41 photon ...... 56 Ruggedization ...... 26 statistics for dynodes ...... 61 Scintillation Output-current control...... 55 counter ...... 77, 80 materials ...... 80, 83 mechanism ...... 79 Parameters, design ...... 30 processes ...... 78 Peak-to-vaHey ratio ...... 46 spectrometer ...... 84 Photocathode ...... 6, 19 Secondary emission ...... 5, 13, 61 contribution to noise ...... 57 statistics, effect of ...... 16 opaque ...... 10 time lag ...... 16 reflection at- ...... 54 Sensitivity ...... 9, 37 ruggedized ...... 26 blue ...... 38 semitransparent ...... 10 integral radiant ...... 38 sensitivity ...... 9, 38 loss ...... 50 spectral response ...... 39 luminous ...... 38 transit-time difference ...... 25 radiant ...... 38 Photoemission ...... 5 Shock ...... 35 time lag ...... 16 Signal compression ...... 110 Photometry ...... 97 Source-detector matching ...... 127, 129 color~balancing ...... 98 Spacial uniformity...... 55 densiometry ...... 99 Spectral response ...... 10, 39, 127 logarithmic ...... 99 Spectrometry ...... 100 units ...... 113 absorption ...... 100 Photomultiplier ...... 5, 19, 37 Raman ...... 101 Index 185 pectroscopy Uniformity, spacial...... 55 energy ...... 84 time ...... 87 tability ...... 49, 51 Valence band ...... 7 tatistical fluctuations 56 Venetian-blind multiplier structure ...... 22 Vibration ...... 35 emperature rating ...... 32 Video amplifier ...... 93 ime-domain reflectometry (TDR) ...... 25 Voltage divider ...... 103 Time lag ...... 16 tapered ...... 109 in photoemission ...... 16 -for pulsed operation ...... 106 'in secondary emission ...... 16 -resistors ...... 104 Time response ...... 19, 26, 48 voltage ratio ...... 104 electron-multiplier ...... 26 wiring ...... 107 Time spectroscopy...... 87 Voltage rating ...... 30 Time stability test ...... 52 . Transit time ...... 25 -spread ...... 49 Wiring techniques ...... 107 186 RCA Photomultiplier Manual Contributors 187 LIST OF MAJOR CONTRIBUTORS TO THE RCA PHOTOMULTIPLIER MANUAL Ralph W. Engstrom Fred A. Helvy Harold R. Krall Thomas T. Lewis W. Dean Lindley Ramon U. Martinelli Robert M. Matheson Anthony G. Nekut Dennis E. Persyk Ronald M. Shaffer Albert H. Sommer 188 RCA Photomultiplier Manual 189 Other RCA Technical Manuals • RCA POWER CIRCUIT8-SP-51- • RCA SOLID·STATE HOBBY CIR· (8" x 511.1")-448 pages. Contains design in CUlTS MANUAL-HM-91 (8%" x 5%") formation for a broad range of power -368 pages. Contains complete construc circuits using RCA silicon transistors, rec tion information on 62 circuits of general tifiers, and thyristors (SCR's, triacs, and interest to all experimenters. Circuits use diacs). Gives design criteria and procedures diodes, transistors, SCR's, triacs, MOS for applications involving rectification, sup transistors, integrated circuits, and light ply filtering, power conversion and regula and heat detectors. Circuit operation is tion, ac line-voltage controls, rf power described in detail; construction layouts, amplifiers, and control and low frequency photographs, schematic diagrams, and parts amplifiers. Shows design examples and prac lists are given; and full-size drilling or tical circuits. Price $2.00. *t printed-circuit templates are included for most circuits to simplify construction. • RCA TRANSISTOR, THYSISTOR & Price $1.95. *t DIODE MANUAL-SC-14 (8" x 511.1")- 656 pages. Contains uJ?-to-date definitive • RCA SILICON CONTROLLED REC· data on over '700 seunconductor devices TIFIER EXPERIMENTER'S MANUAL including tunnel diodes, silicon controlled -KM-71 (8%" x 5%")-136 pages. Con rectifiers, varactor diodes, conventional rec tains 24 practical and interesting control tifiers, and many classes of transistors. circuits that can be built with a comple Features easy-to-understand text chapters, ment of active devices available in kit as well as tabular data on RCA discon form. Includes photographs, schematic dia tinued transistors. Contains 38 practical grams, and descriptive writeups. Also in circuits, complete with parts lists, high cludes brief descriptions of solid-state lighting semiconductor- device applica components used (rectifiers, transistors, tions. Price $2.50. *t SCR's) and short section on trouble shooting. Price 95 cents. • RCA HIGH-SPEED, HIGH-VOLT *t AGE, HIGH-CURRENT POWER TRAN· SISTORS-PM-80 (8" x 511.1")-96 pages. • RCA RECEIVING TUBE MANUAL Provides a basic understanding of the -RC-27 (8" x 511.1',)-672 pages. Contains theory and application of the RCA line of technical data on more than 1400 receiv medium-frequency power transistors. Cov ing tubes for home-entertainment use. In ers physical theory, structures, geometries, cludes six easy-to-read text chapters that packaging, critical application-limiting fac provide basic mformation on electron-tube tors, and the operation and requirements of operation, ratings and characteristics, and power transistors in amplifier, switching, applications. Also features a detailed ap and control applications. Typical circuits plication guide for receiving tubes; quick illustrate the use of transistors in series reference charts for replacement and voltage regulators, linear amplifiers, switch discontinued RCA receiving tubes, black ing regulators, and inverters and con and-white and color picture tubes, and verters, and the application of complemen voltage-reference tubes; and a Circuits sec tary transistor pairs. Selection charts are tion that includes schematic diagrams, de included to facilitate choice of the optimum scriptive writeups, and complete parts lists tYee of power transistor for a variety of for 36 practical electron-tube circuits for military, industrial, and commercial appli a wide variety of applications. Price cations. Price $2.00. *t $2.00.*t 190 RCA Photomultiplier Manual • RCA TRANSMI1TING TUBES-TT-5 index adds to the convenience of this man (8%" x 5%")-320 pages. Gives data on ual. Price $2.50. *t over 180 power tubes having plate-input ratings up to 4 kW and on associated rec • RCA ELECTRO·OPTICS HAND· tifier tubes. Provides basic information on BOOK-EOH-I0 (8\12" x 5W')-IS4 generic types, parts and materials, installa pages. Data from the many technical areas tion and application, and interpretation of involved in electro-optics are compiled and data. Contains circuit diagrams for trans unified in this manual. The presentation is mitting and industrial applications. Fea convenient, featuring more than one hun tures lie-flat binding. Price $1.00*t dred tables, charts, and graphs. A coherent system of symbols and definitions is listed, • RADIOTRON° DESIGNER'S HAND· and data on radiometry, blackbody radia BOOK--4th Edition (8%" x 5W')-1500 tion, and photometry are presented. Infor· pages. Comprehensive reference covering mation on radiant sources, lasers, at the design of radio and audio circuits and mospheric transmittance, and detection, equipment. Written for the design engineer, resolution, and recognition are included. student, and experimenters. Contains 1000 Treatments of detector characteristics, illustrations, 2500 references, and cross image and camera tubes, and optics are referenced index of 7000 entries. Edited followed by a section on electro-optics by F. Langford-Smith. Price $7.00.*t systems analysis. References and an ex tensive index add to the usefulness of this • RCA PHOTOMULTIPLIER MANUAL unique handbook. Price $2.50.*t -PT-61 (814" x 514")-192 pages. The construction, operation, and application of • RCA LINEAR INTEGRATED CIR. photomultipliers are detailed in this well CUITS---IC-42 (814" x 5%")-416 pages. illustrated manual. Discussions of photo Explains the basic principles involved in emission, secondary emission, and the the fabrication, design, and application of design of multiplier structures are followed linear integrated circuits. Includes a dis by a description of photomultiplier per cussion of the basic silicon monolithic fab formance characteristics. An analysis of rication process, analyses of the building statistical fluctuation and noise effects then b!oc~ elements !lsed in linear integrated leads into a section on applications of elf CUlts, explanatIons of the operation and photomultipliers. Photometric and radiant applications of RCA linear integrated cir units, radiant energy sources, spectral re ,:uits, and detailed ratings and characteris sponse, and source-detector matching are tics data and package information for set forth. An extensive tabulation of tech RCA linear integrated circuits. Price nical data includes a selection guide for $2.50.*t over one hundred RCA photomultiplier tubes, with a special section on ruggedized ° Trade Mark Reg. U.S. Pat. Off. • Prices shown a"ply in U.S.A. and are types. Outlines, basing diagrams, and cir subject to chanjle Wlthout notice. cuit information are furnished. A complete t Optional pnce. - - :pectral Response Characteristics 191 RCA TYPICAL PHOTOCATHODE SPECTRAL RESPONSE CHARACTERISTICS (cont'd) D~rk Emission Sensi- Quantum at 22° C Trans- Sensi- tivity Efficiency fA/em" Spectral Cathode Window mission o,aque tivity mA/W % (1 fA= I Response Material Material TYge ype /LA/1m @ A max_ @ A max. 1 x 10-1• Al 117 K-Cs-Sb UV X 67 79 24.4 0.02 Glass 118 K-Cs-Sb UV X 77 88 27.2 0.02 Glass 119 Na-K-Cs-Sb Pyrex X 250 45 9.7 120 K-Cs-Sb UV-grade X 60 71 22 Sapphire 121 Cs-Te Fused X 11.6 7.2 Silica 122 K-Cs-Sb Aluminum X 83 87 26.9 Oxide 123 Cs-Sb UV-grade X 60 48 13.2 Sapphire 124 Cs-Sb UV X 90 72 21.2 Glass 125 Cs-Te Lithium X 11.5 6.1 Fluoride 126 K-Cs-Sb Lime or X 72 82 27.4 Borosil. 127 Ag-Bi-Q.Cs UV X 50 34 8.4 Glass 128 Ga-As UV X 200 35 7.9 0.1 Glass 129 Ga-As-P UV X 200 61 16.8 0.01 Glass 130 Na-K-Cs-Sb Lime or X 200 40 9 Borosil. 131 Na-K-Cs-Sb Lime or X 250 45 9.7 Borosil. 132 Na-K-Cs-Sb Lime or X 200 44 10.4 Borosil. 133 K-Cs-Sb Fused X 60 71 22 Silica 134 Ga-As UV-grade X 200 35 7.9 Sapphire 135 Ga-As-P UV-grade X 200 61 16.8 0.01 Sapphire 192 RCA Photomultiplier Manu RCA TYPICAL PHOTOCATHODE SPECTRAL RESPONSE CHARACTERISTICS The Chart of RCA typical photo RCA's complete capability and, ~ cathode spectral response character clarity reasons, only the most co istics given on the inside front and mon are illustrated. back covers of this Manual contains The table below summarizes t curves representing many of the typical characteristics of the phot photomultiplier-tube spectral response cathodes indicated on the spectr characteristics currently available response chart. from RCA. These curves do not show Da~ Emlssi •• Seasl· Quantum at 22° C Trans Sensl tiylty Eftieieley fA/cmz• Spectral CIt~od8 Window mission Opaque tiylty IIIA/W % (1 fA = RespDlSe Materi.1 Material Type Type pA/11II @ A lIIax. @ A mIX. 1 x 1.... '" AI 101 Ag·O..cs lime or X X 30 2·.8 0.43 900 (S·1) BorosH. 102 Cs·Sb li me or X 40 40 12.4 0.2 (S-4) BorosH. 103 Cs·Sb UV X 60 48 13.2 0.3 Glass 104 Cs·Sb UV X 40 50 18.2 0.3 (S·5) Glass 105 Cs·Bi lime or X 3 2.3 0.78 0.13 (S-8) Borosil. 1116 Ag·Bi·O..cs lime or X 40 20 5.5 70 ($-10) Borosil. 107 Cs·Sb lime or X 70 56 15.7 3 (S-11) Borosil. 108 Cs-Sb Fused X 60 48 13.5 4 (S-13) Silica 109 Cs-Sb Fused X 40 65 24.4 0.3 (S-19) Silica 110 Na-K-Cs-Sb lime or X 150 64 18.8 0.4 (S-20) Borosil. 111 Na-K-Cs-Sb lime or Note 1 300 89 20.8 0.4 BorosH. 112 Na-K-Cs-Sb UV X 180 77 22.7 0.4 Glass 113 Na-K-Cs-Sb Pyrex X 200 77 23.9 0.4 114 Na-K-Cs-Sb Fused X 140 60 17.7 0.4 SHica 115 K-Cs-Sb lime or X 67 79 24.5 0.02 Borosil. 116 K-Cs-Sb Pyrex X 85 97 31.2 0.02 Chart C•• tilled en Precedin, Pa,e. • Based on Dark Current Measured in Typical PMT's. Note 1: Reflective Substrate