Extreme ultraviolet quantum detection efficiency of rubidium bromide opaque photocathodes

Extreme ultraviolet quantum detection efficiency ofrubidium bromide opaque photocathodes

Oswald H. W. Siegmund and Geoffrey A. Gaines

We present measurements of the quantum detection efficiency (QDE) of rubidium bromide opaque photo-cathodes over the 44-1560-A wavelength range. We achieved QDEs of >60% at X = 68 A, and >40% at X 920A, for RbBr photocathode layers applied to the surface of microchannel plates (MCPs). The photoelectricthreshold is observed at X 1560 A, and there is a broad (-100-A) QDE minimum centered at X 775 A whichcorrelates with 2X the band gap energy for RbBr. The QDE is characterized by four peaks centered at X 68A, X 400 A, X 600 A, and -1050 A. The QDE peaks at X 400 A, -600 A, and 1050 A correspond withemission of 3,2, and 1 photoelectrons, respectively. The QDE atthe X 68-A peak is associated with a d-fres-onant absorption feature of RbBr. QDE contributions of the photocathode material inside the channels, andon the interchannel web, have been determined. Measurements of the angular variation of the QDE from 0°to 350 to the channel axis are also presented. We describe a simple QDE model and show that its predictionsare in accord with the QDE measurements. Preliminary assessment of the stability of RbBr indicates that noQDE degradation occurs after limited exposure (20 h) to air at low humidity (<30%). Examination of thephotocathode structure with an electron microscope reveals a rough surface with a scale of the order of 0.5 jrm.

1. IntroductionIt has been known'-' 0 for some time that the QDE of

MCPs may be enhanced by the application of a layer ofphotoemissive material to the MCP surface. This lay-er is usually thick enough to stop all photons at thewavelengths of interest and is termed an opaquellphotocathode. The photemission from the front sur-face of the photocathode layer is detected by the MCPas depicted in Fig. 1, rather than the photoemissionfrom the back surface as is the case for transmissionphotocathodes. 1 Materials that are in general use asphotocathodes for the extreme ultraviolet (EUV) in-clude MgF2,2 '7102- 4 and CsI.1-4,7" 0"" 5 MgF2 is astable7"12"1 3 photocathode, but has a rather low QDE.CsI has higher QDE but is unstable3 7"16"1 7 under expo-sure to humid air. Therefore, several other potentialphotocathode materials have been examined5-9 duringthe last few years in search of high efficiency and stablealternatives. Recent investigations of KBr,5 8 CsBr,6

The authors are with University of California, Space SciencesLaboratory, Berkeley, California 94720.

Received 15 December 1989.0003-6935/90/314677-09$02.00/0.© 1990 Optical Society of America.

and KCl9 have demonstrated that other materials canachieve high EUV QDE. Here we have continued ourphotocathode investigations to find efficient cathodematerials, with emphasis on high, stable QDE betweenX=900 and 1200 A.

The photocathode materials previously mentioneddisplay QDE vs wavelength dependence which is quali-tatively predictable using the band gap energy andabsorption coefficient data. RbBr has a band gap +electron affinity of -8.1 eV,' 8 which is close to thevalue for KBr (8.2 eV). Hence we expect many of theQDE characteristics to be qualitatively similar. Onthe basis of the high QDE achieved for KBr between X= 900 and 1200 A, and favorable photoyield15 andabsorption coefficientl51 9- 24 data for RbBr, we havemeasured the QDE of three sample RbBr photocath-ode layers.

The QDE of RbBr coated MCPs was measured usinga back-to-back Z-stack MCP configuration 8 25 in a de-tector with a wedge and strip position sensitive an-ode.2526 This detector is a photon counting device andaccomplishes single photon imaging through a linearposition dependent division of the MCP output signalbetween three anode segments. A stack of three Phil-ips MCPs with a 12.5-,um channel diameter, 15-Amcenter-to-center channel spacing, a 40:1 channellength-to-diameter ratio, and 25-mm effective aper-ture was employed. Prior to photocathode depositionthe MCPs were cleaned using a previously established

1 November 1990 / Vol. 29, No. 31 / APPLIED OPTICS 4677

Retarding grid

, hv

E

MCPWall

Fig. 1. Schematic i]tion showing the phc

grid, and p(

method.4 ,5,7 -9 I

RbBr was accomrotation/s), andtion was set at 1(cant (Q1000-A)the channel wall'for each of the t]

lowed both translational and rotational motion.NIST standard photodiodes were mounted on a swing

(D (D D o oD Eo arm in the chamber to measure the absolute beambhv intensity. Windowless photodiodes (NIST 305, 320,

h v Cathode A1203 photocathodes) provided a calibration standardhv Cathode e- between X = 56 and 1216 A, while a windowed diode

l ayer <;, / (EMR 15688, CsTe photocathode) accommodated the,*S Adrange of X = 1216-2000 A. Monochromatic radiation

was provided by a 0.5-m grazing incidence monochro-mator with two radiation sources, a hollow cathodedischarge lamp and a filament target soft x-ray source.A range of targets (Al, Si, B, Be, C) in the soft x-raysource allowed selection of monochromatic lines be-tween 44 and 171 A. Various gases (Ar, He, H2, CO,Ne, 02) used with the hollow cathode source providedemission lines between 237 and 1560 A. A calibrated2% transmission mesh was inserted between the MCPdetector and the radiation source. This allowed highflux illumination of the standard photodiodes, whileachieving reasonable (104-s-1) event rates for the MCP

lustration of the photocathode test configura- detector.itocathode layer on the top MCP, the retarding The coated MCPs were initially evaluated to deter-)ssible photon interaction geometries. mine their gain and pulse height distribution (PHD)

characteristics. The MCP gain (5 X 106) and countingthreshold (-10% modal gain) were then set to achieve anarrow PHD (<50%) and a detector event rate inde-

Evaporation of high purity (99.99%) pendent of MCP gain. The channel axis orientationplished while the MCP was rotated (1 was established by determination of the rotationalheated to -110'C. The axis of rota- angle at which a significant low gain peak appears in)0 to the channel axis so that a signifi- the PHD. In this situation events are generated at thelayer of cathode material would coat top of the second MCP in the stack as a result ofS.2 The thickness and deposition rate photons being directly transmitted through the chan-hree RbBr photocathodes studied are nels of the first MCP. The EUV beam (-5-mm diam-

shown in Table I.Transportation of the coated MCPs was accom-

plished in a container filled with dry argon to reducethe possibility of cathode degradation due to reactionswith water vapor. Following installation of the MCPsa 90% transmission etched nickel mesh was attachedover the detector aperture. This enabled a retardingfield to be used which deflects photoelectrons emittedfrom the MCP web areas (Fig. 1) back down onto theMCP. The entire detector was subsequently enclosedin a grounded metal cylinder with a 7.5-cm entranceaperture covered by a 90% transmission mesh. Thisminimized the effects of background noise in the vacu-um tank.

11. ODE Measurement TechniqueThe RbBr coated MCP detector was mounted on a

three-axis manipulator stage in a 0.25- X 0.75-m ultra-high vacuum chamber. The manipulation stage al-

eter) was centered on the detector and the axis ofrotation of the detector was positioned at the photo-cathode surface.

The absolute QDE was evaluated from the ratio ofthe incident photon rate, as determined by the stan-dard photodiodes, to the photon rate observed by theMCP detector. The MCP detector event rate wascorrected for the mesh transmission and the electronicdead time. The off-line spectral continuum back-ground was also measured and subtracted for both theMCP detector and the standard photodiodes. Abso-lute QDE measurements were made at a 17.90 grazeangle to the MCP channel axis, with 0- and 118-V/mmretarding fields. Measurements of the angular varia-tion in QDE were accomplished by evaluation of therelative event rate as a function of angle from a -35°graze angle to a +30° graze angle in 5° steps. Thesedata were normalized using the absolute QDE data fora 17.9° graze angle. The accuracy of the absolute QDE

Table I. Deposition Characteristics of RbBr Test Photocathodes

RbBr 1 RbBr 2 RbBr 3

Deposition rate (A/sec) 1 6.5 11.7 9.4

Thickness (A) 1 4,870 1 0,550 11, 310

4678 APPLIED OPTICS / Vol. 29, No. 31 / 1 November 1990

.5

.6

.4

.2

0

Table II. Errors for the Absolute QDE Measurements

X(A) Error (%, 1a)44-68 30

1 1 4 1 71 7 1 5 1

243-490 15.55 1 2 - 1 21 6 1 3

0 500 1000 1500Wavelength ( A )

Fig. 2. Quantum detection efficiency vs wavelength for RbBr testphotocathode 1 at a 17.90 graze angle to the channel axis using a 118-

V/mm retarding field. Errors are described in Table II.

measurements is determined primarily by the photodiode signal-to-noise ratio and the calibration accura-cy of the standard photodiodes. Table II gives the 1sigma errors for the absolute QDE measurements as afunction of wavelength. The errors for the relativeQDE measurements are -+5%.

IW. ODE Measurements

A. ODE vs Wavelength

The measured QDE of RbBr cathodes 1,2, and 3 areshown in Figs. 2, 3, and 4, respectively. The totalQDE, channel QDE, and the inferred web QDE (total- channel) are displayed on each plot. RbBr cathodes1 and 2 were measured from 44 to 1560 A, and RbBrcathode 3 was measured from X = 44 to 1216 A. Thecorrespondence of QDE curves for the three cathodesis good, generally being within the measurement er-rors. The most significant differences occur betweencathodes 1 and 2 between X = 1216 and 1560 A, justbelow the photoelectric threshold. The QDE curvehas considerable structure, with specific features thatare similar to those found in the QDE curves forKBr57'8 and CsI.-4. The RbBr QDE peaks are cen-tered at X 1050, 600, 400, and 68 A, with QDEminima at X 775, 500, and 171 A. The X 1050 Apeak is associated with single photoelectron emission,and the broad minimum that follows at X 775 A

.8

.ec)

WV

Er.

IIS

ca

.6

.4

.2

00 500 1000 1500

Wavelength ( A )Fig. 3. Quantum detection efficiency vs wavelength for RbBr testphotocathode 2 at a 17.9° graze angle to the channel axis using a 118-


V.

c)S-0

C)ar.0

0

.5

.6

.4

.2

00 500 1000 1500

Wavelength ( A )Fig. 4. Quantum detection efficiency vs wavelength for RbBr testphotocathode 3 at a 17.90 graze angle to the channel axis using a 118-


marks the transition to two-photoelectron production.The peak at X 600 A correlates with two-photoelec-tron emission, and the X 500 A minimum indicatesthe onset of three-photoelectron production. KBr58

and CsI 47 have only small dips in QDE at the onset ofthree-photoelectron production compared to RbBr.The broad QDE peak at X 400 A is due to multiplephotoelectron (3-5) production, then the QDE dropsoff as the absorption coefficient declines. In commonwith other2 -8 bromides and iodides, we see the largeQDE peak at short wavelength (X 68 A) as a result ofa resonant absorption feature (3d-f).24 Photoelectricyield data' 5 for RbBr show features similar to our QDEmeasurements, with a peak at X 1000 A, a prominent


.5 -. 5 - - - - - - - -

._

S-

._

ca

0

C)a

.0

V

.6

.4

.2

0-40 -20 0 20 40

Graze Angle to Channel Axis (Degrees)Fig. 5. Quantum detection efficiency vs graze angle to the channelaxis for RbBr test photocathode 3 at X = 68 A using a 118-V/mm

retarding field. Errors are described in Table II.

.

Sm-

04.'C)

4.'

.5

.4

.3

.2

.1

0-40 -20 0 20 40



minimum at X 790 A, and a rapid increase in yieldfrom X 790 to 600 A.

B. QDE vs Graze AngleThe QDE of RbBr cathode 3 as a function of graze

angle for wavelengths of 68, 171, 462, 616, 736, and 988A is shown in Figs. 5, 6, 7, 8, 9, and 10, respectively.The sharp dropoff in channel QDE, observed at allwavelengths, for graze angles approaching 0 is attrib-utable to increasing cathode layer reflectivity. Thewidth of the 0 minimum becomes smaller as the wave-length decreases in accord with the expected reductionin graze angle required for the onset of high reflectiv-ity. This behavior has also been documented 4' 5' 9 for

.

V.2

.W04 j

CDV

V

.'

.4

.3

.2

.1

0-40 -20 0 20 40



C)

S..m-

0

4.'CDV

4.'

.6

.4

.2

0-40 -20 0 20 40



many other cathode materials. The channel QDE de-creases for graze angles greater than 100 at all thewavelengths examined. It is clear that the mean pho-ton interaction depth in the cathode layer must in-crease as the graze angle increases. Thus the channelQDE drop off may be attributed to a decrease of thephotoelectron escape probability with increasing grazeangle. The angle at which the peak QDE occurs istherefore determined by the rate of change of cathodereflectivity and of photoelectron escape probability asa function of graze angle. Note that the rate at whichthe channel QDE decreases is far more rapid for =171 A (Fig. 6) than for any other wavelength. Thiseffect is also associated with a low contribution of the


I I l '1* Total o Channel-

A Web

I'IIIIIIIIII I, I, ,.,, I,,II,

~ II I II II I I I I II I *1 Total -

o Channel-A Web -

_ I I I I I I I I I I I I I I I I I -

* Totalo Channel-A Web

_ 111 iijg111

fi - a - -f -f -f - -f - -f - - - - -

- - - - - - - - -

.8 .5

.C)

-.

.0

Ca)V

E.'

.3

.2

.1

0 II

I I I I -I I I '* I 'Total Io ChannelA Web

I I I I I I I I

-40 -20 0 20 40Graze Angle to Channel Axis (Degrees)

Fig. 9. Quantum detection efficiency vs graze angle to the channelaxis for RbBr test photocathode 3 at X = 736 A using a 118-V/mm


._

S.

r:

0.'

C)

V

.'

.6

.4

.2

0-40 -20 0 20 40



web area to the overall QDE. As we describe later, thiseffect is seen when the photoelectron mean diffusionlength is much smaller than the mean photon absorp-tion depth. In Figs. 5-10 the web QDE varies verylittle as a function of the graze angle. This resultsbecause the radiation always strikes the web areas atlarge graze angles (50-90°).

IV. Interpretation of QDE Results- RbBr has a bandgap (Eg) of 7.4 eV (Ref. 27) [7.7 eV(Ref. 18)], and an electron affinity (Ea) of 0.5 eV (Ref.27) [0.4 eV (Ref. 18)], which infers a photoelectricthreshold between X = 1530 and 1560 A. The observedphotoelectric threshold (X 1560 A) is thus close to theexpected wavelength. The detailed structure of the

QDE curves may be explained by the use of a simplephotocathode model which we have previously devel-oped.589 This model allows calculation of the QDEusing data on the photoelectron energy distribution,linear absorption coefficient, the photoelectron diffu-sion length and escape probability, and the cathodereflectivity as a function of wavelength and graze an-gle.

A. QDE ModelSeveral recent works2 28-32 have dealt with the de-

tailed aspects of the photoemission process for bothopaque and transmission photocathodes. However,the QDE model that we have employed uses simplevariables, such as the mean photoelectron energy andthe mean photoelectron diffusion length, and assumesa thick, smooth cathode layer. We usually 589 find thisto be adequate for the interpretation of QDE trends,and to give reasonable QDE values. Assuming anexponential transport model the probability, P, thatan absorbed photon results in a detected photoelec-tron is given by

P = cpole(1e + d, sin9)OY, (1)

where Pc is the probability that an emitted photoelec-tron is detected by the MCP, pa is the photoelectronescape probability at the photocathode surface, e isthe characteristic photoelectron diffusion length, andd, is the characteristic photon attenuation length inthe cathode material. d = A-', where , (cm-') is thelinear absorption coefficient. 0 is the radiation grazeangle in the cathode material, which is the graze angleof the incident photons in the absence of refraction.

The QDE of either the cathode material in the chan-nels, or the cathode material on the web areas, is thengiven by

QDE = (1 - Rm ) Z n! (-1)r+lpr, (2)

where R is the reflectivity of the photocathode layer, m= 1 for the web areas, and m 2 1 for the channels sincereflected photons may still be detected. Close to thephotoelectric threshold only a single photoelectron isemitted, but at higher photon energies multiple elec-trons are produced. These may include the primaryphotoelectron, Auger electrons, and secondary elec-trons emitted along the track of energetic photoelec-trons. The sum in Eq. (2) describes the probability ofdetecting one or more of these photoelectrons if n is thenumber of photoelectrons emitted. Here we assume n= EIEg, where E is the photon energy.

B. Model ParametersThe reflectivity of RbBr has been examined in sever-

al studies33 -35 for near normal incidence (0 = 70-82.50)over the range of X = 2100-400 A. These show thatRbBr has prominent (-19%) reflectance peaks at X 1530 and 1080 A, and two smaller (-9%) reflectancepeaks at X 720 and 670 A. Other than these peaksthe reflectivity is generally <4%, and is below 2% whenX < 450 A. The value of Pc is probably >0.5 for the


- - a a w w w .

.

- - - - - - - - - - - - - - - - - - -

. 1000

0)._

._

V 100

0.9

-Cf

.)

100 1000Wavelength (A)

Fig. 11. Characteristic photon attenuation length as a function ofwavelength for RbBr: = 41-70 A,22 X = 70-177 A.24 X = 413-1560

A.33

channel area,36 whilepc for the web area is 0 without aretarding field and.>0.5 with a retarding field.

Considerable data are available for the linear ab-sorption coefficient of RbBr 5 9 -2 4 3 3 covering theranges of X 6000-415 A and 180-25 A. Absorptionpeaks occur at X 1530 A, 1080, 720 A withsignificant minima at X 1380, and 850 A. Atwavelengths shorter than 500 A the absorption coeffi-cient drops off rapidly giving a deep minimum at X\180 A before peaking again at X 83 A. The data for ,uhave been used to calculate the values of d, as a func-tion of wavelength, and these are plotted in Fig. 11.

The photoelectron energy distributions for RbBrhave been investigated over the wavelength range from1087 to 717 A,837 and at 8 A.27 The mean photoelec-tron energy gradually rises as the photon wavelengthdecreases from 1216 to 832 A. Bimodal photoelectronenergy spectra have been observed1837 between X 1000 and 717 A. These clearly show the primaryphotoelectron emission increasing in mean energyfrom 3 eV at X = 1025 A to 7 eV at X = 800agreeing well with the hi-Eg value. Secondary elec-tron emission is also seen with a mean energy of -0.7eV.37 Between X = 800 and 760 A the intensity ofsecondary electron emission rises rapidly and begins todominate by X 717 A. This seems to correlate withthe expected transition to two-electron production at2Eg (X 800 A). At higher energies the photoelectronspectrum peaks at -0.4 eV (secondaries) and has ahigh energy tail extending to >5 eV.27

The photoelectron energy data have been used withthe model predictions of Llacer and Garwin30 3 toestimate the values of 1e and pO as a function of wave-length (Table III). p and le for RbBr are not listedexplicitly in this reference. However, the mean freepath of electrons as a function of electron energy is

Table IlIl. Photoelectron Surface Escape Probability (p) andCharacteristic Diffusion Length () vs Wavelength

(A) le PO Ref1216 260 0.79 301152 295 0.77 301087 345 0.72 18,301026 340 0.73 37,30992 365 0.71 18,30976 350 0.72 37,30920 360 0.71 30911 375 0.69 18,30905 360 0.71 37,30861 385 0.68 37,30834 400 0.67 18,37,30800 395 0.67 18,30760 370 0.69 18,30743 260 0.78 30

determined for RbBr. It is also suggested that themean energy loss for electron collisions decreases withincreasing Z. Therefore the po and 1e for RbBr werederived by interpolation between the model predic-tions30 3' for CsI, KCl, NaF, and LiF on the basis ofmean energy loss as a function of Z and of electronmean free path.

Note that there are considerable variations in thevalues for 1e that are derived from different modelingtechniques,2 2>32 so the estimates for RbBr shown inTable III should be treated with caution. Inaccuraciesmay also have been incurred because we used the emit-ted photoelectron energy spectra rather than the ini-tial photoelectron energy as used in the model.3031This would suggest larger values of 1e, particularly.where the emitted photoelectron energy is small.

C. QDE as a Function of WavelengthModel predictions for the channel QDE at a 17.90

graze angle, based on the information in Fig. 11, TableIII, and reflectivity from Ref. 33, are compared withthe QDE measurements as a function of wavelength inFig. 12. We have assumed that Pc = 0.75 and thechannel area is 0.65. Due to lack of data for the photo-electron energy distributions bel6w X - 700 A, and theabsence of good data for 1e and Pc in this region, only alimited wavelength range has been examined. In ac-cord with the data of Ref. 18 we have assumed thatsecondary electron emission does not dominate until X< 760 A. The model is in good agreement for X > 860A, but predicts rather high QDE close to the 2Eg QDEdip.

The rise in QDE from the photoelectric threshold toX = 1216 A is a result of the steadily increasing le as thephotoelectron energy increases. The flat QDE charac-teristic from X = 1216 A to 860 A is the result of a steadyincrease in both le and d,. A small dip in the QDE at X- 1050 A is caused by a peak in the cathode reflectivity.We expect that the reflectivity is greater than thatused in the model since the reflectivity data33 are fornear normal incidence (75°). Thus the model QDEdip should be more prominent. When double photo-


.6C)

a)._

0

cJ

.'

a

.'

a).'

0

C7

.4

.2

00 500 1000 1500

Wavelength ( A )Fig. 12. Model predictions of the quantum detection efficiency forthe MCP channel areas as a function of wavelength for a 17.90 grazeangle to the channel axis, compared to the measured quantumdetection efficiency for RbBr test photocathode 1. Errors are de-

scribed in Table II.

electron emission begins (X -900 A),18,37 the measuredQDEs drop substantially. We expect the value of e todecrease as a result of the drop in the photoelectronemission energy. However, the model predictions donot work well here because the determinations of thesecondary electron spectrum are not consistent, 83 7

and the interpolations for he are probably inaccuratefor low photoelectron energies. It is also possible thatthe value of Pc changes where the electron energy spec-trum changes significantly. The measured position (A

780 A) of the 2Eg QDE minimum agrees with photo-emission data [A - 790 A (Ref. 15)], and with theexpected 2Eg wavelength [- 800 A (Ref. 18)].

The QDE rises to a peak at X 560 A as the photo-electron energy rises for X < 700 A. This is followed bya drop in QDE at the initiation of three-photoelectronproduction at X 530 A (3Eg X = 535 A). Below X -400 A we expect the QDE to be dominated by second-ary electron emission which will have a relatively con-stant energy spectrum. In this regime the QDE will bedominated by the absorption coefficient. Thus themultiple photoelectron peak at X - 350 A drops awayto a QDE minimum at X 170 A as the RbBr absorp-tion coefficient decreases (Fig. 11). RbBr has a char-acteristic resonant absorption peak at X 83 A,24 asomewhat shorter wavelength than either CsI or KBr,which causes a significant QDE peak as seen in Figs. 1-3. The QDE peak observed is quite narrow, having afull width at half maximum of only 70 A (XX, 44-114 A,Fig. 1) and compares well with the absorption coeffi-cient data (AX 70-90 A, Fig. 11).

D. QDE as a Function of Graze AngleWe do not have detailed measurements of the reflec-

tivity as a function of graze angle and wavelength for

Fig. 13. Scanning electron micrograph of the microchannel platecoated with the RbBr three-test photocathode layer showing a gran-

ularity of -0.5 jim.

RbBr. This makes model predictions of QDE as afunction of wavelength difficult for small graze angleswhere the reflectivity is expected to dominate. How-ever, we can examine the behavior of the QDE as afunction of angle at graze angles greater than the peakQDE angle if we assume that the reflectivity in thisregime is small. At X = 988 A we predict a channelQDE of 28.4% for a graze angle of 17.90 which is a littlehigher than the measured values of 25% at 17.90 and26.5% at-17.9°. Our model predicts that the channelQDE will drop to 26% at a 300 graze angle, whichcompares to measurements of 18% at 300 and 23% at-300. Note that the measurements have a definiteasymmetry, which is largely due to the 13° bias of thechannels of the MCP. Both model and measurementspredict a relatively small QDE drop of •6% over therange of angles examined. This situation is character-istic of when le > dc, as is the case for X = 988 A (TableIII, Fig. 11), and is also characterized by a large webcontribution to the total QDE.

At X = 171 A the model predicts a relative drop of-35% for the channel QDE between graze angles of17.90 and 300 assuming that dc = 1800 A (Fig. 11), anddue to the low secondary electron energy27 1e < 100 A,with pO, - 0.8.30 The measurements show a relativechannel QDE drop of 30% for negative angles, and 50%for positive angles, and are thus in reasonable agree-ment with the model given the expected inaccuracy forthe model at short wavelengths. The measured webcontribution to the total QDE at X = 171 A is small,which is a characteristic of the case where 1e << dc, inaccord with our model assumptions. Also note thatthe angular variation of the web QDE is small (<20%relative) for all the wavelengths examined. This is aresult of the large graze angle for the web (50-90°) andcorrelates with a <25% relative QDE variation over theexpected range of graze angles predicted by the model.


V. Structure and StabilityThe structure of the RbBr 3 test photocathodes was

examined using a scanning electron microscope. Fig-ure 13 shows that the photocathode layer is of a granu-lar form, with granule sizes of -0.5 gAm. This type ofsurface structure is similar to the structure of otherphotocathode layers that we have examined78 and is acharacteristic of the deposition process used. Notethat the granular structure results in a far larger effec-tive surface area than for a smooth uniform photocath-ode layer.

Immediately following the initial QDE measure-ments the RbBr three-test photocathode was exposedto air at -30% humidity for 20 h. The QDE was thenremeasured at a number of wavelengths (304,584,671,920, 1066, and 1216 A) and the results were comparedwith the initial measurements. After the period ofexposure, the QDE values were found to be within themeasurement errors of the original QDE measure-ments. RbBr has a water solubility of 98/100-g H20 at200C, compared with 34.7/100 g for KCl, and 65.2/100 gfor KBr. This would indicate that RbBr should be lessstable to moisture attack than KBr. However, weknow that if interactions with water vapor are impor-tant the effects of photocathode surface area andstructure may play a role in determining stability.Issues such as the deposition rate may be important inthis respect, so we plan to investigate the stability of allthree RbBr test photocathodes.

We would like to thank Philips Components forproviding the MCPs used in this study, also T. Gee forfabrication of the photocathode layers, and T. Boothfor SEM micrography. This work was supported byNASA grant NAGW-1290.

References1. C. Martin and S. Bowyer, "Quantum Efficiency of Opaque CsI

Photocathodes with Channel Electron Multiplier Arrays in theExtreme and Far Ultraviolet," Appl. Opt. 21,4206-4207 (1982).

2. G. W. Fraser, M. A. Barstow, J. F. Pearson, M. J. Whiteley, andM. Lewis, "The Soft X-Ray Detection Efficiency of CoatedMicrochannel Plates," Nucl. Instrum. Methods 224, 272-286(1984).

3. M. P. Kowalski, G. G. Fritz, R. G. Cruddace, A. E. Unzicker, andN. Swanson, "Quantum Efficiency of Cesium Iodide Photocath-odes at Soft X-Ray and Extreme Ultraviolet Wavelengths,"Appl. Opt. 25, 2440-2446 (1986).

4. 0. H. W. Siegmund, E. Everman, J. V. Vallerga, S. Labov, J.Bixler, and M. Lampton, "High Quantum Efficiency OpaqueCsI Photocathodes for the Extreme and Far Ultraviolet," Proc.Soc. Photo-Opt. Instrum. Eng. 687, 117-124 (1986).

5. 0. H. W. Siegmund, E. Everman, J. V. Vallerga, J. Sokolowski,and M. Lampton, "Ultraviolet Quantum Detection Efficiency ofPotassium Bromide as an Opaque Photocathode Applied toMicrochannel Plates," Appl. Opt. 26, 3607-3614 (1987).

6. G. W. Fraser, J. F. Pearson, and J. E. Lees, "Caesium BromideX-Ray Photocathodes," Nucl. Instrum. Methods 256, 401-405(1987).

7. 0. H. W. Siegmund, E. Everman, J. V. Vallerga, and M. Lamp-ton, "Extreme Ultraviolet Quantum Efficiency of Opaque AlkaliHalide Photocathodes on Microchannel Plates," Proc. Soc. Pho-to-Opt. Instrum. Eng. 868, 18-24 (1987).

8. 0. H. W. Siegmund, E. Everman. J. V. Vallerga, and M. Lamp-

ton, "Soft X-Ray and Extreme Ultraviolet Quantum DetectionEfficiency of Potassium Bromide Photocathode Layers on Mi-crochannel Plates," Appl. Opt. 27, 1568-1573 (1988).

9. 0. H. W. Siegmund, E. Everman, J. Hull, J. V. Vallerga, and M.Lampton, "Soft X-Ray and Extreme Ultraviolet Quantum De-tection Efficiency of Potassium Chloride Photocathode Layerson Microchannel Plates," Appl. Opt. 27, 4323-4330 (1988).

10. G. W. Fraser, "The Characterisation of Soft X-Ray Photocath-odes in the Wavelength Band 1-300 A, (B) Caesium Iodide andOther Insulators of High Photoelectric Yield," Nucl. Instrum.Methods 206, 265-279 (1983).

11. B. L. Henke, J. P. Knauer, and K. Premaratne, "The Character-ization of X-Ray Photocathodes in the 0.1-10 keV Photon Ener-gy Region," J. Appl. Phys. 52, 1509-1520 (1981).

12. L. B. Lapson and J. G. Timothy, "Use of MgF2 and LiF Photo-cathodes in the Extreme Ultraviolet," Appl. Opt. 12, 388-393(1973).

13. L. B. Lapson and J. G. Timothy, "Channel Electron Multipliers:Detection Efficiencies with Opaque MgF2 Photocathodes atXUV Wavelengths," Appl. Opt. 15, 1218-1221 (1976).

14. G. W. Fraser, "The Characterisation of Soft X-Ray Photocath-odes in the Wavelength Band 1-300 A, (A) Lead Glass, LithiumFluoride and Magnesium Fluoride," Nucl. Instrum. Methods206, 251-263 (1983).

15. P. Metzger, "On the Quantum Efficiences of Twenty AlkaliHalides in the 12-21 eV Region," J. Phys. Chem. Solids 26,1879-1887 (1965).

16. M. J. Whiteley, J. F. Pearson, G. W. Fraser, and M. A. Barstow,"The Stability of CsI Coated Microchannel Plate Array X-RayDetectors," Nucl. Instrum. Methods 224, 287-297 (1984).

17. K. Premaratne, E. R. Dietz, and B. L. Henke, "The Stability ofCesium Iodide X-Ray Photocathodes," Nucl. Instrum. Methods207, 465-467 (1983).

18. H. Onuki, "Photoelectric Emission from Alkali Halides," Sci.Light Tokyo 23, 54-71 (1974).

19. H. Saito, S. Saito, R. Onaka, and B. Ikeo, "Extreme UltravioletAbsorption of Alkali Halides," J. Phys. Soc. Jpn. 24, 1095-1098(1968).

20. K. Teegarden and G. Baldini, "Optical Absorption Spectra ofthe Alkali Halides at 10°K," Phys. Rev. 155, 896-907 (1967).

21. H. Saito et al., "Absorption Spectra of KCI and RbCl in theExtreme Ultraviolet Region," Solid State Commun. 8, 1861-1864 (1970).

22. A. P. Lukirskii and T. M. Zimkina, "Fotoionisation Absorptionin Ionenkristallen," Rontgenspektren und chemische bindung,Karl-Marx Univ., Leipzig 187-193 (1966).

23. J. E. Eby, K. J. Teegarden, and D. B. Dutton, "UltravioletAbsorption of Alkali Halides," Phys. Rev. 116, 1099-1105(1959).

24. M. Cardona, R. Haensel, D. W. Lynch, and B. Sonntag, "OpticalProperties of the Rubidium and Cesium Halides in the ExtremeUltraviolet," Phys. Rev. B 2, 1117-1131 (1970).

25. 0. H. W. Siegmund, R. F. Malina, K. Coburn, and D. Werth-iemer, "Microchannel Plate EUV Detectors for the ExtremeUltraviolet Explorer," IEEE Trans. Nucl. Sci. NS-31, 776-779(1984).

26. C. Martin, P. Jelinsky, M. Lampton, R. F. Malina, and H. 0.Anger, "Wedge and Strip Anodes for Centroid Finding Position-Sensitive Photon and Particle Detectors," Rev. Sci. Instrum. 52,1067-1074 (1981).

27. B. L. Henke, J. Liesegang, and S. D. Smith, "Soft X-Ray In-duced Secondary Electron Emission from Semiconductors andInsulators: Models and Measurements," Phys. Rev. B 19,3004-3021 (1979).

28. B. L. Henke and J. A. Smith, "0.1-10 keV X-Ray InducedElectron Emissions from Solids-Model and Secondary Elec-tron Measurements," J. Appl. Phys. 48, 1852-1866 (1977).


29. B. L. Henke, "Low Energy X-Ray Interactions: Photoioniza-tion, Scattering, Specular and Bragg Reflection," AIP Conf.Proc. 75, 146-155 (1981).

30. J. Llacer and E. L. Garwin, "Electron-Phonon Interaction inAlkali Halides. I. The Transport of Secondary Electrons withEnergies Between 0.25 and 7.5 eV," J. Appl. Phys. 40,2766-2775(1969).

31. J. Llacer and E. L. Garwin, "Electron-Phonon Interaction inAlkali Halides. II. Transmission Secondary Emission fromAlkali Halides," J. Appl. Phys. 40, 2776-2792 (1969).

32. K. I. Grais and A. M. Bastawros, "A Study of Secondary ElectronEmission in Insulators and Semiconductors," J. Appl. Phys. 53,5239-5242 (1972).

33. C. J. Peimann and M. Skibowski, "Dielectric Properties of theRubidium Halide Crystals in the Extreme Ultraviolet up to 30eV," Phys. Status Solidi 46, 655-665 (1971).

34. G. Baldini and B. Bosacchi, "Optical Properties of Alkali HalideCrystals," Phys. Rev. 166, 863-870 (1968).

35. V. Saile, N. Schwentner, M. Skibowski, W. Steinmann, and W.Zierau, "Optical Excitation of the Rb+4p Level in RubidiumHalides at 8K0 ," Phys. Lett. A 46, 245-246 (1973).

36. G. W. Fraser, "The Electron Detection Efficiency of Microchan-nel Plates," Nucl. Instrum. Methods 206, 445-449 (1983).

37. C. S. Inouye and W. Pong, "Ultraviolet Photoelectron Spectra ofRubidium Halides," Phys. Rev. B 15, 2265-2272 (1977).


Documents

Extreme ultraviolet quantum detection efficiency of rubidium bromide opaque photocathodes