»•M*"«N»)»*«MMMfti ««MMW^W-: i <w unrw lim '%'■***,y*&.*'» rf'Ä&tt-'i.W '•'/;■< *S^ •^•'•w**WHi.r: Security Glastfiinciition ÖOCÜMisNT CONTROL DATA • R-^U

GTK tSylvania Tue.

3. RnF'ORI rtTLI

Infrared Photocathode

,> I! (» ft (»ö H t » h "; Ll HI T V C l. A » 011' I C A T IU N

i/MCLASSIFIEP

^ 4 t>IICfUI*TIVK NOTBI (Typ* of re|wM «ml tnptoiln daltu ,, „« ;f,To

Annual Report "ay 1 Iv/l ~ April JU il.'/-i

Nt ■>. AUTHORS fLul nMn«i nMitl*ms< iniilalj

H. Sonnenberg

iAL Xj April 3Ü, 197;

[,/: CO'-'Tf.'CT Or. GRANT NO.

NOOO.U-70-C-0079 h. PROJKCT NO.

7«. TOTAL NO. OP f'ACll",»

20 la. QNiaiNATOM1! HKi'ofiT NUMftllW

None

;b, Nf OP R«P» 14

tb, STHM in;POMT NOW fXny o(hif numb»« l/i«( m«y b* *<i«'<;ri*d CiJ,» laport)

None

10. AVAtUAitUITV/UMITATtON MOTICW

Unliiiii tod

n. sr-oNi.r>niHO MILITAHY ACTIVITY

Advanced Research Piojectc Agency

II. AB»TRACT

7.^-The ,im0lint 0f cs-0 low-V7ork--func.tion-suri:ace material required to

cmtimiKe the phOtpraspöttM of InAsITVP. .depend» upon the wavelength at which the rMponsa is to be piaÄIVea1?' It is shown that the optimum thickness increaae« exponentially with wavelength.

The effect of thick Cs-0 layers on photocraission from GaAs and Ir^oJ^TP/^cathodor, is experimentally investigated. Simple empirical relafcolfhiis between the yield and thickness and between the escape

probability and thickness are derived.

NATIONAL TECHNICAL INFORMATION SERVICE

Spring«»«, V«. 32ISI

I ,., jw*.irr»i.--. r**

DD .^«1473 UNCLASSIFIED

18

Security Clos;sifk-ation

\

BEST AVAILABLE COPY

UNCLASSiirro. Security Classlftcation

1 ORIOINATING ACTIVITY: Enter the name and Bd&MIl of ,ho ^ntractor, subcontractor, erantre, D^^t.r.enl of D- ?ea"P uctivity or other oreani--.Uion (corpotvte m.thor) UMtKg the rcpott. 5« REPORT SKCURITY CLAPSII-TCATION: Enter the over-

ance with appropriate security regiüations. 2b CROUP: Automatic downGradinR tl specified in DoD Di-

• • Vonri in nnd Ar"u-d Farces Imiustrii.I Manual, hmtr rect.ve 52^;J« "nd ^^'-^^ ap.,Ucable, show that optional •'^S SSS.^*« »A 3 S- Coup 4 as author-

i Rl-PORT TITLE: Ent'.-r the complete repoit tit'.e in ßU • . > M m VMmm in •til cases should be unclassifiod.

cap.tal le to« ^ l ^^o U

tn"se^cted wilhout classifica- ^"ÄSlÄÄioj in all capita.« in parenthests itnincdialely lollowirg the title. 4 DESCRlPVl'E NOTES: If appropriate, enter the type Of

covered. 5 AUTHOR(S): Enter the namc(E) of ^'''^^.^.f ^i."^11

IL^S^lTis^^^^^« requ.ement.

««BAcn« riATr- Fnter the date of the report as day, iJfSS "r month";f/ U more than one date appears on tho rcp.-.l, use dale of publication. 7a. TOTAL NUMBER OF PAGES: The ,o.a. paite count should follow normal paKinaUon procedures, ..e., enter nun-hor of pOK»* containin,- inlormat.o^ 7b. NUMIiFK OF REFERENCES Enter the tote, number of referenceu tiled in the repoit. „ t^MmtACT OR GRANT NUMBER: U appropriate, enter K appi'^U numb^f the contract or grant under which .Me ri'poit was written. ofc a,, fc ßrf PROJECT NUMBER: Enter the appropriate

<>a. OR^lNATOR'S «^^^S^i.rrid^iücd

be unitiueto tnio report.

"r 1 "/.« *pun.so,;, al»o enter thi» numbeKs). ,0. AVAILABtL.TY/L.M.TAT10N NOT.CES: ^-V - t

Vationr. on turtlier dissemination ol the report, otner

imposed by security classification, using standard statements such as:

(1)

(2)

(3)

(4)

(8)

'Qualified requesters may obtain copies of this report from DDC" "ForeiKn announcement and dissemination of this report by DDC is not authorized. "U. 3. Governnienl agencies may obtain copies of .his report directly to« DDC. Other quahf.ed DDC uaeis shall rcqUMt through

"U S militurv agencies may obtain copies of this rirTdTrec.iyNro,. DDC Oth« qualified users shall request through

"All distribution of this report is controlled, ified DDC users shall request Inrough

Qu.-.l-

If the rtport has been furnished to the Office of J"*"""1

Series! Department of Commerce for sale to the public, md,- cal^ this fact and enter the price, if known. U. SUPPLEMENTARY NOTES: Use for additional explana-

tory notes. 1? SPOORING MILITARY ACTIVITY; Kntcr the name of It d.?ar montal project office or laboratory «i~^ ('^ .ng (or) the research and development. Include address. 11 ABSTRACT: Enter an abstract giving a brief and factual

po^ UaddUiLal space is required, a continuation sheet shall

be attached. It U hiuhly desirable that the abstract of cUssified reports

be unclasr n , • Each paragraph of the abstract shal end with

^n rndication of the military --"%'?*^ws" IcTo'W formation in the paragraph, represented a? (TS). (S). (C). ( There is no limitation en the lengf.. of the abstract. How-

ever, the suggested length is from 150 to 22b words. 14 KFY WORDS: Key words ore technically meaningful terns or short phTes'tha, characte.i.e a »P»" "< ««*.^ aS

mdex entries for cataloging tho report. Ley '»•^««•IW» sei cted so that no security classif.c »tion is required. Ide""; neVs' such as ecuipment UK-de! der.^ation, ti.u'o name, military nroiccl code name, geogruphu location, may be user; as key Cds but wil" be ortowedbv an indication of technical con- SSTTM •Wl^.nt of linkt, rules, and weight, is optional.

UNCLASSIFIED

"Security Clnssification

INFRARED rilüTUCATUODE

AlWUAI. TECHNICAL REfORT

iH ARl'A Order Nambcr:

Trogram Code Nmibet:

Contractor:

Effective Date:

Expiration Date:

Contract Amount:

Contract Number:

Principal Investigator

Tclephone:

Scientific Officer:

1806

00014

GTE Sylvsnia Inc.

1 November 1969

30 April 1972

$178,686

NOOO] 4-70-00079

Dr. U. Sonnenberg

(415) 966-3472

Dr. Robert E. Behringer

*. view, ana »nolu.10«. f f ^^.^n^Ä^f.lc^oUci«.

U. S. Government.

Prepared by

Sponsored by

Advanced Research Projects Agency

ARPA Order No. 1806

Approved by

D D C n|E(2[EIEQE

MAY 4 1912

fEiSEinrE c -

£ gj^- -ifmJim .1™^;! ' ... Sonnenbcrg Electro-Optics Research and Development Department

H. L. H. Oatcrink', Manager Electro-Optics Research and Development Department

% TECFIJICAL REP0B2 SUMMARY

The technical objective of thio program is the developoant of an efficient

photocathode for tlie 1.5 micron region of the infrared spectrum. No such

photocathode exists today. The concept of the hetcrojunction photocathode pro-

posed by us earlier , represented a new approach to the development of an

infrared cathode.

This approach, exploited on this contracts led to the.discovery of the

(2) most efficient (at that time) 1.06 micron photocathodev '. Further improvements

in this cathode have been made, and even today it remains the most efficient

1.06 micron photocathode yet developed. Continued development of infrared

cathodes on this project has led to photocathodes having, for the first time,

usable response out to 1.3 microns, see Figure 3.

In Section 1.0 of this report we show that the "optimum thickness" of

Cs-0 low-work-function surfaces used in processing heterojunction-type cathodes

is wavelength dependent. A simple empirical relationship describing this de-

pendence is derived. It is shown that the optimum thickness increases ex-

ponentially with wavelength. In Section 2 the effect of thick Cs-0 layeirs

on photoemission from GaAs and InAs0 i?Q - cathodes is given. Simple empirical

relationships between the yield and thickness and between the escape

probability and thickness are derived. It is shown that the increased Cs-0

thickness required to optimize the infrared response of lower bandgap materials

is in large part responsible for their lower photoresponse.

We have abandoned the simple "heterojunction" cathode approach in favor

of an approach which provides for optical absorption in a small bandgap material

which is in contact with a large bandgap material known to have a high electron

escape probability.

U) Hr~SÖnnenberg, Appl. Phys. Letters 14 289 (1969).

(2) Quarterly Management Report H (Feb. 10, 1970); 11. Sonnenberg, Applied Physics Letters 16, 245 (1970).

Section ]

WAVELENGTH DEPENDENCE OF OfflMUM THICKNESS OF Cs-0

LOW-WOkK-FUNCTION HE RFAGES

In the Cs-O processing of a nesative-olectron-aff i.nity-liypc infrared-

pho to cathode one conmioaly finds that the photorcüponse In tlie visible spectrum

peaks earlier than the infrared response. Continued processing, beyond that

required to optimize the visible response generally leads to continued improve-

ment in the infrared response but at the expense of the response in the visible.

Thus for a given infrared cathode, different thicknesses of Cs-0 low-work-function

material are required to optimize the photoresponse at different wavelengths.

We have investigated this behavior in detail for InA*0 ^PQg-CCe-O) and report

here a very simple empirical relationship between optimum thickness and wave-

length.

InAs .P , with a bandgap less than 1 eV was chooscn since its useful 0.4 0.6

photoresponse extends over a broad spectral range to 1.3 microns. The yield

curves for different levels of Cs and 02 exposure were directly recorded with

a phase-sensitive-detection apparatus and a Perkin-Elmer E-l scanning mono-

chromator. A complete scan from 0.45 microns to 1.4 microns takes about 15

minutes.

The photoresponse of the InAs0 ^PQ . cathode with one monolayer of

Cs^ on the surface Weis first recorded. The simultaneous-exposure technique

was than used to process the cathode with Cs and 02 in approximately one-

monolayer-of-Cs steps. The photoresponse at the end of each step was recorded.

To avoid Cs loss from the cathode between proc^ün^ üteps, the Cs ehamiei.

rather than being turned off completely, was turned down .0 that the arrival

rate of Cs atoms at the cathode surface was in equilibrium with the loss rate.

Stability in the photoresponse at each exposure level could easily be maintained

this way.

To investigate the. effect of increasing Cs-0 coverage on the photo-

response of the cathode, the yield was plotted as a function of Cs exposure

for different wavelengths as shown in Figure 1. Each vertical set of points

represents the photoresponse of the cathode at that particular level of Cs and

02 exposure. The first set of points (at 1.0 monolayers) is for Cs only, and

sLequent sets of points are for increasing Cs-0 exposure, measured in terms

of Cs monolayers.

Figure 1 clearly shows that different thicknesses of Cs-0 layers are required

to optlmi^ the photoresponse at different wavelength.. For example, the

photoresponse at 4500X is optimized by a coverage of only about 2.2 monolayers

of Cs whereas the photoresponse at 10.500S is not optimized until the coverage

has grown to approximately 4.2 monolayers of Cs.

The thickness corresponding to the maximum photoresponse (optimum

thickness) at the different wavelengths is plotted in Figur. 2 as a function

of wavelength. The data points from 0.A5 microns to 1.1 microns are taken from

the position of the maxima of the parametric curves of Figure 1. whereas the

points at 1.3 microns are taken from a number of different experiments where

the processing was not interrupted until the response at 1.3 microns was

optimised.

Fiaure 1 Qunuuu. elficioncy a- a function of ^VlSTavSerth^Tpa^ter- thicknoGS measured in monolayerr, ol Cf, with v.ava.-ai3Ll> |

4500 A0

5000 A0

5500 A0

60C0 A0

6500 If

70D0 A"

7500 A0

8000 A0

0.001 1.0 2.0 3.0 4.0

THICKNESS (MONOLAYERS Cs)

5.0

Figure 2 Thickness of Cs-O low-work-function surlacc required to maxials« the pholoresponse oi InA»0 ^g M a function of wavelength.

to O

CO

uu >- < —J- O O

to LU

o

0.2 0.4 0.6 0.8

WAVELENGTH (MICROMETERS)

.1.0 1.2 1.4

it is apparent fron Figure 2 that tha amount, t, of Ca-0 (neaeurad

monolayers of Cs) required to optisvlae the photoreaponse iü exponentlally

■

>-ndcnt on the waveleugth, X (vnicromatcrs) i.e.,

t T e3A CD

For this particular e*p tlBent, 1 - 1.4 monolayer. of Cs(4) and 3 ■ 1-02 micro-

meters"1. We have verified the exponential dependence for c^tlnuousl^ processed

cathodes as well. We find in ßencral, however, that thinner .Cs-0 coveraSe is

required to optiiuize the photoresponse of continuously processed cathodes, as

demonstrated in Figure 2, and that the overall photoresponse of these cathodes

is much better than that of cathodes for which the processing is interrupted.

To underscore this, compare the yield curve shown in Figure 3 to the data at

4.8 monolayers shown in Figure 1. Figure 3 shows the spectral response of

' the cathode represented by the point in Figure 2 at 1.3 microns and 4.9 mono-

layers of CsK '.

OH» may well ask if the exponential relationship given by equation (1)

is applicable only to InAs^P^ or if it is more universally applicable. Al-

though we have not specifically attempted to determine the wavelength dependence

of the optimum thickness for InAs0i25P0#75, the data that we do have are in

agreement with equation (1). We have also attempted to verify equation (1)

for GaAs. The parametric curves (comparable to those of Figure 1) exhibit a

very broad maximum which makes interpretation difficult. If the optimum thick-

ness for GaAs is exponentially dependent on the wavelength, then ß is very

small. It may be however, that equation (1), and this is purely conjecture, is

applicable only to cathodes in which the top of the interfacial barrier is above

the bottom of the conduction band.

Figure 3 Quantum y laid of o,oiiLiimüur.ly-pi.-oces5cd '^^0,^0.^ '

I X

o UJ —1 UJ

>- O

o

<

Of

0.9 1.0 U

WAVELENGTH (MICROMETERS)

Section 2

EFFECT OF THICK Ca-U LAYERS ON PHOTOrillSSION J FUOli NEÜATIVE-ELECTRÜN-AFFINITY CATUOJAti

The efriciency of small-bandSap-KEA cathodes near threshold :LB much

less than tl^e efficiency near threshold of larfjer-band^.p waterl.-.lr.. For

example, the yield of 1**»^/*^ »« ^ **-****» ic **** *«* l0"* "^ ^

GaAs at 0.85 microns. The major cause for thll difference has be.n correctly

indentified as due to interfacial-barrier effects0'^. However since thicker

cesium-oxide coverage is required to optimise the photore.sponse of smaller-

bandgap materials05, part of the difference in threshold yield should be due

to this difference in thickness. We have investigated the effect of thick

Cs-0 layers on photoemission from KEA cathodes and shov that a substantial

decrease in yield may be expected on the basis of the difference in thickness

alone. A simple empirical relation givinB the effect of thick Cs-0 layers on

photocmission is derived.

The yield data were directly receded with an analoE recording system

which included a phase-sensitive-Jetection apparatus and a scanning monochromator.

The 0 exposure was recorded with a partial pressure analyzer, and was used as

a check on the Cs cover0ge(£) which was estimated by timing the exposure periods

as described in reference 1. Photoemission measurements were made on GaAs and

InAs P . sensitised with Cs and 0, by the simultaneous-exposure technique 1 . ^ 0.A 0.6' •

8

The spectral raspou^e vXth the. infrared yield opfeimtiBcd waa Ciryt:

recorded. Additional yield curves were then recorded ^or increasius Cs 800

0 coveräße. To avoid Cs loss fro» the cathode hetveen processing steps, the

Cs channel was turned dowu so that the arrival rate of Cs aton. at the cathode

was in quilibvium with the loss rate. The point of equilibriun was estimated

from the stability of the photor-spense and remahvod essentially the same at

all exposure levels.

The quantum efficiency of GaAs at 750cß is shove in Figure 4 as a function

of thickness measured in monolayers(2) of Cs. The points represent the experi-

mental values and the solid curve represents the equation

n m r,o (1 - e~Ut)> $>

■where n is the quantum efficiency (electrons/incident photon) as a function of

thickness U (monolayers of Cs); no is the peak quantum efficiency (0.213 electrons/

incident photon for the curve shown) and Ä (0.65 monolayers of Cs for the curve

shown) is defined as the attenuation length. The thickness t. does not represent

the total coverage but rather the coverage beyond that required to optimize the i

infrared photoresponse. The total coverage is given by t =(t + to)» where

t is the coverage required to optimize the infrared response.

Near threshold the yield of UI-V NBA cathodes is given by ^

n - P(l-R) *7[1 + (1/aL)], <3)

where the symbols used have the usual meaning. The factor F/[l + (lM)] which

accounts for the generation and transport of the photoelectrons in the UI-V

semiconductor, clearly does not "change with surface treatment. For the Cs-0

g MiMM ■■■■■■i ■■■■■■■■^■■im! BBI"'

\

Figure 4 Thickness dcpcndouce of the quantum efficiency n. of ^^J500/ 8 The points represent the expcrimental^ata and the sol^d cuive la

a plot of the equation n « Ü.213(l-e ' )•

re FVDnQllDF /MnMniAYFR<^

\ ■

I

layer thicbLscs considerod here, we can probably safely fesume that the

reflectivity R of the surface reHains that of clean'GaAs. Ihla leaves enly

tit.escape probability P. dependent ok surface treatment, Which mtt»t therefore

mirror the tlUckness dcpc.-ndcncc of n. i.e.,

Pd-e-^). & o , -

where * is the escape probability at t - 0. A similar areument vill quickly

show that equation (4) should be valid not only near threshold, but over the

entire spectral range of response of the 1II-V NEA cathode. ',

To verify equation (i) experimentally we have made a least-squares-fit

analysis of equation (3), and oJyield data at each exposure level to determine

■p(t). The points given in Figure 5 represent the escape probabilities at 750oS

obtained from the least-squaJ-fit analysis and the solid curve represents

equation (A) with the escape probability^ P0 - 0.63 and an attenuation length

Ä « 0.65 monolayers of Cs. As expected, the same attenuation length used to fit

the data in Figure 4.' describes the data in Figure 5. Equation 4 with Po = 0.57

and L - 10 "layers of Cs - O" is also in reasonable agreement with the estimated

X-electrpn-escape probability given in Figure ^ of Reference 12.

The attenuation length A, is not constant however, and Figure 6 shows its

wavelength dependence. For the T electrons, A is^approximately constant at

0.65 monolayers of Cs. We would expect a constant attenuation length for this

case since moat of the photoeleetrons arriving at the surface have thermalizcd

-in the r-conduction-band minimum. At shorter wavelengths, i increases and appears

to saturate at about 1.0 monolayer of Cs.

I

11

rigurt; j ifuciuieBH uepünucncc 01 tue escape prounui-Liuy i, CJJ. Lra/uj au /JUU. The pointy represent the cxpcrlracntal data and thu solid curve is

' a plot of the equation P - 0. 63(l-e~0'ei-r,/t:).

QCl < OQ O o.

a, < o

4 6 8 10

Cs EXPOSURE (MONOLAYERS).

12

Figure 6 Wavelength depcnclance of the attenuation length of Cs-Ü low-work-

function surfaceii on GaAs.

WAVELENGTH (MICROMETERS)

13

WB uaX. al.o verified cquaCion (2) for l^o.^o.ö' ^^ 7 ^^

the thickn.^ dcrencleacc oi thu yxcld frou i^^^I'^.-at: 0.6^ «idJOlM (upper

curve) and 0.7 «ictOttt (lover curve). Tbe pelntS on the curve, arc the actual

data points, and the solid line repre.ents oquatioa (2) with r^ - 0.062 and

| . 0.6 for tU upper curvc.aud no - 0.0335 and % - 0.5 for the lower curve.

If, as in reference 1, we M»Utt» that the »ticking coefficient of

C on InAsA /Vn t ard on GaAs doe. not chanye sianificant.y with ccvera^P^ s ü.H U«o then the Cs-0 low-work-function 9VXi*m in an infrared-optiiui^d InAöü>/(P0t6

cathode is approximately 2.3 mnnolayers of Co thxeker than tUat of optimised

Ga.Vs. From equation 2 , with ü - 0.65 monolayer., we find that an additional

2.3 monolayers of Cs would reduce the infrared photoresponse of GaAs by a

factor of about 4. We do not luve sufficie..:-. data on 1^Q^0,G ^ allow^ uc;

to determine its attenuation length near threshold, however from Figure 7, we

' expect that it will be smaller than 0.65 monolayers. Consequently alone on

the basis of the increased thickness required to optimise the infrared response

of InAs ,Pn •. we would expect the threshold response of IHA^^PQ^ to be O.'f O.b ^j

lower than that of GaAs by a factor of 5 or more .

•v

14

Figure 7 '0.^0.6 at thlekftasi d»?cu<Umc« of the guaiittmi efJ.iciencv r\ ol XnA;j(

6300R (upper curve.) and 700ÜA (low.rr CUI-VL;) . The points represent the experimental data and the solid curves are plots .

%Oi the equations n - 0.062(1 - o"0-C/t.) md r, - 0.0335(1 - JT0'"*)

for tlie upper and lower curves rcKpcclively•

, I 2

THICKNESS (M0N0LAYERS Cs)

REFERENCES

1.

2.

3.

5.

6.

11. Sonnenberg, Appl. Phy«. Lett .19, 431 (1971).

A monoiayer of Cs is defined here a. the amount of Ci required to

optimize tV.e photore.ponse at 6323^ with Ol only

We assume that the Cs nad 02 sticking coefficients äo not change

significantly with coverage.

The fact that T is greater than 1 shows that .ome 02 is required to

optimize the photoresponsc at all wavelength., including the ultra

violet range.

Note that the efficiency a. 1.3 microns is almost 0.02%. Further

improvement with better InAs^P^ material can be expected.

If we make the same assumptions about the Cs-0 surface that were made

in reference 1. we find that the low-work-function surface consist.

of approximately 1 monoiayer of Cs and 1.7 monolayers of C.^O.

L. U. James and J. J. Uebbing. Appl- «**•• Liters 16. 370 (1970).

R. L. Bell, L. W. Jame., G. A. Antypas, J. Edgecumbe, and R. L. Moon,

Appl. Phys. Letters 19, 513 (1971).

9. The ratio of Cs to 02 remains constant.

L. W. James and J. L. Holl Phys. Rev. 183. 7A0 (1969).

High escape probabilities such as this were achieved with <110> GaAs

surfaces. Our best yield curves obtained in this material are comparable

to those obtained on <111B> GaAs by.L. W. James, G. A. Antypas,

j. Edgecumbe, R. L. Moon, and R. L. Bell, J. of Appl. Phys. 42, 4976

(1971). Our yield is slightly better than theirs above 1.65 eV but

slightly lower than the.rs below about 1.6 eV. The decreased yield in^ ^

the infrared Is probably due to the low doping of our material (1 x 10 cm ),

L. W. James, J. L. Moll, and W. E. Spicer Symposium on GaAs, 230. (1968).

7.

8.

10.

11.

12.

16

13. The fact that the BAtting o£ the Cr^ channel at the point of equJ llbriura

(Ca loss « CH gain) remains efjsentially the. same at all exposure 1evels,

seems to confirm that the assumption is valid.

14. Since the thickness of the low-work-function surface has such a profound

effect on pbotoemission, it is not surprising that a g^at deal of

progress in the processing of infrared cathodes has recencly been made.

in