Prototype dispenser photocathode: Demonstration and comparison to theory

Article abstract of DOI:10.1063/1.2713341

A method to significantly extend the operational lifetime of alkali-based photocathodes by diffusing cesium to the surface at moderate temperature is presented and shown to restore the quantum efficiency (QE) of cesiated tungsten. Experimental measurements of QE as a function of surface cesium coverage compare exceptionally well with a recent theoretical photoemission model, notably without the use of adjustable parameters. A prototype cesium dispenser cell is demonstrated and validates the concept upon which long-life dispenser photocathodes can be based.

Full text of DOI:10.1063/1.2713341

Prototype dispenser photocathode: Demonstration and comparison to theory
N. A. Moody, K. L. Jensen, D. W. Feldman, P. G. O’Shea, and E. J. Montgomery
Citation: Applied Physics Letters 90, 114108 (2007); doi: 10.1063/1.2713341
View online: http://dx.doi.org/10.1063/1.2713341
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/90/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Modeling the evaporation rate of cesium off tungsten based controlled porosity dispenser photocathodes
AIP Advances 3, 042105 (2013); 10.1063/1.4800700
Evaluation of a cesium iodide photocathode assisted with MgO-coated multiwall carbon nanotubes
Appl. Phys. Lett. 96, 141109 (2010); 10.1063/1.3383220
Factors affecting performance of dispenser photocathodes
J. Appl. Phys. 102, 104901 (2007); 10.1063/1.2805653
The quantum efficiency of dispenser photocathodes: Comparison of theory to experiment
Appl. Phys. Lett. 85, 5448 (2004); 10.1063/1.1827936
Electron diffusion length and escape probabilities for cesiated and hydrogenated polycrystalline diamond
photocathodes
Appl. Phys. Lett. 75, 3533 (1999); 10.1063/1.125379
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.22.89.130 On: Wed, 03 Dec 2014 16:46:33

APPLIED PHYSICS LETTERS 90, 114108 ͑2007͒
Prototype dispenser photocathode: Demonstration and comparison
to theory
N. A. Moodya͒
Los Alamos National Laboratory, Los Alamos, New Mexico 87545
K. L. Jensen
Naval Research Laboratory, Washington, DC 20375-5347
D. W. Feldman, P. G. O’Shea, and E. J. Montgomery
University of Maryland, College Park, Maryland 20742-3511
͑Received 10 January 2007; accepted 8 February 2007; published online 15 March 2007͒
A method to significantly extend the operational lifetime of alkali-based photocathodes by diffusing
cesium to the surface at moderate temperature is presented and shown to restore the quantum
efficiency ͑QE͒ of cesiated tungsten. Experimental measurements of QE as a function of surface
cesium coverage compare exceptionally well with a recent theoretical photoemission model, notably
without the use of adjustable parameters. A prototype cesium dispenser cell is demonstrated and
validates the concept upon which long-life dispenser photocathodes can be based. © 2007 American
Institute of Physics. ͓DOI: 10.1063/1.2713341͔
Photoinjection produces high quality electron beams for
free electron lasers ͑FEL’s͒, energy-recovery linacs, pump-
probe experiments, and other applications.1 A pulsed drive
laser impinges a cathode and acts to photoswitch the electron
beam in synchronicity with an accelerating rf field. Many
photocathode types have been intensely studied, but no
single technology exhibits both high quantum efficiency
͑QE͒ and extended operational lifetime without complicated
surface conditioning. A related problem is the lack of a good
photoemission model for coated surfaces, needed not only
for cathode development but also for end-to-end beam simu-
lation to ascertain the impact of ͑for example͒ intrinsic pho-
tocathode emittance as it relates to emission nonuniformity
and evolution. QE is not the only metric of photocathode
performance: operation under laboratory conditions cannot
be extrapolated to the environment of a rf gun by fiat. Here,
a dispenser cathode capable of establishing and maintaining
a photosensitive cesium layer on the surface via diffusion
through porous sintered tungsten is demonstrated and char-
acterized. Measurements of the QE of a partially covered
surface correlated exceptionally well with a theoretical
model, notably without the use of adjustable parameters.
Based on the temperature-related behavior of the prototype,
it is argued that a controlled porosity dispenser photocathode
͑CPDP͒ is a viable electron source candidate for high power
FEL’s.
Metallic emitters, though extremely robust, require a UV
drive laser and exhibit very low QE. Multialkali photocath-
odes ͑e.g., CsK2Sb and Cs3Sb͒ have impressive QE in the
visible range but degrade quickly even in ultrahigh vacuum,2
believed to be because loss of one or more alkali components
occurs from the cathode surface due to chemical instability.
A remedy in the case of cesium-based photocathodes is re-
coating the surface with trace amounts of cesium to restore
optimal photosensitivity.3 For most photoinjector designs, re-
placement of surface monolayers requires complex proce-
dures at regular intervals. The advantage of the CPDP archi-
tecture is the possibility of in situ monolayer replenishment
of cesium that may, in turn, extend the lifetime of high QE
alkali-based photocathodes.
Dispenser cathodes are a mature thermionic emitter
technology.4–6 The work function is lowered as the alkali
material trapped in a porous tungsten matrix diffuses to the
surface at high temperature. Traditional dispenser thermionic
cathodes operate much hotter ͑Ͼ1000 °C͒ than the break-
down temperature of high QE films ͑ϳ220 °C͒. We advo-
cate instead a prototype dispenser cell which operates at
180 °C using sintered tungsten as a diffusion barrier encap-
sulating a sealed volume of cesium chromate. Initial activa-
tion at 500 °C releases elemental cesium within the cell, and
subsequent heating to 180 °C allows its controlled diffusion
to the cathode surface. The lower operating temperature sig-
nificantly reduces evaporative losses and accommodates
deposition growth of highly photosensitive surface layers
following the initial activation heating.
The effect of cesium on the substrate material ͑tungsten͒
was studied in detail,7,8 by measuring QE while cesium was
evaporated onto the substrate surface using external, com-
mercial alkali sources. The amount of cesium arriving at the
surface was measured using a quartz crystal balance. In the
dispenser cell, cesium arrives from beneath the substrate, so
surface coverage cannot be directly measured as before but
can be theoretically inferred based on the established rela-
tionship between QE and cesium coverage,9 which therefore
provides a method to measure the amount of cesium at the
cathode surface. Using this technique, it was confirmed that
after activation the dispenser cell released cesium to the sur-
face at a controlled rate at temperatures less than 200 °C.
Importantly, QE measurements suggest that the layer uni-
formly coats the surface as in the case where cesium is
evaporated directly onto a solid tungsten metal base.
The sintered tungsten dispenser substrate is roughly 30%
porous and was cleaned prior to deposition using a 40 mA
dose of a 6.4 keV argon ion beam. Photocurrent was mea-
sured using diode lasers ͑Ͻ5 mW͒ at five different wave-
lengths ͑375, 405, 532, 655, and 802 nm͒. Further details of
a͒
Electronic mail: nmoody@ieee.org
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2007/90͑11͒/114108/3/$23.00 90, 114108-1 © 2007 American Institute of Physics
129.22.89.130 On: Wed, 03 Dec 2014 16:46:33

114108-2
Moody et al.
Appl. Phys. Lett. 90, 114108 ͑2007͒
the experimental apparatus are deferred to a later work. The
relationship between QE and cesium coverage is highly re-
peatable when an argon ion beam cleaning is performed at
least once prior to each deposition and agreement with the
updated theoretical model ͑below͒ is excellent. After the ini-
tial argon ion dose, additional cleaning did not produce
higher QE. The base pressure of the UHV system was main-
tained at 3ϫ10−10 Torr, reaching 2ϫ10−9 Torr during the
cesium deposition process because of the slight outgassing of
the source.
Agreement between theory and experiment was mark-
edly improved based on two modifications to the previous
model:9 a revised model of the temperature and energy de-
pendent electron scattering terms and the use of an updated
moment-based approach10 to emission probability. The scat-
tering terms considered are electron-acoustic phonon and
electron-electron ͑for which the latter dominates in metals͒
coupled with a low-temperature residual scattering rate. The
electron-phonon relaxation rate is modeled by11,12
FIG. 1. ͑Color online͒ Theoretical and experimental agreement ͑relative͒ for
QE vs cesium coverage.
the log scale and in an absolute sense, for cesium on tungsten
following an argon ion cleaning, is extraordinarily good. Dis-
crepancies arise for the lower photon energies that may be
related to nonuniform coverage effects and other factors that
will be explored in greater detail elsewhere. To test whether
the agreement was fortuitous, the same experiment was per-
formed for cesium deposited on a silver surface. The level of
agreement was again good ͑see Fig. 4 in Ref. 15͒, suggesting
that the theoretical model adequately accounts for the phys-
ics involved and provides a sufficiently qualified emission
model for complex beam simulation codes.16 For all wave-
lengths and both base metal cases, theory and experiment
agree quite well, an encouraging observation given the com-
plexity of the polycrystalline photoemitting surface, espe-
cially when microscopic features such as grain crystal face,
contaminants, and pores are evident, which the theory at
present treats as uniform. Since work function can vary sev-
eral tenths of an electron volt depending on crystal orienta-
tion, micron-scale resolution in a photoemission electron mi-
croscope system17 as well as temperature variation effect
studies may be required in order to discern systematic varia-
tion between theory and experiment. These experiments are
in preparation.
2
−1
T
/T
␲␳ប3
5
vs TD
x5dx
͑ex − 1͒͑1 − e−x͒
D
␶ac =
, ͑1͒
ͩ ͪ
͵
is the velocity of sound, ⌶ is the deformation po-
ͭ
ͮ
mkBTkF⌶2
T
0
where
vs
tential, បkF is the Fermi momentum, and TD is the Debye
temperature. The electron-electron relaxation time13,14 is
modeled by
8បK2s
␣2␲mc2͑kBT͒
2
−1
⌬E
2kF
␶ee =
1 +
␥
,
ͩ ͩ ͪ ͪ ͩ ͪ
ͫ
ͬ
2
␲kBT
qo
−1
2
x3
x
tan ͑xͱ2 + x ͒
␥͑x͒ =
tan−1x +
−
,
͑2͒
ͩ
ͪ
4
1 + x2
2
ͱ2 + x
where Ks is a dielectric constant modification which accounts
for additional screening by d electrons, ␣ is the fine structure
constant, ⌬E is the electron energy above the Fermi level,
បkF is the Fermi momentum, and qo is the Thomas-Fermi
screening factor. By using a combination of commonly avail-
able temperature-dependent thermal conductivity data and
Monte Carlo simulations of electron-electron scattering rates
as a function of energy from the literature only, we have
inferred appropriate values of ⌶ and Ks and the low tempera-
ture residual scattering rate. Using these values without
modification, we have compared the theoretical model to QE
measurements from tungsten and silver surfaces partially
coated with cesium. The details of the parameter determina-
tion and qualification for the scattering terms based on values
in the literature and the methods for solving the moment-
based integrals in the presence of a work function which
depends on surface coverage of cesium will be deferred to a
separate publication.
The relationship between QE and cesium coverage on
͑atomically clean͒ sintered tungsten can be used to infer the
arrival of cesium to the surface of the dispenser cell by moni-
toring QE. The dispenser cell consists of a thin-walled stain-
less steel canister 1 cm in height and diameter with a 1 mm
thick sintered tungsten disk brazed to the top. Powders of
cesium chromate and titanium ͑1:5 weight ratio͒ were
pressed into a 0.7 gm pellet and inserted into the cell prior to
brazing. The cell was then evacuated and outgassed at
Figures 1 and 2 compare experimental data to theoretical
predictions. The relationship between the experimental coor-
dinate ͑thickness of cesium deposited on the surface͒ and the
theoretical coordinate ͑percentage of surface covered͒ is re-
lated by the cesium covalent radius of 0.52 nm: the mass of
a monolayer corresponds to the density of cesium multiplied
by the area coated and the thickness of the monolayer. All
theoretical values used in the calculation were independently
obtained based on literature sources and not adjusted to op-
timize agreement with experiment. The theory/experiment
FIG. 2. ͑Color online͒ Theoretical and experimental agreement ͑absolute͒
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
agreement, both in the relative dependence highlighted on
for QE vs cesium coverage.
129.22.89.130 On: Wed, 03 Dec 2014 16:46:33

114108-3
Moody et al.
Appl. Phys. Lett. 90, 114108 ͑2007͒
initial cesium layer was placed on the dispenser cell and then
heated to 180 °C continuously while QE was measured for
more than 8 days. The 1/e lifetime of the cathode during this
continuous rejuvenation process was increased by roughly an
order of magnitude to more than 47 days. This confirms that
the rejuvenation process can be sustained over extended pe-
riods.
In conclusion, the QE of a Cs-covered sintered tungsten
surface was characterized, and the data were shown to cor-
relate well with a photoemission model. Experiment and
theory agree to within 30%, and the relationship between QE
and cesium surface coverage was used to investigate and
characterize a prototype dispenser photocathode. The dis-
penser surface was cleaned using an argon ion beam and
activated at 470 °C. Subsequent heating at much lower tem-
peratures ͑140–180 °C͒ allowed a controlled release of ce-
sium to the surface that established and maintained a surface
layer optimal for photoemission. The dispenser cell was
shown to deliver cesium to the surface in a temperature con-
trolled manner over long periods of time. The determination
of whether high QE cathodes can be built based on the dis-
penser cell technology and rejuvenated in situ to provide
uniform surface coverage and good QE at operational tem-
peratures will be the subject of a separate work.
FIG. 3. ͑Color online͒ Cesium coverage vs temperature during dispenser
cathode rejuvenation.
225 °C. Activation occurred upon further heating to 470 °C
when cesium chromate reacted with titanium powder, releas-
ing elemental cesium. Most of the elemental cesium was
contained within the cell, but some escaped and accumulated
on chamber walls and contributed to thermionic emission
from the cathode surface. The cell was cooled to room tem-
perature, allowing cesium to remain within the cell and to
accumulate on its surface. QE was measured continuously
for several days until the cesium layer began to degrade,
resulting in a decreased photoyield from 0.11% to 0.06% in
the UV ͑375 nm͒. Upon heating from room temperature, QE
increased as cesium was brought to the surface to repair the
damaged layer, and at 140 °C, QE reached its previous peak
value of 0.11% in the UV. Argon ion cleaning was used to
remove the entire surface layer and return to an atomically
clean tungsten substrate. The dispenser was again slowly
heated to bring cesium to the surface, and the previous in-
crease and peak in QE was again observed. These results
demonstrate that the dispenser cell can deliver cesium to the
surface in a controlled manner at moderate temperatures
͑ഛ150 °C͒.
Relating QE to cesium coverage quantifies the ability of
the dispenser to deliver cesium to the surface. Figure 3
shows coverage as a function of temperature during the re-
juvenation cycle. Note that at 140 °C ͑the temperature at
which QE peaks͒ the coverage is observed to be about 63%.
Coverage does not increase beyond this value because the
elevated temperature causes desorption which prevents fur-
ther accumulation of cesium at the surface. The 1/e lifetime
for a recesiated layer at room temperature was about 5 days.
QE decreases over time because cesium leaves the surface
either through desorption or because of ion back-
bombardment. If the cell is held at an elevated temperature,
however, the cesium layer can be continuously repaired. An
The authors gratefully acknowledge funding provided by
the Joint Technology Office and the Office of Naval Re-
search. The authors have benefited from interactions with
David Dowell ͑SLAC͒, Jonathan Shaw ͑NRL͒, Joan Yater
͑NRL͒, Dinh Nguyen ͑LANL͒, and John Smedley ͑BNL͒.
1P. G. O’Shea and H. P. Freund, Science 292, 1853 ͑2001͒.
2S. H. Kong, Nucl. Instrum. Methods Phys. Res. A 358, 272 ͑1995͒.
3A. H. Sommer, Photoemissive Materials, 1st ed. ͑Wiley, New York, 1968͒,
Chap. 8, pp. 113–199.
4W. C. Rutledge and E. S. Rittner, J. Appl. Phys. 28, 167 ͑1957͒.
5G. Gärtner, P. Geittner, D. Raasch, and D. U. Wiechert, Appl. Surf. Sci.
146, 22 ͑1999͒.
6R. T. Longo, J. Appl. Phys. 94, 6966 ͑2003͒.
7N. A. Moody, D. W. Feldman, and P. G. O’Shea, J. Directed Energy ͑to be
published͒.
8N. A. Moody, D. W. Feldman, P. G. O’Shea, K. L. Jensen, and A. Balter,
Proceedings of the 27th International Free Electron Laser Conference,
Stanford, CA ͑2005͒, p. MOPP050, www.bessy.de/fel2006/
9K. L. Jensen, D. W. Feldman, N. A. Moody, and P. G. O’Shea, J. Appl.
Phys. 99, 124905 ͑2006͒.
10K. L. Jensen, P. G. O’Shea, D. W. Feldman, and N. A. Moody, Appl. Phys.
Lett. 89, 224103 ͑2006͒.
11B. K. Ridley, Quantum Processes in Semiconductors, 4th ed. ͑Clarendon,
Oxford, 1999͒.
12Harald Ibach, and Hans Lüth, Solid State Physics, 2nd Ed. ͑Springer,
Berlin, 1995͒.
13D. K. Wagner and R. Bowers, Adv. Phys. 27, 651 ͑1978͒.
14A. V. Lugovskoy and I. Bray, J. Phys. D 31, L78 ͑1998͒.
15K. L. Jensen, N. A. Moody, D. W. Feldman, P. G. O’Shea, J. E. Yater, and
J. L. Shaw, Proceedings of the 28th International Free Electron Laser
Conference, Berlin, Germany, 2006 ͑unpublished͒, Paper No. THPPH068.
16J. J. Petillo, presented at 48th Annual Meeting of the Division of Plasma
Physics ͑APS͒, Philadelphia, PA, 2006 ͑unpublished͒.
17J. Shaw, J. Yater, K. L. Jensen, P. G. O’Shea, D. W. Feldman, and N. A.
Moody, presented at Ninth Annual Directed Energy Symposium, Albu-
querque, NM, 2006 ͑unpublished͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.22.89.130 On: Wed, 03 Dec 2014 16:46:33