KRYPTON OVERLAYERS ON GRAPHITE : LOW
ENERGY ELECTRON DIFFRACTION AND AUGER
ELECTRON SPECTROSCOPY MEASUREMENTS
S. Fain, Jr, M. Chinn
To cite this version:
S. Fain, Jr, M. Chinn. KRYPTON OVERLAYERS ON GRAPHITE : LOW ENERGY ELECTRON DIFFRACTION AND AUGER ELECTRON SPECTROSCOPY MEASUREMENTS.
Journal de Physique Colloques, 1977, 38 (C4), pp.C4-99-C4-104. <10.1051/jphyscol:1977415>.
<jpa-00217131>
HAL Id: jpa-00217131
https://hal.archives-ouvertes.fr/jpa-00217131
Submitted on 1 Jan 1977
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
JOWRNAL DE PHYSIQUE
Colloque C4, supplkment au no 10, Tome 38, octobre 1977, page C4-99
KRYPTON OVERLAYERS ON GRAPHITE : LOW ENERGY ELECTRON
DIFFRACTION AND AUGER ELECTRON SPECTROSCOPY MEASUREMENTS (*)
S.C. FAIN, Jr. and M.D. CHINN
Physics Department,
University of Washington,
Seattle, Washington, 98195 U.S.A.
Rhsumh. - La diffraction d'Blectrons lents et la spectroscopie des Blectrons Auger ont BtB
utiliskes pour Btudier I'adsorption de krypton sur la face (0001) du graphite. Les pressions et
temperatures de changement d'ktat des phases bidimensionnelles h p <
torr sont comparBes
avec les isothermes d'adsorption B p >
torr.
Abstract. - Low energy electron diffraction and Auger electron spectroscopy were used to
study the adsorption of krypton on the (0001) plane of graphite. The pressures and temperatures
torr are compared with adsorption isotherms at
for two dimensional phase transitions at p <
p>
torn.
1. Introduction. - The adsorption of Kr on
graphite has been extensively studied by isotherm
measurements for equilibrium vapor pressures
greater than
torr [I-61. Adsorption at pressures
less than
torr was first investigated by Kramer
and Suzanne using low-energy electron diffraction
(LEED) and Auger spectroscopy [7]. Monolayer
condensation was explored by both LEED and
Auger measurements, but the transition from an
in-registry solid to a compressed, out-of-registry
solid first inferred from isotherms at pressures near
1 torr [2] was not observed [7]. We have been able
to observe this transition with a high resolution, low
incident current LEED apparatus [S]. We present
elsewhere [9] lattice parameter measurements near
this transition in a form directly comparable to a
theory of monolayer epitaxy [lo]. Venables and
Schabes-Retchkiman also discuss our measurements
in these proceedings [I 11.
In this paper we present some LEED and Auger
determinations of pressures
(lo-' torr < p <
torr)
and temperatures for various two-dimensional phase
transitions for Kr on the basal plane of a graphite
single crystal. We compare these results with vapor
pressure isotherm results [2-41 and earlier LEED
(*) Research supported by National Science Foundation,
Research Corporation, and University of Washington Graduate
School Research Fund.
and Auger results [7]. In addition we compare the
shape of the deregistry transition observed by
LEED with that observed by isotherms near
1 torr 121 and present our lattice parameter
determination of a two layer film.
2. Procedures. - The LEED apparatus has been
described elsewhere [a]. The most important
modification made for this work was to connect an
AC preamplifier between a 0-5 kV high voltage
supply and the phosphor screen to permit Auger
measurements to be made while observing LEED
patterns. This was done to be certain that both
LEED and Auger observations were being made for
the same area of the sample and for nearly the same
desorption rate associated with electron effects on
the adsorbed layers. In the work of Kramer and
Suzanne, the pressure required for monolayer
condensation was definitely higher as determined by
Auger measurements than by LEED, due to the
much higher energy incident electrons of the
former t73.
Our Auger spectra were obtained with a 342 eV,
2 x lo-' A primary electron beam and a modulation
of 3 eV peak-to-peak. Long data collection times
were necessary because of the low signal to noise
ratio due to shot noise [l2]. The Auger curves
presented here are smoothed averages of three scans
taking 5 minutes each. The LEED pattern at 342 eV
was sufficiently clear to ensure that the electron
beam was focused on the same graphite crystallite
area as at 144 eV. LEED patterns were obtained
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1977415
S. C. FAIN, Jr. AND M. D. CHINN
C4-100
with a 144 eV, 5 x lo-' A primary beam between
Auger scans. Except for the photographs taken
between Auger scans, the LEED observations were
A, 0.2 mm diameter
made with a 144 eV, 3 x
electron beam [9].
Temperatures were taken from the calibration of
Rosenbaum for hard-worked
Au-0.07 % FeChrome1 P thermocouples [13]. The considerable
pressure used to clamp the thermocouple to the front
surface of the graphite and the high in-plane thermal
conduction of graphite assure that the thermocouple
indication is representative of the sample
1 K. Small changes in
temperature to better than
temperature could be measured to a greater
accuracy of -+ 0.3 K. The pressures reported here
are mostly taken from a glass Bayard-Alpert guage
(Consolidated Vacuum Corporation, GIC-017)
calibrated on a dynamic standard which utilized a
capillary flow meter and a fixed orifice (Boeing
Technology Services, Seattle, Washington). This
method permits an absolute accuracy of 2 10 % for
torr. The results
pressures between 10" and
reported here were taken with the ion pumps off and
the entire chamber filled with Kr. Normally a
constant pressure was maintained with the valve to
the Kr bottle essentially closed (static conditions).
An exception was the highest pressure experiment at
47 K where the pressure dropped very quickly
from 3.6 x lo-' to 2.0 x lo-' torr when the Kr valve
was closed. We believe this was due to formation of
bulk Kr layers on the coldest surfaces.
*
3. Results and discussion. - The pressures and
temperatures of two-dimensional phase transitions
observed in this work are given in Table I and as the
points for p < lop4torr in figure 1. The points for
p>
torr are from volumetric isotherm
measurements [2-41 as explained below and in the
figure caption.
Temperature and pressures for two -dimensional
phase transitions of Kr on graphite (not including
possible systematic errors discussed in text).
a ) Fluid-solid condensation
61tlK
62.2 r 0.3 K
65.7 2 0.3 K
5.0 2 0.2 x lo-' torr
1.5 2 0.2 X
torr
7.3 r 0.6 x
torr
b ) Transition from in-registry solid to out-of-registry solid
2.8 r 0.9 x lo-' torr
54.3 r 0.3 K
57.1 &0.3K
1.65 r 0.4 x
torr
7.4 -t 0.5 x
torr
59.3 r 0.3 K
c ) Second layer condensation
46.8 K t 0.5 K
2.8 2 0.8 x
torr
d ) Third layer condensation
46.9 K t 0.5 K
9.8 2 1.3 x
torr
e ) Bulk condensation
46.8 K r 0.5 K
2.3 r 1.2 X lo-' torr
FIG. 1. - Pressure of Kr versus inverse temperature of graphite
substrate for fluid-solid condensation ( a ) , onset of in-registry to
out-of-registry
solid-solid transition ( b ) , second
layer
condensation ( c ) , third layer condensation (d), and bulk
torr)
condensation (e). The data from this experiment (p <
are listed in table I and discussed in the text. The points above
torr are taken from references [2-41. The 2-D triple point
deduced by Larher [3] is shown as an asterisk. Equations for
line ( a ) above and below the asterisk are from Larher [3].
Equations for lines ( b ) and (c) are taken from Regnier, Thorny,
Duval [2]. The bulk vapor pressure equation for line (e) is taken
from Pollack [14]. A vertical dashed line marks 77 K.
3.1 FLUID-SOLIDCONDENSATION.
- Our
measurements of the condensation of a 2-D solid
monolayer from a low density 2-D gas were made by
LEED observations of the appearance of first order
KF ' overlayer.
diffraction
beams
of
the
Measurements made at 61 K with a spot photometer
showed an increase of overlayer LEED intensity
from background level to full intensity for a gradual,
continuous increase of the Kr pressure from 4.8 to
5.2 x lo-' torr in 100 seconds. The width of the
transition observed in this way is narrower than that
observed by Kramer and Suzanne using Auger
measurements [7]. Monolayer condensation was
reversible for temperatures above about 58 K.
The pressures and temperatures we obtain for
monolayer condensation (squares in Fig. 1) are in
reasonable
agreement
with
the
LEED
determinations of Kramer and Suzanne [7] and with
the extrapolation (line a in Fig. 1) of Larher's
measurements [3].
3 . 2 TRANSITION FROM IN-REGISTRY SOLID TO
OUT-OF-REGISTRY SOLID.
Our measurements of
-
Kr-Kr distance near the deregistry transition [9] are presented in figure 2 a as the square of
the ratio of the in-registry Kr-Kr distance do to the
KRYPTON OVERLAYERS ON GRAPHITE
C4-101
Fort [4 and F. A. Putnam, Massachusetts Institute
of Technology, private communication]. Our onset
data are consistent with the extrapolation (line b of
Fig. 1) of the work of Thomy, Regnier, and
Duval [2].
The Auger measurements shown in figures 3 a
and 3 b are taken at pressures well below and well
above the deregistry transition. The increase in
Auger signal in figure 3 b over 3 a is due to the
combined effects of completing a registered
monolayer, adding atoms to compress the
monolayer, and putting some atoms in the second
layer. An Auger spectrum taken just above the
torr) had
deregistry transition (at 53.5 K, 3.6 x
a peak-to-peak height midway between those of
figures 3 a and 36. From this measurement we
conclude that the increase in Kr coverage from
monolayer condensation to the onset of our
deregistry transition is about 10 %, presumably due
to filling in vacancies in the layer.
FIG.2. - Comparison of LEED measurements and vapor
pressure isotherms near the transition from an in-registry solid to
a compressed, out-of-registry solid. a ) do is the registered layer
Kr-Kr distance and d is the mean observed Kr-Kr distance ;
temperatures shown are 54 K (circles), 57 K (squares), and 59 K
(triangles) ; the same solid line is drawn through each set of
experimental data. b ) 0 is the coverage measured by Thomy,
Regnier, and Duval [2]at 77.3 K (circles), 90.0 K (squares), and
96.3 K (triangles) ; 00 = 0.930,0.940, and 0.945, respectively ;the
error bars are the size of the points given in figure 4 of
reference [2] ;the solid lines are guides to the eye.
observed mean Kr-Kr distance d. The data from
Thomy, Regnier, and Duval[2] presented in figure
2 b would be directly comparable to that in figure 2 a
(except for temperature differences) only for a perf ect monolayer with no vacancies, dislocations, or
second layer atoms. For such a perfect monolayer,
the coverage 8 normalized to the coverage 6, at the
onset of the transition would be simply given by
(6/8,) = (do/d)2. However, from figure 1 it is evident that at the higher temperatures the in-registry to
out-of-registry transition is much closer in pressure
to second layer condensation ; thus the number of
atoms in the second layer will be higher for the data
in figure 2 b than for figure 2a. If a correction for
second layer adsorption is applied to the 77.3 K data
in figure 2b, the shape of the corrected data is more
similar, to that of the curves in figure 2 a
[G. D. Halsey, University of Washington, private
communication].
Our pressures and temperatures for onset of the
transition are shown as circles in figure 1. The
circles at higher pressure are taken from Thomy,
Regnier, and Duval [2] and from Putnam and
-
FIG.3.
Smoothed Auger electron spectra obtained with a
342 eV, 2 x lo-' A primary electron beam. a ) In-registry layer
torr), 6 ) Out(near fluid-solid condensation ;60.5 K, 1.7 x
of-registry layer (at a coverage just below second layer
torr), c) Two layer film (at a
condensation ; 47 K, 2.0 x
coverage just above second layer condensation; 47K,
3.6 x lod6torr), d ) Three or four layer film (47 K, 1.1X to-' torr).
3 . 3 MULTILAYERCONDENSATION. - AS the
torr
pressure was increased from 2.0 x
(Fig. 3b) to 3.6 x lo-" torr (Fig. 3 c ) for the
substrate at 47 K, a further increase is observed in
the Auger signal along with a marked flattening of
secondary electron background. An abrupt change
C4- 102
S . C. FAIN, Jr. AND M. D. CHINN
in the intensity of LEED spots and an increased
diffuse scattering was also seen (compare Fig. 4 a
and 4 b ) . As this is the first abrupt change in LEED
patterns observed past the deregistry transition,
both LEED and Auger seem to indicate that second
layer condensation occurred between the pressures
of figure 3 b and 3 c. This point at 47 K is shown as
a triangle in figure 1. Again, our data agrees well
with the extrapolation (line c in Fig. 1) of the data of
Thomy, Regnier, and Duval [2].
Auger electron spectra and LEED photographs
taken at 47 K and 9 x lop6torr are essentially the
same as in figures 3c and 4b. When the pressure
was further increased to 1.1 x
torr, the
secondary electron background flattened even
3d, the nravhite
LEED
further as shown in figure
spots diminished in intensity, and a further increase
in diffuse scattering in LEED was observed. We
believe that three or four Kr layers were present at
this pressure ;our observation is consistent with the
one observation of third layer condensation reported
by Thomy and Duval[2] (see points d in Fig. 1).
For a further increase in pressure to
3.6 x
torr, the Auger spectra were shifted
toward lower energies as expected for an insulator
which is charging up due to a high secondary
electron yield. The LEED pattern showed only
6 first order Kr diffraction spots as expected for a
thick, bulk-like Kr film. As mentioned in section 2,
shutting the Kr valve produced an immediate drop in
pressure, suggesting that thick Kr films were
accumulating on the coldest surfaces. The steady
state pressure observed agrees well with the
extrapolation (line e in Fig. 1) given by Pollack [14].
3 . 4 KRYPTONBILAYER LATTICE CONSTANT. The lateral nearest-neighbor distance in the bilayer
film determined from figure 4b is 4.02 % 0.02 A.
measured relative to the in-registry spacing of
4.26 h;. Our agreement with the bulk Kr-Kr distance
of 4.02 A at 47 K [14] may be due to a cancellation
of various substrate effects. In order to compare our
results with work done at higher temperatures, we
take the bulk expansion coefficient as a first
estimate for the layer expansion. Thus we would
expect the bilayer film to have a Kr-Kr spacing of
4.06 h; at 79 K and 4.08 A at 90 K.
Regnier, Thomy, and Duval [6] deduce from
volumetric isotherms a Kr-Kr spacing of
3.97 2 0.05 A at 90 K, a value much smaller than our
estimate. Moreover, since their coverage is below
second layer condensation, their value should be
greater than for a bilayer film. Their arguments
depend crucially on an interpretation of isosteric
heat measurements [IS] and on the assumption of no
second layer coverage.
A neutron diffraction peak observed by Marti, et
a / . [16] can be interpreted as arising from a
triangular lattice with Kr-Kr spacing of 4.03 h; (no
error stated) for a bilayer at 79 K. However, Marti et
al. prefer an interpretation which requires the
bilayer to retain registry with the graphite substrate
and hypothesize that some kind of imperfection in
their graphite is responsible for the disagreement
between their interpretation and our LEED
results [8].
-
FIG.4. - LEED patterns obtained with a 144 eV, 5 X lo-' A
primary electron beam. Due to non-normal incidence, only 4 of
the 6 graphite first order diffraction beams can be seen near the
edges of the pattern. Two graphite crystallites with slightly
different orientations are present in the electron beam. The six
triplets of spots arise from an out-of-registry Kr overlayer as
e s pressures for a and b are
explained in 191. ~ h e ' t e m ~ e r a t u rand
the same as for figure 3 b and 3c.
KRYPTON OVERLAYERS ON GRAPHITE
4. Conclusions. - Although the agreement of our
observations with extrapolations of vapor pressure
isotherms taken above
torr appears from
figure 1 to be satisfactory, several points should be
kept in mind.
a ) There is no a priori reason to expect the
differential entropy and isosteric heat at the various
two-dimensional phase transitions to be temperature
independent.
b ) Systematic errors in the temperature
measurements of 2 1 K could significantly affect
the agreement.
C ) For
comparison with truly isothermal
measurements the correction to the pressure
measured at the ionization gauge may be as large as
d ~ ~ / 3K
0 0where Ts is the substrate temperature [17].
d ) Due to the many orders of magnitude of
pressure shown in figure' 1 , disagreements of a
factor of 2 in pressure seem small.
C4-103
The most satisfactory result of the measurements
reported here is to confirm directly by diffraction
techniques most of the inferences made by Thomy,
Regnier, and Duval [2] from vapor pressure
isotherms at much higher pressures. Interpretations
of the lattice constant data near the in-registry to
out-of-registry transition are discussed elsewhere [9, 111. The idea that dislocations or structural modulations are important for out-ofregistry layers 19, 11, 181 will certainly receive more
experimental and theoretical attention in the next
few years.
AcknowIedgments. - We wish t o thank
G.D. Halsey for many useful discussions,
A. Thomy, F. Putnam, and P. Thorel for sending
data and articles prior to publication, and R. Diehl
for assistance in data analysis.
References
[I] THOMY,A., DUVAL,X., J. Chim. Phys. 66 (1969) 1966, 67
(1970) 286, 1101.
[2] THOMY,A., REGNIER,J., DUVAL,X., in Thermochimie,
Colloques Internationaux du CNRS (CNRS, Paris) 201
(1972) 511.
[3] LARHER,
Y., J. Chem. Soc. Faraday Trans. I70 (1974) 320.
[4] PUTNAM,
F. A., FORT,T., Jr., J. Phys. Chem. 79 (1975) 459
and in press ; PUTNAM,
F. A., FORT,T., Jr., GRIFFITHS,
R. B., J. Phys. Chem. (in press).
[5] DUVAL,X., THOMY,A., Carbon 13 (1975) 242.
[6] REGNIER,J., THOMY,A., DUVAL,X., 1. Chim. Phys. (in
press).
[7] KRAMER,H. M., SUZANNE,
J., Surf. Sci. 54 (1976) 659.
[8] CHINN,M. D. and FAIN,S. C., Jr., J. Vac. Sci. Technol. 14
(1977) 314.
191 CHINN,M. D., FAIN,S. C., Jr., Phys. Rev. Lett. 39 (1977)
146.
[lo] FRANCK, F. C., VANDER MERWE,J. H., h ~
R. SOC.
. A 198
(1949) 205, 216.
[I 11 VENABLES,J. A., SCHABES-RETCHKIMAN,
P., These proceedings.
[12] TAYLOR,N. J., Rev. Sci. Instrum. 40 (1969) 792.
[13] ROSENBAUM,
R. L., Rev. Sci. Instrum. 40 (1969) 577.
[14] POLLACK,G. L., Revs. Mod. Phys. 36 (1964) 748.
1151 REGNIER,J . , ROUQUEROL,
J., THOMY,A., J. Chim. Phys. 3
(1975) 327.
[I61 MARTI,C. L., CROSET,B., THOREL,P., COULOMB,
J. P.,
Surf. Sci. 65 (1977).
[I71 EDMONDS,
T., HOBSON,
J. P., 3. Vac. Sci. Technol. 2 (1965)
182.
[18] N o v ~ c o ,A., MCTAGUE,J. P., Phys. Rev. Lett. 38 (1977)
1286 and these proceedings.
DISCUSSION
X. DUVAL. - How did you estimate the
correction due to the second layer ?
S. C. FAIN.- We extrapolate the linear part of
the isotherm between B, and the onset of second
layer condensation back to zero pressure. The difference between this line and the intercept at zero
pressure is what we call a second layer correction.
This method was suggested to us by Halsey and is
mentioned in his paper to be published in J. Phys.
Chem. The correction also includes adsorption on
edges and defects.
J. SUZANNE.
- We studied the intensity of
superstructure spots Kr/Gr versus temperature at
constant pressure. It appeared that starting from the
temperature of formation of the first solid layer, the
intensity is first constant then decreased abruptly
(published in Surf. Sci. 54 (1976) 659). We gave as an
explanation the in-registry -+ out-of
registry
transition. We could not see the spot splitting but
what you found could explain what we measured
and our explanation would be right.
S. C. FAIN.- The aperture and positioning of
your photometer are important for interpretation of
C4-104
S. C. FAIN, Jr. AND M. D. CHINN
such a measurement. In some preliminary measurements, in which a photometer was focused on an
area that included all three spots for an out-ofregistry layer, M. D. Chinn and I found a larger total
intensity at constant temperature for an out-ofregistry overlayer with large misfit compared to an
in-registry overlayer. (We did not do measurements
close to the transition). As Bienfait pointed out, if
your photometer saw only the center of the
in-registry spot, then a decrease in intensity for an
out-of-registry layer would be expected as the spots
move away from the in-registry position.
C. MARTI.- What happens on the part of your
surface that does not give an ordered LEED
pattern ?
S. C. FAIN. - We can of course provide no
information on the lattice constant for the
disordered parts of the surface. Auger spectroscopy
will detect atoms in both the ordered and disordered
parts of our surface. For example, Auger spectra
from an area which showed a well defined registered
LEED pattern still showed some Kr after the
pressure was reduced enough to eliminate the
ordered LEED pattern. This could be explained by a
higher adsorption energy on the parts of the surface
that give no overlayer LEED pattern, as expected
for steps and kinks.
F. A. PUTNAM.-Do you have plans to do LEED
intensity measurements to determine the overlayer
substrate spacing ?
S. C. FAIN. -Yes. The amount of measurements
and calculations necessary to determine this
distance can be quite extensive. We have made
measurements of the energy variation of the
specular beam (at one angle) for a clean graphite
crystal and for a Kr overlayer on graphite. The
changes in the energy dependence are rather subtle,
as expected if the overlayer-substrate distance is
close to the graphite interlayer spacing. This result is
consistent with conclusions of Marti et al. regarding
overlayer-substrate spacing.
M. NIELSEN.- Your LEED results on Kr films
show that Kr has structures with a continuous range
of lattice parameter just above the registered phase
density. Can you (as we did for D,, H, layers)
correlate the lattice spacing and the coverage
(filling) ?
S. C. FAIN. - In order to minimize perturbation
of the Kr layers, we must use low current for our
Auger measurements. This prevents our making
coverage measurements of sufficient accuracy to
make detailed conclusions about lattice spacing versus coverage as you are able to do. Your measurements for D, layers are very beautiful ; I hope you
will publish soon the more detailed measurements
near the transition which you mentioned to me
privately.