Adsorption and desorption of Se on Si(100)2×1: Surface restoration

Article abstract of DOI:10.1016/S0039-6028(00)00759-7

This work entails a study of the adsorption of elemental Se on the reconstructed Si(100)2×1 surface. The investigation took place in an ultra high vacuum (UHV) by low energy electron diffraction (LEED), Auger electron spectroscopy (AES), thermal desorption spectroscopy (TDS) and work function (WF) measurements. The adsorption of one monolayer (1 ML) of Se at room temperature (RT) causes the transition of the reconstructed Si(100)2×1 surface to its original bulk terminated Si(100)1×1 configuration, while Se adatoms form a 1×1 structure by breaking the Si-Si dimer bonds. The Si-Se bond is strong (E_b = 2.97 eV/atom), resulting in the formation of a SiSe compound. Above 1 ML, Se forms a SiSe₂ compound with E_b = 2.67 eV/atom. The heating that follows causes the desorption of Se up until 1000 K, where Θ_Se = 0.5 ML, and the Si(100)1×1 structure is changed back to the reconstructed Si(100)2×1 with the Se forming a 2×1 structure. The models of Se(1×1)/Si(100)1×1 and of the Se(2×1)/Si(100)2×1 structures are given.

Full text of DOI:10.1016/S0039-6028(00)00759-7

Surface Science 466 (2000) 173–182
www.elsevier.nl/locate/susc
Adsorption and desorption of Se on Si(100)2×1:
surface restoration
A.C. Papageorgopoulos *, M. Kamaratos
Department of Physics, University of Ioannina, P.O. Box 1186, GR-451 10 Ioannina, Greece
Received 13 April 2000; accepted for publication 9 August 2000
Abstract
This work entails a study of the adsorption of elemental Se on the reconstructed Si(100)2×1 surface. The
investigation took place in an ultra high vacuum (UHV) by low energy electron diffraction (LEED), Auger electron
spectroscopy (AES), thermal desorption spectroscopy (TDS) and work function (WF) measurements. The adsorption
of one monolayer (1 ML) of Se at room temperature (RT) causes the transition of the reconstructed Si(100)2×1
surface to its original bulk terminated Si(100)1×1 configuration, while Se adatoms form a 1×1 structure by breaking
the SiMSi dimer bonds. The SiMSe bond is strong (E =2.97 eV/atom), resulting in the formation of a SiSe compound.
b
Above 1 ML, Se forms a SiSe compound with E =2.67 eV/atom. The heating that follows causes the desorption of
2
b
Se up until 1000 K, where H =0.5 ML, and the Si(100)1×1 structure is changed back to the reconstructed
Se
Si(100)2×1 with the Se forming
a
2×1 structure. The models of Se(1×1)/Si(100)1×1 and of the
Se(2×1)/Si(100)2×1 structures are given. © 2000 Elsevier Science B.V. All rights reserved.
Keywords: Auger electron spectroscopy; Diffusion and migration; Low energy electron diffraction (LEED); Silicon; Surface relaxation
and reconstruction; Thermal desorption spectroscopy; Work function measurements
1
. Introduction
surfaces [7–10]. Sulfur, selenium and tellurium
deposition on semiconducting surfaces proved to
be of particular interest, not only for passivation
purposes, but also for their theoretically predicted
ability to remove the initial dimer reconstruction
on the (100) surface of Si and Ge [11–21]. One
report claims that adsorption of sulfur on clean
Ge(100)2×1 changes the 2×1 structure to 1×1,
which is regarded as an ideal terminated surface
Silicon and other semiconductors, such as GaAs
and InP, are well known for their potentially wide
use in high speed electronics and long wavelength
optoelectrical circuits [1,2]. They have also demon-
strated great value, mainly in space technology, as
solar cells (photovoltaics) [3]. Their efficiency,
however, is reduced by electron, X-ray and
c-radiation damage [4–6]. To protect the semicon-
ducting surfaces without reducing their efficiency,
the above semiconductors are passivated. This is
done by depositing thin dielectric films such as
gallium sulfide, indium sulfide and sulfur on the
[
22] although, according to another study [23],
the ideal terminated surface of Ge(100)1×1 may
not occur with the suppression of the 2×1 recon-
struction upon S adsorption. Moriarty et al. [24]
reported that room temperature adsorption of S
resulted in the formation of an overlayer with an
underlying Si(100) which retained the 2×1 recon-
struction. They also mentioned that annealing of
*
Corresponding author.
0
039-6028/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S0039-6028 ( 00 ) 00759-7

1
74
A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
the S/Si(100)2×1 surface to 325°C led to the
desorption of the sulfur overlayer, and the appear-
ance of the coexisting c(4×4) and 2×1 surface
reconstruction. Recent reported observations [13],
however, suggest that adsorption of elemental S
on Si(100)2×1 initially forms a 2×1 (hemisulfide)
structure up to 0.5 ML. Subsequently, the surface
changes from a 2×1 to a 1×1 (monosulfide)
structure in the coverage range 0.5–1 ML. Higher
amounts of deposited S lead to the formation of
a second layer, whereas part of the S diffuses into
the bulk of the Si substrate. Cakmak and Srivasta
have, very recently, calculated the location of S
atoms above the Si(100) surface, and their results
were in excellent agreement with the above pro-
posed models of the hemisulfide (2×1) and mono-
sulfide (1×1) structures [25]. Bringans and
Olmstead [26] investigated the Se/Si(100)2×1
system with low energy electron diffraction
The uncertainty in the WF measurements was
0.02 eV. Elemental Se was evaporated by thermal
dissociation of WSe single crystal flakes enclosed
2
in envelopes made by thin Ta plates which were
heated by passing current. During dissociation of
WSe the W remained in the Ta envelope, while
2
Se was evaporated through holes made at the edge
of the envelope. By plotting QMS signal intensity
versus time for both atomic and molecular Se,
with a constant current running through the Ta
envelope, and comparing the areas under these
graphs we measured the atomic Se to be 30% of
the total yield. The Si(100) single crystal substrate
was cleaned by heating at 1300 K, while the tem-
perature of the sample was measured by a Cr–Al
thermocouple calibrated by an infrared pyrometer
in a temperature range 900–1300 K. The estima-
tion of the Se coverage on Si(100) surface was
based on a correlation of LEED, AES and TDS
measurements. The surface atomic density of 1 ML
Se on Si(100) is considered to be equal to that of
the outermost layer of Si: 6.8×1014 atoms/cm2.
(
LEED) and ultraviolet photoelectron spectro-
scopy (UPS) measurements, where they concluded
that, at elevated temperatures (300°C), Se forms
SiSe . Annealing at higher temperatures (500–
2
5
50°C) results in only a submonolayer of Se atoms
remaining in bridge sites with a 1×1 structure,
while with theoretical calculations the conclusion
was drawn that both Se and S change the 2×1
reconstructed Si(100) surface to its 1×1 form
3. Experimental results
3.1. Auger measurements — LEED
[
11,12,21]. Tellurium deposition on Si(100)2×1
Fig. 1 shows the peak to peak height (Ap-pH)
of the Se (47 eV) and Si (91 eV) Auger peaks as
a function of the number of Se doses deposited on
the Si(100)2×1 surface at room temperature (RT)
and with subsequent heating in steps of 100 K.
These measurements are shown in correlation with
the corresponding observed LEED patterns. As
seen in Fig. 1, the Ap-pH of the Se (47 eV) peak
increases almost linearly with an increasing
number of Se doses. Near the eighth dose of Se
deposition the curve forms a break and the Se
peak continues to increase linearly with a lower
rate up to 15 doses of deposition. Above the
fifteenth dose the Se peak height increases with an
even lower rate. The Ap-pH of Si decreases in a
similar manner, and the corresponding curve
shows breaks at the eighth and fifteenth doses of
Se deposition (Fig. 1). This behavior is characteris-
tic of layer by layer growth [27]. LEED observa-
tions show that Se causes a weakening of the half
grows in a Stranski–Krastanov mode, and forms
a 1×1 structure [15–20]. From the above discus-
sion it is clear that additional study of the S, Se
and Te on Si(100)2×1 is necessary. In this contri-
bution we investigate the adsorption of Se on
Si(100)2×1 at room and elevated temperatures
by Auger electron spectroscopy (AES), LEED,
thermal desorption spectroscopy (TDS) and work
function (WF) measurements, in ultra high
vacuum (UHV).
2
. Experimental
The experiments were performed in a UHV
chamber, equipped with AES, LEED, a quadru-
pole mass spectrometer for TDS, with a heating
rate of b=15 K/s, and WF measurements. For the
latter measurements we used the diode method.

A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
175
Fig. 1. Auger peak to peak height (Ap-pH) of Se (47 eV) and Si (91 eV) as a function of the number of Se doses deposited on the
Si(100)2×1 surface at RT and with subsequent heating in steps of 100 K.
order spots of the Si(100)2×1 pattern. The latter
disappear near the first break (eighth dose of Se
deposition) of the Se Auger curve (Fig. 1), while
the integer order spots become more intense, and
the pattern changes to a good 1×1 configuration.
Between the eighth and fifteenth doses of Se depos-
ition the LEED pattern does not show any sub-
stantial change and the 1×1 configuration
remains. Higher Se coverages cause a decrease in
the intensity of the integer order spots. The latter
disappear for deposited Se amounts higher than
temperature range the Se peak decreases rapidly,
the peak of the Si substrate increases proportion-
ally to this, while the 2×1 LEED pattern reappears
near 1000 K. The 2×1 LEED pattern becomes
sharper with increasing temperature up to 1300 K.
3.2. Work function measurements — LEED
Fig. 2 shows the work function changes of the
Si(100)2×1 surface during Se deposition at RT,
and after heating up to 1300 K, in correlation with
the LEED observations. The work function
increases almost linearly with Se deposition up to
eight doses of Se. The work function change, at
eight doses of Se deposition, is 0.52 eV. Higher Se
deposition causes a decrease of the work function
by 0.07 eV. Subsequent gradual heating of the
system causes an increase of the work function to
24 doses, while the background of the pattern
becomes almost completely dark (approaching
zero). Subsequent heating (Fig. 1) causes a small
increase of the Si Ap-pH, while the Se Ap-pH does
not show any substantial change up to 800 K. The
LEED pattern depicts the 1×1 configuration once
again in the range 400–800 K. In the 900–1100 K

1
76
A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
Fig. 2. Work function change of the Si(100)2×1 surface during Se deposition at RT, and after heating up to 1300 K, in correlation
with LEED observations.
its maximum value (0.52 eV) at 600 K (point A).
Heating the system to even higher temperatures
decreases the work function and the curve forms
a knee at about 1000 K (point B). The work
function value of this knee is very close to that of
The third peak represents added Se dosage corre-
sponding to a third disordered layer of Se bonded
with Si, by diffusing into the bulk or by drawing
the Si out. It should be noted that the SiSe TD
spectra consisted of the same peaks with approxi-
mately the same intensities as the aforementioned
TD spectra for Se, and are not shown here. Fig. 4
shows a series of thermal desorption spectra of
0.5 ML of Se deposited at room temperature, while
the LEED pattern changes from a 1×1 to a 2×1
configuration. Further heating causes a drastic
decrease of the work function which passes
through a minimum value (−0.08 eV) at 1200 K
and reaches the work function of the clean
Si(100)2×1 surface at 1300 K.
SiSe (186 amu) after Se depositions on the Si(100)
2
surface. As is seen in this figure, the SiSe thermal
2
desorption spectrum initially shows a single peak
at 990 K for Se depositions above eight doses. At
relatively very high Se coverages (greater than 25
doses), new broad peaks appear at about 600 K.
These low temperature peaks increase with greater
amounts of deposited Se.
3.3. TDS measurements
Fig. 3 shows a series of thermal desorption
spectra of Se (79 amu) after Se depositions on
Si(100)2×1 surfaces. The heating rate for all
spectra was 15 K/s and it was constant in the
whole temperature range of 300–1300 K. For an
amount less than or equal to eight doses of Se the
Fig. 5 shows the thermal desorption spectrum
of SiSe (107 amu), after Se deposition on the
Si(100)2×1 surface, compared with those of Se
(79 amu) and SiSe (186 amu). The SiSe thermal
2
desorption spectrum has two peaks. The high
spectrum gives a peak b at 1125 K. For Se cover-
ages greater than the eighth dose, the spectrum
energy peak coincides with the Se peak b at
1125 K for Se coverages equal to or less than eight
1
1
shows a second peak b of Se at a lower temper-
doses. For Se deposition higher than eight doses
this peak saturates and a second peak appears at
1010 K which coincides with the low energy Se
2
ature of 1010 K, and finally a third Se peak b at
3
the relatively low temperature of 610 K. The first
two peaks, from highest to lowest energy, corre-
spond respectively to the thermal desorption of
the first and second layers of Se on Si(100)2×1.
peak b . The intensity of this new SiSe (107 amu)
double peak increases continuously with Se depos-
ition. Selenium deposition higher than 15 doses
2

A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
177
Fig. 4. A series of thermal desorption spectra of SiSe (186 amu)
2
Fig. 3. A series of thermal desorption spectra of Se (79 amu)
after Se deposition on Si(100)2×1 surfaces. The numbers next
to the curves indicate dosage.
after Se deposition on Si(100)2×1 surfaces. The numbers next
to the curves indicate dosage.
causes the appearance of a third, low intensity,
wide peak at about 600 K (not shown here).
4. Discussion
The first issue which needs to be addressed is
the determination of Se coverage. The Auger meas-
urements, according to Fig. 1, indicate that Se on
Si(100)2×1 grows in a layer by layer mode [27].
The first layer is completed near the eighth dose
of Se deposition when the first Auger break occurs,
and the 1×1 LEED pattern appears. The TDS
measurements of this study show, in addition, that
up to the eighth dose of Se on Si(100), only Se
Fig. 5. Thermal desorption spectrum of SiSe (107 amu), after
Se deposition on the Si(100)2×1 surface, compared with those
and SiSe are detected. SiSe is initially detected in
2
the TDS measurements, on the other hand, after
of Se (79 amu) and SiSe (186 amu) at 20 doses of Se.
2

1
78
A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
the twelfth dose (Fig. 4). The WF increases lin-
early up to the eighth dose, as well. This linear
behavior of the work function curve is a
clear indication of a constant dipole moment,
which in turn indicates that the sites occupied by
the Se on the Si substrate are all structurally
similar surface sites up the eighth dose. Combining
the above information, we may conclude that, at
the eighth dose of Se, one Se adatom corresponds
to one Si atom on the substrate’s outermost layer.
We may conclude, therefore, that at the eighth
dose the first layer of Se is completed with a
coverage of 1 ML (6.68×1014 atoms/cm2), which
is consistent with the Auger measurements. Near
the completion of the first Se layer the LEED
pattern gradually changes from a 2×1 to a 1×1
configuration. The adsorption of elemental Se at
room temperature, in other words, causes the
transition of the reconstructed Si(100)2×1 surface
to its original bulk terminated Si(100)1×1 state.
The restoration of the 1×1 structure can be
explained by the adsorption of Se atoms on the
bridge and cave sites which cause the break of the
SiMSi dimer bonds. This model is consistent with
those proposed from previous experimental meas-
urements of S and Te, on Ge(100)2×1 and on
Si(100)2×1 [13,20,26,28], and via theoretical cal-
culations [11,21]. Although the most likely sites
to initially attract divalent Se adatoms are the
dimer bridge sites, this analysis is not conclusive
from the data in this study. The clarity of the 1×1
LEED pattern at a 1 ML Se coverage, however,
points to an ordered 1×1 overlayer. Fig. 6b shows
a side-view schematic, while Fig. 7b shows the top-
view schematic of 1 ML of the Se(1×1) structure
on a Si(100)1×1 surface.
The second monolayer amount of Se is depos-
ited between the eighth and sixteenth doses. The
linearity of the Auger curve between breaks
(Fig. 1) is an indication that this amount forms
an even layer. Between the completion of the first
and the second Se layer, the 1×1 pattern remains
unaltered. This occurrence may lead to the conclu-
sion that the Se atoms of the second layer are also
arranged in a 1×1 structure. The lowering of the
work function for Se coverage above 1 ML
(Fig. 2), however, suggests that the deposited Se
above 1 ML may be submerged into the bulk just
under the surface layer of the Si. It should be
emphasized that the second monolayer of Se,
Fig. 6. Side-view schematic of (a) clean Si(100)2×1, (b) 1 ML of Se with the (1×1) structure on the Si(100)1×1 substrate, (c)
ML of Se on Si(100)1×1, (d) 0.5 ML of Se with the 2×1 structure on Si(100)2×1 after heating the Se(1×1)/Si(100)1×1 surface
to 1000 K.
2

A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
179
explain the double Se(1×1) layer configuration
on the Si(100)1×1 surface is shown by the side-
view schematic in Fig. 6c. The top view appears
the same as that in Fig. 7b, since the Se atoms of
the second layer are correspondingly under those
of the first.
The TD measurements in Fig. 3 indicate that
Se is deposited on Si(100)2×1 in three adsorption
states: b , b and b . Considering the desorption
1
d
2
3
energy E to be equal to the binding energy E of
b
an adsorbate, we can estimate the latter with the
use of Redhead’s equation [29]:
E =RT [ln(nT /b)−3.64]
(1)
d
p
p
where T is the temperature, b=15 K/s is the
p
constant heating rate, R is the gas constant and n
is the pre-exponential factor which for Si substrates
is 1013 s−1. The binding energies which correspond
to b , b and b desorption states are 2.97 eV/atom,
1
2
3
2
.67 eV/atom and 1.58 eV/atom, respectively. The
relatively high binding energies and the formation
of the SiSe and SiSe compounds suggest a cova-
2
lent bond between Se and Si. As shown in Fig. 5,
the two TD peaks formed by the first two Se
monolayers are detected by the mass spectrometer
as SiSe as well as Se. This may indicate that the
b and b thermal desorption peaks are formed
1
2
from the possible dissociation of SiSe . Whether
2
or not this actually occurs, the aforementioned
double layer of Se (at 2 ML) (Fig. 6c) fulfils the
SiSe stoichiometry with the Si substrate, and
2
maintains a different covalent binding state from
that of the SiMSe bond up to a coverage of 1 ML.
Above the sixteenth dose Auger data indicates
the formation of a third Se layer. Adsorption of
Se above 2 ML causes a gradual disappearance of
the integer order spots of the Si(100)1×1 structure
while the background becomes almost completely
dark. For Se coverage above 2 ML (Figs. 3 and
Fig. 7. Top-view schematic of (a) clean Si(100)2×1, (b) 1 ML
of Se with the (1×1) structure on Si(100)2×1 after heating
the Se(1×1)/Si(100)1×1 surface to 1000 K.
according to these observations, is an even
underlayer with a periodic 1×1 arrangement. The
apparent periodicity and even distribution of the
5), TDS gives the b low energy peak of Se and
3
underlayer, and the appearance of the b TD peak
that of SiSe at about 600 K. The presence of the
2
2
after the eighth dose (which is shown further in
latter SiSe may indicate that some adsorbed Se
atoms either diffuse into the bulk, bonding loosely
2
the text to maintain a binding energy of a covalent
nature) point to the likelihood of Se insertion into
the SiMSi bonds. We cannot, therefore, exclude
the possibility that the two layers of Se may
actually form a double-layer SiMSe lattice com-
pound with the substrate. A possible model to
with Si atoms, or reside on the surface where they
cause a diffusion of Si atoms from the bulk to
the surface.
Our data clearly shows a dominating SiSe peak
2
at 990 K, distinct from peaks b and b . It is most
1
2

1
80
A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
likely that the SiMSe covalent bonds at the comple-
while the rest resides in the substrate bulk (Figs. 6b
and 7b).
tion of the second monolayer, when the SiSe
2
stoichiometry is fulfilled, maintain the binding
Further heating (to 990 K) removes the bulk
energy represented by the b TD peak (Fig. 3).
We may thus consider the resulting double layer
SiSe (corresponding to the Fig. 4 TD peak) and
brings the coverage to 2 ML, according to AES
2
2
Si/Se surface lattice structure to be a substrate-
data. LEED observations show a disordered sur-
face state most likely caused by the desorption of
dependent SiSe pseudocompound. The higher
2
binding energy of TD peak b probably corres-
the bulk SiSe , while the WF curve becomes non-
1
2
ponds to the stoichiometry of a coverage of 1 ML
linear. At about 1000 K (approximately equal to
or less. Above the second monolayer the b peak
continues to increase, which indicates (in accor-
the desorption temperature of peak b ), the LEED
pattern shows a strong 2×1 pattern, and the WF
2
2
dance with work function and Auger measure-
ments) that Se continues its bonding with Si below
curve exhibits a characteristic knee as the AES
plot indicates 1 ML. The corresponding room tem-
perature coverage for the WF value at this (point
B) is 0.5 ML. Data here indicates that an amount
of 1 ML of Se is distributed on (and in) the
substrate, while it maintains the strong covalent
the second layer. The b peak saturates at 20 doses
2
(
an equivalent of 2.5 ML). It is at this point that
the 990 K SiSe begins to dominate. It must be
2
noted, however, that the SiSe detected at a cover-
2
age above 2 ML at 990 K may actually form
bonding which characterizes TD peak b . The WF
1
during the thermal desorption process, where
sufficient activation energy through heating is pro-
vided. This high coverage (above 2.5 ML) is thus
marked by Se diffusion into the bulk, which may
information indicates a 0.5 ML Se coverage on the
substrate surface, distributed in such a way as to
allow the observation of a clear 2×1 LEED
pattern. We believe that, at this point, the
Si(100)2×1 reconstruction is reproduced with the
Se atoms bound through the dangling bonds to
the two Si atoms of each dimer. This is consistent
with the fact that Se is divalent. When all the
dimer bridge sites are filled, moreover (on the
Si(100)2×1 surface), the maximum surface cover-
age is 0.5 ML. The model of Se(2×1) on
Si(100)2×1 is shown by the side-view and top-
view schematics in Figs 6d and 7c, respectively.
We can further support the formation of the
Se(2×1) structure on Si(100)2×1 by correlating
it with the work function variation in the 600–
1000 K temperature range between the points A
and B in Fig. 2. Fig. 8 shows the work function
variation during heating the Se(1×1)/Si(100)1×1
surface structure from 600 K (point A) to 1000 K
(point B) where the above structure has changed
to the Se(2×1)/Si(100)2×1. The sizes of the Se
and Si atoms in this figure are considered in
proportion to their real size with the SeMSi bond
being covalent, as shown previously. The covalent
contribute (with the loosely bound Se and SiSe )
2
to the disappearance of the 1×1 LEED pattern.
The SiMSe double lattice layer (with SiSe stoichi-
2
ometry) nevertheless remains over the disordered
layers as a crystal lattice ‘crust’, and is again
‘revealed’ upon heating (as explained below).
Gradual heating of more than three equivalent
monolayers of Se on Si(100)2×1 initially causes
the desorption of the loosely bound SiSe overlayer
2
and the reappearance of the 1×1 pattern. The
work function, at this point, reaches its maximum
value of 0.52 eV, corresponding to a 1 ML room
temperature coverage, while Auger measurements
indicate an equivalent coverage over 3 ML. The
apparent inconsistencies in data give rise to the
question of how LEED and WF data can indicate
a clear 1×1 surface restoration at 1 ML when
both TDS and AES data clearly substantiate the
presence of at least 3 ML of Se on the Si(100)1×1
surface. To reconcile the above data one need only
consider that LEED and WF give information
relating mostly to the state of the sample surface,
while AES and TDS show the total amounts of
the adsorbate. We believe, therefore, that the most
likely answer is that at 600 K 1 ML of Se remains
on the surface, suppressing the reconstruction,
˚
radii of Se and Si are 1.16 and 1.11 A, respectively.
The dipole lengths of the dipole moments which
correspond to the structures at the points A and
˚
B (Fig. 8) are d=1.6 and 2.1 A, respectively. The
work function value for a coverage H of an

A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
181
Fig. 8. The work function variation during heating of the Se(1×1)/Si(100)1×1 surface structure from 600 K (point A) to 1000 K
point B) where the above surface structure has changed to the Se(2×1)/Si(100)2×1 configuration.
(
adsorbate on a substrate is given by
than that of S on the same surface. This mobility
is, most likely, obtained by heating which provides
the required activation energy for Se atoms to rise
up to the bridge sites.
W=2pPN H=2pqN H
(2)
s
s
where P=qd is the electric dipole moment, N the
density of the atoms in the outermost layer of the
s
adsorbate [30] and q the charge of each adsorbed
atom. The charge q remains constant due to the
covalent nature of the SiMSe bonds. If the d=
5. Conclusions
˚
16 A remains constant and the surface H changes
In this paper we study the adsorption of elemen-
tal Se on the Si(100)2×1 surface at room and
elevated temperatures by LEED, AES, TDS and
WF measurements. The adsorption of 1 ML of
elemental Se at RT causes the transition of the
reconstructed Si(100)2×1 surface to its original
bulk terminated Si(100)1×1. The deposited
monolayer of Se forms a 1×1 structure on
Si(100)1×1 by breaking the SiMSi dimer bonds.
from 1 to 0.5 ML, the W should decrease linearly
to a value lower than that at B. Based on the
above argument, the only way to explain the
gradual slope decrease of the W curve between A
and B may be attributed to an increase of the
˚
dipole length to d=2.1 A, which is due to the
formation of a Se(2×1) structure on Si(100)2×1,
as shown in Figs. 6d and 7c.
From our LEED observations, there is no clear
indication of Se(2×1) formation during Se depos-
ition at RT such as for S on Si(100)2×1 [13].
The strong binding energy of Se on Si(100)2×1
The SiMSe bond is strong (E =2.97 eV/atom),
suggesting a covalent bonding which results in the
formation of a SiSe surface compound. Above
1 ML and up to the completion of 2 ML, deposited
Se is submerged into the bulk near the surface of
the Si(100)1×1 substrate. This process results in
b
(
E =2.97 eV/atom) is greater than that of S on
b
Si(100)2×1 (E =2.24 eV/atom) [13], and proba-
bly restricts the mobility of the Se atoms so that
b
the formation of a SiSe stoichiometry with a
2
their diffusion onto the dimer bridge sites is less
binding energy E =2.67 eV/atom. With further
b

1
82
A.C. Papageorgopoulos, M. Kamaratos / Surface Science 466 (2000) 173–182
No. 282, Materials Research Society, Pittsburgh, PA,
993, p. 689.
[11] E. Kaxiras, Phys. Rev. B 43 (1991) 55.
12] P. Kruger, J. Pollmann, Phys. Rev. B 47 (1993) 1898.
[13] A. Papageorgopoulos, A. Corner, M. Kamaratos, C.A.
Papageorgopoulos, Phys. Rev. B 55 (1997) 55.
deposition, the Se atoms are loosely bound on top
1
of the SiSe layer (E =1.58 eV/atom), which may
2
b
continue its formation through a diffusion of Si
atoms from the bulk to the surface. The following
heating procedure caused the gradual desorption
[
[
14] M. Gothelid, G. Leleay, C. Wigren, M. Bjorkqvist, M.
Rad, U.O. Karlosson, Appl. Surf. Sci. 115 (1997) 87.
15] S. Di Nardo, L. Lozzi, M. Passacantando, P. Picozzi, Surf.
Sci. 331 (1995) 569.
of Se (in SiSe or SiSe compounds) up until 1000 K
2
where the Se surface coverage becomes 0.5 ML
[
and the Si(100)1×1 structure changes back to the
reconstructed Si(100)2×1 surface. The models of
clean Si(100)2×1, Se(1×1)/Si(100)1×1, and of
the Se(1×2)/Si(100)2×1 structures are given in
Figs. 6 and 7.
[16] S. Santucci, S. Di Nardo, L. Lozzi, M. Passacantando, P.
Picozzi, Surf. Sci. 352–354 (1996) 1027.
[
[
[
[
[
[
17] M.R. Bennett, A.A. Cafolla, J.W. Cairns, C.J. Duns-
combe, R.H. Williams, Surf. Sci. 360 (1996) 187.
18] K. Tamiya, T. Ohtani, Y. Takeda, T. Urano, S. Hongo,
Surf. Sci. 408 (1998) 268.
19] T. Ohtani, K. Tamiya, Y. Takeda, T. Urano, S. Hongo,
Appl. Surf. Sci. 130–132 (1998) 112.
20] S.A. Yoshikawa, J. Nogami, C.F. Quate, P. Pianetta, Surf.
Sci. 321 (1994) L183.
21] Y.-Y. Zhao, P.-L. Cao, G. Lai, J. Phys. C: Condens.
Matter 10 (1998) 7769.
References
[
1] A.N. MacInnes, M.B. Power, A.R. Barron, Appl. Phys.
Lett. 62 (1993) 711.
22] T. Wesser, A. Bogen, B. Konrad, R.D. Schnell, C.A.
Shung, W. Moritz, W. Steinmann, Surf. Sci. 201 (1988)
[
[
2] R.R. Chang, D.L. Life, Appl. Phys. Lett. 53 (1988) 134.
3] R.K. Jain, G.A. Laundis, in: 22nd IEEE Photovoltaic Spe-
cialists Conference, Las Vegas, NV, 7–11 October (1991) .
4] M. Yamaguchi et al., Appl. Phys. Lett. 4 (1984) 432.
5] A. Yamamoto, M. Yamaguchi, C. Vemura, Appl. Phys.
Lett. 44 (1984) 611.
2
45.
[
[
23] K.T. Leung, L.J. Terminello, Z. Hussain, X.S. Zhang, T.
Hayashi, D.A. Shirley, Phys. Rev. B 38 (1988) 8241.
24] P. Moriarty, L. Koenders, G. Hughes, Phys. Rev. B 47
[
[
(
1993) 15 950.
[
[
25] M. Cakmak, G.P. Srivasta, J. Appl. Phys. 84 (1999) 6070.
26] R.D. Bringans, M.A. Olmstead, Phys. Rev. B 39 (1989)
[
[
6] I. Weinberg, C.K. Swartz, R.E. Half, unpublished results.
7] R. Leonelli, C.S. Sundaraman, J.F. Currie, Appl. Phys.
Lett. 57 (1990) 2678.
1
2 985.
[
[
27] C. Argile, G.E. Rhead, Surf. Sci. Rep. 10 (1989) 277.
28] K. Newstead, A.W. Robinson, A. Patchett, S. d’Addato,
N.P. Prince, R. McGrath, R. Whittle, E. Dudzik, I.T.
McGovern, Surf. Sci. 287/288 (1993) 317.
[
[
8] S. Shicoda, H. Okada, H. Hayashi, J. Appl. Phys. 69
(
1991) 2717.
9] A.N. MacInnes, M.B. Power, A.P. Barron, Chem. Mater.
(1992) 11.
4
[29] P.A. Redhead, Vacuum 12 (1962) 203.
[30] D.P. Woodruff, T.A. Delchar, Modern Techniques of Sur-
face Science, Cambridge University Press, Cambridge,
1986.
[
10] M.B. Power, A.N. MacInnes, A.P. Hepp, A.R. Barron, in:
C.R. Abernathy et al. (Eds.), Chemical Perspectives of
Microelectronic Material IIIMRS Soc. Symp. Proc.