Adsorption chemistry of cyanogen bromine and cyanogen chlorine on silicon(100)

Article abstract of DOI:10.1021/jp022195y

The adsorption and decomposition of cyanogen halides, XCN (X = Br, Cl), on Si(100) is investigated utilizing X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS). For submonolayer exposures, XPS indicates that the CN triple bond of XCN remains intact upon adsorption at 100 K. The UPS spectrum contains two peaks assigned to the π-electrons in the CN triple bond. The splitting indicates that some fraction of the XCN molecules adsorbs molecularly at low temperature. XPS analyses of the C 1s photoelectron peak following submonolayer exposure at low temperature suggest a greater fraction of BrCN (60%) adsorbs molecularly than ClCN (40%). XPS and UPS measurements at room temperature show that the X-CN bond breaks, while the CN bond remains intact during room-temperature adsorption on Si(100). Thus, the UPS spectrum of XCN adsorbed at room temperature on Si(100) contains a peak at 6.0 eV due to the unperturbed π electrons of the CN species. Upon annealing a CN-saturated Si(100) surface to higher temperatures, the UPS spectra indicate that the CN bond remains intact until approximately 700 K. Simultaneous changes in the C 1s and N 1s photoelectron peaks are consistent with the idea that CN bond cleavage is correlated with silicon carbide and nitride formation. These results are compared with a previous study of ICN adsorption on Si(100).

Full text of DOI:10.1021/jp022195y

7726
J. Phys. Chem. B 2003, 107, 7726-7732
Adsorption Chemistry of Cyanogen Bromine and Cyanogen Chlorine on Silicon(100)
P. Rajasekar, Evgueni B. Kadossov, Lucas Ward, Jennifer Lee Baker, and
Nicholas F. Materer*
Department of Chemistry, Oklahoma State UniVersity, Stillwater, Oklahoma 74078-3071
ReceiVed: October 7, 2002
The adsorption and decomposition of cyanogen halides, XCN (X ) Br, Cl), on Si(100) is investigated utilizing
X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS). For submonolayer
exposures, XPS indicates that the CN triple bond of XCN remains intact upon adsorption at 100 K. The UPS
spectrum contains two peaks assigned to the π-electrons in the CN triple bond. The splitting indicates that
some fraction of the XCN molecules adsorbs molecularly at low temperature. XPS analyses of the C 1s
photoelectron peak following submonolayer exposure at low temperature suggest a greater fraction of BrCN
(60%) adsorbs molecularly than ClCN (40%). XPS and UPS measurements at room temperature show that
the X-CN bond breaks, while the CN bond remains intact during room-temperature adsorption on Si(100).
Thus, the UPS spectrum of XCN adsorbed at room temperature on Si(100) contains a peak at 6.0 eV due to
the unperturbed π electrons of the CN species. Upon annealing a CN-saturated Si(100) surface to higher
temperatures, the UPS spectra indicate that the CN bond remains intact until approximately 700 K. Simultaneous
changes in the C 1s and N 1s photoelectron peaks are consistent with the idea that CN bond cleavage is
correlated with silicon carbide and nitride formation. These results are compared with a previous study of
ICN adsorption on Si(100).
Introduction
For both systems, silicon carbides and nitrides are formed after
annealing the Si(100) surface to temperatures greater than 800
K.
Recently, we have investigated the adsorption and decom-
position chemistry of cyanogen iodide (ICN) on Si(100) surfaces
both experimentally and theoretically.1,2 In this paper, the ICN
study is extended to include the surface chemistry of two
additional cyanogen halides, cyanogen bromide (BrCN) and
cyanogen chloride (ClCN), on Si(100). Lu and Lin have recently
reviewed reaction of C, N, and O containing molecules with
the Si(100) surface.3 In particular, Lin and co-workers4-7 have
studied hydrogen cyanide (HCN) and cyanogen (C2N2), on
Si(100) and Si(111) surfaces. Bu and Lin have also studied
another similar compound, s-triazine ((HCN)3), on the Si(100)
surface.8 For HCN, Lin and co-workers find that at high
exposures, HCN dimerizes and possibly polymerizes on the
Si(100) surface at low temperatures.5 Upon warming to 220 K,
iminium (HC-NH) and CN species are identified as decom-
position products. At 560 K, the CN species undergoes a
reorientation from an end-on to a side-on geometry with a
concomitant breakage of one bond to yield a doubly bonded
surface species. After annealing to temperatures above 600 K,
only the CH group, atomic N, and atomic H are left adsorbed
to the surface. Finally, at temperatures greater than 1000 K,
only a mixture of silicon nitride and carbide remains. In contrast,
both (HCN)3 and C2N2 molecules are more stable than HCN
on the Si(100) surface. At low exposures, C2N2 adsorbs in a
side-on geometry that is stable until 550 K. (HCN)3 is believed
to lie flat on the Si(100) and also is stable until 550 K. Above
this temperature, the NC-CN single bond in C2N2 breaks and
leaves the CN group, atomic C, and atomic N on the surface.
(HCN)3 also decomposes above 550 K to a complex mixture
of HCN, NH, and atomic H adsorbed on the Si(100) surface.
The adsorption and thermal chemistry of cyanogen iodide
(ICN) on Si(100) surfaces have been previously investigated
by ultraviolet photoelectron spectroscopy (UPS), X-ray photo-
electron spectroscopy (XPS), and temperature-programmed
desorption (TPD).1 The surface chemistry has also been modeled
by Kadossov et al.,2 utilizing a single-dimer silicon cluster. For
ICN on Si(100), the surface after a 0.2-langmuir ICN exposure
at low temperatures (100 K) consists of intact CN triple bond
species. XPS analysis of the C 1s photoelectron peak indicates
that up to 25% of the adsorbed ICN species have the IC bond
unbroken. Upon annealing to room temperature, the CN bond
remains intact, but all the IC bonds are completely broken on
Si(100). Upon further annealing the Si(100) surface to higher
temperatures, the UPS and XPS spectra indicate that the CN
triple bond starts to disappear at approximately 470 K. By 800
K, all adsorbed CN species decompose completely into silicon
carbide and nitride.
In this paper, we first briefly describe the experimental
apparatus along with details of the syntheses and purification
of BrCN and ClCN. The results are presented and then followed
by a discussion section where the adsorption chemistry of these
three cyanogen halides (ICN, BrCN, and ClCN) is compared
and contrasted. Finally, the conclusion is presented.
Experimental Apparatus
The experimental setup has been previously described.1
Briefly, it consists of a surface analysis system with a typical
base pressure of 2 × 10-10 Torr using an uncorrected nude
ionization gauge. The XPS measurements are performed by
using the Mg anode of a PHI 300 W twin anode X-ray source.
The resulting photoelectrons are detected by a PHI double-pass
* To whom correspondence should be addressed. Phone: (405) 744-
8671. E-mail: materer@okstate.edu.
10.1021/jp022195y CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/15/2003

Adsorption Chemistry of BrCN and ClCN
J. Phys. Chem. B, Vol. 107, No. 31, 2003 7727
cylindrical mirror analyzer (CMA) with a pass energy of 50
eV. UPS is implemented by using a differentially pumped He
discharge ultraviolet source (UG Microtech) to provide He(I)
radiation. The same double-pass cylindrical mirror analyzer with
a pass energy of 5 eV is used to detect the resulting photoelec-
trons. A copper foil provides the fermi level reference for the
UPS photoelectron energies. Silicon (100) wafers from Virginia
Semiconductor (100 mm diameter, arsenic doped, 0.01 Ω cm)
are cleaved into samples approximately 16 mm in length and 6
mm in width. The wafers are 0.5 mm thick. The method for
mounting the silicon sample on the ultrahigh vacuum (UHV)
manipulator, thermocouple mounting techniques, and sample
heating and cooling is found in ref 9. Before each experiment
the silicon samples are cleaned by repeated flashing to temper-
atures greater than 1400 K. This procedure is repeated until no
impurities (C, N, O in particular) are detected by XPS. All
exposures are reported in langmuirs, where 1 langmuir is equal
to 10-6 Torr s. Typical background pressures of 10-7 Torr are
used for the exposures.
To avoid oxygen and nitrogen contamination, the cyanogen
halide is placed in a small UHV compatible container attached
to a precision leak valve mounted on the load-lock chamber. A
dry ice/acetone mixture is used to cool the sample during
repeated freeze-pump-thaw cycles. The pump cycle reduces
the pressure in the container to 10-7 Torr by employing the
turbo molecular pump in the load-lock chamber. ICN is
synthesized using standard techniques10 and is further described
in ref 1. BrCN is obtained from Aldrich. Contact between BrCN
and the stainless steel container is avoided by first placing a
small quantity in a quartz tube, which in turn is slipped into
the container. Even with these precautions, cyanogen is found
to slowly form in the container and is thus a precedent
contaminant. Fortunately, the cyanogen can be removed before
each experiment by additional repeated freeze-pump-thaw
cycles.
Figure 1. Nitrogen and carbon 1s XPS spectra as a function of BrCN
exposure to the Si(100) surface at 100 K. The XPS intensities are
relative.
of a nitrile group. In addition to CN species found in CNx films,
adsorbed intact CN species can be prepared on Rh(110) by a
20-langmuir C2N2 exposure at 100 K followed by a 373 K
annealing treatment.14 The XPS spectrum of this CN covered
Rh(110) surface shows C 1s and N 1s photoelectron peaks
corresponding to adsorbed CN groups at 284.3 and 397.4 eV,
respectively.14
The C 1s photoelectron peak after a 0.3-langmuir BrCN
exposure (Figure 1) is slightly broader and skewed toward higher
binding energy. Following low ICN exposures to Si(100) at 100
K, a similar high-energy tail is observed for the ICN C 1s
photoelectron peak.1 Based on a combination of UPS and XPS
measurements, the high energy tail for the ICN C 1s photo-
electron peak is attributed to the presence of molecular ICN,
atomic I, and CN species on the Si(100) surface.1 Assuming a
similar adsorption model for BrCN, the BrCN C 1s peak can
be fit by using two Gaussians components. One component is
assigned to a molecular species at 286.5 eV and the other to a
dissociated species at 284.5 eV. The peak position of the
dissociated species is justified by the room temperature XPS
and UPS results discussed below. The analysis indicates that
about 60% is adsorbed molecularly (286.5 eV) following a 0.3-
langmuir BrCN exposure with the remaining 40% adsorbed as
atomic Br and CN groups (284.5 eV). The 60% adsorbed
molecularly is much greater than the 25% found for low ICN
exposures to Si(100) at 100 K1 and can be attributed to the
stronger Br-CN bond energy.15
As the exposure is further increased to 0.9 langmuir, Figure
1 also shows a simultaneous intensity increase of the C 1s and
N 1s photoelectron peaks with increasing BrCN exposures. Both
the C 1s and N 1s peaks become more symmetrical and their
peak positions shift slightly upward in energy, approaching a
constant with the highest exposures. No shift with increasing
BrCN exposures is observed for the Si 2p photoelectron peak.
With increasing exposures of molecular BrCN to the Si(100)
surface at 100 K, an increasingly thick molecular BrCN film
with a C 1s binding energy of 286.5 eV is formed. However, a
simple adsorption model in which the BrCN first adsorbs as a
dissociated species followed by molecular adsorption can be
ruled out. The concentration of the dissociated species continu-
ally increases until an exposure of approximately 1.2 langmuirs
is reached.
ClCN is synthesized by using standard inorganic preparative
techniques.11 Briefly, ClCN is formed by bubbling Cl2 gas
through a tetrachloromethane solution of sodium cyanide with
a trace amount of acetic acid at temperatures between -10 and
-5 °C. After the solution turns green, the flowing Cl2 gas is
replaced by He, and the solution is slowly heated to 60 °C.
The resulting gas-phase reaction products are collected in a dry
ice/acetone cooled flask. Excess Cl2 gas is removed from the
final reaction product by repeated freeze-pump-thaw cycles.
The trimer, 1,3,5-triclorotriazine, is found to be a major
contaminant. This contaminant is eliminated by cooling the flask
down to -5 °C in a water/salt bath. The pure gas-phase ClCN
is then collected in a dry ice/acetone cooled receiving flask.
Results and Discussions
XPS Low-Temperature Adsorption Results. The C 1s and
N 1s XPS photoelectron spectra as a function of BrCN exposure
to Si(100) surfaces at 100 K are shown in Figure 1. A 0.3-
langmuir molecular BrCN exposure results in an XPS spectrum
with the C 1s and N 1s photoelectron peaks centered at
approximately 286 and 399 eV, respectively. These binding
energies agree favorably with previous XPS studies of intact
CN species adsorbed on various surfaces. In CNx films prepared
by low-power inductively coupled plasma-activated transport
reactions from a solid carbon source, XPS studies by Popov et
al.12 attribute a C 1s photoelectron peak at 285.9 eV and N 1s
peak at 398.4 eV to the presence of nitrile group (-CtN) in
the carbon nitride film. In another study, Hammer et al.13 also
assign an N 1s photoelectron peak at 398.4 eV to the presence
Figure 2 shows the C 1s and N 1s XPS spectra as a function
of ClCN exposure to the Si(100) surface at 100 K. With
increasing ClCN exposures to Si(100), a simultaneous increase
in intensity for both the C 1s and N 1s photoelectron peaks is
observed. For the C 1s peak, a second peak clearly grows in at
289.2 eV with increasing exposure at 100 K. As in the case of

7728 J. Phys. Chem. B, Vol. 107, No. 31, 2003
Rajasekar et al.
Figure 3. Si(100) XPS spectra for the (a) bromine 3d and (b) chlorine
2p peak as a function of the respective cyanogen halide exposure at
100 K. The XPS intensities are relative.
Figure 2. Nitrogen and carbon 1s XPS spectra as a function of ClCN
exposure to the Si(100) surface at 100 K. The XPS intensities are
relative.
TABLE 1: The XPS C 1s Peak Position after a Saturation
Exposure of ICN, BrCN, and ClCN on Si(100) Surfaces at
Low (100 K) and Room Temperatures (300 K)
low sticking probability of molecular ClCN with respect to ICN
and BrCN results in the disproportional buildup of the dissoci-
ated species. However, the concentration of the dissociated
species continues to increase as the molecular species starts to
build in. Thus, ClCN is similar to BrCN in that ClCN does not
saturate the dissociated state before molecular adsorption occurs.
The Br 3d and Cl 2p XPS spectra provide further support
for the BrCN and ClCN adsorption models. For BrCN exposures
of 0.3 langmuir to Si(100) at 100 K, the Br 3d photoelectron
peak (Figure 3a) is centered at 70 eV with a tail toward higher
binding energies. In addition, for exposures sufficient to form
a multilayer, the Br 3d peak is found at higher binding energies
(72 eV). After a submonolayer exposure of ICN on the Si(100)
surface at 100 K, a comparable high energy tail is observed for
the I 4d photoelectron peak.1 Consistent with the ICN C 1s
photoelectron studies, the tail toward higher energy is due to a
mixture of molecularly adsorbed ICN species and adsorbed
atomic I and CN on the Si(100) surface.1 Therefore, the high-
energy tail observed in the Br 3d spectrum for submonolayer
BrCN exposures to Si(100) at 100 K supports an analogous
adsorption model with both molecular BrCN and dissociated
BrCN present on the surface. This conclusion is also consistent
with the BrCN C 1s photoelectron results discussed above. The
Cl 2p XPS spectrum shown in Figure 3b contains an analogous
trend in peak positions and widths. For a 0.6-langmuir exposure
of ClCN at 100 K (Figure 3b), the Cl 2p photoelectron peak is
observed at 200 eV with a tail toward higher binding energy.
For large exposures (33 langmuirs), multilayers of ClCN are
formed and the Cl 2p peak is shifted upward to 203 eV. Along
with the ClCN C 1s information, these results also support the
adsorption model found for ICN and BrCN, with ClCN adsorbed
both molecularly and as atomic Cl and CN at low exposures
and temperatures on Si(100).
The low-temperature adsorption of BrCN and ClCN is simpler
than that found for HCN on Si(100). When high-resolution
electron energy loss spectroscopy is performed after low HCN
exposures at 100 K, the resulting spectra are complex due to
contributions from HCN, HCNH, and CN species along with
HCN dimers.5 At larger exposures and lower temperatures, HCN
dimers and possibly polymers become the dominate species on
the Si(100) surface.5 In contrast, the low-temperature cyanogen
halide chemistry is similar, with only molecular and one
dissociated species adsorbed on the Si(100) surface.
The driving force for the formation of this dissociated species
is the replacement of the XC bond with stronger XSi and CSi
bonds.16 The resulting CSi and CN bonds are stronger than an
XC or an XN bond. The XC bond in XCN is stronger than the
typical XC bond due to overlap between the p orbital on the
low temperature
room temperature
ICN
BrCN
ClCN
285.5 eV1
286.5 Ev
289.2 eV
284.5 eV1
284.5 eV
284.5 eV
BrCN and ICN,1 these results indicate that a molecular film
forms with increasing ClCN exposures. The C 1s binding energy
for ClCN in the film is 289.2 eV, higher than that for both BrCN
and ICN.1 Table 1 summarizes the measured binding energy
for the C 1s photoelectron peak of molecularly adsorbed ICN,
BrCN, and ClCN. As expected, the increasing electronegativity
of the halogens from I to Cl results in an upward shift in the
binding energy of the C 1s the photoelectron peak.
Compared to the analogous experiments for ICN1 and BrCN,
the uptake rate is significantly smaller for ClCN. Visual
examination of the 3-langmuir exposures for BrCN (Figure 1)
and ClCN (Figure 2) clearly shows the dramatic difference in
the uptake. For BrCN, the multilayer film is clearly formed,
while for ClCN, the multilayer peak is clearly much smaller in
relationship to the peak due to the dissociated species. Unlike
ICN and BrCN, ClCN is prepared as a gas at a pressure greater
than 760 Torr. In addition, the boiling point of ClCN is 13.8
°C, while for ICN and BrCN they are 146 and 61 °C,
respectively.10,11 Thus, possible undetected contributions to the
measured pressure during ClCN dosing, but not ICN or BrCN,
are unlikely. In addition, differences in the ionization cross
sections between these analogous compounds are unlikely to
cause such a dramatic difference in the uptake. The most
probable explanation is that the sticking coefficient for ClCN
is sufficiently smaller than that for ICN and BrCN.
As in the case of BrCN, the C 1s photoelectron peak for a
0.6-langmuir molecular ClCN exposure to Si(100) is slightly
broader than that measured for higher exposures. This low
exposure peak possesses a tail toward higher binding energy
similar to that found for BrCN and ICN.1 Using an analysis
similar to that utilized for BrCN, nearly 40% of ClCN is found
to be adsorbed molecularly (289.2 eV), while the remaining
60% dissociates upon adsorption to form atomic Cl and CN
groups (284.5 eV) after a 0.6-langmuir exposure of ClCN to
Si(100) at 100 K. As in the case of BrCN, this peak position
for the dissociated species is justified by the room-temperature
results discussed below. Although the Cl-CN bond strength is
stronger than the Br-CN bond,15 approximately 20% more
ClCN adsorbs dissociatively than BrCN after low exposures to
Si(100). This contradicts the bond-energy trend inferred from
the BrCN and ICN results. One likely explanation is that the

Adsorption Chemistry of BrCN and ClCN
J. Phys. Chem. B, Vol. 107, No. 31, 2003 7729
halide and the π-system on the CN group.15 However, for
possible adsorption geometries where this overlap is not
possible, the XSi bond is slightly stronger than the XC bond.16
Thus, a side-on adsorption geometry for XCN where the CN
group adds across the Si dimer bond is not energetically
favorable compared to the observed dissociated species. For
ICN, computational studies support this conclusion.2 In contrast
for HCN, the formation of a side-on adsorbed HCNH species
is possible due to the formation of stronger HC and HN bonds
with respect to the HSi bond. Intermediates to this species can
react with other adsorbed HCN molecules leading to dimeriza-
tion or polymerization on the Si(100) surface. Consistent with
these ideas, the C2N2 molecule adsorbs molecularly on the
Si(100) surface at low temperatures due to the stronger CC bond
and the lack of surface H.6 The exclusive molecular adsorption
of the aromatic (HCN)3 species can also be rationalized by the
strong intramolecular bonds.8
Figure 4. He(I) UPS spectra of Si(100) after various exposures to (a)
BrCN and (b) ClCN at 100 K. The UPS intensities are relative.
system contains a feature at 5.0 eV, which is assigned to the
presence of atomic Br on the surface. Similar UPS transition
energies are reported for Br on other surfaces. For Br adsorbed
on Cu(111), Jones and Kadodwala24 attribute an UPS peak at
4.7 eV to a state of majority Br character. On the Ag(110)
surfaces, Kruger and Benndorf 25 identify a peak at 5.2 eV due
to chemisorbed Br.
UPS Low-Temperature Adsorption. The ground state of
the iodide, bromide, and chloride cyanogen halides in the gas
phase has been determined to be a 1Σ+ state with a (1σ)2(2σ)2-
(3σ)2(1π)4(4σ)2(2π)4 electron configuration for the valence
electrons.17-23 An important aspect of the chemical binding
between the halogen and the CN group is the overlap between
the atomic p orbital centered on the halogen and the π-system
of the CN group. This overlap splits the originally degenerate
π-system into two energetically distinct molecular orbitals, an
in-phase (1π) and an out-of-phase (2π) combinations. For a
molecular ICN film on Si(100), the out-of-phase combination
results in one broad peak at 4.4 ( 0.4 eV with respect to the
fermi energy.1 The in-phase combination is energetically close
The UPS spectra as a function of ClCN exposure to the
Si(100) surface at low temperature are shown in Figure 4b. An
interpretation similar to ICN and BrCN can be applied to the
high-exposure ClCN UPS spectra. For a 33-langmuir ClCN
exposure, the spectrum of the resulting multilayer film is shifted
and broadened with respect to lower exposures. The 3-langmuir
ClCN exposure spectrum is better resolved. This spectrum
contains four peaks at 6.5, 8, 9.5, and 13 eV. Using the UPS
spectra of gaseous ClCN as reference,19,22 the peak at 6.5 eV
can be assigned to the out-of-phase (2π) combination, the 8-eV
transition to the nitrogen lone pair (4σ), and the 9.5 eV transition
to the in-phase (1π) combination. The 13 eV transition is
assigned to the 3σ MO of the adsorbed molecular ClCN.
The UPS spectra following low ClCN exposures to the
Si(100) surface at low temperature are more complex than those
of BrCN and ICN. For a 0.3-langmuir ClCN exposure, the
spectrum (Figure 4b) contains a single peak around 5 eV due
to the presence of chlorine from ClCN dissociation. At 0.3
langmuir, no discernible concentration of CN functional groups
is detected by XPS. Due to the higher sensitivity of the Cl 2p
peak, a very small amount of atomic Cl is detected on the
surface. In addition, UPS is found to be very sensitive to atomic
Cl. These results indicate that no molecular ClCN is present
and that Cl is formed by the dissociation of a trace amount of
ClCN. This conclusion is also consistent with the low ClCN
sticking probability with respect to BrCN and ICN found by
the XPS experiment. After 0.6-langmuir ClCN exposures, the
UPS spectrum contains two additional peaks at 6.5 and 9.5 eV.
These transitions are assigned to the out-of-phase (2π) and in-
phase (1π) combinations, respectively. However, XPS results
also indicate the presence of CN groups on the surface, which
have a transition at 6.0 eV. Thus, the 6.5-eV peak also contains
a decreasing contribution due to the surface CN groups with
increasing exposures. This contribution explains the varying
intensity of the 6.5-eV transition with respect to the 9.5-eV peak.
The 4σ molecular orbital is not clearly seen for exposures less
than 3 langmuirs. However, at lower exposures, it contributes
to the almost featureless intensity between 6.5 and 9.5 eV.
XPS Room-Temperature Adsorption. Following an expo-
sure of either BrCN or ClCN to the Si(100) surfaces at room
temperature, the resulting C 1s and N 1s photoelectron peaks
are nearly Gaussian in shape and are centered at approximately
2
to the Σ+ photoelectron peak from the nitrogen lone pair (the
4σ molecular orbital).18 Surface broadening merges these
transitions into one broad peak at 6.7 ( 0.4 eV with respect to
the fermi energy.1 Finally, the 3σ orbital due to the lone pair
on the halogen gives rise to a photoelectron peak at 9.7 ( 0.4
eV.1
The UPS spectra as a function of BrCN and ClCN exposure
to the Si(100) surface at low temperature are shown in Figure
4, parts a and b, respectively. The small shift between the highest
exposures and the 3.0-langmuir exposure is a result of the
changing work-function due to the molecular film. At high
BrCN exposures, three peaks are visible at 5.3, 7.7 and 11.5
eV. Using ICN as a guide, the peak at 5.3 eV originates from
a transition involving the out-of-phase (2π) combinations of
molecular orbitals. The 7.7-eV transition is slightly broader and
is assigned to a mixture of peaks from the in-phase (1π)
combination and to the nitrogen lone pair (4σ). Finally, the 11.5-
eV transition is assigned to the 3σ MO of the adsorbed molecular
BrCN.
For BrCN exposures between 0.3 and 0.6 langmuir, the BrCN
UPS spectra look similar to that at high exposures. Although
features from molecular BrCN after a 0.3-langmuir exposure
are visible, there is both a shoulder toward lower energy and
considerable intensity between the molecular BrCN peaks (5.3
and 7.7 eV). The XPS studies have shown that at low exposure
and temperature a fraction of the BrCN is adsorbed as atomic
Br and CN. For CN generated from room-temperature adsorp-
tion of ICN, a single photoelectron peak at 6.0 eV is visible.1
Thus, CN generated from BrCN adsorption should appear
between the two lowest energy molecular BrCN peaks (5.3 and
7.7 eV). The additional broadening to lower binding energies
(5.3 eV) is due to a contribution of atomic Br at 5.0 eV. To
confirm this assignment, the Si(100) surface is exposed to
molecular Br2 at room temperature. The UPS spectrum of this

7730 J. Phys. Chem. B, Vol. 107, No. 31, 2003
Rajasekar et al.
Figure 6. He(I) UPS spectrum for exposure of (a) BrCN and (b) ClCN
to the Si(100) surface at 300 K. The UPS intensities are relative.
Figure 5. Carbon 1s XPS spectra as a function of (a) 0.3-42-langmuir
BrCN and (b) 0.3-33-langmuir ClCN exposures to the Si(100) surface
at 300 K. The XPS intensities are relative.
room-temperature adsorption contain only one peak due to the
CN π-system centered at 6 eV. The XPS results show the surface
rapidly saturates at relatively low exposures. Unlike XPS which
regulated the power to the anode, the gas flow into the UPS
source is not regulated and the photon flux can drift with time.
Thus, the small increase in intensity of the larger exposures is
attributed to a small drift in the photon flux. Since only one
species exists on the surface, possible photochemistry reaction
leading to an intensity increase is unlikely. A single peak is
also observed after room-temperature adsorption of ICN on
Si(100) and is assigned to the degenerate π molecular orbitals
on the CN groups.1 The UPS shows only a single peak from
the now degenerate π molecular orbitals. After the dissociation
of the I-CN bond, the π molecular orbitals are no longer split
by the overlap between the atomic p orbital centered on the
halogen with the CN π system. For BrCN and ClCN, the room-
temperature UPS results are consistent with the XPS data and
support the room-temperature adsorption model found for ICN.
In this model, the cyanogen halides (XCN) dissociate to form
atomic halogen (X) and intact CN species on the surface.
However, in contrast to atomic I, adsorbed Br and Cl have
discernible features in the UPS spectra. The UPS photoelectron
peak at 5.0 eV resulting from atomic Br originating from the
dissociation of BrCN on the surface imparts a shoulder toward
the lower binding energy on the measured 6 eV peak due to
surface CN groups. The UPS spectrum obtained after a
50-langmuir Br2 exposure to the Si(100) surface at 300 K
contains a broad peak at around 5.5 eV. An examination of the
UPS spectra taken after an exposure of ClCN to the Si(100)
surface at room temperature reveals that the peak at 6.0 eV is
also broader than expected and contains a visible shoulder
toward high binding energies. In addition, a small feature is
observed at approximately 3.0 eV. These additional fractures
are due to the contribution from atomic Cl originating from
ClCN dissociation. The UPS spectrum of a Si(100) surface after
Cl chemisorbed at room temperature has been studied by a
number of authors.27,28 The resulting spectra are complex and
contains many peaks between 3 and 9 eV. Although not as well
resolved, similar spectra after a 32-langmuir Cl2 exposure to
the Si(100) surface at 300 K are obtained by our apparatus.
Thus, the shoulder at 6.0 eV and the broad feature at ap-
proximately 3 eV are due to adsorbed Cl.
284.5 and 398 eV, respectively. Parts a and b of Figure 5 show
the C 1s photoelectron spectrum for BrCN and ClCN, respec-
tively, as a function of exposure to Si(100) at room temperature.
The C 1s peak position is consistent with that found for ICN
after room-temperature exposure to Si(100).1 At room temper-
atures, ICN dissociates into atomic I and CN.1 This C 1s peak
position is also consistent with other XPS studies of intact CN
species adsorbed on various surfaces.12-14 Based on these
observations, both BrCN and ClCN dissociatively adsorb on
Si(100) at room temperature to form an adsorbed halide and an
intact CN functional group. The identification of the 284.5-eV
C 1s photoelectron peak in the room-temperature spectra justifies
the assignment of the lower binding energy component in low-
temperature XPS analysis C 1s of BrCN and ClCN C 1s
photoelectron peak to a dissociated species. After approximately
a 0.2-langmuir BrCN exposure or a 0.6-langmuir ClCN expo-
sure, Figure 5 shows that the C 1s photoelectron peak position
and intensity are unchanging with further exposures of the
respective cyanogen halide. Therefore, room-temperature ex-
posure of either BrCN or ClCN to Si(100) results in a saturated
layer containing the atomic halide species and intact CN
functional groups. For this layer, the ratio of the number of
carbon atoms to nitrogen is found to be approximately 1 for
BrCN and ClCN. This ratio is computed, taking into account
the atomic sensitivity factors,26 from the integrating area ratio
of the C 1s to the N 1s photoelectron peak. Thus, both BrCN
and ClCN dissociatively adsorb on Si(100) at room temperature
to form an adsorbed halide and a CN functional group.
The Br 3d and Cl 2p photoelectron peaks for BrCN and
ClCN, respectively, support the above room-temperature ad-
sorption model. The XPS spectra of the BrCN exposed Si(100)
surface at room temperature show a Br 3d photoelectron peak
at 69 eV. Consistent with the above adsorption model and in
contrast with the low-temperature adsorption results for BrCN,
the room-temperature Br 3d peak lacks the high-energy tail
previously attributed to molecularly adsorbed BrCN. A similar
absence of a high-energy tail is also found for the Cl 2p
photoelectron peak at 200 eV for room-temperature ClCN
exposed Si(100). Again, this result is consistent with the above
adsorption model and in contrast with the low-temperature
adsorption results for ClCN. In conclusion, at room temperature
the surface chemistry of BrCN and ClCN is analogous to that
found for ICN.1 BrCN is adsorbed as atomic Br and CN groups
and ClCN is adsorbed as atomic Cl and CN groups.
XPS Temperature-Dependent Studies. Figures 7 and 8
present the XPS spectra of the C 1s and N 1s photoelectron
peaks for surfaces prepared by room-temperature saturation
exposures of BrCN and ClCN, respectively, and subsequently
annealed for 1 min at indicated temperatures. When the saturated
surface is annealed to 473 K, the C 1s and N 1s photoelectron
peaks start to broaden and shift slightly toward lower binding
UPS Room-Temperature Adsorption. The UPS spectra for
the room-temperature adsorption of BrCN and ClCN as a
function of exposure are given in Figure 6. In contrast to the
low-temperature adsorption (Figure 4), the UPS spectra after

Adsorption Chemistry of BrCN and ClCN
J. Phys. Chem. B, Vol. 107, No. 31, 2003 7731
Figure 7. Nitrogen and carbon 1s XPS spectra from a Si(100) surface
initially exposed to 40 langmuirs of molecular BrCN at 300 K and
annealed to various temperatures for 1 min. The XPS intensities are
relative.
Figure 9. He(I) UPS spectrum from a Si(100) surface initially exposed
to (a) 1 langmuir of BrCN and (b) 2 langmuir of ClCN at 100 K and
annealed to various temperatures for 1 min. The surface is cooled to
100 K before each measurement. The UPS intensities are relative.
vanish as a fraction of the BrCN desorbs and the remaining
dissociates into atomic Br and intact CN groups. At ap-
proximately 323 K, only a single peak appears at 6.0 eV with
a shoulder at the lower binding energy, signifying the complete
dissociation of BrCN into atomic Br and intact CN groups on
the surface. This shoulder is similar to that observed for room-
temperature adsorption and is due to atomic Br. When the
annealing temperature is increased to 873 K, the peak at 6.0
eV disappears leaving only the contribution due to atomic Br
at 5.0 eV. The XPS experiments show that the CN group is
completely dissociated into atomic C and N at 873 K, while
the Br 3d peak intensity drops by about 90%. Thus, the UPS
photoelectron peak at 5.0 eV indicates the presence of atomic
Br on the Si(100) surface after annealing to 873 K. For further
increases in annealing temperature, this peak also vanishes,
indicating the complete removal of surface Br.
Figure 8. Nitrogen and carbon 1s XPS spectra from a Si(100) surface
initially exposed to 33 langmuirs of molecular ClCN at 300 K and
annealed to various temperatures for 1 min. The XPS intensities are
relative.
Figure 9b shows the UPS spectra measured after a 2-langmuir
exposure of ClCN to Si(100) at 100 K and subsequent annealing
at the indicated temperatures. After a 2-langmuir ClCN expo-
sure, the surface contains a mixture of molecular ClCN, atomic
Cl, and CN functional groups. As this surface is annealed to
increasing temperatures, the peaks due to the molecular ClCN
species slowly vanish as some fraction of the molecularly
adsorbed ClCN dissociates into adsorbed Cl and CN species
and the remaining fraction desorbs from the surface. At
approximately 323 K, a distinct peak at 6.0 eV is visible. This
peak is similar to the peak found for a room-temperature
exposure of ClCN and indicates the presence of CN groups on
the surface. Similar to the room-temperature adsorption, this
peak is broadened due to the contribution of adsorbed atomic
Cl to this peak. The UPS spectrum of chemisorbed Cl contains
many peaks between 3 and 9 eV.27,28 Thus the broad feature at
bonding energies higher than 6 eV is also attributed to adsorbed
Cl. This feature also becomes more pronounced with the
decreasing CN surface concentration as inferred by the XPS
temperature-dependent measurements. Finally at 873 K, no Cl
or CN is present on the surface and the UPS spectrum looks
similar to that of clean Si(100).
energies. This shift in binding energy is attributed to the
increasing presence of atomic C and N surface species formed
from partial decomposition of the surface CN groups. The initial
decomposition of the CN groups at 473 K is similar to that
found for ICN1 and other comparable systems.4 As the annealing
temperature is further increased, the C 1s and N 1s peaks
continue to broaden and shift toward lower energies as the
concentration of adsorbed atomic C and N species increases
with the simultaneous decrease in the concentration of adsorbed
CN. Finally at around 873 K, the C 1s and N 1s peaks indicate
only the atomic species with binding energies of 283 and 398
eV, respectively, remain on the Si(100) surface. At this
temperature, all surface CN species are decomposed into Si-C
and Si-N species. A similar decomposition temperature is found
for ICN.1
The decomposition temperature of the adsorbed CN species
on the Si(100) surface is similar to that found for other C and
N containing compounds on the Si(100) surface.3 The CN
surface bonds created by HCN adsorption on Si(100) decompose
between 600 and 800 K.5 The C2N2 and (HCN)3 molecules also
decompose in the same temperature range.6,8
UPS Temperature-Dependent Studies. Figure 9a shows the
UPS spectra measured after a 1-langmuir exposure of BrCN to
Si(100) at 100 K and subsequent annealing at the indicated
temperatures. As discussed above, the two peaks at 5.3 and 7.7
eV for the 1-langmuir BrCN exposed Si surface indicate the
presence of molecular BrCN on the surface. At this exposure,
XPS studies indicate the surface consists primarily of molecular
BrCN with a very small concentration of surface Br and CN.
With increasing annealing temperature, these two peaks slowly
Conclusions
The adsorption and decomposition of BrCN and ClCN on
Si(100) are investigated utilizing XPS and UPS. For submono-
layer exposures, XPS indicates that the CN triple bond remains
intact upon adsorption at 100 K. The π-electrons in the CN
triple bond contribute to two peaks (5.3 and 7.7 eV for BrCN
and 6.5 and 9.5 eV for ClCN) in the UPS spectrum. The splitting
of the π-system shows that some fraction adsorbs molecularly

7732 J. Phys. Chem. B, Vol. 107, No. 31, 2003
Rajasekar et al.
at low temperatures, while the rest adsorbs in a dissociated state.
XPS analyses of the C 1s photoelectron peak following
submonolayer exposure at low temperatures suggest that mo-
lecular adsorption for BrCN is higher (60%) than that of ClCN
(40%). The fraction of molecular BrCN is larger than the 25%
found for ICN adsorption on Si(100),1 possibly due to the
stronger Br-CN bond. However, this possibility does not
explain the large degree of dissociation found for low-temper-
ature ClCN adsorption on Si(100) with respect to BrCN. We
propose, based on the low-temperature XPS uptake experiments,
that molecular ClCN has a lower sticking probability with
respect to ICN and BrCN and that this lower sticking probability
results in the disproportional buildup of the dissociated species.
The dissociated species forms due to the relatively weak X-CN
bond. Consistent with this idea, the C2N2 and aromatic (HCN)3
molecules adsorb without dissociation on the Si(100) surface
at low temperatures.6,8 For low-temperature HCN adsorption
on Si(100), HCN, HCNH, and CN species could be identified
along with HCN dimers.5 In contrast, the Si(100) surface after
low-temperature adsorption of BrCN and ClCN contains only
the molecular and one dissociated species with an intact CN
group on Si(100).
Acknowledgment. We acknowledge the support of this work
by Oklahoma State University Center for Energy Research and
by the Research Corporation. Lucas Ward acknowledges sum-
mer support through OSU’s Department of Physics and its NSF
funded REU program.
References and Notes
(1) Rajasekar, P.; Kadossov, E. B.; Watt, T.; Materer, N. F. Surf. Sci.
2002, 515, 421.
(2) Kadossov, E. B.; Rajasekar, P.; Materer, N. F. Chem. Phys. Lett.
2003, 370, 548.
(3) Lu, X.; Lin, L. M. Int. ReV. Phys. Chem. 2002, 21, 137.
(4) Bacalzo-Gladden, F.; Musaev, D. G.; Lin, M. C. J. Chin. Chem.
Soc. (Taipei) 1999, 46, 395.
(5) Bu, Y.; Ma, L.; Lin, M. C. J. Phys. Chem. 1993, 97, 7081.
(6) Bu, Y.; Ma, L.; Lin, M. C. J. Phys. Chem. 1995, 99, 1046.
(7) Bu, Y.; Ma, L.; Lin, M. C. J. Phys. Chem. 1993, 97, 11797.
(8) Bu, Y.; Lin, M. C. J. Phys. Chem. 1994, 98, 7871.
(9) Materer, N. F.; Watt, T. J. Vac. Sci. Technol. A 2002, 20, 572.
(10) Bak, B.; Hillebert, A. Cyanogen Iodide. In Organic Synthese;
Rabjohn, N., Ed.; John Wiley & Sons: New York, 1963.
(11) Cyanogen Chloride. In Handbook of PreparatiVe Inorganic Chem-
istry, 2nd ed.; Brauer, G., Ed.; Academic Press: 1963; Vol. 1; p 662.
(12) Popov, C.; Plass, M. F.; Bergmaier, A.; Kulisch, W. Appl. Phys.
1999, A69, 241.
The C 1s and N 1s photoelectron peaks found after BrCN
and ClCN exposure at room temperature on Si(100) are similar
to those found for ICN and support the existence of intact CN
function groups.1 As in the case of ICN, no contributions are
found from the molecularly adsorbed species. The UPS spectra
of BrCN and ClCN are more complicated than that of ICN due
to the contribution of adsorbed atomic Br and Cl on the surface.
However, for both BrCN and ClCN, the CN species bonded to
the Si(100) surface contributes only a single peak (6.0 eV) due
to the π-electrons in CN bond to the UPS spectra. This result
indicates that the Br-CN and Cl-CN bonds break, while the
CN bond remains intact during room-temperature adsorption
on Si(100). Upon annealing the CN saturated Si(100) surface
to higher temperatures, the UPS and XPS spectra indicate that
the CN triple bond starts to disappear at approximately 470 K.
Simultaneous changes in the C 1s and N 1s photoelectron peaks
between 470 and 800 K are consistent with the idea that CN
bond cleavage is correlated with silicon carbide and nitride
formation. By 800 K, all adsorbed CN species decompose
completely into silicon carbide and nitride. These annealing
results are similar with the results obtained for the CN-saturated
Si(100) surface prepared by using ICN1 and other analogous
CN-containing systems.5-7
(13) Hammer, P.; Lacerda, R. G.; Droppa, J. R.; Alvarez, F. Diamond
Relat. Mater. 2000, 9, 577.
(14) Bondino, F.; Baraldi, A.; Comelli, G.; Netzer, F. P. Surf. Sci. 2000,
461, 31.
(15) Davis, D. D.; Okabe, H. J. Chem. Phys. 1968, 49, 5526.
(16) Dean, J. A. Lange’s Handbook of Chemistry, 14 ed.; McGraw-
Hill: New York, 1992.
(17) Eland, J. H. D.; Baltzer, P.; Lundqvist; Wannberg, B.; Karlsson,
L. Chem. Phys. 1996, 212, 457.
(18) Eland, J. H. D.; Baltzer, P.; Karlsson, L.; Wannberg, B. Chem.
Phys. 1997, 222, 229.
(19) Heilbronner, E.; Hornung, V.; Muszkat, K. A. HelV. Chim. Acta
1970, 53, 347.
(20) Hollas, J. M.; Sutherley, T. A. Mol. Phys. 1971, 22, 213.
(21) Kovac, B. J. Phys. Chem. 1987, 91, 4231.
(22) Lake, R. F.; Thompson, H. Proc. R. Soc. London Ser. A 1970, 317,
187.
(23) Yee, D. S. C.; Brion, C. E. J. Electron Spectrosc. Relat. Phenom.
1976, 8, 313.
(24) Jones, R. G.; Kadodwala, M. Surf. Sci. Lett. 1997, 370, L219.
(25) Kruger, B.; Benndorf, C. Surf. Sci. 1986, 177, 515.
(26) PHI Handbook of X-ray Photoelectron Spectroscopy; 1977.
(27) Johansson, L. S. O.; Uhrberg, R. I. G.; Lindsay, R.; Wincott, P.
L.; Thornton, G. Phys. ReV. B 1990, 42, 9534.
(28) Martin-Gago, J. A.; Roman, E.; Refolio, M. C.; Lopez-Sancho, J.
M.; Rubio, J.; Hellner, L.; Comelli, G. Surf. Sci. 1999, 424, 82.