Photoinduced fragmentation of multilayer CH3Br on Cu/Ru(001)

Article abstract of DOI:10.1021/jp027857y

The broadband UV (230-420 nm) photoinduced chemistry of CH₃Br adsorbed on Cu(2ML)/Ru(001) in the coverage range of 1-50 ML was studied by monitoring the desorption products (Δp-TPD mode) in combination with post irradiation work function change

Full text of DOI:10.1021/jp027857y

11382
J. Phys. Chem. B 2003, 107, 11382-11390
Photoinduced Fragmentation of Multilayer CH3Br on Cu/Ru(001)
,†
‡
Tsachi Livneh* and Micha Asscher
Department of Physical Chemistry, Nuclear Research Center, NegeV, P.O. Box 9001 Beer-SheVa 84190, Israel,
and Department of Physical Chemistry and the Farkas Center for Light Induced Processes,
The Hebrew UniVersity, Jerusalem 91904, Israel
ReceiVed: December 30, 2002; In Final Form: July 14, 2003
3
The broadband UV (230-420 nm) photoinduced chemistry of CH Br adsorbed on Cu(2ML)/Ru(001) in the
coverage range of 1-50 ML was studied by monitoring the desorption products (∆p-TPD mode) in combination
with post irradiation work function change measurements before and during surface heating (∆æ-ΤPD mode).
∆
æ measurements enabled us to follow multilayer restructure and desorption of parent molecules and
photochemical reaction products in the temperature range of 80-700 K. Methyl radicals accumulated on the
surface are the precursor for the thermal formation of methane and ethylene at 450 K. Dehydrogenation of
the methyl group is the rate-limiting step of the surface reaction resulting in the formation of these molecules.
Based on work function change measurements, an estimate of the adsorbed methyl dipole moment is µ
0
)
0
.48 D. Dissociative electron attachment (DEA) driven CH Br dissociation produced CH and CH fragments
3
3
2
within the parent molecules multilayer matrix. At the multilayer coverage range, ƾ increases by up to 1.1
eV after 10 min UV irradiation. Model calculations qualitatively describe the post irradiation work function
-
changes induced by the embedded photofragments (mostly Br ions) inside the CH
3
Br dielectric film.
Comparison of the ƾ-TPD spectra on Cu(2ML)/Ru(001) to those on clean Ru(001) indicates that the nature
of the molecule-surface interaction and the structure of the first few layers strongly influence the resulting
photochemistry at layer thickness up to at least 20 ML.
1
. Introduction
of the ejected photoelectrons. The dependence of the photo-
fragmentation yield on the number of deposited methyl halide
layers was previously established for CH3Cl/Pt(111) by moni-
toring the dependence of the photolysis rates and photoelectron
yields on the multilayer thickness, where mean free path was
obtained for low energy (>1 eV) photoelectrons.
The photoinduced dissociation of methyl halides on well-
7
1
defined surfaces has been extensively studied. The motivation
of these studies has been mostly to reveal the nature of the
molecule-surface interactions under the influence of UV ir-
radiation. The cross sections of the various processes were found
to depend on the excitation wavelength,2 adsorbate-surface
TOF (time-of-flight) measurements of photoproducts in the
,3
2
system CH3Cl/Ni(111) at excitation wavelengths of 193 and
4
distance and the dielectric properties of a spacer between them,
2
48 nm were conducted. The CH3Cl dissociation threshold
and predeposition of adsorbates5
energy and the high-energy peak of ejected methyl radicals were
attributed to the direct dissociation channel, which is observed
only at 193 nm. Although direct channel was observed at all
coverages above 1-2 ML, the multilayer thickness necessary
to totally quench the DEA process were ∼10 and ∼20 ML for
excitation wavelengths of 248 and 193 nm, respectively. One
may conclude that low energy photoelectrons generated from
the “purely DEA” 248 nm photons could not reach the top layers
Adsorbate photodissociation (at the 1-2 monolayers (ML)
coverage) on metal surfaces is most likely to occur via
1
dissociative electron attachment (DEA) mechanism: CH3X(ad)
-
-
+
hν f CH3X (ad) f CH3 (gas/ad) +X (ad) (X ) Cl, Br, I), in
which a photoexcited substrate electron transfers to a surface
stabilized anionic state of adsorbed CH3X that promptly
dissociates to give methyl radical and halide anion.
Photoexcited substrate electrons are generated with a broad
energy distribution. A fraction of these primary or secondary
electrons can attach to adsorbates molecular (affinity) levels that
may reside under the vacuum level because of image dipole
attraction with the metal substrate. As directly measured by
Sanche and co-workers, electrons at kinetic energy of about 0.5
at thickness that exceeds ∼10 ML, which nicely correlates with
the measured penetration depth for low energy photoelectrons.7
A question arises how dominant is the “cage effect” in
quenching the photodissociation within the molecular multilayers
for the direct dissociation and the DEA channels. Previous
studies have indicated that in the case of the direct dissociation
channel, e.g., in the case of CD3I multilayers on TiO2(110)
surface, methyl fragments that were detected by TOF were
originated from molecules laying very close to the adlayer-
vacuum interface. It was therefore concluded that the cage effect
6
eV are sufficient to attach and cause DEA of methyl halides.
Upon the formation of molecular multilayers, the direct
dissociation route (CH3X(ad) + hν f CH3(gas/ad) +X(gas/ad))
becomes increasingly dominant,2 whereas the DEA channel
is quenched, because of limited penetration into the multilayers
,3
8
is significant and may be dominant within the adlayer. Similar
effects were observed in the case of photodissociation of CH3I
*
To whom correspondence should be addressed. E-mail: tsaliv@
9
within rare gas matrixes. However, this is not necessarily the
netvision.net.il.
2
†
‡
case in the DEA channel. According to Marsh et al., even when
Nuclear Research Center.
The Hebrew University.
electrons cannot reach the topmost layers, DEA may occur in
1
0.1021/jp027857y CCC: $25.00 © 2003 American Chemical Society
Published on Web 09/19/2003

Multilayer CH3Br on Cu/Ru(001)
J. Phys. Chem. B, Vol. 107, No. 41, 2003 11383
the lower layers. This was demonstrated by monitoring the total
cross section for parent molecules removal (CH3Cl/Ni(111)) per
incident photon .
signal transmitted to a computer and converted to Langmuir
units (1 L ) 10 Torr s).
-
6
2
Copper was evaporated onto the Ru(001) sample from a
resistively heated Ta wire wrapped by high purity Cu wire
(99.999%). The Ta filament was covered by a Pyrex shroud
with 5 mm diameter aperture. The Cu source was thoroughly
degassed prior to deposition and was controlled by monitoring
the voltage drop across the Ta wire at constant current. The
The thermal chemistry of methyl and methylene halides on
10-14
copper surfaces was previously studied.
The focus of these
studies has been mainly kinetics and the mechanism of methyl
and methylene dissociation and recombination and the effect
of the copper surface structure. Surface reactions of C1
fragments were studied in various different systems such as
pressure rise during copper evaporation was routinely 1-2 ×
1
0
11,12
10,13
-10
CH3I/Cu(110) and CH3I/Cu(111)
and CH2I2/Cu(110)
10
Torr. The sample was held during evaporation at 640 K
14
and CH2I2/Cu(100) .
in order to avoid CO adsorption and to produce well annealed
surface, thus avoiding three-dimensional clustering of copper
on the Ru(001) surface.
The copper coverage on the Ru(001) surface was kept
constant at 2 ML throughout this work. At this copper coverage,
Unlike the case of CH3I, CH3Br does not thermally dissociate
on Cu(111),1 and only minor dissociation products are detected
5
16
at defects on the Cu(10 ML)/Ru(001) surface. To initiate C1
surface reactions, the C-Br bond had to be cleaved by means
of UV irradiation.1
5-17
the reactivity of the surface toward CH Br thermal decomposi-
Br coverage was shown to play a
3
tion has diminished relative to the clean ruthenium. The
significant role in the photochemical scheme by reducing the
overall DEA cross section.17
remaining minor reactivity is related to defects in the copper
layers. The copper coverage was determined by (∆p-TPD)20 at
a heating rate of 9 K/s. The first layer desorption peak (with a
maximum at 1235 K, at the completion of a monolayer) has a
shape characteristic of zero order desorption kinetics, often taken
as an indication for two-dimensional (2D) island formation
within the adsorbed layer. Upon the full occupation of the
second copper layer the maximum desorption peak is at 1150
K.
Despite the extensive research of the photochemistry of
methyl halides, very few studies explored the processes inside
the multilayer. The chemistry of photogenerated fragments
within the multilayer and irradiation induced multilayer rear-
rangements prior to and during surface annealing are practically
unknown. The focus of this study is to examine the effect of
broadband UV irradiation (230-420 nm) on the photo and
thermal processes within CH3Br multilayers deposited on Cu-
The irradiation source was UV enhanced 450 W Xe lamp
manufactured by Oriel. IR irradiation was eliminated by means
of a water filter. The light was focused at the surface by a series
of quartz lenses covering the entire sample homogeneously. An
Oriel 7-54 filter transmits the spectral range between 230 and
(2ML)/Ru(001) relative to the monolayer. An electrostatic model
we present explains well the experimental data on the basis of
photogenerated charges that are distributed within the molecular
dielectric matrix, with the photoelectrons penetration depth as
a parameter.
4
20 nm. This range avoids direct photodissociation of the
adsorbed CH3Br molecules. The photon flux of the lamp during
2
. Experimental Section
UV illumination at the central wavelength (350 nm) was ∼3 ×
1
6
2
1
0
photons/cm s.
The experiments described here were performed in an
ultrahigh vacuum (UHV) chamber with a base pressure of 3 ×
-
10
3. Results
1
0
Torr obtained by turbomolecular pump (240 L/s). A
+
sputter gun (Ar ions at 600V and sample current of 8mµA)
was employed to clean the Ru(001) surface and a quadrupole
mass spectrometer (QMS-VG MASSTORR DX) was used to
obtain ∆p-TPD spectra. The QMS was surrounded by Pyrex
shroud with 5 mm diameter aperture to minimize detection of
desorbing species from surfaces other than the sample. A Kelvin
probe (Besocke type S) was employed to monitor work function
change (ƾ) as a function of crystal temperature using the same
computer controlled routine. The heating rate was 2 K/s
throughout this study.
18
The thermal chemistry of CH3Br on Ru(001) and over Cu-
19
(
2ML)/Ru(001) in the submonolayer and multilayer range was
discussed elsewhere in detail. Briefly, the saturation coverage
of the CH3Br relative to the number of ruthenium atoms was
18
determined to be 0.22 ( 0.02 for both surfaces with the
14
monolayer coverage (1 ML) consisting of 3.6 ( 0.3 × 10
14
2
and 4.0 ( 0.3 × 10 molecules/cm , respectively. The density
14
of the crystalline CH3Br in the (001) plane is 6.96 × 10
2
21
molecules/cm , as obtained from X-ray studies. Thus, it was
concluded that upon completion of the first layer the density
reaches 52 ( 5% and 58 ( 5%, respectively, of it’s value in
the bulk. We note that our interpretation of a monolayer is
The Ru(001) sample (a square piece, 8 × 8 mm, 1.5 mm
thick) was cut from a single-crystal rod to within (1° of the
2-4,7
(001) crystallographic orientation and then polished by standard
different from other studies
for which the definition of 1ML
is correlated with a definition of a bilayer (2ML) in the present
procedure. Sample cleaning procedure in UHV was described
1
8,19
18,19
elsewhere.
LEED from the clean and annealed surface
study.
showed very sharp hexagonal pattern. The sample was attached
to a liquid nitrogen reservoir via copper feedthroughs directly
welded to the bottom of the dewar. AC resistive heating of two
3
.1. Photochemistry of 1 ML CH3Br on the Cu(2ML)/
Ru(001) Surface. 3.1.1. ∆p-TPD. ∆p-TPD spectra are shown
in Figure 1 of 1.1 ML CH3Br (m/e ) 94) and CH4 (m/e ) 16)
from Cu(2ML)/Ru(001) following exposure to UV light (full
spectral range, 230-420 nm) from a Xe lamp of up to 96 min
of irradiation time. Without irradiation, only minor CH4
desorption is observed, associated with parent molecules dis-
sociation at defects on the copper layer that is subsequently
followed by bimolecular reaction with adsorbed hydrogen to
form methane. Cleavage of the C-Br bond is the result of the
UV irradiation, producing “hot” CH3 radicals. These radicals
are either ejected to the gas phase or accumulated on the surface
0.5 mm diameter tantalum wires, between which the sample is
spot welded, was employed to control the sample temperature.
W5%Re - W26%Re thermocouple wires spot welded to the
edge of the ruthenium sample were used for sample temperature
determination and control.
CH3Br (99.5% pure) was further purified by a few freeze-
pump-thaw cycles to eliminate any noncondensable residual
gases. Exposure was done by filling the chamber through a leak
valve to the desired pressure with the uncorrected ion gauge

11384 J. Phys. Chem. B, Vol. 107, No. 41, 2003
Livneh and Asscher
3 d 3
Figure 2. Extent of CH Br dissociation (N ) and the resulting CH
coverage determined from ∆p-TPD for different irradiation times. In
3 3
the inset: -ln([CH Br](t)/[CH Br](0)) vs UV illumination time is
displayed.
Figure 1. ∆p-TPD spectra of 1.1 ML CH
3
Br (m/e ) 94) and CH
4
(
m/e ) 16) on Cu(2 ML)/Ru(001) following 0, 3, 6, 12, 24, and 96
Quantitative analysis indicates that after 96 min of irradiation
65% of the methyl radicals were ejected to the gas phase.
The photodissociation rate of methyl halides is expected to
follow first-order kinetics in methyl halides concentration;
therefore, it can be expressed
min exposure to a UV Xe lamp at its full spectral range. The spectra
for 0 and 96 min are denoted with dashed and dotted lines, respectively.
In the inset: The C
2 4
H (m/e ) 27) ∆p-TPD after 12 min irradiation
(solid line) and prior to irradiation (dashed line).
and disproportionate around 450 K to CH4 and C2H4.10,11 Finally,
one cannot rule out the possibility of further photochemical
dissociation of the adsorbed CH3 radicals: Their first absorption
band in the gas phase is around 216 nm22 and the threshold of
the adsorbed radicals that dissociate through the DEA mecha-
nism is expected at a higher wavelengths, probably within our
excitation range (λ > 230 nm). However, we expect this channel
to be minor because our photon flux at these wavelengths is
-
ln([CH Br](t)/[CH Br](0)) ) kt
(1)
3
3
with
k )
f(λ)σ(λ)
(2)
∑
λ
2
f(λ) is the wavelength-dependent photon flux (photons/cm s),
σ(λ) is the wavelength-dependent cross section for DEA (cm /
photon), and t is the time of illumination. If the dissociation
cross-section dependence on the surface coverage is negligible,
2
1
5
2
very low (<10 photons/cm s).
As shown in the inset of Figure 1, the C2H4 signal clearly
overlaps that of CH4. A ratio of 3:1 was found between the
integrated areas under the CH4 TPD peak and that of C2H4.
This ratio is higher than the expected 2:1 ratio based on the
-
ln([CH3Br](t)/[CH3Br](0)) is expected to vary linearly with
the time of irradiation. In the inset of Figure 2, we note that
initially, at short irradiation times, the photodissociation rate
detailed mechanism for the methyl radicals disproportionation
-
4
2
was ∼4 × 10 ML/cm s. Beyond 10 min deviation from
linearity is observed, suggesting a decrease in the dissociation
cross-section with time.
that was suggested by Bent and co-workers.1
0,11
Such a
1
1,16
discrepancy that was also found in other studies
may
originate from complete methyl group dehydrogenation followed
by carbon dissolution into the bulk.11 It is interesting to note
that the methane/ethylene ratio decreases in the multilayer
coverage range toward the 2:1 ratio (see below), which may
indicate that the channel of complete methyl dehydrogenation
is becoming less favorable in the presence of accumulated
bromine.
The origin of this saturation like behavior may be due to the
following: (i) The increase in the surface work function due to
the disappearance of methyl bromide and to the deposition of
bromine atoms, see the discussion below. This increase may
have shifted the affinity level of methyl bromide relative to the
energy spectrum of the photoelectrons to a less favorable
position in terms of electron attachment probability. (ii) Br (as
Although the integrated desorption peak of the parent
molecule continuously decreases, the CH4 and consequently
C2H4 peaks increase only for irradiation times below 20 min
and then are kept at constant surface coverage following longer
irradiation times. The extent of CH3Br dissociation (Nd) and
the resulting CH3 coverage are shown for various irradiation
times, see Figure 2. After 24 min, the density of the photoge-
nerated methyl radicals (extracted from the sum of the (cali-
brated) integrated TPD areas of CH4 and C2H4) has saturated
around CH3/Cu(2ML)/Ru(001) ) 0.04 ML. In light of a related
study, where coadsorbed iodine atoms were found to decrease
significantly (by a factor of 3) the saturation coverage of methyl
radicals on Cu (111),23 we conclude that the saturation density
of the methyl radicals is influenced by the coverage of
coadsorbed Br atoms on the Cu(2ML)/Ru(001) surface. Above
a maximum coverage of the photochemicaly produced Br atoms,
the methyl radicals are apparently ejected to the gas phase.
an acceptor) and CH3Br (as a donor) have opposite dipoles on
18,19,24
the surface.
The consistent shift to higher temperatures
of the CH3Br ∆p-TPD (Figure 1) and ∆æ-TPD (see discussion
below) as exposure to UV light increases suggests that attractive
interactions between the coadsorbates (Br atoms and CH3Br
molecules) dominate. Consequently, stabilization of the ground
electronic state may lead to more efficient quenching of the
exited state and therefore the dissociation cross-section should
further decrease.
3.1.2. ƾ-TPD. In Figure 3a, post irradiation ƾ-TPD spectra
are shown of 1.1 ML CH Br/Cu(2ML)/Ru(001) following the
3
indicated exposure times to the UV light. In the absence of
irradiation, the work function is practically constant above 210
K, indicating no significant CH Br dissociation. Upon increasing
3
the time of exposure, a peak around 450 K builds up. This peak
coincides with the detected desorption of CH4 and C2H4, as
revealed in Figure 1. In Figure 3b, the derivative of the work

Multilayer CH3Br on Cu/Ru(001)
J. Phys. Chem. B, Vol. 107, No. 41, 2003 11385
The work function of the system increases by 0.12 eV upon
removal of the methyl species from Cu(2ML)/Ru(001). Based
on the integrated area under the corresponding methane ∆p-
TPD peak, the photochemically generated methyl coverage
reached CH3/Cu ) 0.04 ML (0.18 of an initial 1 ML methyl
bromide). Simple calculation based on the Helmholtz equa-
tion: ∆æ ) 4πeΝµ0, where N is the surface methyl density, µ0
is the isolated methyl dipole moment, e is the unit charge,
provides an estimate of the isolated adsorbed methyl dipole
-
30
moment of µ0 ) 0.48 D (1 D ) 3.34 × 10
C m). The
justification to use this simplified model stems from the low
coverage and insignificant dipole-dipole interaction between
the methyl adsorbates.
As in the ∆p-TPD case, a saturation value is reached also in
the work function increase around 450 K after 12 min irradiation
time. This observation suggests that at longer exposures to the
UV light additional methyl radicals that are generated by the
CH3Br photodissociation, are ejected to the gas phase during
the irradiation. However, further dissociation of CH3Br follow-
ing exposure to the UV light beyond 12 min (resulting in
additional Br deposition on the surface) is evident from the
increase in ƾ at the completion of the ƾ-TPD run around
Figure 3. Post irradiation ƾ-TPD (a) d(ƾ)/dT (b) spectra of 1.1
3
ML CH Br/Cu(2ML)/Ru(001) after exposure times of 0, 3, 6, 12, 24,
4
8, and 96 min (a-g, respectively) to the Xe UV full lamp. In the
inset: A comparison of the ∆p-TPD (solid line) and the d(∆æ)/dT
dotted line) spectra around 450 K, indicating the very good overlap
(
between the two.
function with respect to temperature d(ƾ)/dT is presented. It
is evident that, as bromine atoms accumulate on the surface, a
shift to higher temperatures is observed in the derivative spectra.
This reflects the stabilization of the methyl bromide molecules
due to the presence of the bromine atoms, as a result of attraction
600 K. This result is consistent also with the monotonic decrease
in the ∆p-TPD signal of CH3Br, as shown in Figure 1 at longer
exposure times.
2
4
3.2. Photochemistry of Multilayers of CH Br on the Cu-
between opposite dipoles (bromine) +0.8 eV/0.33 ML and
3
18
(2ML)/Ru(001) Surface. 3.2.1. ∆p-TPD. ∆p-TPD spectra of
CH4 (m/e ) 16) and C2H4 (m/e ) 27) after 10 min UV
irradiation of a Cu(2ML)/Ru(001) surface pre-deposited with
CH3Br ) -1.4 eV/0.22 ML ).
A comparison of the ∆p-TPD and the d(∆æ)/dT spectra
around 450 K is shown in the inset of Figure 3a, indicating a
very good overlap between the two. This is not surprising,
considering that the desorption is rate limited by the methyl
dehydrogenation CH3(a) f CH2(a) + H(a).10 The reactions of
methylene insertion and methyl hydrogenation were found to
1
-50 ML of CH3Br are shown in Figure 4, parts a and b,
respectively. The irradiation source is the same as in section
3.1; therefore, the direct dissociation channel of CH Br can be
3
ignored while the surface temperature during irradiation was
kept at 82 K
10
be considerably faster than the methyl dissociation on Cu(110).
Both the removal of CH3 that is indirectly measured by work
function change and CH4 and C2H4 desorption are rate limited
by the same step. Furthermore, the overlap between the two
measurements suggest that, if methylene fragments simulta-
neously decompose in that temperature range as well, it must
take place without affecting considerably the measured work
function change.
Around 1 ML coverage, ethylene and methane desorption
signals practically overlap, having a single peak around 450 K,
as discussed in section 3.1. As the coverage increases in the
3-22 ML range, the spectra change monotonically in the
following manner:
1. A new desorption peak gradually emerges at 410 K on
account of the 450 K desorption peak. Between 10 and 22 ML,
Figure 4. ∆p-TPD spectra of CH
2K, predeposited with 1-50 ML of CH
nm “window” was transmitted. In the inset: A typical ∆p-TPD spectrum of H
4
(m/e ) 16) (a) and C
Br. The beam produced by the broadband irradiation source (Xe lamp) was filtered so that a 230-420
(m/e ) 2)
2 4
H (m/e ) 27) (b) after 10 min UV irradiation time of a Cu(2 ML)/Ru(001) surface at
8
3
2

11386 J. Phys. Chem. B, Vol. 107, No. 41, 2003
Livneh and Asscher
Figure 5. Dependence of the extent of CH
thickness, as calculated by integrating the area under the ∆p-TPD CH
and C peaks.
3
Br dissociation on the layer
4
2
H
4
the peak around 450 K diminishes, whereas the peak at 410 K
initially grows and then it saturates. The increased population
in the 410 K peak on account of the 450 K peak is more
pronounced for ethylene. 2. New peaks build up simultaneously
around 307 and 330 K for CH4 and around 230 and 280 K for
C2H4.
Figure 6. ƾ-TPD spectra after 10 min UV irradiation time of a Cu-
(
2 ML)/Ru(001) surface predeposited with 1-50 ML of CH
3
Br. For
comparison, the ƾ-TPD of 4 ML CH
3
Br that have not been exposed
In the inset of Figure 4, a typical ∆p-TPD spectrum of H2
to UV light is shown by dotted line. In the inset: A summary of the
increase in the work function resulted from photodissocoation of
(
m/e ) 2) is presented as well. Based on former studies which
showed that hydrogen recombinative desorption from Cu(110)
is expected around 340 K,10 the absence of hydrogen desorption
at this temperature range indicates that the hydrogen-methyl
radical recombination to form methane is faster than hydrogen
recombination and desorption, partially because of the high
density of adsorbed methyl radicals.
3
gradually thicker CH Br layers, following 10 min irradiation.
of its producs from the Cu(2ML)/Ru(001) surface is monitored
at T > 250 K. The increase in ƾ due to methane and ethylene
production gradually shifts from 450 K down to 280 K upon
increasing CH3Br layer thickness up to ∼25 ML. Above this
coverage, no further change in the ƾ profiles is observed,
consistent with the ∆p-TPD results of Figure 4.
Information on surface processes that occur at coverages and
temperatures below the onset for molecular desorption can be
obtained from ƾ-TPD spectra. In Figure 7a,b, we present the
The dependence of the extent of CH3Br dissociation on the
number of deposited layers, calculated by integrating the area
under the CH4 and C2H4 peaks in the ∆p-TPD spectra, is shown
11
in Figure 5. Based on the expected surface reaction pathways,
the CH4 + 2*C2H4 signal is presented as well. The extent of
dissociation steeply increases up to 10 ML, and at a reduced
rate, it grows up to 22 ML coverage. Above 22 ML, no
significant change in the ∆p-TPD spectrum is observed.
∆
p-TPD (Figure 7a) and the d(ƾ)/dT spectra (Figure 7b) of
CH3Br in the 80-150 K range. The temperature ranges for the
desorption of the different layers of CH3Br are indicated.
Comparison of the desorption onset of condensed CH3Br with
the d(ƾ)/dT spectra illustrates that a significant change in the
work function is recorded during sample heating before any
desorption takes place. This process is observed only after the
system has been irradiated. No ƾ change for the nonirradiated
system could be detected prior to the multilayer desorption
temperature (dashed line in Figure 7b). The peaks obtained at
temperatures preceding desorption may arise from diffusion of
DEA products to the surface and multilayer restructure.
However, these processes need further investigation before we
fully understand them.
Differential spectra d(ƾ)/dT at temperatures where the
dissociation products react and desorb as new product molecules
are shown in Figure 7c. We note the shift of the desorption
peaks to lower temperatures, similar to the corresponding ∆p-
TPD spectra shown in Figure 4. The overlap between d(ƾ)/
dT and ∆p-TPD of methane is demonstrated for a layer of 22-
25 ML in Figure 7c. The overlap between the two for the
monolayer coverage range was attributed to the methyl dis-
sociation (CH3 f CH2 + H) being the rate-determining step,
followed by a rapid process of recombinative desorption of
methane due to reaction of adsorbed hydrogen atoms with
3
.2.2. ƾ-TPD. Because the overlap between the excitation
spectrum (Xe lamp source) and the CH3Br absorption spectrum
was reduced to a minimmum (using a proper optical filter) the
main photodissociaction channel left is a DEA surface mediated
-
processes, upon which Br ions are produced within the CH3-
Br matrix.2
,6,25
These ions generate electric field around them
and are attracted to their image charge inside the bulk metal
underneath. To caracterize the dissociation products embedded
inside the multilayer structure after irradiation and to probe the
-
multilayer restructure during the diffusion of the Br ions, a
set of ƾ-TPD spectra were recorded under similar condition
as for the ∆p-TPD. These are shown in Figure 6. For
comparison, the ƾ-TPD of 4 ML CH3Br that have not been
exposed to UV light is shown by dotted line.
The increase in the work function that resulted from photo-
dissocoation of gradually thicker layers of CH3Br following 10
min irradiation at 82 K is shown in the inset of Figure 6. An
increase of the work function of up to 1.1 eV was measured
(
note that ƾ of the 50 ML CH3Br was -0.8 eV prior to
irradiation because ƾ upon adsorption at 82 K gradually
increases after the completion of the second layer ). In addition,
the signature of methyl radical dissociation and the desorption
19

Multilayer CH3Br on Cu/Ru(001)
J. Phys. Chem. B, Vol. 107, No. 41, 2003 11387
Figure 7. ∆p-TPD (a) and the d(∆æ)/dT spectra (b) of CH
3
Br in the 80-150 K range. The temperature range for the desorption of the different
CH Br layers is indicated. (c) d(∆æ)/dT spectra at higher temperature range of 250-550 K, where methyl dehydrogenates, see text. ∆p-TPD of
3
3
methane for irradiated 22 ML and the (d(ƾ)/dT) of 4 ML CH Br that have not been exposed to UV light are also shown for comparison.
adjacent methyl.10 The presence of photo generated Br atoms
as coadsorbates on the surface during this reaction suggests
that despite a shift in the desorption temperature by 40 K, it
did not change the overall kinetic scheme of the methyl
dehydrogenation as a rate-limiting step. This is consistent with
studies by Bent et al.,23 where iodine poisoning was reported
as a minor perturbation to the overall scheme of the methyl
radical dissociation on Cu(111).
In Figure 8a, the ƾ-TPD spectra of CH3Br(9ML)/Ru(001)
at 82 K before and after irradiation are compared. Following
10 min irradiation, an increase of 1.5 eV in the work function
is observed. However, upon annealing the surface to tempera-
tures still below the parent molecule desorption onset at 110
K, the surface work function is practically recovered to the
value in the absence of irradiation. It is evident that photo-
dissociated fragments and multilayer restructure prior to de-
sorption are responsible for a significant work function decrease
of ∼1 eV. In Figure 8b, (d∆æ/dT) spectra are compared for
Ru(001) and Cu(2ML)/Ru(001). It is clear that ƾ prior to
multilayer desorption is more pronounced on Ru(001) than
on the Cu(2ML)/Ru(001) surface. Furthermore, the work func-
tion change flips its sign from positive on the nonirradiated
Ru(001) to negative change upon exposure to the UV light,
as shown for 9 ML CH3Br in Figure 8b. This can be attributed
When multilayers undergo DEA, the density of photoproducts
on the surface is expected to be significantly higher, as a result
of accumulated DEA products from at least 20 ML. Neverthe-
less, the overall behavior at the 250-550 K temperature range
seems very much the same as the monolayer coverage case.
This except for the typical second-order kinetics shift to lower
temperatures as coverage increases.
3
.3. Multilayer CH3Br on Ru(001) vs Cu(2ML)/Ru(001).
-
to the presence of Br union within the multilayer and the
The thermal chemistry and structure of multilayer methyl
bromide on Ru(001)18 and Cu(2ML)/Ru(001) were previously
reported. A question arises as to what is the influence of the
under laying metallic substrate on the photochemistry within
the multilayer: How the dissociation cross section inside the
matrix may be sensitive to it and to what extent the matrix
restructure can be affected by the substrate. Destabilization of
the third layer and then a more stable fourth and thicker layers
on the clean Ru(001) were discussed in terms of bulklike
molecular crystalline structure formation.18 It turns out that on
the copper covered surface destabilization of the third layer is
not observed. Based on work function change measurements
taken during adsorption, we concluded that, unlike the clean
Ru(001) surface, on the Cu(2ML)/Ru(001) surface, CH3Br
molecules do not form the molecular crystalline-like structure
based on antiparallel arrangement. This may be related to the
high density of defects on the Cu(2ML)/Ru(001) surface. Such
defects will often lead to molecular adsorption in a wide
distribution of tilt angles with respect to the surface normal,
thus avoiding the necessary packing that results in the molecular
crystal structure.
graduate process of charge-transfer back to the substrate upon
annealing.
19
Three processes are expected to occur in the matrix:
-
(
i) Multilayers restructure. (ii) Attraction of the Br ions by
their image charge in the bulk metal that results in their
migration toward the surface. (iii) Dissociation of caged negative
-
ion complexes, (CH3Br)n , n g 1 that are created as a solvation-
6
like process that stabilizes the ionic species. Ethane (mass 30)
and ethylene (mass 27, the choice of this mass was to avoid
the large background at mass 28) TPD peaks are observed near
9
5 K. These molecules are absent from the TPD spectra in the
case of the nonirradiated system. The peak near 95 K, shown
in Figure 8c, practically coincides with the minimum of the
dƾ/dT spectrum, shown in Figure 8b. The methane peak (not
shown) was also observed. These ethylene and ethane peaks
probably originate from methyl recombination and methylene
insertion of some of the CH and CH dissociation products
3
2
that are free to diffuse, react, and desorb upon annealing the
multilayer. The correlation between the dissociated fragment
diffusion/reaction/desorption and a significant restructure in the
multilayer reflects a “phase transition-like” behavior that allows

11388 J. Phys. Chem. B, Vol. 107, No. 41, 2003
Livneh and Asscher
Figure 8. (a) ƾ-TPD spectra of CH
before and after irradiation and CH Br(9 ML)/Cu(2 ML)/Ru(001) following irradiation. (c) ∆p-TPD spectra of C
line), C (m/e ) 30, dashed line), and CH Br (m/e ) 94, solid line) after 10 min UV irradiation of a Ru(001) surface, predeposited by 9 ML of
CH Br.
3
Br(9 ML)/Ru(001) at 82 K before and after irradiation. (b) (dƾ/dT) spectra of CH
3
Br(9 ML)/Ru(001)
(m/e ) 27, dashed-dotted
3
2 4
H
H
2 6
3
3
a vast reorganization in the multilayer. However, the fact that
the dƾ/dT spectrum of the nonirradiated multilayer is peaked
The extent of dissociation measured by the integrated signal of
CH4 and C2H4 should reflect the production yield of Br atoms,
based on negligible thermal dissociation. (ii) The work function
change at 82 K that is due to photofragments contribution
(mostly Br ions) that are captured within the CH3Br matrix
following UV irradiation.
(having an opposite sign) in the same temperature range before
the desorption onset of the parent molecules (see Figure 8b)
indicates that this behavior is a unique property of the CH3-
Br-Ru(001) system. This behavior should be associated with
the different packing of the first 4 layers (see refs 18 and 19
for elaborate discussions). Photochemically produced fragments
seem to enhance this phenomenon.
Another interesting difference between the two substrates that
is revealed by photochemistry is the different ƾ-TPD spectra
obtained at temperatures above the CH3Br desorption, T > 210
K. As demonstrated in Figure 8a, the ƾ-TPD profiles of the
irradiated and nonirradiated multilayers are identical above 150
K in the case of clean Ru(001). This indicates that the
photogenerated fragments do not modify the distribution of
species that are formed and stabilized on the surface upon
annealing at temperatures above 210 K. This is not the case for
the Cu(2ML)/Ru(001) surface shown in Figures 4 and 6. It may
be correlated with the lower reactivity of both adsorbed CH3Br
and CH3 radical on this copper character-like surface.
-
Because bromide ions are expected to significantly affect the
work function, we shall treat the ionic influence using a modified
electrostatic model that was originally developed by Tsekouras
2
6
et al. for ions on hydrocarbon films.
A certain concentration of bromide ions are placed on top of
a condensed CH3Br layer deposited on a metal surface. This
layer is expected to act as an insulating linear dielectric medium.
Ions placed on top of it should create a voltage because of their
charge and the capacitance of the film. This voltage creates a
change in the work function æ of the metal+film assembly by
an amount of
-
Aꢀꢀ
eQL
ƾ )
) -eVfilm
0
where Q is the charge deposited, A is the area, L is the film
thickness, ꢀ is the dielectric constant of the film, ꢀ0 is the vacuum
permittivity, and e is the unit charge. If ꢀ is constant throughout
the film, diffusion of ions should lead to a change in the lateral
distribution of the ions inside the film F(x,y,z). As a thermally
activated process, diffusion is accelerated by annealing of the
molecular matrix. At 82 K, we assume that the ions are frozen
at the coordinates where DEA took place. The time dependent
distribution of vertical distances of ions from the surface inside
the film can be followed by monitoring the change in the
work function. Generally, the x and y dependence should be
rather small for a macroscopically homogeneous film, which
is sensitive only to the density average over x and y, to give
F(z). Integration by parts of the Poisson equation over the free
charges in the presence of constant dielectric medium yields a
simple expression for the change in the work function that
4
. Discussion
.1 Multilayer Film Irradiated at 82 K. In the following,
4
we present a simplified electrostatic model that attempts to
address the experimental observations of work function increase
of 1.1 eV upon UV irradiation of multilayer methyl bromide.
For our system of CH3Br on Cu(2ML)/Ru(001) at 82 K, three
assumptions can be made based on earlier studies:2
,3,6,7
(i) No significant direct photodissociation channel takes place.
(
ii) The maximum penetration depth of photoelectrons is ∼20
ML. (iii) For simplicity, we take an average value of the DEA
cross section to represent the fact that we neglect its layer
thickness dependence.
Quantitatively, there are two different aspects of coverage
that were identified in our Cu(2ML)/Ru(001) experiment: (i)

Multilayer CH3Br on Cu/Ru(001)
J. Phys. Chem. B, Vol. 107, No. 41, 2003 11389
number of monolayers results in a penetration depth of 4 ML.
The coverage dependence of the photo fragmentation was
previously established for the CH3Cl/Pt(111) system by measur-
ing the photolysis rates and low energy (<1 eV) photoelectron
yields as a function of thin layer thickness, and a penetration
7
27
depth of 1.8 ML was reported, which is similar to ours. We
may consider the quality of the fit in Figure 9 as supporting
evidence for the claim that ions and electrons have long-range
polarization effects on the neutral dielectric films they are
embedded within and therefore justify the employment of the
ion-dielectric formalism in order to describe embedded charges
inside a multilayer.
The cage effect has often been considered to be the main
reason for relatively small measured cross sections for photo-
9
dissociation of molecules embedded in a matrix. The cage
created by the surrounding atoms or molecules introduces a
barrier for dissociation that originates from the need of the
photofragments to escape the cage walls. However, if a low
energy photoelectron is ejected from the surface to be attached
to a parent molecule within the multilayer, an externally assisted
-
6
anionic bound state, (CH3Br)n , n g 1, may be formed. Such
a “solvation” process should stabilize the anion via polarization
of the surrounding medium. A recent study has reported a DEA
Figure 9. Normalized work function change ∆æ(θ)/∆æ(θmax), fol-
lowing 10 min irradiation time at 82 K as a function of methyl bromide
layer thickness in ML. The dashed line is the calculated curve, see
text.
-16
2
cross section of ∼1 × 10 cm for low energy electrons near
and below 1 eV, for CH3Cl molecules embedded inside a Kr
solid matrix. This value is seven times higher than that of CH3-
results from the voltage across the film:26
6
Cl when deposited on top of a Kr solid surface. The difference
e
L
∆
æ(L) ) -e(V(L) - V(0)) ) -
∫
zF(z) dz (3)
between the surface and bulk photoinduced processes was
attributed to unique product stabilization in the bulk that
overcomes the decrease in the cross section due to the cage
effect. Similar cross sections were reported for CH3Br molecules
condensed on top of a Xe film as well as with a maximum
DEA cross section for electrons at kinetic energy of 0.1 eV.
These values are more than two orders of magnitudes higher
than the equivalent DEA cross section in the gas phase.25
One may conclude that there is an interplay between
stabilization effects and escape probability from a cage of the
dissociated fragments that dictates the net probability for the
DEA process of an embedded CH3Br molecule. Voids in the
amorphous solid that reduce the perfect cage structure may also
reduce the cage effect in favor of enhancing DEA. Because our
measurements are based on macroscopically averaging work
function changes measured at 82 K, it is impossible for us to
0
ꢀꢀ0
Therefore, we can use this picture of a frozen distribution of
ions inside a dielectric medium as a valid description of our
system after irradiation at 82 K. It is necessary to define F(z)
more precisely as follows:
F(z) ) F(λ) N (λ)e-R‚zk
Br
(4)
∑
e
λ
F(λ) is the time integrated flux density of photons weighted by
the wavelengths profile of the irradiation source. Ne(λ) is the
wavelength dependent probability of ejecting a low energy
photoelectron. ∑F(λ)Ne(λ) is therefore the time integrated flux
density of photoelectrons that are ejected from the surface. The
dependence of slow photoelectrons penetration on z is of
exponential form where 1/R is the penetration depth, assumed
-
distinguish between the separate contributions of Br ions (DEA
products) and stabilized (CH3Br)n cluster ions.
7
-
to be independent of their energy below 1 eV. In order to
estimate the number of ions produced at a certain infinitesimally
4.2. Fragmentation within the Multilayer. A notable
difference is observed between the photochemical formation
rates of CH and C H at 1 ML coverage of CH Br on Cu-
small dz distance along the dielectric layer, we introduce kBr as
-
the number of ions (mostly Br ) that are produced per
4
2
4
3
photoelectron. We neglect the dependence of kBr on z, which is
an approximation that becomes more accurate as the ions are
removed away from the surface-film interface, where image
charges do not play a significant role.
In Figure 9, we demonstrate the correspondence between the
calculated expression
(2ML)/Ru(001) to the rates observed at a coverage of 22 ML
on the same substrate. At the high coverage, the photoproducts
may further react only after the annealing/desorption of the
multilayer. A significant difference in the peak desorption
temperatures of these two products is observed. The desorption
peak shifts from 450 K at 1 ML to 410 K at 22 ML, following
exposure to the UV irradiation. This shift is attributed to the
higher density of photogenerated CH3 radicals that were formed
within the ∼25 ML thick film and are accumulated on the
surface upon annealing. Higher density of adsorbed methyl
radicals leads to a faster and, therefore, lower temperature
reaction. We believe that the second-order kinetics appearance
of these products, suggest a disproportionation reaction to take
place between two adjacent methyl groups:
0
∫
zF(z) dz
zF(z) dz
∆
æ(Θ)
Θ
)
0
∆
æ(Θ)max
∫
Θmax
to the experimental values of the work function change at 82 K
after 10 min irradiation as they vary with the layer thickness.
∆æ(Θ) is extracted from the data presented in Figure 6,
normalized to Θmax ) 50 ML, where no changes in ∆æ were
further observed. Converting the fitted parameter R to the
2
CH (a) f CH (g) + CH (a)
3 4 2

11390 J. Phys. Chem. B, Vol. 107, No. 41, 2003
Livneh and Asscher
The dominance of this second order reaction can be understood
only at higher coverages. At the lower ones, methyl dehydro-
genation is the rate-limiting step, as discussed before. Second-
order kinetics explains in a trivial way the increasing yield and
the shift of products appearance to lower temperatures.
revealed by the lower temperatures (230 K) ethylene production
and desorption.
DEA driven CH Br dissociation produced CH and CH
2
3
3
fragments inside the multilayers. The maximum equivalent
coverage of the trapped methyl radicals increases up to 0.06
ML at 22 ML initial parent molecules layer thickness.
Very low temperature for ethylene formation near 230 K, is
observed as a result of the multilayer photochemistry. The C2H4
peak centered at 230 K is attributed to CH2 radicals recombina-
tion: 2CH2(a) f C2H4(a) to form ethylene that desorbs
immediately upon formation. Similar results were obtained on
single-crystal Cu(100).14 The CH2 radicals may originate from
further photodissociation or DEA of methyl radicals at 82 K
inside the multilayer structure. Electron attachment to the CH3
radical is known to stabilize it while removing its planarity.
At the multilayer coverage range, ƾ increases up to 1.1 eV
after 10 min irradiation. We present an electrostatic model that
qualitatively explains the post irradiation work function increase
-
induced by the embedded photofragments (mostly Br ions)
inside the CH Br dielectric film. Using the formalism of the
3
above model to fit the experimental data we obtain a penetration
depth for low energy photoelectrons of 4 ML onto the molecular
2
7
layer, in agreement with prevously reported data for CH Cl/
Pt(111).
3
-
7
The existence of the methide (CH3 ) ion was demonstrated and
shown to have electron affinity of 0.08(0.03 eV.28
Comparing the ƾ-TPD spectra obtained for CH3Br on Cu-
(2ML)/Ru(001) to that on clean Ru(001) surfaces indicate that
the nature of the molecule-surface interaction and structure of
the first few layers significantly influences the photochemistry
up to at least 20 ML.
After the C2H4 desorption peaked at 230 K, another C2H4
production channel is observed, its maximum rate is around 280
K, followed by a CH4 desorption at 310 K. The 280 K C2H4
peak, that was detected also on Cu(100),14 is thought to proceed
via methylene insertion mechanism CH2(a) + CH3(a) f C2H4-
1
0
(
a) + H(a). The subsequent 310 K peak is due to CH3
References and Notes
hydrogenation by the hydrogen atoms that are supplied from
the above methylene insertion.10 It is important to note that no
(1) Zhou, X. L.; Zhu, X. Y.; White, J. M. Surf. Sci. Rep. 1991, 13, 73.
(
2) Marsh, E. P.; Tabares, F. L.; Gilton, T. L.; Meier, W.; Schneider,
M. R.; Cowin, J. P. Phys. ReV. Lett. 1988, 61, 2725.
3) . Marsh, E. P.; Tabares, F. L.; Schneider, M. R.; Gilton, T. L.;
H2 is produced concurrent with CH4 desorption, probably
because CH3 coverage is significantly higher than the H atoms
coverage.
(
Meier, W.; Cowin, J. P. J. Chem. Phys. 1990, 92, 2004.
Around 450K desorption peaks due to CH4 and C2H4, are
observed to shift to lower temperatures with increasing multi-
layer coverage, as discussed above. The C2H4/CH4 ratio
increases by a factor of 2 as the CH3Br initial coverage increased
from 1 to 22 ML. The integrated C2H4 desorption signal at low
temperatures (230 K) is higher than that of CH4. It may be that
both rate constants of CH3 hydrogenation (CH3(a) + H(a) f
CH4(g)) and methylene insertion (CH2(a) + CH3(a) f C2H4(a)
(4) Gilton, T. L.; Dehnbostel, C. P.; Cowin, J. P. J. Chem. Phys. 1989,
1, 1937.
9
(5) Solymosi, F.; Kiss, J.; R e´ v e´ s, K. J. Chem. Phys. 1991, 94, 8510.
(6) Ayotte, P.; Gamache, J.; Bass, D.; Fabricant, I.; Sanche, L. J. Chem.
Phys. 1997, 106 (2), 749.
7) Jo, S. K.; Zhu, X. Y.; Lennon, D.; White, J. M. Surf. Sci. 1991,
241, 231.
(8) Holbert, V. P.; Garrett, S. J.; Stair, P. C.; Weitz, E. Surf. Sci. 1996,
46, 189.
9) Bondybey, V. E.; Burns, L. AdV. Chem. Phys. 1980, 41, 269.
10) Chiang, C.-M.; Wentzlaff, T. H.; Bent, B. E. J. Phys. Chem. 1992,
6, 1836.
(
3
9
(
+
H(a)) are affected in a different way from the increase of the
(
Br coverage and the rate of methylene insertion is amplified
on account of methyl hydrogenation.
(
(
11) Lin, J.-L.; Bent, B. E. J. Vac. Sci. Technol. 1992, 10 (4), 2202.
12) Lin, J.-L.; Bent, B. E. J. Phys Chem 1993 , 97, 9713.
5. Conclusions
(13) Chiang, C.-M.; Wentzlaff, T. H.; Jenks, C. J.; Bent, B. E J. Vac.
Sci. Technol. 1992, 10 (4), 2185.
The broadband UV (230-420 nm) photoinduced chemistry
(14) Kovacs, I.; Solymosi, F. J. Phys Chem B 1997, 101, 5397.
(15) Lamont C. L. A., Conrad H., Bradshaw A. M. Surf. Sci. 1993, 280,
9.
of CH3Br adsorbed on Cu(2ML)/Ru(001) in the 1-50 ML
coverage range was studied by monitoring the desorption
products (∆p-TPD mode) in combination with post irradiation
work function change measurements before and during surface
heating (∆æ-ΤPD mode). The later enabled us to follow
changes in ƾ that are due to multilayer restructure and
desorption of the photoproducts in the 80-700 K temperature
range.
7
(
16) Roop, B.; Zhou, Y.; Liu, Z.-M.; Henderson, M. A.; Lloyd, K. G.;
Campion, A.; White, J. M. J. Vac. Sci. Technol. A7 1989 2121.
17) Lamont, C. L. A.; Conrad, H.; Bradshaw, A. M. Surf. Sci. 1993,
(
2
87/288, 169.
(18) Livneh, T.; Asscher, M. J. Phys. Chem. B 1997, 101, 7505.
(19) Livneh, T.; Asscher, M. J. Phys. Chem. B 1999, 103, 5665.
(20) Wolter, H.; Schmidt, M.; Wandelt, K. Surf. Sci. 1993, 298, 173.
(
21) Kawaguchi, T.; Hijikigawa, M.; Hayafuji, Y.; Ikeda, M.; Fukushima,
R.; Tomiie, Y. Bull. Chem. Soc. Jpn. 1973, 46, 53.
22) Herzberg, G. Molecular Spectra and Molecular Structure III:
Electronic spectra and electronic structure of polyatomic molecules; Van
Nostrand Co. Inc.: Princeton, NJ, 1966; p 609.
Methyl bromide undergoes dissociative electron attachment
(
DEA) at the photoexcitation wavelengths used in this study.
At 1 ML CH3Br, the initial coverage maximum density of
(
the photo generated CH3 radicals is equivalent to 0.04ML. Any
additional DEA produced methyl radical is ejected to the gas
phase. The adsorbed CH3 radicals were found to disproportionate
during surface heating around 450K to produce C2H4 + CH4
in a 1:3 ratio. Simple calculation, which is based on work
function change measurements, provides an estimate of the
isolated adsorbed methyl dipole moment of µ0 ) 0.48 D.
With increasing initial CH3Br coverage from 1 to 22 ML the
CH4 and C2H4 peaks around 450 K are shifted to 410 K and
the C2H4/CH4 ratio increases by a factor of 2. Enhanced CH2
production due to photochemistry within the multilayer is
(23) Chiang, C.-M; Bent, B. E Surf. Sci. 1992, 279, 79.
(
24) Bange, K.; Dohl, R.; Grinder, D. E.; Sass, J. K. Vacuum 1983, 33
(
10-12), 757.
25) Nagesha, K.; Fabrikant, I.; Sanche, L. J. Chem. Phys. 2001, 114
11), 4934.
26) Tsekouras, A. A.; Iedema, M. J.; Cowin, J. P. J. Chem. Phys. 1999,
111 (5), 2222.
(
(
(
(27) We note that our interpretation of a monolayer is different from
other studies [refs 2-4 and 7] for which the definition of 1 ML is correlated
18,19
with a definition of a bilayer (2 ML) in the present study.
(28) Ellison, G. B.; Engelking, P. C.; Lineberger, W. C. J. Am. Chem.
Soc. 1978, 100 (8), 2556.