Isothermal and temperature-programmed oxidation of CH over Pt{1 1 0}-(1 × 2)

Article abstract of DOI:10.1016/S0039-6028(01)01679-X

The isothermal oxidation of adsorbed CH_a on Pt{1 1 0} has been studied using a supersonic molecular beam system. CH_a oxidation was found to be much more facile than C_a oxidation and takes place at surface temperatures as low as 330 K. Over the surface temperature range 350-450 K the gas phase products are CO₂ and H₂O. We conclusively show that the first step of the reaction is CH_a + O_a → CO_a + H_a and that H abstraction to produce OH_a and C_a does not occur. CO₂ is produced by the subsequent surface reaction of CO_a with O_a while H₂O is produced by H_a oxidation. An Arrhenius analysis yields values for the pre-exponential factor and activation energy for the CH_a + O_a reaction of (2.1 ± 1.5) × 10^-6 molecule^-1cm²s^-1 and 71 ± 3 kJmol^-1, respectively. We also studied the temperature programmed reaction CH_a + O_a reaction and found that the product selectivity is very sensitive to the atomic O_a pre-coverage.

Full text of DOI:10.1016/S0039-6028(01)01679-X

Surface Science 505 (2002) 58–70
www.elsevier.com/locate/susc
Isothermal and temperature-programmed oxidation
of CH over Pt{110}-(1 ꢀ 2)
D.T.P. Watson, J.J.W. Harris, D.A. King
*
Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK
Received 31 May 2001; accepted for publication 12 October 2001
Abstract
The isothermal oxidation of adsorbed CH
CH oxidation was found to be much more facile than C
30 K.Over the surface temperature range 350–450 K the gas phase products are CO
that the first step of the reaction is CH ! CO þ H and that H abstraction to produce OH
þ O
occur.CO is produced by the subsequent surface reaction of CO with O while H O is produced by H
Arrhenius analysis yields values for the pre-exponential factor and activation energy for the CH
þ O
ð2:1 Æ 1:5Þ ꢀ 10 molecule cm s and 71 Æ 3 kJmol , respectively.We also studied the temperature programmed
reaction CH reaction and found that the product selectivity is very sensitive to the atomic O pre-cover-
a
on Pt{1 1 0} has been studied using a supersonic molecular beam system.
oxidation and takes place at surface temperatures as low as
and H O.We conclusively show
and C does not
oxidation.An
reaction of
a
a
3
2
2
a
a
a
a
a
a
2
a
a
2
a
a
a
À6
À1
2
À1
À1
a
þ O
a
a
age. Ó 2001 Elsevier Science B.V. All rights reserved.
Keywords: Platinum; Oxidation; Surface chemical reaction
1
. Introduction
(HREELS) and temperature programmed de-
sorption (TPD) they found that on oxygen-
covered Rh{1 1 1}, methylene, CH , is partially
Efficient syngas production from methane over
2a
rhodium and platinum coated ceramic monoliths
was observed by Hickman and Schmidt [1].This
work prompted numerous investigations into the
chemistry of alkyls on oxygen-covered Rh{1 1 1}
and many other transition metals.Friend and co-
workers [2–5] used the thermal decomposition of
the alkyl halides CH I and CH I to study the
oxidised to formaldehyde while methyl, CH3a is
oxidised to form CO, CO , H O, and H , via a
transient methoxy intermediate.They also con-
cluded that oxygen does not insert into the H C–
2
2
2
2
Rh bond to yield formaldehyde but reacts with
a transient CH species which is produced during
2
CH I decomposition, before it equilibrates with
2 2
3
2
2
oxidation of CH3a and CH , respectively.Using
2a
the surface.
high resolution electron energy loss spectroscopy
3
Solymosi and co-workers [6] used a CH radical
source to produce the CH3a overlayer on Rh{1 1 1}
and RAIRS to identify the nature of the reac-
tion products.They found that methoxy was
produced whether CH was pre- or post-oxygen
3
dosed.They found that the amount of methoxy
*
Corresponding author.Tel :. +44-1223-336338/336449; fax:
44-1223-762829.
E-mail address: hg243@cam.ac.uk (D.A. King).
+
0
039-6028/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S0039-6028(01)01679-X

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
59
was considerably less when CH
3
was pre-dosed
cepted to be: hydroxy recombination; and hydrogen
addition to a hydroxy group.Both take place at
surface temperatures as low as 220 K.
and that the yield was extremely sensitive to oxy-
gen exposure.They attributed the differences be-
tween their observations and those of Friend and
co-workers to the low resolution of HREELS, and
the lower oxygen coverages used by Friend and co-
1
The stability and reactivity of C oxygenates on
platinum has been studied extensively.Matsumoto
and co-workers [15] found that on Pt{1 1 1} for-
mate, HCO , was formed by reaction of methanol,
workers [2].The partial oxidation of CH
to
2
2
formaldehyde was reported for the thermal reac-
tion of CH on oxygen-covered Pd{1 0 0} [7] and
ClCH I on oxygen-covered Pt{1 1 1} [8].Since the
CH OH, with adsorbed molecular oxygen.Using
3
À1
2
I
2
RAIRS they observed a band at 1327 cm which
they assigned to the C–O stretch.They found that
2
reaction of oxygen and methylene formed by first
decomposing the halide was not studied in both
cases, we cannot say whether the mechanism sug-
gested by Friend and co-workers [2] occurs on
these metals.Quinlan et al.[9] found that on
Ni{1 0 0} formaldehyde could be produced by ex-
posing an oxygen pre-covered surface to methane.
All of these studies indicate that the reaction of
HCO
270 K to produce CO
2
decomposed at temperatures greater than
and H , exclusively.An-
2
2
other temperature programmed reaction (TPR)-
RAIRS study on Pt{1 1 0}-ð1 ꢀ 2Þ performed by
Ohtani et al.[16], found that HCO formed by the
2
thermal decomposition of formic acid HCO H,
2
decomposed at 300 K.Again, CO
the only products.
2
and H
2
were
C
1
hydrocarbon fragments with oxygen proceed
On the other hand formyl, CHO , has proved
a
via on oxygenated hydrocarbon species and that
stripping of the hydrogens to produce water and
atomic carbon is not significant.
much more difficult to isolate and only a few in-
vestigators have tentatively claimed to have ob-
served it as a transient reactive intermediate.Wang
and Masel [17] found evidence for a HCO inter-
mediate at a surface temperature of 220 K when
To date, the oxidation of methylidyne, CH , has
a
not been studied directly.This is primarily due to
the difficulty in producing pure CH adlayers.Zhou
3 a
CH OH is thermally decomposed on Pt{1 1 0}-
et al.[8] found that on Pt{1 1 1} ClCH
decompose to CH2a at 170 K, and converts com-
pletely to CH by 370 K.Therefore, in the surface
temperature range 320–370 K the CO and H
formation are best described as CH oxidation.
2
I starts to
ð1 ꢀ 2Þ.This assignment is based on EELS bands
À1
at 1100, 1600 and 2500 cm , assigned to the HCO
bend, C–O stretch and C–H stretch, respectively.
Using EELS Anton et al.[18] found that on
a
2
2
O
a
a 2 a
Ru{0 0 1}, CHO , which is formed by H CO de-
Using HREELS they found evidence for O and
a
composition, decomposes at surface temperatures
above 120 K.
Here, we present a study of the oxidation of
CHa on the surface at this temperature range
during the reaction, but none for oxygenated in-
termediates.
a
CH using isothermal reaction techniques, TPR
The reaction of hydrogen adatoms with both
atomic and molecularly adsorbed oxygen has been
studied in great detail on Pt{1 1 1}.Fisher et al.
and mass balance methods.
[
10] studied the reaction of hydrogen with chemi-
2. Experimental
sorbed molecular oxygen using temperature pro-
grammed reaction spectroscopy (TPRS) and X-ray
photoelectron spectroscopy (XPS), and observed
water formation at 200 K.The reaction between
adsorbed atomic oxygen and hydrogen on
Pt{1 1 1} has also been studied by many groups
using HREELS and TPRS [11–14].The reaction
was found to be quite complex and extremely
sensitive to the oxygen and hydrogen pre-cover-
ages.The two major pathways are generally ac-
The experimental apparatus is described in de-
tail elsewhere [19].The sample is mounted cen-
trally on a manipulator in an ultrahigh vacuum
À10
chamber with a base pressure of <2 ꢀ 10 mbar.
The molecular beam is sourced at a supersonic
expansion from a 50 lm nozzle, skimmed, differ-
entially pumped and collimated before entering the
sample chamber.An inert King and Wells (KW)
flag just in front of the crystal is used to control

60
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
the dosing time and to measure sticking proba-
bilities [20].A further flag in the source stage
allows the beam to be turned on and off.The
s
surface temperature (T ) is monitored by a crystal-
mounted thermocouple referenced to an electronic
ice-point and is regulated by programmed resistive
heating.The partial pressures of up to 16 indi-
vidual gases can be monitored using a fixed QMS
situated behind the crystal.
The Pt sample, 11 mm diameter by 1mm thick,
was cut to within 1° of the {1 1 0} plane.Initial
cleaning of the crystal was achieved by repeated
cycles of ion sputtering, annealing and oxygen
treatment.Routine cleaning consisted of annealing
at 1240 K, exposure to oxygen for 5 min while
cooling from 1100 to 950 K and annealing for
Fig.1 .I sothermal reaction of 0 1. 1 ML CH
molecular beam with a flux of 0.023 ML s to produce CO
and H O at a surface temperature of 385 K.
a
À1
adlayer with an O
2
2
2
peratures below 450 K [25].In contrast, we now
find that the CH -oxygen reaction is facile even
1
5 min at 950 K.This procedure yields a clean
a
pt{1 1 0}-ð1 ꢀ 2Þ surface which gives a sharp LEED
pattern and oxygen thermal desorption spectra
that are in good agreement with the literature [21].
at surface temperatures as low as 350 K.The
isothermal reaction between an O2 supersonic
molecular beam and a CH adlayer at a surface
a
The CH
9
4
and O
9.995% pure, respectively, as quoted by the sup-
2
used were >99.9995% and
temperature of 385 K can be seen in Fig.1.The
CH adlayer was prepared by dosing the He see-
a
À1
pliers (Messer (UK) Ltd.). The supersonic meth-
ane beams used in these experiments to prepare the
ded CH beam with a flux of 0.42 ML s for 1 min
4
at surface temperature of 370 K.After 20 s the
source stage flag was lowered allowing the O2
beam to impinge against the inert KW flag, pro-
2
a 4
CH adlayers had a composition of 8% CH –92%
He and the nozzle was heated to 790 K.The oxy-
gen beam used was produced using a room
temperature nozzle.All coverages are quoted in
ducing an immediate rise in the O pressure.After
ꢂ50 s the KW flag was raised allowing the O
2
14
monolayers (ML) (1 ML ꢁ 9:22 ꢀ 10
mole-
beam to impinge directly onto the crystal, with an
immediate drop in the O pressure; a fraction of
À2
cules cm ) and were determined using the spin-
ning rotor gauge which measures absolute beam
fluxes and by comparison of TPD areas with ear-
lier studies [22,23].
2
the O molecules in the beam stick and react with
2
CH to produce gas phase CO and H O.We
a
2
2
found no sign of other gas phase products such as
hydrogen, carbon monoxide, methanol, formal-
dehyde and formic acid.After about 330 s the KW
flag was lowered.This is accompanied by an in-
In the previous paper [24], we demonstrated
a
that a pure adlayer of CH can be formed on
Pt{1 1 0} by dosing from a supersonic molecular
beam source.This procedure is used in the present
crease in the O partial pressure (and a tiny pulse
2
work to investigate the reaction of O
sorbed CH on Pt{1 1 0}.
2
with ad-
of CO which is caused by the mechanical motion
2
of the KW flag in the main chamber).
The source stage flag was closed after 335 s and
the crystal was then temperature ramped at a rate
À1
3
. Results and discussion
.1. Isothermal CH oxidation
of 3 Ks to a surface temperature of 750 K.The
CO
to the reaction of residual C
There are two important points to note.(1) We
took great care to convince ourselves that the
sharp peak in the CO signal which occurs when
the source stage flag is opened (at 20 s, Fig.1) is
2
produced during the temperature ramp is due
with adsorbed O .
a
3
a
a
In earlier studies in this group, we found that
the oxidation of C formed by dissociative ad-
sorption of CH does not occur at surface tem-
a
2
4

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
61
due neither to impurities in the O
reaction with C at the crystal, but results from the
2
beam nor to
a
background on the crystal.(2) The profiles shown
are raw data and are not corrected for the relative
sensitivities of the system to different species.
2
However, since the sensitivity to H O is relatively
low due to the lifetime of adsorbed H O on the
2
walls of the chamber, for clarity the H O signal
2
has been multiplied by a factor of thirty.
To gain insight into the mechanism of CH
a
oxidation, the reaction was studied at different
surface temperatures, ranging from 370 to 500 K.
In each case the CH
dosing the He seeded CH
a
adlayer was prepared by
beam with a flux of 0.42
4
À1
ML s , for 1 min, at a surface temperature of 370
K yielding a CH coverage of 0.11 ML. The re-
a
sulting temporal profiles of the CO
produced are presented in Figs.2 and 3.At a
surface temperature of 370 K the CO partial
2 2
and H O
2
Fig.3.H
CH
at different surface temperatures.
2
O produced by isothermal reaction of a 0.11 ML
pressure rises slowly to a maximum after 85 s and
then decreases slowly to the base value.The peak
shape remains the same up to a reaction temper-
ature of 425 K; the only difference is the increased
reaction rate.The situation becomes more com-
plex as the surface temperature is increased further
À1
a
adlayer with an O
2
molecular beam (0.023 MLs
2
O flux)
derlying mechanism of oxidation, and hence the
peak shapes, are completely different.
because of the increasing conversion of CH
as the surface temperature is raised above 400 K.
The thermal conversion of CH to C has been
discussed in detail elsewhere [24].Above 450 K the
a
to C
a
The shape of the H O temporal profiles are
2
quite different from those of the CO profiles, as
2
shown in Fig.3.The H 2O pressure reaches a
maximum in the first few seconds of the reaction,
stays almost constant and drops quickly to zero
when the reaction is near completion.An impor-
a
a
CH
a
is completely converted to C
a
and the un-
tant point to note is that the CH coverage at a
a
particular surface temperature prior to the reac-
tion is directly related to the total area under the
H
2
O reaction profiles.At surface temperatures of
70 and 385 K the areas are identical, indicating
a minimal conversion of CH to C at 385 K.
However, as the temperature is increased above
00 K the area under the H O peaks decrease,
3
a
a
4
2
until at a surface temperature of 475 K there is no
O production.This shows that above 475 K the
adlayer is composed of adsorbed carbon and not
CH , and CO is produced by oxidation of atomic
H
2
a
2
a
C .
2
The reactive O sticking probability calculated
from data such as that shown in Fig.1 for each of
the above experiments is shown in Fig.4.At a
surface temperature of 370 K the initial sticking
Fig.2 .C O
adlayer with an O
2
produced by isothermal reaction of a 0.11 ML CH
À1
a
2
molecular beam (0.023 MLs
O
2
flux) at
different surface temperatures.

62
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
À1
2
O flux) reacting with a 0.11 ML
Fig.4 .T he time-dependent reactive O
2
sticking probability for a O
2
molecular beam (0.023 MLs
CH adlayer at different surface temperatures.
a
probability is 0.1, which is much lower than the
value of 0.35 on the clean surface under the same
conditions [21].The sticking probability starts to
decrease with time due to site blocking by atomic
Oa and the reaction products.As the reaction
pletion at these low temperatures.However, when
the isothermal reaction temperature is raised to
400 K, the area under the subsequent TPR CO
2
peak is significant and the peak maximum shifts to
500 K which shows there is residual surface species
which cannot be removed by isothermal oxidation
at 400 K.Increasing the isothermal reaction tem-
perature further to 425 and 450 K increases the
TPR area even more.However, after isothermal
oxidation at 500 K, there is no carbon left on the
surface.In the previous studies we have shown the
peak maximum for the TPR reaction of carbon
proceeds, the rate of CO production increases,
2
leaving free sites where O can adsorb, and the O
2
2
sticking probability increases.After about 90 s the
oxygen coverage starts to build up and the stick-
ing probability drops to 0.02. This low sticking
probability was also observed by Walker and King
[
21] on the clean surface for oxygen coverages
above 0.25 ML. A similar behaviour is observed at
85, 400 and 425 K.Another important trend is
with atomic oxygen to produce CO occurs at a
2
3
surface temperature of 505 K [25].We therefore
assign this TPR peak to C oxidation.
the increase of the initial sticking probability as the
reaction temperature is raised.This is due to the
a
The amount of C
mal reaction passes through a maximum at 450 K.
Above this temperature the C reaction starts
þ O
to take place at an appreciable rate.This has im-
portant mechanistic implications for the CH
þ O
reaction.The first step cannot involve the ab-
a
remaining after the isother-
partial conversion of CH
efficient site blocker than CH
a
to C
a
, which is a less
a
[26].
a
a
After each of the isothermal oxidation experi-
ments were finished, the crystal temperature was
a
a
À1
ramped at a rate of 3 K s to 750 K while fol-
lowing the CO partial pressure, shown in Fig.5.
straction of hydrogen from CH to produce atomic
a
2
The peak areas and peak temperatures give a
valuable insight into the mechanism of CH
dation.After isothermal reaction at 370 or at 385
K (not shown), the TPR peak is extremely small
which shows the reaction essentially goes to com-
C and OH , since the C produced would be un-
a
a
a
a
oxi-
reactive below 450 K, yet the reaction goes to
completion at surface temperature of 370 K.
The experiment shown in Fig.6 provides fur-
ther evidence for this conclusion.A CH
a
adlayer

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
63
À1
Fig.5 .C O
different isothermal reaction temperatures.
2
produced by TPR reaction at a heating rate of 3 Ks after the isothermal CH
a
þ O
a
reaction has gone to completion at
À1
2
O flux) for 35 s at a surface temperature of 385
Fig.6 .A CH
K while following the CO
a
adlayer (0.083 ML) is reacted with an O
2
molecular beam (0.025 MLs
2
, CO and H
2
partial pressures.At t ¼ 140 s the crystal is temperature ramped at a rate of 3 KsÀ1 to 650 K.
At this point (t ¼ 240 s) the O
2
beam is again turned on removing all remaining carbon.The inset is a magnification of the temperature
ramp.
À1
(0.083 ML) was prepared and the surface tem-
perature held at 385 K.After 10 s the source stage
of 3 K s to 650 K, where after the residual carbon
is oxidised to produce CO and CO (t ¼ 240 s).
2
flag was opened and an O beam with a flux of
2
The isothermal oxidation reaction at 385 K
produces CO2 and H O When the KW flag is
lowered (t ¼ 65 s) both the CO
decrease slowly to the base value indicating that
the reaction is still taking place.The H O profile is
À1
0
.025 ML s strikes the inert KW flag.At t ¼ 30 s
2
the KW flag was raised and surface oxidation
proceeds.Before oxidation was complete the KW
2 2
and H O signals
flag was lowered (t ¼ 65 s), blocking the O
2
beam.
The source stage flag was then closed (t ¼ 80 s).
2
not included here for the sake of clarity.This
demonstrates that the oxygen involved in the
The crystal temperature was then ramped at a rate

64
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
reaction with CH
, and not the short-lived metastable species O2a
The latter species has been implicated by Walker
and King [27] in the reaction of C with impinging
molecular O to form gaseous CO at 650 K, but
it plays no role in the reaction with CH
a
is the stable dissociated species,
3.2. TPR: CH
a
þ O
a
reaction
O
a
.
We obtained further insight into the mechanism
of CH oxidation by studying the TPR between O
and CH .The CH adlayer (0.09 ML) was pre-
a
a
a
2
2
a
a
.
pared by dosing the He seeded CH beam with a
À1
flux of 0.42 ML s , for 1 min, at a surface tem-
perature of 370 K.The crystal was cooled to 310 K
4
a
During the temperature ramp from t ¼ 140 s
(
(
inset Fig.6) three peaks are evolved: CO
T
2
¼ 410 K), H
2
(T
peak is produced by the reaction
with CO , while H is pro-
decomposition.Since the CO peak
p
¼ 479 K) and CO (T
p
¼ 510
2
and an O beam was dosed for 1 min to co-adsorb
0.094 ML of atomic O
was then ramped at a rate of 3 K s to 650 K,
where the residual C
partial pressures of CO
followed for the whole experiment.Fig.7 shows
the product profiles during the TPR part of the
experiment.
p
K).The tiny CO
2
a
.The crystal temperature
À1
of residual atomic O
duced by CH
a
a
2
a
was titrated with O
2
.The
a
maximum is at 510 K [28] and there are no stable
oxygenate intermediates on the surface at 510 K
2
, CO, H and H O were
2
2
[
mass balance, we find the amount of H produced
15–18] the CO peak is desorption limited.Using
2
by CH
amount of CO and CO
oxidation at 650 K.Therefore, at
¼ 410 K), the only species left on the surface
are CH and CO .This also implies that the first
a
decomposition is exactly equal to the
produced by residual C
t ¼ 155 s
This experiment allows us to draw the following
conclusions:
(1) No gas phase products were observed while
dosing at 310 K.This also leads to the conclusion
that there is no surface reaction at this temperature
2
a
(T
s
a
a
step in the reaction cannot be abstraction of a H
atom to produce C and OH .Therefore we con-
clude that the initial step is
between adsorbed O
possible surface products would be CHO
CO , OH and H ; but Wang and Masel [17] and
Ohtani et al.[16], found that CHO and CHO2a
thermally decompose at surface temperatures of
220 and 300 K, respectively, to produce CO , CO
and H and the latter would instantaneously react
with atomic O to produce gas phase H O(g).Since
a a
and CH , since the only
, CHO2a,
a
a
a
a
a
a
a
CH þ O ! CO þ H
ð1Þ
which probably proceeds though the intermedi-
a
a
a
a
2
a
ate formation of a short-lived CHO
species.
a
adsorbed
a
2
À1
a a 2 2 2
coadsorbed with 0.094 ML O to produce CO , CO, H O and H .The heating rate was 3 Ks
Fig.7 .T PR reaction of 0 0. 9 ML CH
and the H O signal was multiplied by a factor of six for clarity.
2

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
65
Fig.8 .T PR profiles of the CO
and 0.094 ML (solid lines) with a heating rate of 3 Ks
2
and CO produced when 0.09 ML CH
À1
a
is coadsorbed with O
a
pre-coverages of 0.066 ML (dashed lines)
.
no gas phase products are observed while dosing
we conclude that only CH
the surface.
1000 K no recombinative O
observed.
2
desorption peak is
a a
and O are present on
We found that the relative selectivities for the
individual products can be changed by varying the
pre-coverages of atomic O .Fig.8 shows TPR
(
reacts with the O to produce CO and H O, with
2) As the crystal temperature is ramped CHa
a
2
2
a
peak temperatures at around 380 K.We conclude
that the TPR reaction of CH Proceeds first
þ O
to H and CO and the H produced is rapidly
oxidised to H O.This is consistent with the results
of Capitino et al.[11], who showed that H is
oxidised at surface temperatures as low as 220 K.
profiles of the CO and CO produced when 0.09
2
a
a
ML of CH
ages of atomic O
change significantly, but CO
voured at higher O pre-coverages.The relative
a
is coasorbed with different pre-cover-
.The peak temperatures do not
production is fa-
a
a
a
a
2
2
a
a
selectivities for the individual products as a func-
tion of O pre-coverage can be seen in Fig.9;
(
3) Since the TPR CO þ O to produce CO
a
a
2
a
peak maximum occurs at ꢂ390 K [29], the peak
percentages are given with respect to the destina-
tion of adsorbed carbon, including that which is
temperature for maximum CO production, 385
2
K, indicates that the CO
the rate determining step.
As the surface temperature increases surfaces O
is consumed by the two competing processes, CO
and CH oxidation, until at 450 K no O remains.
a
þ O
a
surface reaction is
a
unreacted once all the O has been consumed.At
low oxygen pre-coverages CO is the favoured
product with a low overall conversion.As the
a
O
a
pre-coverage increases to 0.07 ML the CO se-
lectivity reaches a maximum of 32%.The CO
a
a
2
The remaining CHa thermally decomposes to
produce C and gas phase H , with a peak maxi-
selectivity increases to 10% and the overall con-
version is 42%.At the highest O a pre-coverage
studies (0.094 ML) the total conversion increases
to 66% while the CO selectivity decreases to 23%
a
2
mum at 479 K [24].The unoxidised CO a then de-
sorbs in a desorption limited peak at 515 K,
leaving C on the surface.We conclude that there
a
and that of CO
low O pre-coverages CO is produced almost ex-
clusively with a low overall conversion while at O
pre-coverages CO is the major product with a
2
increases to 32%.It is clear that at
is no oxygen left on the surface at 450 K, for the
following two reasons.Firstly, the reaction be-
tween atomic O and either CO or C to produce
a
a
a
a
a
2
CO is facile at a surface temperature of 500 K.
2
high overall conversion.This is consistent with our
model where gas phase CO is produced by CHa
Secondly, if the crystal temperature is ramped to

66
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
Fig.9 .P roduct selectivty as a function of O
tained by integrating the TPR peaks.
a a a
pre-coverage for the TPR of CH (0.09 ML) with atomic O .The selectivities are ob-
Fig.10 .S chematic of the TPR CH
a
a
þ O reaction.
oxidation with subsequent desorption while CO
2
is
a
OH nor the formation of a stable oxygenate
produced by the completing surface reaction, COa
oxidation.
The processes occurring on the surface during
the TPR reaction are shown schematically in Fig.
which is further oxidised.The following mecha-
nism is consistent with all the experimental ob-
servations:
SO2 FO2
1
0.The relative selectivity for each of the indi-
vidual products can be changed by varying the
pre-coverages of CH and O
O
2
ðgÞ ! 2O
a
ð2Þ
a
a
.
k
1
!
CH þ O ! CHO ! CO þ H
ð3Þ
ð4Þ
a
a
a
a
a
4
. Modelling
k2
CO
CO
a
þ O
a
!CO
2
ðgÞ
The isothermal and TPR experiments described
earlier show that the first step of CH oxidation is
neither the abstraction of H to produce C and
a
k3
a
!COðgÞ
ð5Þ
a

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
67
k4
virtually instantaneously converted to OH
a
or
O(g), respectively, at these temperatures, and
H þ O !OH
ð6Þ
ð7Þ
a
a
a
H
2
significant coverages in H and OH cannot occur.
a
a
k5
2
OH !H OðgÞ þ O
a
2
a
Applying the steady state approximation and re-
arranging Eqs.(12), (13) and (15), gives the fol-
k6
OH þ H !H OðgÞ
ð8Þ
are the rate constants for
lowing expression for PH2O
:
a
a
2
k
1
h
CH O
h
where k
2
, k
3
, k
4
, k
5
and k
6
P
H O
2
¼
ð16Þ
each of the individual reactions, while k
overall rate constant for the oxidation of CH
produce CO and H .Steps (6)–(8) represent the
oxidation of H to H O via the mechanism dis-
1
is the
2
a
to
This shows that the CH
termining for H
þ O
reaction is rate de-
O production and that the rate of
a
a
a
a
2
a
2
water production is half that of the overall reac-
tion rate.Inserting this expression and expression
cussed in Refs.[11–14].
Using a simple mean-field kinetic analysis, we
obtained the following set of five coupled differ-
ential equations which describe the variation in
adsorbate coverage as a function of time:
(14) into Eqs.(9)–(13) and integrating gives the
following expression for the time-dependent cov-
erages of the surface species:
Z
dhCH
h
CH ¼ hCHi À 2
P
H O dt
ð17Þ
2
¼
Àk
1
h
CH
h
O
ð9Þ
dt
Z
Z
Z
dh
O
hO ¼ 2FO2 sO2 dt À 3 PH2O dt À PCO2 dt
¼
2F s À k1hCHh À k hCOhO
O2 O2
O
2
dt
2
OH
ð18Þ
À k h h þ k h
ð10Þ
ð11Þ
4
H
O
5
Z
Z
dhCO
dt
h
CO ¼ 2
P
H O dt À
P
CO2 dt
ð19Þ
2
¼
k
1
h
CH
h
O
À k
2
h
CO
h
O
Therefore, the time-dependent reactive O
probability and the CO and H
can be used to determine the time-dependent sur-
face coverages.The product profiles and O up-
2
sticking
dhH
¼
k
1
h
CH
h
O
À k
k h h À 2k h
4
h
H
h
O
À k
6
h
OH
h
H
ð12Þ
ð13Þ
2
2
O partial pressures
dt
2
dhOH
2
OH
¼
À k6hOHh
OH
4
H
O
5
take curve for the reaction of 0.11 ML CH with
a
dt
À1
an O molecular beam (0.023 ML s ) at a surface
2
where hCH, h
O
, h
H
and hCO are the time-dependent
are the flux and reactive
sticking probablity of the incident O molecular
beam.The rates of CO (PCO2 ) and H
formation are given by the following expressions
temperature of 385 K are shown in Fig.1.
O
coverages. F 2
O
and s 2
The signals were subsequently corrected for
pumping speed and mass spectrometer sensitivi-
ties.Additionally, the H O signal was extremely
2
2
2O (PH2O
)
2
noisy due to the low sensitivity of the system
2
to H O.We performed an Igor Pro binomial
P
CO ðgÞ ¼ k
2
h
CO
h
O
ð14Þ
2
smoothing operation with a smoothing factor 2
on the signal to obtain a usable H O trace.To
and
2
2
OH
determine the time-dependent coverage profiles
shown in Fig.11, these signals were integrated and
inserted into Eqs.(17)–(19).
P
H OðgÞ ¼ k
5
h
þ k
6
h
OH H
h :
ð15Þ
2
The above system of coupled differential Eqs.
9)–(13) cannot be solved explicitly and must be
(
a
The O coverage rises rapidly for the first 10 s
solved numerically.A much simpler approach to
obtain meaningful data is to make the steady state
then more slowly at approximately 0.025 ML, but
after about 30 s, it rises more quickly before be-
approximations that dh
This is quite reasonable since any H
H
=dt ¼ 0 and dhOH=dt ¼ 0.
ginning to level off after about 120 s.The CO
coverage increases to a maximum of 0.025 ML at
a
a
or OH is
a

68
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
a 2
After 70 s the O coverage builds up and the O
sticking probability starts to drop again.This is of
course, a qualitative explanation in which the rel-
ative blocking efficiencies have been ignored.
However, it does show that the model is com-
2
pletely consistent with the O sticking profile.
In the model PCO2 and PH2O are given by Eqs.
14) and (16), respectively.Fig.12 shows the PCO2
(
and PH O temporal profiles constructed from the
2
a a a
coverages of the CH , O and CO .The respective
rate constants were determined using a best-fit
procedure.This yields k
ML s , respectively.
1
2
and k as 0.4 and 1.33
Fig.11.The coverages of the O
oxidation of the 0.11 ML CH
molecular beam at surface temperature of 385 K.
a
, CO
a
and CH
a
during the
À1
À1 À1
a
adlayer using a 0.023 MLs
O
2
The modelled profiles show very good agree-
ment with the experimental data, which is sur-
prising considering the simplicity of the model.
One major assumption is that the rate constants,
and hence the activation energies, are independent
5
0 s and then decreases quite sharply.The CH a
coverage drops almost linearly until the reaction is
complete after about 80 s.These coverages can be
used to explain the O
Fig.4.At t ¼ 0 the CH coverage is 0.11 ML. After
5 s, when the sticking probability drops to a
minimum (0.07), the coverages of CH , O and
CO are 0.083, 0.025 and 0.021 ML, respectively.
2
sticking profile at 385 K in
a a a
of O , CH and CO coverages.This is shown to
be valid for the first 80 s, when coverages are rel-
atively low, after which the model breaks down.A
1
a
a
2
close examination of the experimental CO profile
reveals that a secondary (slower) process takes
over after 80 s, where the simple analysis breaks
down.Bonzel and Burton [30] found for Pt{1 1 0}
a
The increased total coverage (0.129 ML) explains
the drop in the O sticking probability.At the
2
maximum (at 70 s) the coverages of CH , O and
a
at high O
and activation energy for the CO
a
coverages the pre-exponential factor
a
CO
The total coverage (0.097 ML) is relatively low,
which is why the O sticking probability is high.
a
are 0.017, 0.064 and 0.016 ML, respectively.
a
þ O
a
reaction
À1
À11
2
À1 À1
were 1:8 ꢀ 10
cm mol
respectively.At a surface temperature of 385 K
s
and 33 kJ mol ,
2
À1
Fig.12 .T he isothermal reaction of a 0 1. 1 ML CH
temperature of 385 K.The O , H O and CO
a
adlayer with an O
signals are corrected for different pumping speed and mass spectrometer sensitivities.
2
supersonic molecular beam of flux 0.023 MLs at a surface
2
2
2

D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
69
Mass balance was used to determine the varia-
tion of adsorbate coverages during the isothermal
oxidation reaction.Using these coverages we
modelled the product profiles and determined
the rate constants for both the CH
a
þ O
a
and
CO reactions.From an Arrhenius analysis
a
þ O
a
the activation energy and pre-exponential factor
for the CH þ O surface reaction are 71 Æ 3 kJ
a
a
À1
À6
À1
2
À1
mol and ð2:1 Æ 1:5Þ ꢀ 10 molecule cm s ,
respectively.
Fig.13 .A n Arrhenius plot for the reaction of CH
a a
with O .
Acknowledgements
À16
2
À1
this gives a rate constant of 6 ꢀ 10
cm mol
s .Using our model for the same reaction, we
À1
We acknowledge an equipment grant from the
UK EPSRC, a TMR Studentship from the Euro-
pean Community (D.T.P.W.) and EPSRC/Cam-
bridge Commonwealth Trust support (J.J.W.H.).
À15
2
À1 À1
obtain 1:4 ꢀ 10
cm mol s , which compares
very favourably.
The above analysis was used to determine the
rate constant for the CH þ O surface reaction at
a
a
temperatures of 370, 385, 400 and 425 K.An Ar-
rhenius analysis yields the pre-exponential factor
and activation energy for the process.The straight
References
ðk1Þ
[
1] D.A. Hickman, L.D. Schmidt, Science 259 (1993) 343.
[2] C.W.J. Bol, C.M. Friend, J. Am. Chem. Soc. 117 (1995)
1572.
3] C.W.J. Bol, C.M. Friend, Surf. Sci. Lett. 337 (1995) L800.
[4] C.W.J. Bol, C.M. Friend, J. Am. Chem. Soc. 117 (1995)
053.
line plot of ln
s
vs 1=T (Fig.13) yields a fre-
À6
À1
quency factor m ¼ ð2:1 Æ 1:5Þ ꢀ 10 molecules
cm s and an activation energy E
¼ 71 Æ 3
1
2
À1
À1
a
[
kJ mol .
8
[
[
5] M.K. Weldon, C.M. Friend, Chem. Rev. 96 (1996) 1391.
6] J.Kiss, A.Kis, F.Solymosi, Surf.Sci.454–456 (2000)
5
. Summary and conclusion
2
73.
7] F.Solymosi, I.Kovacs, K.R ꢀe v ꢀe sz, Surf.Sci.356 (1996)
21.
[
[
[
The isothermal oxidation of CH
and H O, exclusively, over the surface temperature
range 350–450 K.The reaction proceeds through
the dissociative O product, O , and not the di-
a 2
yields CO
1
2
8] X.L. Zhou, Z.M. Liu, J. Kiss, D.W. Sloan, J.M. White,
J.Am.Chem.Soc.117 (1995) 3565.
9] M.A. Quinlan, B.J. Wood, H. Wise, Chem. Phys. Lett. 118
2
a
(
10] G.B. Fisher, J.L. Gland, Surf. Sci. 94 (1980) 446.
1985) 478.
oxygen species, O .The rate of H O formation
2
2c
[
[
reaches a maximum rapidly and remains constant
until the reaction is near completion, while the
CO
occurs near the end of the reaction.The first step
in the overall reaction is the reaction of CH with
atomic O to produce CO and atomic H .Gas-
eous CO2 is produced by a simple Langmuir–
Hinshelwood reaction between CO and O , while
H O is produced via a multistep H oxidation.
11] A.T. Capitano, A.M. Gabelnick, J.L. Gland, Surf. Sci. 419
(
1999) 104.
[12] J.L. Gland, G.B. Fisher, E.B. Kollin, J. Catal. 77 (1982)
63.
2
rate rises more slowly to a maximum, which
2
[
[
[
13] K.M. Ogle, J.M. White, Surf. Sci. 139 (1984) 43.
14] L K. .Verheij, Surf.Sci.371 (1997) 100.
a
a
a
a
15] M.Endo, T.Matsumoto, J.Kubota, K.Domen, C.Hirose,
J.Phys.Chem.B 104 (2000) 4916.
[16] T.Ohtani, J.Kubota, A.Wada, J.Kondo, K.Domen,
C.Hirose, Surf.Sci.368 (1996) 270.
a
a
2
a
[
[
17] J.H. Wang, R.I. Masel, Surf. Sci. 243 (1991) 199.
18] A.B. Anton, J.E. Parmeter, W.H. Weinberg, J. Am. chem.
Soc.108 (1986) 1823.
2 2
During the TPR reaction CO, H , CO and
H
2
O are produced with the complete oxidation
favoured at low CH and high O
product CO
pre-coverages.
2
a
a
[
19] A. Hopkinson, X.-C. Guo, J.M. Bradley, D.A. King,
J.Chem.Phys.9 (1993) 1.

70
D.T.P. Watson et al. / Surface Science 505 (2002) 58–70
[
[
20] D.A. King, M.G. Wells, Surf. Sci. 29 (1972) 454.
21] A.V. Walker, B. Klotzer, D.A. King, J. Chem. Phys. 109
[25] D.T.P. Watson, J.J.W. Harris, D.A. King, J. Phys. Chem.
B, submitted for publication.
(
1998) 6879.
[26] D.T.P. Watson, J.J.W. Harris, D.A. King, Surf. Sci.,
submitted for publication.
[
[
22] Unpublished results.
23] T.E. Jackman, J.A. Davies, D.P. Jackson, W.N. Unertl,
P.R. Norton, Surf. Sci. 120 (1982) 389.
24] D.T.P. Watson, S. Titmuss, D.A. King, Surf. Sci. 505
[27] A.V. Walker, D.A. King, J. Chem. Phys. 112 (2000) 1937.
[28] J.Fair, R J. .Madix, J.Chem.Phys.73 (1980) 3480.
[29] Unpublished results.
[
(2002) 49.
[30] H.P. Bonzel, J.J. Burton, Surf. Sci. 52 (1975) 223.