Interaction of gaseous D atoms with alkyl halides adsorbed on Pt(111), H/Pt(111), and C/Pt(111) surfaces: Hot-atom and Eley-Rideal reactions. I. Methyl bromide

Article abstract of DOI:10.1063/1.479600

The interaction of gaseous D atoms with methyl bromide molecules adsorbed on Pt(111), hydrogen saturated Pt(111), and graphite monolayer covered Pt(111) surfaces was studied in order to elucidate the reaction mechanisms. The reaction kinetics at 85 K surface temperature were measured as a function of the methyl bromide precoverage by monitoring reaction products simultaneously with D atom exposure. On all substrates incoming atoms abstract the methyl group from adsorbed CH₃Br via gaseous CH₃D formation. In the monolayer regime of CH₃Br/Pt(111) pure hot-atom phenomenology was observed in the rates. At multilayer targets the fluence dependence of the kinetics gets Eley-Rideal-like. With coadsorbed H present, the reaction of D with adsorbed methyl bromide revealed in addition to CH₃D a CH₄ product. This and simultaneous abstraction of adsorbed H via gaseous HD and H₂ products clearly demonstrates that hot-atom reactions occur. At CH₃Br adsorbed on a graphite monolayer on Pt(111) the abstraction kinetics of methyl was found to agree with the operation of an Eley-Rideal mechanism. These observations are in line with the expectation that hot-atoms do not exist on a CTPt(111) surface but on Pt(111) and H/Pt(111) surfaces. The methyl abstraction cross-sections in the monolayer regime of methyl bromide were determined as about 0.25 A², irrespective of the nature of the substrate. This value is in accordance with direct, Eley-Rideal or hot-atom reactions.

Full text of DOI:10.1063/1.479600

Interaction of gaseous D atoms with alkyl halides adsorbed on Pt(111), H/Pt(111), and
C/Pt(111) surfaces: Hot-atom and Eley–Rideal reactions. I. Methyl bromide
S. Wehner and J. Küppers
Citation: The Journal of Chemical Physics 111, 3209 (1999); doi: 10.1063/1.479600
View online: http://dx.doi.org/10.1063/1.479600
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/111/7?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Quasiclassical study of Eley–Rideal and hot atom reactions of H atoms with Cl adsorbed on a Au(111) surface
J. Chem. Phys. 122, 074705 (2005); 10.1063/1.1851498
Interaction of gaseous D atoms with alkyl halides adsorbed on Pt(111), H/Pt(111), and C/Pt(111) surfaces: Hot-
atom and Eley–Rideal reactions. III. Isopropyl iodide
J. Chem. Phys. 111, 3225 (1999); 10.1063/1.479602
Interaction of gaseous D atoms with alkyl halides adsorbed on Pt(111), H/Pt(111), and C/Pt(111) surfaces: Hot-
atom and Eley–Rideal reactions. II. Ethyl iodide
J. Chem. Phys. 111, 3218 (1999); 10.1063/1.479601
Interaction of gaseous D atoms with CH 3 I adsorbed on Pt(111), H/Pt(111), and C/Pt(111) surfaces: From hot-
atom to Eley–Rideal phenomenology
J. Chem. Phys. 109, 294 (1998); 10.1063/1.476508
Eley–Rideal and hot-atom reaction dynamics of H(g) with H adsorbed on Cu(111)
J. Chem. Phys. 107, 6420 (1997); 10.1063/1.474302
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

JOURNAL OF CHEMICAL PHYSICS
VOLUME 111, NUMBER 7
15 AUGUST 1999
Interaction of gaseous D atoms with alkyl halides adsorbed on Pt„111…,
H/Pt„111…, and C/Pt„111… surfaces: Hot-atom and Eley–Rideal reactions.
I. Methyl bromide
S. Wehner
¨
Experimentalphysik III, Universitat Bayreuth, 95440 Bayreuth, Germany
a)
¨
J. Kuppers
¨
¨
Experimentalphysik III, Universitat Bayreuth, 95440 Bayreuth, Germany and Max-Planck-Institut fur
Plasmaphysik (EURATOM Association), 85748 Garching, Germany
͑Received 16 February 1999; accepted 6 May 1999͒
The interaction of gaseous D atoms with methyl bromide molecules adsorbed on Pt͑111͒, hydrogen
saturated Pt͑111͒, and graphite monolayer covered Pt͑111͒ surfaces was studied in order to elucidate
the reaction mechanisms. The reaction kinetics at 85 K surface temperature were measured as a
function of the methyl bromide precoverage by monitoring reaction products simultaneously with D
atom exposure. On all substrates incoming atoms abstract the methyl group from adsorbed CH3Br
via gaseous CH3D formation. In the monolayer regime of CH3Br/Pt(111) pure hot-atom
phenomenology was observed in the rates. At multilayer targets the fluence dependence of the
kinetics gets Eley–Rideal-like. With coadsorbed H present, the reaction of D with adsorbed methyl
bromide revealed in addition to CH3D a CH4 product. This and simultaneous abstraction of adsorbed
H via gaseous HD and H2 products clearly demonstrates that hot-atom reactions occur. At CH3Br
adsorbed on a graphite monolayer on Pt͑111͒ the abstraction kinetics of methyl was found to agree
with the operation of an Eley–Rideal mechanism. These observations are in line with the
expectation that hot-atoms do not exist on a C/Pt͑111͒ surface but on Pt͑111͒ and H/Pt͑111͒
surfaces. The methyl abstraction cross-sections in the monolayer regime of methyl bromide were
determined as about 0.25 Å2, irrespective of the nature of the substrate. This value is in accordance
with direct, Eley–Rideal or hot-atom reactions. © 1999 American Institute of Physics.
͓S0021-9606͑99͒71329-2͔
I. INTRODUCTION
*
*
͑ii͒
H atoms impinging at D occupied sites get hot H
atoms with probability 1Ϫpb and generate hot D
atoms with probability pb .
Reactions between gaseous atoms and adsorbed species
are currently described by the Eley–Rideal ͑ER͒ and hot-
1
*
*
͑iii͒ Hot H and D atoms travel across the surface and
stick at empty sites with probability ps and react at
occupied sites with probability pr .
atom mechanisms ͑HA͒.2 Atomic beam/laser spectroscopy
studies performed on abstraction of D adsorbed on Cu͑111͒
surfaces by H atoms towards HD revealed that the HD prod-
uct carries the available reaction energy,3 in line with the
operation of either mechanism. On the other hand, abstrac-
tion of D͑ad͒ ͑͑H͑ad͒͒ by gaseous H͑D͒ on Ni͑100͒,4
Within this purely HA based model, measured HD and D2
kinetics on Ni͑100͒, Pt͑111͒, and Cu͑111͒ surfaces could be
reproduced surprisingly well at fixed pb by variation of the
ratio pr /ps from 1 ͑Cu͒ over 0.1 ͑Ni͒ to 0.01 ͑Pt͒. The
branching probability pb served to adjust the yield of homo-
nuclear products.
Significance of HA based processes in atom–adsorbate
reactions was previously concluded from molecular dynam-
ics calculations, H˜D/Si(100) ͑Ref. 12͒ and H˜D/Cu(111)
͑Ref. 13͒. The origin of hot atoms in these calculations is the
strong attractive atom-surface potential and the inability of
the impinging atoms to transfer their kinetic energy to the
substrate in collisions with heavy surface atoms. This feature
was already stressed in the original work on the HA mecha-
nism by Harris and Kasemo.2
Pt͑111͒,5,6 Pt͑110͒,7 Pt͑100͒,8 and Cu͑111͒ surfaces, inves-
9
tigated through direct product detection, revealed HD kinet-
ics which are not compatible with the operation of ER
mechanisms. Furthermore, in these studies as well as in a
previous investigation on Ni͑110͒,10 homonuclear species,
D2͑H2͒, were obtained as reaction products which are unex-
pected in an ER scenario.
Model calculations of the HD and D2 kinetics in
H˜D͑ad͒ reactions were performed recently11 based on
three assumptions:
͑i͒
H atoms impinging at empty or H occupied sites on a
*
surface get hot H atoms.
HA type processes are not restricted to the H–D ͑ad͒
interaction. A study on the reactions of gaseous H͑D͒ with
methyl iodide ͑CH3I, CD3I͒ adsorbed and coadsorbed with
D͑H͒ on Ni͑100͒ surfaces14 revealed products which could
a͒
Author to whom correspondence should be addressed. Electronic mail:
kueppers@ipp.mpg.de, kueppers@uni-bayreuth.de
0021-9606/99/111(7)/3209/9/$15.00
3209
© 1999 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

¨
3210
J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
S. Wehner and J. Kuppers
only originate through reaction pathways in which hot atoms
were involved. As examples, for the reaction H˜D/CH3I on
Ni͑100͒ these products were in addition to the expected CH4
product CH3D and D2 which occur through the HA processes
ϭ2.4 Å2, ␴2ϭ1 Å2 were deduced. This magnitude of cross-
sections is expected if ER mechanisms operate in these re-
actions.
The observation of HA-type processes on metal surfaces
and ER type processes on C/Pt͑111͒ is in line with the fact
that a graphite monolayer on Pt͑111͒ removes the strong
attractive atom–surface interaction since the potential be-
tween the basal plane of graphite and H is weak. Accord-
ingly, the most important requirement for the generation or
existence of hot atoms on the surface is missing on C/Pt͑111͒
substrates and HA mechanisms are ruled out.
The present study was performed to further investigate the
elementary steps in reactions between gas phase deuterium
atoms and adsorbed alkyl halides. Target molecule methyl
bromide was selected since it allows to study the influence of
the methyl-halide binding energy by comparison with reac-
tions of methyl iodide. In order to confirm the role of the
substrate, metallic vs. nonmetallic, the reactions were inves-
tigated on Pt as well as on H and graphite covered Pt sur-
faces.
*
HϩD͑ad͒˜H͑ad͒ϩD ,
͑1a͒
͑1b͒
͑1c͒
*
D ϩCH3I͑ad͒˜CH3D͑gas͒ϩI͑ad͒,
*
D ϩD͑ad͒˜D2͑gas͒,
with hot atoms ͑stared species͒ as the actual reacting species.
The occurrence and kinetics of the CH3D and D2 products
contradict the operation of an ER mechanism which, for the
addressed case, would predict only a CH4 product through
HϩCH3I͑ad͒˜CH4͑gas͒ϩI͑ad͒.
͑1d͒
These investigations revealed that products and kinetics of
atom–adsorbate reactions on metallic substrates cannot be
explained by the ER scenario in a satisfactory manner. In
contrast, a recent study on the reactions between D atoms
and methyl iodide adsorbed on graphite monolayer covered
Pt͑111͒ surfaces, C/Pt͑111͒,15 revealed kinetics of the CH3D
reaction product which are strictly according to the operation
of an ER mechanism. The measurements utilized simulta-
neous detection of the methane product during application of
a D atom flux ⌽ which was applied at the adsorbate as a step
function of time t, i.e., ⌽ϭ0 at tϽ0 and ⌽ϭconst. at t
II. EXPERIMENT
The experiments were carried out in the UHV system
used for the previous study on the D/adsorbed methyl
͑methylene͒ iodide reactions.15,16 The system is equipped
with LEED/AES instrumentation and a setup for atom/
adsorbate reaction studies. D atoms were generated in an
atom source built according to a published design17 and char-
acterized recently with respect to its efficiency for hydrogen
atom generation.18 It consists of a W tube which is heated at
its front end by electron impact and connected at its cooled
back end to a deuterium gas supply. The atom flux delivered
by this source was calculated from the gas flow through the
tube, tube front temperature ͑1950 K͒, and the D2/2D equi-
librium data. D atom fluxes are given below in units of
monolayers sϪ1 with respect to the Pt͑111͒ surface atom den-
sity, 1 MlsϪ1ϭ1.5ϫ1015 D cmϪ2 sϪ1. The atom source is in-
corporated into a cylindrical, differentially pumped separate
vacuum system ͑source chamber͒ which sticks into the main
chamber. The source chamber has a front aperture which can
be closed by a mechanical shutter. For reaction measure-
ments, the sample with a well defined adsorbate coverage
was placed in front of the closed aperture. Opening of the
shutter defined the reaction start and reaction products were
monitored subsequently by a quadrupole mass spectrometer
͑QMS͒ located in the source chamber. Typically, partial
pressures in 15–20 preselected amu channels were multi-
plexed with appropriate sensitivities. Through fragmentation
analysis unambiguous product identification was possible.
An optical link between the QMS electronics and the PC
which controlled the experiment provided the high data
transfer rate required for fast multiplexing. Since the setup
resembles a pumped reactor, partial pressures monitored by
the QMS are proportional to reaction or desorption rates of
the respective species at or from the surface. With the atom
source switched off the arrangement served for thermal de-
sorption ͑TD͒ spectroscopy.
у0. The gas phase methane product rate, d CH D /dt, was
͓
͔
3
found to precisely match the rate predicted by the solution of
the kinetic equation which describes the reaction in an ER
scheme:
d CH D /dtϭ CH I ␴⌽ exp Ϫ␴⌽t͒,
͑2͒
͓
͔
͓
͔
͑
3
3
0
with CH I
as initial coverage of methyl iodide on
͓
͔
0
3
C/Pt͑111͒, ␴ as reaction cross-section, and ⌽(t) as specified
above. The reaction cross-section ␴ was determined from the
measured kinetics as 0.7 Å2. This magnitude of ␴ is in ac-
cordance with the locality of the reaction event in an ER
scenario.
The conclusion that on nonmetallic C/Pt͑111͒ substrates
atom–adsorbate reactions proceed according to the ER
mechanistic description gained support by a study on the
reactions between D atoms and methylene iodide (CH2I2)
adsorbed on C/Pt͑111͒ surfaces.16 The reaction sequence
DϩCH2I2͑ad͒˜CH2DI͑ad͒ϩI͑ad͒,
DϩCH2DI͑ad͒˜CH2D2͑gas͒ϩI͑ad͒,
͑3a͒
͑3b͒
can be interrupted after the first reaction step ͑3a͒ by per-
forming the reaction at substrate temperatures above the me-
thyl iodide desorption temperature and methyl iodide then is
the gaseous reaction product. Alternatively, below the me-
thyl iodide desorption temperature, the second reaction step
͑3b͒ also occurs and methane is the final product. The rates
of both gas phase products, CH2DI at high T and CH2D2 at
low T, can be calculated from the kinetic equations with the
assumption that ER-type mechanisms with two cross-
sections ␴ ͓3͑a͒ towards CH2DI͔ and ␴ ͓͑3b͒ towards
1
2
CH2D2͔ apply. The measured kinetics were found in excel-
A disk-shaped Pt͑111͒ single crystal was spot-welded
between two Ta wires which were attached to two Cu rods
lent agreement with calculated rates and cross-sections ␴
1
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
Hot-atom and Eley-Rideal reactions. I. Methyl bromide
3211
fixed at the bottom of a small cryostat. Via ohmic heating
and LN2 cooling of the Ta wires, sample temperatures be-
tween 85 and 1200 K could be installed. The sample tem-
perature was monitored and regulated by a PC-controlled
programmer connected to a Ni/CrNi thermocouple which
was welded to the back of the crystal.
Methyl bromide was obtained from Merck ͑99%͒ and
purified by freeze-pump cycles in a gas handling line.
Methyl bromide exposures given below were obtained from
uncorrected ion gauge readings. The sample was cleaned by
sputtering and annealing in oxygen. Graphite monolayers
were deposited on the Pt͑111͒ surface by repeated chemical
vapor deposition ͑CVD͒ of ethane and subsequent annealing
at 1200 K until complete suppression of hydrogen adsorption
on the surface indicated that there were no Pt patches uncov-
ered by C. A STM study has shown that at that stage the Pt
surface is covered by one monolayer high graphite islands.19
These islands withstand repeated heating to 1000 K and are
stable against H atom impact since they are only hydroge-
nated at their boundaries and do not erode chemically.20
III. RESULTS
A. Methyl bromide desorption
In order to establish the monolayer regimes of methyl
bromide on clean as well as on H and C covered Pt͑111͒
surfaces, thermal desorption measurements were performed
on the different substrates. Figure 1 shows spectra measured
in the monolayer and multilayer regimes after exposure of
CH3Br at 85 K substrate temperature. The spectra measured
on Pt͑111͒ are in excellent agreement with those published
recently by French and Harrison.21 Monolayer chemisorbed
methyl bromide desorbs between 120 and 240 K, and above
25 L exposure simultaneous growth of multilayer and mono-
layer features is observed. On C/Pt͑111͒ surfaces the chemi-
sorbed state is replaced by a weaker bound state which des-
orbs via a zero-order process around 125 K. This desorption
order indicates that methyl bromide prefers to adsorb in is-
lands on the C/Pt͑111͒ surface. Methyl bromide multilayer
features start to grow a little earlier than at completion of the
monolayer. On deuterium saturated Pt͑111͒ surfaces the
chemisorbed monolayer state of methyl bromide on Pt͑111͒
is also replaced by a weaker bound state which desorbs
around 128 K. Its desorption features indicate a zero-order
process. The desorption temperatures of the multilayer spe-
cies on Pt͑111͒, D/Pt͑111͒, and C/Pt͑111͒ surfaces are very
similar, as expected. Integration of the desorption spectra
revealed identical uptake as a function of methyl bromide
exposure at the three substrates and suggest a constant stick-
ing coefficient throughout the exposure regime investigated
here. The upper part of Fig. 1 displays the relation between
exposure and coverage obtained by setting the monolayer
capacity on C/Pt͑111͒ surfaces as unity.
FIG. 1. Thermal desorption spectra of methyl bromide adsorbed on Pt͑111͒
͑left͒, deuterium covered Pt͑111͒ ͑middle͒, and graphite monolayer covered
Pt͑111͒ ͑right͒. Adsorption temperature 85 K. Note the scale factors for each
panel. The spectra are stacked according to the respective lists of exposures
in L (1 Lϭ10Ϫ6 Torrϫsec). The relative coverages listed in each panel are
normalized with respect to the saturated monolayer coverage on C/Pt sur-
faces. The top panel illustrates the relation between coverage and exposure.
confirmed that the adlayers had been prepared with the re-
quired purity.
On Pt͑111͒ surfaces, a small methane desorption feature
at 290 K indicates that a few percent of adsorbed methyl
bromide decomposed. Decomposition of methyl bromide
was virtually absent on C/Pt͑111͒ and H/Pt͑111͒, in line with
its weak bonding to those substrates. The apparent modifica-
tion of admolecule Pt͑111͒ interactions through sandwiched
C or H monolayers was observed earlier with several adsor-
bates, benzene,22 methyl iodide,15 methylene iodide,16
isopropanol,23 propene oxide.24 This modification is indica-
tive of the effective shielding of the metal by H and C over-
layers.
For all three adsorbate systems it proved impossible to
prepare pure monolayers without multilayer contributions by
overdosing the surfaces and subsequently ͑even repeatedly͒
flashing the sample to an appropriate temperature in the
range 100 to 120 K. Therefore, abstraction experiments were
performed with as-prepared adsorbed layers at 85 K.
Thermal desorption was monitored simultaneously in the
amu 96 (CH3 81Br), amu 94 (CH3 79Br), and amu 15 ͑CH3
fragment͒ channels. The respective spectra were identical af-
ter proper scaling, as expected. Signals in other channels
monitored during TD spectroscopy, amu 2, 3, 4, 14, 18, 28
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

¨
3212
J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
S. Wehner and J. Kuppers
FIG. 2. Rates of CH3D formation during directing D atoms at methyl bro-
mide adsorbed on C/Pt͑111͒ surfaces. Exposures and coverages listed
specify the initial condition of the CH3Br adlayers. The D flux was started at
tϭ0 and kept constant thereafter. The inset emphasizes the early reaction
period.
FIG. 3. Rates of CH3D formation during directing D atoms at methyl bro-
mide adsorbed on H covered Pt͑111͒ surfaces. Other details see caption of
Fig. 2.
thereafter. The initial rate jumps as well as the amounts of
CH3D produced during subsequent reaction increase with in-
creasing methyl bromide coverage. The inset in Fig. 2 illus-
trates the measured rates on an expanded time-scale which
emphasizes the initial period of the reaction and clarifies that
rate jumps occur at reaction start and not continuous rate
increases from zero.
CH3D formation rates measured with other D atom
fluxes, up to 0.48 MlsϪ1, revealed identical kinetics as those
shown in Fig. 2 with the time axis rescaled to a fluence
B. Abstraction from methyl bromide on C/Pt„111…
Exposure of gaseous D atoms to methyl bromide ad-
sorbed on C/Pt͑111͒ surfaces revealed CH3D as the only gas-
eous product, identified through its tabulated fragmentation
pattern in the amu 17, amu 16, and amu 15 channels.25 Fig-
ure 2 illustrates the evolution of the CH3D rate with reaction
time for increasing methyl bromide coverages established
prior to reaction. The substrate temperature was fixed at 85 K
and the atom flux held constant at 0.28 MlsϪ1 during the
measurements. Since the adsorption energy of methane on
C/Pt͑111͒ is very small, at 85 K reaction temperature desorp-
tion of methane is not rate limiting and the rates shown in
Fig. 2 are rates of formation of CH3D in the D/methyl bro-
mide interaction. The reactions proceed until complete con-
sumption of methyl groups available in methyl bromide on
the surface, verified through post-reaction desorption spectra.
Remaining bromine had to be removed from the surface by
flashing the sample to 1000 K in order to recover clean
C/Pt͑111͒.
*
͑flux time͒ axis and proper adjustment of the rate scale.
CH2D2 and DBr products were never detected.
C. Abstraction from methyl bromide on H/Pt„111…
Reaction measurements with coadsorbed H and methyl
bromide were performed at Pt͑111͒ surfaces which were ini-
tially saturated with H via exposure of 1000 L H2 at 85 K
followed by subsequent installation of a required methyl bro-
mide coverage by exposing CH3Br to the H covered surface.
The kinetics of CH3D formation during D atom exposure
measured at increasing methyl bromide coverages are shown
in Fig. 3. Like on C/Pt͑111͒ surfaces, methyl groups in ad-
sorbed methyl bromide are completely consumed during re-
action. The rates jump at reaction start to values which in-
crease with increasing methyl bromide coverage.
The reaction kinetics for different methyl bromide cov-
erages apparent from the rates shown in Fig. 2 exhibit com-
mon features. At reaction start ͑opening of the shutter͒ the
methane rates jump to their maximum values and fall off
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
Hot-atom and Eley-Rideal reactions. I. Methyl bromide
3213
FIG. 5. Rates of CH3D formation during directing D atoms at methyl bro-
mide adsorbed on Pt͑111͒ surfaces. Other details see caption of Fig. 2.
FIG. 4. Rates of H2, HD, CH4, and CH3D formation during directing D
atoms at methyl bromide adsorbed on H covered Pt͑111͒ surfaces. Note the
different stacking order of the rate curves as a function of the methyl bro-
mide coverage indicated by the exposures and coverages. The spikes at t
ϭ0 in the H2 and HD rates are due to friction desorption caused by the
shutter plate sliding along the front face of the source chamber.
tween D and adsorbed methyl bromide on Pt͑111͒. The mea-
sured rates are shown in Fig. 5. Similar to the kinetics of the
methane product in the reaction of D with coadsorbed
H/CH3Br, at reaction start the rates exhibit a rate step and a
subsequent increase. After achieving a maximum the rates
fall to zero in a late reaction period. Post-reaction desorption
spectra confirmed that methyl abstraction was complete.
Subsequently, in an early reaction period, the rates grow,
achieve maxima at reaction times which increase with me-
thyl bromide coverage, and fall off in late reaction periods.
The inset in Fig. 3 emphasizes these features. A growth of
the CH3D rate in the early reaction period during which the
methyl bromide coverage decreases contradicts the Eley–
Rideal reaction mechanism and suggests the presence of a
more complex reaction scenario.
The collection of kinetic data shown in Fig. 4 confirms
this expectation. In addition to CH3D, H2, HD, and CH4 are
reaction products of the D-methyl bromide H/Pt͑111͒ inter-
action, and each product occurs with its own kinetics. The
occurrence of HD and H2 is required since from the investi-
gations mentioned in the introduction, it is known that the
D/H͑ad͒ interaction results in the release of these products.
Furthermore, a CH4 product is not unexpected in view of the
earlier study on Ni͑100͒.14 The yield and kinetics of the four
species will be analyzed in detail below.
IV. DISCUSSION
The reactions of D atoms with adsorbed methyl bromide
revealed CH3D as product, irrespective of the substrate,
Pt͑111͒, H/Pt͑111͒, or C/Pt͑111͒. In view of previous results
on the abstraction of methyl by H from methyl iodide ad-
sorbed on Cu͑111͒ ͑Ref. 26͒ and Ni͑100͒ ͑Ref. 14͒ surfaces,
this is not surprising. The CH3–Br bond strength, 293
kJ/mol,27 is only a little bigger than that of CH3–I, 234
kJ/mol,28 and the CH3–H bond strength of 438 kJ/mol ͑Ref.
27͒ makes the methyl abstraction reaction substantially exo-
thermic even if the Maxwellian distribution of D atom ener-
gies delivered from the heated source, centered at about 20
kJ/mol, is not taken into account. There is a further contri-
bution to the reaction energy from the gain in adsorption
energy of Br vs CH3Br.
D. Abstraction from methyl bromide on Pt„111…
CH3D, identified through its fragmentation pattern, was
the only gaseous product detected during the interaction be-
A CH2D2 product was never observed, indicating that
abstraction of H by D from the methyl group of CH3Br fol-
lowed by subsequent hydrogenation back to CH2DBr is
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

¨
3214
J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
S. Wehner and J. Kuppers
FIG. 6. Reaction yields of CH3D and CH4 deduced from integrated rates of
methane as a function of initial methyl bromide coverage determined
through TDS ͑top͒. Rate steps of CH3D and CH4 as a function of initial
methyl bromide coverage ͑bottom͒.
FIG. 7. Logarithmic representation of the CH3D rates shown in Figs. 2, 3,
and 5. C/Pt͑111͒ ͑top͒, H/Pt͑111͒ ͑middle͒, Pt͑111͒ ͑bottom͒.
between rate step and initial coverage of the surface reactant,
as expressed in Eq. ͑2͒, is fulfilled. A logarithmic plot of the
CH3D rates against reaction time is shown in Fig. 7 ͑top͒.
Excluding the initial 50 s of reaction time, the logarithms of
the rates decrease linearly with reaction time, which makes
the rate decay proportional to exp(Ϫ␴⌽t), as required by ER.
The scaling of the rates with the fluence ⌽t was confirmed
by the flux dependence of the rates. With this requirement
fulfilled, the extraction of a cross-section from the logarith-
mic rates is justified. The cross-section derived from Fig.
7͑top͒ is ␴ϭ0.25 Å2. This value is in the range expected for
ER type reactions and significantly lower than that measured
in abstraction of methyl from methyl iodide on C/Pt͑111͒,
␴ϭ0.7 Å2.15 Energetic and steric reasons like exothermicity,
target molecule size, and its orientation on the surface might
be the origin of these different cross-sections. The fact that
the rate steps are proportional to the coverage not only in the
monolayer range but also for the second layer suggests that
incoming D atoms can access monolayer species for reaction
even through the second layer. This is plausible since CH3Br
molecules are loosely packed due to their large dipole mo-
ment.
much slower than the abstraction of the methyl group from
methyl bromide. The same was observed in the D-methyl
iodide interaction. With adsorbed CH3Cl, however, the ab-
straction cross-section of H from the methyl group is com-
parable to abstraction of the methyl group.29 The stronger
CH3–Cl bond energy as compared to CH3–Br makes this
feature plausible.
As mentioned, during abstraction all methyl groups of
adsorbed methyl bromide available on the surface were con-
sumed. The methane yields obtained from integrated rates
are displayed in Fig. 6 ͑top͒ as a function of the initial bro-
mide coverage. On the three substrates the CH3D yield is
proportional to this relative coverage. Therefore, the energy
set free upon reaction is not channelled to the adsorbate or
substrate since in that case some of the adsorbed methyl
bromide would desorb and not be available for reaction. This
would induce a nonlinearity between the methane yield and
the coverage of methyl bromide present on the surface at
reaction start.
The kinetics of the D˜CH3Br/C/Pt͑111͒ reaction is
simple to analyze. Figure 6 ͑bottom͒ shows that the rate steps
deduced from Fig. 2 are proportional to the coverage if one
restricts to the coverage range below ⌰ϭ2. Therefore, an
essential condition for the ER mechanism, proportionality
The topmost curves in Fig. 2 and Fig. 7 ͑top͒ correspond
to a thick methyl bromide multilayer, ⌰ϭ5. The cross-
section derived from the exponential CH3D rate decay is
only 0.12 Å2 and the rate step at that coverage is much
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
Hot-atom and Eley-Rideal reactions. I. Methyl bromide
3215
smaller than expected by extrapolation from the monolayer
regime. The observation that abstraction from thick multilay-
ers exhibits a smaller cross-section than those characteristic
for monolayers was also made in abstraction from adsorbed
CH3I and its isotope analogs.15
The steep slope of the logarithmic rates in the first 50 s
of reaction time in Fig. 7͑top͒ suggests that initially abstrac-
tion is faster than at later times. Since abstraction from mul-
tilayers exhibits a smaller cross-section than from monolay-
ers, initial faster abstraction cannot originate from the
multilayer components in the adsorbed methyl bromide
mono- and bilayers.
Whereas on C/Pt͑111͒ the reaction of D with methyl
bromide can be assigned as an ER reaction, the products and
their kinetics shown in Fig. 4 clearly contradict the operation
of an ER mechanism in abstraction from CH3Br on the
H/Pt͑111͒ surface. In reading the kinetic data shown in Fig.
4, it has to be noticed that the H2 and HD rates refer ͑from
top to bottom͒ to increasing methyl bromide coverage,
whereas the methane rates refer ͑from top to bottom͒ to de-
creasing methyl bromide coverage. The marked differences
in the decay of the rates suggests individual pathways for the
various products. According to the products observed, the
overall reactions occurring with mixed H/CH3Br adlayers
are:
Concentrating on the H2 and HD rates in Fig. 4 with the
smallest methyl bromide coverage ͑⌰ϭ0.07 after 5 L methyl
bromide exposure͒ it is apparent that these rates do not
achieve their maximum values right at reaction start, which
they would if ER-type processes would occur. Since through
exposure of molecular H2 to Pt͑111͒ surfaces only a cover-
age of about 0.7 is achieved30 there are still empty sites
available on the surface. Hot-atom species generated through
reactions ͑8a͒ and ͑9a͒ which travel across the surface can
stick at these empty sites and are unavailable for reaction
͑8b͒ or ͑9b͒. Therefore, the HD and H2 rates at reaction start
do not achieve their maximum values. During the early re-
action period after reaction start, through hot-atom sticking
events, the number of empty sites gets smaller and hot-atom
reactions ͑8b͒ and ͑9b͒ get more probable. By this effect, the
rates of H2 and HD increase. They achieve maximum values
if the number of empty sites cannot decrease any more since
stationary conditions with respect to the HϩD coverage is
reached. Behind its maximum the HD rate decreases expo-
nentially, as required for a quasifirst-order reaction, scheme
͑9a͒, ͑9b͒. The cross-section deduced from the rate decay of
HD is ␴ϭ1.4 Å2, in good agreement with the earlier result.16
The H2 rate decays faster since it is a second-order process
with respect to the coverage of H͑ad͒, see reaction scheme
͑8a͒, ͑8b͒.
Approaching higher methyl bromide coverages, the H2
and HD rates decrease since adsorbed H is consumed by the
reactions towards CH4, reaction ͑6͒, and D in reactions to-
wards CH3D, reaction ͑7͒. These competing reactions also
cause the rate maxima of the H2 and HD rates to shift to later
reaction time with increasing methyl bromide coverage.
The phenomenology of the methane rates suggest the follow-
ing pathways via hot-atoms to CH4:
H2 product: DϩH͑ad͒ϩH͑ad͒˜H2͑gas͒,
HD product: DϩH͑ad͒˜HD͑gas͒,
CH4 product: DϩCH3Br͑ad͒ϩH͑ad͒˜CH4͑gas͒ϩBr͑ad͒,
͑4͒
͑5͒
͑6͒
CH3D product: DϩCH3Br͑ad͒˜CH3D͑gas͒ϩBr͑ad͒. ͑7͒
Reactions ͑4͒ and ͑5͒ were investigated previously in this
laboratory by studies of the H˜D͑ad͒/Pt͑111͒ ͑Ref. 5͒ and
D˜H͑ad͒/Pt͑111͒ ͑Ref. 6͒ interaction and were identified as
hot-atom reactions. The kinetics of the H2 and HD products
in Fig. 4 agree with these earlier results and their essential
characteristics were explained in a simplified hot-atom sce-
nario by the model calculations11 mentioned in the introduc-
tion. The elementary reaction steps including hot atom spe-
cies behind reaction ͑4͒ are ͑with starred atoms specifying
hot-atom species͒:
*
DϩH͑ad͒˜D͑ad͒ϩH ,
͑10a͒
͑10b͒
*
H ϩCH3Br͑ad͒˜CH4͑gas͒ϩBr͑ad͒,
and to CH3D:
*
DϩH͑ad͒˜H͑ad͒ϩD ,
͑11a͒
͑11b͒
*
D ϩCH3Br͑ad͒˜CH3D͑gas͒ϩBr͑ad͒.
It is clear that CH4 can only occur through a hot-atom
mechanism. Through reactions ͑10a͒ and ͑10b͒ its rate as-
sumes its maximum right at reaction start, because at that
moment the coverage of H͑ad͒ at the surface is bigger than at
any later time. The consumption of H͑ad͒ towards H2 and
HD products effectively competes with CH4 formation.
Therefore, the rates of CH4 decrease continuously with reac-
tion time and the CH4 yields are much smaller than that of
CH3D, compare Fig. 6. The reaction sequence ͑10a͒, ͑10b͒
allows to describe CH4 production through a quasifirst-order
rate law, and a reaction cross-section of ␴ϭ0.7 Å2 was ob-
tained.
*
DϩH͑ad͒˜H ϩD͑ad͒,
͑8a͒
͑8b͒
*
H ϩH͑ad͒˜H2͑gas͒,
and behind reaction ͑5͒
*
DϩH͑ad͒˜D ϩH͑ad͒,
͑9a͒
͑9b͒
*
D ϩH͑ad͒˜HD͑gas͒.
With ongoing reaction H͑ad͒ gets replaced by D͑ad͒ and D2
products occur through the reaction steps
The CH3D kinetics is also not according to an ER
mechanism. Its rate jumps at reaction start since step ͑11a͒
provides hot D species required for reaction ͑11b͒. At on-
going reaction, through depletion of H͑ad͒, the formation of
CH4 and HD becomes less competitive and the CH3D rate
increases. Through the decrease of methyl groups available
*
DϩD͑ad͒˜D ϩD͑ad͒,
͑9c͒
͑9d͒
*
*
D ϩD͑ad͒˜D2͑gas͒.
The D2 product could not be monitored since the atom
source was operated with D2.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

¨
S. Wehner and J. Kuppers
3216
J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
for reaction a rate maximum is obtained. In the late reaction
period adsorbed hydrogen is replaced by adsorbed deuterium
and the reaction step
*
Through reaction ͑15͒ more hot D atoms are made available
for reaction and the CH3D rate increases. It achieves a maxi-
mum when an optimum condition is achieved concerning the
concentration of adsorbed methyl bromide molecules and
*
DϩD͑ad͒˜D͑ad͒ϩD ,
͑11c͒
*
rate of available hot D atoms. At ongoing reaction, through
*
replaces reaction ͑11a͒ for providing the hot D species nec-
consumption of the methyl groups, the CH3D rate decreases.
Comparing with Fig. 5 it is seen that this pattern is verified
by the experiments in the methyl bromide monolayer cover-
age range where multilayer contributions can be neglected,
exposures less than 65 L. At higher exposures multilayers
develop and shield the metal from the incoming D atoms. An
ER-type mechanism should then contribute, indicated by a
shift of the rate maximum towards the reaction start. Check-
ing with the kinetics in Fig. 5 this expectation is confirmed.
The CH3D yield and rate step in Fig. 6 are as expected from
the previous discussion.
The cross-section revealed from the logarithmic rate plot
in Fig. 7 ͑bottom͒, restricted to the late reaction period, is
␴ϭ0.28 Å2. This value is close to those deduced above for
the other substrates and its magnitude is as expected. The
topmost logarithmic rate in Fig. 7 ͑bottom͒ indicates that
with increasing thickness of the multilayers the cross-section
gets smaller, as already seen in the top curve in Fig. 7 ͑top͒.
Since its corresponding coverage is much bigger than that to
which Fig. 7 ͑bottom͒ refers, this feature is more clearly seen
there.
In the above discussion the concepts of ER and HA
mechanisms were successfully applied to abstraction of me-
thyl from methyl bromide adsorbed in mono- and multilayers
on various substrates. Products and the phenomenology of
their kinetics could be rationalized in a qualitative but satis-
factory manner using the simple principles developed
earlier.5,6,11 According to these principles, on substrates with
weak atom–surface interaction, atom–adsorbate reactions
should proceed via ER mechanisms, whereas strong atom–
surface interaction leads to HA mechanisms. In the latter
case, consequences of hot-atom generation and the interplay
between hot-atom sticking and reaction induces products and
kinetics which are unexpected in an ER scenario. The experi-
mental foundation for support of the general applicability of
these principles is still rather limited since only a few studies
exist from which satisfactory kinetic information on atom–
adsorbate reactions is available. However, the present study
revealed results which are in line with expectations from
these principles.
essary for reaction.
*
*
The branching towards H or D in competing steps
*
͑10a͒ and ͑11a͒ favors the production of D , and therefore
more CH3D molecules than CH4 molecules are formed. This
preference for formation of hot species from the impinging
atom rather than from the adsorbed atom is generally
observed:4–10 in D˜H͑ad͒ reactions, HD is the major prod-
uct and H2 only in the percent range. Likewise, in H˜D͑ad͒
reactions, HD product molecules outnumber D2 molecules.
The rate steps of CH4 and CH3D are quite similar as
depicted in Fig. 6 ͑bottom͒. This illustrates the effectivity of
HD formation, since at reaction start adsorbed H and ad-
*
sorbed CH3Br compete for reaction with hot D atoms.
It is not surprising that the complex reaction scenario in
the early reaction period causes the logarithmic plot of the
CH3D rates in Fig. 7 ͑middle͒ to exhibit a clear linear de-
crease only in the late period. The cross-section deduced in
this period for the abstraction of methyl from methyl bro-
mide by D is ␴ϭ0.25 Å2, very similar to the value obtained
above at the C/Pt͑111͒ surface. This is expected since for the
abstraction process itself it should not matter very much
whether the substrate is a metal or not. The value 0.25 Å2 is
smaller than the cross-sections towards HD ͑1.4 Å2͒ and CH4
͑0.7 Å2͒ and explains why the CH3D product exhibits such a
slow decay, see Fig. 4.
With the reaction between D atoms and adsorbed methyl
bromide on C/Pt͑111͒ and H/Pt͑111͒ surfaces identified as
Eley–Rideal or hot-atom reactions, the discussion of the re-
sults measured with methyl bromide on Pt͑111͒ is quite
simple. Since the substrate is metallic, the following reac-
tions at reaction start are expected:
Hot-atom generation on empty sites:
*
* *
Dϩ ˜D ϩ .
͑12͒
͑13͒
͑14͒
Hot-atom reaction with adsorbed CH3Br:
*
D ϩCH3Br͑ad͒˜CH3D͑gas͒ϩBr͑ad͒.
Hot-atom sticking:
* *
D ϩ ˜D͑ad͒.
V. CONCLUSIONS
At later times in addition the following reactions occur:
The interaction of D atoms with methyl bromide ad-
sorbed on Pt͑111͒, hydrogen covered Pt͑111͒, and graphite
monolayer covered Pt͑111͒ surfaces leads to CH3D as gas-
eous product, with Br remaining on the surface. The kinetics
of the gaseous products were used to identify the reaction
mechanisms leading to these products. On C/Pt͑111͒ pure
Eley–Rideal phenomenology was observed. On H/Pt͑111͒,
products like H2 and CH4 occur simultaneously with CH3D
and HD, indicative of hot-atom reaction mechanisms. On
Pt͑111͒ the kinetics of the CH3D product also is in line with
hot-atom mechanisms. The cross-sections for CH3D forma-
Hot-atom generation on D occupied sites:
*
DϩD͑ad͒˜D ϩD͑ad͒.
͑15͒
͑16͒
Hot-atom reaction with adsorbed D:
*
D ϩD͑ad͒˜D2͑gas͒.
The combination of reactions ͑12͒ and ͑13͒ should produce a
rate step at reaction start. Reaction ͑14͒ is competitive to ͑13͒
and affects the step height. At ongoing reaction empty sites
are filled and the competition of step ͑14͒ gets less important.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34

J. Chem. Phys., Vol. 111, No. 7, 15 August 1999
Hot-atom and Eley-Rideal reactions. I. Methyl bromide
3217
tion are about 0.25 Å2, significantly smaller than that for
abstraction of methyl from adsorbed methyl iodide, 0.7 Å2,
and probably due to the higher methyl–halide bond energy in
methyl bromide. Eley–Rideal phenomenology on C/Pt͑111͒
surfaces and hot-atom phenomenology on H/Pt͑111͒ and
Pt͑111͒ are in line with expectations drawn from role of a
metallic substrate in atom–adsorbate reactions.
Phys. Lett. 237, 468 ͑1995͒; B. Jackson and M. Persson, J. Chem. Phys.
103, 6257 ͑1995͒; S. Caratzoulas, B. Jackson, and M. Persson, ibid. 107,
6420 ͑1997͒.
14
15
16
¨
Th. Kammler and J. Kuppers, J. Chem. Phys. 107, 287 ͑1997͒.
¨
S. Wehner and J. Kuppers, J. Chem. Phys. 109, 294 ͑1998͒.
¨
S. Wehner and J. Kuppers, Chem. Phys. Lett. 288, 873 ͑1998͒.
17 U. Bischler and E. Bertel, J. Vac. Sci. Technol. A 11, 458 ͑1993͒.
18 C. Eibl, G. Lackner, and A. Winkler, J. Vac. Sci. Technol. A 16„5…, 1
͑1998͒.
19 T. A. Land, R. Michely, R. J. Behm, J. C. Hemminger, and G. Comsa,
ACKNOWLEDGMENT
Surf. Sci. 264, 261 ͑1992͒.
20
¨
Th. Zecho, A. Horn, J. Biener, and J. Kuppers, Surf. Sci. 397, 108 ͑1997͒.
Support of this research by the Deutsche Forschungsge-
meinschaft is gratefully acknowledged.
21 C. French and I. Harrison, Surf. Sci. 387, 11 ͑1997͒.
22
¨
¨
C. Lutterloh, J. Biener, K. Pohlmann, A. Schenk, and J. Kuppers, Surf.
Sci. 352-354, 133 ͑1996͒.
J. Biener, C. Lutterloh, A. Schenk, K. Pohlmann, and J. Kuppers, Surf.
Sci. 365, 255 ͑1996͒.
A. Dinger, C. Lutterloh, J. Biener, and J. Kuppers, Surf. Sci. 421, 44
1 D. D. Eley and E. K. Rideal, Nature ͑London͒ 146, 401 ͑1940͒.
2 J. Harris and B. Kasemo, Surf. Sci. 105, L281 ͑1981͒.
23
24
¨
¨
3 C. T. Rettner and D. J. Auerbach, J. Chem. Phys. 104, 2732 ͑1996͒.
¨
4
¨
Th. Kammler, J. Lee, and J. Kuppers, J. Chem. Phys. 106, 7362 ͑1997͒.
͑1999͒.
5
6
7
¨
¨
S. Wehner and J. Kuppers, J. Chem. Phys. 108, 3353 ͑1998͒.
S. Wehner and J. Kuppers, Surf. Sci. 411, 46 ͑1998͒.
Th. Biederer, Th. Kammler, and J. Kuppers, Chem. Phys. Lett. 286, 15
25 A. Cornu and R. Massot, Compilation of Mass Spectral Data, 2nd ed.
͑Heyden, London, 1975͒, Vol. 2.
¨
26 Y.-S. Park, J.-Y. Kim, and J. Lee, Surf. Sci. 363, 62 ͑1996͒; C. T. Rettner,
D. J. Auerbach, and J. Lee, J. Chem. Phys. 105, 10115 ͑1996͒.
27 CRC Handbook of Physics and Chemistry, 75th ed., edited by D. R. Lide
͑CRC, Boca Raton, 1994͒.
͑1998͒.
8
9
¨
Th. Zecho, B. Brandner, and J. Kuppers, Surf. Sci. 418, L26 ͑1998͒.
¨
Th. Kammler and J. Kuppers ͑unpublished͒.
10 G. Eilmsteiner, W. Walkner, and A. Winkler, Surf. Sci. 352-354, 263
28 A. A. Radzig and B. M. Smirnov, Reference Data on Atoms, Molecules,
͑1996͒.
11
and Ions ͑Springer Verlag, Berlin, 1985͒.
¨
Th. Kammler, S. Wehner, and J. Kuppers, J. Chem. Phys. 109, 4071
29
¨
S. Wehner and J. Kuppers ͑unpublished͒.
͑1998͒.
30 K. E. Lu and R. R. Rye, Surf. Sci. 45, 677 ͑1974͒; K. Christmann, G. Ertl,
and T. Pignet, ibid. 54, 365 ͑1976͒; P. R. Norton, J. A. Davies, and T. E.
Jackman, ibid. 121, 103 ͑1982͒.
12 P. Kratzer, J. Chem. Phys. 106, 6752 ͑1997͒.
13 B. Jackson and M. Persson, J. Chem. Phys. 96, 2378 ͑1992͒; M. Persson
and B. Jackson, ibid. 102, 1078 ͑1995͒; M. Persson and B. Jackson, Chem.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.193.242.161 On: Mon, 08 Dec 2014 18:23:34