Interaction of D(H) atoms with physisorbed benzene and (1,4)-dimethylcyclohexane: Hydrogenation and H abstraction

Article abstract of DOI:10.1063/1.470934

Benzene and (1,4)-dimethyl-cyclohexane monolayers were physisorbed on graphite covered Pt(111) surfaces. Exposure of benzene monolayers at 125 K to D atoms (1700 K) initially hydrogenates sp² hybridized C atoms with a cross section of ca. 8 A² producing C₆H₆D intermediates. Additional D atom reactions either transform this intermediate via a second hydrogenation reaction to cyclohexadiene-d₂, C₆H₆D₂, or restore benzene, C₆H₅D, via H abstraction. Once the aromaticity is broken, successive hydrogenation of the diene occurs rapidly generating the saturated cyclohexane-d₆, C₆H₆D₆. The C₆H₅D reaction product can undergo further H/D exchange reactions and, at any level of deuteration, the benzene species might get hydrogenated. Monolayers of the saturated hydrocarbon (1,4)-dimethyl-cyclohexane (DMCH) that are exposed to D atoms produce deuterated DMCH via successive abstraction/hydrogenation reactions. Thermal desorption mass spectra revealed that H atoms at the ring were exchanged with an apparent cross section of 1.7 A². Methyl groups H atoms were exchanged much more slowly than ring H atoms. It was also observed that D exposed molecules/radicals exhibit a tendency to desorb from the surface, which is ascribed to the exothermicity of the reactions which lead to these species.

Full text of DOI:10.1063/1.470934

Interaction of D(H) atoms with physisorbed benzene and (1,4)dimethylcyclohexane:
Hydrogenation and H abstraction
C. Lutterloh, J. Biener, A. Schenk, and J. Küppers
Citation: The Journal of Chemical Physics 104, 2392 (1996); doi: 10.1063/1.470934
View online: http://dx.doi.org/10.1063/1.470934
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/104/6?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Kinetics, mechanism, and dynamics of the gas-phase H(D) atom reaction with adsorbed D(H) atom on Pt(111)
J. Chem. Phys. 113, 2856 (2000); 10.1063/1.1305912
A kinetic study of the interaction of gaseous H(D) atoms with D(H) adsorbed on Ni(100) surfaces
J. Chem. Phys. 106, 7362 (1997); 10.1063/1.473697
The rate of the reaction D+H2(v = 1)→DH+H
J. Chem. Phys. 77, 3478 (1982); 10.1063/1.444292
Fluorescence and Intersystem Crossing of Benzeneh 6, d 6, and 1,4d2
J. Chem. Phys. 51, 182 (1969); 10.1063/1.1671705
The Reaction of Hydrogen Atoms with Dimethyl Mercury
J. Chem. Phys. 13, 559 (1945); 10.1063/1.1723994
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

Interaction of D(H) atoms with physisorbed benzene and (1,4)-dimethyl-
cyclohexane: Hydrogenation and H abstraction
a)
¨
C. Lutterloh, J. Biener, A. Schenk, and J. Kuppers
¨
Max-Planck-Institut fur Plasmaphysik (EURATOM Association), 85748 Garching, Germany
͑Received 13 March 1995; accepted 23 October 1995͒
Benzene and ͑1,4͒-dimethyl-cyclohexane monolayers were physisorbed on graphite covered Pt͑111͒
surfaces. Exposure of benzene monolayers at 125 K to D atoms ͑1700 K͒ initially hydrogenates sp2
hybridized C atoms with a cross section of ca. 8 Å2 producing C6H6D intermediates. Additional D
atom reactions either transform this intermediate via a second hydrogenation reaction to
cyclohexadiene-d2 , C6H6D2, or restore benzene, C6H5D, via H abstraction. Once the aromaticity is
broken, successive hydrogenation of the diene occurs rapidly generating the saturated
cyclohexane-d6 , C6H6D6. The C6H5D reaction product can undergo further H/D exchange reactions
and, at any level of deuteration, the benzene species might get hydrogenated. Monolayers of the
saturated hydrocarbon ͑1,4͒-dimethyl-cyclohexane ͑DMCH͒ that are exposed to D atoms produce
deuterated DMCH via successive abstraction/hydrogenation reactions. Thermal desorption mass
spectra revealed that H atoms at the ring were exchanged with an apparent cross section of 1.7 Å2.
Methyl groups H atoms were exchanged much more slowly than ring H atoms. It was also observed
that D exposed molecules/radicals exhibit a tendency to desorb from the surface, which is ascribed
to the exothermicity of the reactions which lead to these species. © 1996 American Institute of
Physics. ͓S0021-9606͑96͒00605-9͔
I. INTRODUCTION
dressed study, was eliminated by covering the metal surface
with a graphite monolayer prior to adsorption of p-xylene. It
could be confirmed with TDS and HREELS methods that
this Pt͑111͒/C substrate provided an almost ideal physisorp-
tion substrate for the hydrocarbon molecule.
The interaction of H͑D͒ atoms with surfaces and adsor-
bates is a field of growing interest. Various examples have
shown that H atoms adsorb at metal surfaces which exhibit a
barrier towards hydrogen adsorption from molecular H2.1–4
H atoms impinging at Cl covered Au surfaces initiate HCl
formation5 with a subtle energy distribution of the products,
indicative of Eley–Rideal and Langmuir–Hinshelwood
branches in this atom/adsorbate reaction. H or D atoms di-
rected at hydrocarbons adsorbed on metal surfaces react with
these,6,7 although the role of the metallic support is unclear.
Recently, we have shown that thermal D atoms imping-
ing at the surfaces of ion beam deposited C:H films hydro-
genate unsaturated sp2 C centers to sp3 at low temperature,8
whereas at high temperatures H/D exchange due to thermally
induced decomposition of a radical intermediate was
observed.9 D impact induced abstraction of H from saturated
sp3 CHn groups has been identified as an Eley–Rideal type
process.10 Hydrogenation and H abstraction was also ob-
served at H covered diamond surfaces.11 The interplay be-
tween hydrogenation, cross section ␴Hϭ1.1 Å2, and H ab-
straction, cross section ␴Dϭ0.05 Å2, causes a temperature
dependent H impact induced erosion process of the C:H
films.12
Our study revealed that the interaction of D with phys-
isorbed p-xylene is governed by H abstraction and hydroge-
nation reactions, with the important observation that radical
intermediates generated by these reactions from molecular
species exhibit some stability, at least at the temperature em-
ployed in the study, 135 K. This observation is a simple
consequence from the fact that after sufficient atom exposure
to the target species reaction products were detected in ther-
mal desorption spectra which could only be formed via a
radical intermediate.
The Pt͑111͒/C/p-xylene13 study has also shown that the
probabilities of the H impact induced hydrogenation and H
abstraction reactions are affected by the stability of the inter-
mediates and products generated by abstraction and hydro-
genation, fully in line with the experience from the organic
chemistry of these molecules.14
The present work presents a study of the interaction of
thermal D and H atoms with physisorbed benzene and ͑1,4͒-
dimethyl-cyclohexane at Pt͑111͒/C. It will be shown that H
abstraction and hydrogenation, stability of radical intermedi-
ates, and the role of favorable molecular structures govern
this interaction.
In an attempt to investigate whether these reaction
mechanisms are also common to the interaction of thermal H
atoms with adsorbed hydrocarbons, we have studied the re-
actions which proceed at physisorbed p-xylene under impact
of D.13 It was instrumental for this investigation that the
influence of the metal substrate, a Pt͑111͒ surface in the ad-
II. EXPERIMENT
a͒
The experiments were performed in a UHV chamber
equipped with instrumentation for vibrational spectroscopy
¨
Also at Experimentalphysik VI, Universitat Bayreuth, 95440 Bayreuth,
Germany.
2392
J. Chem. Phys. 104 (6), 8 February 1996
0021-9606/96/104(6)/2392/9/$10.00
© 1996 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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

Lutterloh et al.: Interaction of D(H) atoms with benzene
2393
͑HREELS͒, thermal desorption spectroscopy ͑TDS͒, Auger
electron spectroscopy ͑AES͒ and low energy electron diffrac-
tion ͑LEED͒.
The substrates were clean Pt͑111͒ single crystal surfaces
and Pt͑111͒ surfaces covered with a monolayer of graphite
prepared by exposing Pt͑111͒ surfaces to ethane at 900 K
crystal temperature. LEED and Auger spectroscopy were
used to characterize the clean and monolayer graphite cov-
ered surfaces. The graphitization procedure of Pt͑111͒ sur-
face used here is well established and characterized in the
literature.15
Research grade benzene and ͑1,4͒-dimethyl-cyclohexane
͑DMCH͒ were filled into a gas handling line and cleaned by
several freeze cycles. The gases were admitted to the main
chamber by backfilling the chamber. The exposures given
below for benzene and DMCH in Langmuir units, 1 Lϭ10Ϫ6
Torr s, are calculated from uncorrected ion gauge readings.
Deuterium or hydrogen atoms were produced in a heated
Ta tube with an effusion hole ͑temperature 1700 K, pressure
about 1 Pa͒. The H and D atom fluences are given in the
present communication as monolayer ͑ML͒ equivalents rela-
tive to the graphite ͑0001͒ areal density, i.e.,
1 MLϭ3.8ϫ1015 atoms/cm2. The atom flux impinging the
target surface was calculated from the tube temperature and
in-tube D2 ͑H2͒ pressure. Typical atom fluxes were
1–2ϫ1013 atoms/s cm2. The method used to produce atoms
implies that they exhibit a Maxwellian velocity distribution,
which for the present experiments has its maximum at 0.15
eV kinetic energy. As thermal atom sources cannot work at a
dissociation efficiency of unity, it also implies that atom ex-
posure is inevitably connected with simultaneous exposure
of the respective molecules. These molecules, however, rep-
resent an unreactive background in the present experiments.
Thermochemical data suggest that even vibrationally excited
hydrogen molecules with 0.4 to 0.5 eV of vibrational energy
can undergo only significantly endothermic reactions with
the present reactant species.
FIG. 1. Benzene ͑amu 78͒ and hydrogen ͑amu 2͒ thermal desorption spectra
measured after adsorbing benzene at ͑a͒ Pt͑111͒ and ͑b͒ Pt͑111͒/C surfaces
at 125 K. Heating rate: 1 K/s.
clear prior to a TDS measurement which masses had to be
followed.
In order to test their stability and reactivity, H and D
atoms were directed at bare graphite monolayers on Pt͑111͒
surfaces at 120 to 300 K surface temperature. That hydroge-
nation reactions had occurred was confirmed with TDS, but
barely detectable in HREEL spectra. The fraction of hydro-
genated C atoms of the graphite monolayer was only a few
percent in the atom fluence range of significance here. Ap-
parently, only at steps and other imperfections of the C
monolayers hydrogenation reactions can proceed. It is there-
fore safe to assume that the C monolayer does not contribute
to the observed reactions of atoms with adlayer species and
the spectral changes induced by these reactions.
HREEL spectra were recorded in specular direction at an
electron impact energy of 7 eV. The resolution of the double
pass monochromator/analyzer assembly was set at 35 cmϪ1
.
TDS experiments were performed in the spectrum mode,
in which a preselected mass range, typically 50 to 100 amu
͑for benzene͒ or 90 to 130 amu ͑for DMCH͒, was scanned by
the quadrupole mass spectrometer simultaneously with the
application of the temperature ramp, and the resulting signals
were stored in a computer. This computer also controlled the
linear temperature ramp applied to the crystal. Given the
timing requirements of the quadrupole electronics/computer
data exchange, at a heating rate of 1 K/s one spectrum could
be collected every second. This data then allowed to con-
struct accumulated mass spectra by summing up a series of
spectra measured in a specific temperature range. Amu se-
lected thermal desorption spectra could also be extracted
from the original data set. The latter data were very useful
for discrimination between background and sample signal.
Alternatively, up to 9 masses could be preselected and mul-
tiplexed during the TDS runs. This method was used if it was
III. RESULTS
A. Benzene
Prior to the investigation of the reactions of D ͑H͒ atoms
with monolayers of benzene at Pt͑111͒/C it was studied how
monolayers of benzene had to be prepared and in which way
the substrate/adsorbate chemistry was affected by the pres-
ence of a sandwiched graphite monolayer. For this, clean
Pt͑111͒ surfaces and Pt͑111͒/C surfaces were exposed to ben-
zene at 125 K substrate temperature and the adsorption char-
acteristics were studied with TDS and HREELS. Figure 1
shows amu 78 ͑molecular benzene͒ and amu 2 TD spectra
for benzene exposures covering the monolayer and
multilayer range at Pt͑111͒. Three amu 78 desorption peaks
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

2394
Lutterloh et al.: Interaction of D(H) atoms with benzene
at 370, 165, and 142 K can be correlated with first layer
chemisorbed benzene, second layer physisorbed benzene and
third and following layers of condensed benzene. The peak
due to chemisorbed benzene is very broad and indicates a
complicated desorption behavior affected by simultaneous
decomposition of benzene. The adsorption/desorption char-
acteristics observed in the present study are similar to previ-
ous results16 and will be addressed in a separate paper.17
The TD spectrum recorded for amu 2 ͓upper panel, Fig.
1, spectrum ͑a͔͒ exhibits a low temperature feature identical
to that seen at amu 78 due to fragmentation of desorbed
benzene in the mass spectrometer. Additional amu 2 features
observed between 300 and 650 K are due to recombinative
desorption of H atoms. The temperature dependence of these
features is a function of the kinetics of benzene decomposi-
tion and the kinetics of hydrogen desorption. Accordingly,
Auger spectra measured after desorption of chemisorbed
benzene exhibit a carbon signal.
Benzene adsorbed at Pt͑111͒/C exhibits very different
desorption features. Submonolayer and monolayer desorp-
tion peaks are seen around 158 K, at lower temperature than
at Pt͑111͒. These signals occur right at the descending edge
of the multilayer signal which peaks at 142 K after 12 L
benzene exposure. As expected, in Fig. 1 corresponding amu
2 fragmentation signals are observed but decomposition sig-
nals are absent.
The measured desorption spectra at Pt͑111͒/C are in ac-
cordance with weak bonding between benzene and the
graphite monolayer. The observation that the monolayer sig-
nal almost overlaps with the multilayer peak indicates that
adsorption of benzene at the C layer is similar to benzene
adsorbed on a benzene layer, not too surprising in view of
the structural similarities of graphite and benzene. This and
the fact that benzene decomposition is eliminated indicates
that the Pt͑111͒ substrate is effectively shielded by the C
layer.
It is seen in Fig. 1 that in desorption spectra the transi-
tion from the monolayer to the multilayer benzene at
Pt͑111͒/C is smooth. Guided by the spectra shown in Fig. 1
the preparation of monolayers of benzene at Pt͑111͒/C was
achieved by exposing 5.4 L of benzene to the substrate at
125 K. After this procedure, no multilayer desorption signal
could be detected and the measured monolayer signals, as
shown by the respective dotted desorption trace in Fig. 1,
varied only by 5% between different experiments. The ben-
zene coverage of a benzene monolayer expressed with re-
spect to the C monolayer atom density is calculated as 0.07
͓0.18 with respect to the Pt͑111͒ surface atom density͔ from
the amu 78 signal. The calibration of this signal in terms of
coverage was performed with hydrogen desorption spectra
measured at the saturation coverage of hydrogen at Pt͑111͒
surfaces18 and the mass spectrometer sensitivity factors for
hydrogen and benzene, respectively. The benzene/C/Pt
monolayer coverage obtained here is in good agreement with
the value determined for benzene adsorbed at the basal plane
of graphite.19
FIG. 2. Family of HREEL spectra measured at adsorbed and reacted ben-
zene at 125 K. ͑a͒ chemisorbed benzene monolayer at Pt͑111͒ after 3.5 L
exposure, ͑b͒ benzene multilayers at Pt͑111͒ after 250 L, ͑c͒ benzene mono-
layer at Pt͑111͒/C after 5.4 L, ͑d͒ after exposure of 1.3 ML H atoms to a
benzene monolayer at Pt͑111͒/C, ͑e͒ after exposure of 1.3 ML D atoms to a
benzene monolayer at Pt͑111͒/C.
played measured at benzene monolayer and multilayers at
Pt͑111͒, and at a benzene monolayer at Pt͑111͒/C prepared as
noted above. The Pt͑111͒ monolayer spectrum shows a
strong out-of-plane CH bending mode at 830 cmϪ1 whereas
the multilayer and Pt͑111͒/C monolayer spectra exhibit this
mode at 685 cmϪ1. At increased sensitivity the benzene/Pt
monolayer and multilayer spectra show in addition to that
vibrational mode signals from excitation of other vibrations,
even small signals from the sp2 CH stretch mode at 2990
and 3050 cmϪ1, respectively. The Pt͑111͒/benzene mono-
layer spectrum agrees very well with published data.20
A sp2 CH stretch mode is virtually absent for benzene
adsorbed at Pt͑111͒/C surfaces. The vibrational excitation se-
lection rules operating in HREELS21 require that benzene
lies flat in that case. It is remarkable that within the precision
of our experiment the CH out-of-plane mode frequencies of
benzene in multilayers and a monolayer at Pt͑111͒/C are the
same and very close to the gas phase value of 673 cmϪ1 22
.
This corroborates the statement made above on the bonding
of benzene at this surface and supports the view that a graph-
ite monolayer supplies an almost ideal substrate for phys-
isorptive bonding.
Monolayers of benzene at Pt͑111͒/C were prepared in
the above described manner and exposed to D and H atom
fluxes at 125 K substrate temperature. Figure 2 includes
HREEL spectra measured after 1.3 ML D or H atom expo-
These TD results are fully supported by the HREEL
spectra shown in Fig. 2. Here, vibrational spectra are dis-
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

Lutterloh et al.: Interaction of D(H) atoms with benzene
2395
FIG. 3. Mass scan TD spectra measured after exposing D atoms to a ben-
zene monolayer at Pt͑111͒/C at 125 K. Heating rate: 1 K/s. ͑a͒ Unexposed
benzene, ͑b͒ 0.45 ML D, ͑c͒ 1.35 ML D. For easier identification the reac-
tion paths are indicated. Mϩ indicates the peak of the parent molecule ion.
FIG. 4. Amu selected thermal desorption spectra extracted from mass scan
TD spectra measured after exposing a benzene monolayer at Pt͑111͒/C to
0.9 ML D atoms at 125 K.
of peaks above amu 79, up to amu 94 after a 1.35 ML D
atom dose. As already suggested from the HREEL spectra,
these peaks give evidence for reactions between benzene and
impinging D.
sure. It is seen that the vibrational spectrum is markedly
changed as compared to a unexposed benzene layer. The
occurrence of CD bending and stretching modes clearly in-
dicate C–D bonding in reaction products after D exposure.
These products are not simply deuterated ͑in the sense of
H/D exchanged͒ benzene, as from the positions of the CH
stretch band at 2910 cmϪ1 and CD stretch band at 2150 cmϪ1
it is clear that after reaction of D͑H͒ with benzene H and D
are bound to sp3 hybridized C atoms. Apparently, these at-
oms undergo addition to the ring thereby transforming the
sp2 C center at which they have been added to sp3. The
occurrence of bending modes of sp3 CHD groups in the
range 500 to 1500 cmϪ1 suggests that the reaction product
does not lay flat at the Pt͑111͒/C surface. We note for com-
pleteness, that the same effects have been observed in
HREELs spectra of D atom exposed benzene multilayers.
To identify the reactions in more detail, desorption spec-
tra were measured after specific D atom fluences had been
directed to benzene monolayers. In Fig. 3 spectra are dis-
played which illustrate the reactions induced by D exposure.
Here, the sums of about 180 individual mass spectra accu-
mulated between 120 and 300 K upon desorption of unex-
posed and D exposed benzene monolayers are shown. As
expected, the spectral information of the unexposed mono-
layer contains the amu 78 signal from 12C benzene, fragmen-
tation signals at smaller masses, and the amu 79 component
from the 13C contribution to benzene. In contrast to this, the
spectral signals of D exposed monolayers exhibit a sequence
Further data analysis was performed based on amu se-
lected TD spectra extracted from the series of individual
mass scans the sum of which is shown in Fig. 3. Figure 4
gives examples for 78, 79, and amu 90 TD spectra extracted
from a measurement done after 0.9 ML D atom exposure.
The amu 78 spectrum is that of the unreacted fraction of
benzene. The amu 79 spectrum is composed of the 13C con-
tribution of this signal, shown as a dashed line, plus a con-
tribution from a reaction product, obviously C6H5D. This
reaction product must exhibit the same adsorption energy at
the Pt/C surface as C6H6 benzene and should therefore des-
orb with the same kinetics, as observed. In contrast to this,
the TD spectrum of amu 90 exhibits a different peak tem-
perature which excludes a H/D exchanged benzene as the
desorbing species. As the desorption peak temperature has
shifted downwards by 5 K with respect to the benzene peak,
this species should be less aromatic than benzene. The amu
value suggests fully hydrogenated ͑in the sense of D addi-
tion͒ benzene, i.e., cyclohexane C6H6D6, as this species. In
line with this assignment, mass spectrometer fragmentation
products of cyclohexane were detected in mass scan TD
spectra, each with an individual TD spectrum which reflects
that cyclohexane is the parent molecule. In separate experi-
ments with adsorbed cyclohexane it was confirmed that its
desorption temperature is 163 K.
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

2396
Lutterloh et al.: Interaction of D(H) atoms with benzene
C6H6D, in which one ring sp2 CH group is converted to sp3
CHD. In the first pathway a further D impact induces H
abstraction from this latter group accompanied by formation
of HD which escapes into the gas phase. This leads to H/D
exchanged benzenes like C6H5D, C6H4D2, etc. In the second
pathway the C6H6D radical intermediate formed in the initi-
ating step undergoes deuteration ͑in the sense of D addition͒
via cyclohexadiene-d2 ͑amu 82͒ and cyclohexane-d4 ͑amu
86͒ to cyclohexane-d6 ͑amu 90͒ with the appropriate radical
intermediates. As expected, products with amu values above
90 indicate that H/D exchanged benzenes generated by the
first pathway can undergo hydrogenation by D as well, i.e.,
switch to the second pathway. In Fig. 10͑a͒ the reaction steps
are collected. If H atoms are used for the reaction, the ab-
straction pathway specified above should lead back to C6H6
benzene which cannot be distinguished in TD spectra from
unreacted species. The H addition pathway was isolated us-
ing thermal H atoms. Analysis of the data from these experi-
ments in conjunction with data from D atom experiments
discussed previously yields cross sections for both abstrac-
tion from and addition to the C6H6H ͑D͒ intermediate.
In order to check whether D atoms can abstract H from a
saturated ring system, the above experiments were performed
at monolayers of ͑1,4͒-dimethyl-cyclohexane. This molecule
was chosen instead of cyclohexane as it allowed to investi-
gate the reactivity at ring and methyl C atoms towards
atomic D in one experiment.
FIG. 5. Normalized reaction product distributions measured at D exposed
benzene monolayers at Pt͑111͒/C.
B. (1,4)-Dimethyl-cyclohexane
It is seen in Fig. 3 that there are mass signals below amu
78, which could stem from other products than those men-
tioned, e.g., alkanes produced by ring-breaking reactions.
The analysis, however, revealed that these signals in the in-
tegrated spectra are either mass spectrometer fragmentation
products of heavier molecules or do not exhibit particular
desorption peaks and are due to unidentified background spe-
cies.
Figure 5 demonstrates the progress of exchange and deu-
teration reactions as a function of D atom exposure from 0 to
1.35 ML. Here, contributions to the mass integrated TDS by
identified molecular species are represented by bars ͓e.g., the
bar in the top panel of Fig. 5 represents the amu 78 peak
from Fig. 3͑a͔͒. The sums of the bar heights were normalized
to the benzene signal obtained from an unreacted benzene
monolayer ͑i.e., the bar height in the upper panel͒. These
normalized signals were used instead of the measured raw
data as it was observed that after D exposure the number of
molecules collected in thermal desorption spectra had de-
creased. This loss of particles increased linearly with D atom
fluence. As an example, after 1.35 ML D exposure, only
about 10% of the molecules initially adsorbed as benzene
could be detected as desorption species, benzene or reaction
products. The origin of this phenomenon will be addressed in
the discussion.
For experiments with ͑1,4͒-dimethyl-cyclohexane
͑DMCH͒, at first the proper monolayer preparation technique
had to be established, and it had to be assured, that the
Pt͑111͒/C substrate acts as a suitable unreactive carrier of the
molecule, as was demonstrated for benzene.
Figure 6 displays thermal desorption spectra measured
after adsorption of multilayers of DMCH at Pt͑111͒ and
Pt͑111͒/C. On Pt͑111͒ surfaces DMCH ͑amu 112͒ exhibits
complicated desorption features, very similar to those previ-
ously reported for the Pt͑111͒/cyclohexane system.23 In ad-
dition to a multilayer peak at 154 K and a double-peaked
monolayer structure at 245/260 K, a further sharp peak is
observed at 192 K, which might stem from step sites in the
monolayer which offer an increased adsorption energy in the
second layer. The corresponding amu 2 desorption spectrum
shows in addition to the mass spectrometer fragmentation
signal hydrogen desorption above 280 K from decomposition
of products left by decomposition of DMCH. At Pt͑111͒/C
the situation is much less complicated. The peak at 200 K
signals monolayer desorption and a multilayer/second layer
peak is seen at 154 K. The respective amu 2 spectrum exhib-
its the expected fragmentation signals but no hydrogen de-
sorption from decomposition reactions. This suggests that
DMCH in the monolayer at Pt͑111͒/C is less strongly bound
than at Pt͑111͒ and that the sandwich C monolayer elimi-
nates decomposition of DMCH molecules. The above con-
clusion on the action of the C layer for the molecule/
substrate interaction therefore is confirmed with DMCH.
From the progression of mass distributions shown in Fig.
5 two reaction pathways can be identified which are indi-
cated by differently shaded bars. The initial step for both
pathways is D addition resulting in a radical intermediate,
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

Lutterloh et al.: Interaction of D(H) atoms with benzene
2397
FIG. 6. ͑1,4͒-dimethyl-cyclohexane ͑DMCH, amu 112͒ and hydrogen ͑amu
2͒ thermal desorption spectra measured after 10 L DMCH exposures at
Pt͑111͒ and Pt͑111͒/C surfaces.
FIG. 7. HREEL spectra of unreacted and reacted DMCH monolayers at
Pt͑111͒/C at 135 K. The inset shows the CH stretch region on an expanded
scale.
the signal at amu 112 of the parent molecule and the leading
fragment at amu 97 which occurs from methyl split-off upon
ionization. Interaction with D reveals parent peaks at amu
112ϩx accompanied by the respective 112ϩxϪ15 frag-
ments. This observation corroborates the conclusion drawn
from the HREEL spectra that H/D exchange proceeds at the
ring. Completed H/D exchange at the ring leads to a amu 122
product. The spectrum measured after 5.9 ML of D had been
exposed to adsorbed DMCH exhibits products which indi-
cate that finally even the methyl H atoms get exchanged. As
observed in the experiments with benzene, D atom exposure
leads to loss of adsorbed species. Consequently, products
from D atom exchange at the methyl group can only be
observed at mass spectrometer sensitivities at which also
contributions to the spectra from background species are
present ͑e.g., signals below amu 97 in Fig. 8͒.
Results from the analysis of mass spectra measured upon
reaction of DMCH with D are shown in Fig. 9. As with D
exposed benzene the observed loss of particles was signifi-
cant, about half of the particles were lost within the first
monolayer of D exposure. Therefore, in Fig. 9 relative yields
are plotted rather than absolute yields. A comparison of the D
exposures noted in this bar graph with those shown in Fig. 5
suggests a less pronounced reactivity of DMCH as compared
with benzene.
Physisorbed DMCH monolayers could be prepared routinely
by exposing 10 L of DMCH to Pt͑111͒/C followed by a flash
to 144 K. A sample of corresponding monolayer desorption
spectra is displayed in Fig. 6 as a dotted line. The coverage
of a DMCH monolayer is estimated as 0.13 ͓relative to the
Pt͑111͒ atom density͔.
HREEL spectra recorded at a monolayer of DMCH at
Pt͑111͒/C prior and subsequent to exposure to D atoms are
shown in Fig. 7. The unexposed monolayer spectrum exhib-
its a CH stretch band around 2900 cmϪ1 from sp3 CH2 and
CH3 groups at the ring and the methyl groups, respectively.
In addition there are peaks due to bending modes in the
fingerprint frequency region below 1500 cmϪ1. D atom ex-
posure to DMCH monolayers causes the occurrence of a sp3
CD stretch mode peak at 2150 cmϪ1 and a decrease of the
CH band intensity, which is correlated with a shift of the
band position towards the sp3 CH3 frequency value of 2940
cmϪ1, emphasized in the inset of Fig. 7. The presence of a
CD band clearly indicates that DMCH had reacted with D in
H/D exchange reactions. The shift of the CH band peak sug-
gests that H/D exchange had occurred at the ring CH2 groups
and not at the methyl groups. As was observed for benzene
multilayers, HREEL spectra for D exposed DMCH multilay-
ers exhibit the same features as the D exposed DMCH mono-
layers.
IV. DISCUSSION
The results of TDS experiments at various stages of D
atom exposure to adsorbed DMCH are displayed in Fig. 8.
The integrated scan spectrum of unreacted DMCH exhibits
In full analogy to the investigation of D atom induced
reactions at adsorbed p-xylene,13 the present study with ben-
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

2398
Lutterloh et al.: Interaction of D(H) atoms with benzene
zene and DMCH shows that the ongoing reactions can be
classified in the categories H abstraction via HD formation
and hydrogenation ͑D addition͒. The reaction schemes dis-
played in Fig. 10 illustrate the processes identified and serves
as a base for further discussion.
It is evident that all observed reactions are of the Eley–
Rideal type, i.e., the reactions proceed upon impact of the D
atoms at an adsorbed target molecule. Adsorbed D is not
involved as it is not feasible that D adsorbs either at the C
monolayer or at the target molecules.
All but one reaction steps included in Fig. 10 are exo-
thermic. Consider first the reaction numbered as 1 in Fig.
10͑a͒. The heat of reaction can be calculated from the corre-
sponding heats of formation of the radicals H and C6H7 and
the C6H6 molecule24 as 103 kJ/mol. In contrast to this, reac-
tion ͑2͒, H abstraction by D from benzene, is endothermic by
28 kJ/mol as the C6H5–H dissociation energy of 464
kJ/mol24 has to be compared with the HD dissociation en-
ergy, 436 kJ/mol. Accordingly, in gas phase studies on
H/benzene reactions this abstraction was negligible and oc-
curs only under harsh conditions.25 For the present case we
therefore consider reaction ͑2͒ as not contributing. Reaction
͑3͒, the hydrogenation by D of a radical center, e.g., adjacent
to the CHD group, is very exothermic. The gain of C–D
binding energy is given by the full C6H6D–D dissociation
energy, ca. 305 kJ/mol.24 Similar energy gains are obtained
by hydrogenation ͑by D͒ of the other radical species in Fig.
10͑a͒.
FIG. 8. Mass scan TD spectra measured after exposing D atoms to a DMCH
monolayer at Pt͑111͒/C at 135 K. Heating rate: 1 K/s. ͑a͒ Unexposed
DMCH, ͑b͒ 0.9 ML D, ͑c͒ 3.7 ML D, ͑d͒ 5.9 ML D. For easier reading the
fragmentation correlations are indicated.
Competitive to reaction ͑3͒ the restoration of benzene
through reaction ͑4͒ is observed in the present study. The
energy balance of this reaction is determined by the gain of
HD dissociation energy, 436 kJ/mol, minus the C6H5D–H
dissociation energy, 103 kJ/mol. Overall an energetically fa-
vorable reaction. An alternate pathway back to C6H5D ben-
zene would be the split-off of an H atom, which is not fea-
sible at the given substrate temperatures. At elevated
temperatures H split-off could occur via activation.26
Checking the numbers for all reactions in Fig. 10͑a͒ it is
confirmed that with the exclusion of H abstraction from ben-
zene all reactions are considerably exothermic.
The present study cannot supply any information on the
activation energies of the observed reactions. It is known
from gas phase studies that the H/benzene and H/DMCH
reactions are activated. The H abstraction from benzene,
HϩC6H6→C6H5ϩH2, has a barrier of ca. 65 kJ/mol,25 in line
with the endothermicity mentioned above. The H addition to
C6H6 has a barrier of only ca. 18 kJ/mol,14 expected from its
exothermicity. It is clear that the kinetic energy supplied by
the high energy fraction of the 1700 K D atoms suffices to
overcome this barrier. From the fact that the present reac-
tions were performed at 125 K target molecule temperatures
nothing can be concluded, of course.
The cross sections or reaction probabilities for the reac-
tions in Fig. 10 can be roughly estimated from the present
experiments as the kinetic data were obtained under pseudo-
first-order conditions, i.e., constant H flux. Accordingly, the
appropriate rate law can be formulated as
FIG. 9. Normalized reaction product distributions measured at D exposed
DMCH monolayers at Pt͑111͒/C.
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

Lutterloh et al.: Interaction of D(H) atoms with benzene
2399
cross section as slope. Due to the above mentioned loss of
particles raw data were used to get these numbers. This pro-
cedure implies that particles which were lost had reacted and
are accounted for in the calculated cross section. The D re-
action cross section obtained in this way are 8 Å2 for ben-
zene and 1.7 Å2 for DMCH. The H/benzene cross section
was determined at 4 Å2, but one has to keep in mind that in
the H/benzene reactions the H/D exchange reaction is not
included. These numbers are in the range deduced by Bent
et al. from his experiments on interaction of H with hydro-
6,7
carbons adsorbed at Cu͑111͒ and corroborate the above
statement that the reactions are of the Eley–Rideal type. It
has to be observed, however, that the present experiments
deal with energy distributed atoms and as the cross sections
are expected to depend on the atom kinetic energy the given
numbers are averages over the energies determined by the
Maxwellian distribution delivered by our 1700 K source.
Ignoring this complication, the progression of amu val-
ues in Fig. 5 and Fig. 9 suggest the following characteristics.
The bottleneck in the H͑D͒/benzene reaction is the initiating
H͑D͒ addition to the ring. Impact of a further atom with
equal probabilities produces either the diene via addition or
restores benzene via abstraction. These observations empha-
size the strong tendency of the system to keep its aromatic
character. If the aromaticity is broken, further addition reac-
tions are fast, in particular those which occur at radical spe-
cies. It might be of importance here to consider that the
activation barrier for H͑D͒ addition to nonaromatic sp2 CH
groups, e.g., cyclohexadiene: 6 kJ/mol,27 cyclohexene: ca. 10
kJ/mol,27 or ethylene: ca. 9 kJ/mol28 is considerably lower as
compared to 18 kJ/mol for benzene, with the consequence
that more atoms from the thermal sources carry enough en-
ergy to add to the diene. Abstraction reactions at sp3 CH
groups of cyclohexane are less important due to the high
activation barrier of ca. 39 kJ/mol.29 The progression in Fig.
5 can be modeled numerically with sufficient agreement
without considering their contributions.
The progression of products from reactions between
DMCH and D atoms as shown in Fig. 9 and summarized in
Fig. 10͑b͒ suggests that abstraction of H from the ring carbon
atoms is much more probable than from the methyl groups,
in line with the lower C–H bond energy of secondary CH2
groups in comparison with primary CH3 groups.24 The pref-
erential H abstraction at secondary CH2 groups was also ob-
served for propane.30 That abstraction from a methyl group is
not necessarily improbable is, however, clear from our
p-xylol study.13
It has to be observed, that even in the present study, as in
our previous work on p-xylene, radical intermediates are
stable at the Pt͑111͒/C surfaces. It is clear that the special
structure of the present substrate is essential for this stability.
An ideal graphite monolayer does not offer any sites for
bonding of a radical or its dissociation species, only at im-
perfections this is feasible. This makes it also clear that the
present results would have hardly been obtained at a reactive
metal surface, as this offers plenty of sites for bonding of
decomposition products of radical species.
FIG. 10. Detailed map of reaction sequences identified upon reactions of
thermal D atoms with adsorbed benzene ͑a͒ and DMCH ͑b͒.
Ϫdct
* *
ϭct ␴ ⌽
dt
with ct as concentration of the target molecules, ␴ as cross
section, and ⌽ as flux of D atoms.
A logarithmic plot of the decrease of the relative target
molecule concentration against atom fluence, as obtained
from Fig. 5 and Fig. 9, should yield a straight line with the
Finally, we address the observed loss of particles. The
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14

2400
Lutterloh et al.: Interaction of D(H) atoms with benzene
same phenomenon has been observed by Bent et al.6,7 There
are two possibilities to consider. First the above mentioned
imperfections at the graphite monolayer which could act as
sinks of radicals. We do not think that this is of relevance in
the present study as in that case at bigger atom exposures
desorption spectra would exhibit increasing amounts of
background species originating from these trapped radicals,
which was not observed. Second, because many of the reac-
tion steps depicted in Fig. 10 are extremely exothermic it is
possible that dissipation of this energy via a desorption chan-
nel is competitive with dissipation into the substrate. This
effect would lead to a continuous loss of particles, as ob-
served. One expects that the ability to release excess reaction
energy is connected with the adsorption energy of the respec-
tive species. From the small adsorption energy of the present
hydrocarbon molecules at the graphite monolayers this
mechanism is feasible to operate.
5 C. T. Rettner and D. J. Auerbach, Science 263, 365 ͑1994͒; C. T. Rettner,
J. Chem. Phys. 101, 1529 ͑1994͒.
6 M. Xi and B. E. Bent, J. Phys. Chem. 97, 4167 ͑1993͒.
7 M. Xi and B. E. Bent, J. Vac. Sci. Technol B 10, 2440 ͑1992͒.
8 J. Biener, U. A. Schubert, A. Schenk, B. Winter, C. Lutterloh, and J.
¨
Kuppers, J. Chem. Phys. 99, 3125 ͑1993͒.
9
¨
A. Horn, J. Biener, A. Schenk, C. Lutterloh, and J. Kuppers, Surf. Sci.
331–333, 178 ͑1995͒.
C. Lutterloh, A. Schenk, J. Biener, B. Winter, and J. Kuppers, Surf. Sci.
10
¨
316, L1039 ͑1994͒.
11 B. D. Thoms, J. N. Russell, Jr., P. E. Pehrsson, and J. E. Butler, J. Chem.
Phys. 100, 8425 ͑1994͒.
12 A. Horn, A. Schenk, J. Biener, B. Winter, C. Lutterloh, M. Wittmann, and
¨
J. Kuppers, Chem. Phys. Lett. 231, 193, ͑1994͒.
13
¨
C. Lutterloh, J. Biener, A. Schenk, and J. Kuppers, Surf. Sci. 331–333,
261 ͑1995͒.
14 D. L. Baulch, C. J. Cobos, R. A. Cox, C. Esser, P. Frank, Th. Just, J. A.
Kerr, M. J. Pilling, J. Troe, R. W. Walker, and J. Warnatz, J. Phys. Chem.
Ref. Data 21, 411 ͑1992͒; D. Robaugh and W. Tsang, J. Phys. Chem. 90,
4159 ͑1986͒; H. S. Johnston and C. Parr, J. Am. Chem. Soc. 85, 2544
͑1963͒.
15 Hu. Zi-pu, D. F. Ogletree, M. A. van Hove, and G. A. Somorjai, Surf. Sci.
180, 433 ͑1987͒; T. A. Land, T. Michely, R. J. Behm, J. C. Hemminger,
and G. Comsa, Surf. Sci. 264, 261 ͑1992͒.
V. SUMMARY AND CONCLUSIONS
16 M. Abon, J. C. Bertolini, J. Billy, J. Massardier, and B. Tardy, Surf. Sci.
The interaction of thermal D atoms with physisorbed
benzene and ͑1,4͒-dimethyl-cyclohexane proceed via the
Eley–Rideal type reactions D addition and H abstraction.
The initial step in D/benzene interaction is D addition to a
sp2 CH center. The radical intermediate exhibit stability and
is either transferred back to singly deuterated benzene
C6H5D via H abstraction or to cyclohexadiene-d2 C6H6D2
via addition of a second D atom. Repetitive application of
these reactions then leads to multiply deuterated cyclohex-
162, 395 ͑1985͒.
17
¨
C. Lutterloh, J. Biener, A. Schenk and J. Kuppers, Surf. Sci. ͑in press͒.
18 K. Christmann, G. Ertl, and T. Pignet, Surf. Sci. 54, 365 ͑1976͒.
19 U. Bardi, S. Magnanelli, and G. Rovida, Langmuir 3, 159 ͑1987͒.
20 S. Lehwald, H. Ibach, and J. E. Demuth, Surf. Sci. 78, 577 ͑1978͒.
21 H. Ibach and D. L. Mills, Electron Energy Loss Spectroscopy and Surface
Vibrations ͑Academic, New York, 1982͒.
22 T. Shimanouchi, Tables of Molecular Vibrational Frequencies, Consoli-
dated Volume I, NSRDS-NBS 39, p. 152.
23 J. A. Rodriguez and C. T. Campbell, J. Phys. Chem. 93, 826 ͑1989͒.
24 CRC Handbook of Chemistry and Physics, 74th ed, edited by D. R. Lide
͑CRC, Boca Raton, 1993͒.
ane.
D
atoms impinging at physisorbed dimethyl-
cyclohexane induce overall H/D exchange at the ring C at-
oms via sequential H-abstraction/D-addition reactions. H/D
exchange at the methyl groups is also observed, but less
probable.
25 V. S. Rao and G. B. Skinner, J. Phys. Chem. 88, 5990 ͑1984͒; J. H. Kiefer,
L. J. Mizerka, M. R. Patel, and H.-C. Wei, J. Phys. Chem. 89, 2013
͑1985͒.
26 J. M. Nicovich and A. R. Ravishankara, J. Phys. Chem. 88, 2534 ͑1984͒.
27 K. Hoyermann, A. W. Preuss, and H. Gg. Wagner, Ber. Bunsenges. Phys.
Chem. 79, 156 ͑1975͒.
1 A. Winkler, G. Pozgainer, and K. D. Rendulic, Surf. Sci. 251/252, 886
͑1991͒.
28 G. W. Johnston, S. Satyapal, R. Bersohn, and B. Katz, J. Chem. Phys. 92,
206 ͑1990͒; P. D. Lightfoot and M. J. Pilling, J. Phys. Chem. 91, 3373
͑1987͒.
2 P. Sandl, U. Bischler, and E. Bertel, Surf. Sci. 291, 29 ͑1993͒.
3 K. B. Ray, J. B. Hannon, and E. W. Plummer, Chem. Phys. Lett. 171, 469
29 D. Grief and G. A. Oldershaw, J. Chem. Soc. Faraday Trans. 1, 78, 1189
͑1982͒.
¨
͑1990͒; V. Lossev and J. Kuppers, Surf. Sci. 284, 175 ͑1993͒; K. B. Ray,
X. Pan, and E. W. Plummer, ibid. 285, 66 ͑1993͒.
4 P. T. Sprunger and E. W. Plummer, Chem. Phys. Lett. 187, 559 ͑1991͒.
30 F. J. Adrian, J. Bohandy, and B. F. Kim, J. Chem. Phys. 100, 8010 ͑1994͒.
J. Chem. Phys., Vol. 104, No. 6, 8 February 1996
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:
134.176.129.147 On: Thu, 18 Dec 2014 13:35:14