Dehalogenation of 1,1,2-trichloro-1-fluoroethane over ?±-Cr2O3 (101ì?12)

Article abstract of DOI:10.1021/jp022172j

The reaction of CFCl₂CH₂Cl over the stoichiometric Cr₂O₃ (1012) surface yields CFCl=CH₂, HCa??CH, and surface halogen. The 1,2-dihalo-elimination reaction to CFCl=CH₂ is initiated via C-Cl bond cleavage at the CFCl₂-end of the molecule to give a -CFClCH₂Cl haloalkyl surface intermediate. A rate-limiting ?2-chlorine elimination from the surface alkyl gives rise to the CFCl=CH₂ product. Acetylene is formed by the subsequent reaction of CFCl=CH₂ in series. The chlorine liberated from CFCl₂CH₂Cl binds at the five-coordinate surface Cr³⁺ sites on the stoichiometric surface and shuts down the dehalogenation chemistry by site blocking. No carbon buildup is observed on deactivated surfaces, and no evidence is seen for the replacement of surface lattice oxygen by halogen under the conditions of this study. At elevated temperatures, the thermal removal of surface chlorine is observed, and it is attributed to migration into the sample bulk.

Full text of DOI:10.1021/jp022172j

5182
J. Phys. Chem. B 2003, 107, 5182-5189
Dehalogenation of 1,1,2-Trichloro-1-fluoroethane over r-Cr2O3 (101h2)
Steven C. York and David F. Cox*
Department of Chemical Engineering, Virginia Polytechnic Institute & State UniVersity,
Blacksburg, Virginia 24061
ReceiVed: October 1, 2002; In Final Form: March 31, 2003
The reaction of CFCl2CH2Cl over the stoichiometric Cr2O3 (101h2) surface yields CFCldCH2, HCtCH, and
surface halogen. The 1,2-dihalo-elimination reaction to CFCldCH2 is initiated via C-Cl bond cleavage at
the CFCl2-end of the molecule to give a -CFClCH2Cl haloalkyl surface intermediate. A rate-limiting â-chlorine
elimination from the surface alkyl gives rise to the CFCldCH2 product. Acetylene is formed by the subsequent
reaction of CFCldCH2 in series. The chlorine liberated from CFCl2CH2Cl binds at the five-coordinate surface
Cr3+ sites on the stoichiometric surface and shuts down the dehalogenation chemistry by site blocking. No
carbon buildup is observed on deactivated surfaces, and no evidence is seen for the replacement of surface
lattice oxygen by halogen under the conditions of this study. At elevated temperatures, the thermal removal
of surface chlorine is observed, and it is attributed to migration into the sample bulk.
TABLE 1: Halocarbon Reactions Reported To Occur over
Cr2O3
I. Introduction
The reactions of halogen-substituted alkanes with chromia
have been much studied because of their relevance to the vapor-
phase manufacture of hydrofluorocarbon compounds (environ-
mentally friendly refrigerants with low ozone-depleting poten-
tial) and because of interest in the destruction and conversion
of chlorine-containing hydrochlorofluorocarbon (HCFC) ana-
logues.1-3 Typically, these reactions are conducted in the
presence of HF where fluorination (for example, fluorine-for-
chlorine exchange) is an expected result. The halocarbon
reactions most commonly reported over Cr2O3 are summarized
in Table 1.
Observed Reactions
Reaction Conditions
HF addition to CdC double bond Dominant reaction in the
presence of HF
Halogen exchange
Replacement of Cl by F is usually
the desired reaction and is favored
in the presence of HF.
Dehydrohalogenation
(HF and HCl elimination)
Disproportionation
(dismutation)
Reaction may occur in the presence
and absence of HF.
Presence of HF not required.
Isomerization
Presence of HF not required.
Rowley, Webb, Winfield, and co-workers4-7 have studied the
interaction of halogenated compounds with Cr2O3 microcrys-
talline powders using radiolabeled HF, HCl, and halocarbon
compounds. They unambiguously demonstrated the uptake of
halogen by Cr2O3 during the initial “activation” phase of
reaction, and found that 36Cl and 18F could be deposited on both
untreated and pre-halogenated Cr2O3 catalysts by exposure to
HX (X ) halogen) or halocarbon compounds at 623 K. It was
also reported that Cl and F are capable of replacing one another
on the surface, and that halogen exchange occurs between the
surface and various CFC compounds.4,6,7 Direct halogen ex-
change with surface (HF)n-HX oligomers was also proposed.4,6
The resulting halogen exchange sequences yield a mix of
halocarbon compounds over both untreated and pretreated
(halogenated) Cr2O3 catalyst.
In studies of the trichlorotrifluoroethane/Cr2O3 system, Blan-
chard and co-workers focused on the disproportionation and
isomerization of CFCl2CF2Cl to CCl3CF3 over Cr2O3 powder,
supported Cr2O3, and AlF3.8-10 They found that disproportion-
ation and isomerization reactions dominate over Cr2O3 in the
absence of HF. They also observed that pretreating the surface
with pyridine, a Lewis base, inhibits surface activity, indicating
that Lewis acid sites (cation centers) are important to the reaction
mechanism.8 The presence of small amounts of the unsaturated
compounds CCl2dCCl2 and CFCldCCl2 were also reported in
the product mix.
Later work from Blanchard and co-workers focused a broader
spectrum of halocarbon surface reactions including dispropor-
tionation, isomerization, dehalogenation, dehydrohalogenation,
and halogen exchange.11 They reported that halogen exchange
dominates in the presence of HF, but that disproportionation,
isomerization, and dehydrochlorination occurs without HF. It
was suggested that fluorination, in the presence of HF, was
governed by the acidity of the catalyst,11 and that the catalytic
activity for the fluorination of CF3CH2Cl to CF3CH2F is a
function of the number of reversibly oxidizable chromium sites
on the catalyst surface.12,13 XPS measurements of fluorine uptake
by the catalyst following exposures to HF and CF3CH2Cl yielded
a maximum F/Cr ratio of 0.36, regardless of the treatment gas
used.13
Other authors have postulated that the oxidation state of
surface ions plays a crucial role in determining activity. On the
basis of product distributions and thermal desorption data,
Coulson et al.14 propose that the reduction of high-valent
chromium ions to Cr3+ is necessary to create active sites for
the disproportionation of CHF2Cl over powdered Cr2O3 micro-
crystalline catalyst. They argue that the active site is a
coordinately unsaturated Cr-X center, which acts as a strong
Lewis acid. Following exposures at 773 K, Coulson and co-
workers14 report XPS atomic percentages that correspond to F/Cr
ratios of 0.22 and 0.30 for CHF3 and HF, respectively.
Kohne and Kemnitz have studied the reactions of several
halocarbons with chromia, and provided evidence that both
direct halogen exchange and HX abstraction/addition reactions
* Corresponding author. Fax: (540) 231-5022. E-mail: dfcox@vt.edu.
10.1021/jp022172j CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/13/2003

Dehalogenation of 1,1,2-Trichloro-1-fluoroethane
J. Phys. Chem. B, Vol. 107, No. 22, 2003 5183
occur over fluorinated Cr2O3 catalyst in the presence of HF at
573 K.15 Significant amounts of unsaturated haloalkenes were
also found in the product mix, even in the presence of HF.
Kohne and Kemnitz15 argue that isomerization and dispropor-
tionation reactions may be disregarded, with or without HF
present in the reaction mix, and that the product mix observed
is the result of a sequence of reactions: (1) dehydrohalogenation
(HX abstraction) of haloalkanes to olefins, (2) hydrohalogenation
(Markovnikov HX addition) of olefins to haloalkanes, and (3)
direct halogen exchange between haloalkanes and surface X.
Kemnitz and Niedersen16 have also studied the isomerization
reaction of CHF2CHF2 to CF3CH2F (HCFC-134a) over fluori-
nated Cr2O3, without HF. They propose that isomerization
reactions occur in two steps. From deuterium labeling experi-
ments, it was concluded that if chlorine could be excluded from
the reacting system, then the isomerization reaction proceeds
through HF elimination and subsequent HF addition by the
Markovnikov rule. The presence of HCl or any chlorine-
containing halocarbons was found to give rise to a “complex
set of side reactions”.16
Niedersen et al.17 propose that Lewis acid sites are responsible
for the CHF2CHF2 to CF3CH2F isomerization activity of Cr2O3.
They used IR and UV-Vis spectroscopy of adsorbed CO and
NH3 to argue that Lewis acid sites are responsible for the
isomerization. Using IR of adsorbed CO as a probe of the
valence states and coordination numbers of surface chromium
cations for various surface preparations, they report that five-
coordinate Cr3+ is predominant on a fresh Cr2O3 catalyst, and
that fluorination creates active isomerization sites by formation
of CrF3, which increases the Lewis acidity of the cation.
Figure 1. Ball model representation of the Cr2O3 (101h2) surface. The
top view shows the (101h2) surface parallel to the plane of the page. A
surface unit cell is drawn to show the periodicity. The bottom illustration
shows a side view of one stoichiometric repeating layer. The smaller
black spheres represent the chromium cations, and the oxygen anions
are represented by the larger gray spheres.
Changes to Cr2O3 that result from “activation” and fluorina-
tion were studied by Kemnitz and co-workers18 using XPS and
X-ray-excited AES to compare catalyst samples to Cr- and Al-
fluoride standards. They argued that pretreatment of Cr2O3 by
either HX or halocarbons causes the formation of R- and â-CrX3
phases, and the nature of these sites was determined to be
independent of the halogen source, in agreement with the work
of Rowley et al.4 Kemnitz et al. postulated that coordinately
unsaturated (Lewis acid) Cr3+ ions exposed in the â-CrF3
structure are responsible for the catalytic activity because of
their Lewis acidity. They also proposed that all surface activation
(i.e., halogen deposition) reactions occur via HX elimination,
and that over a clean surface, halocarbon compounds decompose
to CO2, HF, and HCl, which in turn halogenate the surface.18
The total halogen uptake (Cl and F) by the surface shows only
a small variation, 0.44 g X/Cr g 0.27, with pretreatment gas.18
surface with the intent of understanding the reaction chemistry
in the absence of HF.
II. r-Cr2O3 (101h2)
The R-Cr2O3 (101h2) surface has been characterized in
previous work.19 The ideal, stoichiometric surface has only one
local coordination environment for the surface cations and
anions. A ball model representation of the ideal, stoichiometric
surface is shown in Figure 1. The topmost atomic layer of the
ideal surface is composed entirely of oxygen anions. The surface
is nonpolar, and one full stoichiometric repeating unit normal
to the surface contains five atomic layers arranged as [O, Cr,
O, Cr, O]. The surface has a rectangular (almost square)
periodicity with a ratio of sides of a/b ) 0.94. At the ideal
(101h2) surface, all O2- anions in the top atomic layer are three-
coordinate with a pyramidal local coordination, and the Cr3+
cations in the second atomic layer are five-coordinate. Both ions
have one degree of coordinative unsaturation relative to their
bulk counterparts.20 All ions below the top two atomic layers
are fully coordinated. A nearly stoichiometric surface can be
prepared by Ar+ ion bombardment and annealing in a vacuum
at 900 K.19
The literature on Cr2O3 contains many suggestions concerning
the reason for the eventual deactivation of the catalyst toward
fluorination reactions. However, a clear picture of Cr2O3
deactivation has not emerged. Coke formation caused by the
degradation of various HCFC compounds is often reported as
the cause of surface deactivation.3 Niedersen et al.17 reported
that surface coke may be burned off in the presence of O2 at
673 K. The formation of halocarbon oligomers and structural
changes to the surface have also been cited as mechanisms for
surface deactivation.10,18 Changes to the oxidation state of
surface cations have also been suggested as a cause for surface
deactivation.10,14
III. Experimental Methods
In the current study, the reaction of CFCl2CH2Cl has been
investigated over the R-Cr2O3 (101h2) surface. CFCl2CH2Cl
(HCFC-131a) is an intermediate in the reaction pathway from
trichloroethene to CF3CH2F (HCFC-134a) over chromia in the
presence of HF.3 The intent of this study is to investigate the
reaction of this intermediate over a well-defined R-Cr2O3 (101h2)
The reaction of CFCl2CH2Cl over a nearly stoichiometric,
(1 × 1), Cr2O3 (101h2) surface was investigated using thermal
desorption spectroscopy (TDS), Auger electron spectroscopy
(AES), and low-energy electron diffraction (LEED). Experi-
ments were conducted in an ion-pumped UHV chamber

5184 J. Phys. Chem. B, Vol. 107, No. 22, 2003
York and Cox
equipped with a Physical Electronics Model 15-155 single-pass
CMA for AES, an Inficon Quadrex 200 mass spectrometer for
TDS and a set of Vacuum Generators 3-grid reverse view LEED
optics. A broad beam ion gun was used for sample cleaning.
The base operating pressure for this study was 1 × 10-10 Torr.
AES data were collected using an incident electron beam of
5 keV for all measurements. Spectra were collected in N(E)
mode and differentiated numerically. All AES measurements
were conducted at 800 K to avoid sample charging. Because of
overlap between the primary oxygen and chromium Auger
peaks, the surface chromium concentration was followed by
measuring the Cr L2,3M2,3M2,3 (490 eV) peak-to-peak height.19
Atomic Cl/Cr ratios were estimated with AES using appropriate
sensitivity factors for the Cl KLL signal21 and the Cr L2,3M2,3M2,3
signal.19 Electron stimulated reduction of the surface (i.e., lattice
oxygen removal) was not observed during AES experiments.
The R-Cr2O3 crystal was oriented to within 1° of the (101h2)
surface using Laue back-reflection and polished to a final mirror
finish with 0.25 µm diamond paste. The sample was mechani-
cally clamped onto a tantalum stage that was fastened to LN2-
cooled copper electrical conductors. A Type K thermocouple
was attached through a hole in the stage to the back of the single
crystal using Aremco #569 ceramic cement. This arrangement
allowed direct measurement of the sample temperature. A nearly
stoichiometric, (1 × 1) surface was prepared by ion-bombard-
ment, followed by annealing to 900 K.19
Figure 2. Desorption features observed following the second 0.03 L
dose of CFCl2CH2Cl in a TDS series initiated on a clean, (1 × 1),
nearly stoichiometric surface. CFCl2CH2Cl (solid line), CFCldCH2
(dashed line), acetylene (dotted line). The baseline for acetylene has
been offset for clarity.
1,1,2-Trichloro-1-fluoroethane (CFCl2CH2Cl) (97% min)
(PCR, Inc.) was used as received. Gas dosing was accomplished
by backfilling the chamber through a variable leak valve. All
doses for CFCl2CH2Cl TDS experiments were conducted at 163
K. The reported dose sizes have been corrected for ion gauge
sensitivity,22 and all desorption traces and quantities have been
corrected for mass spectrometer sensitivity.22-25
Figure 3. The relative quantities desorbed for CFCl2CH2Cl and reaction
products. The TDS experiment consisted of sequential 0.03 L doses
starting on a clean, (1 × 1), nearly stoichiometric Cr2O3 (101h2) surface.
The sum total for all desorbing species (reactant and products) is also
shown.
IV. Results
A combination of thermal desorption and AES results reveals
that CFCl2CH2Cl decomposes into CFCldCH2(g) (HCFC-
1131a), acetylene (HCtCH(g)), and adsorbed halogen. Gas-
phase products were identified by comparison of mass spec-
trometer cracking patterns to thermal desorption peak intensities.
The relative intensities of four m/z signals were used to identify
both CFCldCH2 (80, 45, 26, 31) and acetylene (26, 25, 27,
13). Other products were excluded by a search that included
m/z values ranging from 2 to 200. Specifically, no CO, CO2,
HCl, HF, Cl2, F2, chloroacetylene, fluoroacetylene, or chromium
halide gas-phase reaction products were observed during TDS.
No surface carbon was detected with AES following CFCl2-
CH2Cl decomposition; only surface halogen was observed.
1. CFCl2CH2Cl Thermal Desorption. Figure 2 shows
representative desorption traces for the reactant and gas-phase
products following the second dose of CFCl2CH2Cl in a TDS
series of consecutive 0.03 L doses started on a clean, nearly
stoichiometric (101h2) surface. Nearly 100% of the CFCl2CH2-
Cl reactant is converted to products during initial TDS runs, as
evidenced by the small desorption peak for the reactant molecule
(solid line). The principal product observed is the haloethene
CFCldCH2 (dashed line), formed by a net 1,2-dihalo-elimina-
tion reaction from CFCl2CH2Cl. The CFCldCH2 product
desorption temperature is similar to the desorption temperature
for the CFCl2CH2Cl reactant, suggesting a rate-limiting surface
reaction involving a common surface intermediate for these two
desorbing species (see below). Acetylene (dotted line) is also
formed when the total exposure is low, and yields two desorption
features, one near 335 K and another near 480 K. The acetylene
desorption trace from CFCl2CH2Cl decomposition is identical
to the trace obtained directly from the reaction of CFCldCH2
over Cr2O3 (101h2) in thermal desorption experiments.26
Initially, the nearly stoichiometric Cr2O3 (101h2) surface is
very reactive toward CFCl2CH2Cl, converting nearly 100% of
the reactant to products. Figure 3 shows the relative amounts
of reactant and products desorbed from the surface during a
series of consecutive 0.03 L doses of CFCl2CH2Cl initiated on
a clean, nearly stoichiometric surface. The ratios of CFCl2CH2-
Cl to CFCldCH2 to HCtCH in the desorbing gases from an
initial 0.03 L TDS run is 1:512:40. The amount of the CFCl2-
CH2Cl reactant desorbed from the surface is small following
the first two doses, but increases significantly following 0.10 L
of total CFCl2CH2Cl exposure. The high reactivity of CFCl2-
CH2Cl over clean Cr2O3 (101h2) is evidenced by the small
desorption signal intensity of the reactant molecule during initial
TDS runs.
As shown in Figure 3, the quantity of CFCldCH2 and
acetylene products remains relatively constant for the first three
0.03 L doses of CFClCH2Cl, then the amount of each rapidly
declines. Acetylene production ceases at around 0.15 L of total
exposure, and CFCldCH2 production continues until a total
CFCl2CH2Cl exposure of around 0.22 L is reached. The decline
in product quantities indicates that the surface becomes deac-
tivated toward CFCl2CH2Cl decomposition by a total exposure
of 0.25 L. The sum of all desorbing species decreases by

Dehalogenation of 1,1,2-Trichloro-1-fluoroethane
J. Phys. Chem. B, Vol. 107, No. 22, 2003 5185
Figure 5. AES spectrum of a Cr2O3 (101h2) surface deactivated by
repeated exposure to and thermal desorption of CFCl2CH2Cl. The AES
Cl/Cr ratio for the spectrum is 0.32.
Figure 4b shows desorption traces for the product CFCld
CH2 formed during the same series of 0.03 L doses. The CFCld
CH2 desorption peak initially occurs around 245 K and broadens
to higher temperatures coincident with the shift of the high-
temperature (245-265 K) CFCl2CH2Cl desorption feature in
Figure 4a prior to surface deactivation. A decrease in the amount
of product CFCldCH2 is observed around 0.13 L as the 1,2-
dihalo-elimination reaction of CFCl2CH2Cl to CFCldCH2 shuts
down. The decrease in product CFCldCH2 occurs in tandem
with the increase in reactant (CFCl2CH2Cl) desorption in the
lower temperature range (190-215 K) described above. The
similar temperature ranges for the product CFCldCH2 desorp-
tion and the high-temperature desorption feature of the CFCl2-
CH2Cl reactant suggest that these desorbing species originate
from a common surface intermediate and share the same rate-
limiting surface reaction step.
Figure 4. Desorption traces of CFCl2CH2Cl reactant (a) and CFCld
CH2 product (b) for a series of consecutive 0.03 L doses initiated on a
clean, (1 × 1), nearly stoichiometric surface. The baselines of the
desorption traces have been offset for clarity. The inset in (a) shows
the traces for the first three doses at greater magnification.
2. HC≡CH Product. Dosing CFCl2CH2Cl onto stoichio-
metric Cr2O3 (101h2) yields an acetylene TDS spectrum contain-
ing two features, one around 335 K and the other around 480
K as shown in Figure 2. Acetylene desorption traces obtained
from CFCl2CH2Cl decomposition over a nearly stoichiometric
surface are essentially identical in both shape and peak tem-
perature to the acetylene formed directly from CFCldCH2
decomposition on a nearly stoichiometric surface.26 However,
the quantity of acetylene made from CFCl2CH2Cl decomposition
is small compared to product CFCldCH2, representing less than
10% of the total products prior to surface deactivation.
3. Deposition of Halogen. AES of the sample surface
following the reaction of CFCl2CH2Cl reveals a significant
buildup of surface chlorine. The presence of product acetylene
indicates that a small amount of surface fluorine should also
be seen, but fluorine is not readily seen on this surface using
AES because of a very rapid electron stimulated loss from the
surface.26 The rate of surface chlorine loss is, however,
sufficiently slow to allow for quantification. Figure 5 shows
the AES spectrum of a Cr2O3 (101h2) surface that has been
deactivated by CFCl2CH2Cl exposure and reaction. The spec-
trum was collected at 800 K to avoid sample charging, and gives
an AES Cl/Cr ratio of 0.32 ( 0.01. No carbon is observed on
the deactivated surface, and no evidence is seen for the
replacement of lattice oxygen by halogen for these thermal
desorption conditions.
approximately 50% as the surface becomes deactivated, indicat-
ing a decrease in the CFCl2CH2Cl sticking coefficient as the
surface chemistry shuts down. AES data presented below
demonstrate that the surface Cl/Cr ratio increases as the amount
of products detected in TDS declines. This observation suggests
that deposition of halogen onto the sample surface is responsible
for the observed deactivation.
Figure 4a shows desorption traces of the dosed molecule,
CFCl2CH2Cl, for the same series of consecutive 0.03 L doses
described in Figure 3. Two features are apparent in the
desorption traces of the reactant molecule: a high-temperature
feature near 265 K and a lower temperature feature around 190-
215 K. The high-temperature contribution occurs at low total
exposures and initially has a peak desorption temperature of
245 K (see insert in Figure 4a). The high-temperature feature
increases in intensity and shifts to 265 K during the first three
0.03 L doses. Following the fourth dose, a low-temperature
contribution to the CFCl2CH2Cl desorption feature is observed
around 215 K, while the high-temperature peak remains
relatively unchanged. In subsequent TDS runs, the high-
temperature feature rapidly dies out while the low-temperature
feature increases in intensity and shifts downward to around
190 K. Differences in the peak desorption temperatures and the
shapes of the two CFCl2CH2Cl desorption features suggest that
two separate desorption processes occur for the reactant
molecule.
Figure 6 shows a comparison of the variation in surface AES
Cl/Cr ratio (left axis) and the sum of the amounts of product

5186 J. Phys. Chem. B, Vol. 107, No. 22, 2003
York and Cox
basis of the (1 × 1) LEED pattern observed for a chlorinated
surface, it was assumed that the chlorine adatoms are located
on the surface cations in approximate registry with the positions
of the missing nearest-neighbor oxide anions that would be
found in the bulk structure. A Cr-Cl bond length was estimated
using the combined ionic radii for chromium(III) (0.63 Å) and
chloride ion (1.32 Å) obtained from the literature.28 For a sample
orientation normal to the axis of the CMA and a signal collection
geometry 42° off the spectrometer axis, the expected Cl/Cr ratio
for 100% coverage of surface cations by chlorine is estimated
to be 0.325. This value is in good agreement with the Cl/Cr )
0.32 measured from a halogen-saturated (deactivated) surface.
Therefore, the deactivated (Cl/Cr ) 0.32) surface may be
considered to have a chlorine coverage of essentially one
adsorbed chlorine adatom per surface Cr3+ cation. Additional
thermal desorption runs following adsorption at 163 K produce
no increase in the Cl/Cr ratio of 0.32 and no additional product.
In addition, CFCl2CH2Cl exposures of up to 2.0 L at 773 K
produced no increase in the AES Cl/Cr ratio beyond Cl/Cr )
0.32. Deactivation of the Cr2O3 (101h2) surface toward CFCl2-
CH2Cl decomposition in the thermal desorption experiments is
directly attributed to site blocking of surface cations by deposited
halogen adatoms.
4. Thermal Loss of Surface Chlorine. It was found that
chlorine can be removed from the sample surface by heating to
1100 K. AES spectra of a deactivated surface (AES Cl/Cr )
0.32) taken following five minutes of heating at 1100 K show
that all of the Cl is removed, and the nearly stoichiometric
surface is returned. The desorption of halogen and halogen-
containing species was checked for with TDS, and no desorbing
halogen or halogen-containing products other than CFCldCH2
were detected. The lack of evidence for desorption of surface
halogen or halogen-containing species (ex., CrXn) suggests that
the thermal removal of chlorine from the sample surface occurs
via migration into the sample bulk.
AES experiments were conducted on deactivated surfaces
prepared by exposure to CFCl2CH2Cl to determine the rate of
Cl thermal migration. AES spectra were collected as a function
of time to monitor the loss of chlorine for three different
annealing temperatures. The variation in Cl/Cr ratio with time
is shown in Figure 7a. The time dependence of the surface Cl/
Cr ratio is reasonably fit by first-order kinetics at each
temperature. An Arrhenius analysis (Figure 7b) yields a first-
order activation energy for chlorine migration of 173 kJ/mol.
Given the rapid electron stimulated loss of fluorine, the
kinetics of thermal fluorine removal from the surface cannot
be directly addressed with AES measurements, although the
deposition of a small amount of fluorine is indicated by the
stoichiometry of CFCl2CH2Cl decomposition to acetylene.
However, since no additional fluorinated desorption products
other than CFCldCH2 were observed, it seems reasonable to
expect that any thermal loss of fluorine from the surface would
occur via migration into the sample bulk, as suspected for
chlorine.
Figure 6. The total quantity of products (sum of CClFCH2 and HCt
CH) and variation in surface Cl/Cr ratio for a series of consecutive
0.03 L doses of CFCl2CH2Cl initiated on a clean, (1 × 1), nearly
stoichiometric surface.
(CFCldCH2 and HCtCH) desorbed in TDS (right axis) for a
series of consecutive 0.03 L doses of CFCl2CH2Cl initiated on
a clean, nearly stoichiometric Cr2O3 (101h2) surface. AES spectra
were collected immediately following each TDS run. The
surface Cl/Cr ratio rises rapidly then levels off after a total
exposure of around 0.75 L. The surface Cl/Cr ratio increases
most rapidly for total CFCl2CH2Cl exposures below 0.25 L,
when the conversion to products is highest. As less product is
made, the product sum declines while the surface Cl/Cr ratio
levels off at about 0.32, indicating that surface halogen is
deposited as a result of CFCl2CH2Cl decomposition. The surface
Cl/Cr ratio stops increasing when the product formation ceases.
LEED was used to examine the changes in surface periodicity
with chlorine deposition. The clean, nearly stoichiometric
surface exhibits a (1 × 1) periodicity.19 As chlorine builds up
on the surface, an increase in diffuse background is observed
at low chlorine coverages, but no change is seen in the
underlying (1 × 1) periodicity. As the coverage approaches
saturation, the (1 × 1) LEED pattern sharpens and the diffuse
background decreases. The variation in the LEED pattern
suggests that the chlorine adlayer is disordered at low coverage,
but approaches an integer Cl coverage per surface unit cell at
saturation.
Due to molecular stoichiometry and the product distribution,
the halogen deposited onto the Cr2O3 (101h2) surface from CFCl2-
CH2Cl decomposition should be overwhelmingly chlorine. On
the basis of the relative quantities of CFCl2CH2Cl and HCt
CH formed, the expected amount of surface fluorine is estimated
to be only about 4% of the total amount of deposited surface
halogen. While the surface fluorine cannot be quantified with
AES because of its rapid electron-stimulated removal, disregard-
ing the presence of fluorine at the halogen-saturated surface
should introduce an approximate error of only about 0.01 in
the value of X/Cr. Therefore, the AES Cl/Cr ratio of 0.32 may
be regarded as an excellent approximation to the total halogen
content (X/Cr) of the halogen-saturated (deactivated) surface.
A surface deactivated by CFCl2CH2Cl thermal desorption
treatments yields an experimental AES Cl/Cr ratio of 0.32. Using
an exponential AES signal decay model, an estimate was made
of the expected Cl/Cr ratio for a surface containing one adsorbed
chlorine adatom per five-coordinate surface cation. The calcula-
tion involves conventional layer-by-layer summations assuming
an exponential signal decay with increasing sampling depth,
and ignoring diffraction effects. Interlayer spacings in the
substrate were assumed to be the same as in bulk R-Cr2O3. An
electron mean free path of 9 Å was estimated from the “universal
curve”27 and was used to describe the Auger emission for both
Cr and O lines, which occur at similar kinetic energies. On the
V. Discussion
The decomposition of CFCl2CH2Cl over nearly stoichiometric
Cr2O3 (101h2) yields surface halogen and CFCldCH2 as major
products, and acetylene as a minor product. As mentioned above,
acetylene is the only gas-phase product observed for the reaction
of CFCldCH2 over this same surface.26 Given that the desorp-
tion traces for acetylene from both starting reactants (CFCl2-
CH2Cl and CFCldCH2) are essentially identical, the decom-
position pathway for CFCl2CH2Cl over the Cr2O3 (101h2) is

Dehalogenation of 1,1,2-Trichloro-1-fluoroethane
J. Phys. Chem. B, Vol. 107, No. 22, 2003 5187
temperature desorption state is the result of surface halogenation
associated with the deactivation of the surface under the
conditions of this thermal desorption study. Since molecular
adsorption and desorption of CFCl2CH2Cl is observed on the
deactivated surface after the halogenation (site blocking) of
essentially all the surface Cr3+ cation sites, desorption via the
low-temperature channel is thought to originate from molecular
CFCl2CH2Cl bound at chlorinated cation sites or three-
coordinate oxygen anions via a weak physical interaction. The
total uptake (CFCl2CH2Cl plus products) decreases by about
50% as the surface changes from a clean, stoichiometric
condition to chlorine-saturated (deactivated).
B. 265 K Reaction Channel. Despite the high conversion to
products on a clean surface, a small CFCl2CH2Cl desorption
feature is observed (245-265 K) in the same temperature range
as product CFCldCH2 during initial 0.03 L doses (Figure 4).
The coincident desorption of both the CFCl2CH2Cl reactant and
CFCldCH2 product molecules in TDS indicates that they arise
from a common surface intermediate and share the same rate-
limiting step. Additionally, the desorption temperature observed
for product CFCldCH2 (245 K and broadened to higher
temperatures) from the reaction of CFCl2CH2Cl is 30-50 K
higher in temperature than the desorption signal observed for a
molecular CFCldCH2 adsorbate.26 The higher desorption tem-
perature for product CFCldCH2 also indicates a reaction-limited
process in this temperature range. An apparent first-order
activation energy for this process of 69 kJ/mol can be estimated
with the Redhead equation31 assuming a normal first-order pre-
exponential of 1013 s-1
.
The first step in the surface reaction mechanism is dissociative
adsorption of CFCl2CH2Cl via C-Cl bond cleavage at five-
coordinate Cr3+ cations. The literature indicates that the C-Cl
bond on the more chlorinated carbon (CFCl2- group) is the
weakest bond, and hence the most likely bond to be broken
during dissociative adsorption.4,32-34 In support of this idea, we
have found a pattern of decreasing reactivity with modification
to this end group (CFCl2CH2Cl > CF2ClCH2Cl . CF3CH2-
Cl) due to an increasing barrier to dissociative adsorption on
Cr2O3 (101h2).35 Since there is no change in composition of the
-CH2Cl end group on the opposite end of the molecule for
this sequence of reactants, these results are in agreement with
the idea that the initial attack of the surface is at the C-Cl
bond of the CFCl2- group in CFCl2CH2Cl. Similarly, Gellman
and co-workers33,34 studied the rate constant for dechlorination
of a sequence of haloethanes over Pd(111), and found that the
rate of dechlorination is reduced for haloethanes containing
fewer chlorine atoms per carbon. They also found that the
transition state for the dechlorination reaction is not highly
polarized relative to the initial state, suggesting homolytic C-Cl
bond cleavage. The reaction pathway in this study is attributed
to the dissociative adsorption of CFCl2CH2Cl which results in
a surface alkyl group (-CFClCH2Cl) and adsorbed chlorine,
Cl(s), both bound at surface cation sites. Similar alkyl groups
attached to transition metal cations are well-known in the
organometallic literature, and a strong interaction between
chlorine and microcrystalline Cr2O3 has been documented in
the catalysis literature1,35 and observed in this study.
Figure 7. Kinetics of thermal Cl loss from the Cr2O3 (101h2) surface:
(a) first-order analysis for the rate constant, k, at several temperatures,
and (b) Arrhenius analysis for determination of the activation energy.
attributed to two reactions in series with acetylene formed by
the further reaction of product CFCldCH2:
CFCl2CH2Cl f CFCl)CH2 + 2Cl(s)
CFCl)CH2 f HC≡CH + Cl(s) + F(s)
While HX abstraction is one of the most often reported reactions
for CFCl2CH2Cl and similar HCFC compounds over Cr2O3
powders,13,15,29 no dehydrohalogenation reaction products are
observed under the conditions of this thermal desorption study.
The reaction of CFCl2CH2Cl to CFCldCH2 most closely
resembles the Zn-based chemistry utilized in organic synthesis
for the dehalogenation of vicinal dihalides to alkenes.30
1. CFCl2CH2Cl Adsorption/Desorption and Reaction
Pathway to CFCldCH2. A. 210-190 K CFCl2CH2Cl Desorp-
tion Channel. The low-temperature (210-190 K) CFCl2CH2Cl
desorption channel is first observed after some initial halogena-
tion of the surface as the formation of reaction products declines.
The low peak temperature suggests that desorption of a
molecularly bound CFCl2CH2Cl surface species is responsible
for this desorption feature. The activation energy for desorption
from this state (190 K) is estimated to be around 48 kJ/mol
using the Redhead equation and a normal, first-order pre-
Following the dissociative adsorption of CFCl2CH2Cl, two
elementary reaction pathways are operable:
(1) The surface alkyl may undergo a rate-limiting â-chlorine
elimination step to form the CFCldCH2 product molecule and
a second Cl(s).
exponential factor of 1013 s-1 31
.
It is significant that the low-temperature CFCl2CH2Cl de-
sorption feature is not observed during initial TDS runs at small
doses. AES spectra of the surface taken in conjunction with
TDS experiments show that this feature grows in as the surface
becomes progressively halogenated, suggesting that the low-
Cr-CFClCH2Cl f Cr-Cl + CFCl)CH2

5188 J. Phys. Chem. B, Vol. 107, No. 22, 2003
York and Cox
SCHEME 2: Reaction Pathway from CFCldCH2 to
SCHEME 1: Reaction Pathway from CFCl2CH2Cl to
CFCldCH2
HCtCH from Ref 26
(2) The surface alkyl may recombine with chlorine liberated
by the rate-limiting â elimination to reform the CFCl2CH2Cl
reactant molecule.
nylidene to acetylene then occurs via a 2,1 hydrogen shift.26
The 335 K acetylene feature is desorption limited, and associated
with the first-order desorption of molecularly bound HCtCH,
while the 480 K feature is reaction limited with first-order
kinetics.37,38 The reaction pathway from haloalkene to acetylene
is shown in Scheme 2.
Cr-CFClCH2Cl + Cl(s) f Cr- + CFCl2CH2Cl(g)
Chlorine is provided for the recombination reaction via the same
rate-limiting â-chlorine elimination step, as evidenced by
coincident desorption of CFCldCH2 and CFCl2CH2Cl near 265
K. Note that the desorption rate of CFCl2CH2Cl formed by
recombination of Cl and surface alkyl should be rapid since
the recombination process occurs above the molecular desorption
temperature (210 K). The organometallic literature contains
many examples of â-halogen elimination from alkyl com-
plexes,36-38 and â-halogen elimination has been observed from
haloalkyl surface intermediates on Ag(111)39 and Si(100).40
The first reaction to CFCldCH2 is favored initially, as
evidenced in TDS by the much larger yield of CFCldCH2
compared to CFCl2CH2Cl at 265 K, but the probability for the
recombination reaction increases (slightly) as the extent of
surface chlorination increases, as evidenced by the increase in
CFCl2CH2Cl desorption at high temperatures with the first few
exposures. The combination of TDS and AES data demonstrates
that as the surface becomes increasingly chlorinated, the
recombination to CFCl2CH2Cl becomes increasingly favored.
However, as the halogen content of the surface nears saturation,
both pathways become unavailable as the dissociation of CFCl2-
CH2Cl shuts down due to a lack of exposed (coordinately
unsaturated) cations. It is postulated that the chlorine liberated
via â-elimination bonds to five-coordinate Cr3+ sites when they
are available (i.e., at low surface Cl/Cr ratios). As the surface
becomes increasingly halogenated, bare cation sites become less
available and recombination of chlorine with a surface alkyl
becomes more likely, as evidenced by the increase in CFCl2-
CH2Cl desorption. Consequently, the CFCl2CH2Cl desorption
intensity increases and the product CFCldCH2 intensity de-
creases as the surface halogen content increases. Eventually,
surface deactivation results in essentially 100% of surface
cations being capped by adsorbed halogen and the formation
of CFCldCH2 is completely suppressed. The reaction pathway
from haloalkane to haloalkene is shown in Scheme 1.
3. Surface Chlorine. Various authors have reported the
uptake of halogen by Cr2O3 microcrystalline powder. Kemnitz
et al.18 used XPS to measure the amount of halogen uptake by
microcrystalline Cr2O3 powder exposed to CFCl2CH2Cl at 673
K and atmospheric pressure. They reported a maximum X/Cr
ratio of approximately 0.44. Blanchard and co-workers8,10
studied the uptake of fluorine over microcrystalline Cr2O3
powder exposed to CF3CH2Cl at 653 K and atmospheric
pressure and reported a F/Cr ratio of 0.36. Coulson and co-
workers14 report that exposures of CHF3 and HF at 698 K over
microcrystalline Cr2O3 yielded F/Cr ratios of 0.22 and 0.30,
respectively. The AES Cl/Cr ) 0.32 reported herein lies within
the range of X/Cr values found in the catalysis literature.
Under the conditions of this thermal desorption study, the
deactivation of the Cr2O3 (101h2) surface was found to be
coincident with a 1:1 surface chlorine-to-cation ratio (i.e., a
halogen-saturated surface). Deactivation of the surface occurs
by “capping” surface Cr3+ cations with chlorine (i.e., by site
blocking), thereby preventing reaction by isolating the surface
chromium cations from adsorbates. Site blocking by fluorine
has similarly been reported in our studies of the reaction of
CFCldCH2 to acetylene.26 A similar argument was presented
by Kemnitz et al.18 to explain the deactivation of microcrys-
talline Cr2O3 powder samples toward fluorination reactions. No
evidence was seen for the replacement of surface lattice oxygen
by chlorine under the conditions of this study.
The absence of gas-phase desorption products during thermal
treatments of the chlorine-saturated surface suggests that surface
chlorine loss occurs via migration into the sample bulk. A similar
result has been reported for fluorine.26 While no reports of
halogen migration into Cr2O3 could be found in the surface
science literature, Freund and co-workers report that sodium
adatoms on Cr2O3(0001) thin films also migrate into the sample
bulk.41
4. Comparison to the Catalysis Literature. The chemistry
observed in the thermal desorption experiments in this study
differs significantly from the reported chemistry of haloalkanes
over supported and powdered chromia catalysts. Dehalogenation
chemistry is observed, but no dehydrohalogenation, halogen
exchange, isomerization, or disproportionation chemistry is seen.
The reaction pathway from haloalkane to haloalkene is the
primary chemistry observed in this study, and occurs below
room temperature between 245 and 265 K. In contrast, catalytic
studies are typically conducted between 600 and 700 K.4-9,11
It is likely that the typical catalytic reactions that were not
observed have higher activation energies than the reaction and
desorption steps observed in this study. Given the low activation
2. Reaction Pathway to HCtCH Product. The acetylene
desorption traces observed during CFCl2CH2Cl thermal desorp-
tion are essentially identical to the traces observed for the direct
reaction of CFCldCH2 to acetylene over this same surface.26
Both CFCl2CH2Cl and CFCldCH2 reactants produce traces that
contain two product acetylene desorption features, with similar
peak shapes and peak desorption temperatures. The similarity
in TDS characteristics demonstrates that the kinetics and surface
species involved in both desorption processes are the same.
The second overall reaction step in the sequence to acetylene
has been studied in detail.26 CFCldCH2 undergoes C-Cl bond
cleavage to form a surface fluorovinyl intermediate (Cr-CFd
CH2) followed by R-fluorine elimination to yield a surface
vinylidene intermediate (CrdCdCH2). Isomerization of vi-

Dehalogenation of 1,1,2-Trichloro-1-fluoroethane
J. Phys. Chem. B, Vol. 107, No. 22, 2003 5189
barrier for desorption of CFCl2CH2Cl from the halogenated
surface, the reactant is desorbed at low temperature before the
barrier to the unobserved reactions can be scaled. Under typical
catalytic reaction conditions, the higher partial pressure of
reactants will yield some stationary coverage of reactant at
higher temperatures.
Biosciences Division, Office of Basic Energy Sciences, Office
of Science, U.S. Department of Energy through Grant DE-FG02-
97ER14751.
References and Notes
(1) Wiersma, A.; van de Sant, E. J. A. X.; Makkee, M.; van Bekkum,
H.; Moulijn, J. A. Stud. Surf. Sci. Catal. 1996, 101, 369.
(2) Blanchard, M.; Wendlinger, L.; Canesson, P. Appl. Catal. 1990,
59, 123.
(3) Manzer, L. E.; Rao, V. N. M. AdV. Catal. 1993, 39, 329.
(4) Rowley, L.; Webb, G.; Winfield, J. M.; McCulloch, A. Appl. Catal.
1989, 52, 69.
(5) Kijowski, J.; Webb, G.; Winfield, J. M. Appl. Catal. 1986, 27, 181.
(6) Rowley, L.; Thomson, J.; Webb, G.; Winfield, J.; McCulloch, A.
Appl. Catal. A 1991, 79, 89.
(7) Kijowski, J.; Webb, G.; Winfield, J. M. J. Fluorine Chem. 1985,
27, 213.
(8) Bechadergue, D.; Blanchard, D. M.; Canesson, P. Stud. Surf. Sci.
Catal.; Elsevier: New York, 1988.
(9) Bechadergue, D.; Blanchard, M.; Canesson, P. Appl. Catal. 1986,
20, 179.
(10) Barrault, J.; Brunet, S.; Requieme, B.; Blanchard, M. J. Chem. Soc.,
Chem. Commun. 1993, 35, 374.
(11) Blanchard, M.; Wendlinger, L.; Canesson, P. Appl. Catal. 1990,
The deactivation of the surface due to site blocking of Cr3+
cations by chlorine adatoms also contrasts significantly with
the catalytic literature on haloalkane reactions over chromia.
The level of halogenation observed for the deactivated Cr2O3
(101h2) surface in this study (Cl/Cr ) 0.32) is similar to that
reported for halogen-pretreated (i.e., activated) chromia catalysts.
While it is not necessarily apparent that the majority of surface
cations sites are capped (blocked) at these halogen coverages
on chromia catalysts as they are on Cr2O3 (101h2), the elevated
reaction temperatures in the catalytic studies may give rise to a
higher level of reactivity of halogen adatoms. The surface-to-
bulk diffusion of chlorine observed in this study indicates the
mobility and likely higher reactivity of halogen adatoms at
higher temperatures.
Despite the differences observed between this model thermal
desorption study and the catalysis of chromia in HCFC reactions,
the current results do add insight into possible elementary steps
involved in the catalysis. The net 1,2-dihalo-elimination reaction
(dissociation and â halogen elimination) is an additional pathway
that may be operable in addition to dehydrohalogenation (H-X
elimination) to give haloalkene from haloalkane. The apparent
reversibility of the first elementary step in this reaction sequence
(dissociative adsorption) with kinetics attributable to a â halogen
rate-limiting step suggests a possible elementary reaction
sequence associated with halogen exchange. The recombination
step is only observed for partially halogenated surfaces, and
represents an alternative to halogen trapping at coordinately
unsaturated surface cations sites. Halogen exchange processes
could occur by reaction of surface haloalkyl intermediates with
mobile surface halogen species (for example, HX or halogen
liberated by other processes such as â elimination on a
halogenated surface) to reform haloalkane.
59, 123.
(12) Barrault, J.; Brunet, S.; Requieme, B.; Blanchard, M. J. Chem. Soc.,
Chem. Commun. 1993, 35, 374.
(13) Brunet, S.; Requieme, B.; Colnay, E.; Barrault, J.; Blanchard, M.
Appl. Catal. B: EnViron. 1995, 5, 305.
(14) Coulson, D. R.; Wijnen, P. W. J. G.; Lerou, J. J.; Manzer, L. E. J.
Catal. 1992, 140, 103.
(15) Kohne, A.; Kemnitz, E. J. Fluorine Chem. 1995, 75, 103.
(16) Kemnitz, E.; Niedersen, K.-U. J. Catal. 1995, 155, 111.
(17) Niedersen, K.-U.; Schreier, E.; Kemnitz, E. J. Catal. 1997, 167,
210.
(18) Kemnitz, E.; Kohne, A.; Grohmann, I.; Lippitz, A.; Unger, W. E.
S. J. Catal. 1996, 159, 270.
(19) York, S. C.; Abee, M. W.; Cox, D. F. Surf. Sci. 1999, 437, 386.
(20) Lad, R. J.; Henrich, V. E. Surf. Sci. 1988, 193, 81.
(21) Davis, L. E.; MacDonald, N. C.; Palmberg, P. W.; Riach, G. E.;
Weber, R. E. Handbook of Auger Electron Spectroscopy; Perkin-Elmer:
Eden Prairie, MN, 1976.
(22) An ion gauge sensitivity of 11.4 was used for CFCl2CH2Cl. This
value was calculated using the correlation of S. George reported in ref 25.
A mass spectrometer sensitivity factor of 1.00 was arbitrarily set for m/z )
115. All other mass spectrometer sensitivity factors are relative to this value.
(23) An ion gauge sensitivity of 5.9 was used for CFCldCH2 from the
correlation in ref 25. A relative mass spectrometer sensitivity factor of 1.45
was determined experimentally for m/z ) 80.
VI. Conclusions
(24) An ion gauge sensitivity of 1.7 was used for HCtCH from the
correlation in ref 25. A relative mass spectrometer sensitivity factor of 5.34
was determined experimentally for m/z ) 26.
(25) Brainard, R. L.; Madix, R. J. J. Am. Chem. Soc. 1989, 111, 3826.
(26) York, S. C.; Cox, D. F. J. Catal. 2003, 214, 273.
(27) Riggs, W. M.; Parker, M. J. In Methods of Surface Science Analysis;
Czanderna, A. W., Ed.; Elsevier: Amsterdam, 1975; p 103.
(28) Wells, A. F. Structural Inorganic Chemistry; Clarendon Press:
Oxford, 1984.
CFCl2CH2Cl has been found to decompose over a nearly
stoichiometric Cr2O3 (101h2) surface to CFCldCH2, acetylene,
and surface halogen. The chlorine removed from CFCl2CH2Cl
was found to bind strongly to the surface. Chlorine bound at
five-coordinate surface Cr3+ sites shuts down the dehalogenation
reactions by site blocking. At elevated temperatures, the thermal
removal of surface chlorine is observed and attributed to
migration into the sample bulk. The first-order activation energy
for this chlorine diffusion process is 173 kJ/mol.
(29) Kemnitz, E.; Niedersen, K.-U. J. Fluorine Chem. 1996, 79, 111.
(30) Morrison, R. T.; Boyd, R. N. Organic Chemistry; Allyn and Bacon
Inc.: Boston, 1983.
The reaction proceeds via dissociative adsorption of CFCl2-
CH2Cl at the CFCl2- end of the molecule through a presumably
homolytic C-Cl bond cleavage to give a haloalkyl, -CFClCH2-
Cl, surface intermediate. A rate-limiting â-chlorine elimination
from the surface alkyl intermediate gives rise to the CFCld
CH2 product. A small quantity of acetylene is also formed by
the subsequent reaction of CFCldCH2 in series. No carbon
buildup was observed on deactivated surfaces, and no evidence
was seen for the replacement of surface lattice oxygen by
halogen under the conditions of this study.
(31) Redhead, P. A. Vacuum 1962, 12, 203.
(32) Marangoni, L.; Gervasutti, C.; Conte, L. J. Fluorine Chem. 1981/
82, 19, 21.
(33) Chan, C. W.; Gellman, A. J. Catal. Lett. 1998, 53, 139.
(34) Zhou, G.; Chan, C.; Gellman, A. J. J. Phys. Chem. 1999, 103, 1134.
(35) York, S. C. Ph.D. Thesis, Virginia Tech, 1999.
(36) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Fink, R. G. Principles
and Applications of Organometalic Chemistry; University Science Books:
Mill Valley, CA, 1987.
(37) Bruce, M.; Swincer, A. AdV. Organometallic Chem. 1983, 22, 59.
(38) Stang, P. Chem. ReV. 1978, 78, 383.
(39) Paul, A.; Gellman, A. J. Langmuir 1995, 11, 4433.
(40) Lin, J.-L.; Yates, J. T. J. Vac. Sci. Technol. A 1995, 13, 178.
(41) Dillmann, B.; Rohr, F.; Seiferth, O.; Klivenyi, G.; Bender, M.;
Homann, K.; Yakovkin, I.; Ehrlich, D.; Baumer, M.; Kuhlenbeck, H.;
Freund, H. Faraday Discuss. 1996, 105, 295.
Acknowledgment. The authors gratefully acknowledge
financial support by the Chemical Sciences, Geosciences and