Hydrogen-assisted 1,2-dichloroethane dechlorination catalyzed by Pt-Cu/SiO2: Evidence for different functions of Pt and Cu sites

Article abstract of DOI:10.1021/jp0300389

A monometallic Pt/SiO₂ and two bimetallic Pt-Cu/SiO₂ with Pt/Cu atomic ratio of 1:1 (Pt1Cu1) and 0.33:1 (Pt1Cu3) were examined by a combination of reaction kinetics and FTIR spectroscopic studies to understand the factors that control the selectivity toward ethylene and ethane in the CH₂Cl-CH₂Cl + H₂ reaction. The only product of the CH₂Cl-CH₂Cl + H₂ reaction catalyzed by platinum and Pt1Cu1 in a static reaction was ethane, whereas ethylene formed over the Pt1Cu3 with selectivity of 65-80%, depending on the CH₂Cl-CH₂Cl conversion. Treatment of the Pt1Cu1 catalyst that was already used in the CH₂Cl-CH₂Cl + H₂ reaction with CO at 473 K for 0.5 hr enhanced the ethylene selectivity in the presence of a small amount CO in the reaction mixture. The singleton frequency of CO adsorbed on platinum of the ethylene selective Pt1Cu3 was 2030/cm, while that of the used Pt1Cu1 treated in CO and CH₂Cl-CH₂Cl on the Pt1Cu3 at room temperature showed that 1,2-dichloroethane could compete with CO for adsorption sites on Cu but could not compete for the sites on Pt.

Full text of DOI:10.1021/jp0300389

5568
J. Phys. Chem. B 2003, 107, 5568-5574
Hydrogen-Assisted 1,2-Dichloroethane Dechlorination Catalyzed by Pt-Cu/SiO2: Evidence
for Different Functions of Pt and Cu Sites
Victor Yu. Borovkov,† David R. Luebke, Vladimir I. Kovalchuk, and Julie L. d’Itri*
Department of Chemical Engineering, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15261
ReceiVed: January 10, 2003; In Final Form: April 8, 2003
A monometallic Pt/SiO2 and two bimetallic Pt-Cu/SiO2 with Pt/Cu atomic ratio of 1 (Pt1Cu1) and 0.33
(Pt1Cu3) have been investigated by a combination of reaction kinetics and FTIR spectroscopic studies in
order to understand the factors that control the selectivity toward ethylene and ethane in the CH2Cl-CH2Cl
+ H2 reaction. Carbon monoxide adsorption was used to probe the electronic state (ligand effect) of Pt and
Cu, and the selective poisoning of Pt sites with CO under catalytic reaction conditions was employed to
elucidate the impact of the ensemble effect and the role of Pt and Cu sites in the conversion of
1,2-dichloroethane. It was shown that both Pt and Cu in the bimetallic Pt-Cu catalysts are modified
electronically, but this modification does not impact the catalysts’ selectivity patterns. However, the addition
of CO into the CH2Cl-CH2Cl + H2 reaction mixture at 473 K to block Pt sites resulted in a significant
improvement in the ethylene selectivity of the bimetallic catalysts at the expense of ethane. The observations
are consistent with the idea that with the Pt-Cu catalysts, ethylene forms on Cu sites and the role of Pt is to
produce a limited supply of dissociated H atoms that spill over to Cu and react thereon with adsorbed Cl
atoms.
Introduction
spectroscopy. This technique has been used extensively, espe-
cially for the investigation of Pt-containing bimetallic alloy
catalysts.2,10-13 The value of the singleton frequency of CO
adsorbed on Pt is a measure of the degree of electron density
transfer to or from Pt, whereas the CO band frequency shift
resulting from dipole-dipole coupling provides information
about the size of Pt ensembles.14-18 On the basis of the analysis
of the dipole-dipole coupling frequency shifts, the singleton
frequencies of CO adsorbed on Pt, and the chemical reaction
kinetics, it was concluded that ligand effect does not affect the
ethylene selectivity of Pt-Cu/SiO2 catalysts in hydrogen-assisted
dechlorination of 1,2-dichloroethane.2 It was also suggested that
ethylene forms on Cu sites and the role of Pt is to provide
dissociated H atoms that spill over onto the Cu and react thereon
with adsorbed Cl atoms to regenerate the catalytically active
sites. However, no direct evidence to support this hypothesis
was provided; also, no conclusion was made on the electronic
state of Cu.
Previous research has shown that addition of Cu to Pt catalysts
enhances the selectivity toward olefins in the reaction of vicinal
alkyldichlorides with hydrogen and decreases the selectivity
toward paraffins.1-4 The ethylene selectivity of the Pt-Cu
catalysts in the hydrogen-assisted dechlorination of 1,2-di-
chloroethane can approach 100% depending on the atomic ratio
of Pt to Cu.1,3,4 In addition, the selectivity is a strong function
of time on stream. It is relatively low initially, approaching its
steady-state value in several tens of hours.1,3 The latter was
explained by continuous alloying of the metallic components
of the catalysts initially located apart because of chromato-
graphic separation during catalyst preparation by impregnation
of the support with inorganic salts of Pt and Cu.3 However,
direct experimental evidence for enhancing the degree of
alloying of Pt and Cu in the course of the reaction was not
provided.
There are many examples in which alloying results in
significant changes in catalytic performance compared to that
of the individual alloy components. In the case of Pt-Cu
particles, Cu can donate electron density to the incompletely
filled d-bands of Pt (ligand effect5). Copper may also decrease
the average number of contiguous Pt atoms at the surface that
participate in the reaction by simple dilution (ensemble effect6).
In addition, alloying can influence the adsorption energies of
reactants and reaction products thereby changing the catalyst
activity,7 or it can suppress side reactions resulting in a change
in catalyst selectivity.8,9
The objectives of the present investigation were to probe by
FTIR study of CO adsorption the alloying and the electronic
state of both Pt and Cu in bimetallic Pt-Cu/SiO2 catalysts and
to obtain experimental evidence for different function of Pt and
Cu sites in the hydrogen-assisted dechlorination of 1,2-di-
chloroethane. For the latter, the Pt sites of Pt-Cu alloy particles
were selectively blocked for 1,2-dichloroethane adsorption with
CO.
Experimental Section
One means of effectively probing both ligand and ensemble
effects is the examination of CO adsorption on alloys using IR
The catalysts were prepared by pore volume impregnation
of SiO2 (Aldrich, 99+%, 60-100 mesh, 300 m2 g-1, average
pore diameter, 150 Å) with a 0.1 N aqueous HCl solution
containing H2PtCl6‚6H2O (Alfa, 99.9%) or a mixture of H2-
PtCl6‚6H2O and CuCl2‚2H2O (MCB Manufacturing Chemists,
99.5%), as described elsewhere.1 The concentrations of the
* Author to whom correspondence should be addressed. E-mail:
jditri@pitt.edu.
† Permanent address: N. D. Zelinsky Institute of Organic Chemistry,
117334 Moscow, Russia.
10.1021/jp0300389 CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/20/2003

H2-Assisted 1,2-Dichloroethane Dechlorination on Pt-Cu/SiO2
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5569
metals in the impregnating solutions were adjusted to obtain
metal loadings of 2.4% for Pt/SiO2, and 3.0% Pt + 1.0% Cu
and 3.0% Pt + 2.9% Cu for the two Pt-Cu/SiO2. The bimetallic
catalysts had Pt/Cu atomic ratios of 1:1 (Pt1Cu1) and 1:3
(Pt1Cu3), respectively. The fraction of Pt atoms exposed after
reduction at 493 K, as determined by CO chemisorption,1 was
43, 10, and 4% for the Pt, Pt1Cu1, and Pt1Cu3, respectively.
The average metal particle size for the catalysts, determined by
TEM measurements using electron microscope JEM 2010 with
a resolution of 0.14 nm and accelerating voltage of 200 kW as
described elsewhere,19 was 2.8, 3.4, and 3.2 nm for the Pt,
Pt1Cu1, and Pt1Cu3 catalysts, respectively, reduced at 493 K.
experiments were conducted at a reaction temperature of 523
K. The reaction mixture consisted of 7-10 Torr of CH2Cl-
CH2Cl (Sigma-Aldrich, 99.8%) and a 5-fold excess of H2. For
experiments in which CO was used as a site blocker, approxi-
mately 2 Torr of CO (Praxair, 99.99%) was admitted to the
initial reaction mixture. The CH2Cl-CH2Cl conversion and the
composition of the reaction products were determined from the
band intensities in IR spectra of gaseous CH2Cl-CH2Cl (1238
cm-1), C2H4 (950 cm-1), and C2H6 (2970 cm-1). As the cell
contents were not mixed, it is possible that the kinetics
measurements were influenced by the diffusion of the gases.
Thus, these results were used only for a qualitative analysis,
and a quantitative investigation was conducted using a dif-
ferential flow reactor system.
The flow kinetics experiments were conducted in a stainless
steel, differential, flow reaction system equipped with a quartz
microreactor (10 mm i.d.) in which the catalyst was supported
on a quartz frit. The pretreatment gases and gaseous reactants
were metered using mass flow controllers (Brooks, model
5850E) and mixed prior to entering the reactor. The 1,2-
dichloroethane was metered into the system by flowing He
through a saturator containing the liquid reactant. A constant
inlet concentration was ensured by maintaining the saturator at
a fixed temperature of 273 ( 1 K using a recirculating cooling
system. The catalyst temperature was maintained (1 K using
an electric furnace and a temperature controller (Omega model
CN2011). The reaction products were monitored online using
a gas chromatograph (Varian 3300 series) equipped with a 10
ft 60/80 Carbopack B/5% Fluorocol packed column (Supelco)
and a flame ionization detector. The detection limit for all
products was 2 ppm. Hydrogen chloride was not quantified in
these experiments.
The catalyst pretreatment consisted of flowing He (30 mL/
min) for 5 min at 303 K. The He flow was continued as the
catalyst was heated to 403 K at a rate of 6.7 K/min and held at
403 K for 1 h. Then, the flow was switched to 20% H2 + 80%
He (50 mL/min) and the catalyst was heated to 493 K at a rate
of 6.7 K/min and held at 493 K for 1.5 h. The catalyst was
then cooled in He (40 mL/min) to the reaction temperature.
The reaction was conducted at 473 K and atmospheric
pressure. The total flow rate of the reaction mixture was 41
mL/min and consisted of 7000 ppm CH2ClCH2Cl, 36,600 ppm
H2, and a balance of He. For some experiments, 1450 ppm of
He was replaced with an equal volume of CO. The conversion
was maintained in the differential range between 3 and 4%.
Approximately 200 mg of catalyst were necessary to achieve
this conversion, given the described flow rates and reaction
temperature.
The infrared spectra were recorded with a Research Series II
FTIR spectrometer (Mattson, Inc.) equipped with a liquid N2
cooled MCT detector. The IR cell was similar to that described
elsewhere.20 The cell volume was 200 cm3 and the light path
length was approximately 15 cm. The cell was equipped with
glass stopcocks connected to gas inlet/outlet ports. The spectra
of adsorbed CO were measured with a resolution of 2 or 4 cm-1
,
and 400 scans were accumulated per spectrum. The spectra of
gaseous products of the reaction of 1,2-dichloroethane with H2
were measured with a resolution of 2 cm-1, and 128 scans were
accumulated per spectrum.
The infrared spectra were collected in the transmission mode,
which mandates the use of thin wafers of the catalyst sample.
The self-supporting catalyst wafers (∼20 mg/cm2 thick) were
prepared by powdering the catalyst material in an agate mortar
and then by pressing the powder at 830 atm for 3 min. Such
catalyst wafers are hereinafter referred to as “fresh.” The wafers
were pretreated in situ by evacuating at 403 K for 1 h, then
heating in flowing 5% H2 + 95% He (both Praxair, 99.999+%)
mixture (80 mL/min) to 493 K at 5 K/min and holding at this
temperature for 1 h. Then, the gas phase was evacuated at 493
K to a pressure of 1 × 10-5 Torr.
The dipole-dipole coupling frequency shifts and singleton
frequencies of CO adsorbed on the catalyst were measured by
means of a modified isotopic dilution method21 using mixtures
of 13C18O (Isotec, 99+% 13C, 95+% 18O) and 12C16O (Praxair,
99.99+%) of different compositions. Initially, the mixture of
known composition was adsorbed on the pretreated sample at
room temperature and an equilibrium pressure of 12 Torr. After
20 min the spectrum of adsorbed CO in the presence of gas
phase was recorded, then the gaseous CO was evacuated at the
same temperature for 20 min, and the spectrum of adsorbed
CO was recorded in the absence of gas phase. By repeating the
experiment while systematically changing the composition of
the isotopic mixture, the spectra of adsorbed CO as a function
of 13C18O + 12C16O mixture composition were determined.
Because the number of gaseous CO molecules in the IR cell
significantly exceeded the number of Pt atoms in the wafer, it
was assumed that the composition of adsorbed layer was similar
to that of gaseous 13C18O + 12C16O mixture added. Thus, the
ratio of isotopes in the gas phase was used to approximate that
on the catalyst surface. The singleton frequencies for linearly
adsorbed 12C16O and 13C18O on Pt were the band position for
the corresponding adsorbed CO at its minimal concentration in
the 12C16O + 13C18O gaseous mixture. The dipole-dipole
coupling shifts were the difference between the band position
of adsorbed carbon monoxide after evacuation at maximal
concentration of 12C16O (or 13C18O) in the 12C16O + 13C18O
mixture and the corresponding singleton frequency.
Results
Non-Steady-State Kinetics. With a 5-fold excess of hydro-
gen, 1,2-dichloroethane was converted exclusively to ethane and
HCl with both the Pt and the Pt1Cu1 catalysts (Figure 1). For
the Pt1Cu3 catalysts, the reaction products were not only ethane
and HCl but also ethylene (Figure 2). The initial selectivity of
the Pt1Cu3 catalyst toward the formation of ethylene was 80%,
but it gradually decreased to 65% during the 40 min reaction.
The addition of ∼2 Torr of CO to the reaction mixture (50
Torr of H2 + 10 Torr of CH2Cl-CH2Cl for the Pt, and 35 Torr
of H2 + 7 Torr of CH2Cl-CH2Cl for the Pt1Cu1) decreased
the activity of the Pt and Pt1Cu1 catalysts by ∼50 and 20%,
respectively, in comparison to the results obtained when no CO
was added. However, ethane was still the only carbon-containing
product observed for both catalysts. The addition of CO
The kinetics of 1,2-dichloroethane conversion in the presence
of H2 was studied in both static and continuous flow reactors.
The static system was the IR cell described earlier and

5570 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Borovkov et al.
Figure 3. Dynamics of reaction product accumulation (a sum of ethane
and ethylene) in the CH2ClCH2Cl + H2 reaction catalyzed by Pt1Cu3/
SiO2 at 523 K in a static reactor without (∆) and with addition of 1.5
Torr CO (O). Initial pressures of CH2ClCH2Cl and H2 were 7.8 and
50.0 Torr, respectively.
Figure 1. Dynamics of ethane accumulation in the catalyzed CH2-
ClCH2Cl + H2 reaction at 523 K in a static reactor; (∆) Pt/SiO2, initial
pressures of CH2ClCH2Cl and H2 were 10.0 and 50.0 Torr, respectively;
(O) Pt1Cu1/SiO2, initial pressures of CH2ClCH2Cl and H2 were 7.0
and 35.0 Torr, respectively.
Figure 4. Dynamics of reaction product accumulation, C2H4 (∆) and
C2H6 (O), in the CH2ClCH2Cl + H2 reaction with CO addition in a
static reactor at 523 K catalyzed by Pt1Cu1/SiO2 sample that was used
in the experiment shown in Figure 1 followed by heating in 35 Torr of
CO at 473 K for 0.5 h. Initial pressures of CH2ClCH2Cl, H2, and CO
were 9.0, 45.0, and 1.5 Torr, respectively.
Figure 2. Dynamics of reaction product accumulation in the CH2-
ClCH2Cl + H2 reaction catalyzed by Pt1Cu3/SiO2 at 523 K in a static
reactor to form C2H4 (O) and C2H6 (∆). Initial pressures of CH2ClCH2-
Cl and H2 were 7.8 and 50.0 Torr, respectively.
increased the selectivity of Pt1Cu3 catalyst toward ethylene from
∼80 to 100% and this selectivity was independent of the 1,2-
dichloroethane conversion (not shown). With respect to the
activity of the Pt1Cu3 catalyst, the rate of conversion with and
without CO was approximately the same (Figure 3).
After the Pt1Cu1 catalyst that was used in the CH2ClCH2Cl
+ H2 reaction shown in Figure 1 was heated in 35 Torr of CO
at the temperature of 473 K for 0.5 h, its selectivity behavior
in the presence of CO was significantly different than that of
the fresh catalyst. The initial selectivity toward ethylene when
2 Torr of CO was present was ∼100% (Figure 4). After 20
min, the concentration of ethylene reached a maximum of 5
mol % and remained constant thereafter. The selectivity toward
ethane increased gradually from ∼0% at zero contact time to
70% after 55 min.
It is worth mentioning that the influence of CO on the activity
and selectivity of all the catalysts was reversible. Removing
CO from the reactant mixture resulted in the same kinetics
performance as that observed for catalysts that were not exposed
to CO.
Figure 5. Time-on-stream performance of the Pt1Cu1/SiO2 in the CH2-
ClCH2Cl + H2 reaction at 473 K in a continuous flow reactor with
and without CO in the stream; (O), ethylene, (∆), ethane, (]), ethyl
chloride, (0), conversion.
Differential Flow Kinetics. There was a substantial differ-
ence in the differential kinetics results obtained for the Pt1Cu1
catalyst when CO was added to the reactant stream and when
it was not (Figure 5). Without CO in the reactant stream, ethane
was the main product with selectivity in excess of 95%;
monochloroethane (5%) was also observed. When 1450 ppm
of CO was added into the reaction mixture, ethylene was the

H2-Assisted 1,2-Dichloroethane Dechlorination on Pt-Cu/SiO2
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5571
Figure 7. Dynamics of the exchange of 12C16O pre-adsorbed on the
reduced fresh Pt1Cu1/SiO2 with gaseous 13C18O at ambient temperature.
(A) Spectrum of adsorbed 12C16O in the presence of gas phase; (B) the
spectrum after evacuation at room temperature; (C)-(H) the spectra
of adsorbed CO with the degree of 12C16O exchange varying from 10
(C) to ∼100% (H).
Figure 6. Dynamics of the exchange of 12C16O pre-adsorbed on the
reduced Pt/SiO2 with gaseous 13C18O at ambient temperature. (A)
Spectrum of adsorbed 12C16O in the presence of gas phase; (B) spectrum
after evacuation at room temperature; (C)-(H) the spectra of adsorbed
CO with the degree of 12C16O exchange varying from 10 (C) to ∼100%
(H).
main product (∼90%) and ethane was the minor product. When
the CO was eliminated from the reactant stream, the initial
selectivity was restored. The activity, which decreased by 25%
upon addition of CO, also returned to its original level. The
carbon monoxide was not consumed in the reaction.
Spectral Measurements. With 10 Torr of 12C16O in the gas
phase, the spectrum of 12C16O adsorbed on a reduced fresh Pt
catalyst consisted of an intense absorption band at 2082 cm-1
and a weak band at 1840 cm-1 (Figure 6), corresponding to
linear and bridged modes, respectively.22,23 The evacuation of
the gas phase at room temperature resulted in a shift of the band
assigned to linearly bonded 12C16O from 2082 to 2075 cm-1 as
expected when the CO surface coverage decreases.24 The
admission of 13C18O + 12C16O to the Pt catalyst pre-exposed to
12C16O resulted in the partial replacement of adsorbed 12C16O
molecules by gaseous 13C18O (Figure 6). As the exchange of
adsorbed 12C16O for 13C18O progressed, the intensities of the
bands corresponding to the linear and bridging forms of
adsorbed 12C16O decreased and the band attributed to linearly
adsorbed 12C16O shifted from 2075 to 2040 cm-1. This shift
results from a weakening of dipole-dipole coupling between
adsorbed 12C16O molecules as the coverage decreased by
substitution with the heavier isotopic 13C18O molecules.18,21
Simultaneously, the vibrational bands attributed to 13C18O
adsorbed on metallic Pt in linear and bridged forms appeared
and increased in intensity. The bridging form of adsorbed 13C18O
is characterized by a low intensity band at 1750 cm-1. The band
attributed to 13C18O linearly adsorbed at a low degree of
exchange of 12C16O for 13C18O appeared at 1940 cm-1 and
shifted to 1975 cm-1 as the degree of exchange increased. Thus,
the singleton frequencies for linearly adsorbed 12C16O and
Figure 8. Dynamics of the exchange of 12C16O pre-adsorbed on the
Pt1Cu1/SiO2 that was used in the experiments depicted in the Figure
1 followed by heating the wafer in 35 Torr CO for 0.5 h at 473 K. (A)
Spectrum of adsorbed 12C16O in the presence of gas phase; (B) that
after evacuation at room temperature; (C), (D), and (E) the spectra of
adsorbed CO with the degree of 12C16O exchange of 25, 40, and 65%,
respectively.
band at 2135 cm-1 and the shift of the 2073 cm-1 band to 2065
cm-1. The disappearance of the band at 2135 cm-1 upon room-
temperature evacuation highlights thermal instability of the
complexes of CO with Cu. During the exchange of pre-adsorbed
12C16O for 13C18O, the band of 12C16O linearly adsorbed on Pt
decreased in intensity and shifted from 2065 to 2040 cm-1. At
a low degree of exchange, the band of linearly adsorbed 13C18O
on Pt appeared at 1945 cm-1 and with increasing degree of
exchange shifted to 1970 cm-1. Thus, similar to those for the
monometallic Pt catalyst, the singleton frequencies for linearly
adsorbed 12C16O and 13C18O on Pt of the PtCu1 catalyst were
2040 and 1945 cm-1, respectively; however, the dipole-dipole
13C18O on Pt were determined to be 2040 and 1940 cm-1
respectively. The dipole-dipole shifts for both isotopes were
the same, 35 cm-1
,
.
shift for both molecules was 25 cm-1
.
When the exchange experiment was conducted with the
reduced Pt1Cu1 catalyst, very different adsorption behavior was
observed than that for Pt sample (Figure 7). At a 12C16O
equilibrium pressure of 10 Torr, the spectrum of adsorbed
12C16O consisted of two absorption bands with maxima at 2073
and 2135 cm-1. These bands have been previously assigned to
12C16O adsorbed linearly on metallic Pt (2073 cm-1) and Cu
(2133 cm-1).18,24 Evacuation of the 12C16O at room temperature
for 15 min resulted in the disappearance of the high-frequency
After the reaction of CH2Cl-CH2Cl with H2 was performed
on the Pt1Cu1catalyst wafer (Figure 1), it was heated in 35 Torr
of CO at 473 K for 0.5 h. Figure 8 shows the spectrum evolution
of 12C16O and 13C18O adsorbed on this catalyst during the
exchange of pre-adsorbed 12C16O for 13C18O. In the presence
of gaseous 12C16O, the spectrum consisted of two bands at 2055
and 2135 cm-1. These bands are attributed to 12C16O molecules
linearly adsorbed on metallic Pt and Cu, respectively. As in
the case of the reduced fresh Pt1Cu1 catalyst, the high-frequency

5572 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Borovkov et al.
Figure 9. IR spectra of 12C16O adsorbed on reduced Pt1Cu3/SiO2 at
equilibrium pressures of 4.3 (A), 2.15 (B), 1.07 (C), 0.53 (D), 0.26
(E), 0.13 (F), 0.07 (G) Torr, and after evacuation of the sample with
pre-adsorbed CO at room temperature for 12 h (H). Rotational structure
seen in the spectra is a contribution from gas-phase CO.
Figure 10. IR spectra of CO adsorbed on catalyst Pt1Cu3/SiO2 in the
presence in the gas phase of 6.7 Torr CO (A) and of the mixture 6.7
Torr CO + 6.7 Torr CH2Cl-CH2Cl (B)
in the presence of H2.1-4,26-28 The first one, hydrodechlorination,
results in the formation of alkanes, whereas the second,
hydrogen-assisted dechlorination, leads to olefins. With 1,2-
dichloroethane, the primary reaction intermediate is most likely
the di-σ complex of ethylene that forms via the homolytic
dissociation of both C-Cl bonds of CH2Cl-CH2Cl.1,26 This
intermediate can either desorb into the gas phase as ethylene
(hydrogen-assisted dechlorination pathway) or react with ad-
sorbed H atoms to form ethane (hydrodechlorination pathway).
With little or no Cu in the catalyst, the hydrodechlorination
chemistry dominates. At 523 K in the presence of excess of
hydrogen, both monometallic Pt- and bimetallic Pt1Cu1 exclu-
sively converted CH2Cl-CH2Cl into HCl and ethane (Figure
1). The metallic Pt is responsible for the formation of this
hydrocarbon.1-3,28 Not surprisingly, the electronic state of Pt
in the Pt1Cu1 is similar to the monometallic Pt-sample, viz.,
the singleton frequencies (2040 cm-1) of the IR band of linearly
adsorbed 12C16O on metallic Pt were the same for both catalysts
(Figures 6, 7).
One may question whether the Pt and Cu are alloyed in the
reduced fresh Pt1Cu1 catalyst, because of the similarities in the
results for CO adsorption on Pt and the catalytic performance.
However, the different effect of CO in the reaction mixture on
the activity of the Pt and Pt1Cu1 catalysts in the CH2Cl-CH2-
Cl + H2 reaction suggests that Pt and Cu interact in the Pt1Cu1.
The formation of bimetallic Pt-Cu particles (probably, with
the surface highly enriched in Pt) is also evident from the
analyses of the spectra of CO adsorbed on Cu. There is a single
absorption band with a maximum at 2135 cm-1, independent
of CO surface coverage (Figure 7). The maximum frequency is
40 cm-1 higher than that for CO linearly adsorbed on metallic
Cu supported on alumina18 or silica.24 This peak cannot be
assigned to complexes of CO with Cu+ ions because those
species are stable in a vacuum at room temperature. Moreover,
after reduction of the Pt1Cu1 catalyst in hydrogen at temperature
as high as 773 K, conditions at which Cu+ is definitely
reduced,29 the spectrum of CO adsorbed on Cu is also
represented by a single band at ∼2135 cm-1 (not shown). Thus,
this band is attributed to complexes of CO with metallic Cu
modified by Pt.18,24,29
band at 2135 cm-1 disappeared after the sample was evacuated
at room temperature for 15 min. During the exchange of pre-
adsorbed 12C16O for 13C18O, the band maximum positions of
both linearly adsorbed 12C16O and 13C18O remained essentially
unchanged.
The spectra of 12C16O adsorbed at different equilibrium
pressures on the reduced fresh Pt1Cu3 catalyst are shown in
Figure 9. The band at 2030 cm-1 has been assigned to linearly
adsorbed CO on Pt, and the band at 2135 cm-1 has been
assigned to CO adsorbed on Cu.18,24 A sharp decrease in
intensity of the 2135 cm-1 band occurred when the CO
equilibrium pressure was decreased, demonstrating that the CO-
Cu species are unstable. Decreasing the CO equilibrium pressure
did not affect the positions of the band maxima for CO adsorbed
on both Pt and Cu. However, after evacuation of the sample
with pre-adsorbed CO for 12 h at room temperature, the 2030
cm-1 band intensity decreased sharply and its maximum shifted
2025 cm-1 (spectrum H in the Figure 9). Nevertheless, it is
more consistent to consider 2030 cm-1 as the singleton
frequency for linearly adsorbed CO on Pt of the Pt1Cu3 catalyst
because, unlike the other singleton frequency measurements,
the value of 2025 cm-1 was obtained for CO coverage
approaching zero, which is characteristic of the thermal de-
sorption method,25 different than that of the isotopic dilution.
It is worth noting that the position of the 12CO band on Pt when
mixtures of 12CO and 13CO of different compositions were
adsorbed on a low metal loading Pt-Cu catalyst with the Pt/
Cu atomic ratio of 0.3 was independent of the 12CO concentra-
tion in the 12CO + 13CO mixture and the band was also located
at 2030 cm-1 2
.
The spectra of CO adsorbed on Pt1Cu3 in the presence of
6.7 Torr of CO (spectrum A) and of the mixture 6.7 Torr CO
+ 6.7 Torr CH2Cl-CH2Cl (spectrum B) are shown in Figure
10. Each spectrum consisted of two bands with maxima at 2030
and 2136 cm-1 corresponding to CO linearly adsorbed on Pt
and Cu, respectively. The presence of CH2Cl-CH2Cl in the IR
cell did not affect the CO absorption band associated with Pt.
However, the addition of CH2Cl-CH2Cl resulted in the decrease
of the 2136 cm-1 band intensity.
The Pt in the Pt1Cu3 catalyst has a different electronic
signature than monometallic Pt or the Pt in the Pt1Cu1. The
Pt1Cu3 is characterized by a low singleton frequency of 2030
cm-1 and the absence of dipole-dipole shift for linearly
adsorbed CO (Figure 9). Such Pt modifications result most likely
from the alloying of Pt and Cu such that the surface Pt atoms
Discussion
There are two reaction pathways for the conversion of vicinal
alkyldichlorides by bimetallic Pt- and Pd-containing catalysts

H2-Assisted 1,2-Dichloroethane Dechlorination on Pt-Cu/SiO2
J. Phys. Chem. B, Vol. 107, No. 23, 2003 5573
are strongly diluted by atoms of Cu. Chandler and Pignolet
studied CO adsorbed on a cluster-derived, silica-supported Pt-
Cu catalyst with DRIFTS.30 They attributed the small CO
dipole-dipole interaction, with respect to that observed by
others for coimpregnated catalysts, to more complete alloying
in their bimetallic cluster-derived sample. Similarly, Grunert
and co-workers interpreted the smaller CO dipole-dipole
interactions obtained with ZSM-5 supported Pt-Cu catalysts
after repeated oxidation and reduction cycles in terms of alloy
formation and consequently, a diminished Pt ensemble size.31
In both these works, the change in the dipole-dipole interactions
of approximately 10 cm-1 was considered to be substantial. For
the present work, the monometallic Pt and the reduced fresh
Pt1Cu1 had dipole-dipole shifts of 35 and 25 cm-1, respec-
tively, and Pt1Cu3 displayed no dipole-dipole interaction
whatsoever.
dual site mechanism is certainly reasonable, considering the
difference in the heat of adsorption for CO on Pt and on Cu at
room temperature is 27 kcal/mol (46 and 19 kcal/mol for Pt
and Cu, respectively39). This difference is sufficiently great that
CO remains adsorbed strongly on Pt while it is removed from
Cu by simple evacuation.29 Moreover, under reaction conditions
CO adsorbed on Cu atoms can be substituted for CH2Cl-CH2-
Cl, as it was observed at room temperature (Figure 10). The
blocking Pt sites with CO forces 1,2-dichloroethane to chemisorb
and dissociate exclusively on Cu.40,41
The size of both Cu and Pt ensembles on the surface of the
catalysts is important for high ethylene selectivity in the CH2-
Cl-CH2Cl + H2 reaction. When the concentration of Cu on
the surface of Pt-Cu particles is low, as in the case of the
reduced fresh Pt1Cu1 catalyst, the Cu sites do not contribute
significantly to the conversion of 1,2-dichloroethane (Figure 1).
The reaction occurs mainly on Pt sites. If any reaction
intermediate does form on the limited number of Cu sites, the
difference in the heat of adsorption would drive the intermediate
to the Pt sites, on which hydrogenation occurs. The CO in the
reaction mixture substantially decreases activity by competing
strongly with CH2Cl-CH2Cl for the Pt sites.
The fact that the CO adsorbed on Pt in the Pt1Cu3 can be
removed by simple evacuation (Figure 9) at ambient temperature
provides further evidence that Pt and Cu in this catalyst
atomically mixed. For monometallic Pt, adsorbed CO is stable
in a vacuum at ambient temperature.32 Similar to the Pt1Cu1,
the band of CO adsorbed on Cu at 2135 cm-1 is independent
of CO surface coverage and is 40 cm-1 higher than for CO
linearly adsorbed on supported metallic Cu.18,24 Thus, it is
reasonable to conclude that the Pt and Cu in the Pt1Cu3 catalyst
are alloyed.
As the concentration of Cu on the surface of Pt-Cu particles
increases, the copper forms larger ensembles and its role in
improving ethylene selectivity becomes pronounced (Figures
2, 4, 5). The 1,2-dichloroethane cannot compete with CO for
the Pt adsorption sites (Figure 10), and it is “forced” to dissociate
on Cu; ethylene is thus formed on the Cu sites. If the Cu
ensemble is large enough, the surface diffusion of the ethylene
precursor will not result in immediate capturing it by a Pt site
followed by hydrogenation to ethane, and the precursor would
desorb into the gas phase as an ethylene molecule. Naturally
we are assuming the H2 can still dissociate on Pt when CO is
present, because Cl is not accumulating on the catalyst;42 the
catalyst is still active. In the integral reactor, the decrease in
the ethylene selectivity with time for the Pt1Cu3 without CO
(Figure 2) and for the CO pretreated Pt1Cu1 in the presence of
CO with conversion (Figure 4) is explained by ethylene
readsorption on Pt sites followed by hydrogenation. Apparently
ethylene can successfully compete with CO for Pt sites. The
It is also important to consider why preheating the Pt1Cu1
catalyst used in the CH2Cl-CH2Cl + H2 reaction in a CO
atmosphere at 473 K results in an increase in the singleton
frequency for linearly adsorbed CO on Pt to 2052 cm-1 (ligand
effect) and the disappearance of CO dipole-dipole interactions
(ensemble size effect) in comparison to the reduced fresh PtCu1
catalyst (Figure 8). The increase in the singleton frequency can
be attributed to the modification of electronic properties of Pt
by chemisorbed chlorine atoms formed during hydrodechlori-
nation of CH2Cl-CH2Cl or chemisorbed oxygen resulting from
CO dissociation on Pt at the temperature as high as 473 K.33,34
Because of their high electronegativity, chemisorbed chlorine
and oxygen atoms will increase the singleton frequency of
linearly adsorbed CO by withdrawing electron density from Pt.
Dissociation of CO on the used PtCu1 may cause an enrichment
of the catalyst surface in Cu because of the higher binding
energy of oxygen to Cu than to Pt.35-38 The Cu atoms together
with the carbon formed in the CO dissociation reaction may
split Pt ensembles, which results in the elimination of dipole-
dipole interaction between the CO molecules linearly adsorbed
on Pt in the Pt1Cu1 catalyst preheated in CO at 473 K.
heat of ethylene adsorption on Pt is within 22-30 kcal mol-1 43
,
as opposed to 46 kcal mol-1 for CO adsorbed on Pt.39 The fact
that the ethylene selectivity for the Pt1Cu3 catalyst with CO in
the reaction mixture does not depend on the 1,2-dichloroethane
conversion suggests that the Pt ensembles in this catalyst are
very small. It is well-known that the energy of ethylene
adsorption on Pt depends on the metal ensemble size.44-46
Apparently, ethylene cannot compete with CO for an adsorption
site containing less than four adjacent Pt atoms, the minimal
number to form di-σ-complex of ethylene42,45,46 with reasonably
high adsorption energy.
The idea that blocking, or limiting, the size of the Pt
ensembles suppresses hydrogenation activity is borne out from
the experiments when CO is added to the reaction stream. A
small amount of CO in the reaction mixture increases the
selectivity of the Pt1Cu3 catalyst toward ethylene, up to 100%,
at the expense of ethane. The used Pt1Cu1 catalyst preheated
in CO also selectively forms ethylene when CO is present in
the reaction mixture. Moreover, the kinetics of hydrocarbon
product accumulation shown in Figure 4 provides evidence that
ethylene is a primary product of the CH2Cl-CH2Cl dehaloge-
nation reaction. The reversible effect of CO addition on the
selectivity of both Pt1Cu1 and Pt1Cu3 in the conversion of 1,2-
dichloroethane suggests that there are two types of catalytic
sites: one type is responsible for the formation of ethylene (Cu)
and the other type of centers catalyzes hydrogenation chemistry
(Pt). The sites responsible for ethylene hydrogenation are
poisoned by the addition of CO into the reaction mixture. A
Thus, ensemble size effect plays a decisive role in determining
ethylene selectivity of Pt-Cu/SiO2 catalysts in the CH2Cl-
CH2Cl + H2 reaction. The ligand effect does not seem to
influence strongly the catalyst performance, because changes
of the Pt electronic state of opposite character accompany similar
changes in selectivity. The Pt in the Pt1Cu3 catalyst is electron-
rich, while the Pt after treatment of Pt1Cu1 in CO is electron-
poor compared to the monometallic Pt catalyst (Figures 6, 8,
9), yet both catalysts in the presence of CO in the reaction
mixture produce ethylene instead of ethane as the main product.
It is particularly difficult to visualize, then, how electronic effects
can be important to selectivity change.

5574 J. Phys. Chem. B, Vol. 107, No. 23, 2003
Borovkov et al.
(5) Sachtler, W. M. H. Vide. 1973, 164, 67.
Conclusion
(6) Sachtler, W. M. H.; van Santen, R. A. AdV. Catal. 1977, 26, 69.
(7) Peden, C. H. F.; Goodman, D. W. J. Catal. 1987, 104, 347.
(8) Dautzenberg, F. M.; Helle, J. N.; Biloen, P.; Sachtler, W. M. H. J.
Catal. 1980, 63, 119.
(9) Carter, J. L.; McVicker, G. B.; Weissman, J.; Kmak, W. S.; Sinfelt,
J. H. Appl. Catal. 1982, 3, 327.
(10) Anderson, J. A. J. Catal. 1993, 142, 153.
(11) Zanier-Szydlowski, N.; Moisson, B.; Blejean, F.; Didillon, B. Proc.
SPIEsInt. Soc. Opt. Eng. 1993, 2089, 410.
(12) De Menorval, L.; Chaqroune, A.; Coq, B.; Figueras, F. J. Chem.
Soc., Faraday Trans. 1997, 93, 3715.
Silica-supported monometallic Pt and bimetallic Pt-Cu
catalysts with a Pt-to-Cu atomic ratio of 1 (Pt1Cu1) and 0.33
(Pt1Cu3) exhibit different selectivity patterns when exposed to
1,2-dichloroethane and H2 at 473 K and atmospheric pressure.
The only product of the CH2Cl-CH2Cl + H2 reaction catalyzed
by Pt and Pt1Cu1 in a static reactor is ethane, whereas ethylene
forms over the Pt1Cu3 with selectivity of 65-80% depending
on the CH2Cl-CH2Cl conversion. Treatment of the Pt1Cu1
catalyst that was already used in the CH2Cl-CH2Cl + H2
reaction with CO at 473 K for 0.5 h enhances the ethylene
selectivity in the presence of a small amount CO in the reaction
mixture. The initial ethylene selectivity of the Pt1Cu1 in this
case is 100%, gradually decreasing to 30% with the CH2Cl-
CH2Cl conversion. The selectivity of the Pt1Cu3 in the presence
of CO is 100% independent of the conversion.
According to an IR study of CO adsorption on the ethylene
unselective Pt and Pt-Cu catalysts, the electronic state of Pt in
the Pt and Pt1Cu1 is similar with the singleton frequency for
linearly adsorbed CO on Pt being 2040 cm-1 for both catalysts.
The singleton frequency of CO adsorbed on Pt of the ethylene
selective Pt1Cu3 is 2030 cm-1, whereas that of the used Pt1Cu1
treated in CO (also ethylene selective) is 2055 cm-1. The
coadsorption of CO and CH2Cl-CH2Cl on the Pt1Cu3 at room
temperature showed that 1,2-dichloroethane can compete with
CO for adsorption sites on Cu but cannot compete for the sites
on Pt.
The experimental observations described in the present paper
are consistent with the idea that Pt and Cu in the reduced fresh
Pt1Cu1 form alloyed Pt-Cu particles significantly enriched with
Pt. The Pt and Cu in the used, CO-treated Pt1Cu1 and Pt1Cu3
catalysts are also alloyed and modified electronically. However,
the electronic modification of Pt and Cu does not correlate with
ethylene selectivity in the CH2Cl-CH2Cl + H2 reaction. Instead,
ensemble effect plays a decisive role in determining ethylene
selectivity of the Pt-Cu catalysts. Small ensembles of Pt with
less than four contiguous atoms surrounded with Cu are
suggested to be the active sites for the hydrogen-assisted 1,2-
dichloroethane dechlorination toward ethylene. Copper is the
site for C-Cl bond dissociation and Pt supplies dissociated H
atoms to clean the Cu of adsorbed Cl atoms regenerating
metallic sites.
(13) Chander, B. D.; Schabel, A. B.; Rubinstein, L. I.; Pignolet, L. H.
Chem. Ind. 1998, 75, 607.
(14) Crossley, A.; King, D. A. Surf. Sci. 1976, 58, 379.
(15) Crossley, A.; King, D. A. Surf. Sci. 1977, 68, 528.
(16) Crossley, A.; King, D. A. Surf. Sci. 1980, 95, 131.
(17) King, D. A. Springer Ser. Chem. Phys. 1980, 15, 179.
(18) Toolenaar, F. J. C. M.; Stoop, F.; Ponec, V. J. Catal. 1983, 82, 1.
(19) Zaikovskii, V. I.; Ryndin, Yu. A.; Koval’chuk, V. I.; Plyasova, L.
M.; Kuznetsov, B. N.; Yermakov, Yu. I. Kinet. Catal. 1981, 22, 340.
(20) Deshmukh, S. S.; Borovkov, V. Yu.; Kovalchuk, V. I.; d’Itri, J. L.
J. Phys. Chem. B 2000, 104, 1277.
(21) Hammaker, R. M.; Francis, S.; Eischens, R. P. Spectrochim. Acta
1965, 21, 1295.
(22) Sheppard, N.; Nguyen, T. T. AdV. Infrared Raman Spectrosc. 1978,
5, 67.
(23) De La Crus, C.; Sheppard, N. Spectrochim. Acta 1994, A50, 271.
(24) Toolenaar, F. J. C. M.; Reinalda, D.; Ponec, V. J. Catal. 1980, 64,
110.
(25) Primet, M. J. Catal. 1984, 88, 273.
(26) Heinrichs, B.; Delhez, P.; Schoebrechts, J.-P.; Pirard, J.-P. J. Catal.
1997, 172, 322.
(27) Heinrichs, B.; Schoebrechts, J.-P.; Pirard, J.-P. J. Catal. 2001, 200,
309.
(28) Rhodes, W. D.; La´za´r, K.; Kovalchuk, V. I.; d’Itri, J. L. J. Catal.
2002, 211, 173.
(29) Sokolova, N. A.; Barkova, A. P.; Furman, D. B.; Borovkov, V.
Yu.; Kazansky, V. B. Kinet. Katal. 1995, 36, 434.
(30) Chandler, B. D.; Pignolet, L. H. Catal. Today 2001, 65, 39.
(31) Shpiro, E. S.; Tkachenko, O. P.; Jaeger, N. I.; Schulz-Ekloff, G.;
Grunert, W. J. Phys. Chem. B 1998, 102, 3798.
(32) Ponec, V.; Bond, G. C. Catalysis by Metal and Alloys; Elsevier:
Amsterdam, 1995.
(33) Boudart, M. AdV. Catal. 1969, 20, 153.
(34) McCrea, K. R.; Parker, J.; Chen, P.; Somorjai, G. Abstr. Pap.s
Am. Chem. Soc. 2001, 221st COLL-183.
(35) Shustorovich, E.; Bell, A. T. Surf. Sci. 1992, 268, 397.
(36) Grimblot, J.; Luntz, A. C.; Fowler, D. E. J. Electron Spectrosc.
Relat. Phenom. 1990, 52, 161.
(37) Luntz, A. C.; Grimblot, J.; Fowler, D. E. Phys. ReV. B. 1989, 39,
12903.
(38) Anderson, J. R. Structure of Metallic Catalysts; Academic Press:
London, 1975.
(39) Somorjai, G. A. Introduction to Surface Chemistry and Catalysis;
Wiley: New York, 1994.
Acknowledgment. Financial support from the Department
of EnergysBasic Energy Sciences (DE-FG02-95ER14539) is
gratefully acknowledged.
(40) Walter, W. K.; Jones, R. G.; Waugh, K. S.; Bailey, S. Catal. Lett.
1994, 24, 333.
(41) Yang, M. X.; Sarcr, S.; Bent, B. E. Langmuir 1997, 13, 229.
(42) Linke, R.; Schneider, U.; Busse, H.; Becker, C.; Schroder, U.;
Castro, G. R.; Wandelt, K. Surf. Sci. 1994, 307, 407.
(43) Passos, F. B.; Schmal, M.; Vannice, M. A. J. Catal. 1996, 160,
118.
(44) Windham, R. G.; Koel, B. E.; Paffett, M. T. Langmuir 1988, 4,
1113.
(45) Paffett, M. T.; Gebhard, S. C.; Windham, R. G.; Koel, B. E. Surf.
Sci. 1989, 223, 449.
References and Notes
(1) Vadlamannati, L. S.; Kovalchuk, V. I.; d’Itri, J. L. Catal. Lett. 1999,
58, 173.
(2) Vadlamannati, L. S.; Luebke, D. R.; Kovalchuk, V. I.; d’Itri, J. L.
Stud. Surf. Sci. Catal. 2000, 130, 233.
(3) Luebke, D. R.; Vadlamannati, L. S.; Kovalchuk, V. I.; d’Itri, J. L.
Appl. Catal. B 2002, 35, 211.
(4) Ito, L. N.; Harley, D. A.; Holbrook, M. T.; Smith, D. D.; Murchison,
C. B.; Cisneros, M. D. Eur. Patent 0 640 574 A1, 1994.
(46) Cortright, R. D.; Dumesic, J. A. J. Catal. 1994, 148, 771.