Reaction Mechanism and Kinetics of Olefin Metathesis by Supported ReOx/Al2O3 Catalysts

Article abstract of DOI:10.1021/acscatal.5b02233

The self-metathesis of propylene by heterogeneous supported ReO_x/Al₂O₃ catalysts was investigated with in situ Raman spectroscopy, isotopic switch (D-C₃⁼ → H-C₃⁼), temperature-programmed surface reaction (TPSR) spectroscopy, and steady-state kinetic studies. The in situ Raman studies showed that two distinct surface ReO₄ sites are present on alumina and that the olefins preferentially interact with surface ReO₄ sites anchored at acidic surface sites of alumina (olefin adsorption: C₄⁼ > C₃⁼ > C₂⁼). The isotopic switch experiments demonstrate that surface Re?CH₃ and Re?CHCH₃ are present during propylene metathesis, with Re? representing activated surface rhenia sites. At low temperatures (<100 °C), the rate-determining step is adsorption of propylene on two adjacent surface sites (rate ≈ [C₃⁼][Re?]². At high temperatures (>100 °C), the rate-determining step is the recombination of two surface propylene molecules (rate ≈ [C₃⁼]²[Re?]). To a lesser extent, the recombination of surface Re?CH₃ and Re?CHCH₃ intermediates also contributes to self-metathesis of propylene at elevated reaction temperatures.

Full text of DOI:10.1021/acscatal.5b02233

Research Article
pubs.acs.org/acscatalysis
Reaction Mechanism and Kinetics of Olefin Metathesis by Supported
ReOx/Al2O3 Catalysts
Soe Lwin and Israel E. Wachs*
Operando Molecular Spectroscopy and Catalysis Laboratory, Department of Chemical and Biomolecular Engineering, Lehigh
University, Bethlehem, Pennsylvania 18015, United States
S
* Supporting Information
ABSTRACT: The self-metathesis of propylene by heteroge-
neous supported ReOx/Al2O3 catalysts was investigated with
=
in situ Raman spectroscopy, isotopic switch (D−C3 → H−
=
C3 ), temperature-programmed surface reaction (TPSR)
spectroscopy, and steady-state kinetic studies. The in situ
Raman studies showed that two distinct surface ReO4 sites are
present on alumina and that the olefins preferentially interact
=
=
=
with surface ReO4 sites anchored at acidic surface sites of alumina (olefin adsorption: C4 > C3 > C2 ). The isotopic switch
experiments demonstrate that surface Re*CH3 and Re*CHCH3 are present during propylene metathesis, with Re*
representing activated surface rhenia sites. At low temperatures (<100 °C), the rate-determining step is adsorption of propylene
on two adjacent surface sites (rate ≈ [C3 ][Re*]2. At high temperatures (>100 °C), the rate-determining step is the
=
= 2
recombination of two surface propylene molecules (rate ≈ [C3 ] [Re*]). To a lesser extent, the recombination of surface Re*
CH3 and Re*CHCH3 intermediates also contributes to self-metathesis of propylene at elevated reaction temperatures.
KEYWORDS: catalyst, supported, rhenia, Al2O3, metathesis, olefins, mechanism, kinetics
membered reaction intermediate. During steady-state ethylene
self-metathesis, Aldag et al. found that the reaction rate had a
1.4 reaction order with respect to the ethylene partial pressure,
consistent with more than one ethylene molecule participating
in the rate-determining step (rds) at elevated temperatures.9
Spinicci et al. concluded that desorption of propylene from
supported ReOx/Al2O3 catalysts follows second-order kinetics
with temperature-programmed desorption (TPD) studies.10
Some kinetic studies suggested that olefin metathesis proceeds
via the Langmuir−Hinshelwood kinetic model,8,11,12 but
Kapteijn et al. considered the L-H model inadequate and
developed a first-order kinetic model based on intermediate
carbene species (ReCH2 and ReCHCH3) with product
desorption being the rate-determining step.13 This mechanism
is more in line with Chauvin’s metallacyclobutane mechanism
that was proposed for homogeneous and supported organo-
metallic catalysts.14
The lack of agreement among researchers about the reaction
mechanism and kinetics of olefin metathesis by supported
ReOx/Al2O3 catalysts is due to the use of different catalysts
(variable BET surface areas and rhenia loadings) and reaction
conditions (different olefins, olefin partial pressures, and
reaction temperatures). The motivation of this study is to
resolve these issues by systematically applying in situ Raman
spectroscopy, temperature-programmed surface reaction
(TPSR) spectroscopy, and steady-state propylene self-meta-
1. INTRODUCTION
Olefin metathesis has been a sustainable reaction to produce
important chemicals for the past 50 years.1 Supported ReOx/
Al2O3 catalysts for olefin metathesis were originally discovered
by British Petroleum (BP) in 1964.2 The high reactivity and
selectivity of supported ReOx/Al2O3 catalysts under ambient
conditions allows performing metathesis of functionalized
olefins, which is assisted by promoters.3 The high price and
volatility of rhenium oxides limit the use of ReOx catalysts only
to production of fine chemicals,3 but recent patent activity
indicates a resurgence of interest in this catalytic system due to
a shortage of both linear and functionalized olefins.4
Although olefin metathesis by supported ReOx/Al2O3
catalysts has been investigated over the years, conflicting
reports still exist about the reaction mechanism and kinetics of
this catalytic reaction. Supported ReOx/Al2O3 catalysts can be
activated upon exposure to propylene and 2-butene at room
temperature, but activation by ethylene is a slow process or is
even not possible.5−7 Aldag et al., however, concluded from
self-metathesis of C2H3D, yielding both C2H4 and C2H2D2, that
ethylene can initiate metathesis by supported ReOx/Al2O3
catalysts at slightly higher temperatures (95 °C).8 By
combining steady-state kinetic studies with simultaneous
thermal gravimetric analysis (TGA) to determine the quantity
of adsorbed ethylene during the reaction, the metathesis
reaction rate was found to vary with the square of the adsorbed
ethylene.8 It was proposed that the ethylene metathesis reaction
involves two mobile adsorbates, resulting in second-order
kinetics with respect to the adsorbate concentration.8,9 This
suggests that ethylene self-metathesis proceeds via a four-
Received: October 5, 2015
Revised: November 3, 2015
© XXXX American Chemical Society
272
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
thesis kinetic studies under various conditions to develop a
comprehensive model of olefin metathesis by supported ReOx/
Al2O3 catalysts.
3. RESULTS
3.1. In Situ Raman Spectroscopy. The in situ Raman
spectra of the supported 9.4% ReOx/Al2O3 catalyst exposed to
flowing Ar, C2 /Ar, C3 /Ar, and 2-C4 /Ar at 30 °C are shown
in Figure 1. The dehydrated catalyst in flowing Ar exhibits two
=
=
=
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. The supported ReOx/Al2O3 and
ReOx/TaOx/Al2O3 were prepared by impregnation of a 65−70 wt %
aqueous solution of perrhenic acid, HReO4 (Sigma-Aldrich) and a
toluene solution of tantalum ethoxide (Ta-(OC2H5)5, Alfa Aesar,
99.999%) onto the Al2O3 support (Engelhard batch no. H5433C).
The 9.4% ReOx/Al2O3 (Harshaw) catalyst was used for the in situ
Raman studies because lower fluorescence from this alumina support
gave rise to higher quality in situ Raman spectra at lower temperatures.
The reaction studies were performed with the 15.6% ReOx/Al2O3
(Engelhard) catalyst, since only a limited amount of the discontinued
Harshaw alumina was available. The percentage of Re above 6% would
not change in situ studies, since all of these catalysts would contain the
same two species and their interactions with olefins will not be
different. The full procedure of incipient-wetness impregnation,
drying, and calcination can be found in another paper.15
2.2. In Situ Raman Spectroscopy. The Raman spectra of the
supported ReOx/Al2O3 catalysts were obtained with a Horiba-Jobin
Ybon LabRam HR instrument equipped with three laser excitations
(532, 442, and 325 nm) and a liquid N2 cooled CCD detector
(Horiba-Jobin Yvon CCD-3000 V). The 442 nm laser was chosen,
since it minimized sample fluorescence. Spectral resolution was
approximately 1 cm−1, and the wavenumber calibration was checked
using the silica standard line at 520.7 cm−1. The lasers were focused on
the samples with a confocal microscope using a 50× objective
(Olympus BX-30-LWD). Typically, the spectra were collected at 30 s/
scan and 5 scans with a 200 μm hole. Approximately 5−25 mg of each
catalyst in powder form was loaded into an environmental cell
(Harrick, HVC-DR2) with a SiO2 window and O-ring seals which was
kept cool by flowing water. The catalysts were initially dehydrated at a
heating rate of 10 °C/min up to 500 °C and held for an hour under a
30 mL/min flow of 10% O2/Ar (Airgas, certified, 9.989% O2/Ar
balance). After cooling in Ar to 30 °C, a spectrum was taken before 1%
Figure 1. In situ Raman spectra of the supported 9.4% ReO4/Al2O3
catalyst after 60 min exposure to olefins at 30 °C.
bands from surface dioxo ReO4-I (∼1003 cm−1) and ReO4-II
(∼1013 cm−1) sites anchored at basic and acidic surface
hydroxyls of alumina, respectively.15 The interaction of the
olefins with the surface rhenia sites increases with olefin size
=
=
=
(C2 < C3 < C4 ), as reflected by the decreasing intensity of
the Raman ReO oxo bands. The decrease in intensity of the
Raman ReO oxo bands is from both bonding of the olefins
to the surface rhenia sites and the corresponding darkening of
the sample, with the latter dependent on the amount of
coordinated olefins. In all cases, olefins preferentially coordinate
to the surface ReO4-II sites, which show a greater decrease in
intensity upon olefin adsorption in comparison to the surface
RO4-I sites. Previously, it was shown that the surface ReO4-II
sites are the catalytically active sites and that catalysts with
100% of surface ReO4-II sites can be synthesized by promotion
with surface TaOx sites that block formation of inactive surface
ReO4-I sites (ReTaAl catalysts).15
=
=
=
C2 (ethylene), 1% C3 (propylene), or 1% C4 (trans-2-butene) in
argon or helium (balance) was introduced (three separate experi-
ments). After 60 min of each olefin flow at 30 °C, another spectrum
was taken.
2.3. Steady-State Kinetics Studies. The catalytic activity
measurements were performed in a fixed-bed catalytic reactor under
differential conditions (propylene conversion <15%). A separate
molecular sieve moisture trap was installed in the inlet propylene gas
line to purify the reactants. Both inlet and outlet gas lines were heated
using external electric heaters to ∼200 °C to prevent condensation of
the reactants and products. The catalysts were pretreated in 10% O2/
Ar at 500 °C for 30 min before cooling in Ar to either 70 or 150 °C.
Then a gas mixture of 1−10% C3H6/Ar was introduced at the rate of
∼100 mL/min. The desired concentration was achieved by diluting
the 10% C3H6/Ar gas with Ar. The products were analyzed using an
online gas chromatograph (Agilent GC 6890) equipped with flame
ionization (Agilent Serial No. USC250823H) and thermal con-
ductivity (Restek Product No. PC3533) detectors. Conversion was
normalized with propylene flow rate and catalyst weight to obtain
reactivity, reported in mmol/(g h).
2.4. Temperature-Programmed Surface Reaction (TPSR)
Spectroscopy. The temperature-programmed surface reaction
experiments were performed using an Altamira Instruments system
(AMI-200). The outlet gases were connected to an online Dymaxicon
Dycor mass spectrometer (DME200MS) for analysis. Typically
∼100−200 mg of catalyst was loaded into the U-tube reactor. The
cracking patterns were carefully adjusted with blank gas runs. The
supported 15.6% ReOx/Al2O3 (15.6ReAl) catalyst was used for the
following experiments. As in other experiments, the catalyst was always
initially pretreated with 10% O2/Ar at 500 °C for 30 min and then
cooled in Ar to the conditions of interest. Details of the specific
experimental protocols are given in the Supporting Information.
3.2. Steady-State Kinetics for Self-Metathesis of C3H6.
3.2.1. Reaction Order. The steady-state kinetics for self-
metathesis of C3H6 to C2H4 and C4H8 by the supported 15.6%
=
ReOx/Al2O3 catalyst was examined as a function of C3 partial
pressure and reaction temperature and are presented in Figure
2. The conversions of propylene were maintained below 15% to
maintain differential reaction conditions and emphasize the
forward metathesis reaction. The self-metathesis of propylene
=
reaction order in C3 partial pressure is a strong function of
temperature with ∼0.9 at 70 °C and increases to ∼1.8 at 150
°C. The propylene self-metathesis reaction products of C2= and
=
C4 were formed in equimolar amounts and also follow the
same reaction orders (see Figure S1 in the Supporting
Information).
3.2.2. Apparent Activation Energy. The apparent activation
energy for the steady-state self-metathesis of propylene was
found from an Arrhenius plot (see Figure S2 in the Supporting
Information). The apparent activation energy values below and
above 100 °C, corresponding to the different reaction order
regions, were found to be ∼7 and ∼21 kJ/mol, respectively.
273
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
Figure 2. Self-metathesis of propylene as a function of C3H6 partial
pressure and temperature. The reactivity was determined from
propylene molecules converted.
Figure 4. Reaction products from steady-state transient isotope kinetic
=
analysis (SSITKA): catalyst equilibration with D−C3 for 30 min
=
followed by switch to H−C3 .
3.2.3. Number of Surface ReOx Sites Involved in the Rate-
Determining Step (rds). The number of surface rhenia sites (n)
involved in the rds of the self-metathesis of C3H6 to C2H4 and
C4H8 can also be determined from the slope of the
ln(reactivity/(g h)) vs ln(ReO4-II) loading because the surface
ReO4-II catalytically active sites are completely dispersed on the
alumina support, giving the general relationship of rate ≈
[ReOx]n. Note that TaOx promotion blocks formation of the
inactive surface ReO4-I sites, allowing for a direct comparison of
the reaction rate and the number of active surface ReO4-II
sites.15 The ln(rate)−ln([ReOx]) plot presented in Figure 3 for
=
metathesis. Shutting off the flow of D−C3 causes an
=
=
instantaneous drop in the formation of D−C2 and D−C4 ,
indicating that the self-metathesis reaction is first order in D−
=
C3 partial pressure at 30 °C. Furthermore, the rate-
=
determining-step (rds) is the adsorption of D−C3 , since if
decomposition of a surface intermediate were the rds, then the
=
=
formation of D−C2 and D−C4 would lag the removal of the
D−C3= feed from the gas phase, which does not seem to occur.
=
=
Replacement of D−C3 with H−C3 initially causes a
perturbation in the flow that leads to a small increase in the
formation of D2CCH2 > CH2CH2 > CD3CDCH2. As
the H−C3= partial pressure increases, the transient D2CCH2,
CH3CHCD2, and CD3CDCH2 isotopomers are formed
with the exact same time-resolved pattern, suggesting the same
kinetic pattern (see Figure S3 in the Supporting Information).
The formation of CD3CDCHCH3 and CD3CDCDCD3,
however, was not observed. MS signals at m/z 29 and 31 were
also observed and are primarily from cracking of CH3CH
CH2 in the MS. The MS m/z 29 and 31 signals could also be
from CH2CHD and CHDCD2 reaction products, but the
time-dependent behavior of these signals followed that of
CH3CHCH2 and not of the labeled D2CCH2. Thus, there
is no evidence that significant CH2CHD and CHDCD2
ethylene products were formed. The ratio of products
=
containing CD2/CDCD3 during titration with H−C3 is
=
surprisingly ∼7, as indicated in Table 1. This suggests that the
concentration of surface CD2 intermediates is significantly
greater than that of surface CDCD3 intermediates during
Figure 3. ln−ln plot of reactivity (mmol/(g h) for C3 consumption)
vs ReO4-II loading for the ReTaAl catalyst that only contains the active
surface ReO4-II sites.15
=
self-metathesis of D−C3 at 30 °C.
=
Subsequent H−C3 -TPSR studies from 30 to 200 °C titrates
propylene self-metathesis yields a slope of ∼2.1, suggesting that
n = 2 and implying that two surface ReOx-II sites are involved
in the rds of the self-metathesis of C3H6 reaction.
3.3. Self-Metathesis of C3H6/C3D6. 3.3.1. Steady-State
Isotope Switch C3D6 → C3H6. The self-metathesis of propylene
the strongly bound surface intermediates, yielding the
isotopomers CH3CHCD2 > CD2CH2 > CD3CDCH2,
which are presented in Figure 5. The CD2/CDCD3
containing product ratio of ∼22 during TPSR (see Table 1)
further indicates that surface CDCD3 intermediates are a
minority surface intermediate. These findings suggest that
surface CD2 species are the most abundant reaction
=
was investigated by first equilibrating the catalyst with D−C3
=
for 30 min and then switching to a flow of H−C3 , as shown in
=
Figure 4. This steady-state isotope switch experiment provides
many key mechanistic and kinetic insights into the propylene
intermediates during D−C3 self-metathesis. The pronounced
=
reaction between H−C3 and strongly bound surface CD2
=
self-metathesis reaction. Before the isotope switch, D−C2 and
D−C4 were produced in a 1:1 ratio from D−C3 self-
above 200 °C indicates that this is a preferred reaction pathway
in this temperature range and is also reflected by the
=
=
274
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
Table 1. Number of Surface Intermediates Determined for
Titration Studies with the Supported 15.6% ReOx/Al2O3
a
Catalyst
b
b
adsorbate
titrant
reaction product monitored
30 °C
TPSR
=
=
D−C3
H−C3
CD3CDCH2
CD2CH2
0.2
0.4
1.0
0.0
0.4
2.1
7.3
0.0
CD2CHCH2
CD3CDCHCH3
a
Titration initially performed at 30 °C and continued with increasing
temperature during TPSR. Percentage of Re sites producing each
product.
b
Figure 6. Isotope-labeled reaction products from Ar/C3D6(30 °C)/
C3H6(100 °C)-TPSR.
=
Figure 5. Reaction products from H−C3 -TPSR after 30 min exposure
to D−C3= at 30 °C followed by 30 min exposure to H−C3= at 30 °C.
only a minor amount of CH2CH2 formed. For propylene, the
major isotopomer products were CD3CDCD2, CD3CD
CH2, and CH3CHCD2 with only minor amounts of
CH3CHCH2. For 2-butene, the major isotopomer products
were CD3CDCDCD3, CD3CDCHCH3, and CH3CH
CHCH3. All of the olefin products appear at the same reaction
temperature (Tp = 167 °C), reflecting that breaking of C−H/
C−D bonds does not take place during this olefin metathesis
surface reaction at elevated temperatures, since such a rds
would exhibit a kinetic isotope effect. This experiment nicely
demonstrates that the surface CD2/CH2 and CDCD3/
CHCH3 intermediates can readily undergo bimolecular
reactions to form the olefin isotopomer products in the
absence of gas-phase propylene.
=
simultaneous significant increase in formation of H−C3 . The
titration results in Table 1 also demonstrate that more D-
containing surface intermediates were titrated above room
temperature than at room temperature (factor of ∼6 (84%/
=
16%)). Between 30 and 200 °C, the formation rates of H−C2
and H−C4= are ∼1:1. Furthermore, the self-metathesis reaction
=
=
rate shows a kinetic isotope effect (KIE) of H−C3 /D−C3
≈
1.96. Although H−C/D−C bonds are not directly involved in
the self-metathesis reaction, surprisingly, they appear to have an
influence on the olefin metathesis transition state. Grubbs et al.
also observed such an H−C/D−C isotope effect in the early
stages of decadiene metathesis by the supported CoO-MoO3/
Al2O3 catalyst and attributed the isotope effect to formation of
a surface π-allyl intermediate during the initiation stage, which
The bimolecular surface kinetics for propylene self-meta-
thesis in the absence of gas-phase propylene was confirmed
with TPSR studies that varied the coverage of the surface
intermediates. After adsorption of C3= at 30 °C, the catalyst was
briefly preheated to 50, 75, 100, or 125 °C to desorb a portion
of the surface intermediates to vary their surface coverage. The
resulting TPSR spectra are shown in Figure 7 and reveal a
strong dependence of propylene formation kinetics upon the
surface coverage. The decreasing Tp value with increasing
surface coverage indicates second-order surface reaction
kinetics (see section S5 in the Supporting Information for a
detailed discussion and data analyses). The same exact trend
was also found for production of ethylene and 2-butene, with all
olefins appearing at the same Tp for a given coverage (see
Figure S5 in the Supporting Information). This TPSR
experiment reveals that surface CH2 and CHCH3
intermediates are also able to recombine by bimolecular
is the precursor to the active system.16
=
3.3.2. Self-Metathesis of Surface Intermediates (H−C3
+
=
D−C3 ). The self-metathesis of surface intermediates formed
during propylene metathesis was also examined in the absence
=
of gas-phase C3 . Self-metathesis of H3CCHCH2 was initially
conducted at 100 °C to equilibrate the catalyst with stable
surface intermediates, then the catalyst was cooled to room
temperature in flowing Ar, and finally CD3CDCD2 was
adsorbed at room temperature. Performing the TPSR experi-
ment in flowing Ar assured that gas-phase propylene would
minimally influence the surface reactions. Almost all
isotopomer metathesis products were formed at high temper-
ature, as shown in Figure 6. The D-containing products
=
dominated because D−C3 was adsorbed at 30 °C, which gave
a higher surface concentration of intermediates than initial
=
=
=
=
adsorption of H−C3 at 100 °C. The major ethylene
reactions to form C2 , C3 , and C4 metathesis reaction
products in the absence of gas-phase propylene.
isotopomer products were CD2CD2 and CH2CD2, with
275
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
propylene. This indicates that surface CD2/CH2 inter-
mediates are the most abundant reactive intermediates during
propylene self-metathesis by supported ReOx/Al2O3 catalysts.
The greater surface concentration of Re*CD2 in comparison
to Re*CDCD3 suggests that the reaction between surface
Re*CDCD3 and Re*CDCD3 intermediates is just barely
faster that the reaction between surface Re*CD2 + Re*
CD2 intermediates, as previously reported for olefins with
different lengths.18,19
4.3. Reaction Mechanism(s) and Kinetics. 4.3.1. First-
Order Kinetics at Low Temperatures. At low temperatures
(30−70 °C), the catalyst surface has a high concentration of
surface intermediates with the coverage of surface CH2 being
greater than that of surface CHCH3. The first-order steady-
state kinetics with regard to propylene partial pressure is a
consequence of propylene adsorption being the rds. This
=
=
conclusion is supported by the isotopic D−C3 → H−C3=
=
switch, showing the direct dependence of D−C2 and D−C4
Figure 7. Self-metathesis of C3H6 in flowing Ar after C3H6 adsorption
at 30 °C as a function of preheating temperature (50, 75, 100, and 125
°C) to control coverage of surface intermediates. The full experimental
procedure can be found in section S1 in the Supporting Information.
The C2H4 and C4H8 signals follow the same Tp values as C3H6 (see
Figure S5 in the Supporting Information).
=
formation on the D−C3 partial pressure (Figure 4) and the
direct dependence of CD2CH2, CH3CHCD2, and
=
CD3CDCH2 on the H−C3 partial pressure. In addition,
the reaction rate dependence on the surface rhenia sites varies
as ∼[Re*]2 (Figure 3), where Re* represents an activated,
coordinatively unsaturated surface rhenia site. This suggests
that the adsorption site consists of a dual site (requiring two
adjacent surface Re* sites). These insights support the
following forward reaction mechanism for the low-temperature
self-metathesis of propylene, which dominates at low propylene
conversions.
4. DISCUSSION
4.1. Nature of Catalytic Active Sites. In situ Raman
spectroscopy reveals that two distinct isolated dioxo surface
ReO4 sites (ReO4-I 1003 cm−1 and ReO4-II ∼1013 cm−1 in
Figure 1) are present on the Al2O3 support15 and that all olefins
preferentially coordinate to the initial surface ReO4-II sites. The
At steady-state and low temperatures, the surface is saturated
with surface CH2Re* and CH3CHRe* intermediates
resulting in
=
=
extent of interaction increases with olefin size (C4 > C3
>
=
C2 ). This suggests that only the initial surface ReO4-II sites
anchored to the acidic surface hydroxyls of the alumina support
become activated upon exposure to olefins and are the
catalytically active sites for olefin metathesis.15 The nature of
the activated surface rhenia sites is elaborated upon below.
4.2. Surface Reaction Intermediates. The in situ IR
spectra were previously presented and revealed that adsorbed
C3= π complexes are also present on the supported ReOx/Al2O3
catalyst. The in situ IR spectra could not provide any insights
about surface reaction intermediates because of the strong
contributions to the IR signal from a surface propylene π
complex and propylene adsorbed on the Al2O3 support.17 The
IR spectra, however, did demonstrate that the surface
intermediates react and/or desorb by ∼150 °C in the absence
of gas-phase propylene.
CH3−CHCH2 + CH2Re* + CH3CHRe*
→ CH2Re*CH2 + CH3CHRe*CHCH3
(1)
CH2Re*CH2 → Re* + CH2CH2
CH3CHRe*CHCH3 → Re* + CH3CHCHCH3
(2)
(3)
The molecular details of the surface intermediates in eqs 2
and 3, as well as the other reaction pathways, are not kinetically
significant since the rds is the adsorption step (eq 1).
Consequently, the reaction rate kinetic expression is given by
rate = krds[C3=][CH2Re*][CH3CHRe*]
Assuming that the surface concentrations of the surface
[CH2Re*CH2] and [CH3CHRe*CHCH3] inter-
mediates are small with the dominant surface intermediates
being the activated sites, then
(4)
=
=
The isotopic D−C3 → H−C3 switch experiment clearly
indicates that both surface CD2 and CDCD3 intermediates
=
are present on the catalyst during self-metathesis of D−C3
(Figure 4). Similarly, surface CH2 and CHCH3
intermediates are present on the catalyst from adsorption of
=
=
[Re*]0 = [CH2Re*] + [CH3CHRe*]
Assuming that the ratio of surface intermediates is constant
since propylene always reacts with each of the two surface
intermediates during the metathesis reaction according to eq 1
both H−C3 and D−C3 (Figure 6). The appearance of the
isotopomer olefin products at several different temperatures
during the isotopic switch-TPSR experiment suggests that
multiple surface CD2/CH2 and CDCD3/CHCH3
intermediates are present that differ in their bonding strength
to the catalytically active sites. The origin of the weakly and
strongly bound surface intermediates is currently not fully
understood: it is related to different surface rhenia oxidation
states or surface intermediates with different bonding strengths.
Surprisingly, the surface concentrations of the CD2/CH2
intermediates are significantly greater than those of the surface
CDCD3/CHCH3 intermediates during self-metathesis of
(5)
[CH2Re*]/[CH3CHRe*] = n (n > 1)
which leads to
(6)
=
2
=
2
rate = (krds/n)[C3 ][Re*]0 = k′rds[C3 ][Re*]0
(7)
The absence of first-order dependence on the number of
activate surface sites, rate ≈ [Re*]0, indicates that the
276
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
decomposition of the surface intermediates (eqs 2 and 3) is not
the rate-determining step at low temperatures.
propylene adsorption also recombine above ∼100 °C to form
all three olefin products:
4.3.2. Second-Order Kinetics at High Temperatures. The
steady-state kinetic studies reveal that the high-temperature
reaction pathway proceeds via a bimolecular mechanism
involving reaction between two surface species. The apparent
reaction order of 1.8 rather than 2 in propylene partial pressure
at elevated temperatures may be reflecting the minor
contribution of the first-order reaction kinetics at the higher
reaction temperature (see more below on this additional
pathway). The steady-state second-order reaction kinetics,
however, can only be derived by reaction between two
CH2Re* + CH2Re* ⇌ CH2CH2 + 2Re*
(15)
CH3CHRe* + CH3CHRe*
⇌ CH3CHCHCH3 + 2Re*
CH2Re* + CH3CHRe* ⇌ CH3CHCH2 + 2Re*
(16)
(17)
These surface bimolecular recombination pathways also
occur above 100 °C and probably are also contributing to the
overall reaction kinetics during steady-state propylene self-
metathesis at elevated temperatures. In the presence of gas-
=
adsorbed C3 molecules, since a bimolecular reaction between
surface Re*CH2 and surface Re*CHCH3 does not give
rise to second-order kinetics (see below). This suggests that the
surface intermediates present on the catalyst at high temper-
=
phase C3
=
C3H6 + 2Re* ⇌ CH2Re* + CH3CHRe*
(18)
ature are adsorbed C3 π complexes and that the high-
temperature reaction pathway proceeds via a four-centered
intermediate proposed in some of the early metathesis
studies.8,9,11,12 According to this bimolecular reaction mecha-
nism, the olefins can exchange their individual half-components
to yield the metathesis products. The kinetics for the second-
order self-metathesis of propylene are derived by assuming
Kads = [CH3CHRe*][CH2Re*]/[C3H6][Re*]2
(19)
surface site balance [*]0
= [Re*] + [CH2Re*] + [CH3CHRe*]
At high temperatures, surface intermediate coverage is low
and
(20)
=
reaction between two surface C3 π complexes:
=
=
C3 + Re* ⇌ Re*−C3
Kads
(8)
=
=
[Re*]0 ≈ [Re*]
(21)
Re*−C3 + Re*−C3 → 2Re* + CH2CH2
+ CH3CHCHCH3 krds
From adsorption equilibrium (eq 19)
[CH2Re*][CH3CHRe*] = Kads[C3H6][Re*]2
(9)
(22)
with the bimolecular reaction step eq 9 being the rds, which
gives the rate expression
rate = krds[CH2Re*][CH3CHRe*]
= 2
rate = krds[Re*C3 ]
From the equilibrium step eq 8
2
(10)
= krdsKads[C3H6][Re*]0
(23)
This bimolecular reaction involving recombination of surface
intermediates was earlier proposed by Spinicci et al. from TPD
studies of propylene desorption.10 The bimolecular recombi-
nation of surface intermediates cannot be dominant at elevated
temperatures, since it would give a reaction order of 1 with
respect to propylene partial pressure, which is much lower than
the measured reaction order of ∼1.8. The steady-state apparent
reaction order of slightly less than 2, ∼1.8, may reflect the
partial contribution of this surface bimolecular reaction pathway
at elevated temperatures.
The bimolecular recombination of surface intermediates for
olefin metathesis has been previously proposed for both
supported metal oxide heterogeneous and homogeneous
systems.8−12,20 The current kinetic study demonstrates that,
although this bimolecular reaction pathway does take place at
elevated temperatures for supported metal oxide catalysts, this
reaction pathway is not the dominant reaction pathway at
elevated temperatures.
=
=
[Re*C3 ] = Kads[Re*][C3 ]
From the surface site balance
(11)
(12)
=
[Re*]0 = [Re*] + [Re*C3 ]
with [Re*]0 representing all the activated surface sites, [Re*]
=
representing vacant sites, and [Re*C3 ] representing the
=
sites occupied by adsorbed C3 . At elevated temperatures, the
surface coverage of the surface intermediates is low, as
previously shown by in situ IR spectroscopy.17 Consequently
[Re*]0 = [Re*]
(13)
which gives
rate = krds{Kads[C3 ][Re*]0 }2
=
(14)
The second-order kinetics suggests that at high temperatures
the reaction involves two adsorbed C3 π complexes and
=
5. CONCLUSIONS
proceeds via a four-centered intermediate. Aldag et al.
previously proposed that the ethylene self-metathesis reaction
at elevated temperatures involves two mobile adsorbates that
form four C-membered surface intermediates, resulting in
second-order kinetics with respect to the adsorbate concen-
tration.8
4.3.3. Self-Metathesis of Surface Intermediates at High
Temperatures. The TPSR studies in flowing Ar reveal that the
surface CH2Re* and CH3CHRe* intermediates from
The olefins preferentially interact with the surface ReO4-II sites
on acidic sites of the alumina support. The surface Re*CH2
and Re*CHCH3 intermediates are present on activated
supported ReOx/Al2O3 catalysts, with the Re*CH2 inter-
mediate being dominant. The propylene metathesis reaction by
supported ReOx/Al2O3 catalysts occurs via multiple reaction
pathways that depend on reaction temperature. At low
temperatures (<100 °C), the catalyst is saturated with surface
277
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278

ACS Catalysis
Research Article
Re*CH2 and Re*CHCH3 intermediates and the rds
involves adsorption/reaction of gas-phase propylene with the
=
surface intermediates (rate ≈ [C3 ]). At elevated temperatures
(>100 °C), the catalyst exhibits low coverage of surface
intermediates and the rds proceeds via recombination of two
=
= 2
molecularly adsorbed C3 π complexes (rate ≈ [C3 ] ). To a
lesser extent, bimolecular recombination of surface Re*CH2
and Re*CHCH3 intermediates also contributes to propylene
= 1
self-metathesis at elevated temperatures (rate ≈ [C3 ] ). The
presence of multiple reaction mechanisms at different temper-
atures accounts for the variation of the reaction orders in
propylene partial pressure with reaction temperature.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b02233.
■
S
Additional steady state and TPSR graphs (Figures S1−
S5) with detailed analysis (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail for I.E.W.: iew0@lehigh.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
S.L. and I.E.W. acknowledge financial support from U.S. DOE
Basic Energy Sciences (Grant No. FG02-93ER14350).
■
REFERENCES
■
(1) Mol, J. C. J. Mol. Catal. A: Chem. 2004, 213, 39−45.
(2) Disproportionation catalyst. British Petroleum Corporation.
Great Britain Patent GB1054864 (A), September 8, 1964.
(3) Mol, J. C.; van Leeuwen, P. W. N. M. Metathesis of Alkenes.
Handbook of Heterogeneous Catalysis; Wiley-VCH: New York, 2008;
Vol. 14, pp 3240−3256.
(4) Lwin, S.; Wachs, I. E. ACS Catal. 2014, 4, 2505−2520.
(5) Salameh, A.; Coperet, C.; Basset, J.; Bohm, V. P. W.; Roper, M.
Adv. Synth. Catal. 2007, 349, 238−242.
(6) McCoy, J. R.; Farona, M. F. J. Mol. Catal. 1991, 66, 51−58.
(7) Olsthoorn, A. A.; Boelhouwer, C. J. Catal. 1976, 44, 207−216.
(8) Aldag, A. W.; Lin, C. J.; Clark, A. J. Catal. 1978, 51, 278−285.
(9) Lin, C. J.; Aldag, A. W.; Clark, A. J. Catal. 1976, 45, 287−296.
(10) Spinicci, R.; Tofanari, A. J. Therm. Anal. 1983, 27, 391−400.
(11) Aldag, A. W.; Clark, A. J. Catal. 1978, 54, 98−101.
(12) Luckner, R. C.; McConchle, G. E.; Wills, G. E. J. Catal. 1973,
28, 83−91.
(13) Kapteijn, F.; Bredt, L. H. G.; Homburg, E.; Mol, J. C. Ind. Eng.
Chem. Prod. Res. Dev. 1981, 20, 457−466.
(14) Her
176.
́
isson, J.-L.; Chauvin, Y. Makromol. Chem. 1971, 141, 161−
(15) Lwin, S.; Keturakis, C. J.; Handzlik, J.; Sautet, P.; Li, Y.; Frenkel,
A. I.; Wachs, I. E. ACS Catal. 2015, 5, 1432−1444.
(16) Grubbs, R. H.; Swetnick, S. J. J. Mol. Catal. 1980, 8, 25−36.
(17) Lwin, S.; Li, Y.; Frenkel, A.; Wachs, I. E. ACS Catal. 2015, 5,
6807−6814.
(18) Szeto, K. C.; Mazoyer, E.; Merle, N.; Norsic, S.; Basset, J.;
Nicholas, C. P.; Taoufik, M. ACS Catal. 2013, 3, 2162−2168.
(19) Spronk, R.; Mol, J. C. Appl. Catal. 1991, 70, 295−306.
(20) Hughes, W. B. J. Am. Chem. Soc. 1970, 92, 532−537.
278
DOI: 10.1021/acscatal.5b02233
ACS Catal. 2016, 6, 272−278