Acid site densities and reactivity of oxygen-modified transition metal carbide catalysts

Article abstract of DOI:10.1016/j.jcat.2016.09.012

Acidic properties of β-Mo₂C, α-Mo₂C, W₂C, and WC were quantified by assessing the kinetics of isopropanol (IPA) dehydration at 415?K either (i) under inert He/Ar atmosphere or (ii) with 13?kPa O₂ co-feed. Dehydration kinetics were zero-order with respect to IPA for all catalysts and under all reaction conditions. Intrinsic activation energies were similar across all catalysts (89–104?kJ?mol^?1). Acid site densities calculated via in situ 2,6-di-tert-butylpyridine (DTBP) titration were used to normalize dehydration turnover frequencies (TOF). O₂ co-feed increased dehydration rates per gram by an order of magnitude for all catalysts tested, but TOF remained invariant within a factor of ～2. Mo- and W-based carbides showed similar dehydration kinetics regardless of O₂ co-feed, and O₂ co-feed did not alter bulk carbidic structure as noted by X-ray diffraction. Br?nsted acid site provenance results from the oxophilicity of Mo and W carbides.

Full text of DOI:10.1016/j.jcat.2016.09.012

Journal of Catalysis 344 (2016) 53–58
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Acid site densities and reactivity of oxygen-modified transition metal
carbide catalysts
⇑
Mark M. Sullivan, Aditya Bhan
Department of Chemical Engineering and Materials Science, University of Minnesota - Twin Cities, 421 Washington Avenue SE, Minneapolis, MN 55455, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Acidic properties of b-Mo2C,
a-Mo2C, W2C, and WC were quantified by assessing the kinetics of
Received 24 June 2016
Revised 26 August 2016
Accepted 11 September 2016
isopropanol (IPA) dehydration at 415 K either (i) under inert He/Ar atmosphere or (ii) with 13 kPa O2
co-feed. Dehydration kinetics were zero-order with respect to IPA for all catalysts and under all reaction
conditions. Intrinsic activation energies were similar across all catalysts (89–104 kJ molÀ1). Acid site den-
sities calculated via in situ 2,6-di-tert-butylpyridine (DTBP) titration were used to normalize dehydration
turnover frequencies (TOF). O2 co-feed increased dehydration rates per gram by an order of magnitude for
all catalysts tested, but TOF remained invariant within a factor of ꢀ2. Mo- and W-based carbides showed
similar dehydration kinetics regardless of O2 co-feed, and O2 co-feed did not alter bulk carbidic structure
as noted by X-ray diffraction. Brønsted acid site provenance results from the oxophilicity of Mo and W
carbides.
Keywords:
Brønsted acid
Dehydration
Molybdenum carbide
Tungsten carbide
Interstitial carbide
Catalysis
Ó 2016 Elsevier Inc. All rights reserved.
Site density
Oxophilic
Transition metal
1. Introduction
mechanism in addition to ex situ NH3 temperature programmed
desorption (TPD) studies; these acid sites were shown to be similar
Multifunctional interstitial transition metal carbide catalysts
have been shown to catalyze hydrogenolysis [1–3], hydrogenation
[2,4–7], isomerization [8–12], and hydrodeoxygenation (HDO)
[13–17] reactions. The diverse functionalities of transition metal
carbide bulk formulations arise from different catalytically active
sites generated by modification of the catalyst surface. Ribeiro
et al. [18,19] reported a bifunctional metal-acid surface of an
oxygen-modified tungsten carbide formulation that catalyzed
isomerization of C5-C7 alkanes at atmospheric pressure and 430–
630 K via sequential dehydrogenation/methyl shift/hydrogenation
pathways as evinced by terminal isotopic enrichment in n-
to acid sites on WOx catalysts because comparable product distri-
butions of 3,3-dimethylpentane isomerization over O⁄-WC and
Al2O3-supported Pt/WO3 catalysts were observed. Isomerization
site time yields, normalized by ex situ irreversible CO chemisorp-
tion uptakes, were similar over 700-K oxygen-modified WC and
W2C catalysts (244 Â 10À3 vs. 125 Â 10À3 sÀ1) [8,18,19]. Lamic
and coworkers [20] also demonstrated isomerization of
n-heptane over both bifunctional WCx and mixed Mo2C/WO2 cata-
lysts, hypothesizing that the carbidic Mo2C or WCx acts as the
de/hydrogenation catalyst and the WO2 provides the Brønsted acid
site necessary for the isomerization of n-heptene via a carbenium
ion intermediate.
heptane-1-13
C isomerization reactions in contrast to the C5
hydrogenolysis mechanism prevalent over noble metals that
would also yield isotopic enrichment of interior carbon atoms.
WC and W2C catalysts were oxygen-modified via slow O2 introduc-
Iglesia and coworkers [21,22] demonstrated the presence of
both Brønsted and Lewis acid sites on supported WOx catalysts that
were active for both isomerization and dehydration through the
use of in situ pyridine and 2,6-di-tert-butylpyridine (DTBP) titra-
tions. Ruddy and coworkers [23,24] demonstrated the presence
tion (0.1 l
mol sÀ1 gÀ1) at room temperature followed by heating
under 20 kPa O2 to 300–800 K, and catalysts were finally pre-
treated under H2 flow at 673 K for 1–2 h prior to exposure to reac-
tion conditions (95 kPa H2). The presence of Brønsted acid sites on
O⁄-WCx catalysts was implicated by the methyl shift isomerization
of acid sites on bulk b-Mo2C and on SBA-15-supported
a-MoC1Àx
nanoparticles via both NH3 temperature programmed desorption
(TPD) studies and density functional theory (DFT) calculations that
showed comparable NH3 binding energies between O⁄-MoCx and
proton-form zeolites. Although acid sites have been reported to
exist across transition metal oxides and oxygen-modified
⇑
Corresponding author.
E-mail addresses: sulli931@umn.edu (M.M. Sullivan), abhan@umn.edu (A. Bhan).
http://dx.doi.org/10.1016/j.jcat.2016.09.012
0021-9517/Ó 2016 Elsevier Inc. All rights reserved.

54
M.M. Sullivan, A. Bhan / Journal of Catalysis 344 (2016) 53–58
transition metal carbides, quantification and comparison of acid
site densities and strengths are lacking, and similarities between
oxidic and carbidic acid sites have yet to be characterized.
We have previously investigated the acidic and metallic charac-
teristics of molybdenum carbide catalytic formulations during
HDO of biomass-derived oxygen-containing molecules including
anisole, furfural, m-cresol, and acetone [25–31]. Metallic sites
responsible for anisole, furfural, and m-cresol HDO can be selec-
tively titrated using CO, and Brønsted acid sites responsible for
the rate-determining isopropanol (IPA) dehydration during the
sequential hydrogenation/dehydration of acetone could be selec-
tively titrated using DTBP. These in situ titrations allow for site
density and TOF measurements of the catalytically active sites
without ex situ methods, yielding both the identity and quantity
of active sites.
in 40 mL deionized H2O, and the mixture was homogenized and
magnetically stirred at ambient temperature for 0.25 h. The mix-
ture was then placed in an oven at 393 K for 24 h. A dark gray por-
ous solid mass was formed; the resulting solid was crushed and
used as a precursor. Approximately 2 g of this precursor was trea-
ted under flowing He, heated at 0.25 K sÀ1 to 1073 K, held at
1073 K for 0.5 h, and then cooled to ambient temperature under
the same flow. This product was shown to have bulk a-Mo2C crys-
tal structure (see discussion in Section 3 below). Approximately
0.2 g of this precursor was then heated under flowing He to
415 K for IPA dehydration measurements and was subsequently
treated with 13 kPa O2 co-feed under reaction conditions as dehy-
dration rates rose until they reached a stable value.
WC and W2C catalysts were synthesized according to a modi-
fied method of Delannoy and coworkers [35]. Both catalysts were
synthesized with 0.03–0.1 g WO3 precursor (Sigma Aldrich,
99.995% trace metal basis) using the quartz tube flow system
described above under 1.82 cm3 sÀ1 total flow of 20% CH4/H2 and
heated at 0.0167 K sÀ1 until their final synthesis temperature
(1073 K for WC, 903 K for W2C), and finally held isothermally for
8 h. CH4 gas flow was then removed, and the catalyst was allowed
to cool to 415 K under H2 flow.
In this work, we use IPA dehydration activation energy, TOF,
and site density measurements as probes of acid sites on freshly
synthesized as well as oxygen-modified (O⁄) transition metal car-
bide catalysts of molybdenum and tungsten (a-Mo2C, b-Mo2C,
W2C, and WC). Catalysts demonstrated extreme oxophilicity as
noted by zero-order kinetics with respect to IPA pressure, indica-
tive of a catalyst surface saturated with oxygenate. Intrinsic acti-
vation energies and dehydration turnover frequencies (TOF) were
calculated using acid site densities measured via in situ DTBP
titration. Selective DTBP titration both directly demonstrates the
presence of Brønsted acid sites on O⁄-modified carbides and
allows for calculation of intrinsic dehydration TOF values. Acid
site densities increased with O2 co-feed-induced surface oxida-
tion, but the nature of the active sites, as quantified via intrinsic
activation energies and TOF, was similar across all measured cat-
alysts regardless of bulk carbide structure, the presence of O2 co-
feed, or bulk transition metal (Mo or W). The nature of the acid
sites on TPR-synthesized transition metal carbides is uniform
across formulations oxidized by both reactant alcohol O⁄ incorpo-
ration and direct O2 co-feed at reaction conditions. The persistent
and consistent nature of acid site identity with O⁄ source, site
density, transition metal, and bulk crystal structure serves to fur-
ther the understanding of catalytic applications of oxophilic tran-
sition metal carbides or other oxophilic catalytic materials from
both a fundamental perspective and for any potential acid catal-
ysis/HDO applications.
Catalysts were passivated after reaction by treatment in flowing
1% O2/He (Matheson, Certified Standard Purity) at 1.0 cm3 sÀ1 at
ambient temperature for at least 0.75 h in an effort to avoid violent
bulk oxidation with atmospheric O2 [36,37]. Samples were passi-
vated even if treated with O2 co-feed during reaction.
2.2. Catalyst characterization
The mass fraction of molybdenum in the a-Mo2C was analyzed
by inductively coupled plasma optical emission spectroscopy
(ICP-OES); experimental details are provided in Section S.1 of
Supplementary Information.
X-ray diffraction (XRD) was performed with a Bruker D8
Discover 2D X-ray diffractometer equipped with a 2-D VÅNTEC-
500 detector and a 0.8-mm collimator. CuK and K radiation
a
1
a2
was used in conjunction with a graphite monochromator. Scans
were performed in two or three frames for 300 s per frame. Total
scan range was typically 2h = 20–80° with the scan frames being
adjusted so as to center the primary intensity peaks in each frame.
The 2D scans were converted to 1D intensity vs. 2h with a step size
of 0.04° 2h and merged for analysis. A zero background holder was
used with a small amount of vacuum grease for sample support.
Surface area and porosity of the passivated catalyst samples
were measured using an ASAP Micromeritics 2020 analyzer. BET
surface area and porosity measurements using N2 were performed
at liquid nitrogen boiling temperature. Prior to N2 physisorption
2. Experimental
2.1. Catalyst synthesis and passivation
All catalysts were prepared in a quartz tubular reactor (I.D.
10 mm). Catalyst precursors were heated either in a three-zone
or in a one-zone split tube furnace (Applied Test Systems) under
gas flows including CH4 (Matheson, 99.97%), H2 (Matheson,
99.999%), or He (Minneapolis Oxygen, 99.997%).
Synthesis of orthorhombic b-Mo2C was carried out using a
temperature-programmed reaction method reported previously
[25,32,33]. Batches of catalyst were synthesized using 0.03–
measurements, samples were degassed to <6 lm Hg and heated
to 523 K at 0.17 K sÀ1 and held for 2–4 h. BET Surface areas of
oxygen-modified carbide samples were typically <5 m2 gÀ1, and
surface areas of post-reaction samples without O2 co-feed were
typically ꢀ20–60 m2 gÀ1; surface areas were not used for quantita-
tive comparison due to their evolution with O2 exposure [27] or
aging time [28].
1.0 g ammonium paramolybdate (sieved, 177–400 lm, (NH4)6-
Mo7O24Á4H2O, Sigma, 99.98%, trace metal basis). The paramolyb-
date was heated at 0.06 K sÀ1 to 623 K and held at 623 K for 12 h
under total flow of 3.0 cm3 sÀ1 of 15 vol% CH4/H2, followed by a
temperature ramp at 0.046 K sÀ1 from 623 K to 873 K and a tem-
perature hold at 873 K for 2 h. The reactor was then cooled
under the same gas flow to reaction temperature (415 K) to
begin reaction.
Cubic a-Mo2C (also labeled as a-MoC1Àx) [24] was synthesized
according to a modified method of Vitale and coworkers [34].
Sucrose (5 g) and ammonium paramolybdate (5 g) were dissolved
2.3. Reaction and titration methods
Reactions were carried out in an apparatus described previously
[32]. Briefly, reactants were fed to stainless steel reactor lines via
either an M6 Valco syringe-free liquid handling pump or a KD
Scientific syringe pump (KDS120). Reactant lines were heated by
resistive heating tape to prevent condensation. 2-Propanol (ACS
99.97%), argon (Matheson, 99.999%), and helium (Minneapolis
Oxygen, 99.997%) were used as received. Products and reactants
at the reactor effluent were quantified using a flame ionization

M.M. Sullivan, A. Bhan / Journal of Catalysis 344 (2016) 53–58
55
detector (FID) with an online gas chromatograph (Agilent 7890
GC); gas-phase separation prior to the FID was accomplished with
a J&W HP-1 dimethylpolysiloxane column (50.0 m length, 320 lm
out any exposure to oxygen. Rates are normalized by the total
moles of metal atoms in the oxide precursor. Rates were measured
in the absence of heat and mass transfer limitations (Section S.2 of
Supplementary Information). Dehydration rates typically varied
<10% over the course of a ꢀ8 h experiment to measure IPA pressure
dependence, activation energy, and TOF. Fresh carbides that had
not been exposed to O2 co-feed exhibited increasing dehydration
rates at early times on stream, hypothesized to be attributable to
surface oxidation from deposited oxygen during an initial transient
in IPA catalytic dehydration. Propylene formation rates over all car-
bides exhibit zero-order dependence upon IPA pressure over two
orders of magnitude (ꢀ0.044–4.4 kPa). The zero-order dependence
is in line with the expected oxophilicity of these carbides; the cat-
alyst surface is saturated with IPA even at low reactant partial
pressures. Propylene was the primary product formed. Acetone
was formed in small quantities (<5% carbon selectivity) over fresh
catalysts and formed negligibly once O2 co-feed was introduced.
Di-isopropylether (DIPE) was formed with typically <10% carbon
selectivity. DIPE formation rates showed a positive order depen-
dence upon IPA pressure. DIPE formation rates were not used for
characterization and quantification of acid sites on the catalyst in
this study due to the nonzero order kinetics and the low product
formation rates. A linear dependence of carbon conversion on con-
tact time evinced an absence of product inhibition [32].
I.D., 0.52 lm film). Helium was used as the column carrier gas at a
flow rate of 0.15 cm3 sÀ1. GC oven temperature profiles for reactor
effluent separation typically began at 213 K for 0.058 h and were
ramped to 423 K at 0.42 K sÀ1
.
IPA partial pressure was varied from ꢀ0.044 to 4.4 kPa by alter-
ing liquid flow rates from ꢀ10 to 800
l
L hÀ1. For temperature vari-
ation experiments, reactor temperature was measured at the
moment of sample injection into the GC with reactor temperature
measured by a thermocouple in a thermal well on the external sur-
face of the quartz reactor. Temperatures were raised and lowered
ꢀ15 K from the standard reaction temperature of 415 K; repeats
of rates at multiple temperatures were performed over the course
of the experiment to ensure consistency and lack of catalyst history
effects.
Titration experiments were performed with liquid mixtures of
IPA and 2,6-di-tert-butylpyridine (DTBP). After catalyst synthesis
and possible O2 co-feed introduction, IPA was introduced and
steady state propylene production rates were obtained (typically
ꢀ1 h of stabilization time). IPA partial pressure variation experi-
ments were performed first, followed by temperature variation
experiments. Temperature and propylene production rates were
subsequently stabilized at 415 K. The flow of reactants was then
switched from pure IPA flow to the IPA/DTBP mixture at an identi-
cal volumetric flow rate. DTBP mole fraction in the IPA/DTBP mix-
ture was ꢀ3 Â 10À4. The propylene formation rate was monitored
using online gas chromatography and was linearly extrapolated to
zero during titration to estimate the total DTBP uptake as a mea-
sure of acid site density.
Carbides synthesized under identical conditions were also
tested for IPA dehydration with a co-feed of 13 kPa O2, and the
results are plotted in Fig. 1(b). O2 co-feed increased dehydration
rates per bulk metal atom by an order of magnitude for all car-
bides tested in this work. The rate dependence upon IPA pressure
remained invariant (zero-order), indicative of
a
reactant-
saturated surface. These results for b-Mo2C have been published
previously [32], and the three additional carbides tested in this
work behaved similarly as characterized by zero-order kinetics
and increased dehydration rate with O2-induced surface
oxidation.
3. Results and discussion
Temperature variation experiments were performed on
carbides with both zero and 13 kPa O2 co-feed, and the rates of
Fig. 1(a) shows the rate of IPA dehydration to form propylene
over in situ synthesized molybdenum and tungsten carbides with-
Fig. 1. Propylene synthesis rates over interstitial carbide catalysts of tungsten and molybdenum, normalized per mole of W or Mo, as a function of IPA partial pressure (a)
with no O2 co-feed, or (b) with 13 kPa O2 co-feed. T = 415 K, P = 114 kPa, PCH4 = 4 kPa (internal standard), balance He or Ar. Total gas flow rate of 1.83 cm3 sÀ1. Approximately
0.1 g of oxide precursor was loaded for each synthesis and subsequent reaction tests (0.2 g for a-Mo2C). Conversion varied from 0.1–19% for (a) and 0.3–66% for (b).

56
M.M. Sullivan, A. Bhan / Journal of Catalysis 344 (2016) 53–58
propylene formation as a function of temperature are plotted in
Fig. 2(a) and (b). The activation energies that can be extracted from
the slopes of Fig. 2(a) and (b) are intrinsic because of the zero-order
dehydration kinetics. The data presented in Fig. 1(a) and (b) exhibit
an order of magnitude increase in the measured propylene rates
with oxygen co-feed, but the IPA dehydration kinetics (as quanti-
fied by the activation energy) remained invariant even with an
order of magnitude increase in rate. The activation energies
extracted from the slopes of Fig. 2(a) and (b) and tabulated in
Table 1 are shown to be similar.
Propylene turnover frequencies (TOF), normalized by the acid
site density calculated from the extrapolated time to complete
titration based on DTBP co-feed experiments, are plotted in Fig. 3
and also tabulated in Table 1. Acid site densities per bulk metal
atom are also reported due to difficulties in accurately ascertaining
oxidized sample surface area as mentioned in Section 2.2. A typical
DTBP titration experiment is shown in Section S.3 of Supplemen-
tary Information. TOF values for all carbides evaluated in this work
are plotted in Fig. 3 along with TOF values calculated from inter-
Propylene TOF values for all MoCx samples varied within a fac-
tor of ꢀ2 regardless of O2 co-feed pressure; WCx propylene TOF
values also varied within a factor of ꢀ2 for O2 co-feed experiments.
WCx propylene TOF values were also typically higher than MoCx by
a factor of ꢀ2. This implicates a similarity in the nature of the acid
sites for IPA dehydration across fresh and oxygen modified car-
bides of molybdenum and tungsten.
Propylene rates were not entirely inhibited by DTBP co-feed;
DTBP was able to irreversibly titrate 85–94% of the dehydration
rates of O⁄-carbides, but DTBP could only titrate 40–50% of the
dehydration rate of fresh carbides that had never been exposed
to O2. This could be explained by the presence of acid sites that
DTBP cannot chemically bind to, such as (i) Lewis acid sites, and/
or (ii) sites that are sterically inaccessible to DTBP either due to
the presence of catalyst pores on the order of 1–2 nm for b-Mo2C
[32] or due to closely localized acid sites (crystal edge sites). The
propylene TOF was calculated by assuming that all sites are chem-
ically identical Brønsted acid sites. The presence of Lewis acid sites
or a non-uniform distribution of sites would cause a distribution in
TOF values that is not captured in this single-value TOF estimation
technique and would also account for the un-titrated rate; we are
unable to distinguish between the relative effects of possible Lewis
acid sites or sterically inaccessible sites. However, the measured
similarity in reaction order dependences, activation energies, and
turnover frequencies across all catalyst samples implies the lack
of a broad spectrum of active sites with varying reactivity.
The post-reaction X-ray diffraction patterns of all catalysts
tested for IPA dehydration in Figs. 1–3 are presented in Fig. 4.
mediate O2 co-feed pressures for W2C and b-Mo2C. A TOF for
a-
Mo2C with no O2 co-feed could not be calculated due to the inabil-
ity of DTBP to titrate the extremely low dehydration rate. The nat-
ure of the as-synthesized a-Mo2C surface was noticeably different
from the other catalysts used in this study due to the synthesis
procedure (described in the experimental section). The surface
prior to O2 co-feed was expected to be covered in carbon residue
due to the large carbon content of the sucrose in the synthesis pre-
cursor. ICP-OES of the
compared to the theoretical 94 wt% Mo of a pure
a
-Mo2C sample showed ꢀ37 wt% Mo as
a
-Mo2C sample.
Fig. 4 shows that the bulk structures of WC, W2C, and a-Mo2C were
O2 co-feed caused a ꢀ130 h transient during which the propylene
production rate rose and finally stabilized; IPA dehydration pres-
sure dependence, activation energy, and TOF were measured at
this stable dehydration rate (Section S.4 of Supplementary
Information).
unaffected by O2 co-feed at 415 K under reaction conditions. Tran-
sition metal carbide catalysts demonstrate Brønsted acidity due to
surface oxidation, and the acid sites generated are similar regard-
less of the extent of oxidation or bulk transition metal composition
as evinced by similar kinetics and propylene TOF values. Surface
Fig. 2. Propylene synthesis rates over interstitial carbide catalysts of tungsten and molybdenum, normalized per mole of W or Mo, as a function of temperature (a) with no O2
co-feed, or (b) with 13 kPa O2 co-feed. P = 114 kPa, PCH4 = 4 kPa (internal standard), PIPA = 0.54 or 1.1 kPa, balance He or Ar. Total gas flow rate of 1.83 cm3 sÀ1. Approximately
0.1 g of oxide precursor was loaded for each synthesis and subsequent reaction tests (0.2 g for a-Mo2C). Conversion varied from 0.1–20% for (a) and 0.9–78% for (b).

M.M. Sullivan, A. Bhan / Journal of Catalysis 344 (2016) 53–58
57
Table 1
Kinetic parameters of IPA dehydration over carbides of Mo and W with and without O2 co-feed. Tabulated numbers were obtained from data plotted in Figs. 1–3.
No O2 co-feed
13 kPa O2 co-feed
a
a
Ea
ab
TOFc
Ld (/1 Â 10À5
)
Ea
ab
TOFc
Ld(/1 Â 10À5
)
W2C
WC
104
89
99
6
À0.05
0.06
0.08
À0.02
0.14
0.14
–
17
6
–
100
98
99
7
0.01
0.11
0.02
0.05
0.31
0.30
0.19
0.13
78
82
4
9
8
4
4
5
3
a-Mo2C
b-Mo2C
92
0.07
6
94
66
a
b
c
Activation energy in units of kJ molÀ1
.
a
IPA partial pressure dependence on propylene formation rate (r = kPIPA).
Propylene turnover frequency with units of (propylene sÀ1 DTBP siteÀ1).
d
Brønsted acid site density, as calculated by DTBP uptake extrapolated to titrate dehydration rates completely, normalized per bulk metal atom (Mo or W).
Fig. 3. Propylene turnover frequency (TOF) as a function of propylene synthesis rate
normalized by catalyst loading for carbides of W and Mo. TOF values are listed in
Table 1 and were measured on the catalyst formulations used for IPA pressure and
temperature variation experiments shown in Figs. 1 and 2. TOF values with O2 co-
feed pressures in between 0 and 13 kPa are also plotted for b-Mo2C and W2C. TOF
was normalized using Brønsted acid site density measured by extrapolation of
in situ DTBP titration experiments. T = 415 K, ꢀ0.1 gcat of oxide precursor,
P = 114 kPa, PIPA = 0.54 or 1.1 kPa, PCH4 = 4 kPa (internal standard), balance He, total
gas flow rate of 1.83 cm3 sÀ1
.
oxidation increases surface acid site density, but bulk carbide crys-
tal structure remains unchanged with O2-induced surface
alterations.
Fig. 4. X-ray diffraction patterns of WC, W2C, and
with or without 13 kPa O2 co-feed. Peak indices were obtained from the JCPDS
database for WC (JCPDS PDF # 01-089-2727), W2C (JCPDS PDF # 00-035-0776), and
a-Mo2C catalysts after reaction
The IPA dehydration kinetic parameters as reported in Table 1
are relatively independent of bulk carbidic crystal structure, imply-
ing that (i) the structure and composition of the bulk carbide min-
imally alter the acidic properties of the catalyst surface, and (ii)
oxygen modification with O2 co-feed increases the number of
acidic active sites without significantly altering the nature of the
acid sites. XPS of carbide catalysts after 13 kPa of O2 co-feed for
over 12 h showed no transition metal in the carbidic oxidation
state (<10 atom% Mo3+) with the bulk of the surface atoms residing
as oxides or suboxides (Mo5+ and Mo6+) as shown in Section S.5 of
Supplementary Information. This increase in surface oxidation
with concurrent increase in acidic active site density as measured
by in situ titrations suggests that the Brønsted acid sites on the car-
bide surfaces, both fresh and oxygen modified, are similar in nature
to acid sites present on transition metal oxides.
a-Mo2C (JCPDS PDF # 03-065-0280).
13 kPa over W2C induced nearly identical TOF values (0.30 vs.
0.31 sÀ1
as well as acid site densities (rates of 241 vs.
)
241 mol sÀ1 molWÀ1). Dehydration rates over b-Mo2C increased
gradually over the range of O2 co-feed pressures from 0 to 13 kPa
with O2 pressures above ꢀ10 kPa not affecting the dehydration
rate. Conversely, O2 co-feed of 0.4 kPa over W2C induced a maxi-
mum dehydration rate, and higher O2 co-feed pressures caused
no additional increase in the dehydration rates as noted in Fig. 3
and in Section S.6 of Supplementary Information. DeMaria et al.
[38] used a Knudsen cell in conjunction with mass spectroscopic
methods to correlate partial pressures of vaporized metal oxides
to enthalpies of metal–oxygen bond formation; the greater com-
parative oxophilicity of tungsten with respect to molybdenum is
evident from the enthalpies of formation of Mo–O and W–O bonds
(485 63 and 644 42 kJ molÀ1, respectively). Oxygen co-feed
over both MoCx and WCx catalysts generates Brønsted acid sites
due to the oxophilicity of these formulations, and the acid sites
generated on the catalyst surface saturate with increasing O2 pres-
sure, but the greater oxophilicity of WCx in comparison with MoCx
catalysts causes the maximum acid site density to saturate at
lower O2 co-feed pressure.
While both fresh and oxygen-modified MoCx and WCx catalyze
IPA dehydration at nearly identical rates, it is worth noting the
effects of different oxophilicities of the MoCx and WCx catalysts
shown by the effects of O2 co-feed pressure in Fig. 3. An increase
from 0.3 kPa to 13.5 kPa O2 co-feed over b-Mo2C did not affect
the dehydration TOF (0.16 vs. 0.13 sÀ1), but the acid site density
was nearly doubled as noted by the dehydration rate per Mo atom
(45 vs. 84 mol sÀ1 mol ). Conversely, O2 co-feed of 0.4 kPa and
À1
Mo

58
M.M. Sullivan, A. Bhan / Journal of Catalysis 344 (2016) 53–58
[3] J.H. Sinfelt, D.J.C. Yates, Nat. Phys. Sci. 229 (1971) 27–28.
4. Conclusions
[4] J.-S.S. Choi, G. Bugli, G. Djéga-Mariadassou, J. Catal. 193 (2000) 238–247.
[5] H. Shou, R.J. Davis, J. Catal. 306 (2013) 91–99.
Carbides of molybdenum and tungsten exhibit Brønsted acidity
following carbide synthesis without exposure to O2 as evinced by
irreversible DTBP titration. Dioxygen co-feed increases the rate of
acidic dehydration of IPA to propylene at 415 K by an order of
magnitude over interstitial carbides of molybdenum and tungsten;
O2-induced surface oxidation does not alter bulk carbide structures
as evidenced by the unchanged post-reaction X-ray diffraction pat-
terns. The oxophilic characteristics of fresh and O2-modified (O⁄)
carbides result in IPA surface saturation with partial pressures as
low as 0.044 kPa and in zero-order dehydration kinetics to propy-
lene. Intrinsic activation energies of unimolecular dehydration
[6] M.-L. Frauwallner, F. López-Linares, J. Lara-Romero, C.E. Scott, V. Ali, E.
Hernández, P. Pereira-Almao, Appl. Catal. A Gen. 394 (2011) 62–70.
[7] K.J. Leary, J.N. Michaels, M. Stacy, J. Catal. 101 (1986) 301–313.
[8] E. Iglesia, J.E. Baumgartner, F.H. Ribeiro, M. Boudart, J. Catal. 131 (1991) 523–
544.
[9] A.F. Lamic, C.H. Shin, G. Djéga-Mariadassou, C. Potvin, Appl. Catal. A Gen. 302
(2006) 5–13.
[10] E.A. Blekkan, C. Pham-Huu, M.J. Ledoux, J. Guille, Ind. Eng. Chem. Res. 33
(1994) 1657–1664.
[11] R.B. Levy, M. Boudart, Science 181 (1973) 547–549 (80-.).
[12] V. Keller, P. Wehrer, F. Garin, R. Ducros, G. Maire, J. Catal. 166 (1997) 125–135.
[13] H. Ren, W. Yu, M. Salciccioli, Y. Chen, Y. Huang, K. Xiong, D.G. Vlachos, J.G.
Chen, ChemSusChem 6 (2013) 798–801.
[14] H. Ren, Y. Chen, Y. Huang, W. Deng, D.G. Vlachos, J.G. Chen, Green Chem. 16
(2014) 761–769.
[15] K. Xiong, W. Yu, D.G. Vlachos, J.G. Chen, ChemCatChem 7 (2015) 1402–1421.
[16] A.L. Jongerius, R.W. Gosselink, J. Dijkstra, J.H. Bitter, P.C.A. Bruijnincx, B.M.
Weckhuysen, ChemCatChem 5 (2013) 2964–2972.
[17] S. Boullosa-Eiras, R. Lødeng, H. Bergem, M. Stöcker, L. Hannevold, E.A. Blekkan,
Catal. Today 223 (2014) 44–53.
over
fresh
and
O⁄-modified
carbides
were
similar
(89–104 kJ molÀ1) across carbide bulk structure, with or without
O2 co-feed. Propylene TOF values for fresh and O⁄-carbides, nor-
malized by the calculated Brønsted acid site density obtained from
extrapolation of DTBP rates of titration, were similar within a fac-
tor of ꢀ2 for both molybdenum and tungsten carbides. The
oxophilicity of the transition metal carbide formulations results
in the genesis of Brønsted acidity with catalyst surface oxidation
without the alteration of bulk crystal structure.
[18] F.H. Ribeiro, R.a. Dalla Betta, M. Boudart, J.E. Baumgartner, E. Iglesia, J. Catal.
130 (1991) 86–105.
[19] F.H. Ribeiro, M. Boudart, R.A.D. Betta, R.A. Dalla Betta, E. Iglesia, J. Catal. 513
(1991) 498–513.
[20] A.F. Lamic, T.L.H. Pham, C. Potvin, J.M. Manoli, G. Djéga-Mariadassou, J. Mol.
Catal. A Chem. 237 (2005) 109–114.
[21] J. Macht, C.D. Baertsch, M. May-Lozano, S.L. Soled, Y. Wang, E. Iglesia, J. Catal.
227 (2004) 479–491.
Acknowledgments
[22] C. Baertsch, K.T. Komala, Y. Chua, E. Iglesia, J. Catal. 205 (2002) 44–57.
[23] J.A. Schaidle, J. Blackburn, C.A. Farberow, C. Nash, K.X. Steirer, J. Clark, D.J.
Robichaud, D.A. Ruddy, ACS Catal. 6 (2016) 1181–1197.
[24] F.G. Baddour, C.P. Nash, J.A. Schaidle, D.A. Ruddy, Angew. Chemie Int. Ed. 55
(2016) 1–5.
[25] W. Lee, Z. Wang, R. Wu, A. Bhan, J. Catal. 319 (2014) 44–53.
[26] C.-J. Chen, W.-S. Lee, A. Bhan, Appl. Catal. A Gen. 510 (2015) 42–48.
[27] M.M. Sullivan, A. Bhan, ACS Catal. 6 (2016) 1145–1152.
[28] W.-S. Lee, Z. Wang, W. Zheng, D.G. Vlachos, A. Bhan, Catal. Sci. Technol. 4
(2014) 2340–2352.
[29] K. Xiong, W.-S. Lee, A. Bhan, J.G. Chen, ChemSusChem 7 (2014) 2146–2149.
[30] W.-S. Lee, A. Kumar, Z. Wang, A. Bhan, ACS Catal. 2 (2015) 4104–4114.
[31] M.M. Sullivan, C.-J.C.-J. Chen, A. Bhan, Catal. Sci. Technol. 6 (2016) 602–616.
[32] M.M. Sullivan, J.T.J.T. Held, A. Bhan, J. Catal. 326 (2015) 82–91.
[33] J.A. Schaidle, N.M. Schweitzer, O.T. Ajenifujah, L.T. Thompson, J. Catal. 289
(2012) 210–217.
This research was supported by Office of Basic Energy Sciences,
the U.S. Department of Energy under award number no.
DE- SC0008418 (DOE Early Career Program). Part of this work
was carried out in the College of Science and Engineering Charac-
terization Facility, University of Minnesota, which has received
capital equipment funding from the NSF through the MRSEC
program. We thank Javier Garcia Barriocanal for assistance with
the X-ray diffraction studies, Rick Knurr for assistance with the
ICP-OES characterization, and Bing Luo for assistance with the
XPS characterization.
[34] G. Vitale, M.L. Frauwallner, C.E. Scott, P. Pereira-Almao, Appl. Catal. A Gen. 408
(2011) 178–186.
Appendix A. Supplementary data
[35] L. Delannoy, J.-M. Giraudon, P. Granger, L. Leclercq, G. Leclercq, Catal. Today 59
(2000) 231–240.
[36] J.S. Lee, S.T. Oyama, M. Boudart, J. Catal. 106 (1987) 125–133.
[37] S.K. Bej, L.T. Thompson, Appl. Catal. A Gen. 264 (2004) 141–150.
[38] G. De Maria, R.P. Burns, J. Drowart, M.G. Inghram, J. Chem. Phys. 32 (1960)
1373.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2016.09.012.
References
[1] J.S. Lee, S. Locatelli, S.T. Oyama, M. Boudart, J. Catal. 125 (1990) 157–170.
[2] G.S. Ranhotra, A.T. Bell, J.A. Reimer, J. Catal. 108 (1987) 40–49.