Functional assessment of the strength of solid acid catalysts

Article abstract of DOI:10.1016/j.jcat.2009.03.005

We describe here a rigorous method to estimate the deprotonation energy (DPE) and acid strength for solid Bronsted acids with uncertain structure using rate constants for reactions involving cationic transition states. The approach exploits relations between turnover rates for dehydration and isomerization reactions and DPE values on Keggin polyoxometalates and H-BEA solids with known structures. These relations are used to estimate the strength of acid sites in SO₄-ZrO₂(SZr), WO_x-ZrO₂(WZr), and perfluorosulfonic resins (SAR) from their alkanol dehydration and alkane isomerization rate constants. Alkanol dehydration and alkane isomerization proceed via pathways independent of acid identity and are limited by steps involving late transition states. Turnover rates (per accessible acid sites measured by titration during catalysis) are related to the relevant rate constants and are used to estimate DPE values for SZr, WZr, and SAR. Isomerization data estimate DPE values of 1110 kJ mol^-1 and 1120 kJ mol^-1 for SZr and WZr, respectively, while dehydration rate data lead to slightly higher values (1165 kJ mol^-1 and 1185 kJ mol^-1). The DPE value for SAR was 1154 kJ mol^-1 from dehydration reactions, but diffusional constraints during reactions of non-polar alkanes precluded isomerization rate measurements. SZr and SAR contain stronger acid sites than zeolites (1185 kJ mol^-1), but weaker than those in H₃PW₁₂O₄₀ and H₄SiW₁₂O₄₀ (1087 kJ mol^-1 and 1105 kJ mol^-1). Acid sites present in WZr during alkane isomerization are stronger than those present in zeolites, but these become similar in strength in the polar environment prevalent during dehydration catalysis. These effects of reaction media (and treatment protocols) reflect differences in the extent of dehydroxylation of catalytic surfaces. OH groups remaining after dehydroxylation are stronger acid sites because of a concomitant decrease in electron density in the conjugate anion and the formation of Bronsted-Lewis acid conjugate pairs. The method proposed and used here probes acid strength (as DPE) on sites of uncertain structure and within solvating media inherent in their use as catalysts. It can be used for any Bronsted acid or reaction, but requires reactivity-DPE relations for acids of known structure, the mechanistic interpretations of rates, and the measurement of accessible protons during catalysis. The resulting DPE values provide a rigorous benchmark for the structural fidelity of sites proposed for acids with uncertain structure, a method to assess the consequences of the dynamic nature of active sites in acid catalysis, and a connection between theory and experiment previously unavailable.

Full text of DOI:10.1016/j.jcat.2009.03.005

Journal of Catalysis 264 (2009) 54–66
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Functional assessment of the strength of solid acid catalysts
Josef Macht, Robert T. Carr, Enrique Iglesia *
Department of Chemical Engineering, University of California at Berkeley, Berkeley, CA 94720, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
We describe here a rigorous method to estimate the deprotonation energy (DPE) and acid strength for
solid Brønsted acids with uncertain structure using rate constants for reactions involving cationic transi-
tion states. The approach exploits relations between turnover rates for dehydration and isomerization
reactions and DPE values on Keggin polyoxometalates and H-BEA solids with known structures. These
Received 13 December 2008
Revised 24 February 2009
Accepted 8 March 2009
Available online 22 April 2009
4 2 x 2
relations are used to estimate the strength of acid sites in SO –ZrO (SZr), WO –ZrO (WZr), and perfluoro-
sulfonic resins (SAR) from their alkanol dehydration and alkane isomerization rate constants. Alkanol
dehydration and alkane isomerization proceed via pathways independent of acid identity and are limited
by steps involving late transition states. Turnover rates (per accessible acid sites measured by titration
during catalysis) are related to the relevant rate constants and are used to estimate DPE values for SZr,
Keywords:
Keggin polyoxometalates
2
-Butanol dehydration
n-Hexane isomerization
Sulfated zirconia
Tunstated zirconia
Perfluorosulfonic acid resins
Deprotonation energy
ꢀ1
ꢀ1
WZr, and SAR. Isomerization data estimate DPE values of 1110 kJ mol and 1120 kJ mol for SZr and
ꢀ
1
WZr, respectively, while dehydration rate data lead to slightly higher values (1165 kJ mol
and
ꢀ1
ꢀ1
1185 kJ mol ). The DPE value for SAR was 1154 kJ mol from dehydration reactions, but diffusional con-
straints during reactions of non-polar alkanes precluded isomerization rate measurements. SZr and SAR
ꢀ1
contain stronger acid sites than zeolites (1185 kJ mol ), but weaker than those in H
3
PW12
O40 and
ꢀ1
ꢀ1
H SiW12O40 (1087 kJ mol and 1105 kJ mol ). Acid sites present in WZr during alkane isomerization
4
are stronger than those present in zeolites, but these become similar in strength in the polar environment
prevalent during dehydration catalysis. These effects of reaction media (and treatment protocols) reflect
differences in the extent of dehydroxylation of catalytic surfaces. OH groups remaining after dehydroxy-
lation are stronger acid sites because of a concomitant decrease in electron density in the conjugate anion
and the formation of Brønsted–Lewis acid conjugate pairs. The method proposed and used here probes
acid strength (as DPE) on sites of uncertain structure and within solvating media inherent in their use
as catalysts. It can be used for any Brønsted acid or reaction, but requires reactivity-DPE relations for
acids of known structure, the mechanistic interpretations of rates, and the measurement of accessible
protons during catalysis. The resulting DPE values provide a rigorous benchmark for the structural fidelity
of sites proposed for acids with uncertain structure, a method to assess the consequences of the dynamic
nature of active sites in acid catalysis, and a connection between theory and experiment previously
unavailable.
Ó 2009 Elsevier Inc. All rights reserved.
1
. Introduction
The catalytic properties of solid Brønsted acids have often been
environments relevant to catalytic reactions or provide direct mea-
sures of the accessibility, number, or strength of acid sites during
catalysis. These approaches are often limited by the broad nature
of infrared O–H bands in non-zeolitic acids of catalytic interest,
such as Keggin-type polyoxometalate (POM) clusters [8]
(H PW12O40) and perfluorosulfonic acid resins [3,9], restricting,
3
as a result, their current ability to contrast the acid properties
among the broad range of useful solid acid catalysts.
The use of adsorption enthalpies for basic molecules, accessible
by temperature-programmed desorption or calorimetry, to
measure acid strengths is not unambiguous. The assessment of
desorption dynamics requires kinetic analyses inadequate for
non-uniform ensembles of acid sites [10]; it also requires probe
molecules that desorb before they or the solid acid decompose, a
ubiquitous problem for Keggin-type POM clusters [11], sulfonic
attributed to the strength of their acid sites [1]. Yet, methods re-
quired for the rigorous assessment of acid strength have remained
elusive. For H-form zeolites, correlations between acid strength
and the shifts in O–H stretching frequencies or 1H-lines upon
adsorption of basic molecules (by infrared [2–4] and nuclear mag-
netic resonance [5–7] spectroscopies, respectively) have been pro-
posed. These methods lack, however, a rigorous connection to a
specific value of deprotonation energies accessible to theoretical
treatments. Also, they cannot assess acid strength within solvating
*
Corresponding author. Fax: +1 510 642 4778.
E-mail address: iglesia@berkeley.edu (E. Iglesia).
0
021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.03.005

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
55
acid resins, and sulfated oxides. Significant contributions to
-
+
A + H + R
adsorption enthalpies from non-specific van der Waals interac-
tions, especially for molecules confined within zeolites, and spe-
cific charge-transfer interactions via hydrogen bonding, which
are ubiquitous for adsorbed molecules but not for cationic species
at transition states, cause adsorption enthalpies to depend only in
part on acid strength.
-
PA
-
+
A + RH
ΔEint
DPE
Acid strength rigorously reflects the energy required to remove
a proton from a solid acid; this deprotonation energy (DPE) is a
probe-independent intrinsic property of an acid [12]. The ubiqui-
tous involvement of cationic transition states and the more neutral
character of adsorbed intermediates (relative to transition states)
in solid acids make it essential that we discern the respective
and often independent roles of DPE and of specific and non-specific
interactions in the stabilization of adsorbed reactants and products
and especially of ion-pairs at transition states, which depend on
the combined effects of van der Waals, hydrogen-bonding, and
electrostatic interactions for stabilization.
We propose here a method for assessing acid strength (as spe-
cific DPE values) from the dynamics of any reaction for which
mechanistic interpretations implicate kinetically-relevant elemen-
tary steps catalyzed by Brønsted acid sites. Our strategy involves
turnover rate and selectivity measurements on solid acids of
known structure, such as zeolites and POM clusters, for which
DPE values can be estimated accurately, under conditions of strict
kinetic control [12–14]. The resulting relations are then used to
estimate DPE values for solid acids with uncertain or non-uniform
site structures at conditions prevalent during each catalytic reac-
tion. Our recent studies have established the rigor and accuracy
of such relations for alkanol dehydration and (bifunctional) alkane
isomerization on acid forms of Keggin-type POM clusters and zeo-
lites, which are used as solid acids with known structure [12,13].
Brønsted acid sites stabilize cationic transition states formed via
protonation of either reactants or reactant-derived intermediates.
Quantum chemical methods [13,15–18], the strong effects of carbe-
nium ion stability on reaction barriers and rate constants [12], and
the small kinetic isotope effects observed [12] indicate that proton
transfer is typically complete at the transition state, which can there-
fore be accurately treated as an ion pair. As a result, activation barri-
-…
+
[
A RH ]
AH + R
ΔHads
E =DPE-PA+ΔE -ΔH
ads
a
int
…
AH R
Scheme 1. Thermochemical cycle for acid–base reactions on Brønsted acid
catalysts. The activation barrier (E ) is determined by the deprotonation energy
(DPE) of the acid (AH), the ‘‘proton affinity” (PA), the ion-pair stabilization energy
int), and the reactant adsorption energy (
a
(D
E
DHads).
Alkanol dehydration and bifunctional n-hexane isomerization
are used here as specific probe reactions, but the strategy described
and implemented is general for chemical reactions with kineti-
cally-relevant elementary steps catalyzed by Brønsted acids, as
long as rate measurements can be interpreted rigorously in terms
of kinetic and/or thermodynamic parameters for elementary steps
and the number of accessible protons is measured during catalysis.
Elimination and isomerization elementary steps and their kinetic-
relevance are well understood [12,14,19–21] and allow rates and
selectivities to be interpreted as rate and equilibrium constants
and consequently in terms of the stability of adsorbed species
and cationic transition states. As a result, the effects of DPE and
DEint can be rigorously assessed for materials with well-defined
structures, which are amenable to quantum chemical calculations.
We have recently reported accurate and rigorous connections be-
tween rate (and equilibrium) constants and DPE for elementary
steps in dehydration and isomerization reactions on acid zeolites
and POM clusters [12,14,19]. These studies showed that DPE and
transition state stabilization energies determine activation barriers
and that rate constants decrease exponentially with increasing DPE
values.
ers (E
deprotonation energyofthe catalyst(DPE), onthe reactantorproduct
proton affinities (PA), on the ion-pair interaction energies ( int), and
on the enthalpy of adsorption of gas-phase reactants to form bound
intermediates preceding the transition state ( ads) (Scheme 1).
The types and properties of acid sites determine transition state
energies through the magnitudes of DPE, int, and ads, indicat-
a
) for elementary steps in acid catalysis depend on the
DE
D
H
x
Sulfated zirconia (SZr) [22–25], WO domains supported on zir-
conia (WZr) [26–28], and perfluorosulfonic acid resins (SAR; e.g.
Nafion ) [29] catalyze some chemical reactions more effectively
D
E
D
H
Ò
ing that activation barriers depend on the ability of an acid to do-
nate a proton (DPE), but also on its role in stabilizing the reactants
than acidic zeolites. Here, we probe the acid properties of these so-
lid acids, for which structures are uncertain and DPE estimates,
which depend on the structures assumed in calculations, are unre-
liable. We show that the mechanisms involved in Keggin POM and
H-BEA catalysts also apply to SZr, WZr, and SAR for alkanol dehy-
dration (Section 3.1) and for SZr and WZr for bifunctional alkane
isomerization (Section 3.4). We also report measurements of acid
site densities during catalysis using titration with 2,6-di-tert-butyl-
pyridine on these materials (Sections 3.2 and 3.5) and use them to
measure their intrinsic site reactivity in terms of turnover rates.
The number of accessible Brønsted acid sites in SZr and WZr
depends on treatment and reaction conditions, specifically on
(
D
H
ads) and the cationic species in the ion-pair (
predominantly via electrostatic interactions. The extent to which
DPE values are compensated by int defines the energy relevant
DEint), for the latter
D
E
to the stability of the transition state and thus to catalytic reactiv-
ity [12]; yet, these are not available from characterization methods
that exploit the binding of probe molecules, often stabilized via
covalent or van der Waals interactions, within environments extra-
neous to catalysis. The contributions from
DHads to the relevant
activation barriers depend on whether adsorbed reactants are
present as hydrogen-bonded alkanols [12,14] or as covalently
bound alkoxides [19]. Differences in
acids in 2-butanol dehydration [14] and n-hexane isomerization
19] were, however, much smaller than the differences among
DHads values among various
2
temperature, H O concentration (on SZr) [30], and reductant and
[
oxidant concentrations and identity (on WZr) [31]. The relations
between reactivity and DPE developed for well-defined acids are
then used here to provide DPE estimates for SZr, WZr, and SAR cat-
alysts with unknown active site structure (Sections 3.3 and 3.6).
These DPE estimates can then be used to benchmark the fidelity
the respective transition state energies. Therefore, the range of
barriers and rate constants for an elementary step on solid acids
is predominantly determined by the catalyst DPE and by its ability
to stabilize the transition state by electrostatic interactions as an
ion-pair.
4 2
of the various acid site structures considered for SO –ZrO and

5
6
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
WO
x
–ZrO
2
catalysts (Section 3.7). For these materials, the number
2.3. 2-Butanol dehydration catalysis
and strength of acid sites, therefore, do not depend solely on the
composition or initial structure of solids, but also on solvation
effects during catalysis and on their thermal history during
treatments.
Rates and selectivities were measured at 343 K in a quartz flow
cell (1.0 cm inner diameter; 0.02–0.2 g catalyst; 125–180 lm
aggregates) using protocols similar to those for n-hexane isomeri-
zation. 2-Butanol reactants (Sigma–Aldrich, 99.5%) were used
without purification and introduced as a liquid using a syringe
pump (Cole Parmer, 74900 series) by vaporization into flowing
He (Praxair, UHP) at 393 K. Molar rates of He and of all gaseous
species were adjusted to give the desired reactant pressures and
to maintain low and relatively constant reactant conversions
2
. Experimental and theoretical methods
2.1. Catalyst synthesis and characterization
Synthesis protocols for SZr [32], WZr [31], and SAR [29] have
(
<10%). Only the alkene products of dehydration reactions were
been reported earlier. Sulfated zirconium hydroxide (Magnesium
Electron, Inc. (XZO 1077/01)) was heated in static ambient air
detected (1-butene, cis-2-butene, trans-2-butene). Deactivation
by oligomerization reactions was not observed at the low alkanol
conversions and low alkene concentrations prevalent in these
experiments. Brønsted acid sites were titrated by introducing
ꢀ
1
(
0.17 K s , 3 h, 873 K); the resulting material had a N
2
BET sur-
[32]. WZr
2
ꢀ1
ꢀ1
face area of 109 m g
and 0.44 (mmol sulfate) g
was prepared by aqueous impregnation of zirconium oxyhydrox-
ide, precipitated from a 0.5 M aqueous solution of zirconyl chloride
2
,6-di-tert-butylpyridine (Aldrich, 97%) dissolved in 2-butanol
reactants (Sigma–Aldrich, 99.5%, anhydrous) into flowing He and
vaporized at 393 K to give a stream containing 0.5 kPa 2-butanol
and 0.9 Pa 2,6-di-tert-butylpyridine.
(
ZrOCl
using NH
ous ammonium metatungstate solution ((NH
Chemicals, 99.9%) to give a material containing 15 wt% of WO
2
ꢁ 8H
OH, Aldrich 99%) with the required amount of an aque-
40, Strem
2
O, Aldrich, >98 wt%) at constant pH (10, controlled
4
4 6 2 12
) H W O
3
2.4. Theoretical calculations and methods
[
31]. The impregnated WZr powders were dried at 383 K and trea-
3
ꢀ1 ꢀ1
ted in flowing dry air (Matheson, zero grade, 1 cm s
1
density was 7.9 nm . The Nafion -SAC13 material was used as
received (Sigma–Aldrich) and is denoted as SAR. A surface area of
g
) at
2
ꢀ1
Density functional calculations were performed using the
2 x
023 K; its N BET surface area was 49 m g and the WO surface
ꢀ2
Ò
Gaussian 03 program [34]. Geometries were optimized and ener-
gies calculated at the B3LYP/6-31 + G(d,p) level of theory. The
DPE values for Keggin-type POM clusters have been reported
earlier [14,35].
2
ꢀ1
3
0
50 m g
.14 mmol g have been reported for this material [33]. Physical
(Pt/Al prepared
as reported [19]) were prepared by mixing and grinding the metal
and a concentration of Brønsted acid sites of
ꢀ1
mixtures of SZr, SAR, and WZr with Pt/Al
2
O
3
2 3
O
3. Results and discussion
+
and acid co-catalysts into small aggregates (Pt/H = 1; <10
diameter). These mixtures were pressed into wafers, crushed,
and sieved to retain 125–180 m aggregates; they were used as
bifunctional catalysts in alkane isomerization studies.
lm
3.1. Kinetics and mechanism of 2-butanol dehydration on SO
4 2
–ZrO ,
l
WO –ZrO and sulfonic acid resins
x
2
2
-Butanol dehydration rates decreased with increasing reactant
2
.2. n-Hexane isomerization catalysis
pressure on SZr, WZr, and SAR catalysts (Fig. 1a). Similar dehydra-
tion rate dependence on 2-butanol partial pressures on SAR cata-
lysts has previously been reported [52]. This reaction proceeds
via elementary steps that include quasi-equilibrated molecular
adsorption of 2-butanol on Brønsted acid sites via hydrogen bond-
ing (Step 1, Scheme 2) to form ‘‘monomers” [12–14]. These mono-
Isomerization rates and selectivities were measured at 473 K in
a tubular quartz flow reactor (1.0 cm inner diameter; 0.1–0.3 g cat-
alysts). SAR and WZr samples were held onto a porous quartz disc
and heated to 473 K at 0.083 K s in flowing He (0.83 cm s ) be-
fore rate measurements. SZr catalysts were treated at 673 K in He
ꢀ1
3 ꢀ1
mers then decompose irreversibly via E1-type pathways, in which
3
ꢀ1 ꢀ1
þ
(
0.83 cm g
s
) for 2 h before n-hexane isomerization reactions.
C
a
—OH2 bonds are cleaved to form water molecules and adsorbed
Sample temperatures were measured using a K-type thermocouple
located within a dimple in the external quartz wall of the reactor
and were controlled (±0.2 K) using a Watlow controller (Series
butoxides in kinetically-relevant steps involving carbenium ions at
late transition states (Step 2, Scheme 2). The catalytic dehydration
cycle is completed by deprotonation of adsorbed butoxides (Step 3,
Scheme 2) to form alkenes. Butanol monomers can also interact
with another 2-butanol molecule to form unreactive dimers, which
are stabilized by hydrogen bonds [–O–H O–] between the two
molecules (Step 4, Scheme 2).
These elementary steps lead to a rate equation that accurately
describes all 2-butanol dehydration rate data on Keggin clusters
and H-BEA zeolites [12,13]:
9
3
82) and a resistively heated furnace. Transfer lines were held at
93 K to prevent adsorption or condensation of reactants, prod-
ꢁ
ꢁ ꢁ
ucts, or titrants. n-Hexane reactants (Sigma–Aldrich, 99%, Fluka,
9%) were used without additional purification and were intro-
duced as a liquid using a syringe pump (Cole Parmer, 74900 series)
by vaporization into a He (Praxair, UHP) and H (Praxair, UHP)
stream at 393 K. Molar flow rates of He, H and n-hexane were con-
trolled independently to give the desired reactant molar ratios
n-hexane/H = 0.05–1) and space velocities and to maintain low
9
2
2
þ
k
2
½H ꢃ
(
2
r ¼ 1
ð1Þ
þ K
4
½C
4
H
9
OHꢃ
and relatively constant n-hexane conversions (1–6%). Reactant
and product concentrations were measured by gas chromatogra-
phy using flame ionization detection (Agilent 6890N GC, 50 m
In this equation, k2 is the elimination step rate constant (Step 2,
Scheme 2) and K4 is the equilibrium constant for dimer formation
+
HP-1 column, methyl silicone, 50 m ꢂ 0.32 mm ꢂ 1.05
l
m).
(Step 4, Scheme 2); [H ] represents the number of accessible pro-
Brønsted acid sites were titrated by introducing 2,6-di-tert-butyl-
tons. The form of Eq. (1) reflects the presence of 2-butanol mono-
mers and dimers as most abundant surface species during
catalysis. For Keggin-type POM clusters, the proposed elementary
steps and their kinetic relevance are consistent with all kinetic
and isotopic data and with DFT estimates of activation barriers
[12,13]. This rate equation also describes all available 2-butanol
pyridine (Aldrich, 97%) dissolved in the n-hexane reactants into
He/H streams at 393 K. The amount of titrant adsorbed on the cat-
2
alyst was measured from its residual concentration in the effluent
stream using the same chromatographic protocols as for catalytic
reactions.

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
57
5
4
3
2
1
(a)
HO
H
K1
H
A
+
A
HO
(
Step 1) Adsorption
OH2
HO
H
k2
H O
A
+
2
A
(
Step 2) E1 Elimination
k3
H
+
A
(
Step 3(a)) Butene Desorption
A
0
0
.0
0.2
0.4
0.6
0.8
1.0
2
-Butanol Pressure (kPa)
K4
H+
A-
O
O
HO
H
HO
3
2
2
1
1
0
0
.0
+
H
H
(
Step 4)
Dimer Formation
(
b)
A
.5
.0
.5
.0
.5
.0
Scheme 2. Proposed sequence of elementary steps for 2-butanol dehydration on
Brønsted acid catalysts (HA) [12,13]. The sequence of elementary steps is also
applicable to the dehydration reactions of other alkanol reactants and ether
cleavage reactions [12].
di-tert-butylpyridine during 2-butanol dehydration; 2,6-di-tert-
butylpyridine adsorbs to Brønsted acid sites in competition with
alkanol reactants, but cannot coordinate to Lewis acid sites be-
cause of steric hindrance around its nitrogen atom [36].
Fig. 2 shows 2-butanol dehydration rates on SZr as Brønsted
acid sites are titrated by 2,6-di-tert-butylpyridine. The left panel
2.0
Start of Titration
0
.0
0.2
0.4
0.6
0.8
1.0
2-Butanol Pressure (kPa)
+
Fig. 1. 2-Butanol dehydration rates (per H ) (a) and inverse 2-butanol dehydration
1.5
rates (b) as a function of 2-butanol pressure for SZr (
343 K, conversion < 10%). Dashed lines represent the predictions from Eq. (1).
N), WZr (j), and SAR (d)
(
dehydration data on H-BEA, as well as the rates of dehydration for
other alkanols and ethers on all Keggin-type POM clusters [12].
Fig. 1a shows 2-butanol dehydration rates predicted using val-
1
0
0
.0
.5
.0
2 4
ues for k and K obtained by regressing rate data to the form of
Eq. (1) for each catalyst (dashed curves). The excellent agreement
between predicted and measured rates in Fig. 1a shows that
Eq. (1) accurately describes 2-butanol dehydration rate data also
on SZr, WZr, and SAR. Reciprocal 2-butanol dehydration rates show
a linear dependence on 2-butanol pressure (Fig. 1b), as expected
from the functional form of Eq. (1) and also as observed on POM
clusters and acid zeolites [12].
3
2
.2. Measurement of the number of accessible protons during
-butanol dehydration by 2,6-di-tert-butylpyridine titration on
0
20
0
0.0 5
0.1 0
0.1 5
0.2 0
4 2 x 2
SO –ZrO , WO –ZrO , and sulfonic acid resins
Reaction
Time (min)
Cumulative Adsorbed Titrant
(molecules / S-atom)
The regression of these kinetic rate data gives values for k
and K ; therefore intrinsic elimination rate constants (k ) require
[H+]
2
4
2
+
Fig. 2. Titration of acid sites on SZr during 2-butanol dehydration at 343 K. Reaction
rates (per H ) at 343 K as a function of the cumulative amount of titrant (2,6-di-tert-
independent measurements of the number of protons ([H ]) acces-
+
sible during catalysis. We measure these here by titration with 2,6-
butylpyridine) adsorbed (0.2 kPa 2-butanol, 0.9 Pa 2,6-di-tert-butylpyridine).

5
8
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
in Fig. 2 shows dehydration rates before adding the titrant, while
the right panel shows rates as a function of the amount of titrant
adsorbed. 2-Butanol dehydration rates decreased as 2,6-di-tert-
butylpyridine irreversibly titrates Brønsted acid sites. The titrant
uptake required to completely suppress dehydration rates (0.16
100
ꢀ1
(
S-atom) , Fig. 2) provides an upper bound for the number of
Brønsted acid sites responsible for measured reaction rates; it is
an upper bound because we cannot rule out that sites inactive
for catalysis are titrated concurrently and irreversibly. 2-Butanol
dehydration rates decreased linearly with 2,6-di-tert-butylpyridine
uptakes (Fig. 2); these trends do not give unequivocal evidence for
acid sites uniform in reactivity or strength, because 2,6-di-tert-
butylpyridine titrates all acid sites that bind it irreversibly, irre-
spective of their reactivity, before it titrates sites farther along
the bed.
10
SAR
The titration of Brønsted acid sites by 2,6-di-tert-butylpyridine
led to a linear decrease in dehydration rates and to their complete
suppression also on WZr and SAR catalysts. Measured titrant up-
SZr
ꢀ1
takes were 0.06 (W-atom) on WZr. The number of accessible
protons in WZr depends strongly on WO surface density [31].
These titrant uptakes are slightly larger than reported earlier dur-
WZr
DPE estimates
x
1
ꢀ1
ing 2-butanol dehydration at 373 K (ꢄ0.04 (W-atom) ) [31] for
samples with similar WO surface density, possibly because of loss
x
of Brønsted acid sites via hydroxyl recombination at the higher
reaction temperatures used in the previous study (373 K vs. 343 K).
The titrant uptake on SAR was one titrant per proton, where the
number of protons was determined independently from its Na ex-
change capacity [33]. The stoichiometric titration of all protons in
SAR is surprising because this material consists of agglomerates of
1200
1150
Deprotonation Energy (kJ mol )
11100
-
+
2
Fig. 3. 2-Butanol dehydration rate constants (k ) (Scheme 2, Step 2) as a function of
ꢀ
+
deprotonation energy (
D
E
rxn for HA ? A + H (non-interacting), where HA is the
PW/Si (d),
). Measured
-butanol dehydration rate constants for SZr, WZr, and SAR are shown as dashed
ꢀ
acid and
A
is the conjugate base) calculated by DFT for 0.04H
AlW/Si ( ), 0.04H CoW/Si ( ), and H-BEA (r
3
0.04H
4
SiW/Si (j), 0.04H
5
N
6
.
2
the resin (20–60 nm diameter) and SiO particles [29,33]. Thus,
2
some protons may reside within inaccessible regions. The 2,6-di-
tert-butylpyridine uptakes indicate, however, that these titrants
can access protons within the perfluorosulfonic acid resin particles
in spite of their large size, at least when polar 2-butanol reactants
are also present.
horizontal lines. k2 values were obtained by fitting Eq. (1) to the kinetic data and
normalization by the 2,6-di-tert-butylpyridine uptake (see also Table 1), which
corresponds to the number of accessible protons.
ionic species by the anionic conjugate base (DEint), and (iv) the en-
We discuss next how the k
2
values from the 2-butanol dehydra-
thalpy of the adsorbed intermediates immediately preceding the
transition state (relative to the isolated reactant(s) in the contact-
tion rate data and 2,6-di-tert-butylpyridine uptakes can be used to
estimate DPE values (and probe acid strength) for SZr, WZr, and
SAR catalysts. These DPE values cannot be estimated reliably by
theory because of significant uncertainties in the structure of the
active acid sites present in these materials.
ing gas phase and the acid) (
¼ DPE ꢀ PA þ
Changes in the central atom in Keggin-type POM clusters increased
DHads) according to (Scheme 1):
E
a
D
E
int
ꢀ
D
H
ads
ð2Þ
ꢀ
1
ꢀ1
DPE values by 34 kJ mol as its valence decreased (1087 kJ mol
;
ꢀ1
3
.3. Estimates of deprotonation energies from 2-butanol dehydration
H
3
PW12O
40 and 1121 kJ mol ; H
5
AlW12O40) [14]. This range of DPE
rate constants on SO –ZrO , WO –ZrO , and sulfonic acid resins
4
2
x
2
values increased activation barriers (derived from DFT) for 2-buta-
ꢀ1
nol dehydration by only 13.4 kJ mol because of more effective
electrostatic stabilization as DPE values increased [14]. For example,
Fig. 3 shows 2-butanol dehydration rate constants (with the
ordinate as a logarithmic scale) as a function of DPE on Keggin-type
ꢀ
1
D
E
int values derived from DFT are 318 kJ mol for H
3
PW12
O
40 and
n+
5+
4+
3+
ꢀ1
ꢀ1
POM clusters supported on SiO
2
(H8ꢀnX W12
O
40, X = P , Si , Al ,
335 kJ mol for H
5
AlW12
O
40, a difference of 17 kJ mol [14]. Sim-
2
+
Co ; in order of increasing DPE) and on H-BEA. The well-defined
structure and composition of these materials allow reliable DPE
estimates using density functional theory (DFT) [12,14]. The line
regressed through these 2-butanol dehydration rate constants
shows that there is a common exponential dependence of these
rate constants on DPE, consistent with thermochemical cycle treat-
ments of elimination elementary steps (Scheme 1), as discussed
next.
ilar compensation effects were observed for the experimental
2-butanol dehydration barriers [12]. For Keggin-type POM clusters
and various mineral acids, weaker acids (with higher DPE values)
typically form anionic clusters with a higher charge density, often
as a consequence of their smaller size and larger number of protons,
and these anionic species effectively stabilize cations and lead to
more negative
DEint values [12]. The observation that DEint values
become more negative as the number of charge-balancing protons
ꢀ
1
The effects of DPE on dehydration rate constants and activation
barriers are rigorously addressed by thermochemical cycles that
describe the stability of cationic transition states in terms of the
relevant reactant and catalyst properties [12,37]. The barrier for a
increases (e.g.
int = 335 kJ mol ) for acids of identical anion size can be under-
stood by considering the charge of the anionic conjugate base as
H
3
PW12
O
40
,
D
E
int = 318 kJ mol
;
5 40
H AlW12O ,
ꢀ
1
D
E
dþ
nþ
ð1þð8ꢀnÞdÞꢀ
40
H
ðX
ꢀn
n+
W
12
O
Þ
. This shows that the electron density in
40 conjugate base increases with decreasing valence
8
reaction step catalyzed by Brønsted acids (E
a
) depends on (i) the
the X
12
W O
energy required to deprotonate the acid (DPE), (ii) the energy of
corresponding gas-phase reactions of reactants with protons (e.g.,
ROH + H ? R + H O for dehydration reactions with late transition
2
of X and that the electrostatic stabilization of cations becomes
stronger as the number of charge-balancing protons concurrently
increases.
+
+
states), (iii) the interaction energy of the ion-pair at the transition
state, resulting mainly from electrostatic stabilization of the cat-
The compensation found between
nature of DPE values as the energy cost for charge separation
DEint and DPE reflects the

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
59
to form the non-interacting proton and anionic conjugate base.
3.4. Kinetics and mechanism for n-hexane isomerization on SO
4 2
–ZrO
Weaker acids (with larger DPE values) require a large energy in-
put for charge separation, but this energy is recovered in part by
the interactions of the resulting anion with other cations (as
transition states or intermediates). The distance between anionic
clusters and cations is smallest for the proton removed in the
process described by DPE, because protons lack core electrons
and their repulsive effects. Thus, DPE values can only be partially
and WO –ZrO
x
2
Elementary rate constants for isomerization steps can only be
accurately measured when n-hexane isomerization occurs on
bifunctional catalysts containing a metal function used to maintain
hydrogenation–dehydrogenation equilibrium [51]. Monofunction-
al alkane isomerization on Brønsted acids is limited by initiation
steps that involve protolytic alkane dehydrogenation, instead of
isomerization steps. These dehydrogenation steps may also be cat-
alyzed by redox sites, provided either by the solid acid (or by spu-
rious impurities) and are unrelated to acid chemistry, and often
require temperatures that degrade the inorganic acid or deplete
the sites or the spurious impurities [38,39]. SZr and WZr catalysts
dehydrogenate alkanes via stoichiometric reduction processes of
compensated by
tion states [19].
DEint upon formation of the ion pairs at transi-
For 2-butanol dehydration on Keggin-type POM clusters, DFT
calculations have shown that ads values (see also Scheme 2, Step
) changed from ꢀ77 kJ mol
AlW12 40) as DPE values and barriers increased by 34 kJ mol
and 13.4 kJ mol , respectively [14]. Similar weak correlations
between DPE and ads are expected in general for Brønsted acids,
D
H
ꢀ
1
ꢀ1
1
3
(H PW12
O
40) to ꢀ73 kJ mol
ꢀ1
(
H
5
O
ꢀ1
D
H
x x
SO and WO species, respectively, a property often misinterpreted
because alkanol adsorption involves hydrogen-bonding interac-
tions between protons and oxygen atoms in alkanols; these inter-
actions are ubiquitous for other Brønsted acids. This suggests that
the effect of composition on rate constants and activation barriers
for alkanol dehydration reactions predominantly reflects the sta-
bilization of cationic transition states and, therefore, the effects
of DPE, instead of differences in the stability of adsorbed precursors
on sites with different DPE values.
as evidence for very strong acid sites [38,39]. Thus, monofunctional
alkane isomerization reactions on SZr and WZr are controlled by
their ability to form alkenes via catalytic or stoichiometric redox
processes that do not depend on the strength of their Brønsted acid
sites. The presence of metal function, in contrast, provides constant
and known alkene concentrations along the catalyst bed.
Fig. 4 shows n-hexane isomerization turnover rates (per acces-
sible proton titrated by 2,6-di-tert-butylpyridine; Table 1) on phys-
Reliable DPE estimates are unavailable from theory for SZr, WZr,
and SAR catalysts because of the uncertain structure of their acid
sites. 2-Butanol dehydration rate constants on SZr, WZr, and SAR
are indicated by horizontal lines at their respective values in the
ordinate of Fig. 3. This line intersects that defined by elimination
rate constants on POM clusters and zeolites at a value correspond-
ing to the DPE values for SZr, WZr, and SAR catalysts.
ical mixtures of SZr (Fig. 4a) or WZr (Fig. 4b) with Pt/Al
H = 1) as a function of the (nC /H
6 2
2
O
3
(Pt/
) molar ratio in the inlet stream.
Reaction rates increased with increasing (nC /H ) molar ratios, lin-
+
6
2
early at first and then more gradually, on both SZr and WZr cata-
lysts, as also observed on POM and H-BEA catalysts [19] (Fig. 4a
and b). Varying H or nC pressures at constant (nC /H ) molar ra-
2 6 6 2
tios did not influence isomerization turnover rates. These data are
consistent with quasi-equilibrated dehydrogenation–hydrogena-
tion equilibrium and with the catalytic sequence discussed next.
Quasi-equilibrated hydrogenation–dehydrogenation steps on
the Pt function and elementary steps involved in the isomerization
of the resulting alkenes on Brønsted acid sites are depicted in
Scheme 3. Dehydrogenation forms equilibrated mixtures of linear
hexene isomers (Scheme 3, Step 1), which are protonated on
Brønsted acid sites to form the corresponding alkoxides (Scheme
3, Step 2). The position of the double bond in n-hexene isomers
is kinetically-irrelevant because hydride shifts within alkoxide
intermediates are fast and quasi-equilibrated. Alkoxides isomerize
via protonated dialkylcyclopropane transition states [17] to form
monobranched isomers (Scheme 3, Step 3), which deprotonate to
form isoalkenes (Scheme 3, Step 4) that hydrogenate on Pt (Scheme
3, Step 5) to complete the bifunctional isomerization cycle. At (suf-
In this manner, we can rigorously estimate a value for the acid
strength of these materials, specifically at the conditions in which
they operate as catalysts and as a specific quantity. These values
rank a broad range of solid acids in terms of experimentally elusive
acid strengths; they also provide a method to benchmark the fidel-
ity of various structures suggested for acid sites in SZr, WZr, and
SAR solids by comparing these DPE values with those calculated
for these hypothetical site structures. The DPE estimates are
ꢀ1
ꢀ1
ꢀ1
1
165 kJ mol , 1185 kJ mol , and 1154 kJ mol
for SZr, WZr,
and SAR. SZr and SAR are similar in acid strength, but slightly
stronger than zeolites and weaker than Keggin-type POM clusters.
The acid strength of WZr resembles that for sites in H-BEA at the
conditions of alkanol dehydration catalysis. DPE values for Keggin
POM and H-BEA represented in Eq. (2) correspond to sites with
uniform structure. A distribution of sites (and DPE values), possible
for SZr, WZr, and SAR, makes the DPE values reported here the
reactivity-weighed averages that are consequential for their cata-
lytic behavior.
+
ficiently) high Pt/H ratios, dehydrogenation occurs reversibly and
much more rapidly than isomerization; therefore, alkene and iso-
alkene concentrations remain in equilibrium with their respective
alkanes. On the sample with the highest volumetric n-hexane
These DPE estimates assume that the extent to which
DEint com-
ꢀ2
pensates DPE terms, as reflected by Eq. (2) and the data shown in
Fig. 3 for materials of known structure, is also applicable to SZr,
3 2
isomerization rate (H PW/SiO ; 0.04 POM nm ), isomerization
+
rates became independent of Pt content for Pt/H ratios above
+
WZr, and SAR. We consider this to be a reasonable assumption.
ꢄ0.2; [19] thus, equilibrium is certain at the Pt/H ratios of unity
+
n+
D
E
int estimates calculated for the stabilization of Na cations by
2
used here. Co-impregnated Pt/H8ꢀnX W/SiO samples and physi-
the deprotonated anionic conjugate base for several acids with
3 2 2 3
cal mixtures of H PW/SiO with Pt/Al O gave identical isomeriza-
tion rate constants, thus ruling out artifacts caused by intrapellet
alkene gradients [19].
The assumptions of alkane dehydrogenation (Scheme 3, Steps 1
and 5) and alkene protonation/alkoxide deprotonation equilibria
(Scheme 3, Steps 2 and 4) are consistent with data [19,21] and
DFT calculations [17]. Irreversible and kinetically-relevant alkoxide
isomerization steps (Scheme 3, Step 3) give the rate equation:
ꢀ1
ꢀ1
DPE values ranging from 1080 kJ mol (HC
5
(CN)
5
) to 1480 kJ mol
ꢀ1
ꢀ1
(
HF) gave
D
E
int values of ꢀ397 kJ mol to ꢀ643 kJ mol ; these
ꢀ1
values showed a maximum deviation of ±15 kJ mol from the
estimated by the best linear fit of int as a function of DPE (see sup-
porting information for calculated values of DPE and int). These
D
E
int
D
E
D
E
deviations provide an upper bound for the expected deviations of
the DPE estimates for SZr, WZr and SAR from true values, because
SZr, WZr, and SAR all contain oxo-type species, as in the case of
½
nC
6
ꢃ
ꢃ
þ
the POM clusters and zeolites from which
were established.
D
Eint–DPE correlations
kisomKprotKdehy
½H ꢃ
½
H
2
r
isom
¼
ð3Þ
½
K
dehy
H
nC
6
ꢃ
1
þ Kprot
½
2
ꢃ

6
0
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
12
and Kprot were determined by regressing rate data to the form of
Eq. (3), while Kdehy was obtained from thermodynamic data [41].
Fig. 4 shows n-hexane isomerization rates predicted from Eq.
(3) using values for kisom and Kprot obtained by regressing rate data
to the form of Eq. (3) for SZr (Fig. 4a) and WZr (Fig. 4b) (dashed
curves). The excellent agreement between predicted and measured
rates (Fig. 4) confirms the accuracy of Eq. (3) in describing n-hex-
ane isomerization rates on SZr and WZr. SAR showed a very differ-
ent kinetic behavior, which we address at the end of this section.
Hydrogen transfer from alkanes to alkoxides, required to desorb
the latter as alkanes, can propagate carbenium ion cycles without
requiring alkane dehydrogenation on SZr [25]. These hydride
transfer steps would lead to isomerization rates proportional to al-
1
0
8
6
4
2
0
kane concentrations and independent of H
diction with the linear dependence on (nC
2
pressures, in contra-
/H ) ratios observed
6
2
and predicted from Eq. (3). This kinetic dependence reflects the fact
that alkanes (instead of alkenes) determine surface hexoxide con-
centrations. Thus, we conclude that hydride transfer pathways are
inconsequential at the conditions of our measurements.
(
a)
0
.0
0.1
0.2
0.3
0.4
0.5
(
n-Hexane/H ) Molar Ratio
Brønsted acid sites form on WZr via stoichiometric reduction of
2
WO
31,42,43]. The formation of these temporary acid sites via reduc-
tive processes led to a first-order dependence of o-xylene isomer-
ization rates on H pressure, even though H is not required
mechanistically or stoichiometrically) for o-xylene isomerization
turnovers [43]. The different kinetic dependences on H pressures
for o-xylene isomerization (first-order) and n-hexane isomeriza-
/H )) reflect the essential absence of unsaturated
x
domains with concomitant formation of acidic O-H groups
8
6
4
2
0
[
2
2
(
2
tion (ꢄ(nC
6
2
species during isomerization catalysis (<1 Pa hexenes); such unsat-
urated hydrocarbons act as oxidants by scavenging hydrogen from
surfaces, thus reversing stoichiometric reduction steps that form
Brønsted acid sites. The higher oxidant concentrations (ꢄ1 kPa)
prevalent during o-xylene isomerization remove hydrogen atoms
2
more rapidly and thus require H activation to control the number
of Brønsted acid sites during steady-state catalysis. A more de-
tailed discussion of the effects of oxidant/reductant concentrations
on the number of temporary Brønsted acid sites is included in the
supporting information section.
n-Hexane isomerization rates cannot be described by Eq. (3) on
6 2
sulfonic acid resins (SAR). The (nC /H ) ratios used on SZr and WZr
catalysts (0.05–1; Fig. 4) did not give detectable isomerization
rates on SAR at 473 K. The use of pure n-hexane reactants, how-
ever, led to detectable formation of isohexanes. Fig. 5a shows n-
hexane isomerization rates as a function of equilibrium n-hexene
pressures (combined for all n-hexene isomers) on SAR and SZr.
Above ꢄ35 Pa n-hexenes pressures, n-hexane isomerization rates
resembled those measured on SZr at much lower pressures
(b)
0.0
0.5
1.0
1.5
2.0
(
n-Hexane/H ) Molar Ratio
2
+
Fig. 4. n-Hexane isomerization rates (per accessible H ) as a function of (nC
6
/H
), 60 kPa H
(d), 60 kPa H (s),
)). The dashed lines represent fits to the rate data using Eq. (3).
2
) for
+
a physical mixture of (a) SZr and Pt/Al
) or (b) WZr and Pt/Al (473 K, Pt/H = 1, 30 kPa H
0 kPa H
2
O
3
(473 K, Pt/H = 1, 30 kPa H
2
(
N
2
+
(
8
D
2
O
3
2
2
2
(O
Table 1
,6-Di-tert-butylpyridine uptake measured by titration during 2-butanol dehydration
0.2 kPa 2-butanol, 0.9 Pa 2,6-di-tert-butylpyridine) and n-hexane isomerization
reaction (473 K, 0.2 (nC /H ), 1.4 Pa 2,6-di-tert-butylpyridine) (Figs. 2 and 6).
(
ꢄ0.04 Pa). Isomerization rates increased with increasing n-hexene
2
(
concentrations on SAR, first linearly and then more gradually. In
contrast with SZr, which deactivates at n-hexene pressures above
ꢄ0.1 Pa (Fig. 5b), because of the high prevalent alkoxide concentra-
tions and concomitant oligomerization side reactions, SAR cata-
lysts did not deactivate even at hexene pressures >35 Pa.
6
2
2
2
,6-Di-tert-butylpyridine uptake during
-Butanol dehydration
n-Hexane isomerization
ꢀ
1
ꢀ1
SZr
WZr
SAR
0.16 (S-atom)
0.06 (W-atom)
1.00 (H )
0.28 (S-atom)
SAR catalysts consist of agglomerates of polymeric persulfonic
ꢀ
1
ꢀ1
0.05 (W-atom)
2
acid resins and colloidal SiO particles, a morphology that differs
+
ꢀ1
Not measured
markedly from that in inorganic solids, such as POM clusters, zeo-
lites, SZr, and WZr. The protons within perfluorosulfonic acid resin
particles cannot be reached by non-polar species [33], but become
accessible when polar reactants, such as alkanols and ethers, ex-
pand the resin structures, as shown by the similar kinetic behavior
of SAR, SZr, and WZr materials in the case of 2-butanol dehydration
reactions (Fig. 1). The restricted access of hexenes (and hexanes)
causes intraparticle concentration gradients and low local hexane
and hexene concentrations. These lead, in turn, to low isomeriza-
tion rates, but also to even larger effects on the rates of bimolecular
oligomerization reactions (Fig. 5b).
which also accurately describes rate data also on Keggin-type POM
clusters [19] and zeolites [40]. In Eq. (3), kisom is the isomerization
rate constant (Scheme 3, Step 3), Kprot is the equilibrium constant
for n-alkoxide formation (Scheme 3, Step 2), Kdehy is the dehydroge-
nation equilibrium constant (Scheme 3, Step 1), and [H ] is the con-
centration of protons accessible during catalysis, reported here from
+
titration of acid sites by 2,6-di-tert-butylpyridine. Values for kisom

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
61
Kdehy
H2
+
(
Step 1) Dehydrogenation
Metal catalyzed
Kprot
H
A
+
(Step 2) Protonation
A
CH
H
+
A-
CH2
CH
k
isom
(
Step 3) Isomerization
A
A
A
Kdeprot
H
A
+
(
Step 4) Deprotonation
Khy
H2
+
(
Step 5) hydrogenation
Metal catalyzed
Scheme 3. Proposed sequence of elementary steps for n-hexane isomerization on bifunctional metal–Brønsted acid catalysts [19].
3
.5. Measurement of the number of accessible protons during n-hexane
isomerization by 2,6-di-tert-butylpyridine titration on SO –ZrO and
WO –ZrO
In contrast, titrant uptakes required to suppress reaction rates
4
2
on WZr were smaller during n-hexane isomerization
þ
ꢀ1
x
2
(0:05H WO
x
,
Table 1) than during 2-butanol dehydration
+
ꢀ1
(
0.06H W-atom , Table 1). These data appear to reflect OH
Fig. 6 shows n-hexane isomerization turnover rates as Brønsted
acid sites were titrated by 2,6-di-tert-butylpyridine on SZr–Pt/
Al O mixtures. The left panel in Fig. 6 shows n-hexane isomeriza-
2 3
recombination events, which remove Brønsted acid sites and occur
more readily at the higher temperatures (473 K vs. 343 K for 2-
butanol dehydration) and anhydrous conditions of isomerization
catalysis. These reaction-dependent titrant uptakes on WZr and
SZr illustrate the essential requirement that the number of accessi-
ble Brønsted acid sites be measured during reaction and that theo-
retical treatments consider hypothetical structures consistent with
these effects of reaction environment on the structure of acid sites.
tion rates before introducing the titrants, while the right panel
shows rates as a function of the amount of titrant adsorbed. n-
Hexane isomerization rates decreased linearly as 2,6-di-tert-butyl-
pyridine titrated Brønsted acid sites (Fig. 6). The titrant uptake
required to suppress isomerization rates was 0.28 per S-atom in
the sample (Fig. 6 and Table 1). 2,6-di-tert-Butylpyridine titrations
on WZr also led to a linear decrease and to the complete suppres-
sion of isomerization rates. The measured titrant uptake required
to suppress isomerization rates was 0.05 per W-atom in WZr
3.6. Deprotonation energy estimates from n-hexane isomerization rate
4 2 x 2
constants for SO –ZrO and WO –ZrO
(Table 1).
Fig. 7 shows measured n-hexane isomerization rate constants
The amount of titrant required to titrate all sites on SZr was
(in a log scale) on Keggin-type POM clusters supported on SiO
(H8ꢀnX W12O40, X = P , Si , Al , Co ; in order of increasing
2
ꢀ1
n+
5+
4+
3+
2+
slightly larger during n-hexane isomerization (0.28 S-atom
,
ꢀ1
Fig. 6) than during 2-butanol dehydration (0.16 S-atom , Fig. 2).
These results may seem surprising at first, because water has been
implicated in the conversion of Lewis acids into Brønsted acid sites
on these materials; [44] thus, we expect a higher density of
Brønsted acid sites during 2-butanol dehydration (ꢄ10–20 Pa
DPE) and on H-BEA zeolite as a function of their DPE values (calcu-
lated from DFT). As in the case of 2-butanol dehydration reactions,
the line through the data suggests a common exponential depen-
dence of isomerization rate constants on DPE [19], as expected
from the thermochemical cycle formalism in Section 3.3. n-Hexane
isomerization rate constants on SZr and WZr are shown as horizon-
tal lines defined by their measured rate constants at the ordinate,
because their uncertain structures make theoretical DPE estimates
unreliable. The intersection of these lines with that regressed
through the data on POM and zeolite catalysts of known structure
allows DPE to be estimated for SZr and WZr at the conditions of
2
H O formed) than in the anhydrous environment prevalent during
isomerization catalysis. It is possible that any Brønsted acid sites
formed by H O interactions with Lewis acid sites are too weak to
2
bind 2,6-di-tert-butylpyridine irreversibly and to catalyze 2-buta-
nol dehydration. DPE estimates derived from 2-butanol dehydra-
tion rate constants are higher (and acid sites are thus weaker)
than those derived from hexane isomerization rate constants (see
Section 3.7 for data and detailed discussion). Thus, we conclude
that the reaction environment influences the number and strength
of acid sites and, by inference, their local structure.
ꢀ1
isomerization catalysis. These DPE values are 1110 kJ mol for
ꢀ1
SZr and 1120 kJ mol for WZr, indicating that SZr and WZr are
stronger acids than zeolites and similar in strength to the weakest
6 5
POM acids (H CoW12O40 and H AlW12O40; Table 2). These DPE

6
2
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
5
4
4
3
3
2
2
1
1
.0
.5
.0
.5
.0
.5
.0
.5
.0
1
1
2
0
8
6
4
2
0
(
a)
Start of Titration
0
.0 0.1 0.2
40.0 60.0 80.0 100.0
=
nC Pressure (Pa)
6
0.5
.0
0
1
1
1
1
6
4
2
0
8
6
4
2
0
0
10
20
03 .00
0.1
0.2
0.3
SZr
(
b)
=
Reaction
Time (min)
Cumulative Adsorbed Titrant
(molecules / S-atom)
(
0.1 Pa nC )
6
Fig. 6. Titration of acid sites for a physical mixture of SZr and Pt/Al
2
O
3
during n-
function of the
cumulative amount of titrant (2,6-di-tert-butylpyridine) adsorbed (473 K, 0.2 (nC
), 1.4 Pa 2,6-di-tert-butylpyridine).
+
hexane isomerization at 473 K. Reaction rates (per H ) as
a
6
/
H
2
SAR
60 Pa nC )
=
ꢀ1
ꢀ1
(
1110 kJ mol
for SZr and 1120 kJ mol
for WZr) (Table 2).
(
6
Such differences in DPE values are significant and of the same
magnitude as the entire range of DPE values covered by the POM
compositions examined (56 kJ mol
ꢀ1
ꢀ1
range; 1087 kJ mol
to
ꢀ1
n+
1
143 kJ mol
for H8ꢀnX W12O40, with X = P, Si, Al, Co). These
reaction-dependent DPE values may reflect inaccurate data or
inappropriate assumptions in our approach, neither of which are
evident from our data or from their analysis. They may reflect
instead effects of reaction environment or treatment conditions,
which differ between elimination and isomerization catalysis, on
the number, structure, and strength of Brønsted acid sites.
One assumption in our method for estimating DPE for unknown
structures is that the
depend on DPE either weakly for all materials or in the same man-
ner as in the case of Keggin clusters and H-BEA samples, on which
the reactivity-DPE relations in Figs. 3 and 7 are based. For example,
different extents of reactant stabilization by various acids, reflected
0
5
10 15 20 25 30
Reaction Time (min)
+
¼
6
Fig. 5. (a) n-Hexane isomerization rates (per H ) as a function of n C (n-hexene)
pressure on physical mixtures of SAR (d) or SZr (
N
2 3
) and Pt/Al O mixtures (473 K,
Pt/H+ = 1). The dashed line represents a fit to the rate data using Eq. (3). (b) n-
¼
Hexane isomerization rates on SAR (d) (ꢄ60 Pa nC ) or SZr and Pt/Al
2 3
O mixtures
6
DHads terms (Eq. (2)) for SZr, WZr, and SAR
¼
¼
(
N
) (ꢄ0.1 Pa nC ) as a function of reaction time. nC pressures were calculated
6
6
from nC
equilibrium.
6
(and H
2
pressures) at reaction temperatures assuming (de)hydrogenation
values differ somewhat from those obtained from 2-butanol dehy-
dration rate constants on WZr and SZr (Section 3.3). Next, we inter-
pret these differences in DPE values in terms of plausible effects of
treatment or reaction conditions on the identity and strength of
Brønsted acid sites. These effects, taken together with DPE esti-
mates, provide useful benchmarks for possible structures of
Brønsted acid sites at conditions imposed by specific catalytic
reactions.
in the DHads term in Eq. (2), could lead to different DPE estimates
from n-hexane isomerization and 2-butanol dehydration reactions
(Eq. (2)). Hexoxide intermediates, with significant covalent charac-
ter, may respond to catalyst composition or structure more sensi-
tively than the H-bonded butanol species involved in dehydration
ꢀ1
ꢀ1
reactions. The Kprot values for SZr and WZr (4.4 Pa and 3.1 Pa
,
respectively) are smaller than those for POM clusters with similar
ꢀ
1
acid strength (8.9 Pa
(
,
H
5
AlW12
40
O ) and those for H-BEA
ꢀ
1
21.9 Pa ) (Table 3). These data show that the stability of cova-
3
.7. Implications of DPE values and of their dependence on reaction
condition for the nature of Brønsted acid sites on SO –ZrO and
WO –ZrO
lently-bound alkoxides is not related to the acid DPE [19]. DFT cal-
culations showed that the stability of H-bonded alkanol reactants
(interacting with H-atoms in the acid) decreased only slightly with
increasing DPE for POM clusters [14]. On both POM clusters and H-
BEA, covalent n-hexoxide species became slightly more stable with
increasing DPE (Table 3) [19]. These weak effects indicate that the
different sensitivities of 2-butanol dehydration and hexane isom-
4
2
x
2
The DPE estimates of 1165 kJ mol 1 (SZr) and 1185 kJ mol
WZr) obtained from 2-butanol dehydration rate constants are lar-
ger than those determined from isomerization rate constants
ꢀ
ꢀ1
(

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
63
Table 3
DPE values or estimates and equilibrium constants for alkoxide formation (Kprot
)
)
(Scheme 3, Step 2).
DPE (kJ mol 1
ꢀ
)
Kprot (Pa
ꢀ1
1
00
H
H
H-BEA
SZr
3
PW12
AlW12
O
O
40
40
1087
1121
1185
4.1a
a
5
8.9
21.9a
4.4
SZr
b
1110
WZr
WZr
1120b
3.1
1
0
a
See Ref. [19].
See also Table 2.
b
1
SZr and WZr. We discuss next, first for WZr and then for SZr,
other possible causes of reaction-dependent DPE values.
On WZr, DPE values are smaller during n-hexane isomerization
than during 2-butanol dehydration catalysis. These differences
may reflect some dehydroxylation, resulting from the anhydrous
environment and the higher temperatures prevalent in isomeriza-
tion catalysis. Dehydroxylation is known to increase acid strength
0
.1
DPE estimates
1150
(
decrease DPE) for residual protons in POM clusters, because of a
concomitant decrease in the charge density of the anionic conju-
gate base as the number of protons decreases [14]. This leads to
a decrease in the energy required for charge separation because
of the greater stability of the anion, as discussed in Section 3.3.
0.01
1200
1100
-
1
Deprotonation Energy (kJ mol )
Dehydroxylation of H
3
PMo12
O40 (to HPMo12O39) and of reduced
Fig. 7. n-Hexane isomerization (kisom) (Scheme 3, Step 3) rate constants as a
function of deprotonation energy (
HA is the acid and A is the conjugate base) calculated by DFT for physical mixtures
H
5
PMo12 40 (to H PMo12O
O
3
39) clusters decreased DPE values by
ꢀ
+
D
E
rxn for HA ? A + H (non-interacting), where
ꢀ1
ꢀ
ꢄ16 kJ mol in each case [14]. Partial dehydroxylation is consis-
tent with the smaller number of protons measured by 2,6-di-
tert-butylpyridine during isomerization than during dehydration
+
of Pt/Al
2
O
3
(473 K, Pt/H = 1) and 0.04H
3 4 5
PW/Si (d), 0.04H SiW/Si (j), 0.04H AlW/Si
(N
), 0.04H
6
CoW/Si ( ), and H-BEA (r) Measured n-hexane isomerization rate
.
constants on SZr and WZr are shown as solid horizontal lines. kisom values where
obtained by fitting Eq. (3) to the kinetic data and normalization by the 2,6-di-tert-
butylpyridine uptake, which corresponds to the number of accessible protons.
+
ꢀ1
catalysis (0.05 vs. 0.06H W-atom ; Table 1).
ꢀ1
On SZr, DPE values were 55 kJ mol smaller from n-hexane
isomerization than from 2-butanol dehydration rate data, suggest-
ing that, as in the case of WZr, SZr is a stronger acid under the
anhydrous conditions of isomerization catalysis. n-Hexane isomer-
ization rates increased ꢄ50-fold when SZr catalysts were treated in
He at 673 K before reaction (the procedure used for the data shown
in Figs. 4–7) than when treated in He only at 473 K. The isomeriza-
tion rate constant measured on SZr samples treated at 473 K, when
placed on the regressed line in Fig. 7, leads to a DPE estimate of
erization to DPE do not arise from different effects of DPE on
adsorbed reactants as a result of their different charge.
We can exclude any significant role of DPE in reactant stability
even more convincingly by examining the dependence of the
k
isom
Kprot product (instead of kisom) on DPE for known acid struc-
ꢀ1
ꢀ1
tures. The value of kisom prot depends only on the enthalpy and en-
K
1157 kJ mol , which is 47 kJ mol larger than on the SZr sample
ꢀ1
tropy of transition states relative to gas-phase hexenes and does
not include any contributions from the energetics of adsorbed hex-
treated at 673 K (1110 kJ mol ). In fact, this DPE estimate for the
sample treated at 473 K is very similar to that obtained on SZr dur-
ꢀ1
oxides. These kisom
K
prot data (Table 2; see also kisom
K
prot vs. DPE fig-
ing 2-butanol dehydration at 343 K (1165 kJ mol ). We note that
2-butanol dehydration rate constants were unaffected by thermal
treatment before reaction (673 K for 2 h vs. untreated), apparently
because the extent of hydroxylation (and the acid strength) is
established by the alkanol dehydration environment itself. The
conditions during alkanol dehydration favor more hydroxylated
surfaces and weaker acid sites than the anhydrous conditions of
ꢀ1
ure in supporting information) give DPE values of 1112 kJ mol
ꢀ1
and 1130 kJ mol for SZr and WZr, respectively; these are the
ꢀ1
same as those obtained from kisom on SZr and ꢄ10 kJ mol higher
on WZr (Table 2). DPE estimates from kisom prot, which do not
depend on the stability of adsorbed intermediates, remain much
smaller (ꢄ55 kJ mol ) for isomerization than for elimination on
both WZr and SZr. We conclude, therefore, that DPE effects on
reactant stability are not responsible for the differences in DPE val-
ues derived from isomerization and dehydration rate data on both
K
ꢀ1
isomerization reactions. DFT calculations show that H O adsorbs
3 40
weakly even on strongly acidic H PW12O clusters
2
ꢀ1
(D
H
ads = ꢀ67.5 kJ mol ) [35]. Thus, thermal treatments are not
likely to increase acid strength by merely causing water desorp-
tion. These strong effects of thermal treatment and reaction envi-
ronment on hydroxylation and acid strength indicate that
estimates of acid strength remain equivocal in value, and perhaps
even kinetically-inconsequential, except when measured during
catalysis. We conclude that the removal of OH groups by thermal
treatment at 673 K leads to stronger Brønsted acid sites, which
are preserved during anhydrous alkane isomerization; OH groups
can re-form, however, during dehydration reactions of protic reac-
tants, such as alkanols.
Table 2
DPE estimates from 2-butanol dehydration and n-hexane isomerization rate con-
stants (from measured data and correlations between reactivity and DPE in Figs. 3 and
7
).
Deprotonation energy values (kJ molꢀ1)
2
-Butanol dehydration
n-Hexane isomerization
from kisom
from k
2
from kisomK
prot
SZr
WZr
SAR
1165
1185
1154
1110
1120
1112
1130
We discuss next how interactions of H
cules such as H O and SO (relevant to Brønsted acid sites in SZr)
influence their DPE value. Similar interactions may give rise to
2 4
SO species with mole-
2
3

6
4
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
the observed effects of pretreatment on rates and to the differences
in DPE values derived from alkanol dehydration and alkane isom-
erization rate data (Table 2).
(ꢀ55 kJ molꢀ1). Treatment at high temperatures and anhydrous
conditions favor Brønsted–Lewis conjugate acids via dehydroxyla-
3
tion, which forms SO -like species from sulfonic acid groups in SZr.
DPE values derived from 2-butanol dehydration and n-hexane
In contrast, protic species, such as alkanols and water, coordinate
to Lewis centers and disrupt Brønsted–Lewis acid pairs, via reac-
tions analogous to:
ꢀ
1
ꢀ1
isomerization rate constants on SZr (1165 kJ mol , 1110 kJ mol
respectively) are smaller than for isolated gas-phase H SO mono-
mers (1293 kJ mol ). Two types of interactions among H SO
O, and SO decrease DPE values for H SO monomers. Coopera-
tivity effects decrease DPE by forming (H SO dimers
) (Scheme 4a), in which
the deprotonated anion is stabilized by another H SO molecule via
H-bonding. Strong effects of H O molecules on acid strength for
liquid H SO –H O mixtures are also well-known [45]. The DPE of
SO monomers decreases also upon formation of Brønsted–
Lewis conjugate acids, such as in complexes consisting of a
SO monomer interacting with one (H , Scheme 4b) or
two (H
,
2
4
ꢀ1
2
4
,
H
2
S
2
O
7
þ H
2
O ! 2H
2
SO
4
ð5Þ
H
2
3
2
4
4 2
)
which increases the number of Brønsted acid sites but also their
2
ꢀ1
ꢀ1
ꢀ1
46
ꢀ1
(
1220 kJ mol vs. 1293 kJ mol for H
2
SO
4
DPE values (from 1177 kJ mol for H
2 2
S O7 to 1293 kJ mol for
H
2
4
SO monomers [46]). These trends are consistent with the stron-
2
4
ger acids found in SZr at the anhydrous conditions of alkane isom-
erization than when butanol is present in dehydration reactions.
These conclusions provide an interpretation for the thermal treat-
ments required to detect catalytic reactivity in SZr samples. Ther-
mal treatments cause dehydroxylation and a consequent increase
in the acid strength of residual OH groups via cooperativity effects
2
2
4
2
H
2
4
H
2
4
2 2 7
S O
S
2 3
O10, Scheme 4b) SO
3
molecules, which act as Lewis acids:
x
as SO groups form and interact with sulfonic acid groups (instead
of water) (Scheme 4a). Dehydroxylation can also form Brønsted–
Lewis conjugate acids via condensation (analogous to the reverse
H
2
SO þ nSO
4
3
$ H
2
S
nþ1
O
4þ3n
ð4Þ
ꢀ1
These Brønsted–Lewis acid pairs give DPE values of 1177 kJ mol
46
ꢀ1
46
10
ꢀ1
for H
SO
larger anionic clusters, caused by association or complexation of
SO monomers, increases acid strength, either via cooperativity
effects or via the formation of Brønsted–Lewis conjugate acids.
O, in contrast, decomposes (H SO dimers and Brønsted–Lewis
2
S
2
O7 and 1127 kJ mol for H
2
S
3
O
(cf. 1293 kJ mol for
of Eq. (4)) or decomposition of H
sulfonic acid groups to form grafted analogs of strongly acidic
species (Eq. (4)) (Scheme 4b).
Cooperativity effects and reversible formation of Brønsted–
Lewis acid pairs are possible only for weakly-bound SO species,
which can interact with Lewis acids (e.g. SO ) to decrease DPE or
with Lewis bases (e.g. H O) to increase DPE. Active acid sites in
2 4 3
SO to SO , which interacts with
H
2
4
monomers [46]). More effective electron delocalization over
2 2 7
H S O
H
2
4
x
H
2
2
4
)
2
3
acid pairs and leads to an increase in the electron density at anionic
clusters, thus decreasing acid strength.
2
SZr may resemble those in supported phosphoric acid catalysts,
for which P-NMR [31] indicate that species that interact with basic
probe molecules consist of H PO monomers and H PO (PO ) olig-
3 4 3 4 3 n
omers [47]. Silicon phosphate species with strong support interac-
tions were unaffected by these molecules because they do not act
The effects of cooperativity (ꢀ73 kJ molꢀ1) and Brønsted–Lewis
ꢀ
1
complexation (ꢀ115 kJ mol ) on DPE are stronger for H
2
SO
4
monomers than the effects of reaction environment on 2-butanol
dehydration and n-hexane isomerization on DPE reported here
as Brønsted acids [47]. Similarly, species resembling H
mers and H SO (SO oligomers, instead of SO species strongly
bound on ZrO , may act as the active acid sites on SZr, consistent
2 4
SO mono-
2
4
3
)
n
x
2
with the strong effects of pre-treatment and reaction conditions
and with the arguments proposed here to explain these effects.
Some sulfate groups in SZr (ꢄ40%) can be removed by contact
with liquid water at 298 K, indicating that they interact weakly
with ZrO surfaces [25]. Such SO can react with SO or H O to form
2 x 3 2
or consume Brønsted–Lewis conjugate acids. The formation of such
Brønsted–Lewis conjugate acids was proposed to account for the
weakening of antisymmetric S–O stretches in oligomeric SO
x
spe-
cies with increasing H O concentration, as a result of hydrolysis
2
of S–O–S bonds and decomposition of Brønsted–Lewis conjugate
acids [48].
We conclude that the strengthening of acid sites via Brønsted-
acid/Lewis-acid complexation upon dehydroxylation and their
weakening by interactions with Lewis bases account for the effects
of reaction environments on DPE values for WZr. Anhydrous isom-
ꢀ1
erization conditions give lower DPE values (1120 kJ mol , Table 2)
ꢀ
1
than those prevailing during dehydration reactions (1185 kJ mol
,
Table 2) because the extent of dehydroxylation differs in these two
reaction environments.
Rate constants for n-hexane isomerization and 2-butanol dehy-
ꢀ
ꢀ
1
1
dration on SZr indicate that DPE values (1120 and 1165 kJ mol
are significantly smaller than for H SO monomers (1293 kJ mol
)
)
2
1
4
ꢀ
or (H
lated for H
(
2 4 2
SO ) dimers (1220 kJ mol ), but lie between those calcu-
ꢀ1
2
SO
4
–SO
3
(1177 kJ mol ) and H
1127 kJ mol ) in the gas phase. Our DPE estimates are signifi-
cantly smaller than those reported from theoretical estimates for
hypothetical SZr structures at various SO and H O pressures
1353–1499 kJ mol ), which are also larger than for H SO mono-
2 4 3 2
SO –(SO ) complexes
ꢀ
1
3
2
ꢀ1
(
2
4
Scheme 4. H dimer (cooperativity effect) (a), Brønsted–Lewis conjugate acids
2
SO
4
mers [49]. These previous studies have suggested that SZr contains
H
2
2
S O
7
(b) and H S O10 (c). The colored surfaces correspond to the van der Waals
2 3
2
7
ꢀ
active sites consisting of pyrosulfate anions ðS
2
O
Þ bound in a tri-
surfaces of the molecules. The color scheme shows whether areas on these surfaces
are electropositive (blue) or negative (red).
dentate configuration on ZrO
2
with two vicinal bridging hydroxyl

J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
65
groups on ZrO
this structure, the protons are associated with ZrO
the SO moiety and are very weakly acidic (1418 kJ mol DPE)
49,50]. These DPE values derived from DFT for SZr [49,50] would
lead to undetectable isomerization rates according the correlations
shown in Fig. 7. Thus, we conclude that such species may indeed
represent the most stable structures, but their stability and strong
2
support and one adsorbed water molecule [49]. In
reaction environment, WZr acid sites are either stronger (n-hexane
isomerization) or nearly equal (2-butanol dehydration) to those of
acid zeolites. For SZr and WZr, lower DPE estimates during n-hex-
ane isomerization relative to 2-butanol dehydration demonstrate
the subtle effects of reaction environment on catalytic acid sites
and the importance of utilizing methods that measure acid
strength at catalytically relevant conditions.
2
and not with
ꢀ1
x
[
interactions with ZrO
2
render them essentially unreactive and
inconsequential in Brønsted acid catalysis by SZr.
Acknowledgments
DPE estimates from the n-hexane isomerization and 2-butanol
dehydration rate constants indicate that the identity of the active
Brønsted acid sites of SZr is sensitive to reaction conditions and
to treatment protocols. Their acid strength is determined by the
extent to which Brønsted–Lewis conjugate acid-type interactions
occur as the extent of dehydroxylation evolves in catalytic
environments.
Support by the Chemical Sciences, Geosciences, Biosciences
Division, Office of Basic Energy Sciences, Office of Science US
Department of Energy under grant number DE-FG02-03ER15479
is gratefully acknowledged. We thank Professors Matthew Neurock
and Michael Janik (University of Virginia) for their contributions to
the theory that underpinned our study, Dr. Cindy Yin (UC-Berke-
We have shown here that the acid strength can be assessed in
terms of quantitative values for deprotonation energies for
Brønsted acids with unknown site structures from the dynamics
of catalytic reactions. This approach requires mechanistic interpre-
tations of measured rates, knowledge about the number of acid
sites present during catalysis, and a correlation between rate con-
stants for the kinetically-relevant elementary steps and deprotona-
tion energies derived from data on solid acids with known
structure. The method proposed is not specific to any catalyst or
reaction. It probes acid strength at the prevailing reaction condi-
tions from the stability of cationic transition states, accessible
through thermochemical cycles and consequential to catalysis on
solid acids. These DPE estimates provide a rigorous benchmark to
assess the fidelity of structures proposed for acid sites in these
materials and a method to probe how such structures depend on
reaction environments.
5 6
ley) for the synthesis of bulk H A1W and H CoW clusters, and Pro-
fessor Johannes Lercher from the Technical University of Munich,
Professor Chelsey Baertsch, and Ms. Cathy Chin for the SZr, WZr
2 3
and Pt/Al O catalysts.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.03.005.
References
[1] A. Corma, Chem. Rev. 95 (1995) 559, and references therein.
[2] A. Zecchina, F. Geobaldo, G. Spoto, S. Bordiga, G. Ricchiardi, R. Buzzoni, G.
Petrini, J. Phys. Chem. 100 (1996) 16584.
[
3] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Spano, F. Geobaldo, J. Chem.
Soc., Faraday Trans. 92 (1996) 4863.
[
4] C. Paze, S. Bordiga, C. Lamberti, M. Salvalaggio, A. Zecchina, G. Bellussi, J. Phys.
Chem. B 101 (1997) 4724.
4
. Conclusions
[
[
5] F. Haase, J. Sauer, J. Phys. Chem. 98 (1994) 3083.
6] M. Hunger, Catal. Rev. – Sci. Eng. 398 (1997) 345.
DPE values for solid acids of unknown structure have been esti-
[7] J.F. Haw, T. Xu, Adv. Catal. 42 (1998) 115.
[
[
8] C. Paze, S. Bordiga, A. Zecchina, Langmuir (2000) 8139.
9] R. Buzzoni, S. Bordiga, G. Ricchinardi, G. Spoto, A. Zecchina, J. Phys. Chem. 99
mated by comparing their measured rate constants to previously
developed rigorous correlations between rate constants and DPE
on Keggin-type POM and H-BEA. This method is possible because
the structures of POM clusters and acid zeolites are known, allow-
ing their DPE values to be calculated by DFT. We have demon-
strated this approach by measuring turnover rates on SZr, WZr,
and SAR for 2-butanol dehydration and n-hexane isomerization,
which are interpreted as kinetic rate constants by independent
measurements of acid site density by chemical titration using 2,6
di-tert-butylpyridine. Although these provide specific examples,
this technique can be widely applied to any reaction whose kinet-
ically-relevant step is catalyzed by Brønsted acid sites. The kinetics
of 2-butanol dehydration on SZr, WZr, and SAR catalysts and of
n-hexane isomerization on SZr or WZr as physical mixtures with
(
1995) 11937.
[10] W.E. Farneth, R.J. Gorte, Chem. Rev. Chem. Rev. 95 (1995) 615.
[
[
[
11] N. Mizuno, M. Misono, Chem. Rev. 98 (1998) 199.
12] J. Macht, M.J. Janik, M. Neurock, E. Iglesia, J. Am. Chem. Soc. 130 (2008) 10369.
13] J. Macht, M.J. Janik, M. Neurock, E. Iglesia, Angew. Chem., Int. Ed. 46 (2007)
7864.
[14] M.J. Janik, J. Macht, E. Iglesia, M. Neurock, J. Phys. Chem. C 113 (2009) 1872.
15] M. Boronat, P. Viruela, A. Corma, J. Phys. Chem. A 102 (1998) 982.
16] V.B. Kazansky, Catal. Today 51 (1999) 419.
[
[
[
17] T. Demuth, X. Rozanska, L. Benco, J. Hafner, R.A. van Santen, H. Toulhoat, J.
Catal. 214 (2003) 68.
[
[
[
18] M.J. Janik, R.J. Davis, M. Neurock, Catal. Today 116 (2006) 90.
19] J. Macht, R.T. Carr, E. Iglesia, J. Am. Chem. Soc., in press.
20] P.B. Weisz, in: D.D. Eley, P.W. Selwood, P.B. Weisz (Eds.), Advances in Catalysis
and Related Subjects, vol. 13, Academic Press, New York, 1962, p. 137.
21] F.J.M.M. de Gauw, J. van Grondelle, R.A. van Santen, J. Catal. 206 (2002) 295.
22] M. Hino, K. Arata, Chem. Commun. (1980) 851.
[
[
[
Pt/Al
clusters and H-BEA catalysts, allowing for their interpretation as
kinetic rate constants. Physical mixtures of SAR with Pt/Al
2 3
O were identical to those obtained for Keggin-type POM
23] E. Iglesia, S.L. Soled, G.M. Kramer, J. Catal. 144 (1993) 238.
[24] X. Song, A. Sayari, Catal. Rev. – Sci. Eng. 38 (1996) 329.
[
[
25] X. Li, K. Nagaoka, L.J. Simon, R. Olindo, J.A. Lercher, J. Catal. 232 (2005) 456.
26] K. Arata, M. Hino, in: M.H. Phillips, M. Ternan (Eds.), Proceedings of the 9th
International Congress on Catalysis, The Chemical Institute of Canada, Ottawa,
1988, p. 1727.
27] E. Iglesia, D.G. Barton, S.L. Soled, S. Miseo, J.E. Baumgartner, W.G. Gates, G.A.
Fuentes, G.D. Meitzner, in: J.W. Hightower, W.N. Delgass, E. Iglesia, A.T. Bell
2 3
O ,
however, exhibited a low reactivity in n-hexane isomerization,
presumably because of its low Brønsted-acid site accessibility in
non-polar environments. The deprotonation energy (DPE) values
estimated using the measured rate constants and the correlations
established earlier for Keggin-type POM clusters and zeolite H-
[
(
Eds.), Proceedings 11th International Congress of Catalysis Stud. Surf. Sci.
Catal., vol. 101, Elsevier, Amsterdam, 1996, p. 533.
[28] J.G. Santiesteban, J.C. Vartuli, S. Han, R.D. Bastian, C.D. Chang, J. Catal. 168
1997) 431.
ꢀ1
ꢀ1
BEA are 1110 kJ mol and 1120 kJ mol for SZr and WZr, respec-
(
tively, based on the n-hexane isomerization rate constants, and
[
[
29] M.A. Harmer, W.E. Farneth, Q. Sun, J. Am. Chem. Soc. 118 (1996) 7708.
30] C. Morterra, G. Cerrato, C. Emanuel, V. Bolis, J. Catal. 142 (1993) 349.
ꢀ1
ꢀ1
ꢀ1
1
165 kJ mol , 1185 kJ mol and 1154 kJ mol for SZr, WZr and
SAR, based on their 2-butanol dehydration rate constants. From
these estimates, we conclude acid sites on SZr and SAR are stronger
than those found in acid zeolites, but are weaker than those found
[31] C.D. Baertsch, K.T. Komala, Y.-H. Chua, E. Iglesia, J. Catal. 205 (2002) 44.
[
[
[
32] X. Li, K. Nagaoka, J.A. Lercher, J. Catal. 227 (2004) 130.
33] Q. Sun, W.E. Farneth, M.A. Harmer, J. Catal. 164 (1996) 62.
34] M.J. Frisch et al., Gaussian 03, revision C.02, Gaussian Inc., Wallingford, CT,
2004.
3 4
on H PW12O40 and H SiW12O40 Keggin clusters. Depending on its

6
6
J. Macht et al. / Journal of Catalysis 264 (2009) 54–66
[
35] M.J. Janik, K.A. Campbell, B.B. Bardin, R.J. Davis, M. Neurock, Appl. Catal. A 56
2003) 51.
36] D. Farca ßs iu, R. Leu, A. Corma, J. Phys. Chem. B 106 (2002) 928.
37] M.T. Aronson, R.J. Gorte, W.E. Farneth, J. Catal. 98 (1986) 434.
38] X.B. Li, K. Nagaoka, L.J. Simon, R. Olindo, J.A. Lercher, A. Hofmann, J. Sauer, J.
Am. Chem. Soc. 127 (2005) 16159.
39] S. Kuba, P.C. Heydorn, R.K. Grasselli, B.C. Gates, M. Che, H. Knözinger, Phys.
Chem. Chem. Phys. 3 (2001) 146.
40] A. van de Rundstraat, J.A. Kamp, P.J. Stobbelaar, J. van Grondelle, S. Krijnen, R.A.
van Santen, J. Catal. 171 (1997) 77.
[44] J.F. Haw, J. Zhang, K. Shimizu, T.N. Venkatraman, D.-P. Luigi, W. Song, D.H.
Barich, J.B. Nicholas, J. Am. Chem. Soc. 122 (2000) 12561.
[45] V.B. Kazansky, Catal. Today 73 (2002) 12.
[46] I.A. Koppel, P. Burk, I. Koppel, I. Leito, T. Sonoda, M. Mishima, J. Am. Chem. Soc.
122 (2000) 5114.
[47] T.R. Krawietz, P. Lin, K.E. Lotterhos, P.D. Torres, D.H. Barich, A. Clearfield, J.F.
Haw, J. Am. Chem. Soc. 120 (1998) 8502.
[48] C. Morterra, G. Cerrato, F. Pinna, M. Signoretto, G. Strukul, J. Catal. 139 (1994)
181.
[49] C. Breitkopf, H. Papp, X. Li, R. Olindo, J.A. Lercher, R. Lloyd, S. Wrabetz, F.C.
Jentoft, K. Meinel, S. Förster, K.M. Schindler, H. Neddermeyer, W. Widdra, A.
Hofmann, J. Sauer, Phys. Chem. Chem. Phys. 9 (2007) 3600.
[50] A. Hofmann, J. Sauer, J. Phys. Chem. B 108 (2004) 14652.
[51] P.B. Weisz, E.W. Swegler, Science 126 (1957) 31.
[52] R. Thorton, B.C. Gates, J. Catal. 34 (1974) 275.
(
[
[
[
[
[
[
[
[
41] D.R. Stull, E.F. Westrum, G.C. Sinke, The Chemical Thermodynamics of Organic
Compounds, Robert E. Krieger Publishing Company, Malabar, 1987.
42] D.G. Barton, S.L. Soled, G.D. Meitzner, G.A. Fuentes, E. Iglesia, J. Catal. 181
(
1999) 57.
43] R.D. Wilson, D.G. Barton, C.D. Baertsch, E. Iglesia, J. Catal. 194 (2000) 175.