Compensation effect in isopropanol dehydration over heteropoly acid catalysts at a gas-solid interface

Article abstract of DOI:10.1016/j.jcat.2012.06.021

Kinetics of isopropanol dehydration at a gas-solid interface over a range of bulk and supported Bronsted acid catalysts based on H₃PW ₁₂O₄₀ (HPW) Keggin-type heteropoly acid was studied in a continuous flow fixed-bed reactor in the temperature range of 55-135 °C and isopropanol partial pressure of 0.94-5.52 kPa. The bulk catalysts included HPW itself and its Cs acid salts Cs_nH_3-nPW ₁₂O₄₀ (Cs_nH_3-nPW). The supported catalysts comprised of 15 wt.% HPW supported on SiO₂, TiO₂, ZrO₂ and Nb₂O₅. Under the conditions studied, the reaction was found to be of zero order in isopropanol. The turnover rate decreased in the order HPW > Cs_2.5H _0.5PW ≈ Cs₂HPW > HPW/SiO₂ > HPW/TiO₂ > HPW/Nb₂O₅ ≈ HPW/ZrO ₂, which is in line with the acid strength of these catalysts. The true activation energies E and pre-exponential factors A obtained from zero-order kinetics, were found to exhibit a compensation linear relationship ln A = mE + c. Moreover, the bulk and supported catalysts exhibited different compensation plots. This is suggested to be due to the different chemical structure of Bronsted acid sites in these catalysts, resulting in differing reaction mechanisms. The bulk catalysts possess strong surface H⁺ sites located on peripheral (bridging) oxygen atoms in the Keggin unit, whereas supported HPW catalysts have weaker H⁺ sites probably located on oxygen atoms of support. Consequently, on the bulk catalysts, isopropanol dehydration is suggested to occur via E1 elimination pathway, whereas on the supported catalysts it might also involve a contribution of E2 elimination pathway.

Full text of DOI:10.1016/j.jcat.2012.06.021

Journal of Catalysis 293 (2012) 158–164
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Compensation effect in isopropanol dehydration over heteropoly acid catalysts
at a gas–solid interface
Geoffrey C. Bond a, Sarah J. Frodsham b, Paul Jubb b, Elena F. Kozhevnikova b, Ivan V. Kozhevnikov b,
⇑
a 9 Townfield, Rickmansworth WD3 7DD, UK
b Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Kinetics of isopropanol dehydration at a gas–solid interface over a range of bulk and supported Brønsted
acid catalysts based on H3PW12O40 (HPW) Keggin-type heteropoly acid was studied in a continuous flow
fixed-bed reactor in the temperature range of 55–135 °C and isopropanol partial pressure of 0.94–
5.52 kPa. The bulk catalysts included HPW itself and its Cs acid salts CsnH3ꢀnPW12O40 (CsnH3ꢀnPW).
The supported catalysts comprised of 15 wt.% HPW supported on SiO2, TiO2, ZrO2 and Nb2O5. Under
the conditions studied, the reaction was found to be of zero order in isopropanol. The turnover rate
decreased in the order HPW > Cs2.5H0.5PW ꢁ Cs2HPW > HPW/SiO2 > HPW/TiO2 > HPW/Nb2O5 ꢁ HPW/
ZrO2, which is in line with the acid strength of these catalysts. The true activation energies E and pre-
exponential factors A obtained from zero-order kinetics, were found to exhibit a compensation linear
relationship ln A = mE + c. Moreover, the bulk and supported catalysts exhibited different compensation
plots. This is suggested to be due to the different chemical structure of Brønsted acid sites in these cat-
alysts, resulting in differing reaction mechanisms. The bulk catalysts possess strong surface H+ sites
located on peripheral (bridging) oxygen atoms in the Keggin unit, whereas supported HPW catalysts have
weaker H+ sites probably located on oxygen atoms of support. Consequently, on the bulk catalysts, iso-
propanol dehydration is suggested to occur via E1 elimination pathway, whereas on the supported cat-
alysts it might also involve a contribution of E2 elimination pathway.
Received 31 May 2012
Revised 21 June 2012
Accepted 22 June 2012
Available online 21 July 2012
Keywords:
Heteropoly acid
Heterogeneous acid catalysis
Isopropanol dehydration
Compensation effect
Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction
interface over a broad range of bulk and supported catalysts based
on the strongest heteropoly acid H3PW12O40 (HPW) and its acidic
Heterogeneous acid catalysis by heteropoly acids (HPAs) has at-
tracted much interest both in fundamental and applied research
because of its potential to generate significant economic and envi-
ronmental benefits [1–5]. The majority of catalytic applications use
the most stable and easy available Keggin HPAs comprising hetero-
Cs salts CsnH3ꢀnPW12O40 (CsnH3ꢀnPW) (Table 1). Previously, these
catalysts have been characterised by XRD, FTIR, 31P MAS NMR
and NH3 adsorption calorimetry [7].
2. Experimental
poly anions of the formula [XM12O40]
nꢀ, where X is the heteroatom
(PV, SiIV, etc.) and M is the addendum atom (MoVI, WVI, etc.). HPAs
possess strong Brønsted acidity superior to that of conventional
solid acid catalysts such as acidic oxides and zeolites. Amongst var-
ious heteropoly acids, tungsten HPAs are usually chosen as acid
catalysts because of their stronger acidity and higher thermal sta-
bility compared to molybdenum HPAs [1–5]. The acid and catalytic
properties of HPAs have been the subject of many studies ([1–6]
and references therein), but the structure of surface acid sites
and its relation to the activity of solid HPA catalysts are still
unclear.
2.1. Chemicals and catalysts
Chemicals were purchased from Aldrich and used as supplied
without further purification. Preparation of catalysts and supports
and catalyst characterisation techniques have been described else-
where [7]. Commercial Hombikat anatase titania (TiO2-H), Degussa
P25 titania (anatase/rutile = 3:1) (TiO2-D) and Aerosil 300 silica as
well as in-house made ZrO2 and Nb2O5 hydrous oxides [7] were
used as supports. TiO2 hydrous oxide (TiO2-I) was prepared by pre-
cipitating TiIV hydroxide from 0.5 M TiIV oxochloride aqueous solu-
tion (prepared from TiCl4) by adding dropwise 1 M ammonia
aqueous solution at 50 °C with stirring until pH 7.3 was reached,
followed by ageing at 50 °C for 1 h. The precipitate was filtered
off, washed with distilled water until Clꢀ free, oven dried at
100 °C overnight and air calcined at 400 °C for 2 h. Cs salts of
We now attempt to tackle this problem using a kinetic approach
through the investigation of isopropanol dehydration at a gas–solid
⇑
Corresponding author.
E-mail address: I.V.Kozhevnikov@liverpool.ac.uk (I.V. Kozhevnikov).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.06.021

G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
159
Table 1
Catalyst texture from nitrogen adsorption.a
b
Catalyst or support
SBET (m2 gꢀ1
)
Pore diameterc (Å)
Pore volumed (cm3 gꢀ1
)
H3PW12O40
Cs2.5H0.5PW12O40
Cs2HPW12O40
2.0
111
55
229
41
81
24
20
204
154
230
196
36
52
80
45
24
0.04
0.07
0.03
1.17
0.16
0.23
0.24
0.15
0.22
0.19
0.13
0.07
0.08
0.05
15%H3PW12O40/SiO2
15%H3PW12O40/TiO2-H
15%H3PW12O40/TiO2-H-500e
15%H3PW12O40/TiO2-D
15%H3PW12O40/TiO2-I
15%H3PW12O40/Nb2O5
15%H3PW12O40/Nb2O5-500e
15%H3PW12O40/Nb2O5-PCf
15%H3PW12O40/ZrO2
15%H3PW12O40/ZrO2-500e
15%H3PW12O40/ZrO2-PCf
40
49
165
166
96
118
121
54
56
31
66
a
Prior to N2 adsorption the catalysts were evacuated at 200–250 °C for 2.5 h.
BET surface area.
Average BET pore diameter.
Single point total pore volume.
The catalysts were calcined in air at 500 °C for 3 h.
Support was calcined at 300 °C before HPA impregnation.
b
c
d
e
f
HPW were prepared as described previously [7]. Zirconia- and nio-
bia-supported HPW were prepared by wet impregnation of ZrO2
and Nb2O5 hydrous oxides with an aqueous HPW solution [7]
and also by impregnation of supports pre-calcined at 300 °C/1 h
in air. The latter are labelled ZrO2-PC and Nb2O5-PC. Catalyst tex-
ture is given in Table 1. Phosphorus and tungsten elemental anal-
ysis (ICP) of HPA catalysts was in good agreement with the
preparation stoichiometry.
studies, e.g. [6,9,12]. The reaction is suggested to occur through E1
or E2 elimination pathways. The E2 pathway (bimolecular
elimination) involves concerted cleavage of C–O and C–H bonds
in alcohol using a pair of acid and base catalyst sites. The E1 path-
way (monomolecular elimination) breaks C–O bond in alcohol in
the rate-limiting step using an acid site to form water and carbe-
nium ion (possibly as a surface alkoxide); the latter undergoes pro-
ton abstraction to form alkene.
With catalysts possessing strong proton sites, such as HPA, iso-
propanol dehydration is likely to occur through E1 elimination,
with a plausible sequence of steps represented by Scheme 1 [6,7].
It includes equilibrated isopropanol adsorption on surface Brønsted
acid sites HB (where Bꢀ is the conjugate base) via hydrogen bonding
followed by isopropanol decomposition involving its protonation
followed by immediate C–O bond breaking through a carbenium-
ion transition state to form water and isopropyl carbocation; the
latter, by H+ abstraction, transforms into propene or reacts with an-
other isopropanol molecule to form diisopropyl ether (competing
SN1 pathway). Assuming that formation of carbenium ion is the
rate-limiting step and the product forming steps are irreversible,
with no re-adsorption of products taking place, the rate of isopropa-
nol conversion is given by the following equation:
2.2. Catalyst testing
Isopropanol dehydration was carried out under atmospheric
pressure in a Pyrex fixed-bed reactor (9 mm internal diameter) fit-
ted with on-line GC analysis as described previously [7]. Typical
reaction conditions: 0.1–0.2 g catalyst (45–180 lm particle size),
0.94–5.52 kPa i-PrOH partial pressure, 55–135 °C temperature
range ( 0.1 °C) and 20 ml minꢀ1 gas flow rate (N2 as a carrier
gas). Before reaction, the catalysts were pre-treated in situ at a
specified temperature. Reaction rates (in mol hꢀ1
g
) were calcu-
ꢀ1
cat
lated using the equation r = XF/W, where X is the isopropanol con-
version (typically X = 2–25%), is the molar flow rate of
F
isopropanol and W is the catalyst weight. For each catalyst, the rate
was measured at least at four different temperatures in the tem-
perature interval of 20–30 °C at a specified partial pressure of iso-
propanol (0.94 or 5.52 kPa), and at one temperature the effect of
isopropanol pressure on the reaction rate was examined. From
these results the reaction order in isopropanol was found to be
zero ( 0.1), hence isopropanol conversion was equivalent to the
reaction rate. Turnover rates (per accessible Brønsted site) were
calculated as explained in the text. From zero-order kinetics, the
values of true activation energy E ( 4 kJ molꢀ1) and pre-exponen-
tial factor lnA ( 1) (A in hꢀ1) were obtained using the Arrhenius
equation.
kE1K1½Hþꢂ½i-PrOHꢂ
r ¼
ð1Þ
1 þ K1½i-PrOHꢂ
Here K1 is the equilibrium constant for isopropanol adsorption
on Brønsted acid sites, kE1 is the rate constant of the E1 rate-limit-
ing step and [H+] is the density of accessible proton sites in catalyst
(in mol gꢀca1t). In agreement with the above mechanism, a good lin-
ear relationship between the logarithm of turnover rate (per
Brønsted acid site) and HPA acid strength has been found for iso-
propanol dehydration catalysed by HPA [7]. At high [i-PrOH], when
catalyst active sites are saturated with isopropanol, the reaction
becomes zero-order in isopropanol, i.e., r = kE1[H+]. In this case
the observed activation energy is the true value E. Gas-phase iso-
propanol dehydration over a wide range of oxide catalysts has been
found zero-order at isopropanol partial pressures p > 0.6 kPa and
120–300 °C [9].
3. Results and discussion
3.1. Mechanism of isopropanol dehydration
The dehydration of isopropanol has been widely employed as a
test reaction for probing acid–base properties of various catalysts
[8,9], including HPA catalysts [1,6,7,10,11]. The mechanism of
acid-catalysed dehydration of secondary alcohols, such as isopro-
panol and 2-butanol, in the gas phase has been addressed in many
3.2. Kinetic parameters
We measured isopropanol dehydration rates with HPA catalysts
at p = 0.94–5.52 kPa in the temperature range of 55–135 °C at

160
G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
Scheme 1. Proposed mechanism of isopropanol dehydration through E1 elimination pathway.
isopropanol conversions in the range of 2–25%. At these conditions,
the dehydration proceeded without catalyst deactivation, as repre-
sented by the reaction over 15%HPW/TiO2-H at 90 °C (Fig. 1). The
conversions were well below equilibrium limits, as estimated
using thermodynamic properties for the reaction components
available from the reference sources [13,14]. The reaction was
zero-order in isopropanol over 15%HPW/TiO2-H since the rate
practically did not change when p varied from 0.94 to 5.52 kPa
(Fig. 1). Also isopropanol conversion increased linearly with the
contact time (Fig. 2), as expected for a zero-order reaction. It
should be noted that for the reaction over silica-supported HPA
catalysts at lower i-PrOH pressures, p = 0.05–0.5 kPa, Macht et al.
[6] have reported rate inhibition with increasing p. No such effect
was observed under our reaction conditions.
30
25
20
15
10
5
5
4
3
2
1
0
Conversion
Propene/Ether
0
0
0.5
1
1.5
Propene was the major reaction product (67–97% selectivity),
with diisopropyl ether also formed. At a constant temperature,
the propene/ether molar ratio increased with isopropanol conver-
sion due to a decrease in isopropanol partial pressure (Fig. 2), in
agreement with the mechanism (Scheme 1). The propene/ether ra-
tio depended on the catalyst (Table 2), as may be expected for a
surface reaction. Amongst the catalysts studied, bulk HPW showed
the highest ether selectivity. This can be attributed to the strong
ability of bulk HPAs to absorb light alcohols [1,2], which would in-
Contact time W/F (h gcat-1mol-1)
Fig. 2. A plot of i-PrOH conversion and propene/ether molar ratio versus contact
time for i-PrOH dehydration over 15%HPW/TiO2-D (90 °C, p = 5.52 kPa).
and Cs2HPW (55 m2 gꢀ1, 0.063 mmol(H+) gꢀ1). The same approach
was used to obtain the number of surface protons in bulk HPW,
since in this case, too, the surface mechanism in more likely than
the bulk mechanism, as discussed elsewhere [7,16]. This gave for
HPW (2.0 m2/g) a proton density of 0.0069 mmol(H+) gꢀ1. The true
activation energies E and pre-exponential factors A were deter-
mined from the Arrhenius equation in the form lnkE1 = lnA ꢀ E/RT.
The activation parameters and the values of r and kE1 are given in Ta-
ble 3 together with the results obtained previously under similar
conditions [7]. Changing the isopropanol pressure 6-fold gives ki-
netic parameters that are within experimental error of each other
(entries 9 and 10 for Cs2HPW in Table 3). The E values obtained
are within the range of activation energies for isopropanol dehydra-
tion reported in the literature [9].
crease isopropanol concentration at the catalyst surface.
From the values of r (mol hꢀ1
g
) obtained, the rate constants
ꢀ1
cat
kE1 (hꢀ1) were calculated using the density of accessible proton sites
which was estimated as follows [7]. For supported HPW catalysts,
which all contained 15% HPW at sub-monolayer coverage, all HPA
protons were assumed to be equally available for reaction
(0.156 mmol gꢀ1 proton density for all supported catalysts). This
has been proved correct for silica-supported HPW by titration with
NH3 [15] and pyridine [6]. For bulk Cs salts of HPW, which have been
suggested to catalyse isopropanol dehydration through the surface-
type mechanism [1,2], the number of surface protons was calculated
using a Keggin unit cross section of 144 Å2 [1,2] and catalyst surface
The kE1 values obtained fall in the order: HPW > Cs2.5H0.5PW ꢁ
areas from Table 1: Cs2.5H0.5PW (111 m2gꢀ1, 0.064 mmol(H+) gꢀ1
)
Cs2HPW > HPW/SiO2 > HPW/TiO2 > HPW/Nb2O5 ꢁ HPW/ZrO2. This
A
100
90
80
70
60
50
40
30
20
10
0
B 90
80
70
Conversion
Propene
Ether
60
50
40
30
20
10
0
Conversion
Propene
Ether
0
20
40
60
80
100
0
20
40
60
80
Time (min)
Time (min)
Fig. 1. Time course for i-PrOH dehydration over 15%HPW/TiO2-H at 90 °C (0.1 g catalyst, pre-treatment at 300 °C/1 h, 20 ml/min flow rate): (A) p = 0.94 kPa,
r = 0.0024 mol hꢀ1 gꢀ1; (B) p = 5.52 kPa, r = 0.0028 mol hꢀ1 gꢀ1
.

G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
161
Table 2
Propene/ether molar ratio for different catalysts.a
Catalystb
15%HPW/TiO2-D (14)
15%HPW/ZrO2-PC (12)
Cs2.5H0.5PW (5)
HPW (3)
Propene/ether (mol/mol)
i-PrOH conversion (%)
4.2
12
7.3
25
3.7
23
2.0
15
a
At 90 °C and p = 5.52 kPa.
Numbers in brackets correspond to the entry in Table 3.
b
Table 3
Activation parameters for isopropanol dehydration.a
Catalystb
E (kJ molꢀ1
)
lnA (hꢀ1
)
lnkE1 kE1 (hꢀ1
)
lnrc r (mol hꢀ1 gꢀ1
)
c
Bulk HPW and CsnH3ꢀnPW catalysts
1
2
3
4
5
6
7
8
HPW (150 °C/1 h, 5.52 kPa, 60–90 °C)
HPW (200 °C/1 h, 5.52 kPa, 60–90 °C)
HPW (300 °C/1 h, 5.52 kPa, 60–90 °C)
Cs2.5H0.5PW (150 °C/1.5 h, 5.52 kPa, 60–90 °C)
Cs2.5H0.5PW (200 °C/1 h, 5.52 kPa, 60–90 °C)
HPW (300 °C/3 h, 5.52 kPa, 60–90 °C) [8]
Cs2.5H0.5PW (300 °C/1 h, 5.52 kPa)
Cs2HPW (200 °C/1 h, 5.52 kPa, 55–80 °C)
Cs2HPW (150 °C/1.5 h, 5.52 kPa, 55–75 °C)
Cs2HPW (150 °C/1.5 h, 0.94 kPa, 55–75 °C)
Cs2.5H0.5PW (300 °C/3 h, 5.52 kPa, 60–90 °C) [8]
147
107
104
97
95
90
82
73
73
70
55.6
43.1
39.4
36.3
35.6
35.7
31.8
28.7
28.4
27.7
27.0
8.4
8.5
6.0
5.0
5.0
6.7
5.2
5.0
5.0
5.1
5.1
ꢀ3.5
ꢀ3.4
ꢀ5.9
ꢀ4.7
ꢀ4.7
ꢀ5.0
ꢀ4.5
ꢀ4.7
ꢀ4.7
ꢀ4.6
ꢀ4.6
9
10
11
68
Supported HPW catalysts
12
13
14
15
16
17
18
19
20
21
22
23
24
25
15%HPW/ZrO2-PC (300 °C/1 h, 0.94 kPa, 55–85 °C)
100
97
90
92
90
86
83
81
76
72
71
63
59
43
36.8
34.8
32.4
32.2
32.4
31.5
27.3
28.4
25.6
24.1
24.5
20.7
20.3
13.6
4.6
3.7
3.3
2.7
3.4
3.8
0.6
2.3
1.0
0.9
1.6
0.4
1.3
ꢀ0.3
ꢀ4.2
ꢀ5.1
ꢀ5.5
ꢀ6.1
ꢀ5.3
ꢀ4.9
ꢀ8.2
ꢀ6.5
ꢀ7.8
ꢀ7.9
ꢀ7.1
ꢀ8.4
ꢀ7.4
ꢀ9.1
15%HPW/TiO2-H (300 °C/1 h, 0.94 kPa, 70–90 °C)
15%HPW/TiO2-D (300 °C/1 h, 0.94 kPa, 70–90 °C)
15%HPW/TiO2-I (300 °C/1 h, 0.94 kPa, 70–90 °C)
15%HPW/TiO2-H (300 °C/3 h, 5.52 kPa, 100–130 °C) [8]
15%HPW/SiO2 (300 °C/1 h, 5.52 kPa, 65–90 °C) [8]
15%HPW/Nb2O5 (300 °C/1 h, 0.94 kPa, 95–115 °C)
15%HPW/TiO2-H-500 (5.52 kPa, 100–130 °C) [8]
15%HPW/ZrO2 (300 °C/1 h, 0.94 kPa, 75–105 °C)
15%HPW/Nb2O5-500 (5.52 kPa, 105–135 °C) [8]
15%HPW/Nb2O5 (300 °C/3 h, 5.52 kPa, 100–130 °C) [8]
15%HPW/ZrO2-500 (5.52 kPa, 105–135 °C) [8]
15%HPW/ZrO2 (300 °C/3 h, 5.52 kPa, 100–130 °C) [8]
15%HPW/Nb2O5-PC (300 °C/1 h, 0.94 kPa, 105–135 °C)
a
The values of E ( 4 kJ molꢀ1) and lnA ( 1) were obtained using the Arrhenius equation from at least four rate measurements in the temperature range shown. For bulk
HPW, the error in A may be larger due to the low surface area of HPW (Table 1).
b
In brackets are catalyst pre-treatment and reaction conditions.
At 100 °C; calculated from the Arrhenius equation.
c
order is in line with the acid strength of these catalysts [7]. It
should be noted that pre-calcination of ZrIV hydrous oxide at
300 °C prior to HPW impregnation increased catalyst activity 40
times (cf. entries 12 and 20 in Table 2). The possible explanation
is that the fresh ZrIV hydrous oxide interacts strongly with HPW
reducing its acid strength, as demonstrated previously [7]. But
when its structure is hardened by pre-calcination, zirconia proba-
bly becomes less interactive with HPW. This, however, did not
work for niobia.
reactions that compensation is an automatic consequence of the
use of apparent Arrhenius parameters, where activation energies
are a function of surface coverage and reactant concentration
[17]. Recent discussion of the enthalpy–entropy relation can be
found in [19,20] and references cited therein.
A plot of lnA versus E for the supported HPA catalysts is shown
in Fig. 3. There is an excellent linear relation lnA = mE + c (correla-
tion coefficient R2 = 0.985), which implies that all the Arrhenius
plots should intersect at a single temperature (the isokinetic tem-
perature Ti = 298 K) derived from the slope m of the compensation
plot by m = (RTi)ꢀ1. This compensation plot is important because
with many such plots found in the literature of heterogeneous
catalysis there is at least an element of doubt as to whether the
Arrhenius parameters are true or apparent. In the present case,
the argument for their being true is very strong, so that explana-
tions relevant to compensation based on apparent Arrhenius
parameters are not applicable, and alternatives need to be consid-
ered. Before going into a discussion on this, let us introduce the
unsupported HPA catalysts.
3.3. Activation parameters for supported HPA catalysts
Activation energies E for supported HPA catalysts vary from 43
to 100 kJ molꢀ1 and rates r by a factor of 150 (Table 3). From this it
follows that the changes in E must be at least partly compensated
by changes in the pre-exponential factor lnA. Such ‘‘compensation’’
has been known to occur with heterogeneously-catalysed reac-
tions for almost 100 years [17,18]; it is not unique to catalysis,
and similar phenomena have been observed in a variety of differ-
ent fields of science ([17,18] and references therein). The existence
of linear free-energy relations between enthalpies and entropies of
the reaction is found throughout organic chemical reactions, and
the lnA term in catalytic processes will embrace the entropy
change. It has been demonstrated for a number of catalytic
3.4. Activation parameters for unsupported HPA catalysts
Activation energies again vary widely (68–146 kJ molꢀ1),
although the rates only change by about 4.5 times (Table 3); in

162
G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
60
50
40
30
20
10
y = 0.3692x + 1.6192
Bulk catalysts
R2 = 0.9901
1
y = 0.4034x - 4.3021
R2 = 0.9848
Supported catalysts
2
3
4
5
12
6
13
14
15
16
7
8
9
17
18
19
20
10
11
22
21
23
24
25
40
60
80
100
E (kJ mol-1)
120
140
Fig. 3. Compensation plots for i-PrOH dehydration catalysed by bulk and supported
HPA catalysts (data numbering as in Table 3). The isokinetic temperature Ti is 326 K
for bulk and 298 K for supported catalysts.
particular all the Cs-containing catalysts have similar activities
(lnr = ꢀ4.6 0.1 and ln kE1 = 5.1 0.1); nevertheless the activation
energies range between 68 and 97 kJ molꢀ1. Therefore compensa-
tion between E and lnA is again implied, as shown by Fig. 3; but
most of the points lie consistently above the line for the supported
HPA catalysts due to their generally higher activities, and on a line
of lower slope that gives Ti = 326 K. The use of the compensation
plot therefore provides a graphic demonstration of the different
characters of the two sets of catalysts, and a ready means of sepa-
rating the two.
3.5. The reason for different compensation plots for bulk and supported
HPA catalysts
In the case of supported catalysts it seems likely that the effec-
tive protons are located on the support rather than on the Keggin
polyanions. This may be the reason for the difference in compensa-
tion behaviour of bulk and supported HPA catalysts. In bulk HPW
and its Cs salts the surface protons are located on the bridging out-
er oxygen atoms, O2 and/or O3, in the Keggin anion (Fig. 4) ([2,21]
and references therein). Since protons generally tend to reside on
the most basic sites available, in the case of supported HPW cata-
lysts, the HPW protons are likely to be located on the neighbouring
oxygen atoms of support, which are more basic than the outer oxy-
gen atoms in the Keggin anion. It has been demonstrated that
Fig. 5. 1H MAS NMR spectra of H3PW12O40 supported on SiO2 (surface area
390 m2 gꢀ1): (1) bulk H3PW12O40, (2) 50% H3PW12O40/SiO2, (3) 37% H3PW12O40
/
SiO2, (4) SiO2 (samples evacuated at 200 °C at 10ꢀ3 Pa for 3 h, 3 kHz spinning rate,
chemical shifts relative to external TMS, sidebands asterisked) [22].
interaction of HPW with oxide support reduces the strength of pro-
ton sites in such catalysts [7]. The different location of protons in
bulk and silica-supported HPW is evidenced by 1H MAS NMR
(Fig. 5) [22]. Bulk HPW exhibits a signal of strong proton sites at
9.3 ppm. In the spectrum of HPW/SiO2 only the signal of silanol
groups at 1.8 ppm can be seen up to 20 wt.% HPW loading. The sig-
nal of strong protons at 9.3 ppm appears only above 50% HPW
loading. This clearly shows that proton sites in supported HPW
have a different chemical structure from the bulk HPW. In HPW/
SiO2, the proton sites may be located on silanol (A) and siloxane
(B) groups of silica rather than on the outer oxygen atoms in the
3ꢀ
PW
O
polyanion [23]:
12
40
3ꢀ
Fig. 4. The Keggin structure of [PW12O40
]
polyanion with three types of outer
oxygen atoms: terminal (O1), edge-bridging (O2) and corner-bridging (O3).

G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
163
Scheme 2. Proposed mechanism of isopropanol dehydration involving E1 and E2 pathways.
energy levels for the reactants to another for the products, but these
will not be equally spaced. So the reaction occurs most readily when
there are matching levels in each set (Fig. 6) [17]. The spacing be-
tween levels decreases with increasing energy, and therefore in-
creases the chance of finding a matching or closely similar level in
the product set. Thus the higher the activation energy, the greater
the probability of reaction occurring. It has been shown ([17,19]
and references therein) that the necessity of accumulating a num-
ber of excitations in order to provide a large
D
H– must also lead
to a contribution to
D
S–. This entropy–enthalpy relation leads to ki-
netic compensation. This is a simplified statement of one of the
ways in which a correlation between E and lnA can be understood.
4. Conclusions
Fig. 6. Reaction as a change from one set of vibrational levels for the reactants (left)
to another for the products (right) [17].
Isopropanol dehydration at a gas–solid interface over a wide
range of bulk and supported acid catalysts based on H3PW12O40
heteropoly acid was found to exhibit a compensation relationship
between the Arrhenius activation parameters, true activation en-
ergy and pre-exponential factor, obtained from zero-order kinetics.
This compensation relationship is important because with many
such relationships found in the literature of heterogeneous cataly-
sis there is an element of doubt as to whether the Arrhenius
parameters are true or apparent. Moreover, the bulk and supported
HPA catalysts exhibited different compensation plots. This may be
explained by different chemical structure of Brønsted acid sites in
these catalysts, resulting in differing reaction mechanisms. On the
bulk catalysts, isopropanol dehydration is suggested to occur via
E1 elimination pathway, whereas on the supported catalysts it
might also involve a contribution of E2 elimination pathway.
Therefore, supported HPW catalysts, as compared to HPW-
based bulk catalysts, contain weaker proton sites as well as possess
basic sites located on support. It is conceivable that in the case of
supported HPW catalysts E1 elimination could be assisted to some
extent by E2 elimination involving proton abstraction from C–H
bond in isopropanol aided by the conjugate base Bꢀ or another
neighbouring basic site, as represented by step kE2 in Scheme 2.
A similar mechanism has been proposed for 2-butanol dehydration
over silica-supported HPAs [6].
The different compensation behaviour for bulk and supported
HPA catalysts may therefore be attributed to the different chemical
structures of surface proton sites in these catalysts, which may
have mechanistic implications resulting in a contribution of E2
pathways to the E1 elimination reaction of isopropanol in the case
of supported HPW catalysts.
References
[1] T. Okuhara, N. Mizuno, M. Misono, Adv. Catal. 41 (1996) 113–252.
[2] I.V. Kozhevnikov, Catalysis by Polyoxometalates, Wiley & Sons, Chippenham,
2002.
3.6. The origin of compensation
[3] M. Misono, Catal. Today 100 (2005) 95–100.
[4] I.V. Kozhevnikov, J. Mol. Catal. A 305 (2009) 104–111.
[5] M.N. Timofeeva, Appl. Catal. A 256 (2003) 19–35.
[6] J. Macht, M. Janik, M. Neurock, E. Iglesia, J. Am. Chem. Soc. 130 (2008) 10369–
10379.
[7] A.M. Alsalme, P.V. Wiper, Y.Z. Khimyak, E.F. Kozhevnikova, I.V. Kozhevnikov, J.
Catal. 276 (2010) 181–189.
Compensation between true Arrhenius parameters may find its
origin in a relation between the enthalpies and entropies of activa-
tion [15 and references therein]. The fundamental equation of the
transition state theory gives the rate constant as
[8] K. Tanabe, M. Misono, Y. Ono, H. Hattori (Eds.), New Solid Acids and Bases:
Their Catalytic Properties, Kodansha, Tokyo, 1989.
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
kBT
h
D
R
S–
D
H–
RT
k ¼
exp
exp
ꢀ
ð2Þ
[9] A. Gervasini, J. Fenyvesi, A. Auroux, Catal. Lett. 43 (1997) 219–228.
[10] L.R. Pizzio, C.V. Carceres, M.N. Blanco, Appl. Catal. A 167 (1998) 283–294.
[11] E. Lopez-Salinas, J.G. Hernandez-Cortez, I. Schifter, E. Torrez-Garcia, J.
Navarrete, A. Gutierrez-Carrillo, T. Lopez, P.P. Lottici, D. Bersani, Appl. Catal.
A 193 (2000) 215–225.
[12] S. Delsarte, P. Grange, Appl. Catal. A 259 (2004) 269–279.
[13] CRC Handbook of Chemistry and Physics, 89th ed., 2008–2009.
[14] K.S. Pitzer, J. Chem. Phys. 5 (1937) 433–479.
where kB is the Boltzmann’s constant, h is the Planck’s constant, T is
the absolute temperature,
S– is the entropy of activation, H– is
the enthalpy of activation and R is the gas constant [17,18,24]. Thus
S– and H– must be supplied for the reaction to reach the transi-
D
D
D
D
[15] E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 224 (2004) 164–169.
[16] J.B. Moffat, Metal-Oxygen Clusters: The Surface and Catalytic Properties of
Heteropoly Oxometalates, Kluwer, NY, 2001.
[17] G.C. Bond, M.A. Keane, H. Kral, J.A. Lercher, Chem. Rev. – Sci. Eng. 42 (2000)
323–384.
tion state. For a catalysed reaction, the pre-exponential factor A
equates to the first two terms on the right-hand side, multiplied
by the concentration of active sites. Very simply and qualitatively,
the reaction may be viewed as a change from the set of vibrational

164
G.C. Bond et al. / Journal of Catalysis 293 (2012) 158–164
[18] G.C. Bond, Metal-Catalysed Reactions of Hydrocarbons, Springer, NY, 2005.
[19] A. Yelon, E. Sacher, W. Linert, Phys. Chem. Chem. Phys. 14 (2012) 8232–8234.
[20] P.J. Barrie, Phys. Chem. Chem. Phys. 14 (2012) 318–336.
[21] B.B. Bardin, S.V. Bordawekar, M. Neurock, R.J. Davis, J. Phys. Chem. B 102
(1998) 10817–10825.
[22] V.M. Mastikhin, S.M. Kulikov, A.V. Nosov, I.V. Kozhevnikov, I.L. Mudrakovsky,
M.N. Timofeeva, J. Mol. Catal. 60 (1990) 65–70.
[23] I.V. Kozhevnikov, K.R. Kloetstra, A. Sinnema, H.W. Zandbergen, H. Van Bekkum,
J. Mol. Catal. A 114 (1996) 287–298.
[24] K.J. Laidler, Chemical Kinetics, third ed., Harper & Row, New York, 1987.