Shape-selective diisopropylation of naphthalene in H-Mordenite: Myth or reality?

Article abstract of DOI:10.1016/j.jcat.2009.12.005

Selective diisopropylation of naphthalene to 2,6-diisopropylnaphthalene is a challenging goal in sustainable catalysis. Ultrastable Y and H-Mordenite zeolites are the best catalysts reported in the literature with respect to 2,6-diisopropylnaphthalene selectivity. It is generally accepted that in the case of H-Mordenite, shape-selectivity is responsible for the observed 2,6-diisopropylnaphthalene selectivity, while on Ultrastable Y-zeolite, the observed selectivity reflects the internal thermodynamic equilibrium of positional isomers. Revisiting both the experimental and the computational work in this field now leads to the conclusion that shape-selectivity of whatever kind can be ruled out in the case of H-Mordenite. H-Mordenite catalysts produce usually a kinetically controlled mixture of diisopropylnaphthalene isomers which can shift to the direction of a thermodynamical distribution at high reaction temperatures or over more active catalysts.

Full text of DOI:10.1016/j.jcat.2009.12.005

Journal of Catalysis 270 (2010) 60–66
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Shape-selective diisopropylation of naphthalene in H-Mordenite: Myth or reality?
a
a,*
a
a
b
Christophe Bouvier , Wim Buijs , Jorge Gascon , Freek Kapteijn , Bogdan C. Gagea ,
b
b
Pierre A. Jacobs , Johan A. Martens
a
Catalysis Engineering, Delft Chem Tech, TU Delft, Julianalaan 136, 2628 BL Delft, The Netherlands
Center for Surface Chemistry and Catalysis, Catholic University of Leuven, Kasteelpark Arenberg 23, Leuven B-3001, Belgium
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 July 2009
Revised 2 December 2009
Accepted 5 December 2009
Available online 12 January 2010
Selective diisopropylation of naphthalene to 2,6-diisopropylnaphthalene is a challenging goal in sustain-
able catalysis. Ultrastable Y and H-Mordenite zeolites are the best catalysts reported in the literature with
respect to 2,6-diisopropylnaphthalene selectivity. It is generally accepted that in the case of H-Mordenite,
shape-selectivity is responsible for the observed 2,6-diisopropylnaphthalene selectivity, while on Ultrast-
able Y-zeolite, the observed selectivity reflects the internal thermodynamic equilibrium of positional iso-
mers. Revisiting both the experimental and the computational work in this field now leads to the
conclusion that shape-selectivity of whatever kind can be ruled out in the case of H-Mordenite. H-Mord-
enite catalysts produce usually a kinetically controlled mixture of diisopropylnaphthalene isomers which
can shift to the direction of a thermodynamical distribution at high reaction temperatures or over more
active catalysts.
Keywords:
H-USY
H-MOR
Shape selectivity
2
,6-Diisopropylnaphthalene
Thermodynamical distribution
Kinetic mixture
Ó 2009 Elsevier Inc. All rights reserved.
Molecular modeling
1
. Introduction
Selective diisopropylation of naphthalene to 2,6-diisopropyl-
seems to be the determining factor. This prompted us to reinvesti-
gate the two cases, both experimentally and computationally.
naphthalene (2,6-DIPN) is an important step towards an attractive
and sustainable route to 2,6-naphthalene dicarboxylic acid. H-
Mordenite and H-USY [1–18] are the best with respect to selectiv-
ity, though they largely differ in activity.
2
. Experimental part
2.1. Materials
Naphthalene dialkylation is a complex reaction. The dialkyl-
naphthalenes are prone to many types of reactions such as isomer-
ization, transalkylation and dealkylation [2,19–23].
Shape-selectivity, based on the fit of 2,6-DIPN inside the chan-
nel of H-Mordenite, has been widely accepted as the explanation
for the high yield to 2,6-DIPN [23]. So far, many research teams
have focused on H-Mordenite, its dealuminated versions
Naphthalene 98% was purchased from Aldrich, isopropanol (ACS
reagent), cyclohexane (99+%) and 2,6-DIPN (99%) were supplied by
Acros. All reagents were used without further purification. Com-
mercially available zeolites used were H-USY (CBV760) Si/Al = 30
from PQ and H-MOR Si/Al = 5.9 from Zeocat. H-MOR Si/Al = 24
was synthesized according to the procedure described by Lu
et al. [38].
[
1,3,10,24–27], specimens with surface modifications [28–30], as
well as on theoretical studies on the isomer formation [19,23,31–
6]. At the same time, studies using H-USY [2,4,5,8,13,37] showed
H-MOR 5.9: Crystal size: 0.8–1.0
7.36; Al : 12.59; Na O: 0.05; Overall atomic Si/Al (chemical
composition): 5.9, Framework atomic Si/Al (27Al MAS NMR): 6.9,
2
lm, composition (wt.%): SiO :
8
2
O
3
2
3
even higher yields towards DIPNs in general and 2,6-DIPN
particularly.
We were intrigued by the fact that in the case of H-Mordenite,
shape-selectivity is used as the dominant explanation for the ob-
served yield, and selectivity while in the case of H-USY thermody-
namic control, based on the relative stability of the DIPN isomers,
2
3
Crystallinity% (XRD): 97 BET, 500 m /g: Bulk density, g/cm : 0.63
Brønsted acid site concentration: mmol H+/g of dry zeolite: 1.8.
The acid site concentration was determined using Temperature-
Programmed Desorption of ammonia (NH
carried out on a Micromeritics TPR/TPD 2900 apparatus equipped
with thermal conductivity detector (TCD). Approximately,
5 mg of each sample was pretreated at 125 °C and rapidly cooled
to 100 °C followed by loading with ammonia applying a flow of
0 mL/min for about 1 h. A He flow of 30 mL/min was applied to
remove weakly adsorbed NH . A linear temperature program was
3 3
-TPD). NH -TPD was
a
2
3
*
Corresponding author. Fax: +31 (0) 15 27 85006.
E-mail address: w.buijs@tudelft.nl (W. Buijs).
3
0
021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.12.005

C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
61
started (100–600 °C at 10 °C/min), and the desorbed amount of
ammonia was analyzed by the TCD. The TPD spectra were used
to analyze the amount of acidic hydroxyl groups present in the cat-
alyst samples.
on a 25 m  0.25 mm VF-5MS column. Temperature was set at
50 °C for 1 min and then raised till 300 °C at 10 °C/min. Mass spec-
tra were recorded with a VG SE spectrometer at 70 eV. More details
about the analytical procedures can be found elsewhere [39].
The following yields and selectivities were calculated based on
naphthalene: selectivity towards Naphthalene (N; in isomerization
experiments only), IsoPropylNaphthalenes (IPN), DiIsoPropylNaph-
thalenes (DIPN), PolyIsoPropylNaphthalenes (PIPN), polyalkylated
unsaturated compounds (Unsaturates) and Others (not identified
naphthalenic compounds). In addition, the isomer distributions
in the IPN and DIPN fractions were also determined.
H-MOR 24: Crystal size: 1.0–1.5
2
lm, composition (wt%): SiO :
8
7.36; Al : 12.59; Na O: 0.05; Overall atomic Si/Al (chemical
2
O
3
2
composition): 24, Framework atomic Si/Al (27Al MAS NMR): 20,
2
Crystallinity% (XRD): 98 BET, 546 m /g Brønsted acid site concen-
tration: mmol H+/g of dry zeolite: 1.1. A photograph of a typical
crystal is shown below in Fig. 1.
2.2. Reaction procedure
2.4. Molecular modeling
Unless mentioned otherwise, all reactions were performed in a
5
0 mL stainless steel Parr autoclave using the following standard
procedure. For alkylation reactions, 320 mg of naphthalene
2.5 mmol), 500 mg of catalyst and 0.38 mL (5.0 mmol) of isopro-
All quantum chemical calculations were carried out using the
Spartan ’06 molecular modeling suite of programs [40]. First, the
most stable DIPN conformers were determined using molecular
mechanics (MMFF), next the resulting structures were fully opti-
(
panol were stirred for 24 h at 200 °C in 25 mL of cyclohexane. At
the beginning of a catalytic run, samples were taken hourly. For
isomerization reactions, 530 mg of 2,6-DIPN (2.5 mmol) and
*
mized at the B3LYP/6-31G level, using semi-empirical PM3 data
as input. A frequency calculation was performed to allow thermo-
dynamic corrections. Finally, single point calculations on the
5
00 mg of catalyst in 25 mL of cyclohexane were introduced into
the reactor vessel and heated at 200 °C for 24 h. Pressure was built
up autogenously. Catalysts were calcined prior to use under an air
atmosphere according to the following procedure: 3 °C/min from
room temperature to 550 °C and kept at that temperature for 5 h
before cooling to room temperature. High amounts of catalyst
were used to reach a significant conversion, even with poorly ac-
tive catalysts. The catalyst was removed from the reaction mixture
by filtration over cotton, and the reaction samples were analyzed
directly. Generally, mass balances were about 95%. At room tem-
perature, none of the catalysts showed significant preferential
absorption of naphthalenic compounds; however, isopropanol
was adsorbed up to 25% in H-USY.
B3LYP/6-311++G(2df,2p) level were performed on the B3LYP/6-
*
3
1G optimized structures. Transition States were fully optimized
and checked by animation of their unique imaginary vibration.
3
. Results
The consecutive reaction steps in the isopropylation of naph-
thalene are presented in Fig. 2. Naphthalene with up to 4 isopropyl
substituents could be detected in the reaction mixture depending
on the reaction conditions. Some polyalkylated unsaturated prod-
ucts (Unsaturates) with mass 252 described by Moreau et al.
[
5,41] were observed as well.
In close analogy to isobutene [42], isopropanol can be activated
2.3. Product analysis
by an acidic zeolite to either the isopropyl cation or the isopropyl
alkoxide after the elimination of a water molecule. Reaction with
naphthalene with the release of a proton leads to an isopropyl
naphthalene. Dealkylation can be considered as the reversed reac-
tion. The isopropyl cation can release a proton, leading to the for-
mation of propene. The reaction of propene with an isopropyl
GC Â GC analyses were performed using a VF-1 capillary col-
umn with dimensions of 50 m  0.25 mm and film thickness
0
.40
lm, combined with a VF-17 capillary column with dimensions
1
.5 m  0.1 mm and film thickness 0.20
l
m. Helium was used as
carrier gas and the flow was kept constant at 1 mL/min. The split
moiety is the first step of C
-oligomers can occur as well. Reaction products with methyl or
ethyl substituents on the naphthalene were detected as well as
propene and C -alkenes. Evidence for the formation of these light
3
-oligomerization. Cracking of these
ratio was 1:100. The injection of 1 l of the sample was performed
l
C
3
at 280 °C. The temperature program was as follows: 1 min at 70 °C
and a ramp at 5 °C/min till 300 °C. The second dimension time was
s and the hot pulse time 0.2 s. GC–MS analyses were performed
5
2
compounds was obtained by leading nitrogen through the reaction
mixture. The resulting gas flow was contacted with a solution of
bromine in dichloromethane, and the corresponding dibrominated
species were identified with GC–MS. Levels were ꢀ0.1% on a molar
3
basis. However, ‘‘unsaturates”, originating from a C -dimer and
naphthalene, were present at a detectable level in all experiments
and can be considered as an indication for C
coke formation.
3
-loss and possibly
3.1. Isopropylation of naphthalene and isomerization of 2,6-DIPN over
H-USY
Under standard conditions at 200 °C, H-USY showed an excel-
lent 87% conversion already after 4 h, increasing to 91% after
2
4 h, combined with a yield to DIPNs of 38.4%, and the intermedi-
ate IPNs of 37.5%. PIPNs, Unsaturates and Others account for 15%
Fig. 3 and Table 1). Up to 4 h, the (molar) sum of the products
(
and naphthalene was very close to 100%. Later on, the molar sum
of identified naphthalene-based reaction products decreased
slightly due to the formation of some heavy side products. With
respect to the isomeric distribution, in the IPNs, 2-IPN accounted
Fig. 1. A typical crystal of H-Mordenite Si/Al = 24.

6
2
C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
8
5
1
4
2
(n-2) isopropanol
isopropanol
7
isopropanol
n
3
6
IPN
Fig. 2. Consecutive reaction steps in catalytic naphthalene isopropylation.
Naphthalene
DIPN
PIPN
Fig. 3. Yields (%) of various products in the isopropylation of naphthalene over H-USY against reaction time.
for 93% and 1-IPN to 7%. Actually, this resembles the thermody-
namic distribution as obtained by computation. In the DIPNs,
for 72%. In the DIPNs, the
DIPN (15.8%) were dominant, followed by
a
,
a
-products 1,4-DIPN (33%), and 1,5-
,b-products 1,3-DIPN
a
2
,6-DIPN and 2,7-DIPN accounted for 36% and 42%, respectively.
Table 1 also lists the result of an experiment at 100 °C, sampled
(11.6%), 1,6-DIPN (12.2%) and 1,7-DIPN (18.9%). b,b-products 2,3-
DIPN (3.6%), 2,6-DIPN (3.0%) and 2,7-DIPN (2.4%) were formed only
in minor amounts.
after 4 h. The conversion was still 27%, but the product pattern and
isomeric distribution of IPNs and DIPNs were completely shifted.
The yield to IPNs was 22%, and the yield to DIPNs was 4.3% only.
PIPNs, Unsaturates and Others could only be hardly detected. With
respect to the isomeric distribution, in the IPNs, 1-IPN accounted
Under standard conditions at 200 °C in isomerization, the con-
version of 2,6-DIPN reached a maximum of ca. 85% after 4 h and
decreased to ca. 81% after 24 h (Table 1 and Fig. 4). The main prod-
ucts were 2-IPN and (non-2,6-) DIPN isomers with yields of 38.9%
and 18.2%, respectively, after 24 h. Within isomeric DIPNs, the for-
mation of 2,7-DIPN with a yield of 10.6% was dominant. Product
formation in time revealed that naphthalene 2-IPN and 2,7-DIPN
were intermediate products in the reaction network. The naphtha-
lene yield showed a minimum of 7.0% after 2 h. 2,6-DIPN showed a
weak minimum after 4 h and 2,7-DIPN a distinct maximum of
Table 1
Composition of the reaction mixture of the isopropylation of naphthalene, and
isomerization of 2,6-DIPN over H-USY after 24 h. Y: yield; ID: isomer distribution.
Catalyst: H-USY
T (°C)
Isopropylation
100 (4 h)
Isopropylation
200
Isomerization
200
2
1.7% after 3 h. It is noteworthy to mention that the highest naph-
Compound
Y (%)
ID (%)
Y (%)
ID (%)
Y (%)
ID (%)
thalene yield (13.0%) was obtained after just 1 h. Thus, dealkylation
seems to be the fastest process in the beginning of the reaction.
Furthermore, Fig. 4 strongly suggests, by the reciprocal decrease
in 2,6-DIPN and increase in 2,7-DIPN, that a transalkylation process
contributes to the shift of 2,6- to 2,7-DIPN. In a study to be pub-
lished separately, more attention will be paid to these kinetic phe-
nomena and the influence of temperature, combined with the
results of molecular modeling.
Naphthalene
73.2
22.2
9.0
37.5
9.1
38.9
1
2
1
1
1
1
1
1
1
2
2
2
-IPN
-IPN
72.1
22.9
0.0
11.6
32.6
15.8
12.2
18.9
0.0
6.6
93.4
0.0
5.5
0.7
2.0
5.2
4.9
0.0
5.1
94.9
0.0
5.7
0.8
2.0
3.2
4.8
0.0
,2-DIPN
,3-DIPN
,4-DIPN
,5-DIPN
,6-DIPN
,7-DIPN
,8-DIPN
,3-DIPN
,6-DIPN
,7-DIPN
4.3
38.4
37.3
In isomerization, the selectivity to
a-substituted products was
low. At the end the product distribution showed a striking resem-
blance with the one obtained in naphthalene isopropylation. Final-
ly, as H-USY remained active over the whole period in both
experiments, the level of Unsaturates remained stable after 4 h at
3.6
3.0
2.4
3.5
36.3
42.2
0.1
51.2
28.3
PIPNs
Unsaturates
Others
0.1
0.1
0.1
5.0
4.3
5.8
4.4
4.2
6.1
ꢀ
4.3%, and the overall mass balance was ꢀ95%, coke formation
does not seem to be an important factor.

C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
63
3
.2. Isopropylation of naphthalene and isomerization of 2,6-DIPN over
Within the DIPN isomers, the relative yields of 2,6- and 2,7-DIPN
were 15% and 10% only. 1,3- 1,4-, 1,5-, 1,6- and 1,7-DIPN account
for ꢀ14%, 18%, 9%, 15% and 17%, respectively. From Fig. 5, it is clear
that the formation of 1-IPN and 2-IPN, followed by the DIPNs and
some side products, remain almost linear in time.
As H-MOR with a Si/Al = 5.9 despite its relative high activity
might not be the best candidate to show shape-selectivity, an H-
MOR with Si/Al = 24 was used as well. At 200 °C, the decreased
acidity and the slightly larger crystal size led to a conversion of
ꢀ2% only, and therefore the temperature was raised to 250 °C. At
250 °C, a 4% conversion was reached. The ratio of 2-IPN/1-IPN
dropped to 0.5 at 200 °C and 0.6 at 250 °C, compared to 0.9 for
H-MOR with Si/Al = 5.9. No major changes were observed in the
DIPN isomer distribution. No coke formation was observed, and
the mass balance was very close to 100%.
H-Mordenite
In naphthalene isopropylation over H-Mordenite Si/Al = 5.9, the
naphthalene conversion reached ꢀ20% after 20 h in a rather linear
way (Fig. 5, Table 2). The main reaction products were 1- and 2-IPN
with a total IPN yield of ꢀ17%. The 1-IPN and 2-IPN isomers were
present in a ratio close to one. This result can be considered as an
indication for a more kinetic product distribution instead of the
thermodynamic one obtained with H-USY. The selectivity towards
DIPNs was only 1.1%, while PIPNs, Unsaturates and Others ac-
3
counted only for 0.4%, 1.9% and 0.2%, respectively. Thus, C -oligo-
merization is relatively more prominent compared to DIPN
formation as in the case of H-USY, though its absolute level is >2
times lower at 200 °C. However, from the mass balance and the lin-
ear growth in time of all products, it might be concluded that coke
formation, if significantly present at all, does not seem to play a
role in the selectivity pattern obtained under these conditions.
In the isomerization of 2,6-DIPN over H-Mordenite Si/Al = 5.9, a
similar ꢀ20% conversion was obtained after 20 h as in the case of
diisopropylation. The main reaction products were 2,7-DIPN, 2-
Fig. 4. Yields (%) of products in the isomerization of 2,6-DIPN over H-USY against time.
Fig. 5. Yields (%) of various products in the isopropylation of naphthalene over H-Mordenite (Si/Al = 5.9) against time.

6
4
C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
Table 2
3.3. Computational results
Composition of the reaction product mixture in the isopropylation of naphthalene
over H-Mordenite Si/Al = 5.9 and Si/Al = 24, and the isomerization of 2,6-DIPN over H-
MOR 5.9. Y: yield; ID: isomer distribution.
As mentioned already in the introduction, it is generally ac-
cepted that H-USY yields a thermodynamic mixture of DIPN iso-
mers. Table 3 summarizes experimental results with H-USY, H-
MOR and computational results using different methods, obtained
by other authors and by us.
Catalyst
H-MOR 5.9
H-MOR 5.9
H-MOR 24
Isopropylation Isomerization Isopropylation
T (°C)
200
200
200
250
Experimental values for the DIPN isomer distribution over H-
USY are quite similar. The largest difference in the experimental
values is ꢀ2.5% for 2,6-DIPN. This is mainly due to our new analyt-
ical technique (GC Â GC) which allows better peak separation and
avoids baseline integration errors which might occur in the older
technique [23]. Baseline integration errors lead to a slight overes-
timation of the large peaks (2,6- and 2,7-DIPN) relative to the smal-
ler peaks of the other DIPN isomers. The computational results
obtained with all methods are in quite fair agreement with the
experimental data, especially if one takes into account that MP2 re-
sults are obtained at 25 °C, B3LYP results at 200 °C and B3PW91 re-
sults at 250 °C.
At higher temperature, relative amounts of 2,6- and 2,7-DIPN
decrease slightly in favor of other isomers. Both B3LYP and
B3PW91 assign 2,7-DIPN as the most abundant isomer in accor-
dance with the experimental data; however, energy differences
are so small that no absolute conclusion should be drawn from
these findings.
Compound
Y (%)
ID (%)
Y (%) ID (%) Y (%)
ID
Y (%)
ID
(%)
(
%)
Naphthalene 79.2
2.0
5.2
98.4
1.3
96.1
3.5
1
2
1
1
1
1
1
1
1
2
2
2
-IPN
-IPN
17.2
47.3
52.7
0.0
13.6
18.0
8.8
15.5
16.8
0.0
5.1
94.9
0.0
67.3
32.7
0.0
14.3
19.7
6.5
19.1
17.5
0.0
62.7
37.3
0.0
14.7
11.8
5.0
17.1
18.3
0.0
,2-DIPN
,3-DIPN
,4-DIPN
,5-DIPN
,6-DIPN
,7-DIPN
,8-DIPN
,3-DIPN
,6-DIPN
,7-DIPN
1.1
91.5
ꢀ0.05
0.10
0.3
0.0
0.0
0.2
0.4
0.0
0.1
89.0
10.2
2.1
15.1
10.1
1.6
12.8
8.5
1.3
19.6
12.2
PIPNs
Unsaturates
Others
0.4
1.9
0.2
1.3
0.0
0.1
<0.01
0.25
<0.01
<0.01
0.30
<0.01
IPN, PIPNs and naphthalene with yields of 9.3%, 4.9%, 1.3%, and
.0%, respectively. Other DIPN isomers besides 2,7-DIPN did appear
The relative amounts of 1-IPN and 2-IPN are also in agreement
with the experimental results obtained with H-USY. Experimen-
tally, distributions of 6.6% and 93.4%, respectively, were found,
while 4.5% and 95.5% were calculated.
2
but in very low amounts. The 2-IPN/1-IPN ratio was close to 20.
Fig. 6 shows a regular build-up in time of most products, except
for naphthalene which gave a maximum yield of 5% after 2 h and
then gradually decreased to 2%, and 2,6-DIPN which showed a local
maximum after 3 h of 95% and then decreased to 81%. This behav-
ior is more or less similar to what was observed with H-USY. Deal-
kylation seems to be the fastest process in the beginning of the
reaction. Transalkylation of 2,6- to 2,7-DIPN seems to play a role
also. Whereas in the isopropylation with H-Mordenite Si/Al = 5.9,
a significant amount of Unsaturates were formed, in the isomeriza-
tion of 2,6-DIPN, Unsaturates were not observed at all, but PIPNs
were, albeit mainly triisopropylnaphthalenes. Again coke forma-
tion was very unlikely.
Furthermore, it should be noted that single point calculations
on the B3LYP/6-311++G(2df,2p) level performed on the B3LYP/6-
*
31G optimized structures did not lead to significantly different re-
sults, and these results are therefore not listed.
On the other hand, the isomeric distribution obtained with the
H-MOR catalysts seriously deviate from the thermodynamic val-
ues. In an attempt to find a possible explanation for these results,
electrostatic charges were calculated on different naphthalenic
compounds. The results are listed in Table 4.
From Table 4, it is clear that isopropylation of the 1- or
a-posi-
tion of naphthalene is kinetically preferred over the thermodynam-
Fig. 6. Yields (%) to various products in the isomerization of 2,6-DIPN over H-Mordenite (Si/Al = 5.9) against time. The amount of 2,6-DIPN relates to the second axis on the
right side of the graph.

C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
65
Table 3
Moreau et al. [11], under identical experimental conditions as
we applied, found that Mordenite catalysts were less active than
Y-zeolites. Yields of 2,6- and 2,7-DIPN ranged from 4% to 14% in
a ratio ꢀ1/1 and approaching 80% in the DIPN isomer distribution
with Si/Al ranging from 6.9 to 20.6. The relative yield to 2-IPN ran-
ged from ꢀ55% to 84%. However, all these relative yields to both 2-
IPN and 2,6-DIPN are in fact significantly below the yield that
would result from a thermodynamic distribution, so there is no
reason to use the shape-selectivity argument.
Summary of experimental results on H-USY and H-MOR, and computational results
on the thermodynamic equilibrium composition of DIPN isomers.
Compound Method T (°C)
Catalyst Method
T (°C) T (°C)
B3PW91c H-
Catalyst T (°C)
MP2a B3LYPb H-USY
b
H-
H-
MOR
24
250
USYc MOR
b
25
200
200
250
b
2
50
5.9
2
00
1
1
1
1
1
1
1
2
2
2
,2-DIPN
,3-DIPN
,4-DIPN
,5-DIPN
,6-DIPN
,7-DIPN
,8-DIPN
,3-DIPN
,6-DIPN
,7-DIPN
0.0
2.1
0.0
0.0
1.1
1.7
0.0
0.1
0.0
4.0
0.1
0.1
4.5
3.8
0.0
0.3
0.0
6.0
0.2
0.2
8.4
4.7
0.0
0.4
0.0
5.1
0.5
0.4
5.0
2.5
0.0
1.5
0.0
3.8
0.1
0.1
7.6
4.9
0.0
0.4
0.0
0.0
14.7
11.8
5.0
17.1
18.3
0.0
1.3
19.6
12.2
13.6
18.0
8.8
15.5
16.8
0.0
2.1
15.1
10.1
Song et al. [3,33] reported relative high amounts of both 2-IPN,
and 2,6-DIPN over Mordenite catalysts with Si/Al ratios ranging
from 17 to 35 using decalin and trimethylbenzene as solvents at
2
50 °C. However, at their highest conversion (28%) in trimethyl-
benzene, the relative amount of 2-IPN is 89%, still slightly below
the thermodynamic value of 95%. Furthermore, the fused silica col-
umn applied for their analytics is insufficient to resolve the differ-
ent DIPN isomers between 41 and 42 min.
Though quite far away from our experimental conditions,
Cutrufello et al. [9] revealed the importance of coke formation on
the performance of H-MOR in the isopropylation of naphthalene
at 350 °C. Conversions decreased from 30% to 20% in time for naph-
thalene and from 80% to 60% for isopropanol. It was concluded that
48.7
46.0
40.5
46.8
37.8
42.4
42.0
43.0
40.3
42.9
a
b
c
Ref. [23].
Our results.
Ref. [19].
ically more stable 2- or b-position, as the negative charge on C
À0.286 vs. À0.094 on C . The latter is reflected in the energies of
their corresponding Transition States. On the same B3LYP/6-31G
1
is
fast oligomerization of C
3
-species and cracking of oligomers caused
2
*
almost immediate pore blocking, suggesting that the actual isopro-
pylation occurred on or near the outer surface. Their experimental
ratio 2-IPN/1-IPN was ꢀ4, and the DIPN product mixture consisted
of ꢀ8% of 1,3-DIPN, <5% of 1,7-DIPN, <5% of 1,6-DIPN, up to 60% of
level, the Transition State for the isopropylation at the 1-position
of naphthalene lies ꢀ8 kJ/mol lower as the one for isopropylation
at the 2-position. 1-IPN itself shows a kinetic preference for isopro-
pylation at the 4- and 5-position, leading to 1,4- and 1,5-DIPN. 2-
IPN shows a kinetic preference for the 4-, 5-, and 8-position, lead-
ing to 1,3-, 1,6-, and 1,7-DIPN. Only further net isomerization can
lead to the thermodynamically more stable 2,6- and 2,7-DIPN. In
a separate experimental and computational study to be reported
elsewhere, a more quantitative assessment of the observed DIPN
isomer distributions with the various catalysts and temperatures
will be presented.
2
,6-DIPN and up to 30% of 2,7-DIPN. There was a marked influence
of on-stream-time on the 2,6-DIPN/2,7-DIPN ratio, starting from
ꢀ
1 after 50 min and reaching ꢀ3 after 350 min. These results can-
not be contributed to shape-selectivity as it should be expected
that with increasing pore blocking the shape-selective formation
of 2,6-DIPN would go down with time on-stream, and the opposite
occurs.
Tasi et al. [23] tried to relate the molecular dimensions of the dif-
ferent DIPN isomers to the elliptical main channel size of H-Morde-
nite (6.5 Å Â 7.0 Å) [44]. They stated that 2,6-DIPN (6.61 Å Â 6.61 Å)
would fit far better as 2,7-DIPN (6.62 Å Â 7.26 Å) in the Mordenite
main channel. All other DIPN isomers yielded overall much larger
dimensions, ranging from 6.22 Å Â 9.46 Å for 1,4-DIPN to
4
. Discussion
4.1. Thermodynamic control in H-USY vs. shape-selectivity in H-MOR
6
.89 Å Â 9.96 Å for 1,3-DIPN. We were able to confirm these molec-
For H-USY, Brzozowski et al. [19] showed that the partition of
ular dimensions of the isomeric DIPNs, except for 2,7-DIPN, which
turned out to be equal to 2,6-DIPN in our hands.
all possible 10 DIPN isomers observed experimentally agreed very
well with the one obtained computationally (Table 4. Our isopro-
pylation and isomerization experiments with H-USY, and compu-
However, contrary to Tasi et al., we state that the sizes of the
equilibrium conformers of the different isomers cannot predict
the observed isomeric distribution in H-MOR. Except for 2,6-DIPN
and 2,7-DIPN, no other isomer should be formed at all, as they are
all far too big to fit into the channel. If this approach would be ap-
plied to H-USY (7.4 Å Â 7.4 Å) [44] also, the isomeric DIPNs, except
for 2,6-DIPN and 2,7-DIPN, would not even fit into the faujasite
pore entrance. However, even 1,4-DIPN can easily enter the fauja-
site pore by the rotation of its isopropyl groups, making the mole-
cule much smaller.
*
tational results using standard DFT (B3LYP/6-31G ), fully
confirmed the results of Brzozowski et al. Finally, our observed
partition of 2- and 1-IPN is also very close to its computed thermo-
dynamic value at 200 °C. Thus, thermodynamic control explains
the observed partition of IPN and DIPN isomers, the resulting high
selectivity to 2,6- and 2,7-DIPN, and the effect of temperature, very
satisfactorily.
H-Mordenite has been used by several authors as catalyst for
the diisopropylation of naphthalene [1–3,5,7,9,11,23,26,33,36,43].
Different kinds of shape-selectivity were used to explain the appar-
ent high selectivity to both 2-IPN and 2,6-DIPN.
In summary, none of these data do support a shape-selectivity
explanation. Our computational data both on the thermodynamic
DIPN isomer distribution and on the kinetic preferences actually
Table 4
*
B3LYP/6-31G calculated electrostatic charges on C
n
in naphthalene, 1-IPN and 2-IPN.
Compound
Electrostatic charge on C
n
C
1
C
2
C
3
C
4
C
5
C
6
C
7
C
8
Naphthalene
À0.286
À0.176
À0.485
À0.094
À0.113
+0.183
À0.094
À0.109
À0.168
À0.286
À0.303
À0.287
À0.286
À0.296
À0.267
À0.094
À0.073
À0.102
À0.094
À0.115
À0.104
À0.286
À0.259
À0.296
1
-IPN
-IPN
2

6
6
C. Bouvier et al. / Journal of Catalysis 270 (2010) 60–66
support a rather classical explanation that places the product dis-
tributions in between a kinetically and a thermodynamically con-
trolled composition. Our experimental results point into the
same direction: the isomeric distributions obtained with H-MOR
catalysts are located in between the results obtained with H-USY
at 100 °C, yielding a kinetic mixture, and at 200 °C, yielding a ther-
modynamic mixture. The precise outcome of the different experi-
ments with H-MOR depends on the activity of the catalyst, the
temperature (including coke formation and pore blocking) and a
minor role of the solvent (cyclohexane, decalin, trimethylbenzene).
Our experimental data are in line with the results of the cited
authors, and we do have a more accurate analytical procedure
Acknowledgment
This work is financially supported by the IWT-Vlaanderen
(Instituut voor de aanmoediging van Innovatie door Wetenschap
en Technologie in Vlaanderen) within the SBO-framework.
References
[
[
[
1] C. Song, C.R. Acad. Sci. Paris, Ser. IIc 3 (2000) 477.
2] Y. Sugi, M. Toba, Catal. Today 19 (1994) 187.
3] C. Song, S. Kirby, Microporous Mater. 2 (1994) 467.
[4] R. Brzozowski, W. Tecza, Appl. Catal. A 166 (1998) 21.
[
[
[
5] P. Moreau, C. He, Z. Liu, F. Fajula, J. Mol. Catal. A 168 (2001) 105.
6] Z. Liu, P. Moreau, F. Fajula, Appl. Catal. A 159 (1997) 305.
7] A. Katayama, M. Toba, G. Takeuchi, F. Mizukami, S. Niwa, S. Mitamura, J. Chem.
Soc., Chem. Commun. 1991 (1991) 39.
(
GC Â GC). Routine DFT-calculations support explanations for the
catalytic performance of both H-USY and H-MOR based on the nor-
mal interplay of thermodynamics and kinetics. The shape-selectiv-
ity argument is neither sufficient nor needed to explain our and
former experimental data.
[8] S.J. Chu, Y.W. Chen, Appl. Catal. A 123 (1995) 51.
[
9] M.G. Cutrufello, I. Ferino, R. Monaci, E. Rombi, V. Solinas, P. Magnoux, M.
Guisnet, Appl. Catal. A 241 (2003) 91.
[
[
[
10] C. Song, Stud. Surf. Sci. Catal. 113 (1998) 163.
11] P. Moreau, A. Finiels, P. Geneste, J. Solofo, J. Catal. 136 (1992) 487.
12] Ch. Subrahmanyam, B. Viswanathan, T.K. Varadarajan, J. Mol. Catal. A 226
(2005) 155.
5
. Conclusions
[
[
13] J. Wang, J.-N. Park, Y.-K. Park, C.W. Lee, J. Catal. 220 (2003) 265.
14] I. Mathew, S. Mayadevi, S. Sabne, S.A. Pardhy, S. Sivasanker, React. Kinet. Catal.
Lett. 74 (2001) 119.
From all experimental and computational results, it can only be
concluded that:
[15] I. Mathew, S. Sabne, S. Mayadevi, S.A. Pardhy, S. Sivasanker, Indian J. Chem.
Technol. 8 (2001) 469.
[
16] G. Kamalakar, S.J. Kulkarni, K.V. Raghavan, S. Unnikrishnan, A.B. Halgeri, J. Mol.
Catal. A 149 (1999) 283.
1
2
. H-USY indeed produces a thermodynamic mixture of DIPN iso-
mers, which can be predicted nicely by even routine DFT-calcu-
lations. The same calculations also yield qualitative insight in
the expected isomeric distribution of DIPN isomers under
kinetic conditions, using the electrostatic charges on the rele-
vant carbon atoms.
. Neither a thermodynamic controlled distribution of DIPN iso-
mers nor a shape-selective one is obtained in the experiments
over the various Mordenite catalysts at different temperatures.
Arguments against shape-selectivity can be divided into four
categories:
[17] E. Kikuchi, K. Sawada, M. Maeda, T. Matsuda, Stud. Surf. Sci. Catal. 90 (1994)
91.
[
[
3
18] R. Brzozowski, W. Skupinski, J. Catal. 210 (2002) 313.
19] R. Brzozowski, J.C. Dobrowolski, M.H. Jamroz, W. Skupinski, J. Mol. Catal. A 170
(2001) 95.
20] R. Brzozowski, W. Skupinski, J. Catal. 220 (2003) 13.
21] R. Brzozowski, J. Catal. 232 (2005) 366.
[
[
[
22] R. Brzozowski, Stud. Surf. Sci. Catal. 158 (2005) 1693.
[23] G. Tasi, F. Mizukami, I. Palinko, M. Toba, A. Kukovecz, J. Phys. Chem. A 105
(2001) 6513.
[
24] J.-H. Kim, Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, X. Tu, M. Matsumoto,
Microporous Mater. 5 (1995) 113.
25] A.D. Schmitz, C. Song, Catal. Today 31 (1996) 19.
[
1
. There is no reason to use the shape-selectivity argument
because the amount of 2,6-DIPN is well below the thermody-
namic one, in the presence of high amounts of the other
isomers.
. The analytical procedure might be insufficient.
. Coke formation, leading to pore blocking, actually disables
shape-selectivity related to the fit of 2,6-DIPN in the main
elliptical channel of H-Mordenite. So, these results cannot
be used to advocate shape-selectivity.
. Consequences of computational data: the data obtained by
Tasi et al. [23] and our data, on the molecular size of all DIPN
isomers indicate that apart from 2,6- and 2,7-DIPN no other
isomers should be formed if the catalytic reaction would
occur in the main channel of H-Mordenite, except for a minor
contribution of the outer surface.
[26] A.D. Schmitz, C. Song, Catal. Lett. 40 (1996) 59.
[
[
[
[
27] Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, J.-H. Kim, X. Tu, M. Matsumoto,
Stud. Surf. Sci. Catal. 90 (1994) 397.
28] J.-H. Kim, Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, X. Tu, M. Matsumoto, S.
Nakata, A. Kato, Appl. Catal. A 131 (1995) 15.
29] Y. Sugi, K. Nakajima, S. Tawada, J.-H. Kim, T. Hanaoka, T. Matsuzaki, Y. Kubota,
K. Kunimori, Stud. Surf. Sci. Catal. 125 (1999) 359.
30] Y. Sugi, J.-H. Kim, T. Matsuzaki, T. Hanaoka, Y. Kubota, X. Tu, M. Matsumoto,
Stud. Surf. Sci. Catal. 84 (1994) 1837.
2
3
[
[
[
31] P. Moreau, Z. Liu, F. Fajula, J. Joffre, Catal. Today 60 (2000) 235.
32] G. Tasi, I. Palinko, F. Mizukami, React. Kinet. Catal. Lett. 74 (2001) 317.
33] C. Song, X. Ma, H.H. Schobert, ACS Symp. Ser. 738 (2000) 305.
4
[34] C. Song, X. Ma, A.D. Schmitz, H.H. Schobert, Appl. Catal. A 182 (1999) 175.
[
35] J.A. Horsley, J.D. Fellmann, E.G. Derouane, C.M. Freeman, J. Catal. 147 (1994)
31.
2
[
36] G. Tasi, F. Mizukami, M. Toba, S. Niwa, I. Palinko, J. Phys. Chem. A 104 (2000)
1337.
[37] P. Moreau, A. Finiels, P. Geneste, F. Moreau, J. Solofo, Stud. Surf. Sci. Catal. 78
1993) 575.
38] B. Lu, T. Tsuda, Y. Oumi, K. Itabashi, T. Sano, Microporous Mesoporous Mater.
6 (2004) 1.
(
[
7
[
[
39] C. Bouvier, N. Reumkens, W. Buijs, J. Chromatogr. A 1216 (2009) 6410.
40] Wavefunction, Inc., 18401 Von Karman Avenue, Suite 370, Irvine, CA 92612,
USA.
41] C. He, Z. Liu, F. Fajula, P. Moreau, Chem. Commun. 18 (1998) 1999.
42] C. Tuma, J. Sauer, Angew. Chem. 117 (2005) 2.
We thus conclude that shape-selective formation of 2,6-DIPN in
H-MOR is a myth.
[
[
[
43] P. Moreau, A. Finiels, P. Geneste, J. Joffre, F. Moreau, J. Solofo, Catal. Today 31
3
. The actual DIPN mixture obtained with H-MOR catalysts has a
distribution in between that of a kinetic and a thermodynamic
controlled mixture and is dependent on activity of a specific
catalyst, temperature and to some extent of the solvent.
(
1996) 11.
[
44] Database of zeolite structures. Choose MOR (FAU) in the table of frameworks,
and next select Type Material to yield the channel dimensions. <http://
www.iza-structure.org/databases/>.