Enantioselective dehydration of butan-2-ol using zeolite Y modified with dithiane oxides

Article abstract of DOI:10.1006/jcat.1997.1576

Modification of zeolite H-Y by dithiane oxides (2-R-1,3-dithiane 1-oxide; R = H, CH₃, C₆H₅) is shown to enhance significantly its activity for the acid catalysed gas phase dehydration of butan-2-ol. The rate enhancement is observed for catalysts that are prepared by adding the dithiane oxide to the zeolite synthesis gel or by adsorption of the dithiane oxide onto commercial samples of zeolite H-Y. The origin of the rate enhancement is considered to result from a specific interaction between the dithiane oxide modifier with both the extra-framework and framework aluminium in the zeolite. Modification of zeolite H-Y with (R)-1,3-dithiane 1-oxide enhances the conversion of (S)-butan-2-ol compared to (R)-butan-2-ol in the temperature range 110-150°C when the two enantiomers are reacted separately. Modification with (S)-2-phenyl-1,3-dithiane 1-oxide gives a catalyst for which (R)-butan-2-ol is the most reactive of the two enantiomers. Reaction of racemic butan-2-ol over these chirally modified H-Y zeolites demonstrates that this modification procedure makes the zeolite enantiomerically discriminating and one enantiomer preferentially reacts, although both are present in the micropores under the reaction conditions. This effect is considered to be due to enantioselective rate enhancement, since, although the rate of dehydration of both enantiomers is enhanced in the chiral environment, the dehydration rate of one enantiomer is accelerated relative to the other. It is suggested that the effect is due to preferential adsorption at the chiral active site.

Full text of DOI:10.1006/jcat.1997.1576

JOURNAL OF CATALYSIS 167, 533–542 (1997)
ARTICLE NO. CA971576
Enantioselective Dehydration of Butan-2-ol Using Zeolite Y Modified
with Dithiane Oxides
�
�
�
Saskia Feast, M. Rafiq, H. Siddiqui, Richard P. K. Wells,† David J. Willock, Frank King,‡ Colin H.
�
�
�
Rochester,† Donald Bethell, Philip C. Bulman Page, and Graham J. Hutchings
�
Leverhulme Centre for Innovative Catalysis, Department of Chemistry, University of Liverpool, Liverpool L69 3BX, United Kingdom;
†
Department of Chemistry, University of Dundee, Dundee DD1 4HN, United Kingdom; and ‡Research and Technology Group,
ICI Katalco, P.O. Box 1, Billingham, Cleveland TS23 1LB, United Kingdom
Received September 25, 1996; revised December 10, 1996; accepted December 10, 1996
heterogeneous asymmetric catalyst must not only activate
Modification of zeolite H–Y by dithiane oxides (2-R-1,3-dithiane the substrates but must exert stereochemical control, and
-oxide; R = H, CH3, C6H5) is shown to enhance significantly its this requires the preferential formation of a particular di-
1
activity for the acid catalysed gas phase dehydration of butan-2-ol.
The rate enhancement is observed for catalysts that are prepared
by adding the dithiane oxide to the zeolite synthesis gel or by ad-
sorption of the dithiane oxide onto commercial samples of zeolite
H–Y. The origin of the rate enhancement is considered to result
astereoisomeric transition state. To date, three approaches
have been taken in the design of heterogeneous enantiose-
lective catalysts: (i) the use of a chiral support for an achiral
metal catalyst, (ii) modification of an achiral heterogeneous
catalyst using a chiral cofactor, and (iii) the immobilisa-
tion of a homogeneous catalyst. Early studies in this field
adopted the first of these approaches. In 1932, Schwab et al.
from a specific interaction between the dithiane oxide modifier with
both the extra-framework and framework aluminium in the zeo-
lite. Modification of zeolite H–Y with (R)-1,3-dithiane 1-oxide en-
(
8, 9) demonstrated that supporting Cu, Ni, Pd, and Pt on
hances the conversion of (S)-butan-2-ol compared to (R)-butan-2-ol
�
in the temperature range 110–150 C when the two enantiomers the enantiomers of quartz gave a catalyst capable of the
are reacted separately. Modification with (S)-2-phenyl-1,3-dithiane enantioselective dehydration of butan-2-ol, and the best re-
1
-oxide gives a catalyst for which (R)-butan-2-ol is the most reactive sults were obtained from a catalyst with submonolayer cov-
of the two enantiomers. Reaction of racemic butan-2-ol over these
chirally modified H–Y zeolites demonstrates that this modification
erage of the metal. Subsequently, natural fibres were used
as catalyst supports including silk and cellulose. Synthetic
chiral polymers were also evaluated as supports, but the
early work experienced a range of problems including lack
of reproducibility and low enantioselection. These studies
have, however, recently led to the eventual development of
polypeptides as catalysts, for example, in the epoxidation
of chalcones with enantiomeric excesses of ca. 99% (10).
The creation of a chiral catalyst surface by the ad-
sorption of a chiral modifier onto an achiral catalyst has
been successful in a number of studies, particularly for
enantioselective hydrogenation. For example, the modifi-
cation of platinum catalysts with cinchona alkaloids for the
procedure makes the zeolite enantiomerically discriminating and
one enantiomer preferentially reacts, although both are present in
the micropores under the reaction conditions. This effect is consid-
ered to be due to enantioselective rate enhancement, since, although
the rate of dehydration of both enantiomers is enhanced in the chiral
environment, the dehydration rate of one enantiomer is accelerated
relative to the other. It is suggested that the effect is due to prefer-
ential adsorption at the chiral active site.
�c 1997 Academic Press
INTRODUCTION
Interest in asymmetric synthesis continues to increase, hydogenation of prochiral �-ketoesters (11–13), and the
and this has highlighted the need for the design of highly modification of Raney nickel catalysts with diethyl tartrate,
selective asymmetric catalysts. Most of this research activ- used for the hydrogenation of prochiral �-dicarbonyl
ity concerns homogeneous catalysts and has led, for exam- compounds (14), have been extensively studied. The im-
ple, to the design of catalysts for the synthesis of
L
-Dopa mobilisation of an enantioselective homogeneous catalyst
(
1) and to the Sharpless epoxidation and dihydroxylation on to an achiral support, such as a zeolite, has also proved
processes (2). More recently, increased attention has been to be a viable approach. For example, Wan and Davis
given to the identification of suitable heterogeneous asym- (15) have immobilised a homogeneous hydrogenation
metric catalysts (3–7), since such catalysts readily overcome catalyst with retention of both the high activity and the
the problems typically encountered with homogeneous sys- enantioselection of the nonimmobilised catalyst. Very
tems, namely product recovery and catalyst separation. A recently Bein and co-workers (16) have encapsulated
533
0021-9517/97 $25.00
Copyright �c 1997 by Academic Press
All rights of reproduction in any form reserved.

5
34
FEAST ET AL.
�
chiral Mn (III) salen complexes in zeolite Y and have an autoclave (95 C, 48 h), cooled, and the product recov-
reported high enantiomeric excesses for epoxidation of ered by filtration, washed with distilled water and calcined
�
cis-�-methylstyrene with sodium hypochlorite.
(400 C, 12 h). Incorporation of a dithiane oxide modifier
These early studies demonstrate that very high levels was achieved by dissolving 1,3-dithiane 1-oxide (0.153 g) in
of enantioselection can be achieved using heterogeneous the minimum amount ofwater and addingthistogether with
asymmetric catalysts. Most of these catalysts tend to be very solutionsA and B prior to the formation ofthe gelwith com-
specific, however, and small changes to the nature of the position 11SiO2 :Al2O3 : 10NaOH : 0.02C4H8S2O : 162H2O.
catalyst, the modifier or the substrate can lead to a loss of
enantioselectivity. We therefore consider that a generic ap-
Modification of zeolite Y by adsorption of dithiane oxides.
Two commercial samples of zeolite Y were used in this
proach to the design of heterogeneous asymmetric catalysts
study:(a) H–Y (Union Carbide, LZY82) with a SiO2 :Al2O3
would have an immense benefit for this field. We believe
ratio of 4.8, and (b) Na–Y (Crosfield) with a SiO2 :Al2O3 ra-
that this can be provided by heterogeneous catalysts with
tio of 5.52. The latter sample was ion-exchanged three times
well-defined structures. For this reason we have concen-
with aqueous ammonium nitrate (0.1 mol/liter, 10 mL/g
trated our initial research (17) on the chiral modification
of zeolite catalysts since these microporous materials are
well-defined crystalline solids and the nature of the mod-
ified surface can be probed with a range of spectroscopic
techniques. In this paper we describe a method by which
zeolite Y can be made enantiomerically discriminating for
the dehydration of racemic butan-2-ol.
�
zeolite), recovered by filtration, and calcined (400 C, 8 h)
to convert it into the H–Y form.
Zeolite Y (2.0 g) was treated with a solution of 1,3-
dithiane 1-oxide (0.144 g) in deionised water (30 mL) at
�
�
5
2
1
0 C for 2 h. The modified zeolite was filtered, dried (100 C,
h), pelleted, and sieved to give particles with diameter 0.6–
mm. This procedure achieved an average loading of one
dithiane molecule per zeolite supercage and appropriate
modifications were used to obtain materials with a higher
loading.
EXPERIMENTAL
Catalyst Preparation
Catalyst characterisation. The catalysts were charac-
terised using powder X-ray diffraction, C, Al, and
Si MAS NMR spectroscopy, thermogravimetric analysis,
X-ray photoelectron spectroscopy (XPS), and electron im-
pact mass spectroscopy (EIMS).
Preparation of dithiane oxides. Enantiomerically pure
or enriched (R)- or (S)-2-substituted-1,3-dithiane 1-oxides
I, which are asymmetric at sulphur, were prepared accord-
ing to standard procedures (18).
13
27
29
|!
S
a|
|!a|
Sa!S
Catalyst Testing
+
S
a
a!
|
�
O
Catalysts (0.3 g) were tested in a glass microreactor and
all reactant and product lines were heated. Butan-2-ol was
introduced into the reactor using a calibrated syringe pump
and prevaporised prior to dilution with a nitrogen car-
R
I
II
Samples where R = H, CH3, or Ph were prepared in rier gas. The reactor exit gases were analysed with on-line
both the racemic and enantiomerically enriched forms gas chromatography using a flame ionisation detector. A
(
ee of enantiomerically enriched forms: R = H, 83% (R)- heated 16-loop sample valve was used to store samples prior
enantiomer; R = Ph, 99% (S)-enantiomer). When R = CH3 to analysis. Two analytical methods were adopted. First, for
or Ph a second chiral centre is present but the relative con- analysis of the hydrocarbons together with racemic butan-
figuration was always controlled and in the materials used 2-ol, a GS-Q megabore column (30 m) was employed.
in this study only the trans configuration was utilised. The Second, for separation of the butan-2-ol enantiomers, a
achiral 1,3-dithiane II was obtained commercially.
� -cyclodextrin trifluoroacetyl capillary column (40 m) was
employed with a column temperature of 25 C and split
ratio of 120 : 1. These conditions reliably give baseline sepa-
ration of the (R)- and (S)-butan-2-ol; for example, the sep-
aration of racemic butan-2-ol can be achieved with an ac-
curacy for each enantiomer of � 0.05% .
�
In situ incorporation of a chiral modifier during ze-
olite synthesis. Zeolite Y was synthesised from a gel
formed by the addition of two solutions. Solution A was
formed by mixing fumed silica (33.04 g) with distilled wa-
ter (120 mL) with stirring to form an opaque gel. So-
lution B was formed by dissolving aluminium powder
Infrared Spectroscopy Studies
(
2
1.45 g) in aqueous sodium hydroxide (19.80 g NaOH,
5.67 mL H2O). Solutions A and B were mixed thor-
Catalyst (ca. 100 mg) was compressed at 80 MN m 2 be-
�
oughly with stirring to form a homogeneous gel with com- tween two polished stainless-steel dies into self-supporting
position 11SiO2 :Al2O3 : 10NaOH : 162H2O, and this was discs of diameter 25 mm. These were mounted in an in-
�
aged (25 C, 120 h). After ageing, the gel was heated in frared cell fitted with fluorite windows and glass blown to

ENANTIOSELECTIVE DEHYDRATION OF BUTAN-2-OL
535
conventionalglassvacuum apparatusfitted with grease-free
TABLE 1
taps. The infrared cell could be heated to a temperature of
Catalytic Performance of Zeolite Y Modified by the Addition of
�
4
00 C with the disc in the infrared beam, and contained a
1
,3-Dithiane 1-Oxide to the Synthesis Gel for the Dehydration of
thermocouple used for temperature control.
a
Butan-2-ol
Once mounted in the cell, the catalyst was evacuated at
ambient temperature for 1 h before raising the temperature
to 150 C until no further evidence of adsorbed water was
Temper- Conver-
Product composition (% )
ature
sion
(% )
�
Catalystb
�
( C)
But-1-ene c-But-2-ene t-But-2-ene
observed (approximately 2 h). The sample was then allowed
to cool to ambient, before the disc was exposed to butan-
Na–Y
H–Y
190
7.1
16.1
36.4
47.5
140
160
1.0
4.0
16.0
15.7
17.0
17.4
28.7
36.0
37.4
55.6
47.0
45.2
2
-ol vapour. Infrared spectra of the discs were recorded
�
1
with a resolution of 4 cm at varying temperatures with a
Perkin–Elmer 1750 FTIR spectrometer.
180
Na–SO–Y
H–SO–Y
190
3.7
17.4
36.3
46.3
140
1.8
16.4
42.4
25.9
22.6
22.0
27.7
29.8
33.2
46.4
47.6
44.8
RESULTS
1
2
60
00
Conversion of Butan-2-ol over Zeolite Y Modified with
Dithiane Oxide during Synthesis
a
� 3
� 1
Catalyst (0.3 g) reacted with butan-2-ol (3.2 � 10 mol h ) prevap-
�
2
� 1
orised in diluent nitrogen (4.3 � 10 mol h ).
Samples of zeolite Y were prepared in the presence and
b
Zeolites synthesised as described under Experimental. Na– denotes
absence of a racemic sample of 1,3-dithiane 1-oxide. The that the zeolite is in the sodium-exchanged form, H– denotes that the
samples were characterised by powder X-ray diffraction zeolite is in the proton-exchanged form, and SO– denotes that 1,3-dithiane
27
1-oxide has been added to the synthesis gel.
and both were found to be crystalline; AlMASNMR spec-
troscopy confirmed that all aluminium atoms were tetrahe-
drally coordinated. 29Si MAS NMR spectroscopy was used
to determine the framework Si/Al ratio of the zeolite. For
the zeolite Y made in the absence of the dithiane oxide the
Si/Al ratio was 2.7, which is similar to commercial samples
of zeolite Y. For the zeolite synthesised in the presence of
dithiane oxide, the framework Si/Al ratio was significantly
lower, at 2.0, indicating that the presence of the dithiane in
the synthesis gel increases the incorporation of aluminium
into the framework. C MAS NMR spectroscopy of a sam-
ple prepared using 1,3-dithiane 1-oxide in the synthesis gel
showed signals characteristic of the 1,3-dithiane 1-oxide,
confirming that the modifier was incorporated in the zeo-
lite. These materials were found to be inactive for the dehy-
the modifier are incorporated by this method, as evidenced
by the integrated intensities of the 13C MAS NMR spec-
tra. In addition, to prepare catalytically active materials it
+
is necessary to carry out ion exchange of the Na form of
+
the zeolite with NH4 cations and to form the active proton
form by a calcination procedure. During this process, it can
be expected that some of the protonated dithiane oxide will
also be exchanged as the cation and an equilibrium will be
13
+
+
established between the concentrations of Na , NH , and
4
+
C4H9S2O as counter-cations in the zeolite. In addition, the
calcination procedure required for the conversion of the
+
+
NH form to the H form could lead to thermal decompo-
4
sition of the dithiane oxide. For these reasons, we decided
not to pursue this method of catalyst synthesis, and instead
a postsynthesis modification procedure was investigated.
dration of racemic butan-2-ol, and conversions typically of
�
<
8% were observed even at temperatures as high as 200 C
(
Table 1). The presence of the dithiane oxide modifier leads
to a decrease in activity, and no effect on the chemoselec-
tivity was observed. In this case we consider that the zeolite
is present in the sodium ion-exchanged form, and hence the
concentration of Brønsted acid sites is very low. The mate-
Conversion of Butan-2-ol over Zeolite Y Modified
with Dithiane Oxide Postsynthesis
Catalyst stability. Samples of zeolite H–Y modified by
rials were then ion-exchanged with ammonium nitrate and adsorption of 1,3-dithiane 1-oxides at one molecule per
�
calcined for a short time at 400 C to give the proton ex- supercage were prepared and characterised by powder X-
changed forms. The dehydration of racemic butan-2-ol was ray diffraction and 13C MAS NMR spectroscopy. Powder
then examined usingthe ion-exchanged materials(Table 1). X-ray diffraction showed that the modification process had
+
27
The proton forms are much more active than the Na forms, not caused anysignificant lossofcrystallinity. The AlMAS
and the proton form of the dithiane oxide-modified zeo- NMR spectra of both the unmodified zeolites indicated the
lite is much more active than the unmodified zeolite, and presence of nonframework aluminium, but this was not sig-
also gives a somewhat higher selectivity for production of nificantly increased by the adsorption of the dithiane oxide.
13
but-1-ene.
C MAS NMR spectroscopy of the modified zeolite, when
These results show that modified catalysts can be pre- compared with the solution NMR spectra of the equivalent
pared by the incorporation of dithiane oxide modifiers in dithiane oxide (Fig. 1), indicated that the dithiane oxide
the synthesis gel of zeolites. However, only low levels of was molecularly adsorbed within the zeolite. The stability

5
36
FEAST ET AL.
TABLE 2
EIMS Analysis for 2-Phenyl-1,3-dithiane 1-oxide-Modified
Zeolite Y
m/z
�
Temperature ( C): 40–100
200–300
300–400 400–500
Zeolite–HY
46, 44
46, 44
—
ND
ND
—
—
2-Phenyl-1,3-
212, 121, 90, 45
dithiane 1-oxide
Zeolite HY
46, 45, 44
—
ND
106, 91, 78,
64, 45
2-Phenyl-1,3-
dithiane 1-oxide
Note. ND, not detected.
did not show any desorption or decomposition products in
this temperature region. Secondly, the modified samples
�
were heated (180 C, 2 h) in a fixed-bed microreactor un-
der flowing nitrogen. The samples were cooled and then
extracted with deuteriated chloroform using a soxhlet pro-
1
cedure. Analysis of the extract by H NMR spectroscopy
confirmed that the dithiane modifier could be recovered
unchanged by this procedure. Thirdly, thermogravometric
�
analysis (heating rate 10 C/min) was employed to assess
further the stability of the modifier in the zeolite. A com-
parative analysis of the unmodified zeolite H–Y and zeolite
H–Y modified with 1,3-dithiane 1-oxide is shown in Fig. 2.
�
Both samples show similar weight loss below ca. 250 C due
to water loss. The modified zeolite does show an additional
�
weight loss at ca. 450 C. This behaviour is consistent with
the results of the EIMS study.
To assess whether the modifier was evenly distributed
throughout the microporousstructure ofthe zeolite, an XPS
study was conducted for zeolite H–Y modified with 1,3-
dithiane 1-oxide at one molecule per supercage. Analysis
of the peak areas of the S2p and the Si2p peaks were used to
determine the S/Si ratio (19), which was found to be 0.07.
The calculated value of the ratio expected for the bulk of the
FIG. 1. 13C NMR spectra (a) 1,3-dithiane 1-oxide dissolved in CDC13;
TMS; (b) 1,3-dithiane 1-oxide adsorbed in zeolite Y (LZY 82, Union
*
Carbide; one molecule/supercage).
of the modified zeolites was then investigated with a range
of techniques. First, the modified zeolites were examined
with EIMS. The samples were gradually heated at a con-
�
stant rate (2 C/min), and the evolved gases were analysed.
The results for 2-phenyl-1,3-dithiane 1-oxide as modifier
are shown in Table 2, and the results are compared with the
pure modifier and the unmodified zeolite. The pure dithiane
�
oxide decomposed in the region 200–300 C, whereas the
zeolite-modified sample did not show any decomposition
�
products below 400 C; at this temperature there was no ev-
idence for the molecular ion (m/e 212), which indicates that
the dithiane oxide decomposes at this temperature rather
FIG. 2. TGA of (a) zeolite Y and (b) 1,3-dithiane 1-oxide adsorbed
than simply desorbing. The unmodified zeolite, as expected, in zeolite Y (LZY 82, Union Carbide; one molecule/supercage).

ENANTIOSELECTIVE DEHYDRATION OF BUTAN-2-OL
537
TABLE 3
molecule/supercage), and these were used as catalysts for
the dehydration of racemic butan-2-ol. The results are
shown in Table 3 together with those for the unmodi-
fied zeolites. It is apparent that modification of zeolite
H–Y (LZY 82, Union Carbide) with 1,3-dithiane 1-oxide
givesconsiderablyhigher catalyticactivitythan zeolite H–Y
(Crosfield) modified in the same way. However, both mod-
ified samples displayed an enhanced catalytic activity when
compared with their respective unmodified zeolite. This en-
hanced activity was maintained for many hours, whereas
the unmodified zeolites were rapidly deactivated under the
reaction conditions. Turnover numbers of >1200 based on
the dithiane 1-oxide were observed for the conversion of
butan-2-ol over the modified zeolite, indicating that the en-
hanced activity is a sustained catalytic effect. Modification
with 2-methyl-1,3-dithiane 1-oxide gave similar results, but
Catalytic Performance of Dithiane Oxide-Modified Zeolite Ya
Temper- Conver-
Product composition (% )
ature
( C)
sion
(% )
Catalystb
US–Y
�
But-1-ene c-But-2-ene t-But-2-ene
115
0
—
—
47.4
51.2
—
36.6
33.2
1
1
40
50
6.0
18.2
16.0
15.6
SOUS–Y
100
140
11.6
95.4
97.3
24.4
15.0
13.3
51.3
46.0
44.2
24.3
39.0
42.5
(
R = H)
1
80
SOUS–Y
115
150
10.0
98.8
100
15.9
13.7
12.4
33.6
38.5
35.3
50.5
47.8
52.3
(
R = CH3)
1
90
H–Y
115
0
—
—
36.0
37.0
—
47.0
46.0
1
1
40
70
2.0
5.0
17.0
17.0
2
-phenyl-1,3-dithiane 1-oxide did not lead to such a pro-
nounced rate enhancement, although it still gave a higher
activity than the unmodified zeolite. Following reaction, all
the catalysts were analysed by powder X-ray diffraction,
and some loss of crystallinity was apparent. The zeolites
were found to have coked during the reaction, but the dithi-
ane oxide modifier could still be recovered by soxhlet ex-
traction using deuteriated chloroform.
SOH–Y
110
140
1.7
10.6
14.6
12.5
28.3
29.2
57.0
58.3
(
R = H)
SOUS–Y
120
153
0.5
3.5
9.1
17.2
15.9
16.5
31.0
34.5
36.5
51.7
49.6
47.0
(
R = C6H5)
1
70
a
� 3
� 1
Catalyst (0.3 g) reacted with butan-2-ol (3.2 � 10 mol h ) prevapor-
�
2
� 1
ised in diluent nitrogen (4.3 � 10 mol h ).
Key to catalysts: US–Y, LZY 82, Union Carbide; H–Y, ion-exchanged
NaY, Crosfield; SO denotes the zeolite has been modified by addition of
b
A range of control experiments was carried out to probe
the origin of this rate enhancement (Table 4). Modifica-
2
-R-1,3-dithiane 1-oxide (one molecule per supercage, R = H, CH3, C6H5). tion of zeolite Y with 1,3-dithiane (compound II) using
an analogous procedure led to a small rate enhancement,
significantly lower than that observed for modification by
sample for uniform distribution is 0.07, and the agreement
between these values indicates that the modifier is indeed
uniformly distributed.
the equivalent dithiane oxide, indicating the importance of
the sulphoxide oxygen atom as a vital component of the
catalytically active site. Investigation of a nonmicroporous
Catalytic performance of dithiane oxide-modified zeolites SiO2/Al2O3 catalyst indicated that in this case the dithi-
for the dehydration of butan-2-ol. A range of catalysts ane oxide modifier acted as a poison. In addition, the same
was prepared by the adsorption of racemic 2-substituted- reaction carried out over dithiane oxide supported on an
1
,3-dithiane 1-oxides (substituent R = H, CH3, C6H5; one inactive support, BN, indicated that the dithiane oxide by
TABLE 4
Reaction of Racemic Butan-2-ol over Modified Catalystsa
Yb
Y–SOc
Y–Sd
SiO2/Al2O3e
SiO2/Al2O3 f
BN–SO g
115 225
Catalyst:
�
Temperature ( C)
Conversion (% )
Selectivity (% )
But-1-ene
trans-But-2-ene
cis-But-2-ene
115
0
225
90
115
90
110
3.6
175
35.6
200
31.5
200
24.5
0
0
—
—
—
17.7
40.6
41.6
8.3
53.8
37.8
15.4
53.8
30.7
14.7
46.5
38.8
16.5
31.7
51.8
21.6
33.9
46.5
—
—
—
—
—
—
a
� 3
� 1
)
Catalysts (0.3 g) tested in a conventional glass microreactor using on-line GC analysis with butan-2-ol (3.3 � 10 mol h
�
2
� 1
prevaporized in a nitrogen diluent (3.7 � 10 mol h ).
b
Zeolite Y (ultrastabilized LZY 82, Union Carbide).
c
Zeolite Y modified with (� )-1,3-dithiane 1-oxide, one molecule per supercage (7.6 wt% ).
d
Zeolite Y modified with 1,3-dithiane, one molecule per supercage (7.2 wt% ).
e
Nonmicroporous silica alumina, SiO2/Al2O3 = 5.7.
f
Nonmicroporous silica alumina, SiO2/Al2O3 = 5.7, modified with 1,3-dithiane 1-oxide (6.9 wt% ).
g
Boron nitride modified with 1,3-dithiane 1-oxide (7.4 wt% ) under these conditions unmodified BN was also inactive.

5
38
FEAST ET AL.
itself does not exhibit any catalytic activity for this reac-
tion. These control experiments demonstrate that it is the
combination of the microporous zeolite together with the
dithiane oxide that is important for the creation of the en-
hanced activity catalytic site observed with the modified
zeolites.
In situ FTIR spectroscopic investigation of butan-2-ol
dehydration over dithiane oxide-modified zeolite H–Y.
The adsorption of racemic butan-2-ol on unmodified ze-
�
olite H–Y (Crosfield) at 50 C leads to the formation of
hydrogen-bonded species. Figure 3 shows the IR spectra
�
of H–Y at 50 C, before and after exposure to butan-2-ol
vapour. Two hydroxyl stretching vibrations are observed,
�
1
at 3632 and 3543 cm , denoted high (HF) and low (LF)
frequency bands, respectively. van Santen et al. have shown
that the protons of the LF band are hydrogen bonded to
other lattice oxygen atoms (20). Adsorption of butan-2-ol
leads to the formation of hydrogen bonds between the ad-
sorbate and the Brønsted acidic zeolite hydroxyl groups
as evidenced by the HF zeolite hydroxyl band decreas-
ing in integrated intensity with respect to the LF band.
Direct comparison of band intensity is not valid in this
case as the extinction coefficients of bands originating from
hydrogen-bonded species are significantly increased com-
pared to those of the free molecular species (21). The broad
FIG. 4. FTIR spectra of zeolite H–Y after exposure to butan-2-ol
133 N m ) at (a) 50 C, (b) 80 C, and (c) 90 C.
�
2
�
�
�
(
�
1
bands centred around 2400 and 1700 cm are two of the
three characteristic bands formed by species strongly hy-
drogen bonded to zeolites (22–24).
On heating under a static vacuum, some of the adsorbed
butan-2-ol desorbs. A small amount of butanone is also
formed, indicated by the presence of a C == O stretch
�
1
�
band centred at 1705 cm . However, at 90 C water is
eliminated, signified by the large increase in intensity of the
�
1
water deformation band at 1645 cm . At the same time,
a shift occurs in the CH3 and CH2 band positions (Fig. 4).
For hydrogen-bonded molecules, the asymmetric and
symmetric CH3 stretch bands of adsorbed butanol appear
at 2976 and 2888 cm ; above 90 C, these bands shift to
960 and 2873 cm . This is indicative of the formation of
�
1
�
�
1
2
butoxide species (25).
When butan-2-ol was adsorbed on to the zeolite modified
with 1,3-dithiane 1-oxide (one molecule/supercage), the
same initial hydrogen-bonded species was formed. How-
ever, on heating, the elimination of water and the band shift
characteristic of butoxide formation, described above, oc-
�
�
curred at 70 C, rather than 90 C. Modification with higher
loadings of dithiane-1-oxide did not lead to a further de-
crease in the temperature of butoxide formation. A blank
experiment in which the zeolite H–Y was treated with wa-
�
ter at 50 C for 2 h in a procedure analogous to that used in
the preparation of the dithiane oxide modified zeolite did
not give this effect. For the water treated zeolite H–Y, the
�
butoxide species was still formed at 90 C, indicating that
the observed decrease in the temperature of formation of
the butoxide species was due to the modification of zeolite
Y by the 1,3-dithiane 1-oxide.
�
FIG. 3. FTIR spectra of zeolite H–Y at 50 C (a) before and (b) after
exposure to butan-2-ol (133 N m ).
�
2

ENANTIOSELECTIVE DEHYDRATION OF BUTAN-2-OL
539
TABLE 5
Effect of Reactant Feed Rate on the Conversion of (R)-
and (S)-Butan-2-ol to Alkenes over Zeolite Y Modified
a
with (R)-1,3-Dithiane 1-oxide
Conversion (% )
Reactant flow rate
�
2
� 1
(
10 mol h )
(R)-Butan-2-ol
(S)-Butan-2-ol
2
6
.1
.6
5.7
4.5
12.2
8.0
1
1
3.0
8.0
1.48
0.01
4.75
1.05
a
Zeolite Y (ion-exchanged Na–Y ex Crosfield modified with
one molecule per supercage of (R)-1,3-dithiane-1-oxide, 0.3 g)
reacted separately with (R)- and (S)-butan-2-ol prevaporised
�
2
� 1
�
in diluent nitrogen (5.6 � 10 mol h ) at 110 C.
FIG. 5. Effect of temperature on the conversion of (R)-butan-
-ol and (S)-butan-2-ol over zeolite H–Y (Crosfield) modified with (R)-
,3-dithiane 1-oxide (one molecule/supercage); reaction conditions as in
2
1
for conversion and selectivity are shown in Figs. 5 and 6.
It is clear that when the zeolite is modified with (R)-1,3-
dithiane 1-oxide, (S)-butan-2-ol is more reactive over the
entire temperature range compared with (R)-butan-2-ol.
In a further set of experiments, the temperature was main-
Table 6.
Reaction of R-butan-2-ol and S-butan-2ol over zeolite
Y modified with enantiomerically enriched dithiane oxides.
A catalyst was prepared by adsorbing enantiomerically en-
riched (R)-1,3-dithiane 1-oxide (enantiomeric excess 83% )
in zeolite H–Y (Crosfield, one molecule/supercage). A se-
�
tained at 110 C, and the flowratesofthe (R)-and (S)-butan-
2
-ols were separately varied (Table 5). Enhanced reactivity
of (S)-butan-2-ol was again observed.
A similar set ofexperimentswasconducted with zeolite Y
ries of control experiments were carried out in which enan-
tiomerically pure (R)- and (S)-butan-2-ol were reacted over
the modified zeolite, and it was determined that, in the tem-
(
LZY, 82, Union Carbide) modified with enantiomerically
enriched (S)-2-phenyl-1,3-dithiane 1-oxide (enantiomeric
excess 99% ) at a loading of one molecule/supercage; the
results for the separate reactions of (R)- and (S)-butan-2-
ols with increasing temperatures are shown in Fig. 7. In
this case it is apparent that (R)-butan-2-ol is more reac-
tive than (S)-butan-2-ol over the temperature range. Over
both modified zeolites, the activation energies for the sepa-
rate conversions of (R)- and (S)-butan-2-ols are very similar
�
perature range investigated (110–150 C), no racemisation
of the butan-2-ol occurred. In addition, following reaction
with either enantiomerically pure (R)- or (S)-butan-2-ol, or
racemic butan-2-ol, the chiral modifier was extracted with
deuteriated chloroform, as described above, and was inves-
1
tigated by H NMR spectroscopy in the presence of a chiral
shift reagent. In this way it was determined that these treat-
ments had not induced racemisation of the chiral modifier.
A series of experiments were then carried at increasing
�
1
(
105 � 5 kJ mol ).
Reaction of racemic butan-2-ol over zeolite Y modified
temperatures for the separate reactions of (R)- and (S)- with enantiomerically enriched dithiane oxides. The dif-
butan-2-ol over the chirally modified zeolite; the results ference in reaction rate observed for the reaction of the
FIG. 6. Effect of temperature on selectivity (a) (R)-butan-2-ol and (b) (S)-butan-2-ol over zeolite H–Y (Crosfield) modified with (R)-1,3-dithiane
1
-oxide (one molecule/supercage); reaction conditions as in Table 6; product selectivity: ᭹, but-1-ene; ᭡, cis-but-2-ene; , trans-but-2-ene.

5
40
FEAST ET AL.
DISCUSSION
Modification of zeolite H–Y with dithiane oxides leads to
the creation of a catalyst that has enhanced activity for the
acid-catalysed dehydration of butan-2-ol when compared
with the unmodified zeolite. This effect is observed both
for samples prepared by addition of the modifier to the
synthesis gel and for the postsynthesis modification of two
different commercial samples of zeolite Y (Crosfield and
Union Carbide). When the modifier is added to the synthe-
sis gel, the resulting zeolite contains only a small quantity
of the modifier, but the effect is still apparent.
The interaction of alcohols with the acid forms of zeolites
has been particularly well studied in recent years (26, 27),
and it is generally considered that the adsorbed species that
FIG. 7. Effect of temperature on the conversion of (R)-butan-2-ol
and (S)-butan-2-ol over zeolite H–Y (LZY 82, Union Carbide) modified
with (S)-2phenyl-1, 3-dithiane 1-oxide (one molecule/supercage); reaction is formed initially is an alkoxide. In situ FTIR spectroscopy
conditions as in Table 6.
confirmed that, in the present study, an alkoxide species is
formed on reaction of the butan-2-ol with both the unmod-
ified zeolite and the dithiane oxide-modified zeolite. Con-
enantiomerically pure (R)- and (S)-butan-2-ol suggests that sistent with the enhanced reactivity observed with the flow
this catalyst system should be able to discriminate between reactor studies, the in situ FTIR spectroscopy experiments
the two enantiomers of racemic butan-2-ol. Hence, a se- reveal that the alkoxide species is formed at a significantly
ries of experiments were carried out for the reaction of lower temperature for the dithiane oxide modified zeolite
racemic butan-2-ol over zeolite H–Y modified with enan- than for the unmodified zeolite.
tiomerically enriched dithiane oxides; the results are given
Dehydration of butan-2-ol leads to the formation of
in Table 6. As observed previously with the enantiomer- but-1-ene, cis-but-2-ene and trans-but-2-ene. At equilib-
ically pure (R)- and (S)-butan-2-ols, when the zeolite is rium, but-1-ene becomesthe major product at temperatures
�
modified with (R)-1,3-dithiane 1-oxide, it is apparent that >950 C, and, for the conditions used in this study, the equi-
S)-butan-2-ol reacts in preference to (R)-butan-2-ol. In librium compositions are 47.5 : 35 : 17.5, for trans-but-2-ene,
addition, when the zeolite is modified with (S)-2-phenyl- cis-but-2-ene, and but-1-ene, respectively. When the mod-
,3-dithiane 1-oxide, (R)-butan-2-ol reacts in preference to ified zeolites are used as catalysts, product compositions
S)-butan-2-ol. In both these sets of experiments, enan- that are very close to the equilibrium values are observed,
(
1
(
tioselection was observed for a number of turnovers, al- and consequently it is not possible to draw any mechanistic
though the effect was short-lived at the higher tempera- information from the observed chemoselectivity. However,
tures (turnover numbers, 17 and 2, respectively (based on it is interesting to note that for the sulphoxide-modified
�
dithiane 1-oxide) at 110 and 150 C, respectively).
catalyst as the temperature increases the yield of but-1-ene
TABLE 6
Reaction of Racemic Butan-2-ol over Zeolite Y Modified by Enantiomerically Enriched Dithiane Oxides
b
� 3
� 1
)
Catalyst
modifier
Product composition (� 10 mol h
Butan-2-ol conversion (% )
Temp.
( C)
Conversion
(% )a
�
zeolite
But-1-ene
But-2-ene
R-Butan-2-ol
S-Butan-2-ol
R
S
Relative rate
c
Y
Y
Y
Y
(R-)-I, R = H
(R-)-I, R = H
(R-)-I, R = H
(S-)-I, R = Ph
(S-)-I, R = Ph
110
120
150
110
120
0.5
1.3
9.9
4.2
7.5
—
0.037
0.085
0.620
0.271
0.507
3.673
3.669
3.657
3.399
3.260
3.640
3.581
2.968
3.660
3.594
0.002
0.006
0.018
0.276
0.415
0.035
0.094
0.707
0.015
0.081
1 : 17.5
1 : 15.7
1 : 39.3
18.4 : 1
5.1 : 1
c
0.015
0.105
0.020
0.044
c
d
d
Y
a
Total conversion of R- and S-butan-2-ol.
Flow rate, 10 mol h
Zeolite Y (Crosfield NaY, ion exchanged with NH4NO3 and calcined at 550 C), modified with R-1,3-dithiane 1-oxide, one molecule per supercage
b
c
� 3
� 1
.
�
�
3
� 1
� 3
� 1
(
0.1 g), tested in a convential glass microreactor with racemic butan-2-ol (7.35 � 10 mol h ), prevaporized in a nitrogen diluent (6.7 � 10 mol h ).
d
Zeolite Y (ultrastabilized LZY 82, Union Carbide) modified with S-2-phenyl-1,3-dithiane 1-oxide, one molecule per supercage (0.1 g), tested in a
convential glass microreactor with racemic butan-2-ol (7.35 � 10� mol h ), prevaporized in a nitrogen diluent (6.2 � 10 mol h ).
3
� 1
� 3
� 1

ENANTIOSELECTIVE DEHYDRATION OF BUTAN-2-OL
541
increases slightly whereas the yield of trans-but-2-ene de- able to discriminate between the enantiomers of butan-2-ol.
creases (Fig. 6). This is the reverse of the expected trend This important effect is achieved by enantioselective rate
based on equilibrium data (28, 29) and indicates that the enhancement, i.e., both enantiomers react faster in the chi-
modified catalysts are exerting some slight kinetic control. ral environment than in the absence of the chiral modifier,
All these data indicate that when the dithiane oxide is but one reacts faster than the other. The observation that
added to the zeolite, a new high activity site is formed. modification leads to the formation of a high activity site
It is apparent that the enhancement in activity is great- has also been observed for the cinchona alkaloid modifi-
est for the sample of zeolite Y supplied by Union Carbide cation of platinum for the enantioselective hydrogenation
(
LZY 82). This is an ultrastabilised zeolite Y that has been of prochiral �,�-ketoesters. In the case of cinchona modi-
steamed and consequently contains significant amounts of fication, rate enhancements of three orders of magnitude
nonframework aluminium. The rate enhancement is less are observed (11), and hence the observation of a rate en-
pronounced, but still significant, with the modification of hancement effect may be a prerequisite for the design of
the zeolite Y supplied by Crosfield. The proton form of effective enantioselective catalysts.
this zeolite was obtained by ion exchange of the commer-
It is observed that the activation energies for the de-
cially supplied Na–Y sample which originally contained no hydration of the separate enantiomers over the zeolites
extraframework aluminium. The ion-exchange procedure doped with the enantiomerically enriched dithiane oxides
-1
adopted led to dealumination of the framework, but the ex- are identical within experimental error (105 � 5 kJ mol ).
tent wasmuch lessmarked than with zeolite H–Y (LZY-82). In addition, it should be noted that the enantiomeric dis-
Hence, the degree of enhancement observed may be related crimination obtained from the reaction of racemic butan-
to the presence of extraframework aluminium. This must, 2-ol is higher than that based on the conversions of the
however, act in combination with the Brønsted acid site separate enantiomers. The apparent inconsistency can be
of the bridging hydroxyl group associated with the zeolite resolved by regarding the modified zeolite as if it were an
framework aluminium, since the dehydration reaction in- enzyme (a zeozyme?) operating with its active sites satu-
vestigated is catalysed by Brønsted acidity. A number of rated with substrate. The conversions of the separate enan-
studies have indicated that a combination of extraframe- tiomers then yield the relative reactivities in the catalytic
work (Lewisacid sites) and framework (Brønsted acid sites) step, whereas the results using the racemate reflect the
aluminium can give enhanced acidity in zeolite catalysts. enantioselection not only of the catalytic step but also of
�
Haag and Lago (30) showed that steaming zeolites at 500 C the competitive binding of the substrates. Present indica-
with low levels of water vapour can lead to an increase tions, supported by molecular modelling studies, are that
in the activity of the zeolite for acid-catalysed reactions. the latter effect is the larger.
Mirodatos and Bartemeuf (31) showed that superacid sites
It is important to note that the reactants and products in
could be created in mordenite by a steaming procedure. this reaction are gas phase. However, the reactant butan-
These effects were subsequently explained by Fritz and 2-ol is present at high partial pressures, and, at the lower
�
Lunsford (32) in terms of the initial dealumination of the temperature used in this study, which is only 12 C above
zeolite to form nonframework aluminium that imparts, pre- the boiling point of butan-2-ol, it is possible that condensa-
sumably through an electrostatic effect, strong acidity of tion may occur in the micropores of the zeolite. This effect
the remaining framework Brønsted acid sites. This leads to would not, however, be expected to contribute significantly
an increase in acidity of the zeolite, and, for simple acid- to the results obtained at the higher temperatures, where
catalysed reactions such as cracking, an enhanced activity enantioselection is still observed. We consider that these
is observed. In the present work, the high activity site is results provide the first example of a gas-phase enantios-
considered to be formed by the specific interaction of the elective reaction heterogeneously catalysed by a zeolite.
dithiane oxide with both the extraframework aluminium Although the initial approach has been to design a cata-
and the Brønsted acid site associated with the framework lyst that consumes chiral molecules, the study is, in effect,
aluminium. Supporting evidence for this proposed model a proof of the concept that a zeolite can be modified so
comes from detailed 27Al MAS NMR spectroscopy studies that it preferentially catalyses reaction of one enantiomer
(
33) and molecular simulations (34) that have shown that of a chiral substrate in the presence of both. In view of the
the dithiane oxide interacts strongly with the extraframe- immense range of microporous zeolites available together
work aluminium atoms. with the large number of enantiomerically pure modifiers,
The formation of high activity acid sites enables these we consider that this approach will provide the basis for the
modified zeolites to be used as enantioselective catalysts, generic design of a new type of enantioselective catalysts.
since, although only one active site per supercage is mod-
ACKNOWLEDGMENTS
ified, it is many orders of magnitude more active than the
remainingunmodified sites. When the dithiane oxide isused
in the enantiomerically enriched form, the active sites are for financial support.
We thank the SERC Catalysis and Interfaces Initiative and ICI Katalco

5
42
FEAST ET AL.
REFERENCES
18. Page, P. C. B., Gareh, M. T., and Porter, R. A., Tetrahedron: Asymmetry
139 (1993); Page P. C. B., Wilkes, R. D., Namwindwa, E. S., and Willy,
2
1
. Knowles, W. S., Sabacky, M. J., Vineyard, B. D., and Weinkauf, D. J.,
J. Am. Chem. Soc. 97, 2565 (1975).
. Katsuki, T., and Sharpless, K. B., J. Am. Chem. Soc. 102, 5974
M., Tetrahedron 52, 2125 (1996).
19. Seah, M. P., in “Practical Surface Analysis” (D. Briggs and M. P. Seah,
Eds.), p. 181. Wiley, New York, 1983.
2
(
1980).
20. Brugmans, M. J. P., Kleyn, A. W., Lagendijk, A., Jacobs, W. P. J. H.,
and van Santen, R. A., Chem. Phys. Lett. 217, 117 (1994).
21. Pimentel, G. C., and McClellan, A. L., “The Hydrogen Bond.”
Freeman, San Francisco, 1960.
3
4
5
. Izumi, Y., Adv. Catal. 32, 215 (1983).
. Brunner, H., Topics Stereochem. 18, 129 (1988).
. Bartock, M., “Stereochemistry of Heterogeneous Metal Catalysts,”
p. 511 Wiley, New York, 1985.
22. Bratos, S., and Ratajczak, H., J. Chem. Phys. 76, 77 (1982).
6
. Morrison, J. D., “Asymmetric Synthesis,” Vol. 5. Academic Press, 23. van Santen, R. A., Rec. Trav. Chimiques Pays-Bas. 113, 423 (1984).
New York, 1985.
24. Claydon, M. F., and Sheppard, N., J. Chem. Soc., Chem. Commun. 1431
(1969).
25. Williams, C., Makarova, M. A., Malysheva, L. V., Paukshtis, E. A., and
Zamaraev, J. Chem. Soc., Faraday Trans. 86, 3473 (1990).
26. Chang, C. D., Stud. Surf. Sci. Catal. 61, 393 (1991).
7
8
9
. Blaser, H. U., Tetrahedron: Asymmetry 843 (1991).
. Schwab, G. M., and Rudolph, L., Naturwisserschaft 20, 362 (1932).
. Schwab, G. M., Rost, F., and Rudolph, L., Kolloid Z. 68, 157
(
1934).
1
0. Kroutil, W., Mayon, P., Lasterra-Sanchez, M. E., Maddrell, S. J., 27. Hutchings, G. J., Johnson, P., Lee, D. F., Warwick, A., Williams, C. D.,
Roberts, S. M., Thornton, S. R., Todd, C. J., and Tuter, M., J. Chem.
Soc., Chem. Commun. 845 (1996).
1. Blaser, H. U., Jalett, H. P., Monti, D. M., Baiker, A., and Wehrli, J. T.,
Stud. Surf. Sci. Catal. 67, 147 (1991).
2. Webb, G., and Wells, P. B., Catal. Today 12, 319 (1992).
3. Baiker, A., Stud. Surf. Sci. Catal. 101, 51 (1996).
4. Keane, M. A., and Webb, G., J. Catal. 136, 1 (1992).
5. Wan, K. T., and Davis, M. E., Nature 370, 449 (1994).
and Wilkinson, M., J. Catal. 147, 177 (1994).
28. Kilpatrick, J. E., Prosen, K. S., Pitzer, K. S., and Rossini, F. D., J. Res.
Natl. Bur. Stand. 36, 559 (1946).
29. Butler, A. C., and Nicolaides, C. P., Catal. Today 18, 443 (1993).
30. Haag, W. O., and Lago, R. M., U.S. Patent 4,326,994.
31. Mirodatos, C., and Bartemeuf, D., J. Chem. Soc., Chem. Commun. 39
(1981).
1
1
1
1
1
1
32. Fritz, P. O., and Lunsford, J. H., J. Catal. 118, 85 (1989).
6. Bein, T., and Ogunwumi, S. B., “Chiral Catalysis,” paper 006. ACS 33. Wells, R., Hutchings, G. J., King, F., and Rochester, C. H., manuscript
Symposium, Orlando, FL, 1996. in preparation.
7. Feast, S., Bethell, D., Page, P. C. B., Siddiqui, M. R. H., Willock, D. J., 34. Feast, S., Bethell, D., Page, P. C. B., Siddiqui, M. R. H., Willock, D. J.,
1
King, F., Rochester, C. H., and Hutchings, G. J., J. Chem. Soc., Chem.
Commun. 2409 (1995).
Hutchings, G. J., King, F., and Rochester, C. H., Stud. Surf. Sci. Catal.
101, 211 (1996).