New transition-state models and kinetics of elimination reactions of tertiary alcohols over aluminum oxide

Article abstract of DOI:10.1021/jo971609u

A new transition-state model was developed in order to justify the anti intramolecular E2 elimination with cis (Z)-preference over pure alumina and interinolecular E2 elimination with trans (E)-preference over doped alumina. The reactions of model compounds 1,2,3-triphenyl-2-propanol (1), 1,2-diphenyl-2-propanol (2), and 3,3,3-trideuterio-1,2-diphenyl-2-propanol (3) with aluminum oxides with a pH range of 4.5-9.5 and thorium oxide in the temperature range of 200-350 °C in 2-hexanol have been investigated. Over acidic alumina (pH = 4.5 ±0.5), the ratio of E-isomer to Z-isomer (E/Z ? 2) for 2 was found to remain unchanged in this temperature range. At 300 °C, however, Saytzeff elimination favored Hofmann. Over pure alumina the E/Z ratio was equal to 0.650 (2-alkene/1-alkene = 0.750). At equilibrium, the E/Z ratio for 2 was equal to 4.5 with the formation of trace amounts of Hofmann adducts. The ratio of Saytzeff to Hofmann elimination was found to be pH independent. Any decrease in pH caused a slight increase in the E/Z ratio. The average primary kinetic isotope effect (k_H/k_D) for elimination at 230 °C was equal to 3.775 ± 0.227. The ratio of E/Z over thorium oxide at 300 and 350 °C was similar to that of aluminum oxide at 300 °C, but the Saytzeff elimination was surprisingly favored over Hofmann! The energy of activation (E_a), entropy of activation (AS?), selectivity, isotope effect (k^H/k_D), and semiempirical calculation (AM1) all agreed with concerted E2 elimination.

Full text of DOI:10.1021/jo971609u

J . Org. Chem. 1998, 63, 7619-7627
7619
New Tr a n sition -Sta te Mod els a n d Kin etics of Elim in a tion
Rea ction s of Ter tia r y Alcoh ols over Alu m in u m Oxid e
H. A. Dabbagh* and J . Mohammad Salehi
Catalysis Research Laboratory, Collage of Chemistry, Isfahan University of Technology, Isfahan, Iran
Received August 28, 1997
A new transition-state model was developed in order to justify the anti intramolecular E2 elimination
with cis (Z)-preference over pure alumina and intermolecular E2 elimination with trans (E)-
preference over doped alumina. The reactions of model compounds 1,2,3-triphenyl-2-propanol (1),
1
,2-diphenyl-2-propanol (2), and 3,3,3-trideuterio-1,2-diphenyl-2-propanol (3) with aluminum oxides
with a pH range of 4.5-9.5 and thorium oxide in the temperature range of 200-350 °C in 2-hexanol
have been investigated. Over acidic alumina (pH ) 4.5 ( 0.5), the ratio of E-isomer to Z-isomer
(E/Z = 2) for 2 was found to remain unchanged in this temperature range. At 300 °C, however,
Saytzeff elimination favored Hofmann. Over pure alumina the E/ Z ratio was equal to 0.650 (2-
alkene/1-alkene ) 0.750). At equilibrium, the E/ Z ratio for 2 was equal to 4.5 with the formation
of trace amounts of Hofmann adducts. The ratio of Saytzeff to Hofmann elimination was found to
be pH independent. Any decrease in pH caused a slight increase in the E/ Z ratio. The average
primary kinetic isotope effect (k
ratio of E/ Z over thorium oxide at 300 and 350 °C was similar to that of aluminum oxide at 300
C, but the Saytzeff elimination was surprisingly favored over Hofmann! The energy of activation
H D
/k ) for elimination at 230 °C was equal to 3.775 ( 0.227. The
°
q
(E
a
H D
), entropy of activation (∆S ), selectivity, isotope effect (k /k ), and semiempirical calculation
(AM1) all agreed with concerted E2 elimination.
In tr od u ction
oxide catalysts used in different laboratories and/or from
different reaction conditions. These catalysts can vary
widely in their activity for skeletal isomerization and for
double-bond shift of olefinic hydrocarbons.
Recent interest in stereoselective synthesis of olefins1-3
via the reaction of triphenylphosphine in tetrachloro-
methane promoted a study of the reaction of model
compounds 1,2,3-triphenyl-2-propanol (1), 1,2-diphenyl-
-propanol (2), and 3,3,3-trideuterio-1,2-diphenyl-2-pro-
panol (3) with aluminum oxide and thorium oxide.
The nature of the acidity of various catalysts and
catalyst supports on the mode of elimination has been a
matter of controversy for many years. A variety of
techniques have been used to measure acidity, including
titration, measurement of quantity of chemisorbed am-
monia at various temperatures, use of dealkylation rate
of alkyl aromatics, the use of isomerization and dehydra-
tion rate, and determination of IR spectra of chemisorbed
ammonia or pyridine.11 The application of these tech-
niques has given conflicting and widely differing results
from the same materials. A decrease in surface area has
been noted with an increase in the amounts of the dopant
(sodium). The high surface area catalysts have been
reported to be more selective for dehydration, whereas
2
Metal oxides are excellent and widely used catalysts
for the dehydration of alcohols.4
-16,18-23
It has been
shown that many discrepancies can result from different
catalytic properties of the aluminum oxide and thorium
(
1) Dabbagh, H. A.; Faghihi, KH. Int. Sci. Technol. Iranica Sci.,
submitted.
(
2) Dabbagh, H. A.; Malekpoor, S.; Faghihi, KH. Iranian Polymer
J . 1998, 7, in press.
(
(
(
(
(
(
3) Dabbagh, H. A. Int. Sci. Technol. Iranica Sci., in press.
4) Blanc, E.J .; Pines, H. J . Am. Chem. Soc. 1968, 33, 2035.
5) Pines, H.; Pillai, C. N. J . Am. Chem. Soc. 1961, 83, 3270.
6) Schappell, F. G.; Pinees, H. J . Org. Chem. 1966, 31, 1735.
7) Pines, H.; Haag, W. O. J . Am. Chem. Soc. 1961, 83, 2847.
8) Kibby, C. L.; Lande, S. S.; Hall, W. K. J . Am. Chem. Soc. 1972,
2
2
low-surface catalysts were selected for dehydrogenation.
The effect of temperature on stereoselectivity and/or
regeoselectivity has been reported: an increase in the
percentage of Hofmann adducts is observed at higher
temperatures.16
9
1
4, 1.
(
(
9) Knozinger, H.; Buhl, H.; Kochloefl, K. J . Catal. 1972, 24, 57.
10) Dabbagh, H. A.; Hughes, H. G.; Davis B. H. J . Catal. 1992,
33, 445.
(
(
(
(
(
(
(
11) Parry, E. P. J . Catal. 1963, 2, 371.
The mode of elimination in the dehydration reaction
over alumina catalyst has been shown to be anti from
antiperiplanar conformation with the preference of the
cis (Z)-isomer, where there is a possibility of formation
of geometric isomers. The explanations for anti elimina-
tion and cis preference have been a matter of controversy.
A syn mode of elimination similar to pyrolysis of esters
has been reported for the dehydration of alcohols over
metal oxides.1
12) Lumneen, A. J .; Hoozer, R. V. J . Org. Chem. 1967, 32, 3386.
13) Dabbagh, H. A.; Davis, B. H. J . Org. Chem. 1990, 55, 2011.
14) Dabbagh, H. A.; Davis, B. H. J . Catal. 1988, 110, 416.
15) Dabbagh, H. A.; Davis, B. H. J . Mol. Catal. 1988, 47, 123.
16) Davis B. H. React. Kint. Catal. Lett. 1979, 11, 1.
17) Dabbagh, H. A.; Franzus, B.; Haung, T. TS.; Davis, B. H.
Tetrahedron 1991, 949, 47.
18) Noller, H.; Low, W.; Andrew, P. Ber. Bunsen-Ges. Phys. Chem.
964, 68, 663.
(
1
(
(
19) Notari, B. Chim. Ind. (Milan) 1969, 51, 1200.
20) Alunni, S.; Baciocchi, E.; Perricci, P. J . Org. Chem. 1977, 42,
0,13,16
2
05.
(
21) Chuang, T. T.; Dalla Lana, I. G. J . Chem. Soc., Faraday Trans.
Stereochemistry of elimination (anti or syn) in con-
certed homogeneous reactions depends on the properties
of the system and mainly upon those of the substrate and
1
1972, 773.
(
(
22) Siddhan, S.; Narayanan, K. J . Catal. 1979, 59, 405.
23) Sedlacek, J . J . Catal. 1979, 57, 208.
1
0.1021/jo971609u CCC: $15.00 © 1998 American Chemical Society
Published on Web 10/09/1998

7
620 J . Org. Chem., Vol. 63, No. 22, 1998
Dabbagh and Salehi
Ta ble 1. Rea ction of 2 in 2-Hexa n ol w ith Str on gly
a ,b
Acid ic Alu m in u m Oxid e (p H ) 4.5)
3
1
/T × 10 (K-1) 4/5 4 + 5/6
% 4
% 5
% 6
1
1
1
2
.7450
.8760
.9880
.1140
2.0
2.450 48 (1.80) 23 (1.40) 29 (1.50)
2.25 0.630 27 (1.40) 12 (1.10) 61.5 (1.80)
2.30 0.550 25 (1.40) 11 (1.0) 64 (1.80)
2.20 0.470 22 (1.30) 10 (1.0) 68 (1.80)
a
Values in (log % alkenes). b The relative percent was calculated
by 1H NMR.
the steric interactions of the intermediates and/or the
transition states.1 Anti elimination occurs from gauche
conformations of the molecule and syn elimination from
eclipsed conformations.1
2,14,15,20-23
The stereochemistry of elimination reactions of second-
ary and tertiary alcohols are meaningful with respect to
both regeoselectivity (Hofmann vs Saytzeff) and/or ste-
reoselectivity (anti vs syn) only when it is obtained under
the conditions where primary products are produced with
minimum secondary isomerization.
We believe we have found just the right system, which
can shed more light on the mechanism for the dehydra-
tion reactions over aluminum oxide; the system can also
help gain more insight into the reaction mechanism and
selectivity for alcohol dehydration over metal oxides. We
have undertaken a study of isotope effect, kinetics, and
semiempirical calculation (AM1)2 of molecular geometry
of the monolayer and the bilayer of a small intermediate
and the mechanism of the reactions of model compounds
F igu r e 1. Effect of temperature on alkene distribution for
the reactions of 2 over 2 g of strongly acidic alumina (pH )
3
3
4
.5 ( 0.5), flow rate 30 cm /h (8 cm /h, unfilled symbols).
5
1
,2,3-triphenyl-2-propanol (1), 1,2-diphenyl-2-propanol
(
2), and 3,3,3-trideuterio-1,2-diphenyl-2-propanol (3).
Kinetics of the dehydration reaction were investigated
in the temperature range of 200-300 °C. The effect of
temperature and pH (surface area, mesh size, flow rate,
amounts of catalyst, reactor type, and media were kept
constant) on the stereoselectivity as well as regioselec-
tivity have been investigated in detail. It is the purpose
of the present paper to provide the evidence that serves
to establish the fact that the mechanism of the reaction
of tertiary alcohols with aluminum oxide is concerted anti
E2, the cis preference is not universal, and the Saytzeff
rule depends on the steric interactions of the intermedi-
ates and/or the transition states. Results are compared
with the homogeneous reagent triphenylphosphine in
carbon tetrachloride, with cross-linked diphenylphos-
phine polystyrene in carbon tetrachloride, and with
thorium oxide.
F igu r e 2. Effect of temperature on anti/syn and Saytzeff-
Hofmann elimination for the reactions of 2 over strongly acidic
γ-alumina (pH ) 4.5 ( 0.5), flow rate 30 cm /h (8 cm /h,
unfilled symbols).
3
3
Resu lts
Spectroscopy (PMR). The rates of disappearance of
alcohol (2) and the rate of formation of intermediate were
fast and nonmeasurable (Scheme 1). We accepted the
statements made by several investigators that the alco-
Rea ction of 1,2-Dip h en yl-2-p r op a n ol (2). Alcohol
(
2) in 96% 2-hexanol was passed over acidic aluminum
3
oxide (pH ) 4.5 ( 0.5) with a flow rate of 30 cm /h at a
temperature range of 200-300 °C. Results are sum-
hols first adsorb on the catalyst surface forming an
2 3
marized in Table 1 (results with pure γ-Al O prepared
intermediate.8
,9,11,19,20,23,24
Of course, for alcohol 2 this
from aluminum isopropyloxide are presented in Table 4)
and Figures 1 and 2. The results for 2-hexanol (conver-
sion, dehydration vs dehydrogenation, and 2-hexenes/1-
hexenes ratio) for several aluminum oxides and thorium
oxide are summarized in Table 2.
Kin etics. The rates of appearance of (E)-alkene (4),
Z)-alkene (5), and 1-alkene (6) were followed at various
temperatures by proton nuclear magnetic resonance
intermediate appears to have a very short life (a very
weak Al-O-CR bond, 100% conversion for 2 vs 5%
conversion for 2-hexanol under similar conditions), con-
verting rapidly to a transition state and then in a rate
determination step to alkenes. According to the Ham-
mett postulate then, the transition state must resemble
the intermediates 2A-C. This is substantiated by the
optimized structure shown in Scheme 5. The extended
bond-making-bond-breaking is equal in the more polar
alumina backbone (transition state-like), but bond-break-
ing-bond-making is not enhanced in the alcohol 2
(
(
24) Peri, J . B. J . Phys. Chem. 1965, 69, 211, 220, 231.
(25) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . P. J .
Am. Chem. Soc. 1985, 107, 3902.

Elimination Reactions of Tertiary Alcohols
J . Org. Chem., Vol. 63, No. 22, 1998 7621
Ta ble 2. Con ver sion of 2-Hexa n ol over Alu m in u m
Ta ble 3. Cor r ected P r im a r y Isotop e Effect for th e
Con ver sion (4.7 ( 0.3) of Alcoh ols 2 (d 0) a n d 3 (d 3) w ith 2
g of Str on gly Acid ic (p H ) 4.5) Alu m in u m Oxid e a t 230
a
Oxid es a n d Th or iu m Oxid e
2
-hexenes/
hexenes/
°
C (30 m L/h )
catalysts
T (°C)
convn (%)
1-hexene
2-hexanone
%
E (4)
%Z (5)
%1-alkene (6)
Al2O3-Pb
230
230
230
230
230
200
230
350
4.5
5.0
5.0
15
45
4.30
17.0
42.0
2.70
1.30
1.40
1.70
1.80
1.50
2.0
30
15
13
f
f
f
c
kH/kDa
Al2O3-N
time (min)
d0
d3
(d0)
(d3)
(d0)
(d2)
d
Al2O3-LA
2
7
27
27
25
25
44
44
44
44
10
10
11
11
21
17
17
21
63
63
64
64
35
39
39
35
4.0
e
Al2O3-H
Al2O3-H
Al2O3-H
Al2O3-B
3.50
3.60
4.0
g
1
2
4
2
h
22
4.0
i
a
ThO2
0.10
Corrected values (see eq 1).
a
b
Flow rate is 30 mL/h unless otherwise stated. Prepared from
c
d
aluminum isopropoxide. Neutral alumina, pH ) 7. Low acidic
be utilized as a basis for obtaining the kinetic isotope
effect. The isotope effects are corrected for 15% protium
using eq 1.
e
f
alumina, pH ) 6. Acidic alumina, pH 4.5. Trace amount of
g
h
2
)
-hexanone was formed. Flow rate 8 mL/h. Basic alumina, pH
i
9.5. Flow rate 4 mL/h.
Ca lcu la tion of Kin etic Isotop e Effect. The cor-
rected kinetic isotope effects for the elimination step were
calculated using the relative rate of formation of (R-
Sch em e 1
benzylstyrenes (Table 3) using eq 1, where k′
E Z
, k′ , and
k′R are the relative rates of formation of (E)-isomer (4),
(
Z)-isomer (5), and 1-alkene (6), respectively.
k′R
(
)
k′ + k′
E
Z
H
[
k /k ]
)
H
D elim
k′R - 0.15k′R
(
)
k′ + 0.15k′R(B) + k′ + 0.15k′R(B′)
E
Z
D
(1)
k′E
k′ + k′
k′Z
k′ + k′
Z
B )
B′ )
(
)
(
)
E
Z
E
Effect of p H. The effect of acidity and/or basicity of
catalysts on stereoselectivity (E/ Z) and regioselectivity
Hofmann-Sytzeff) was tested on various aluminum
(
oxides with a wide range of pH values (4.5-9.5). Gener-
ally, the pH values refer to aqueous solution (in dehydra-
tion of alcohols water molecules are produced at reaction
sites!); however, the relative acidity-basicity and the
amounts of alkali that alumina contains (when compared
with pure alumina) is what we are concerned with but
not the absolute pH values. To minimize the deviation
backbone (intermediate-like). Activation energies for the
alkene formation step were calculated by the slope of the
plot of log % alkenes versus reciprocal of temperatures
2
in product distribution, the surface area (150 m /g), the
3
(Table 1, Figure 4, a typical plot). The entropies of
mesh size (150, 58 Å), the flow rate (30 cm /h), the media
(4% 2 in 2-hexanol), the reactor type, and the amounts
activation were calculated using the intercept of the
of catalyst (2.0 g fresh) were kept constant (Figure 3).
Rea ction of 2 w ith P u r e γ-Alu m in u m Oxid e.
Results of the reaction of 2 via pure γ-aluminum oxide
Arrhenius plot (log A) using the following expressions:
∆
Sq/R
log ) -E
a
/RT + log A, A ) (κkT/h)e
. Data points at
3
00 °C were ignored in the calculation of the energy of
(prepared from aluminum isopropoxide followed by cal-
activation due to large divination from linearity. Data
points at 200-260 °C produced slopes with good correla-
tion coefficients (r) (Table 8, Figure 4).
cination at 600 °C for 6 h) are reported in Table 4. To
lower the conversion of alcohols, 0.50 g of pure alumina
was diluted with 1.50 g of powderd glass beads. Results
of the reaction of 2-hexanol with pure γ-aluminum oxide
is presented in Table 2.
Rea ction of 2 w ith Th or iu m Oxid e. Results of the
reaction of 2 with thorium oxide (prepared from thorium
nitrate through calcination of thorium nitrate at 600 °C
for 6 h) are reported in Table 5. Results of the reaction
of 2-hexanol with thorium oxide is presented in Table 2.
Rea ction of 1,2,3-Tr ip h en yl-2-p r op a n ol (1). Alco-
hol (1) in methyl acetate when passed over strongly acidic
alumina (pH ) 4.5 ( 0.5) at 200 °C under similar
conditions as for 2 converted 100% to a mixture of 48%
(E)-1,2,3-triphenyl-2-propene (1E) and 52% (Z)-1,2,3-
triphenyl-2-propene (1Z).
Isotop e Effects in th e Elim in a tion Rea ction s of
1
,2-Dip h en yl-2-p r op a n ol in 2-Hexa n ol. The kinetic
isotope effect was run on a 4% solution of 3,3,3-trideu-
terio-1,2-diphenyl-2-propanol (3) in 2-hexanol on 2 g of
fresh strongly acidic γ-Al O (pH ) 4.4 ( 0.5) at 230 °C.
2 3
The relative percent of alkenes were then compared with
those obtained from alcohol 2 in a separate run but
exactly under similar conditions using 2 g of fresh
catalyst, Table 3. Under these conditions, for both
reactions the conversion of 2-hexanol is 4.7 ( 0.3%. In
other words, when the conversion of these mixtures is
effected under similar conditions, the relative rate of
formation of 1-alkene-d
0
(k′R)
H
to 1-alkene-d
2
D
(k′R) can

7
622 J . Org. Chem., Vol. 63, No. 22, 1998
Dabbagh and Salehi
Ta ble 4. Con ver sion of 2 to Olefin s via P u r e
a ,b
γ-Alu m in u m Oxid e a t 230 °C
run no.
time (min)
% 4
% 5
% 6
4/5
4 + 5/6
1
2
3
c
4
8
12
80
120
20
15.50
17
76
80.5
25
29
26
24
55
55.50
57
0
1
0.80
0.50
0.60
3.20
4.30
0.80
0.80
0.750
d
18.5
99
a
Aluminum oxide was prepared from aluminum isopropoxide
b
(
(
0.5 g Al2O3 + 1.5 g of powder glass beads). Average conversion
93%). c 2.0 g of catalyst was used. d Equilibrium composition.
Ta ble 5. Con ver sion of 2 to Olefin s via Th or iu m Oxid e
a t 350 °C
run no. % convn time (min) % 4 % 5 % 6 4/5 4 + 5/6
1
2
3
a
80
78
86
0
4
8
12
30
25
55
49
58
26
21
30
19 2.10
30 2.30
14 1.90
4.30
2.30
6.30
b
30
47
21
32 2.20
2.10
a
Reaction temperature at 230 °C. b Reaction temperature at 300
F igu r e 3. Effect of pH on anti/syn elimination (E/ Z) and on
Saytzeff-Hofmann (2-alkenes/1-alkene) elimination for the
reactions of 2 over aluminum oxides at 230 °C.
°
C.
Ta ble 6. Com p a r ison of th e Deh yd r a tion Rea ction of
2
-Bu tyl System
intermediates
2-B/1-Ba
cis/trans
ref
+
- b
[
[
[
[
[
[
[
P-OR] X
2.570
1.270
3.760
0.060
8.10
4.30
3.0
4
4
3
4
3
1
4
+
- c
I-OR] X
+
- d
H2O-OR] TSO
1.70
1.45
0.70
0.50
smalle
+
-
(CH3)3N-OR] HO
+
-
H2O-OR] SO4
+
-
Ph3P-OR] Cl
5.30
0.350
+
-
(CH3)2SOR] C2H5O
a
Ratio of 2-butenes/1-butene. b P ) pure alumin. c I ) sodium-
d
doped alumina. Thermodynamic equilibrium, TSO ) C7H7SO3.
Mostly trans.
e
Ta ble 7. Com p a r ison of P r od u ct Distr ibu tion of
Deh yd r a tion Rea ction of (2)
reagent
T (°C)
%4
%5
%6
E/ Z 4 + 5/6
Ph2P-(P)a
80
78
70
230
350
300
230
230
=100
trace trace VLb
VL
1.10
Ph3P-CCl4 c
TSOH-C6H6
51.5 <1
48.5
0
large
4.20
2.30
2.30
2.20
1.20
0.65
d
81
25
49
47
19
17
19
11
21
21
16
26
e
γ-Al2O3-H
64
30
32
56
57
0.560
2.30
2.10
0.630
0.750
f
ThO2
ThO2g
h
γ-Al2O3-B
i
F igu r e 4. Typical Arrhenius plot for the rate of appearance
γ-Al2O3
3
3
of 4 with strongly acidic γ-Al
2
O
3
, flow rate 30 cm /h (8 cm /h,
a
b
Cross-linked diphenylphosphine polystyrene (see ref 2). VL
unfilled symbols).
very large. c See ref 1. d Equilibrium composition (see ref 3).
)
e
f
g
Strongly acidic alumina (pH ) 4.5). Conversion 78%. Conver-
Discu ssion
sion 30%. h Basic alumina (pH ) 9.5). i Pure alumina prepared
from aluminum isopropoxide.
Both homogeneous and heterogeneous dehydration of
secondary and tertiary alcohols are among the most
controlled R-benzylstyrene (6) approximately in equal
amounts (1:1) with trans-R-methyl stilbene (4) in 0.4 h,
rapidly converting to an equilibrium mixture of olefins
in 2 h. They reported that the reaction of 2 with cross-
linked diphenylphosphine polystyrene in carbon tetra-
chloride mainly produced trans-R-methylstilbene 4 with
trace amounts of 5 and 6 (Table 7).2 Steric interaction
at the transition state or intermediate is shown to be
studied of all major organic processes.1
-23
There have
been a number of attempts to relate the mechanisms of
these two systems; many discrepancies, however, have
resulted from the different reaction conditions and
reagents used (Tables 6 and 7).
Hom ogen eou s System s. Dabbagh et al.1,3 recently
reported that triphenylphosphine in carbon tetrachloride
converted 100% of 2 to a mixture of 28% 1,2-diphenyl-
1
7
responsible for the observed selectivity. Dabbagh et al.
2
-chloropropane, 37% 4, 35% 6, and a trace of 5 at 78
have reported that elimination is the minor pathway for
the reaction of erythro-3-deuterio-2-butanol with tri-
phenylphosphine (>98% anti elimination in favor of
Saytzeff orientation with a primary deuterium isotope
effect of 2) (Table 6); 2-methyl-2-propanol produced
°
C. They demonstrated that stereoselectivity and regio-
selectivity of elimination in concerted homogeneous reac-
tions depend on the properties of the system and mainly
on the steric interaction of the intermediate and/or
transition state. It was also reported3 that, under
homogeneous acidic conditions, 2 produced kinetically
H D
almost 90% 2-methylpropene (k /k ) 2).

Elimination Reactions of Tertiary Alcohols
J . Org. Chem., Vol. 63, No. 22, 1998 7623
Heter ogen eou s System s. Pine and co-workers4-7
studied the dehydration of alcohols over aluminum oxide
extensively and provided evidence that initial products
arise from concerted trans elimination. Participation of
two catalyst surfaces (crevices or pores of the solid) was
isomerization preventer. Fifth, we ran the dehydration
on thorium oxide at several temperatures. Finally, we
optimized the geometry of adsorbed 2 and 2-butanol on
2
5
alumina via the semiempirical calculation (AM1).
The major factors contributing to the inconsistencies
of the dehydraton mechanism of alcohols over alumina
are variation in surface area, mesh size, temperature,
acidity-basicity, flow rate, amounts of catalyst, media,
and reactor type. In our investigation, the effects of two
major contributors, i.e., temperature or acidity-basicity,
on product distribution are sought; other variables are
kept constant.
6
proposed to achieve the required steric arrangement. Of
course, if the dehydration proceeds only in crevices of
molecular size, then the reaction must be diffusion-
controlled, a phenomenon that has never been observed.
7
Pine and Haag explained the cis preference by as-
suming an intermediate proton-olefin complex, since the
cis π-complex is more stable than the trans π-complex.
Noller and co-workers12 and Notari explained the cis
preference in the E1-like mechanism on the grounds of
steric considerations.
13
Effect of Tem p er a tu r e. The product distribution (%
olefins) shows a small variation at the temperature range
of 200-260 °C (a kinetically controlled process) with a
sudden change at 300 °C (Figure 1). The ratio of (E)-
alkene/(Z)-alkene (stereoselectivity) shows a small change
at higher temperatures; however, the regioselectivity,
e.g., the ratio of 2-alkene/1-alkene, increased drastically
Hall et al.8 rejected the proposed model that anti
elimination must be due to a reaction between surfaces
in a pore or a crevice. They suggested that large O2 ions
adjacent to the Lewis acid site alcolate are responsible
for the abstraction of the â-hydrogen.
-
3
at 300 °C to a flow rate of 8 cm /h (approaching the
9
Knozinger and co-workers proposed a new transition-
thermodynamically controlled process). Two factors may
have contributed to this effect: the weakening of the
CR-O bond of the adsorbed alcohol on the surface at
higher temperatures or secondary isomerization of 1-
alkene to 2-alkenes. To test the latter effect, pure 4
passed over acidic alumina at 230 °C. No isomerization
was observed. All attempts to isolate pure 5 or 6 failed.
However, no isomerization was observed when a mixture
of mostly 1-alkene 6 (68% and 32% of 4 and 5) was
allowed to pass over fresh acidic alumina at 230 or 250
°C under conditions similar to those in each of the
previous runs. At 300 °C, 6 (68%) isomerized to near-
equilibrium composition (4 + 5/6 ) 7.3, with the ratio of
4/5 ) 2.60). To prevent isomerization of olefins, conver-
sion of 2-hexanol was maintained in the range of 1-15%
at the temperature range of 200-300 °C. Alcohol 2,
however, converted 100% at all temperature and pH
values (unless stated otherwise).
state model that explains the anti mode of elimination
with the cis preference. They assumed a localized
adsorption of the reacting molecule on the surface and
further that the steric restrictions in the transition state
are most likely to determine the cis/trans ratio of the
respective isomeric olefins.
A possible alternative is suggested by the Peri model
of the alumina surface.24 Here the dual acid-base sites
are aluminum ions, exposed where the terminal OH
groups are missing, and large O2 ions replacing the OH
groups and covering adjacent lattice positions. If it is
assumed that an alcoholic hydroxyl group covers the
aluminum ion with the alcoholic hydrogen pointing away
from the surface, then hydrogen atoms on the â carbon
atom in the trans position to the hydroxyl group can be
discharged on top of the adjacent O2 ions more readily
than those in the cis positions.
-
-
Thoria-catalyzed dehydration of 2-butanol is shown to
be similar to the pyrolysis of sec-butyl acetate and of
trimethyl-sec-butylammonium hydroxide. The mode of
elimination is stereospecific syn elimination in favor of
Hofmann orientation. The dehydration of tertiary-2-
alkanols over thoria is not stereoselevtive.12,13
Effect of p H. The second major factor that has been
reported to have pronounced effect on the mechanism of
the dehydration of alcohols is the acidity-basicity of the
catalyst. The major problem in the investigation on this
effect has been the inconsistency in the preparation of
the catalyst.
The question raised here was whether aluminum oxide
with tertiary alcohols could be used in a stereoselective
synthesis of hindered olefins. If so, what is then the
mechanism of elimination for these reactions? What is
the effect of temperature and pH on the stereoselectivity
and regioselectivity? How comparable are the dehydra-
tions of 2 in a heterogeneous system with that of
homogeneous triphenylphosphine in carbon tetrachloride
and with that of semiheterogeneous cross-linked di-
phenylphosphine polystyrene in carbon tetrachloride?
To answer these questions, we first studied the effect
of varying the pH and temperature on reactivity, stereo-
selectivity (e.g., the ratio of syn/anti elimination), and
regioselectivity (e.g., the ratio of Saytzeff/Hofmann elimi-
nation). Second, we calculated the kinetics of elimination
The effect of acid-base strength upon the structure of
transition states in elimination reactions has not been
convincingly explained in the literature. Aluni et al.
2
0
reported that with an increase in base strength there was
a significant decrease in the CR-leaving group bond and
the transition state subsequently leaned toward E1cB.
A decrease in BET surface area was noticed with an
increase in the amount of dopant; high surface-area
catalysts were more selective for dehydration, but low
surface-area catalysts were selective for dehydrogena-
tion.21 Siddhan and Narayanan studied the change in
the acidity-basicity balance in the sodium-impregnated
alumina samples vis- a` -vis pure alumina. In our inves-
tigation, we purchased several catalysts with the same
22
2
surface area (150 m /g) and mesh size (150, 58 Å) with a
reaction [e.g. the energy of activation (E
a
) and the entropy
wide range of pH values (4.5-9.5) from Aldrich Chemical
Co. The reactivity and selectivity of these catalysts were
compared with that of prepared catalysts. Within our
experimental error at 230 °C, the ratio of 2-alkene/1-
alkene (regioselectivity) remained unchanged in this pH
range. A slight increase for the ratio of (E)-alkene/(Z)-
alkene (stereoselectivity) was observed at lower pH
q
of activation (∆S )]. Then, we studied the elimination
reaction of 3,3,3-trideutero-1,2-diphenyl-2-propanol (3)
and obtained a primary kinetic isotope effect. In the next
step, we compared the reactivity and selectivity of pure
alumina with those of doped alumina (surface areas of
2
1
50 m /g) with 2-hexanol both as a solvent and as

7
624 J . Org. Chem., Vol. 63, No. 22, 1998
Dabbagh and Salehi
On the basis of our new definition of stereoselectivity1
i.e., a reaction should be considered stereoselective if the
Ta ble 8. En er getics (ca l/m ol) for th e Ra te of
Ap p ea r a n ce of Alk en es over γ-Alu m in a a t 503 Ka
(
q
comp no.
Ea (cal/mol)
rb
Ac
∆S (cal/mol)
product distribution is far from that of the thermody-
namic equilibrium composition), the reactions of doped
aluminum oxides (pH ) 4.5-9.5) are less stereoselective
than pure alumina. The ratio of 4/5 is equal to 2.2
[approaches that of equilibrium composition (4/5 ) 4.3
at 70 °C)] and 0.60, respectively. A cis preference (1E/
(
(
E)-alkene (4)
Z)-alkene (5)
-alkene (6)
1.720
1.530
0.760
0.9930
0.9990
0.9990
137. 0
51.0
30.0
-53.80
-55.70
-56.90
1
a
pH ) 4.5. b Correlation coefficient. c Arrhenius parameter.
1
Z ) 0.920) was observed for the dehydration of 1 over
values (Figure 3). Addition of dopant to the alumina
surface in order to change the acidity-basicity nature
of the catalyst would occupy certain reaction sites but
not all. The reactivity is decreased, but enough free
acidic sites and basic sites are available for adsorbing
the alcohols and abstraction of â-hydrogen(s), respec-
tively. Dopant also make the catalyst more rigid (in-
creases the ionic nature of the catalyst), less flexible, and
less selective. Pure alumina is less rigid (less ionic in
nature), more flexible and more selective, Table 7.
strongly acidic aluminum oxide (pH ) 4.5). This ratio
was equal to 5.7 for reactions with triphenylphosphine
in carbon tetrachloride and very large for cross-linked
1
diphenylphospine polystyrene in carbon tetrachloride
(
Table 7).2 Reactivity of doped alumina is much lower
than that of pure alumina. To obtain the same percent-
age of 2-hexanol conversion, 0.1 g of pure alumina and
2
.0 g of doped alumina were used.
Another interesting observation is that the reaction of
-butanol over aluminum oxide favored the Saytzeff
2
Kin etics. The mechanism is consistent with the
elimination, but in the case of 2 Hofmann elimination,
it showed unusual predominance over the Saytzeff (reac-
tion of 2 with triphenylphosphine produced equal amounts
of Saytzeff-Hofmann adducts). This substantiates our
energetics of the reaction, Table 8. The entropy of
q
activation (∆S ) is appreciably low. The two reacting
centers, e.g., the basic site on alumina and the â-hydro-
gens, must attain a specific geometry to permit the
bonding intermediates that occur as the transition state
1
earlier prediction that the Hofmann-Saytzeff ratio does
not depend on the reagent nor on the product stability,
but instead, it depends on the steric interactions in the
transition state and/or intermediate. This is demon-
strated in Scheme 5. â-Hydrogens producing 1-alkene
or (Z)-alkene and Base-4 are molecularly constrained as
close neighbors and energetically favorable when they are
compared to the â-hydrogen producing the most stable
q
is approached. The entropy of activation (∆S ) for the
formation step of all three alkenes ((E)-alkene, (Z)-alkene,
and 1-alkene) is similarly small and negative. The
Arrhenius parameter is directly proportional to number
of collisions. Collisions of reaction sites with specific
geometry decreases the entropy of activation (more
negative) and subsequently decrease the Arrhenius pa-
rameter (Table 8). This substantiates the fact that the
reaction of basic center (Base-4, Scheme 5) with â-hy-
drogen (s) are molecularly unconstrained. The low
activation energies for dehydration of 2 are due to
weakness of the bond between Al-O-CR. This is
substantiated further by a very high conversion (100%)
and also in Scheme 5 (extend of bond-breaking bond-
making is enhanced). For 2-hexanol the Al-O-CR is
much stronger (10-15% conversion at the same condi-
tions), Scheme 6 (extend of bond-breaking bond-making
is not enhanced).The less hindered alcohols show much
higher activation energy. Therefore, the dehydration
reaction of 2 over pure γ-alumina is a kinetically con-
trolled process (leaning toward the thermodynamically
controlled for doped alumina) with low activation energy
for the intermediate decomposition step (the rate-
determining step). The less stable alkenes (1-alkene and
(E)-alkene.
The next step is to compare the dehydration reactions
of 2-butanol (Table 6) and 2-hexanol (Table 2) with that
of 2 (Table 7) in both homogeneous and heterogeneous
systems. The regeoselectivity (2-alkenes/1-alkene) for
2
-hexanol (2.60, this work) and 2-butanol (2.70, ref 4) are
similar over pure alumina. Over doped alumina the
values of 1.270-2.0 (2.0 for basic alumina, pK ) 9.5)
are obtained (this work and ref 4). This value for
-hexanol is equal to 0.10 over thorium oxide and
a
2
matches the value of 0.060 for 2-butanol (basic homoge-
neous Hofmann elimination). The data demonstrate that
by increasing the basicity of alumina the regeoselectivity
does not approach that of thorium oxide. This is further
shown for the elimination reaction of 2 over pure alu-
mina, doped alumina, and thorium oxide. In fact, the
regeoselectivity for acidic alumina and for that of basic
alumina oxide is similar within the experimental error
(Z)-alkene) are formed in favor of the more stable (E)-
(Figure 3). The ratio of 4/5 is decreased (lowest for pure
alkene at low temperatures and then equilibrate to the
thermodynamic equilibrium composition at higher tem-
peratures.
alumina) when the amount of dopant (basicity) is in-
creased. This ratio for acidic alumina (4/5 ) 2.3) ap-
proaches that of equilibrium composition (4/5 ) 4.20).
Geom etr y. Geometric conditions for the dehydration
of alcohols on γ-alumina and thorium oxide have been
studied by Sedlacek23 with the aid of a computer program.
He has concluded that the reaction may proceed on a
catalyst surface not only as the syn elimination but also
as the energetically more favorable anti elimination.
In the construction of the models used in this work,
we used general geometric conditions suggested by Sed-
lacek.23 (a) The molecule of alcohol adsorbed via its
oxygen atom on a surface by a passive hydrogen bond
(Scheme 2) at the electrophilic site on the surface, i.e.,
the coordinatively unsaturated Al3 ion. Adsorption of
alcohol follows quite quickly to the dissociation of the
The position of the transition state along the reaction
coordinate can be estimated by the magnitude of kinetic
isotope effect. Reactions that have linear symmetrical
transition states with hydrogen positioned centrally
between the two acceptor centers are shown to have a
maximum isotope effect. The observed averaged deu-
terium isotope effect (Table 3) equal to 3.775 ( 0.227
indicates that the rate-determining step is the proton loss
in a concerted E2-like but not synchronous transition
state (Scheme 5). The extent of bond-making bond-
breaking is not equal in the dehydration reaction of
+
2
-butanol, a fact that is revealed in the low k
H D
/k ( 2.0
and in Scheme 6.

Elimination Reactions of Tertiary Alcohols
J . Org. Chem., Vol. 63, No. 22, 1998 7625
Sch em e 2. Six-Mem ber ed Rin g Hyd r ogen
Bon d in g of 2 w ith Neu tr a l Alu m in a
Sch em e 4
Sch em e 3
alcohol molecule, yielding the surface alcolate and a
hydroxyl group (Scheme 3). According to quantum
chemical calculations, the adsorption by passive hydrogen
bond increases the ability of the γ-hydrogen to be
attacked by nucleophilic species. (b) The alcolate inter-
acts in the adsorption complex through its γ-hydrogen
with a surface oxygen ion. (c) The O-CR and Câ-H
bonds in the transition state are mutually anti and syn
periplanar, which determines concerted anti and syn
eliminatioms, respectively. According to quantum chemi-
cal calculations, the activation energy is lowest for these
conformations.
Sch em e 5
Generally speaking, the transition-state model has to
take into account the properties of the reactant molecule
as well as those of the products besides the very specific
geometry requirements brought about by the transition
state developing on the surface of the solid. The rate of
syn elimination is strongly dependent upon the dihedral
angle formed between the tetrahedral centered on the
carbon atoms R and â. Thus, it would appear that
structural considerations within the alcohol molecules
control the course of the reaction. The fixed geometry of
the surface places constraints on the transition state;
however, such constaints are absent in homogeneous
media. In the latter, the acid-base can approach the
substrate from all directions, and the lower energy
pathway leads to complete stereoselectivity. In the
heterogeneous reaction, compromises have to be made
and complete stereoselectivity cannot be expected. J ust
as anti elimination is expected to be the rule in homo-
geneous media (in the absence of special circumstances),
geometric constraints are expected to require syn elimi-
nation for the surface-catalyzed reaction.
reaction of 2 and 2-butanol over aluminum oxide. The
structure of adsorbed alcohol (intermediate) on a small
monolayer pure alumina and bilayer doped alumina was
optimized by semiempirical calculation (AM1), Schemes
5
-7.
These models clearly demonstrate that the partial
dehydration on the surface of alumina leads to a situation
We prepared pure alumina by the method presented
in Scheme 4. The structure of pure γ-alumina can be
depicted as a polymerization process. The major differ-
ence is that bonds in aluminum oxide are ionic in
character.
8
2-
where, as proposed by Hall and co-workers, large O
ions and Lewis acid sites cover adjacent lattice positions.
The alcohol 2 or 2-butanol is adsorbed through the
alcoholic oxygen atom at Lewis acid site and the basic
2-
center, an O ion, does the dehydration process. The
Dabbbagh and co-workers recently1 optimized the
geometry of intermediates for the reaction of 2 with
triphenylphosphine in carbon tetrachloride by semi-
empirical calculations (AM1, CNDO, and MNDO)25 and
showed that reaction proceeds via intermolecular anti-
E2 elimination and substitution (Schemes 4 and 5, ref
anti â-hydrogen atoms and the R-methyl hydrogens are
more easily discharged, hence the observed stereospeci-
ficity, regioselectivity, and intramolecular anti-E2 elimi-
nation with cis preference. Thus, two different centers
function and the expanded system does not utilize some
of the gauche configurations that result in a loss in
stereoselectivity. The adsorbed structure 2 with two
1
). We applied the same method to the dehydration

7
626 J . Org. Chem., Vol. 63, No. 22, 1998
Dabbagh and Salehi
Sch em e 6
surface (strong attraction between cations and anions)
has a less flexible surface. In doped alumina, bonds
between Al3 and O are much stronger, O is a weaker
+
2-
2-
1
-
base, and therefore, the O of hydroxyl group is the basic
site of â-elimination. There is less steric restriction
between the adsorbed alcohol and the surface. Therefore,
the more stable anti configuration with trans preference
is formed. The trans-intermolecular E2 elimination
(higher activation energy with lower entropy of activa-
tion) with Hofmann preference is favored over cis-
intramolecular E2 elimination (Scheme 7). This explains
the lower observed reactivity of doped alumina over pure
alumina. A similar observation is reported for cross-
linked polystyryl diphenyl phosphine.2
,28,29
Stronge evi-
Sch em e 7
dence is presented that the polymer-bound Wilkinson’s
catalyst can aggregate to form binuclear cluster. The 2%
cross-linked are mobile enough to allow the interaction
of two active sites. The 20% cross-linked are less mobile
(more rigid), and the interaction of the active sites is
substantially reduced.
The most intriguing fact is that the new models exactly
complement the experimental observations!
Con clu sion
Studies of alcohols with proper â-substituents revealed
that the anti elimination with cis preference (justified
by optimized geometry calculated through simiempirical
AM1 method) is not universal in all catalytic eliminations
but that it, in fact, depends on the mode of elimination,
temperature, and the catalyst. While anti elimination
with syn preference was noticed in pure alumina, a trans
olefin was formed to a major amount in doped alumina
and thoria at 300 and 350 °C. Kinetic isotope effect and
product distribution both indicate that at lower temper-
atures reaction in pure aluminum oxides proceeds via
intermolecular E2 elimination and favors the Hofmann
orientation with much lower selectivity than in homoge-
neous nonacidic dehydrations. At higher temperatures,
product distribution is similar to that of equilibrium
composition, which indicates an ionic intermediate (e.g.,
E1). Therefore, we conclude that the dehydration of
alcohols over metal oxides depends on not only the steric
interaction of the intermediate and/or transition state but
is also strongly dependent on the preparation conditions
of the catalyst and reaction conditions.
phenyl groups facing away from the surface (in a cis
conformation) forces the R-methyl group in a position
close to the Base-4 site where elimination leads to the
less stable Hofmann olefin. However, there is a strong
van der Waals’ attraction between the â-phenyl of 2 and
the terminal hydroxyl group of alumina, which stabilizes
the cis configuration (Scheme 5). In the case of 2-butanol,
the two methyl groups (there is a strong van der Waals’
repulsion between the R-methyl and terminal hydroxyl
groups) are forced away from the surface (in a cis
conformation), pushing the â-hydrogen near the Base-4
site, which leads to the less stable cis-2-butene and
1
-butene, respectively. The â-hydrogen forming the most
stable t-2-butene is in unfavorable position with respect
to the Base-4 site, Scheme 6.
The new models, however, disagree with the proposal
made by Knozinger and co-workers that the adsorbed
Exp er im en ta l Section
9
Gen er a l. The catalytic reactions were performed in a
vertical plug flow reactor made of Pyrex glass and fitted with
thermal well extended to the center of the catalyst bed. About
structure with cis conformation may easily incline back-
wards toward the surface, whereas a structure with trans
conformation would bring about a high steric restriction
to inclination toward each side. In fact, it is the combi-
nation of steric interactions between the surface and the
adsorbed species that determine the configuration of the
adsorbed alcohol which, in turn, leads to an observed
selectivity. An interesting observation is that the struc-
ture depicted in Scheme 5 resembles half intermediate
0-30 cm3 of the reactor volume above the catalyst bed
2
contained Pyrex glass beads to serve as preheater. Liquid
products were collected at room temperature with increasing
time intervals. 2-Hexanols (96%) were used as solvent with
4
% 2 to prevent alkene isomerization and to increase the vapor
pressure of the alcohol 1-3 and alkenes. It has been shown
that 2-octanol is adsorbed strongly enough to essentially
eliminate the alkene isomerization reaction at conversion
(
alcohol 2 backbone) and half transition state (aluminum
26,27
levels of about 50%.
Conversion of 2-hexanol was moni-
oxide backbone). There are an extended bond breaking
for Al-O-CR and the extension of the Al-Base 4 oxygen
for abstraction of â-hydrogen (s). No extended bond-
making bond-breaking is observed for the alcohol back-
bone. The more ionic bonds in alumina are more fexible
than alcohol 2 covalent bonds. A bilayer model for doped
alumina with more ionic character and, therefore, a rigid
tored by Shimadzu GC-14A gas chromatography using an SE-
(
26) Davis, B. H. J . Org. Chem. 1972, 37, 1240.
(27) Davis, B. H. J . Catal. 1979, 58, 493.
28) Reed, J .; Eisenberg, P.; Teo, B. K.; Kincard, B. M. J . Am. Chem.
Soc. 1977, 99, 5217.
29) Reed, J .; Eisenberg, P.; Teo, B. K.; Kincard, B. M. J . Am. Chem.
Soc. 1978, 100, 2375.
(
(

Elimination Reactions of Tertiary Alcohols
J . Org. Chem., Vol. 63, No. 22, 1998 7627
Sch em e 8
Meth od of Ca lcu la tion s. All the present semiempirical
calculations were carried out with complete geometry optimi-
zation unless otherwise stated. The AM125 calculations were
performed with the HyperChem (V.3) program.
1
,2,3-Tr ip h en yl-2-p r op a n ol (1).3 The reaction of ethyl
benzoate (84.94 g, 0.566 mol), benzyl chloride (130 g, 1.24 mol),
and magnesium 30.3 g, 1.24 mol) in diethyl ether produced a
viscous oil. When crystallized in n-hexane, this oil produced
1
7
3
3
0 g of 2: mp 83.5-85 °C; H NMR (CCl
4
) δ ppm 1.65 (s, 1H),
.15 (AB-quartet, 4H J ) 15 Hz), 6.90-7.45 (m, 15H); IR (KBr)
300-3400 br.
1
,2-Dip h en yl-2-p r op a n ol (2).3 The reaction of aceto-
phenone (10 g, 0.0833 mol), benzyl chloride (10.5 g, 0.0916
mol), and magnesium (2.23 g, 0.0916 mol) in diethyl ether
produced a viscous oil. When crystallized in n-hexane, this
1
oil produced 9.0 g of 2: mp 44-46.5 °C; H NMR (CCl
4
) δ ppm
1
6
.5 (s, 3H), 1.55 (s, 1H), 2.95 (AB-quartet, 2H, J ) 15 Hz),
.80-7.40 (m, 10H); IR (KBr) 3300-3400.
r-Tr id eu ter ioa cetop h en on e. Tetradeuterioacetic acid
(
(
11.10 g, 0.1734 mmol) reacted with phosphorus pentachloride
4.33 g, 0.021 mmol) to give 10.60 g (75% yield) of trideute-
rioacetyl chloride, bp 44-50 °C. Trideuterioacetyl chloride
(8.70 mL, 0.1215 mmol) reacted with 40 mL (0.45 mol) of
benzene in the presence of 40 g (0.30 mol) of freshly sublimed
aluminum chloride to give 7.70 g (52% yield) of R-trideuterio-
acetophenone, mp 202 °C.
3
,3,3-Tr id eu ter io-1,2-d ip h en yl-2-p r op a n ol (3). The re-
action of R-trideuterioacetophenone (0.4872 g, 0.00406 mol)
with 20% excess benzyl chloride in magnesium (1.170 g,
0.0.04875 mol) in diethyl ether produced a viscous oil. When
3
0 column, and then the excess 2-hexanol was removed under
reduced pressure. Conversion of alcohols 1-3 and the relative
product percent were measured by H NMR spectra (Scheme
8
9
1
crystallized in n-hexane, this oil produced 6.11.0 g of 3: mp
1
1
-3
47-50 °C; H NMR (CCl ) δ ppm 1.50 (s, 3D), 2.0 (s, 1H), 2.95
), Varian, EM390 (90 MHz).
Catalysts (with pH ) 4.5-
4
2
(AB-quartet, 2H, J ) 15 Hz), 6.80-7.40 (m, 10H); IR (KBr)
.5, surface area ) 150 m /g, mesh size ) 150, 58 Å) and
1
3
300-3400. Analysis by H NMR showed 85% deuterium and
aluminum isopropoxide were obtained from Aldrich Chemical
2
22
15% protium in the methyl group. Correction for kinetic
isotope effect for 15% protium were made using eq 1.
Co. Pure alumina (BET surface area 150 m /g) was prepared
by hydrolysis of aluminum isopropoxide followed by calcination
at 600 °C for 6 h. Thorium oxide was prepared through
calcination of thorium nitrate at 600 °C for 6 h. All runs
were performed over 2 g of fresh catalysts (to maintain low
2
2
(1)
2 3
conversion 0.50 g of Al O prepared from aluminum isopro-
pyloxide and 1.5 g of powder glass beeds were used) and
pretreated in situ in air for 2 h at 400 °C and then with dry
nitrogen for 30 min. IR spectra were obtained by Shimadzu
ZU-435. Melting points were taken by the Gallenkemp
melting point apparatus and are uncorrected. Yields were
Ack n ow led gm en t. This work was supported by
Isfahan University of Technology Research Council
Grant IUT-ICHE762.
1
calculated from H NMR and/or isolated products are not
optimized. All starting material or solvents were obtained
from Merck or Fluka Chemicals companies and were purified
with proper purification techniques.
J O971609U