Ammoximation of cyclic ketones on TS-1 and amorphous SiO2-TiO2

Article abstract of DOI:10.1006/jcat.1996.0219

Ammoximation of cyclic ketones on TS-1 and amorphous SiO₂-TiO₂ with H₂O₂ and NH₃ was studied. The reactivity of cyclic ketones over TS-1 and amorphous SiO₂-TiO₂ showed the following order: cycloheptanone > cyclopentanone > cyclohexanone > cyclooctanone. Shape-selective poisoning studies revealed that the active sites of TS-1 for ammoximation of cycloheptanone are mostly inside the molecular sieve. The reactivity of cyclohexanone and methylcyclohexanones over TS-1 followed the order cyclohexanone > 2-methylcyclohexanone 3-methylcyclohexanone > 2,6-dimethylcyclohexanone, reflecting the difference in the diffusivity among the reactants or products inside TS-1. Besides the diffusivity the reactivity seemed to be affected by the difficulty in access of the carbonyl group to the Ti active site inside the zeolite; from the comparison of reactivity of dimethycyclohexanone isomers, the steric hindrance of the substituent methyl group to the access of carbonyl group was considered to decrease in the following order: β-equatorial > α-equatorial > β-axial > α-axial. Diffusion of cyclohexanone or its oxime has been found to play an important role in the ammoximation when the TS-1 crystals larger than 1.5 μm were employed as a catalyst. For the TS-1 crystals smaller than 0.3 μm, the ammoximation activity was independent of the diffusivity factor and was determined by the number of active Ti sites present in the zeolite framework.

Full text of DOI:10.1006/jcat.1996.0219

JOURNAL OF CATALYSIS 161, 570–576 (1996)
ARTICLE NO. 0219
Ammoximation of Cyclic Ketones on TS-1 and Amorphous SiO2–TiO2
Takashi Tatsumi1,2 and Nizamidin Jappar1
Department of Applied Chemistry, Faculty of Engineering, University of Tokyo, Hongo, Tokyo 113, Japan
Received March 27, 1995; revised March 7, 1996; accepted March 25, 1996
molecular sieves, TS-1 and TS-2, are titanium isomorphs of
ZSM-5 and ZSM-11 and have the MFI and MEL structures,
respectively. The titanium silicates were found to catalyze
the conversion of cyclohexanone into cyclohexanone
oxime in the presence NH3 and H2O2 with high selectiv-
ities without the coproduction of any ammonium sulfate
(16–20).
We report here ammoximation of various ketones cata-
lyzed by various TS-1 and amorphous SiO2–TiO2. The in-
fluences of steric conformation of ketone and its diffusion,
and the reaction site of TS-1 are also discussed.
Ammoximation of cyclic ketones on TS-1 and amorphous
SiO2–TiO2 with H2O2 and NH3 was studied. The reactivity of cyclic
ketones over TS-1 and amorphous SiO2–TiO2 showed the fol-
lowing order: cycloheptanone > cyclopentanone > cyclohexanone
> cyclooctanone. Shape-selective poisoning studies revealed that
the active sites of TS-1 for ammoximation of cycloheptanone are
mostly inside the molecular sieve. The reactivity of cyclohexanone
and methylcyclohexanones over TS-1 followed the order cyclohex-
anone > 2-methylcyclohexanone Լ 3-methylcyclohexanone > 2,6-
dimethylcyclohexanone, reflecting the difference in the diffusivity
among the reactants or products inside TS-1. Besides the diffusivity
the reactivity seemed to be affected by the difficulty in access of
the carbonyl group to the Ti active site inside the zeolite; from
the comparison of reactivity of dimethycyclohexanone isomers,
the steric hindrance of the substituent methyl group to the access
of carbonyl group was considered to decrease in the following
order: ꢀ-equatorial > ꢁ-equatorial > ꢀ-axial > ꢁ-axial. Diffusion
of cyclohexanone or its oxime has been found to play an important
role in the ammoximation when the TS-1 crystals larger than
1.5 ꢂm were employed as a catalyst. For the TS-1 crystals smaller
than 0.3 ꢂm, the ammoximation activity was independent of the
diffusivity factor and was determined by the number of active Ti
c
EXPERIMENTAL
TS-1 was prepared by a method reported in a literature
(21). Tetraethyl orthosilicate (TEOS) (Tokyo Chemical In-
dustry), tetrabutyl orthotitanate (TBOT) (Kanto Chem-
ical), and tetrapropylammonium hydroxide (TPAOH)
(20 wt% aq., Tokyo Chemical Industry) were used. Aque-
ous TPAOH (one-fifth of the total amount) was first added
to TEOS. Then TBOT in 2-propanol and finally the rest
of TPAOH were added to give a clear solution with the
composition
sites present in the zeolite framework.
ꢀ 1996 Academic Press, Inc.
TEOS : TBOT : TPAOH : H2O = 1 : 0.02 : 0.34 : 15.
INTRODUCTION
After alcohols were removed at 353 K, this solution was
heated in a Teflon-lined autoclave at 423 K for 50 h. The
product recovered by filtration was calcined in air at 823 K
for 6 h. The Si/Ti ratio of this sample (TS-1-S1) was 79.
Other samples (TS-1-S2-S6) of TS-1 with a different crys-
tal size and Si/Ti ratio were prepared in a similar man-
ner and by changing the compositions of reactant mixture,
hydrolysis conditions, and temperature program in hy-
drothermal synthesis. A recently reported method (22) us-
ing a mixture of tetrapropylammonium and tetraethylam-
monium hydroxides (TEAOH, 10 wt% aq., Tokyo Chemi-
cal Industry) was also adopted to obtain TS-1 (TS-1-S7) of
increased crystal size. Following the method described in
a patent (23), TS-1(TS-1-S8) of large crystal size was syn-
thesized; colloidal silica (Snowtex 40, Nissan Chemical In-
dustries) was employed as a silica source and hydrothermal
synthesis was carried out at 428 K for 6 days.
The titanium silicalite molecular sieves TS-1 and TS-2
have been reported to catalyze a variety of oxidation reac-
tions in the presence of aqueous hydrogen peroxide (1–9).
The production of cyclohexanone oxime is an important
step in the production of nylon-6. The methods presently
employed in the manufacture of cyclohexanone oxime can-
not avoid the production of ammonium sulfate and the use
of hazardous chemicals such as oleum, halides and nitrogen
oxides (10). There have been many attempts to change the
process into environmentally benign ones. Gas-phase am-
moximation using NH3 and O2 has been reported (11–15),
but the yields were relatively low. The titanium silicate
1
Present address: Engineering Research Institute, Faculty of Engi-
neering, University of Tokyo, Yayoi, Tokyo 113, Japan.
2
To whom correspondence should be addressed. Fax: 81-3-5800-6825.
E-mail: tatsumi@komiyama.t.u-tokyo.ac.jp.
570
0021-9517/96 $18.00
c
Copyright ꢀ 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

AMMOXIMATION OF CYCLIC KETONES
571
Amorphous SiO2–TiO2 was prepared by hydrolysis of
TABLE 1
TEOS and tetraethyl orthotitanate (Tokyo Chemical In-
dustry) at 423 K. The product recovered by filtration was
calcined in air at 823 K for 6 h. The Si/Ti ratio of this sample
was 85.
a
Ammoximation of Ketones on TS-1 and Amorphous SiO2–TiO2
Oxime yield (%)
Reactant
TS-1b
Amorphous SiO2–TiO2
c
Chemical composition of the samples was obtained by
ICP (Japan Jarrell-Ash ICAP 575 Mark II) after solubiliza-
tion of the samples in HF, HCl, and HNO3. X-ray pow-
der diffraction patterns were collected on a Rigaku Denki
RU-200A diffractometer using CuKꢁ radiation. SEM mi-
crographs were obtained with a JEOL JSM 5400 micro-
scope equipped with an ED probe. The Fourier transform
infrared spectra (IR) were obtained on a Perkin Elmer
1600 spectrometer using the KBr pellet technique. UV–Vis
spectra were obtained on a Hitachi 340 spectrophotome-
ter. Liquid-phase adsorption experiments were carried out
at 303 K using 1,3,5-triisopropylbenzene, whose dimensions
were too large to enter the pores of the MFI zeolite, as a
solvent (24). Adsorption measurements of o-xylenes were
conducted on a highly sensitive microbalance. After zeolite
samples (0.25 g) were evacuated at 473 K, the measure-
ments were performed at 393 K and an o-xylene pressure
of 0.39 kPa to estimate the time taken to reach 30% of the
amount of o-xylene adsorbed at infinite time.
Cyclopentanone
Cyclohexanone
Cycloheptanone
Cyclooctanone
30
24
39
21
3.2
3.3
4.0
2.3
a Catalyst, 50 mg; ketone, 10 mmol; ketone : NH3 : H2O2 = 1 : 2 : 1 (mo-
lar ratio); H2O2 (5 wt% aq. solution); NH3 (28 wt% aq. solution); tem-
perature, 353 K; reaction duration, 5 h.
b TS-1-S1, Si/Ti = 79.
c Si/Ti = 85.
yield was included in the oxime yield in Table 1. The reac-
tivity of the cyclic ketones on the TS-1 catalyst followed the
decreasing order
cycloheptanone > cyclopentanone
> cyclohexanone > cyclooctanone.
The reactivity of the cyclic ketones on amorphous SiO2–
TiO2 followed the order
The oxidation reactions were carried out in a three-neck
flask (50 ml). One neck was fitted with a condenser and
another neck was fitted with a rubber septum. All ketones
were purchased from Tokyo Chemical Industry. To a mix-
ture of 10 mmol of ketone and 2 ml of water was added
1.4 ml of NH3 (28 wt% aq.) and then 50 mg of catalyst.
The resulting mixture was heated with a water bath kept at
353 K under vigorous stirring. Hydrogen peroxide (5 wt%)
was injected dropwise through the septum with a feed pump
for 2 h. The ketone : NH3 : H2O2 molar ratio was 1 : 2 : 1. Af-
ter addition of hydrogen peroxide, the reaction was contin-
ued for 3 h. After the reaction, the products were extracted
with ether and analyzed on a Shimadzu GC 14A gas chro-
matograph equipped with a 50-m OV-1 capillary column.
Cyclopentanone, 2-pentanone, or cyclohexanone was used
as an internal standard for gas chromatographic analysis.
cycloheptanone > cyclopentanone
Լ cyclohexanone > cyclooctanone.
Althoughsmallamountsofperoxydicycloalkylamineand
aldol condensation products were also formed on TS-1, the
selectivity to oximes including lactams was >97%. In the
reaction of cyclohexanone on TS-1 the conversion of H2O2
was 94% and hence the yield of cyclohexanone oxime based
on the converted H2O2 was only 26%, indicating that H2O2
extensively decomposed without oxidizing the substrate.
However, it is unlikely that this nonproductive decompo-
sition occurs on extraframework titanium on TS-1; even in
the absence of catalyst H2O2 conversion (decomposition)
was 92%. Because of low concentration (4.4 g/liter) of TS-1
nonproductive thermal decomposition of H2O2 was an in-
evitable concomitant. It is noteworthy that our catalyst con-
centration is much lower than that of Roffia et al. (18) who
observed that lowering catalyst concentration results in a
considerable decrease in the cyclohexanone conversion and
the oxime yield based on H2O2.
RESULTS AND DISCUSSION
Ammoximation of Cyclic Ketones on TS-1
and Amorphous SiO2–TiO2
Except for cycloheptanone, the reactivity of cyclic ke-
Ammoximation of various cyclic ketones on a TS-1 tones on the TS-1-S1 catalyst decreased with the increases
sample (TS-1-S1, Si/Ti = 79) and amorphous SiO2–TiO2 in their size; it was in good agreement with that on amor-
(Si/Ti = 85) was carried out under the standard conditions. phous SiO2–TiO2, where cycloheptanone still showed the
The results are presented in Table 1. The experimental error highest activity. This is consistent with the observation by
was about ꢁ 0.15%. On TS-1 ꢃ-valerolactam (0.2% yield) Armor et al. (13), who found that the cycloheptanone was
and ε-caprolactam (0.1% yield) were also obtained from more reactive than other cyclic ketones in the ammoxima-
cyclopentanone and cyclohexanone, respectively. Since tion of ketones with NH3 and O2. Thus the reactivity of
these lactams are considered to be formed via oximes, their cyclic ketones on the TS-1 sample apparently reflects the

572
TATSUMI AND JAPPAR
inherent reactivity, as yet not clarified, in the absence of can enter the pore of the molecular sieve and poisoned both
steric restriction in the MFI structure. the outer and the inner surface, resulting in remarkably de-
Amorphous SiO2–TiO2 was found to be much less ac- creased activity. It is concluded that the ammoximation of
tive than crystalline zeolite TS-1 in the ammoximation of cycloheptanone took place mostly inside the TS-1 molecu-
cyclic ketones. These results suggest that Ti4+ present in lar sieve. The poisoning was not due to suppression of NH3
zeolite TS-1 is different from Ti4+ present in amorphous activation, since oxyfunctionalization of hexane on TS-1
SiO2–TiO2, resulting from the difference in dispersion and was similarly inhibited by triethylamine (28).
coordination state of Ti (vide infra). By-products such as
On the other hand, the cycloheptanone oxime yield was
lactam, peroxydicycloalkylamine, and aldol condensation largely decreased by addition of either triphenylamine or
products were negligible on amorphous SiO2–TiO2.
triethylamine over amorphous SiO2–TiO2. Since the aver-
age pore diameter of amorphous SiO2–TiO2 was 22 nm,
it is conceivable that both amines easily entered the pore,
inhibiting the ammoximation.
Influence of Organic Amine on Ammoximation
of Cycloheptanone
The pore sizes of titanium silicate TS-1 with an MFI
structure are 0.53 ꢂ 0.56 nm for the straight channel and
0.51 ꢂ 0.55 nm for the sinusoidal channel (25), and these
dimensions are about the same as those of cross-section of
C5–C8 cyclic ketones. It is important to distinguish between
the reactions on outer and inner surfaces. It is proposed that
hydroperoxy or peroxy Ti species over TS-1 is the active site
for ammoximation reaction (16, 26). To clarify contribution
from inner and outer surfaces of the TS-1 catalyst, organic
amines of different size were used to poison catalyst surface
selectively.
The ammoximation of cycloheptanone was carried out
in the presence of triphenylamine or triethylamine. The
results are compared in Table 2. It was found that when
triethylamine was added, the reactivity of cycloheptanone
was severely retarded and no aldol condensation products
were formed. Reddy et al. (27) reported similar inhibition of
aldol condensation when NH3 was added in excess. The cy-
cloheptanoneoximeyieldwasonlyslightlydecreasedbythe
addition of triphenylamine. As the size of triphenylamine
is too large to enter the pore of MFI zeolite, only the outer
surface of TS-1 would be poisoned by triphenylamine. The
reactivity of cycloheptanone was only slightly decreased be-
cause the triphenylamine-poisoned outer surface is consi-
derably small compared to the inner surface. Triethylamine
Reactivity of Methylcyclohexanones on TS-1
The reactivity and adsorbed amount of cyclohexanone
and methylcyclohexanones on the TS-1-S1 catalyst are
listed in Table 3. The reactivity of the methylcyclohex-
anones followed the decreasing order
cyclohexanone > 2-methylcyclohexanone
Լ 3-methylcyclohexanone
> 2,6-dimethylcyclohexanone.
Methylcyclohexanones showed lower reactivity than
cyclohexanone, indicating the steric restriction im-
posed by the methyl branch. However, except for 2,6-
dimethylcyclohexanone the amount of adsorption of
these methylcyclohexanones was almost the same as that
of cyclohexanone. The low adsorbed amount of 2,6-
dimethylcyclohexanone accounts for the low reactivity.
The reactivity of 2-methylcyclohexanone and 3-methyl-
cyclohexanone was about half of that of cyclohexanone.
To explain different reaction behavior of ammoximation of
methylcyclohexanones on the TS-1 catalyst, adsorption rate
of methylcyclohexanones on the catalyst was measured.
The results are shown in Fig. 1. The adsorption of cyclohex-
anone was fast and nearly saturated adsorption (58 mg/g)
was attained only after 15 min. This capacity happened to
be almost equal to the cyclohexane adsorption capacity
TABLE 2
TABLE 3
Influence of Organic Amine on Ammoximation of Cycloheptanone
a
Reactivity and Adsorbed Amount of Cyclic Ketones on TS-1
on TS-1 and Amorphous SiO2–TiO2
Oxime yielda
(%)
Adsorbed amountb
(mg/g-cat)
Oxime yield (%)
Reactant
Catalyst
None
Triphenylamine
Triethylamine
Cyclohexanone
24
66
65
67
6.0
TS-1b
SiO2–TiO2
39
4.0
32
1.0
5.4
1.4
2-Methylcyclohexanone
3-Methylcyclohexanone
2,6-Dimethylcyclohexanone
12.4
12.6
5.0
c
a Catalyst, 50 mg; ketone, 10 mmol; ketone : NH3 : H2O2 = 1 : 2 : 1 (mo-
lar ratio); H2O2 (5 wt% aq. solution); NH3 (28 wt% aq. solution); organic
amine : ketone = 1 : 1, temperature, 353 K; reaction duration, 5 h.
b TS-1-S1, Si/Ti = 79.
a Catalyst (TS-1-S1), 50 g; ketone, 10 mmol; ketone : NH3 : H2O2 =
1 : 2 : 1 (molar ratio); H2O2 (5 wt% aq. solution); NH3 (28 wt% aq. so-
lution); temperature, 353 K; reaction duration, 5 h.
c Si/Ti = 85.
b Solvent, 1,3,5-triisopropylbenzene; 303 K; 3 h.

AMMOXIMATION OF CYCLIC KETONES
573
FIG. 1. Time course of adsorption of ketone on TS-1. Catalyst (TS-
1-S1), 100 mg; solvent, 1,3,5-triisopropylbenzene; 303 K. ( ) cyclohex-
anone; (᭡) 2-methylcyclohexanone; (᭹) 3-methylcyclohexanone; (᭜) 2,6-
dimethylcyclohexanone.
FIG. 2. Influence of methyl substituents on ammoximation of di-
methylcyclohexanoneonTS-1. Catalyst(TS-1-S1), 50mg;ketone, 10mmol
(2,5-dimethylcyclohexanone:cis/trans = 0.28; 2,6-dimethylcyclohexanone:
cis/trans = 5.3); ketone : NH3 : H2O2 = 1 : 2 : 1 (molar ratio); H2O2 (5 wt%
aq. solution); NH3 (28 wt% aq. solution); temperature, 353 K; reaction
duration, 5 h.
(59 mg/g) obtained by Mirajkar et al. (29), although our
experimental conditions (cyclohexanone in liquid phase at
303 K) considerably differed from theirs (cyclohexane in
vapor phase at 298 K).
The adsorbed amount of 2-methylcyclohexanone and
3-methylcyclohexanone gradually increased to reach the
level similar to cyclohexanone after 3 h. The adsorbed
amount of 2,6-dimethylcyclohexanone increased very
slowly. The order of adsorption rate (diffusivity) was
Fig. 2. The reactivity of cis- and trans- forms followed the
order
trans-2,6-dimethylcyclohexanone
> cis-2,5-dimethylcyclohexanone
> cis-2,6-dimethylcyclohexanone
> trans-2,5-dimethylcyclohexanone.
cyclohexanone > 2-methylcyclohexanone
Լ 3-methylcyclohexanone
The stable conformation of each isomer is shown in Fig. 3.
A molecule of trans-2,6-dimethylcyclohexanone exists as
a pair of conformational enantiomers. In either confor-
mation, only one methyl group can occupy an equatorial
position. On the other hand, the cis-isomer exists as two
conformational diastereomers. In the more stable con-
formation of cis-2,6-dimethylcyclohexanone, both methyl
groups occupy equatorial positions. Similarly, the more
stable conformation of trans-2,5-dimethylcyclohexanone is
> 2,6-dimethylcyclohexanone.
The reactivity of the methylcyclohexanones on the TS-1
catalyst obeyed the same order as their adsorption rate.
These results show that the diffusivity of the reactant con-
trolled the rate in the ammoximation of methylcyclohex-
anones over the TS-1 catalyst.
Influence of Steric Factor in Ammoximation
of Dimethylcyclohexanones on TS-1
It was thought that the low reactivity of dimethyl-
cyclohexanones may also result from the difficulty in the
access of carbonyl group to active sites of TS-1 caused
by the steric hindrance of their methyl substituents. To
investigate the effect of the methyl substituent on the
reactivity, a comparison was made among the ammox-
imation rates of 2,6- and 2,5-dimethylcyclohexanones.
Dimethylcyclohexanone contains cis- and trans-isomers.
The cis/trans ratios of 2,6-dimethylcyclohexanone and
2,5-dimethylcyclohexanone are 5.25 and 0.28, respectively.
The results of ammoximation of mixture of cis- and trans-
isomers on the TS-1 catalyst (TS-1-S1) are presented in
FIG. 3. The stable conformation of dimethylcyclohexanone isomers.

574
TATSUMI AND JAPPAR
the diequatorial one. In the cis-2,5-dimethylcyclohexanone ple TS-1-S7 was synthesized using a mixture of TPAOH
isomer there is one equatorial and one axial methyl group. and TEAOH and consisted mostly of hexagonal crystals of
Dimethylcyclohexanones have not only difficulty in dif- 1.5–3 ꢂm. With colloidal silica as a silica source TS-1-S8
fusion inside the TS-1 pore, but also their methyl sub- was obtained, TS-1-S8 comprised mainly crystals of over
stituents causes steric hindrance to the access of carbonyl 5 ꢂm and in the shape of hexagonal coffins. Thagaraj et al.
group to active sites of TS-1. Comparison of the reactiv- observed that the intensity ratio of the 960 cmꢃ1 band to the
ity shown in Fig. 3 suggests that the steric hindrance of the 550 cmꢃ1 band (I960/I550) in the IR spectra increased with
methyl group to the access of carbonyl group decreases in the Ti content of the TS-1 framework (21). The ratio can be
the order
employed to present the relative concentration the Ti con-
tent in the framework. The UV–Vis spectra of TS-1-S1-S8
samples exhibited the ligand to metal charge transfer band
assigned to isolated Ti in tetrahedral framework position in
the range from 210 to 230 nm. The TS-1-S3-S5 sample had
also a broad shoulder due to extraframework Ti in the re-
gion of 240 to 250 nm. The low relative intensity I960/I550, for
the samples TS-1-S3-S5 is in agreement with the presence of
extraframework Ti suggested by their UV–Vis spectra. The
presence of a broad shoulder at about 270 nm in the case
of TS-1-S6 could be due to hexacoordinated Ti species con-
taining Ti–O–Ti bonds. The absence of adsorption of about
340 nm in all samples indicate the absence of titanium ox-
ide (anatase) impurities (32). It is noteworthy amorphous
SiO2–TiO2 showed the band assigned to anatase.
With these TS-1 catalysts, ammoximation of cyclohex-
anone was performed. Figure 4 illustrates the dependence
of the turnover number (mol-products/mol-Ti) of ammox-
imation of cyclohexanone on the characteristic diffusion
time, t0.3. It is found that high turnover numbers were at-
tained for the low t0.3 samples (lower than 100 s), where
the diffusion should occur effectively. For the large t0.3 sam-
ples, the turnover numbers were severely suppressed. Since
there is a clear correlation between the turnover number
and t0.3 for the TS-1 samples with t0.3 > 180 s, it is suggested
ꢀ-equatorial > ꢁ-equatorial > ꢀ-axial > ꢁ-axial.
This is contrary to what would be expected; it is gener-
ally considered that the ꢁ-methyl group, which is close to
the carbonyl group, causes more steric hindrance than the
ꢀ-methyl group. The large retarding effect of the ꢀ-methyl
group in the equatorial position may be due to the remote
geometrical restriction inside the zeolite pore accommo-
dating the reactant.
Factors Determining the Activity in Ammoximation
of Cyclic Ketones
Diffusion may be an important step in catalytic reactions,
but it is difficult to determine the diffusivity directly under
the reaction conditions. According to the report of Olson
and Haag, the time necessary for 30% adsorption (t0.3) of
o-xylene can be used as a criterion to measure the diffusivity
of reactant and product inside zeolite; t0.3 is proportional to
r2/D, where r is the crystal radius and D is the diffusivity
of the molecule (30). Van der Pol et al. have reported that
smaller crystals of TS-1 have higher activity in the hydro-
xylation of phenol (31).
Seven more samples of TS-1 with varying crystal size and
Si/Ti ratio were prepared. The samples TS-1-S2-S6 were
prepared in a similar manner to TS-1-S1. X-ray diffrac-
tograms showed good crystallinity for all the TS-1 sam-
ples. The results of ICP, SEM, IR, and UV–Vis of all sam-
ples the shown in Table 4. SEM observation revealed that
these samples were spherulitic particles under 0.3 ꢂm. Sam-
TABLE 4
Characterization of TS-1 Samples and Amorphous SiO2–TiO2
Crystal size
UV-Vis
(nm)
Catalyst
Si/Ti
(ꢂm)
I960 cmꢃ1 /I550 cm
ꢃ1
TS-1-S1
TS-1-S2
TS-1-S3
TS-1-S4
TS-1-S5
TS-1-S6
TS-1-S7
TS-1-S8
SiO2–TiO2
79
92
61
52
58
120
56
53
0.1–0.3
0.1–0.3
0.1–0.3
0.1–0.3
0.1–0.3
0.3–0.5
1.5–3.0
>5
0.25
0.16
0.15
0.14
0.10
0.03
0.10
0.13
—
210–230
210–230
210–230
210–250
210–240
210–270
210–230
230
FIG. 4. Relationship between t0.3 and activity in ammoximation
cyclohexanone. Catalyst (TS-1-S1-S8), 50 mg; cyclohexanone, 10 mmol;
ketone : NH3 : H2O2 = 1 : 2 : 1 (molar ratio); H2O2 (5 wt% aq. solution);
NH3 (28 wt% aq. solution); temperature, 353 K; reaction duration, 5 h.
Numbers in parentheses indicate the sample number n of TS-1-S n.
85
—
300–350

AMMOXIMATION OF CYCLIC KETONES
575
that the ammoximation of cyclohexanone on the catalysts
TABLE 5
with large crystal size was strongly affected by diffusion of
cyclohexanone or its oxime products. Scatter in the data
for t0.3 < 100 s suggests that for the TS-1 samples with small
crystal size, the reaction was not governed by diffusional
factors.
Oxime Yield from Cyclic Ketones on TS-1 with Different
Crystal Sizea
Oxime yield/(I960 cmꢃ1 /I550 cm
)
ꢃ1
Catalyst
Si/Ti
t0.3
Cyclohexanone
Cyclooctanone
C8/C6
The oxime yield is plotted against the I960/I550 ratio in
Fig. 5. Except for the samples TS-1-S7 and TS-1-S8 whose
t0.3 was >100 s, the oxime yield was found to be virtually
proportional to the relative intensity I960/I550. Therefore,
for the zeolite crystals with a size under 0.3 ꢂm the oxime
productivity was simply determined by the number of ac-
tive sites present in the TS-1 framework. There was no
clear relation between the presence/absence of extraframe-
work Ti species deduced from the UV–Vis and the catalytic
performance. For the samples having no extraframework
Ti, TS-1-S1 and TS-1-S2 exhibited relatively high turnover
number and oxime yield but TS-1-S6 and TS-1-S7 not.
The ammoximation of cyclohexanone and cyclooctanone
was carried out using the catalyst with longest t0.3 (TS-1-S8)
and the standard catalyst (TS-1-S1). Results are compared
in Table 5. The specific activity of TS-1-S1, expressed in
terms of oxime yield per relative intensity I960/I550, for the
ammoximation of cyclohexanone and cyclooctanone was
lower than that of TS-1-S8. Ammoximation of cyclooc-
tanone was slower than that of cyclohexanone, probably
due to the low diffusivity of cyclooctanone and its oxime
compared to that of cyclohexanone and its oxime. Compar-
ison of the reactivity ratio of cyclooctanone/cyclohexanone
for the two TS-1 samples shows more suppression of reac-
tivity for cyclooctanone for the sample with larger crys-
tal size (long t0.3), suggesting that the ammoximation of
cyclooctanone also takes place mainly inside the TS-1 zeo-
A
B
79
53
58
891
96
38
84
22
0.88
0.58
a
Catalyst (A = TS-1-S1, B = TS-1-S8), 50 mg; ketone, 10 mmol; ke-
tone : NH3 : H2O2 = 1 : 2 : 1 (molar ratio); H2O2 (5 wt% aq. solution); NH3
(28 wt% aq. solution); temperature, 353 K; reaction duration, 5 h.
lite; if the reaction occurs on the external surface, one would
not expect significant difference in the reactivity ratio be-
tween the two samples.
CONCLUSIONS
The reactivity of ketones over TS-1 shows the order
cycloheptanone
> cyclopentanone > cyclohexanone
> cyclooctanone > 2-methylcyclohexanone
Լ 3-methylcyclohexanone
> 2,6-dimethylcyclohexanone.
The active sites of TS-1 for ammoximation of cyclo-
heptanone are mostly inside the molecular sieve. From
the relative reactivity of dimethylcyclohexanone isomers,
it is concluded that the steric hindrance of the substituent
methyl group to the access of carbonyl group inside the ze-
olite decreases in the order ꢀ-equatorial> ꢁ-equatorial > ꢀ-
axial > ꢁ-axial. Diffusion of cyclohexanone or its oxime has
been found to play an important role in the ammoximation
when the TS-1 crystals larger than 1.5 ꢂm are employed as a
catalyst. For the TS-1 crystals smaller than 0.3 ꢂm, the am-
moximation activity is determined by the number of active
Ti sites present in the zeolite framework.
ACKNOWLEDGMENT
T.T. acknowledges Tokuyama Science Foundation for financial support.
REFERENCES
1. Notari, B., Stud. Surf. Sci. Catal. 37, 413 (1987).
2. Clerici, M. G., Bellussi, G., and Romano, U., J. Catal. 129, 159 (1991).
3. Tatsumi, T., Nakamura, M., Negishi, S., and Tominaga, H., J. Chem.
Soc. Chem. Commun. 476 (1990).
4. Huybrechts, D. R. C., DeBruycker, L., and Jacobs, P. A., Nature
(London) 345, 240 (1990).
FIG. 5. Relationship between oxime yield an I960/I550. Catalyst (TS-
1-S1-S8), 50 mg; cyclohexanone, 10 mmol; ketone : NH3 : H2O2 = 1 : 2 : 1
(molar ratio); H2O2 (5 wt% aq. solution); NH3 (28 wt% aq. solution);
temperature, 353 K; reaction duration, 5 h. Numbers in parentheses indi-
cate the sample number n of TS-1-S n.
5. Hari Prasada Rao, P. R., Thangaraj, A., and Ramaswamy, A. V.,
J. Chem. Soc. Chem. Commun. 1139 (1991).
6. Tatsumi, T., Nakamura, M., Negishi, S., and Tominaga, H., Chem. Lett.
297 (1990); Tatsuni, T., Nakamura, M., Negishi, S., and Tominaga, H.,
Catal. Lett. 10, 259 (1991).

576
TATSUMI AND JAPPAR
7. Tatsumi, T., Nakamura, M., and Tominaga, H., in “Catalytic Science 21. Thagaraj, A., Kumar, R., Mirajkar, S. P., and Ratnasamy, P., J. Catal.
and Technology” (S. Yoshida, N. Takezawa, and T. Ono, Eds.), Vol. 1,
p. 213. Kodansha-VCH, Tokyo, 1991.
103, 1 (1991).
22. Tuel, A., and Ben Taarit, Y., Microporous Mater. 1, 179 (1993).
8. Reddy, J. S., Kumar, R., and Ratnasamy, P., Appl. Catal. 58, L1 (1991). 23. U.K. Patent GB 2071071, 1981.
9. Clerici, M. G., Appl. Catal. 68, 249 (1991).
24. Namba, S., Kanai, Y., Shoji, H., and Yashima, T., Zeolite 4, 77 (1984).
10. “Kirk–Othmer Encyclopedia of Chemical Technology,” Vol. 18, p. 432. 25. Meier, W. M., and Olson, D. H., in “Atlas of Zeolite Structure Types,”
Wiley, New York, 1982.
11. Armor, J. N., J. Catal. 70, 72 (1981).
12. Armor, J. N., Carlson, E., Soled, S., Conner, W. C., Laverick, A.,
DeRites, B., and Gates, W., J. Catal. 70, 84 (1981).
p. 138. Butterworth–Heinemann, Stoneham, 1992.
26. Bellussi, G., and Rigutto, M. S., Stud, Surf. Sci. Catal. 85, 177 (1994).
27. Reddy, J. S., Sivasanaker, S., and Ratnasamy, P., J. Mol. Catal. 69, 383
(1991).
13. Armor, J. N., Zambri, P. M., and Leming, R., J. Catal. 72, 66 (1981).
14. Dreoni, D. P., Pinelli, D., and Trifiro, F., J. Mol. Catal. 69, 171 (1991).
28. Nakamura, M., Negishi, S., Yuasa, K., Tatsumi, T., and Tominaga, H.,
Shokubai 32, 99 (1990).
15. Dreoni, D. P., Pinelli, D., and Trifiro, F., Buska, G., Lorenzelli, G., Catal. 29. Mirajkar, S. P., Thangaraj, A., and Shiralkar, V. P., J. Phys. Chem. 96,
Lett. 11, 285 (1991). 3073 (1992).
16. Thangaraj, A., Sivasanker, S., and Ratnasamy, P., J. Catal. 131, 394 30. Olson, D. H., and Haag, W. O., in “Catalytic Materials,” (T. E. Whyte,
(1991).
R. A. Dalla Betta, E. Derouane, and R. T. K. Baker, Eds.), ACS Sym-
posium Series, Vol. 248, p. 275. Am. Chem. Soc., Washington, DC,
1984.
17. Reddy, J. S., Sivasanker, S., and Ratnasamy, P., J. Mol. Catal. 70, 335
(1991).
18. Roffia, P., Stud. Surf. Sci. Catal. 55, 43 (1990).
19. Tvaruzkova, Z., Habersberger, K., Zilkova, N., and Jiru, P., Appl.
Catal. A 79, 105 (1991).
20. Tvaruzkova, Z., Petras, M., Habersberger, K., and Jiru, P., Catal. Lett.
13, 117 (1991).
31. Van der Pol, A. J. H. P., Verduyn, A. J., and Van Hooff, J. H. C., Appl.
Catal. A 92, 113 (1992).
32. Bordiga, S., Boscheerini, F., Buffa, F., Coluccia, S., Genoni, F.,
Lamberti, C., Lefanti, G., Matchese, L., Petrini, G., Vlaic, G., and
Zecchina, A., J. Phys. Chem. 98, 4125 (1994).