Oxidation of benzene to phenol with hydrogen peroxide catalyzed by a modified titanium silicalite (TS-1B)

Article abstract of DOI:10.1002/adsc.200600371

The hydroxylation of benzene to phenol with hydrogen peroxide was investigated using different solvents and a series of catalysts, obtained by modification of titanium silicalite (TS-1). The best results were obtained after post-synthesis treatment of TS-1 with NH₄HF₂ and H ₂O₂. The new catalyst (TS-1B), used in the presence of a particular co-solvent (sulfolane) is able to protect the produced phenol from over-oxidation and dramatically enhanced the selectivity of the reaction.

Full text of DOI:10.1002/adsc.200600371

FULL PAPERS
DOI: 10.1002/adsc.200600371
Oxidation of Benzene to Phenol with Hydrogen Peroxide
Catalyzed by a Modified Titanium Silicalite (TS-1B)
a,
a
a
a
Daniele Bianchi, * Luigi Balducci, Rossella Bortolo, Rino D’Aloisio,
a
a
a
a
Marco Ricci, Guido Spanò, Roberto Tassinari, Cristina Tonini,
and Raffaele Ungarelli
a
a
Enitecnologie S.p.A. – Centro Ricerche di Novara – Istituto Donegani, Via Fauser 4, 28100 Novara, Italy
Fax : (+39)-0321-447-241; e-mail: daniele.bianchi@polimerieuropa.com
Received: July 21, 2006
Abstract: The hydroxylation of benzene to phenol vent (sulfolane) is able to protect the produced
with hydrogen peroxide was investigated using differ- phenol from over-oxidation and dramatically en-
ent solvents and a series of catalysts, obtained by hanced the selectivity of the reaction.
modification of titanium silicalite (TS-1). The best
results were obtained after post-synthesis treatment
of TS-1 with NH HF and H O . The new catalyst Keywords: hydrogen peroxide; hydroxylation; phe-
4
2
2
2
(
TS-1B), used in the presence of a particular co-sol- nols; solvent effects; titanium; zeolites
Introduction
Results and Discussion
Phenol is one of the most important intermediates of The aim of this work was the development of a selec-
the chemical industry with a global production esti- tive method for the oxidation of benzene to phenol
mated at around 8 Mt/year, almost completely based with hydrogen peroxide. In the first stage of the study,
upon the cumene process which also co-produces ace- the hydroxylation of benzene was investigated using
[
3]
tone. Since the expected market growth is much titanium silicalite (TS-1) as the catalyst, on the basis
larger for phenol than for acetone, the development of its well established efficiency in a series of oxida-
of a process based on the direct oxidation of benzene tion processes with hydrogen peroxide, such as alkene
[
4]
[5]
to phenol is an important goal in industrial organic epoxidation , alkane oxidation, alcohol oxida-
[
6]
[7]
chemistry. Unfortunately, the oxidation of benzene is tion, cyclohexanone ammoximation, and phenol
[
8]
usually affected by a poor selectivity due to the lack hydroxylation.
of kinetic control: indeed, phenol is more easily oxi-
dized than benzene itself and substantial amounts of free medium (benzene/water),
The reactions were carried out either in a solvent-
[
9,10]
or using different
over-oxygenated by-products (catechol, hydroqui- co-solvents in order to homogenize the hydrophobic
none, benzoquinones, and tars) also form. Moreover, substrate and the aqueous hydrogen peroxide in the
only few oxidants are environmentally acceptable, the liquid phase.
most promising among them appearing to be nitrous
As shown in Table 1, the selectivity of the reaction
oxide (exploited in the Solutia process, based on cat- was strongly affected by the nature of the solvent. In
[
1]
alysis by iron containing zeolites) and hydrogen per- our hands, the selectivity of the reaction observed op-
oxide. erating under the biphasic, solvent-free conditions
With the latter oxidant, remarkable results have was sensibly lower than that reported in the litera-
[
9]
been recently obtained by combining the use of a pe- ture. Namely, the selectivity to phenol (expressed as
culiar solvent (sulfolane) and a modified titanium sili- moles of produced phenol/moles of reacted benzene”
[
2]
calite (TS-1B) as the catalyst. We now report full 100) measured for 4.5% and 15.1% benzene conver-
details about this new, acetone-free route to phenol sions were 42% and 24%, respectively.
and about the synthesis and the performance of the
new catalyst.
In the case of the most commonly used co-solvents
(acetone, acetonitrile, methanol, tert-butyl alcohol and
acetic acid) the selectivity to phenol (expressed as
moles of produced phenol/moles of reacted benzene”
100) was lower than 50% even at a benzene conver-
Adv. Synth. Catal. 2007, 349, 979 – 986
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
979

Daniele Bianchi et al.
FULL PAPERS
Table 1. Effect of the co-solvent on benzene oxidation.
[a]
[b]
[c]
Co-solvent
H O conv. [%]
H O sel. [%]
Benzene conv. [%]
Product selectivity [%]
2
2
2
2
PH
CA
HY
BQ
TAR
[
d]
none
acetone
100
90
93
98
90
85
100
12
94
97
20
4.5
4.2
4.5
4.7
4.1
5.1
0.3
0.1
7.3
7.8
42
41
43
40
44
50
>99
>99
80
14
13
16
15
8
16
0
0
17
12
12
15
10
11
0
0
5
4
-
27
30
26
26
35
12
0
0
7
5
19
21
19
20
30
3
6
62
67
4
3
4
3
11
0
0
1
1
acetonitrile
methanol
tert-butyl alcohol
acetic acid
dimethylformamide
N-methylpyrrolidone
dimethylsulfone
sulfolane
7
7
83
[
a]
Reaction conditions: benzene/co-solvent=2/8 w/w, H O /benzene molar ratio=0.1, TS-1/benzene=0.1 w/w, tempera-
2
2
ture=1008C. PH =phenol, CA=catechol, HY=hydroquinone, BQ=benzoquinone.
[
[
[
b]
c]
d]
Moles of phenol/moles of reacted H O ”100.
2
2
Moles of product/moles of reacted benzene ” 100.
Reaction conditions: benzene/water=1/5 w/w, H O /benzene molar ratio=0.1, TS-1/benzene=0.15 w/w, temperature=
2
2
[9]
6
08C.
sion lower than 5%, mainly due to the formation of
dioxygenated products and tars.
This effect was confirmed by the estimated loading
of the free phenol molecule and the phenol-sulfolane
Even worse results were obtained using nitrogen- complex (expressed as number of loaded molecules in
containing solvents, such as dimethylformamide, that a crystal elementary unit of TS-1), which turned out
promotes the unproductive decomposition of hydro- to be 13.6 and 0.8, respectively, calculated using the
gen peroxide, and N-methylpyrrolidone, that depress- software ꢁSorptionꢂ (Accelrys Inc., San Diego).
es the catalytic activity of TS-1, affording only negligi-
ble yields to phenol.
Additional evidence for the protective effect exert-
ed by sulfolane was obtained by measuring the reac-
Conversely, a dramatic improvement of selectivity tion rate, expressed as turn-over frequency (TOF:
was observed using co-solvents bearing a sulfone moles of reacted substrate/moles of Ti”hour) for the
group, such as dimethyl sulfone and tetramethylene oxidation of benzene and phenol, carried out sepa-
sulfone (sulfolane), which allowed a conversion of rately in acetone and sulfolane as co-solvents. As
benzene close to 8% with a selectivity to phenol shown in Figure 2, in the case of acetone, the phenol
À1
higher than 80%. In the case of sulfolane, the detect- oxidation rate (TOF=190 hour ) was 10 times faster
ed by-products were catechol (7%), hydroquinone
(
4%), 1,4-benzoquinone (1%), and only a small
amount of tars (5%).
Sulfolane is a solvent with high dipole moment
(
4.81 Debye) and dielectric constant (43.26), showing
the peculiar property of forming complexes with phe-
[
11]
nolic compounds.
These complexes are not stable
enough to be isolated but have been identified in so-
[
12]
lution by IR spectroscopy.
The increased selectivity observed running the ben-
zene oxidation in sulfolane may be due to the forma-
tion of a sterically hindered species (Figure 1) which
cannot enter the titanium silicalite pores, thus allow-
ing phenol to remain relatively protected towards fur-
ther oxidation.
Figure 2. Oxidation rate of benzene and phenol. The oxida-
tions of benzene and phenol were carried out separately
using acetone and sulfolane as co-solvents and TS-1 (TiO =
2
3
1
.6% by weight) or TS-1B (TiO =0.9%, obtained from TS-
2
by post-treatment with (NH )HF and H O ). ( ) TOF of
&
4
2
2
2
Ti measured on benzene (moles of reacted benzene/moles
of Ti”hour); ( ) TOF of Ti measured on phenol (moles of
reacted phenol/moles of Ti”hour)
&
Figure 1. Protecting effect of sulfolane.
980
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 979 – 986

Oxidation of Benzene to Phenol with Hydrogen Peroxide
FULL PAPERS
with respect to that measured for benzene (TOF=19
Further treatments were carried out under acidic
À1
hour ). Conversely, operating in sulfolane the rate conditions, in presence of hydrogen peroxide, in order
À1
measured for phenol (TOF=51 hour ) was only 1.6 to promote the extraction of titanium as a soluble per-
time faster than that measured for benzene (TOF=31 oxide species. The efficiency of the titanium removal
À1
hour ), in agreement with the higher value of the ob- was strongly dependent on the acid used for the treat-
served selectivity.
ment. The extraction was negligible (2%) with phos-
Despite the improvement of the selectivity ob- phoric acid, moderate (10%) with sulfuric acid and
tained by the solvent engineering, the overall perfor- massive (close to 50%) with hydrofluoric acid.
mance of the reaction was still not suitable for the de-
In the case of sulfuric acid, some titanium was solu-
velopment of a new process for phenol production, bilized but no significant effects were observed on the
mainly due to the low productivity and the conspicu- performances of the resulting catalyst.
ous amount of by-products.
Conversely, using hydrofluoric acid, the modified
In particular, the formation of over-oxidation by- TS-1 turned out to be a less selective catalyst than the
products, such as 1,4-benzoquinone and tars, suggests starting material, with a drop of the selectivity based
the possible presence of a heterogeneous population on hydrogen peroxide from 67 to 56%. This was
of catalytic sites on TS-1, differing in reactivity and probably due to a severe damage of the crystalline
selectivity.
structure of the zeolite, after the aspecific removal of
In order to remove the less selective catalytic sites, a consistent fraction of both silicon and titanium
a catalyst engineering study, based on post-synthesis atoms, under the strong treatment conditions.
treatments, was carried out on TS-1.
Operating under milder conditions, using ammoni-
It has been reported that titanium atoms are not um fluoride as the fluorine source and hydrogen per-
equally distributed in the 12 crystallographically inde- oxide, the extraction of titanium was equally effective
pendent sites of the MFI framework of TS-1, but are (46%), but, in this case, a dramatic enhancement of
[
13]
hosted in 4 preferential positions, which, in princi- selectivity was observed. Even better results were ob-
ple, might display different catalytic properties. tained using ammonium bifluoride and hydrogen per-
Different titanium species can also have different oxide, with the selectivity based on benzene rising
structural properties, therefore they could be extract- from 83 to 91% and that based on hydrogen peroxide
ed from the zeolite framework at different rates. As from 67 to 80%.
listed in Table 2, several post-synthesis treatments of
The treatment carried out with ammonium bifluor-
the catalyst were carried out in water suspension at ide and hydrogen peroxide (F/Ti molar ratio=7 and a
808C, using different methods.
H O /Ti molar ratio=30) was highly selective in the
2 2
The simple treatment with ammonium acetate, cur- removal of titanium atoms, leaving unchanged the
[
14]
rently used to clean-up the catalyst,
did not pro- crystalline structure of the zeolite. Up to 75% of tita-
mote any titanium leaching, with negligible change of nium was extracted as (NH ) Ti(O )F , as confirmed
4
3
2
5
the catalytic activity of TS-1, tested on benzene oxida- by UV-Vis, IR and Raman spectroscopy, in compari-
tion. Conversely, using ammonium bifluoride, the ex- son with an authentic sample prepared as reported in
[
15]
traction of a small amount of titanium was detected, the literature.
Only a negligible molar fraction of
but, also in this case, with only a minor effect on the silica (1%) was removed as (NH ) SiF .
4
2
6
selectivity of the catalyst.
Table 2. Post-synthesis modification of TS-1.
[a]
[b]
[c]
[d]
Treatment
H O conv. [%]
H O sel. [%]
Benzene conv. [%]
Product selectivity [%]
2
2
2
2
PH
CA
HY
BQ
TAR
none
97
67
7.8
7.3
7.6
7.8
7.4
6.7
8.4
8.6
83
82
81
84
84
80
90
91
7
7
8
7
5
6
6
5
4
4
5
4
3
3
3
3
1
1
1
1
2
4
0
0
5
6
5
4
6
7
1
1
A
C
H
T
R
E
U
N
G
(NH )OAc
94
95
95
95
96
98
98
64
65
69
65
56
77
80
4
A
C
H
T
R
E
U
N
G
(NH )HF
4 2
H PO /H O
3
4
2
2
H SO /H O
2
4
2
2
HF/H O2
2
A
C
H
T
R
E
U
N
G
(NH )F/H O
4 2
2
2
A
C
H
T
R
E
U
N
G
(NH )HF /H O
4
2
2
[
[
a]
b]
The post-synthesis treatments of TS-1 were carried out as described in the experimental section.
Reaction conditions: benzene/sulfolane=2/8 w/w, H O /benzene molar ratio=0.1, TS-1 (or modified TS-1)/benzene=0,1
2
2
w/w, temperature=1008C. PH=phenol, CA=catechol, HY=hydroquinone, BQ=benzoquinone.
[
[
c]
Moles of phenol/moles of reacted H O ”100.
2
2
d]
Moles of product/moles of reacted benzene ” 100.
Adv. Synth. Catal. 2007, 349, 979 – 986
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
981

Daniele Bianchi et al.
FULL PAPERS
As reported in Figure 3 the extent of titanium ex-
The TOF measured for the oxidation of phenol
traction in the treatment was tuned using different also increased, but to a lesser extent, from 51 to 81
À1
F/Ti molar ratios (moles of fluoride, added as ammo- hour , according to the higher selectivity observed in
nium bifluoride, versus moles of titanium present in the reaction.
the given sample).
The catalyst performances were strongly dependent
on the temperature at which the treatment was car-
ried out. As shown in Figure 5, a mild and selective
removal of titanium, was achieved up to 808C, con-
versely, at 908C, the titanium sites were severely dam-
aged and a sharp decrease of the catalytic activity was
observed.
Figure 3. Post-synthesis modification of TS-1: extraction of
titanium. The treatment of TS-1 (TiO =3.6% by weight)
2
was carried out in aqueous solution using (NH )HF (F/Ti
4
2
molar ratio ranging from 0.5 to 8) and H O (H O /Ti molar
2
2
2
2
ratio=30) at 808C for 4 h. The F/Ti ratio is expressed as
moles of fluoride, added as ammonium bifluoride, versus
moles of titanium present in the given sample.
Figure 5. Effect of the post-synthesis treatment temperature
on the catalytic performances of TS-1B. The treatment of
TS-1 (initial Ti/Si molar ratio=0.028) was carried out in
aqueous solution using (NH )HF and H O (F/Ti molar
As reported in Figure 4, the selectivity of the cata-
lyst modified with ammonium bifluoride and hydro-
gen peroxide, now named TS-1B, increased at lower
residual titanium concentrations. After removal of
4
2
2
2
ratio=7 and H O /Ti molar ratio=30) at different tempera-
2
2
ture. All the samples showed a similar extent of titanium
leaching (75%) after 4 h, as determined by XRF analysis.
75% of the initial titanium, the selectivity to phenol
raised up to 94%, with formation of catechol (4%),
The catalytic properties of TS-1B (TiO =0.97%)
2
and hydroquinone (2%) as the only by-products, were compared with those of the starting TS-1
without any evidence for further oxidation reactions. (TiO =3.60%) and of an untreated TS-1 preparation
2
The activity of the catalyst was also affected by the with a similar titanium content (TiO =0.93%). As
2
treatment, and the TOF of the residual metal in the shown in Figure 6, the reaction profiles obtained for
À1
oxidation of benzene increased from 31 to 80 hour ,
as shown in Figure 2.
Figure 6. Effect of the catalyst modification on benzene oxi-
dation. Reaction conditions for benzene oxidation: benzene/
Figure 4. Effect titanium removal on the catalyst selectivity.
) Selectivity based on H O ; (*) selectivity based on ben-
zene. The reactions were carried out as previously described
in Table 2.
sulfolane=2/8 w/w, catalyst/benzene=0.1 w/w, tempera-
ture=1008C. An aqueous solution of H O (30% w/w) was
(
&
2
2
2
2
gradually added. Catalysts: (*) TS-1 (TiO =3.60%), (~)
2
TS-1 (TiO =0.93%), ( ) TS-1B (TiO =0.97%).
2
&
2
982
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 979 – 986

Oxidation of Benzene to Phenol with Hydrogen Peroxide
FULL PAPERS
benzene oxidation clearly demonstrate the superior sidual titanium levels are compared. The treated sam-
À1
performances of TS-1B in terms of selectivity, which ples show a new feature centered near 40,000 cm
became even more evident at higher benzene conver- and evident erosion of the tetrahedral Ti(IV) absorp-
sion.
tion. The new species is consistent with a Ti center
[17,18]
The selectivity of the oxidation was not affected by having an enlarged coordination sphere
but prob-
the temperature, in the range from 80 to 1008C while ably still isolated in the silicalite matrix while no ag-
it started to decrease above 1108C due to the unpro- gregated TiO species seem to be formed upon the
2
ductive decomposition of hydrogen peroxide to treatment.
À1
oxygen and water. Conversely, the catalytic activity of
The intensity of the 960 cm band in the FT-IR
TS-1B sharply increased with the temperature, the spectrum allows the quantitative estimation of the tet-
TOF measured in the oxidation of benzene rising rahedral Ti sites by means of a suitable titration
À1
[19]
from 15 to 80 h on operating at 80 and 1008C, re- curve. In Figure 8 we report the FT-IR spectra, in
spectively.
Some effort has been devoted to characterize the
new catalyst, TS-1B, in comparison with samples of
the parent TS-1, with different titanium loadings. The
characterization studies were carried out using as
starting material a well crystallized TS-1 sample with
a lower titanium content with respect to the industrial
catalyst (2.8% vs. 3.6%) bearing all the titanium
atoms in the tetrahedral form.
The parent TS-1, with 2.8% TiO , has been treated
2
with ammonium bifluoride and hydrogen peroxide,
stopping the reaction at different extents of titanium
extraction, in order to obtain samples of TS-1B con-
taining 2.4, 1.7, 1.4, 1.0, 0.9 and 0.8% of residual
TiO , respectively.
2
TS-1 and TS-1B samples have been characterized
by i) DRS UV-Vis; ii) FT-IR; and iii) argon adsorp- increases from top to bottom. All spectra have been normal-
tion.
Figure 8. IR spectra of TS-1 and TS-1B. Ti extracted fraction
À1
ized on overtone absorption in the 1500–2100 cm range.
It is well known that the UV-Vis spectrum of TS-1
[
16]
shows two important features. One is the band near
À1
À1
50,000 cm
ascribed to LMCT (ligand to metal the range 1000–700 cm , of the above mentioned
charge transfer) of the tetrahedral Ti(IV) center and samples of TS-1B at different extents of extraction to-
À1
the other, which occurs at 30,000 cm , is evidence of gether with the parent untreated sample. The label re-
an impurity due to a colloidal or crystalline anatase ported in the Figure identifies each sample according
À1
TiO phase.
to the decreasing of the 960 cm band. All the spec-
2
In Figure 7 the spectra of a pure TS-1 (2.8% TiO2) tra have been normalized on overtone absorption in
À1
and those of two samples of TS-1B with different re- the 1500–2100 cm range (not shown); the reliability
of the normalization procedure is confirmed by the
À1
good superimposition of the 800 cm band which
does not change its intensity and shape upon the
treatment (although a minor emerging of a fine com-
plex structure upon increasing the extent of extraction
is visible).
À1
In Figure 9 we plot the results of the 960 cm band
measurements of the TS-1B treated samples vs. titani-
um content (by XRF analysis) together with a series
of samples of TS-1 at different Ti loadings (directly
from synthesis) taken as reference. The two series fit
two good straight line correlations but even if for the
unextracted series (including the parent sample) the
straight line passes through the origin, this is not the
case for the line fitting the TS-1B extracted samples.
Figure 7. DRS UV-VIS spectra of TS-1 and TS-1B. TS-1
Figure 9 shows that, for the treated samples, the
(
Full line); TS-1B after extraction of 13% of TiO (dashed
2
À1
line); TS-1B after extraction of 48% of TiO (dot-dashed 960 cm band does not account for all analytical tita-
2
line).
nium. Therefore, a new species of titanium, which
Adv. Synth. Catal. 2007, 349, 979 – 986
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
983

Daniele Bianchi et al.
FULL PAPERS
À1
3
000 cm ) defects decrease with increasing extent of
extraction.
In this way a low concentration of active sites is ob-
tained in a structure with a very low concentration of
defects, the latter being a property commonly associ-
ated with a high hydrophobic character of the catalyt-
[
20]
ic materials.
The low defects and titanium concentration reflect
a very regular and highly ordered MFI structure. De-
spite the hydrolyzing treatment, leading to extraction
of up to 70% of Ti atoms, the crystalline structure
does not suffer any evident damage according to
XRD and TEM characterization (XRD patterns and
micrographs not shown for the sake of brevity).
Figure 9. Estimation of tetrahedral Ti site concentration. IR Moreover, an overall increase in surface homogeneity
À1
9
60 cm band measurement vs. analytical (XRF) Ti content
arises as a consequence of the treatment. In Ar ad-
sorption isotherms (see Figure 11), a substep in the
for ( ) samples at different Ti loading directly from the syn-
thesis and (~) TS-1B at different Ti residual concentration.
&
À1
does not contribute to the intensity of the 960 cm
band, and which may be tentatively associated to that
found in UV-Vis spectra, must be present. Surprising-
ly, the two straight lines in Figure 8 run quite parallel
indicating that the amount of the new species does
not increase with the Ti extraction.
It is known that, in as-synthesized TS-1, the
number of internal defects decreases upon increasing
[
13]
the titanium loading. Conversely, in the case of TS-
B samples the number of silanols decreases with re-
1
sidual titanium content. In Figure 10 we report the
Figure 11. Ar adsorption at 87 K. Isotherms in the relative
À2
pressure range 0–1”10 are reported for clarity, being the
0
further adsorption up to p/p =1 almost negligible.
micropores filling region becomes more and more evi-
dent as a higher fraction of titanium is extracted. The
parent TS-1 does not show this substep at all, the
transition being masked by a broadening of the step
0
along p/p . The treatment initially leads to a slight de-
crease in the adsorbed volume, probably reflecting a
minor smoothing of the external surface (a decrease
of micropore volume being unlikely). Then the sub-
step appears and moves towards higher partial pres-
sures as the strength of the treatment increases.
Figure 10. IR spectra in the OH stretching range. From top
to bottom: Ti content as well as silanols defects decrease
upon strengthening of the treatment.
[
21,22]
This is reminiscent of a well known effect,
typi-
cal of MFI structure. The features reported are very
similar to those observed for ZSM5 at different Al
content (Si/Al molar ratio from 20 to 10,000) and for
same normalized spectra of Figure 8 plotting the OH dealuminated ZSM5. During the step, argon adsorbed
À1
stretching range from 3800 to 3000 cm in which in micropores undergoes a fluid to crystalline-like
both isolated and defective silanols (titanols) are phase transition with a consequent density increase.
known to absorb. A quantitative determination of The presence and the sharpness of the transition step
defect concentration is not straightforward but it is reflect a high homogeneity of the surface which, in
evident from reported spectra that both isolated turn, points to a very low concentration of preferen-
À1
(
3800–3650 cm ) and hydrogen bonded (3650– tial and highly energetic adsorption sites.
984
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 979 – 986

Oxidation of Benzene to Phenol with Hydrogen Peroxide
FULL PAPERS
In summary, we observed the appearance of a new
titanium species, other than the former tetrahedral
one, upon the treatment of TS-1 with ammonium bi-
fluoride and hydrogen peroxide. This species is char-
acterized by a larger coordination sphere and seems
not to contribute to the characteristic vibrational fea-
HPLC analysis of the oxidation products were carried out
using an RP-18 column (250”4 mm, from Merck) and a
mixture of acetonitrile and 0.5 mM aqueous trifluoroacetic
acid (46/54 v/v) as the eluent, with a flow rate of
À1
1
mLmin . The UV detector was set at 212 nm (for phenol,
catechol and hydroquinone) and 260 nm (for 1,4-benzoqui-
none).
À1
ture at 960 cm in the IR spectrum of the TS-1. We
do not know the actual role of this species as catalytic
site but it cannot be excluded that it contributes, to-
gether with the modified surface property (lower de-
Preparation of TS-1B
Titanium silicalite TS-1 (10 g, containing 4.5 mmol of Ti, cor-
fectivity and the consequent enhanced hydrophobici- responding to 3.60% w/w of TiO ) was suspended in an
2
ty), to the improved catalytic performances reported aqueous solution (300 mL) containing NH HF (0.9 g,
4
2
for TS-1B.
15.8 mmol, corresponding to a F/Ti molar ratio=7) at 808C.
A 30% (w/w) aqueous solution of hydrogen peroxide
(
15.4 g, 135.5 mmol) was added in a single step and the sus-
pension was mechanically stirred for 4 h, keeping the tem-
perature at 808C. The mixture was then cooled at 258C and
the catalyst filtered off and washed with 100 mL of water.
The wet catalyst was dried at 1008C and calcined in air at
Conclusions
The selective oxidation of benzene to phenol with hy-
drogen peroxide was achieved by using a peculiar co-
solvent (sulfolane) and an improved zeolite catalyst
5
508C for 4 h. The TiO content of the modified titanium
2
silicalite (TS-1B) was 0.9% (w/w) as determined by XRF
(
TS-1B), obtained by modification of titanium silica- and ICP analysis, both the techniques giving fully compara-
lite with NH HF and H O .
ble results.
4
2
2
2
In a further development study, the reaction was
carried out under continuous operation, by using a
continuous stirred-tank reactor (CSTR) equipped
with filter, to maintain the catalyst inside the vessel.
Running the process at 1008C, only a negligible loss
of catalytic activity was observed after 100 h of con-
tinuous operation.
The simplicity of this one-step process, the high se-
lectivity achieved, and the demonstrated possibility of
a continuous operation process in CSTR reactors,
make this technology particularly attractive for indus-
trial application.
Preparation of TS-1B with Different Titanium
Contents
Titanium silicalite TS-1 (10 g, containing 4.5 mmol of Ti, cor-
responding to 3.60% w/w of TiO ) was suspended in an
2
aqueous solution (300 mL) containing different amounts of
NH HF , ranging from 0.06 g, (1.13 mmol) to 1.0 g
4
2
(
8
(
18 mmol) corresponding to an F/Ti molar ratio from 0.5 to
. A 30% (w/w) aqueous solution of hydrogen peroxide
15.4 g, 135.5 mmol) was added in a single step and the sus-
pension was mechanically stirred for 4 h, keeping the tem-
perature at 808C. The samples were isolated and analyzed
for the titanium content as described above. The titanium
content of the resulting catalysts (expressed as % of TiO by
2
weight) is reported in brackets for each F/Ti ratio used in
the treatment: F/Ti=0.5 (TiO =3.2%, corresponding to
2
Experimental Section
1
2
3
4
6
0% extraction), F/Ti=1 (TiO =2.8%, corresponding to
2
2% extraction), F/Ti=2 (TiO =2.2%, corresponding to
2
General Remarks
8% extraction), F/Ti=2.5 (TiO =1.8%, corresponding to
2
9% extraction), F/Ti=4 (TiO =1.3%, corresponding to
Titanium silicalites, with different titanium contents (3.60,
2
4% extraction), F/Ti=6 (TiO =1.0%, corresponding to
3.00, 2.80, 2.55, 2,00, 1,40 and 0.93% w/w of TiO ) were syn-
2
2
3]
[
71% extraction), F/Ti=7 (TiO =0.9%, corresponding to
thesized according to a literature procedure. All other
chemicals (Aldrich) were used as purchased.
2
7
5% extraction).
The loading of the free phenol molecule and the complex
phenol-sulfolane in the crystal elementary unit of TS-1 was
calculated using the software ꢁSorptionꢂ (Cerius 2, Accelrys
Inc., San Diego).
OtherPost-Synthesis Modifications of Titanium
Silicalite
FT-IR spectra have been recorded with a System 2000
Perkin–Elmer equipped with an MCT detector cooled at
Using the same procedure above described, TS-1 was treat-
ed at 808C for 4 h with the following methods:
7
7 K with liquid nitrogen on samples pressed in self support-
a) (NH )OAc (18 mmol): The resulting TiO content of
the modified catalyst was 3.60% (w/w).
4
2
À5
ed pellets, and outgassed at room temperature up to 10
Torr. Ar adsorption measurements at 87 K have been car-
ried out on an ASAP 2010 Micromeritics. DRS UV-Vis
spectra have been recorded with a Lambda 5 Perkin–Elmer
spectrophotometer equipped with reflectance sphere on
b) (NH4)HF2 (18 mmol): The resulting TiO content of
2
the modified catalyst was 3.20% (w/w).
c) H PO (18 mmol) and H O (67.5 mmol): The resulting
3
4
2
2
TiO content of the modified catalyst was 3.53% (w/w).
2
À5
powder samples previously outgassed at 1008C up to 10
d) H SO (18 mmol) and H O (67.5 mmol): The resulting
2
4
2
2
Torr.
TiO content of the modified catalyst was 3.24% (w/w).
2
Adv. Synth. Catal. 2007, 349, 979 – 986
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
985

Daniele Bianchi et al.
FULL PAPERS
e) HF (13.5 mmol) and H O (67.5 mmol): The resulting References
2
2
TiO content of the modified catalyst was 1.87% (w/w).
2
f) (NH )F (13.5 mmol) and H O (67.5 mmol): The result-
[1] G. I. Panov, A. S. Kharitonov, V. I. Sobolev, Appl.
Catal. A 1993, 98, 1–20.
[2] L. Balducci, D. Bianchi, R. Bortolo, R. D’Aloisio, M.
Ricci, R. Tassinari, R. Ungarelli, Angew. Chem. Int.
Ed. 2003, 42, 4937–4940.
4
2
2
ing TiO content of the modified catalyst was 2.02% (w/w).
2
g) (NH )HF (13.5 mmol) and H O (67.5 mmol): The re-
4
2
2
2
sulting TiO2 content of the modified catalyst was 1.84%
w/w).
(
[
3] M. Taramasso, G. Perego, B. Notari, (to Enichem), U.
Benzene Oxidation (Solvent Free)
S. Patent 4,410,501, 1983.
[
[
[
[
4] M. G. Clerici, P. Ingallina, J. Catal. 1993, 140, 71–83.
5] M. G. Clerici, Appl. Catal. 1991, 68, 249–261.
6] F. Maspero, U. Romano, J. Catal. 1994, 146, 476–482.
7] P. Roffia, G. Leofanti, A. Cesana, M. A. Mantegazza,
M. Padovan, G. Petrini, S. Tonti, P. Gervasutti, Stud.
Surf. Sci. Catal. 1990, 55, 43–52.
8] A. Esposito, M. Taramasso, C. Neri, (to Enichem), U.
S. Patent 4,396,783, 1983.
9] A. Bhaumik, P. Mukhrjee, R. Kumar, J. Catal. 1998,
Titanium silicalite (TS-1, 6 g) was suspended in a mixture of
benzene (40 g, 512.8 mmol) and water (200 g) in a glass re-
actor at 608C. Then, an aqueous solution of 30% w/w hy-
drogen peroxide (5.8 g, 51.3 mmol or 19.2 g, 170 mmol) was
fed stepwise within 1 hour and the suspension stirred for an
additional hour keeping the temperature at 608C. At the
end of the reaction, the hydrogen peroxide concentration
was determined by titration with potassium permanganate
and the oxidation products concentration was determined
by HPLC analysis. The amount of tars was determined after
separation of the products by silica gel chromatography, first
eluting phenol, catechol and hydroquinone with ethyl ace-
tate, then the tars with methanol.
[
[
1
78, 101–107.
10] R. Kumar, P. Mukhrjee, A. Bhaumik, Catal. Today
999, 49, 185–191.
[
[
[
1
11] R. S. Drago, B. Wayland, R. L. Carlson, J. Am. Chem.
Soc. 1963, 85, 3125–3128.
12] V. V. Prezhdo, O. V. Prezhdo, E. V. Vaschenko, J. Mol.
Struct. 1995, 356, 7–13.
Benzene Oxidation (Co-solvent)
Titanium silicalite (either TS-1 or TS-1B, 5 g) was suspended
in a solution of benzene (40 g, 512.8 mmol) and one of a co-
solvent: acetone, acetonitrile, methanol, tert-butyl alcohol,
acetic acid, dimethylformamide, N-methylpyrrolidone, dime-
thylsulfone or sulfolane (160 g). The suspension was charged
into a stainless steel autoclave, and pressurized at 5 bar with
nitrogen. Then, an aqueous solution of 30% w/w hydrogen
peroxide (5.8 g, 51.3 mmol) was fed stepwise within 1 hour
and the suspension stirred for an additional hour keeping
the temperature at 1008C. At the end of the reaction, the
hydrogen peroxide concentration was determined by titra-
tion with potassium permanganate and the oxidation prod-
ucts concentration was determined by HPLC analysis. The
amount of tars was determined after separation of the prod-
ucts by silica gel chromatography, first eluting phenol, cate-
chol and hydroquinone with ethyl acetate, then the tars with
methanol.
[13] C. Lamberti, S. Bordiga, A. Zecchina, G. Artioli, G.
Marra, G. Spanò, J. Am. Chem. Soc. 2001, 123, 2204–
2
212.
[
14] V. Bolis, S. Bordiga, C. Lamberti, A. Zecchina, A.
Carati, F. Rivetti, G. Spanò, G. Petrini, J. Catal. 1992,
1
37, 497–503.
15] W. Massa, G. Pausewang, Mater. Res. Bull. 1978, 13,
61–368.
[
3
[
[
16] G. N. Vayssilov, Catal. Rev. Sci. Eng. 1997, 39, 209–251.
17] G. Leofanti, F. Genoni, M. Padovan, G. Petrini, G.
Trezza, A. Zecchina, Studies in Surface Science and
Catalysis 1991, 62, 553–564.
[
18] I. Halasz, M. Agarwal, E. Senderov, B. Marcus, App.
Catal. A: 2003, 241, 167–184.
[
19] G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G.
Spanò, F. Rivetti, A. Zecchina, J. Am. Chem. Soc. 2001,
1
23, 11409–11419.
[
[
20] A. Corma, J. Catal. 2003, 216, 298–312.
21] P. L. Llewellyn, J. P. Coulomb, Y. Grillet, J. Patarin, H.
Lauter, H. Reichert, J. Rouquerol, Langmuir 1993, 9,
1846–1852.
Turn-Over Frequency (TOF) Determination
The turn over frequency of the titanium was expressed as
moles of reacted substrate/(moles of Ti”hour).
The TOFs were determined by carrying out the oxidation
reactions as described above, but adding the total amount of
hydrogen peroxide at the beginning of the reaction and
stopping it (by filtering off the catalyst and fast cooling) at
an oxidant conversion lower than 20%.
[22] J. C. Groen , L. A. A. Peffer, J. Perez-Ramꢃrez, Micro-
por. Mesopor. Mater. 2003, 60, 1–17.
986
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 979 – 986