Partial oxidation of 4-tert-butyltoluene catalyzed by homogeneous cobalt and cerium acetate catalysts in the Br-/H2O 2/acetic acid system: Insights into selectivity and mechanism

Article abstract of DOI:10.1002/chem.200700371

The partial oxidation of 4-fert-butyltoluene to 4-tert-butylbenzaldehyde by hydrogen peroxide in glacial acetic acid, catalyzed by bromide ions in combination with cobalt(II) acetate or cerium(III) acetate, has been studied in detail. Based on the observed differences in reaction rates and product distributions for the different catalysts, a reaction mechanism involving two independent pathways is proposed. After the initial formation of a benzylic radical species, either oxidation of this intermediate by the metal catalyst or reaction with bromine generated in situ occurs, depending on which catalyst is used. The first pathway leads to the exclusive formation of 4-tert- butylbenzaldehyde, whereas reaction of the radical intermediate with bromine leads to formation of the observed side products 4-tert-butylbenzyl bromide and its hydrolysis and solvolysis products 4-tertbutylbenzyl alcohol and 4-tert-butylbenzyl acetate, respectively. The cobalt(II) catalysts Co(OAc) ₂ and Co(acac)₂ are able to quickly oxidize the radical intermediate, thereby largely preventing the bromination reaction (i.e., side-product formation) from occurring, and yield the aldehyde product with 75-80% selectivity, In contrast, the cerium catalyst studied here exhibits an aldehyde selectivity of around 50% due to the competing bromination reaction. Addition of extra hydrogen peroxide leads to an increased product yield of 72% (cerium(III) acetate) or 58% (cobalt(II) acetate). Product inhibition and the presence of increasing amounts of water in the reaction mixture do not play a role in the observed low incremental yields.

Full text of DOI:10.1002/chem.200700371

FULL PAPER
DOI: 10.1002/chem.200700371
Partial Oxidation of 4-tert-Butyl
ACHTREUNG
ꢀ
and Cerium Acetate Catalysts in the Br /H O /Acetic Acid System: Insights
2
2
into Selectivity and Mechanism
A
H
R
U
G
Leon G. A. van de Water,* Arati Kaza, James K. Beattie, Anthony F. Masters, and
[
a]
Thomas Maschmeyer*
Abstract: The partial oxidation of 4-
tert-butyltoluene to 4-tert-butylbenzal-
dehyde by hydrogen peroxide in glacial
acetic acid, catalyzed by bromide ions
in combination with cobalt(II) acetate
used. The first pathway leads to the ex-
clusive formation of 4-tert-butylbenzal-
dehyde, whereas reaction of the radical
intermediate with bromine leads to for-
mation of the observed side products
4-tert-butylbenzyl bromide and its hy-
drolysis and solvolysis products 4-tert-
butylbenzyl alcohol and 4-tert-butyl-
benzyl acetate, respectively. The cob-
radical intermediate, thereby largely
preventing the bromination reaction
(i.e., side-product formation) from oc-
curring, and yield the aldehyde product
with 75–80% selectivity. In contrast,
the cerium catalyst studied here exhib-
its an aldehyde selectivity of around
50% due to the competing bromina-
tion reaction. Addition of extra hydro-
gen peroxide leads to an increased
or cerium(III) acetate, has been stud-
A
C
H
T
R
E
U
N
G
ied in detail. Based on the observed
differences in reaction rates and prod-
uct distributions for the different cata-
lysts, a reaction mechanism involving
two independent pathways is proposed.
After the initial formation of a benzylic
radical species, either oxidation of this
intermediate by the metal catalyst or
reaction with bromine generated in situ
occurs, depending on which catalyst is
A
H
R
U
G
A
H
R
U
A
H
R
U
G
product yield of 72% (cerium(III) ace-
A
H
R
U
G
2
tate) or 58% (cobalt(II) acetate).
Product inhibition and the presence of
increasing amounts of water in the re-
action mixture do not play a role in the
observed low incremental yields.
Keywords: cerium · cobalt · homo-
geneous catalysis · oxidation · reac-
tion mechanisms
[1,2]
Introduction
dioxygen or hydrogen peroxide.
Transition metal cata-
lyzed oxidation reactions have attracted much recent re-
search interest, as less harsh reaction conditions can be ap-
plied to reach conversions similar to those observed in unca-
talyzed oxidation reactions. Apart from this, side reactions
can be suppressed to some extent, as the relative reactivity
of a hydrocarbon substrate and that of its partial oxidation
products is known to change considerably in the presence of
The selective oxidation of hydrocarbons is a major challenge
in industrial chemistry. The low reactivity of hydrocarbon
substrates requires forcing reaction conditions, which often
lead to side reactions and overoxidation of the desired, par-
tially oxidized, product. Mild conditions are mandatory to
prevent overoxidation, but the resulting low conversions
leave large quantities of hydrocarbon substrate to be recy-
cled continuously in these processes. The ultimate synthetic
and industrial aim is high conversion without side reactions
under mild reaction conditions with benign oxidants such as
[3]
a catalyst. Several transition metal catalyzed autoxidation
reactions have been applied to industrial processes, for ex-
ample, the production of terephthalic acid, used in the pro-
duction of polyesters, from p-xylene, which is catalyzed by
[4]
cobalt(II) acetate bromide, with air as oxidant.
The cobalt(II) acetate catalyzed oxidation of hydrocar-
bons with hydrogen peroxide as oxidant, bromide ions as co-
catalyst, and acetic acid as solvent has been reported for the
partial oxidation of (substituted) toluenes to the correspond-
[
a] Dr. L. G. A. van de Water, A. Kaza, Dr. J. K. Beattie,
Dr. A. F. Masters, Prof. Dr. T. Maschmeyer
School of Chemistry
University of Sydney
Sydney, NSW 2006 (Australia)
Fax : (+61)293513329
[5–8]
ing benzaldehydes.
al interest, as 4-tert-butylbenzaldehyde is a key intermediate
This reaction is of significant industri-
E-mail: Vandew_l@chem.usyd.edu.au
th.maschmeyer@chem.usyd.edu.au
[8]
in the production of fragrances. The aldehyde product can
Chem. Eur. J. 2007, 13, 8037 – 8044
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8037

be obtained with high selectivity only at low conversion, be-
cause the benzoic acid overoxidation product is inevitably
Results and Discussion
formed as the aldehyde product accumulates. Cerium
A
H
R
U
G
Results of the catalytic oxidation of 4-tert-butyltoluene after
acetate also catalyzes the reaction, albeit with lower conver-
sion and aldehyde selectivity than with cobalt(II) acetate.
2 h are summarized in Table 1. In the first 40 min of the Ce-
[7]
A
H
R
U
G
No aldehyde product was detected in the absence of bro-
mide ions in this study, and this, together with the fact that
mide is formed rapidly and is the main reaction product
(see Figure 1). Formation of this product then ceases and,
after reaching a maximum of 0.74 mmol, the amount de-
creases to 0.62 mmol after 2 h. Benzaldehyde formation
starts from the moment of the first hydrogen peroxide addi-
tion at a constant, but somewhat slower, rate compared to
the formation of the benzylic bromide. Aldehyde formation
almost stops after the addition of the last hydrogen peroxide
aliquot at t=60 min. The other products of the reaction, 4-
tert-butylbenzyl alcohol and 4-tert-butylbenzyl acetate, are
detected merely in small quantities and only during the
second hour of the reaction. Interestingly, formation of
these products seems to be related to the decrease in the
amount of 4-tert-butylbenzyl bromide. Indeed, from t=
53 min onwards, the combined amounts of benzylic bromide,
alcohol, and acetate remain almost constant at around
0.85 mmol (Figure 1b). Formation of the aldehyde product
4-tert-butylbenzyl bromide was observed as a side product,
prompted those authors to suggest that the reaction occurs
[7]
via a benzylic bromide intermediate. However, the exact
role of cerium and bromide could not be established in that
study. The deactivation mechanism of the cobalt(II) cata-
lyzed oxidation of 4-tert-butyltoluene was studied by Amin
[8]
et al. Water, part of the aqueous hydrogen peroxide solu-
tion, was suggested to transform cobalt into a less active
form, thereby limiting the overall aldehyde yield. Moreover,
accumulation of the 4-tert-butylbenzaldehyde product was
found to prevent progress of the reaction beyond 50% con-
version. Another conclusion from this work was that 4-tert-
butylbenzyl bromide is not an intermediate in the reaction,
[6]
[7]
in contrast to the reports by Jones et al. and Auty et al.
The aim of the present study is to improve cobalt(II)- or
cerium(III)-catalyzed toluene
oxidation in terms of aldehyde
selectivity and conversion
ACHTREUNG
Table 1. Results after 2 h for the oxidation of 4-tert-butyltoluene with different catalysts in the presence of
through a detailed understand- KBr with hydrogen peroxide as oxidant and glacial acetic acid as solvent (T=708C).
ing of the reaction pathway. We Entry Catalyst
studied the differences between
Conversion
[%]
Aldehyde
selectivity [%]
Product yields [%]
aldehyde 5 bromide 7
2.3 3.9 0.4
alcohol 4
0.4
acetate 8
the cobalt(II)- and cerium(III)-
A
H
R
U
G
1
2
3
4
5
6
7
8
9
–
6.9
28
28
35
35
47
52
74
49
65
0.2
31
51
52
59
82
79
79
catalyzed reactions and at-
tempted to identify reaction in-
termediates and reaction path-
ways. Based on the catalytic
data presented here and those
reported previously by others,
we propose the existence of
two separate reaction pathways,
one of which leads to formation
of the aldehyde main product,
A
H
R
U
G
4
)
2
Ce
A
H
R
U
G
3
)
6
14
15
20
28
11
10
7.3
5.8
1.2
1.7
4.2
0.0
1.5
1.7
2.8
0.4
Ce
Ce
Ce
A
T
E
N
(OAc)
3
3
3
A
H
R
U
G
A
H
R
U
G
Co
Co
Co
Co
A
T
E
N
2
2
2
37
2.10.0
0. 0. 2
0.4
0.0
A
H
R
U
G
412.9
511 0.2
36
50
0.2
1
A
H
R
U
G
69
73
77
.8
A
H
R
U
G
2
2
0.3
0.3
2.10.3
2.5
1
0
A
H
R
U
G
0.2
0.0
A
H
R
U
G
100
0.0
0.0
[
a] 0.50 g of water added at the start of the reaction. [b] Double amount of hydrogen peroxide added (addition
and the other to the side prod- continued during second hour of reaction).
ucts of this reaction.
A
H
R
N
3 2 2
Figure 1. a) Product distribution for the oxidation of 4-tert-butyltoluene in acetic acid at 708C over Ce(OAc) by slow addition of H O during the first
6
(
4
0 min. Substrate:Ce:H
”), 4-tert-butylbenzyl alcohol (~), and 4-tert-butylbenzyl acetate (*). b) Side-product formation. Total amount (+) of 4-tert-butylbenzyl bromide (”),
-tert-butylbenzyl alcohol (~), and 4-tert-butylbenzyl acetate (*).
2 2
O
:KBr ratio=45:1:92:8. Mass balance (+), 4-tert-butyltoluene (^), 4-tert-butylbenzaldehyde (&), 4-tert-butylbenzyl bromide
8038
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8037 – 8044

Catalytic Partial Oxidation of 4-tert-Butyltoluene
FULL PAPER
appears to follow a course independent of the formation of
the bromide, alcohol, and acetate products. A similar pat-
tern of side-product formation was observed by Auty et al.
(rather than following an S-shaped curve) during the first
hour of the reaction, that is, when hydrogen peroxide was
added slowly to the reaction mixture. In contrast to the
[7]
in their study on the Ce
The result of the Co(OAc) -catalyzed reaction (entry 6,
2
A
H
R
U
G
cerium
alcohol, or acetate side products are observed in the experi-
ments with Co(OAc) and Co(acac) catalysts. The small
amount of benzylic bromide in the cobalt(II)-catalyzed reac-
tions, compared to the cerium(III)-catalyzed case, may origi-
A
H
R
U
G
3
A
H
R
U
Table 1) is depicted in Figure 2. The aldehyde selectivity is
A
H
R
U
G
ACHTREUNG
2
2
AHCTREUNG
nate from fast conversion of the intermediate benzylic bro-
mide to the benzaldehyde product in the cobalt(II)-cata-
lyzed reactions. We tested the hypothesis of a benzyl bro-
mide intermediate by using 4-tert-butylbenzyl bromide as
substrate in the reaction, with either Ce
A
H
R
U
G
A
C
H
T
R
E
U
N
G
3
or no catalyst at all. With Ce(OAc) , the aldehyde was pro-
AHCTREUNG
3
duced in 10% yield, which is lower than the 15% yield
when 4-tert-butyltoluene was used as substrate. The differ-
ence in aldehyde yield was even more pronounced when Co-
A
T
E
N
strate, compared to 37% when 4-tert-butyltoluene was used.
When the reaction was performed without a catalyst, the
yield of 4-tert-butylbenzaldehyde was still 14%. This blank
experiment shows that aldehyde formation from the benzyl-
ic bromide occurs to the same extent as in the reactions in-
volving a catalyst, that is, hydrolysis of the benzylic bromide
to the alcohol and subsequent oxidation to the aldehyde do
not require a metal catalyst. More importantly, the low con-
version of the bromide substrate to the aldehyde product in
Figure 2. Product distribution for the oxidation of 4-tert-butyltoluene in
acetic acid at 708C over Co
first 60 min. Substrate:Co:H
-tert-butyltoluene (^), 4-tert-butylbenzaldehyde (&), 4-tert-butylbenzyl
A
H
R
U
G
2 2 2
by slow addition of H O during the
2
O
2
4
bromide (”), 4-tert-butylbenzyl alcohol (~
), and 4-tert-butylbenzyl ace-
tate (*).
7
9% at a conversion of 47%, although the mass balance of
this reaction was not as good as in the case of Ce(OAc) .
Only 92% of the starting material could be recovered after
h, and the 8% unaccounted for was found to be the over-
the presence of Co(OAc) suggests that this species is not an
A
H
R
U
G
2
A
H
R
U
G
intermediate in the Co-catalyzed reaction, as more uncon-
verted bromide would have been found in the reaction mix-
ture when 4-tert-butyltoluene was used as substrate. In the
3
2
oxidation product 4-tert-butylbenzoic acid (vide infra). Alde-
hyde formation increased linearly with time during the first
hour, during which hydrogen peroxide was added. In con-
Ce(OAc) -catalyzed reaction, aldehyde formation via the
A
H
R
U
G
3
benzylic bromide intermediate cannot be ruled out com-
pletely, as a significant amount of the benzylic bromide is
observed in this reaction.
We propose that 4-tert-butylbenzaldehyde and 4-tert-bu-
tylbenzyl bromide are formed via two different mechanistic
pathways. The proposed reaction mechanism is depicted in
Scheme 1.
trast to the cerium
zylic bromide, alcohol, or acetate was observed in this
period. A similar activity was observed with Co(acac) : 73%
A
C
H
T
R
E
U
N
G
(III)-catalyzed reaction, hardly any ben-
AHCTREUNG
2
aldehyde selectivity at 49% conversion (Table 1, entry 9),
with 11% not detected. The conversion of 4-tert-butyl-
A
C
H
T
R
E
U
N
G
toluene in the presence of Mn
A
H
R
U
G
(
Table 1, entry 11); in fact this “catalyst” is an inhibitor of
Formation of benzylic radical intermediate 2: The first step
in the mechanism is the reaction of a bromine radical with a
methyl hydrogen atom of substrate 1 to form a benzylic rad-
ical and HBr [Eq. (3), i.e., reaction (i) in Scheme 1]. The
bromine radical is formed by reaction of bromide ions with
the reaction, as the blank experiment in which no catalyst
was used (entry 1) still resulted in a conversion of 6.9%.
Mechanism: The mechanism of aldehyde formation cata-
lyzed by Ce
III
IV
A
C
H
T
R
E
U
N
G
(OAc) was studied by Auty et al. 4-tert-Butyl-
Co or Ce [Eq. (2)] which, in turn, are formed by oxida-
3
II
III
benzyl bromide was claimed to be a reaction intermediate,
which is thought to be subsequently hydrolyzed to the ben-
zylic alcohol, which, in turn, is oxidized to the benzaldehyde
by bromine (formed in situ on reaction of hydrogen perox-
ide with bromide ions). The evidence for this reaction se-
quence was an S-shaped curve for aldehyde formation,
which indicated that the aldehyde originated from a precur-
sor species formed during the reaction. The fact that no re-
action occurred in the absence of bromide ions supported
this interpretation. Our observations with Ce
tion of Co or Ce by hydrogen peroxide [Eq. (1)].
II
III
þ
III
IV
2
Co =Ce þ 2 H þ H O ! 2 Co =Ce þ 2 H O
ð1Þ
ð2Þ
ð3Þ
2
2
2
[7]
III
IV
ꢀ
II
III
C
Co =Ce þ Br ! Co =Ce þ Br
C
C
ArꢀCH
þ Br ! Arꢀ CH þ HBr
3
2
[7]
Evidence for this reaction is the observed color change in
the case of Co(OAc) on addition of hydrogen peroxide,
from deep blue (tetrahedral Co species) to blue-green,
A
H
R
U
G
ACHTREUNG
3
2
II
lyst differ in that the aldehyde product forms gradually
Chem. Eur. J. 2007, 13, 8037 – 8044
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8039

T. Maschmeyer, L. G. A. van de Water et al.
Scheme 1. Proposed reaction mechanism for the oxidation of 4-tert-butyltoluene to 4-tert-butylbenzaldehyde over cobalt or cerium catalysts, with hydro-
gen peroxide as oxidant, bromide ions as cocatalyst, and acetic acid as solvent.
III
[9,10]
where the green color is indicative of Co species.
In
Pathway I: formation of 4-tert-butylbenzaldehyde (5) from
benzylic radical intermediate 2: The formation of 4-tert-bu-
tylbenzaldehyde (5) without involvement of benzyl bromide
intermediate 7 may involve oxidation of the benzylic radical
to benzylic carbocation 3 [Eq. (4), i.e., reaction (ii) in
the case of Ce(OAc) , a change from colorless to orange is
ACHTREUNG
3
IV
[7]
observed, which may be indicative of Ce species. This as-
signment is not unambiguous, however, as the color could
also be due to the presence of molecular bromine, formed
on reaction of bromide ions with hydrogen peroxide. The
UV/Vis spectra of a bromine solution and Ce salts in
acetic acid both showed absorption maxima in the 375–
III
IV
Scheme 1] by Co or Ce .
IV
C
III
IV
þ
II
III
Arꢀ CH þ Co =Ce ! Arꢀ CH þ Co =Ce
ð4Þ
2
2
400 nm region.
II
III
III
IV
Thus, the catalytic species Co or Ce undergoes cycles
of successive oxidation and reduction, by reducing hydrogen
peroxide and by oxidizing bromide ions, respectively. When
Both Co and Ce are strong oxidants, capable of oxidizing
the benzylic radical. Reaction of benzylic carbocation 3 with
water generates 4-tert-butylbenzyl alcohol (4), which is
quickly oxidized by hydrogen peroxide or bromine (formed
in situ on reaction of bromide ions with hydrogen peroxide)
to benzaldehyde 5. Other, more complicated reaction se-
quences from benzylic carbocation 3 to 4-tert-butylbenzyl al-
cohol (4) cannot be excluded, but we have no evidence of
such intermediates. Alternative routes may involve a benzyl-
ic hydroperoxide intermediate, formed on reaction of 3 with
hydrogen peroxide, as this type of species has been reported
IV
ceric ammonium nitrate [(NH ) Ce (NO ) ] was used, cata-
A
H
R
U
G
4
2
3 6
lytic results similar to those with Ce
A
H
R
U
G
that is, the oxidation state of the initial cerium catalyst can
be III or IV. Because of the suitability of Co /Co and
III
II
IV
III
III
II
Ce /Ce catalysts, a Mn /Mn catalyst was also tested in
the oxidation reaction. No substrate conversion was ob-
served at all with Mn
without any catalyst, which still showed a conversion of
.9%. This indicates reaction inhibition by the manganese
A
H
E
N
[12]
6
to give the corresponding benzylic alcohol.
Reaction of
species, probably by catalyzing peroxide decomposition and/
or by interfering with the free-radical species involved in the
reaction sequence, since manganese complexes are known
to act as radical scavengers. This result is therefore consis-
tent with the free-radical mechanism proposed in Scheme 1.
The addition of a radical scavenger, namely, 2,6-di-tert-butyl-
phenol, indeed resulted in complete inhibition of the reac-
tion.
benzylic carbocation 3 with bromide ions or acetate appears
not to play a significant role here, as formation of benzylic
bromide 7 and benzylic acetate 8 as side products appears
to occur independently of formation of benzaldehyde prod-
uct 5.
[11]
Pathway II: side-product formation from benzylic radical in-
termediate 2: In competition with the generation of 4-tert-
8040
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8037 – 8044

Catalytic Partial Oxidation of 4-tert-Butyltoluene
FULL PAPER
butylbenzaldehyde, the 4-tert-butylbenzyl radical (2) may
react with molecular bromine (formed on reaction of hydro-
gen peroxide with bromide ions) to give 4-tert-butylbenzyl
bromide (7) (reaction (vi) in Scheme 1). This side reaction
can ultimately lead to 4-tert-butylbenzaldehyde (5) after hy-
drolysis of bromide 7, followed by oxidation with hydrogen
peroxide or bromine, as described by Auty et al. The water
needed for hydrolysis is introduced into the reaction in the
aqueous hydrogen peroxide solution and is also formed on
catalytic decomposition of this reagent. Alternatively, 4-tert-
butylbenzyl bromide (7) may react with the acetic acid sol-
vent to give 4-tert-butylbenzyl acetate (8).
to proceed via Pathway I. Instead, 4-tert-butylbenzaldehyde
(5) formation involves initial formation of the 4-tert-butyl-
benzyl radical (2), which is formed on reaction of toluene
substrate 1 with bromine radicals, stemming from bromine,
or with a hydroxyl radical, which may be formed on decom-
position of the added hydrogen peroxide. The radical inter-
mediate then reacts with bromine to give benzylic bromide
7. The uncatalyzed reaction, which proceeds exclusively
through Pathway II, thus results in an aldehyde selectivity of
31% at 6.9% conversion.
Formation of benzylic bromide 7 in the Ce-catalyzed reac-
tion slows down significantly after the first 20 min, either
due to the lowering of the bromide concentration in the re-
action mixture or, possibly, due to the increasing amount of
water in the system. The possibility that the presence of
water increases 4-tert-butylbenzaldehyde selectivity was
[7]
Product distribution: The aldehyde selectivities in the
cobalt- and cerium-catalyzed reactions differ markedly, with
the cobalt catalysts showing higher aldehyde selectivity than
the cerium ones (75 vs. 50%, respectively). The aldehyde se-
lectivity in the experiment without a catalyst is only 31%,
much lower than in the catalyzed reactions. These selectivity
differences can be understood by considering the proposed
reaction scheme: benzylic radical intermediate 2 may be oxi-
tested by performing the Ce(OAc) -catalyzed reaction in the
A
H
R
U
G
3
presence of a small amount of water, added at the start of
the reaction (Table 1, entry 4). Indeed, the benzaldehyde
yield increased from 15 to 20%, while the combined yield
of the bromide, alcohol, and acetate side products was simi-
lar to that of the original experiment (entry 3): 14.3 versus
13.4%, respectively. The yield of benzylic bromide was re-
duced to 7.3%, versus 10% in the experiment without
added water, which is related to the increased amount of
benzylic alcohol product, as the extra amount of water is
thought to shift the equilibrium of reaction (vii) towards the
alcohol product. The change in aldehyde selectivity, from 52
to 59%, points to the existence of two different pathways,
one leading to 4-tert-butylbenzyl bromide (7) and its hydrol-
ysis and solvolysis products 4-tert-butylbenzyl alcohol (4)
and 4-tert-butylbenzyl acetate (8), and the other to 4-tert-bu-
tylbenzaldehyde (5).
III
IV
dized by Co or Ce to ultimately give benzaldehyde prod-
uct 5 (Pathway I), or react with bromine to give benzylic
bromide 7 and its associated byproducts (Pathway II). The
4
-tert-butylbenzyl radical (2) appears to react much faster
III
IV
with Co than with Ce . In fact, in the case of cobalt, reac-
tion of 2 with bromine according to Pathway II is almost
completely suppressed due to the fast oxidation reaction
(
Pathway I), as hardly any bromide, alcohol, or acetate
products are observed. With the cerium catalysts, in con-
trast, the relatively slow rate of Pathway I allows the side re-
action (Pathway II) to occur to a considerable extent, espe-
cially in the first 20 min of the reaction. Thus, aldehyde se-
lectivity is favored by fast oxidation of benzylic radical inter-
mediate 2, that is, by the presence of a strongly oxidizing
The effect of water addition on the Co(OAc) -catalyzed
A
H
R
U
G
2
reaction was also tested (Figure 3). In this case, the rate of
aldehyde formation was identical for both reactions. Howev-
er, in the experiment in which water was added, an in-
creased amount of the bromide product was observed in the
early stages of the reaction (see Figure 3). The amount of
this side product then decreases, and after 2 h the aldehyde
III
catalyst. The higher oxidizing power of Co compared to
IV
Ce is then the most likely explanation for the higher alde-
hyde selectivity in the cobalt-catalyzed reaction. In the un-
catalyzed reaction, the aldehyde selectivity is still 31%. In
the absence of a metal catalyst, the reaction is not expected
Figure 3. a) Product distribution for the oxidation of 4-tert-butyltoluene in acetic acid at 708C over Co
A
H
R
U
G
2
(OAc) without added water (a) and with addition
of 0.50 g water at the beginning of the reaction (b). Substrate:Co:H :KBr ratio=45:1:92:8. 4-tert-Butylbenzaldehyde (&), 4-tert-butylbenzyl bromide
2
O
2
(
”), 4-tert-butylbenzyl alcohol (~), and 4-tert-butylbenzyl acetate (*).
Chem. Eur. J. 2007, 13, 8037 – 8044
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8041

T. Maschmeyer, L. G. A. van de Water et al.
selectivity was 79%, just as in the original experiment. Thus,
in the cobalt(II) reaction, addition of water initially favors
in this case. This may explain the imperfect mass balance of
the Co(OAc) -catalyzed reaction (see Figure 2). As 4-tert-
A
H
R
U
G
2
Pathway II, the opposite to the cerium
A
H
R
U
G
butylbenzoic acid could not be analyzed directly by GC-MS,
a sample of the reaction mixture was derivatized with BF3/
MeOH to obtain the methyl ester of any 4-tert-butylbenzoic
acid present. The GC-MS analysis of the derivatized reac-
tion mixture showed the presence of a substantial amount of
methyl 4-tert-butylbenzoate, that is, overoxidation of 4-tert-
tion. Significantly, the reaction mixture to which no water
was added had a deep blue color, which slowly changed to
brownish pink as the reaction progressed. A much paler
blue was observed at the start of the reaction when water
was added, and the color change to pink occurred much
faster. These observations point to differences in coordina-
butylbenzaldehyde had indeed occurred in the Co
catalyzed reaction.
A
C
H
T
R
E
U
N
G
(OAc) -
2
II
tion environment for the Co catalyst, whereby a blue color
II
is indicative of tetrahedral Co species, and a pink color of
II [9,10]
octahedral Co .
The different effect of adding water at
Continued peroxide addition: One of the questions remain-
ing is whether the reaction, and in particular the 4-tert-butyl-
benzaldehyde yield, can be improved by the addition of
the start of the cerium- and cobalt-catalyzed reactions illus-
trates that the observed product distributions are the result
of a delicate balance between the two competing reaction
pathways presented in Scheme 1. Addition of water changes
the relative rates of the two routes in different ways for the
two catalysts, and results in increased aldehyde selectivity in
the cerium-catalyzed reaction and a decreased aldehyde se-
lectivity in the early stage of the cobalt-catalyzed case.
more hydrogen peroxide. In a reaction with Ce(OAc) as
A
H
R
U
G
3
catalyst, addition of hydrogen peroxide was continued
during the second hour of the reaction (Table 1, entry 5), so
the total amount of hydrogen peroxide was twice that used
in the standard experiment (Table 1, entry 3). The benzalde-
hyde yield was 28%, compared to 15% in the original ex-
periment. Higher aldehyde selectivity was observed in the
experiment in which peroxide addition was continued (82 vs
52%). This is thought to be due to the extra amount of
water present in the reaction (in the form of aqueous hydro-
gen peroxide solution and its decomposition product), which
shifts equilibrium (vii) in Scheme 1towards benzylic alcohol
4 and thereby reduces the yield of benzylic bromide 7. Alco-
hol 4 is subsequently oxidized by the extra equivalent of
peroxide to give 4-tert-butylbenzaldehyde (5). In the reac-
tion in which addition of hydrogen peroxide was continued,
2.5% of starting material was unaccounted for after the re-
action, compared to 2.7% in the standard experiment. This
may be ascribed to overoxidation of the aldehyde product
to the benzoic acid derivative. If this mass-balance differ-
ence is added to the aldehyde yield in both cases, the com-
bined product yield (i.e., 4-tert-butylbenzaldehyde plus 4-
tert-butylbenzoic acid) on addition of double the amount of
hydrogen peroxide would be 31%, an increase of 72% com-
pared to the 18% obtained in the standard experiment.
The assumption that the differences in mass balance are
due to overoxidation of the aldehyde to the benzoic acid is
valid, as 1) the methyl ester of this product has been ob-
served with GC-MS after derivatization of the reaction mix-
Overoxidation: Selective oxidation reactions inherently
suffer from overoxidation, as the desired (partially oxidized)
product is generally more easily oxidized than the substrate.
In the oxidation of toluenes to benzaldehydes, formation of
the corresponding benzoic acids (reaction (v) in Scheme 1)
may reduce the aldehyde yield. The good mass balance (see
Figure 1) of the cerium-catalyzed reaction suggests that 4-
tert-butylbenzoic acid (6) is not formed in this case. The rel-
ative oxidizabilities of methyl-substituted aromatics and the
corresponding benzylic alcohols and benzaldehydes depend
on the oxidant and the type of catalyst. For example, in the
uncatalyzed autoxidation of methyl-substituted aromatics,
the reactivity of the resulting benzaldehyde is much higher
than that of the corresponding benzylic alcohol. In con-
ꢀ
trast, in the Co/Mn/Br -catalyzed autoxidation reaction, the
reactivity of the alcohol is higher than that of the aldehyde,
which means that no overoxidation of the aldehyde will
occur in the presence of the alcohol, that is, the aerobic oxi-
dation of benzyl alcohol yields benzaldehyde exclusively as
long as benzyl alcohol is still present in the system. Only
after complete conversion of benzyl alcohol is formation of
[3]
benzoic acid observed. The hydrogen peroxide/bromide
oxidation system (i.e., in situ generation of bromine) has
tures with BF /MeOH; 2) addition of hydrogen peroxide to
3
[13]
been studied in detail by Amati et al. Their investigation
revealed that benzylic alcohols are more easily oxidized by
hydrogen peroxide/bromide than the corresponding benzal-
dehydes. This is in agreement with the current study: no
overoxidation in the cerium-catalyzed reaction is observed,
as a small amount of 4-tert-butylbenzyl alcohol is always
present in the system due to hydrolysis of the 4-tert-butyl-
benzyl bromide byproduct. As this alcohol is more easily
oxidized than 4-tert-butylbenzaldehyde, overoxidation of the
a mixture containing bromide ions and 4-tert-butylbenzalde-
hyde is expected to result in 4-tert-butylbenzoic acid, as no
easily oxidizable 4-tert-butylbenzyl alcohol is present. The
aldehyde is then the most likely species to be oxidized, as it
is more reactive than the 4-tert-butyltoluene substrate.
When a second amount of hydrogen peroxide was added
during the second hour of the Co(OAc) -catalyzed reaction
A
H
R
U
G
2
(Table 1, entry 8), the aldehyde yield was 51% (see
Figure 4). This is an increase of only 38% from the 37% al-
dehyde yield measured in the standard experiment. Howev-
er, when the decrease in mass balance in both reactions is
ascribed to overoxidized 4-tert-butylbenzaldehyde, values of
71versus 45% (aldehyde yield plus unaccounted-for prod-
aldehyde is not expected to occur. In the Co(OAc) -cata-
A
H
R
U
G
2
lyzed reaction, hardly any 4-tert-butylbenzyl bromide, and
hence no 4-tert-butylbenzyl alcohol, is formed, and this sug-
gests that 4-tert-butylbenzaldehyde is prone to overoxidation
8042
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8037 – 8044

Catalytic Partial Oxidation of 4-tert-Butyltoluene
FULL PAPER
(
Table 1, entry 6 and Table 2, entry 2). The combined
amount of 4-tert-butylbenzaldehyde and 4-tert-butylbenzoic
acid formed during the reaction was 2.76 mmol, very similar
to the value observed for the original experiment
(
2.74 mmol). This suggests that product inhibition does not
play a prominent role in this reaction. In the experiments in
which 4-tert-butylbenzaldehyde was added at the start of the
reaction, the degree of overoxidation is much higher than in
the standard experiment. This is as expected, as oxidation of
the aldehyde may occur from the first addition of hydrogen
peroxide onwards. To investigate the effect of a smaller
amount of available substrate and the presence of some
product on the rate of substrate conversion, a test reaction
was conducted in which 2.64 mmol of 4-tert-butyltoluene
was mixed with 2.35 mmol of 4-tert-butylbenzaldehyde at
the start of the reaction (Table 2, entry 3). This starting
point is comparable to the situation after one hour in the re-
Figure 4. Product distribution for the oxidation of 4-tert-butyltoluene in
acetic acid at 708C over Co
A
H
R
U
G
2 2 2
(OAc) by slow addition of H O over
1
20 min. Substrate:Co:H :KBr ratio=45:1:184:8. Mass balance (+), 4-
2 2
O
tert-butyltoluene (^), 4-tert-butylbenzaldehyde (&), 4-tert-butylbenzyl
bromide (”), 4-tert-butylbenzyl alcohol (~), and 4-tert-butylbenzyl ace-
tate (*).
uct for 2 and 1h of hydrogen
peroxide addition, respectively)
Table 2. Results for the oxidation of 4-tert-butyltoluene over Co
peroxide in glacial acetic acid at 708C with different compositions of the reaction mixture.
A
H
R
U
G
2
(OAc) in the presence of KBr and hydrogen
are obtained, which corre- Entry Catalyst
Amount of 1 at Amount of 5 added Conversion Yield of aldehyde 5+
t=0 [mmol]
at t=0 [mmol]
of 1 [%]
overoxidized product [mmol (%)]
sponds to an incremental prod-
1
2
3
4
5
Co
A
T
E
N
2
2
2
2
2
6.13
6.07
2.64
2.88
–
48
47
87
2.76 (45)
2.74 (45)
2.20 (84)
uct yield of 58% on continued
addition of hydrogen peroxide
during the second hour. The al-
dehyde selectivity drops from
AHCTREUNG
A
H
R
U
G
2.35
[a]
[a]
[a]
[a]
[a]
A
T
E
G
3.36
2.16
51
1.71 (51)
A
H
R
U
G
2.82
–
80
2.16 (77)
79
to 69% if double the
[a] Double amount of hydrogen peroxide added (addition continued during second hour of reaction).
amount of hydrogen peroxide is Amounts of 1 and 5 are those present in the reaction mixture at t=65 min. Conversion and yield are calculat-
added. Close inspection of the ed over the second hour of the reaction.
decrease in amount of substrate
in Figure 4 reveals that sub-
strate conversion in the second hour is significantly lower
than in the first hour: indeed, 4-tert-butyltoluene conversion
action in which addition of hydrogen peroxide was contin-
ued in the second hour of the reaction (Table 2, entry 4).
After 2 h 2.30 mmol (87%) of substrate was converted,
which is more than the 1.73 mmol (51%) of remaining sub-
strate that was converted in the second hour of the experi-
ment involving the double amount of hydrogen peroxide
(Table 1, entry 8). The amount of product formed in the test
reaction was 2.20 mmol (amount of aldehyde plus benzoic
acid products), considerably more than in the second hour
of the reaction when the double amount of hydrogen perox-
ide was used (1.71 mmol). Finally, an experiment involving
2.82 mmol of 4-tert-butyltoluene substrate resulted in the
formation of 2.16 mmol of product (77%, Table 2, entry 5),
comparable to the 2.20 mmol when the aldehyde was added
at the start of the reaction. Once again, this result suggests
that product inhibition cannot explain the lower conversions
observed during the latter stages of the reaction.
in the case of Co(OAc) in the second hour was 28%, much
A
C
H
T
R
E
U
N
G
2
less than the 45% converted during the first hour.
These experiments indicate that adding double the
amount of hydrogen peroxide to the reaction results in 72%
(Ce
A
C
H
T
R
E
U
N
G
(OAc) case) or 58% (Co(OAc) ) more product, if the
A
H
R
U
G
3
2
unaccounted-for amount of material is ascribed to overoxi-
dized product (the benzaldehyde). Thus, the observed lower
rate of product formation on addition of the additional
amount of hydrogen peroxide during the second hour of the
reaction cannot be explained sufficiently by the occurrence
of overoxidation and other side reactions, but is due to
lower degrees of substrate conversion. The lower conversion
in the second hour may be due to the decreased substrate
concentration, which has dropped from 6.15 mmol at the
start of the reaction to 3.36 mmol after 1h in the Co (OAc)
A
H
R
U
G
2
reaction (entry 8). Alternatively, product inhibition may
play a role.
Conclusion
Product inhibition: The possibility of product inhibition was
assessed in a reaction in which 2.88 mmol of 4-tert-butylben-
zaldehyde was added to 6.13 mmol of 4-tert-butyltoluene at
the start of the reaction (Table 2, entry 1). The results can
The mechanism of the partial oxidation of 4-tert-butylto-
II
III
ꢀ
luene to the corresponding aldehyde in the Co or Ce /Br /
hydrogen peroxide/acetic acid system has been studied in
detail. The reaction seems to proceed by two different path-
be compared with the original Co(OAc) experiment
A
H
R
U
G
2
Chem. Eur. J. 2007, 13, 8037 – 8044
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8043

T. Maschmeyer, L. G. A. van de Water et al.
ways, one of which leads to exclusive formation of 4-tert-bu-
tylbenzaldehyde, and the other to formation of the benzylic
bromide and its hydrolysis, solvolysis, and oxidation prod-
ucts. The role of the metal catalyst is to oxidize the inter-
mediate 4-tert-butylbenzyl radical species, whereby the rate
of this reaction relative to the side reaction determines the
product distribution in the reaction mixture. Cobalt catalysts
exhibit higher aldehyde selectivities due to their ability to
oxidize the benzylic radical intermediate before it can react
with bromine to generate the side products.
nese(II) acetate tetrahydrate (Merck), and anhydrous cobalt(II) acetate
(
Aldrich) were obtained from commercial sources and used as received.
Cobalt(II) acetylacetonate (Co(acac) ) was prepared by following a liter-
A
H
R
U
G
2
[14]
ature procedure. 4-tert-Butylbenzyl acetate was prepared by esterifica-
tion of 4-tert-butylbenzyl alcohol with acetic acid followed by purification
by column chromatography (silica, hexane:ethyl acetate 12:1).
Catalytic testing: In a typical oxidation reaction, 5.4 mmol of 4-tert-butyl-
toluene, 1.0 mmol of KBr, 0.12 mmol of cerium or cobalt catalyst, and
chlorobenzene (internal standard for GC analysis) were mixed with
1
3
0 mL of glacial acetic acid, and the mixture was heated to 708C. A
0 wt% aqueous solution of hydrogen peroxide (11.0 mmol) was added
to this mixture in six equal aliquots over 1h, and the mixture was stirred
Overoxidation of 4-tert-butylbenzaldehyde is prevented
when a small amount of 4-tert-butylbenzyl alcohol is present
in the reaction mixture, as this species is more easily oxi-
dized than the aldehyde. Due to the presence of small quan-
tities of the alcohol in the cerium-catalyzed reaction, no
overoxidation to the benzoic acid product is observed in
that case. Overoxidation does occur to a significant extent in
the cobalt-catalyzed reactions due to the high selectivity for
the aldehyde product and hence the absence of the alcohol
product.
and heated for another hour. The overall molar ratio 4-tert-butyltolue-
n+
2 2
ne:M :H O :KBr was 45:1:92:8 in all cases, similar to that in the study
[
7]
by Auty et al. Samples for GC-MS analysis were taken 5 min after each
addition of hydrogen peroxide and after 2 h. The samples were neutral-
ized with aqueous NaOH, extracted with ethyl acetate and toluene, and
subsequently analyzed on a Shimadzu GC-MS [QP 2010, using a fused
silica Rtx-5Sil MS column (30 m, 0.25 mm i.d.]. The total flow was
ꢀ
1
6
9.3 mLmin and a split ratio of 50 was applied. The initial column tem-
perature was 908C (hold time: 3 min), which was increased to 2008C
ꢀ1
(
ramp rate: 208Cmin ) and held at that temperature for 1.5 min. In
some experiments, water or 4-tert-butylbenzaldehyde was added at the
start of the reaction, and in some other experiments twice the amount of
hydrogen peroxide was added by addition of a further six equal aliquots
during the second hour of the reaction.
The addition of water at the start of the reaction in the
case of Co(OAc) initially leads to decreased aldehyde se-
A
C
H
T
R
E
U
N
G
2
lectivity. After 2 h, however, conversion and aldehyde selec-
tivity are very similar to those of the experiment without
added water. Doubling the amount of added hydrogen per-
[
[
[
[
1] R. A. Sheldon, Green Chem. 2005, 7, 267.
2] M. Ziolek, Catal. Today 2004, 90, 145.
3] W. Partenheimer, Adv. Synth. Catal. 2006, 348, 559.
4] W. Partenheimer, Catalysis of Organic Reactions, Vol. 62 (Eds.:
M. G. Scaros, M. L. Prunier), Marcel Dekker, New York, 1995,
p. 307.
oxide leads to an increase in product yield of 72 (Ce
A
H
R
U
G
and 58% (Co(OAc) ). The low incremental yield is not due
ACHTREUNG
2
to the presence of increasing amounts of water in the reac-
tion mixture, nor due to the buildup of the aldehyde reac-
tion product. Efforts to reduce the amount of water and al-
dehyde product in the reaction mixture are thus not expect-
ed to result in higher product yields. A possible explanation
for the lower rate of product formation on continued perox-
ide addition is the formation of small amounts of side prod-
ucts such as dialkyl phenol species (not detected in this
study) that act as radical scavengers. The ability of this type
of phenolic compound to inhibit the reaction completely has
been illustrated in this study.
[
5] G. M. Dugmore, G. J. Powels, B. Zeelie, J. Mol. Catal. A: Chem.
1
995, 99, 1 .
[
6] C. W. Jones, A. Hackett, I. Pattinson, A. Johnstone, S. L. Wilson, J.
Chem. Res. Miniprint 1996, 2501.
[7] K. Auty, B. C. Gilbert, C. B. Thomas, S. W. Brown, C. W. Jones,
W. R. Sanderson, J. Mol. Catal. A: Chem. 1997, 117, 279.
[
[
8] A. A. Amin, J. K. Beattie, Org. Process Res. Dev. 2003, 7, 879.
9] G. H. Jones, J. Chem. Soc. Chem. Commun. 1979, 536.
[
[
10] G. H. Jones, J. Chem. Res. Synop. 1981, 228.
11] O. Vajragupta, P. Boonchoong, L. J. Berliner, Free Radical Res. 2004,
38, 303.
12] B. B. Wentzel, M. P. J. Donners, P. L. Alsters, M. C. Feiters, R. J. M.
Nolte, Tetrahedron 2000, 56, 7797.
13] A. Amati, G. Dosualdo, L. Zhao, A. Bravo, F. Fontana, F. Minisci,
H.-R. Bjorsvik, Org. Process Res. Dev. 1998, 2, 261.
14] R. G. Charles, M. A. Pawlikowski, J. Phys. Chem. 1958, 62, 440.
[
[
[
Experimental Section
Starting materials: 4-tert-Butyltoluene, 4-tert-butylbenzyl bromide, 4-tert-
butylbenzyl alcohol, 4-tert-butylbenzaldehyde (Fluka), hydrogen peroxide
Received: March 5, 2007
Revised: May 4, 2007
(
30 wt% aqueous solution, Univar), potassium bromide (Merck), cerium-
A
C
H
T
R
E
U
N
G
(III) acetate (Aldrich), cerium(IV) ammonium nitrate (Merck), manga-
Published online: July 10, 2007
8044
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 8037 – 8044