The complex synergy of water in the metal/bromide autoxidation of hydrocarbons caused by benzylic bromide formation

Article abstract of DOI:10.1002/adsc.200303187

One of the most active and selective catalysts in homogeneous liquid phase oxidation using molecular oxygen (O₂) is a mixture of cobalt, manganese and bromide salts in acetic acid. It has been used to produce hundreds of different carboxylic acids in high yield and purity including the commercial production of terephthalic acid from p-xylene. Water is normally a by-product in these reactions and it is shown here that its concentration is an important reaction variable. In anhydrous acetic acid, with reagents with sufficiently strong electron-withdrawing substitutents (toluene, 4-carboxytoluene, 4-chlorotoluene), all of the active bromide becomes inactive via benzylic bromide formation. The Co/Mn/ Br catalyst is therefore converted to a Co/Mn catalyst which is dubbed 'catalyst failure' because of its undesirable characteristics of lower activity, decreased selectivity especially towards over-oxidation and color formation. For 4-chlorotoluene, increasing the water concentration to 5 weight % initially decreases the rate of reaction but eventually is more active and selective because the oxidation and hydrolysis of the benzylic bromide allows for sufficient active catalytic bromide. It is shown that benzylic bromides do not 'promote' the reaction and that both oxidation and solvolysis of the benzylic bromide occurs during autoxidation. During polymethylbenzene oxidation, benzylic bromide formation occurs only with the most reactive methyl group. The complex factors during metal/bromide autoxidation -some favored by increased water concentration and others detrimental - are outlined.

Full text of DOI:10.1002/adsc.200303187

FULL PAPERS
The Complex Synergy of Water in the Metal/Bromide
Autoxidation of Hydrocarbons Caused by Benzylic
Bromide Formation
Walt Partenheimer
Central Research and Development, E. I. DuPont de Nemours & Co., Inc., Experimental Station, Wilmington,
Delaware 19880 ± 0328, USA
Fax: ( ‡ 1)-302-695-8347, e-mail: Walter.Partenheimer@usa.dupont.com
Received: October 24, 2003; Accepted: January 18, 2004
Abstract: One of the most active and selective and color formation. For 4-chlorotoluene, increasing
catalysts in homogeneous liquid phase oxidation the water concentration to 5 weight % initially
using molecular oxygen (O ) is a mixture of cobalt, decreases the rate of reaction but eventually is more
2
manganese and bromide salts in acetic acid. It has active and selective because the oxidation and
been used to produce hundreds of different carbox- hydrolysis of the benzylic bromide allows for suffi-
ylic acids in high yield and purity including the cient active catalytic bromide. It is shown that
commercial production of terephthalic acid from p- benzylic bromides do not −promote× the reaction
xylene. Water is normally a by-product in these and that both oxidation and solvolysis of the benzylic
reactions and it is shown here that its concentration bromide occurs during autoxidation. During polyme-
is an important reaction variable. In anhydrous acetic thylbenzene oxidation, benzylic bromide formation
acid, with reagents with sufficiently strong electron- occurs only with the most reactive methyl group. The
withdrawing substitutents (toluene, 4-carboxytoluene, complex factors during metal/bromide autoxidation ±
4-chlorotoluene), all of the active bromide becomes some favored by increased water concentration and
inactive via benzylic bromide formation. The Co/Mn/ others detrimental ± are outlined.
Br catalyst is therefore converted to a Co/Mn catalyst
which is dubbed −catalyst failure× because of its Keywords: autoxidation; bromide; catalytic hydro-
undesirable characteristics of lower activity, de- carbon oxidation; CÀH activation; cobalt; manganese;
creased selectivity especially towards over-oxidation oxygen
Introduction
In homogeneous liquid-phase reactions, invariably the
hydrocarbon and solvent −overoxidize× to produce
The high reactivity and selectivity of metal/bromide carbon dioxide, carbon monoxide and additional wa-
[
1]
catalysts for the homogeneous liquid-phase air oxida- ter. In metal-catalyzed autoxidations there are reports
tion of hydrocarbons has been industrially exploited in which increasing water concentration decreases the
[
1±3]
since the 1960 s for the large-scale production of overall rate of reaction
and others in which the rate
aromatic acids, such as terephthalic acid. Literally initially increases with concentration then decreases.[
hundreds of different hydrocarbons have been shown In metal/bromide systems, the decrease in rate is very
to produce oxygenates by this method and at least 30 general since it occurs for a number of different
different combinations of metals with bromide have reactants (p-xylene, m-xylene, 2,2-bis(tolyl)hexafluoro-
4±6]
[
2]
been shown to have often different and unique propane, and p-nitrotoluene ), and different catalysts
[
1,9,18]
catalytic behavior.
This has been summarized in (Mn/Br, Co/Br, Ni/Mn/Br, Co/Mn/Br, Mn/Zr/Br, Ni/
detail.[
1]
Mn/Zr/Br, Co/Mn/Br/Zr).
[1]
Water forms as the product of the autoxidation of
The general statement that high water concentrations
hydrocarbons. For methyl aromatic compounds one during autoxidation in acetic acid solvent are detrimen-
mole of water is produced for every methyl group tal is not true. As will be described, there are reagents of
oxidized (where Ar is an aromatic ring such as benzene, specific reactivity where higher water concentrations
naphthalene, biphenyl, etc.):
are necessary to produce high yields such as with 4-
chlorotoluene. The effect of water is quite subtle. There
are more phenomena ± some which are favored by low
water concentration and others that are favored by high
ꢀ
1†
Adv. Synth. Catal. 2004, 346, 297 ± 306
DOI:10.1002/adsc.200303187
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
297

FULL PAPERS
W. Partenheimer
Figure 1. General description of the autoxidation of toluene
or substituted toluenes using metal/bromide catalysts.
Figure 2. Effect of water on the Co/Mn/Br-catalyzed aerobic
oxidation of p-carboxytoluene (p-toluic acid) and 4-chloro-
tolune.
concentrations ± which are beyond the scope of a single
paper. The phenomena are briefly mentioned in the Commercial processes operate at approximately 5 ± 20
conclusion of this paper and will be more fully detailed weight % water in the reactors.
in a subsequent paper.
The presence of benzylic bromides during the autox-
idation of hydrocarbons is well-known.[ They form Comparison of the Autoxidation of Chlorotoluene in
with a variety of catalysts (Mn/Br, Ce/Br, Co/Br, Co/Ce/ 0 and 5 weight % Water in Acetic Acid, Experiments
Br, Co/Mn/Br, Co/Mn/Ce/Br, Co/Mn/Zr/Br) and a 3 and 4
1]
[
1,7,8,9]
variety of different alkylbenzenes.
The mechanism
of benzylic bromide disappearance during autoxidation Figure 3 illustrates the effect of the batch reaction of 4-
[
7,8]
has generally been ascribed to oxidation
although chlorotoluene with a Co/Mn/Br catalyst. At a initial
hydrolysis of the benzylic bromide to benzyl alcohol and water concentration of 5 weight %, the rate of oxygen
[
9]
hydrobromic acid has also been proposed. The hy- uptake as a function of time is typical ± the initial activity
drolysis of alkyl bromides to produce hydrobromic acid is due to the oxidation of the methyl group and the latter
[
10]
[11]
from bromoform,
acetylene tetrabromide,
and maximum to formation of the more reactive benzalde-
[
12]
even bromobenzene are commonly used as sources hyde. What was surprising was the oxygen uptake when
of ionic bromide in metal/bromide-catalyzed reactions anhydrous acetic acid was initially used. The initial rate
including the industrial manufacture of terephthalic of oxygen uptake was higher in anhydrous acetic acid ±
acid from water.
as expected since 5% water in acetic acid depresses the
rate of oxidation (Figure 2) ± but then quickly decreased
to a rate less than in 5% water. From the total oxygen
uptake it was obvious that the reaction was not
complete. This is confirmed with the conversion of the
Results and Discussion
4
-chlorotoluene, see Figure 4. The initial rate of con-
The Effect of Water on 4-Chlorotoluene and
version is higher in anhydrous acetic acid but the final
conversion is only 60% ± in contrast to that in 5% water,
where the final conversion is higher than 90%. The rate
4-Carboxytoluene (p-Toluic Acid) Oxidations
The sequence of reactions that occurs during the metal/ of formation of the intermediate 4-chlorobenzaldehyde
bromide autoxidation of toluenes is given in Figure 1. is also consistent with still substantial amounts of the
The effect of water concentration on the maximum rate unreacted benzaldehyde at the end of the experiment,
of oxygen uptake of the two reagents used in this study is see Figure 5. (The 4-Cl-benzoic acid yield is not reported
given in Figure 2 and experiments 1and 2 in Table 1.
because the acid is quite insoluble at room temperature
These reactions are complicated by benzylic bromides and only the soluble compounds were measured by GC).
formation, see below. The effect of water is substantial The benzaldehyde yield in the presence of water is
with the rate essentially being reduced by one-half when slightly higher than when starting with anhydrous acetic
the water is increased from 0 to 5 weight % and then acid.
again when increased from 5 to 10 weight %. A 2 M
A striking observation that indicates that something is
solution of toluene completely oxidized to benzoic acid drastically different in anhydrous acetic acid is that the
will generate about 4 weight % water in acetic acid. color of the solution was becoming opaque in anhydrous
298
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 297 ± 306

Synergy of Water in the Metal/Bromide Autoxidation of Hydrocarbons
FULL PAPERS
Table 1. Initial conditions of the aerobic oxidation experiments.
#
[Reagent], M
Catalyst
Catalyst M
[H O], wt %
2
1
2
3
5
p-toluic acid, 0.75
Co/Mn/Br
Co/Mn/Br
Co/Mn/Br
Co/Mn/Br
0.01/0.01/0.02
0.01/0.01/0.02
0.01/0.01/0.02
0.01/0.01/0.02
0.0, 5.0, 10.0
0.0, 5.0, 10.0
0.0, 5.0
4-chlorotoluene, 0.80
4-chlorotoluene,0.80
1,2,4-trimethylbenzene, 0.83
,4
10.0
Br as BPTA
Co/Mn
Co/Mn/Br
Co/Mn/Br
Co/Mn/Zr/Br
Co/Mn/Br
6
7
8
9
0
1
1,2,4-trimethylbenzene, 0.83
1,2,4-trimethylbenzene, 0.83
p-toluic acid
toluene, 0.65
p-tolualdehyde
0.04/0.002
0.01/0.01/0.02
0.005/0.005/0.010.0
0.02/0.02/0.001/0.04
0.0025/0.0025
10.0
10.0
0.0
0.0
0.0
1
1
p-tolualdehyde
Co/Mn/Br
0.0025/0.0025/0.0050
HBr added at 124 min
Co/Mn/Br‡ BPTA
Co/Mn/Br
1
1
2
3
1,2,4-trimethylbenzene
1,2,4-trimethylbenzene, 0.83
p-toluic acid,0.37
0.01/0.01/0.02 ‡ 0.02
10.0
10.0
0.01/0.01/0.02
All performed at 958C except #9 at 758C.
Figure 3. Effect of water on the rate of molecular oxygen
uptake during the Co/Mn/Br-catalyzed aerobic oxidation of
Figure 4. Effect of water on the conversion during the Co/
Mn/Br-catalyzed aerobic oxidation of 4-chlorotoluene.
4-chlorotoluene.
acetic acid while the normal light pink color was chloride ion, which does not promote metal-catalyzed
[
1]
observed in 5% water throughout the reaction. The reactions, to a Co/Mn/Br catalyst has been shown to
absorbances of the samples were taken at 450 nm as a enhance the rate of autoxidation by forming additional
function of time, see Figure 6. The samples were diluted ionic bromide via the displacement reaction:
prior to measurement and Beer×s law was assumed to be
valid. Finally, the rate of carbon dioxide generation was
exceptionally high when the solution began to turn
opaque, see Figure 7. The carbon dioxide selectivity is
ꢀ
2†
defined as the rate of carbon dioxide formation/rate of These experiments suggest that either benzylic bromide
oxygen uptake  100.
is a weaker −promoter× during autoxidation or does not
have any catalytic activity whatsoever. An experiment
was designed to demonstrate whether or not benzylic
bromides have catalytic properties.
Inactivity of Benzylic Bromides as a Catalyst
Homogeneous cobalt(II) catalysts at 0.01M are very
A decrease in catalytic activity attributed to benzylic slow to initiate with methylbenzenes at 958C, often
[
14]
bromide formation has been reported during the taking 2 ± 3 h for significant activity to occur. Cobalt/
[
7,9,13]
autoxidation of toluene.
Also, the addition of the manganese catalysts are even less active because any
Adv. Synth. Catal. 2004, 346, 297 ± 306
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
299

FULL PAPERS
W. Partenheimer
Figure 5. Effect of water on the benzaldehyde yield during
the Co/Mn/Br-catalyzed aerobic oxidation of 4-chloroto-
luene.
Figure 7. Effect of water on the carbon oxide selectivity
during the Co/Mn/Br-catalyzed aerobic oxidation of 4-
chlorotoluene.
half-life of 24 h at 958C in 10% water/acetic acid. The
reaction is:
ꢀ
4†
The reagent chosen is 1,2,4-trimethylbenzene because it
is 5 times more reactive than 4-chlorotoluene, see
Table 2, hence more easily initiated.
The experimental result, #5 in Table 1, is that when
,2,4-trimethylbenzene was mixed with a Co/Mn cata-
1
lyst, no recordable oxygen uptake occurred for two
hours, see Figure 8. Repeating the experiment with a
Co/Mn/Br catalyst (0.01/0.01/0.02 M) results in an
immediate high rate of oxygen uptake, see Figure 8,
Figure 6. Absorbance during the Co/Mn/Br-catalyzed aerobic
oxidation of 4-chlorotoluene with 0% initial water concen-
tration.
and a rate of disappearance of the reagent of 6.95 Â
À5 À1
10
s . The calculated amount of hydrobromic acid
provided by solvolysis of a-Br-p-toluic acid is only
Co(III) that is generated is rapidly reduced (half-life of 0.0005 M during this time period. It is concluded that
.0056 sec at 608C in 10% water/acetic acid) by the benzylic bromide has virtually no activity or promo-
0
reaction:
tional effect in metal-catalyzed autoxidations. After 2 h,
the oxygen uptake slowly increased as sufficient hydro-
bromic acid was supplied by hydrolysis of the BPTA to
initiate the reaction.
ꢀ
3†
In contrast, addition of bromide to a cobalt catalyst
immediately initiates the reaction and the activity is Rationalization of the Observations During the
[
14]
increased by a factor of 3.8. A further synergistic Autoxidation of Chlorotoluene at 0 and 5% water/
interaction occurs when manganese is added to a cobalt/ acetic mixtures
bromide catalyst as its activity is increased from 3.8 to
[
15]
1
6. Thus an experiment which will confirm whether or The above empirical observations become understand-
not a benzylic bromide −promotes× an autoxidation able when the yield of the benzylic bromide of 4-
would be to initially use a cobalt/manganese catalyst chlorotoluene was measured ± where the yield is of
with a benzylic bromide. One would want a benzylic benzylic bromide it is based on the amount of sodium
bromide that does not hydrolyze readily under the bromide added to the reaction, see Figure 9. At 120 min
experimental conditions used. a-Br-p-toluic (BPTA) reaction time, 95% of the bromide is benzylic bromide
acid was chosen because the measured solvolysis of and at 180 min 100% of the bromide has been converted
st
À6 À1
BPTA has a 1 order rate constant of 8.0 Â 10 s with a to the catalytically inactive benzylic bromide. One has
300
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 297 ± 306

Synergy of Water in the Metal/Bromide Autoxidation of Hydrocarbons
FULL PAPERS
Figure 8. Effect of different bromine sources on the rate of
oxygen uptake during the Co/Mn/Br-catalyzed aerobic oxi-
dation of 1,2,4-trimethylbenzene. Br-PTA is a-bromo-p-
toluic acid.
Figure 9. Effect of water on the benzylic bromide formation
during the Co/Mn/Br-catalyzed aerobic oxidation of 4-
chlorotoluene.
The active catalytic species may be a metal(II)-bro-
[
17]
essentially gone from a more active Co/Mn/Br catalyst mine(0) compound or HBr2. As a result the steady
to a much less active Co/Mn catalyst. For example, in a state concentrations of cobalt and manganese remain
[
14]
Co/Mn-catalyzed autoxidation of 1,2,4-trimethylben- largely in their lower oxidation states. With a Co/Mn
zene the rate of disappearance of the reagent is catalyst, reactions 5and 6will occur but not 7, hence high
À6 À1
2
.31(0.49) Â 10
s
while with a Co/Mn/Br catalyst, steady states of Mn(III) are expected to occur. In the
À5 À1
the rate is 6.9(0.036) Â 10
s
or 30 times more presence of water disproportionation of the Mn(III) can
[
18]
reactive, experiments 6 and 7.
also occur:
A Co/Mn/Br catalyst has a redox cascade in which
Co(II) is rapidly oxidized to Co(III) via the peroxy acid,
ClPhC(ˆO)OOH. This then rapidly oxidizes Mn(II) to
ꢀ
8†
Mn(III) which in turn oxidizes the Br( À I) to Br(0) at
16]
6
08C[ (ligands omitted for simplicity):
The colors of these species in acetic acid and acetic acid/
water solutions are given in Table 3. Authentic samples
of these metals in their different oxidation states can be
generated from reactions 5, 6 and 7 although reaction 8
is more complex. The color of the solutions during the
oxidation reactions are consistent with the conversion of
a Co/Mn/Br catalyst to a Co/Mn catalyst. In experiment
, when 4-chlorotoluene is reacted in anhydrous acetic
acid, the initial color was an intense blue due to the
tetrahedral Co(II) species present. [The solution is
ꢀ
ꢀ
ꢀ
5†
6†
7†
[
19]
3
Table 2. Reactivity and the rates of solvolysis of benzylic bromides in 10 weight % water and 958C correlated with −catalyst
failure×. Reactivity and rates of solvolysis predicted by Hammett relationships.
Substituent
Reactivity (relative to toluene)
PhCH Br Solvolysis, t [h]
Catalyst failure?
2
1/2
2
4-OCH3
5.610.1
5.02
3.94
1.98
1.78
1.25
1.00
0.78
0.72
0.40
0.35
0.23
0.17
3
1
4
4
3
,4-OCH3
,2,4-CH3
-CH3
-C(CH3)3
-CH3
0.15
no
no
no
0.24
0.94
1.2
2.4
H
3.7
6.0
yes
yes
4
4
4
4
4
4
-Cl
-Br
-COOH
7.2
24
yes
-C(ˆO)OCH3
31
69
123
-CN
-NO2
Adv. Synth. Catal. 2004, 346, 297 ± 306
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
301

FULL PAPERS
W. Partenheimer
actually composed mostly of octahedral Co(II) species the carbon dioxide selectivity at the end of a Co/Mn/Br-
but the small amount of tetrahedral that exists accounts catalyzed reaction in 5 weight % water seen in Figure 7
for the majority of the absorbance.] As the reaction is normal. At the end of the experiment, the substrate is
proceeds, the color turns to a deep brown color used up and the operating catalyst starts to attack the
characteristic of Mn(III). Manganese(IV) will not acetic acid solvent. This occurs for only a short period of
form in significant quantities at these low water time. The number is also magnified because the rate of
concentrations. In experiment 4 when 4-chlorotoluene reaction is very slow and the carbon dioxide selectively
is oxidized with a high initial water concentration, the is a ratio of rate of carbon dioxide formed divided by the
solution has a pale blue color throughout the reaction total rate of oxygen uptake.)
which is indicative of very low Co(III) and Mn(III)
concentrations and is basically due to octahedral catalyst due to all of the active ionic bromide being
cobalt(II) species. converted to benzylic bromide, the resultant catalyst is
When a Co/Mn/Br catalyst is converted to a Co/Mn
Use of the molar absorptivities in Table 3 at 450 nm much less active, much less selective, and imparts a deep
and assuming that all of the metal present is oxidized to coloration to the system. Consequently, we have dubbed
these higher oxidation states results in a calculated this phenomena −catalyst failure× because it converts to
absorbance less than that observed in the experiment. the characteristics of a non-metal/bromide catalyst.
The observed absorption at 450 nm, Figure 6, cannot be
just due to Co(III), Mn(III) and Mn(IV), i.e., there must
be some other absorbing species. The probably cause is Catalyst Failure for 4-Carboxytoluene (p-Toluic Acid,
radical coupling associated with the organic reagents PTA) and Toluene
and subsequent oxidation which are known to produce
highly colored species.
[
18]
Similar to 4-chlorotoluene in anhydrous acetic acid, the
The disappearance of benzylic bromides during these reaction of 4-carboxytoluene (p-toluic acid) also termi-
reactions is due both to the oxidation and hydrolysis of nates due to a-Br-p-toluic acid formation, see experi-
the benzylic bromides, see below. The lower steady state ment 8 and Figure 10. The slightly greater than 100%
concentrations of the benzylic bromide during the yields shown in Figures 9 and 10 are probably due to
oxidation of 4-chlorotoluene in the presence of 5 experimental error since the calculations are based on
weight % water, Figure 9, would be expected from the small GC areas. The changes were mimicked by the
rate of hydrolysis alone since the half-lives of a- oxygen uptake (not shown). The autoxidation of toluene
bromotoluene decrease from 14 to 3.7 to 1.l h at 5, 10 with the highly active Co/Mn/Zr/Br catalyst[
20]
in
and 20 weight % water/acetic acid, respectively, at anhydrous acetic acid at 758C also terminates prema-
58C.[
9]
turely due to benzylic bromide formation, experiment 9,
9
The high rate of carbon dioxide formation is expected see Figure 11.
from the high steady state concentrations of Co(III) and
Mn(III), see Figure 7, since both Co(III) and Mn(III)
are strong oxidants that readily decarboxylate acids. A Re-initiation of a Catalyst Failure with Hydrobromic
Co catalyst has high steady state concentrations of Acid for Toluene and p-Tolualdehyde (4-CH PhCHO)
3
Co(III) giving it an intense green color. Whenbromide is
added to a Co catalyst, the color instantly changes from If the termination of the reaction is due to benzylic
deep green to pale blue. Similarly, the rate of carbon bromide formation, then addition of HBr after a catalyst
[
14]
dioxide is greatly reduced, by a factor 5.9. (The rise of
Table 3. Visible absorption characteristics of cobalt, manga-
nese and zirconium in acetic acid-water mixtures.
Metal
Color
Molar absorptivity
cobalt(II), tetrahedral intense blue
cobalt(II), octahedral pale pink
550 at 600 nm[a]
15 at 450 nm
cobalt(III)
intense green 320 at 450 nm
colorless - - -
intense brown 239 at 450 nm
intense brown 320 at 450 nm
manganese(II)
manganese(III)
manganese(IV)
zirconium(IV)
colorless
- - -
Figure 10. Effect of benzylic bromide formation on the
conversion of 4-carboxytoluene during the Co/Mn/Br-cata-
lyzed aerobic oxidation of 4-carboxytoluene.
[
a]
For the monobromo compound, see: K. Sawada, M.
Tanaka. J. Inorg. Nucl. Chem. 1977, 39, 339.
302
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 297 ± 306

Synergy of Water in the Metal/Bromide Autoxidation of Hydrocarbons
FULL PAPERS
Figure 13. Effect of hydrobromic acid addition on carbon
dioxide selectivity during the Co/Mn-catalyzed aerobic
oxidation of p-tolualdehyde.
Figure 11. Effect of hydrobromic acid addition during the Co/
Mn/Br-catalyzed aerobic oxidation of toluene. Additional
HBr was added at 750 min.
catalysts in glass reactors. Metal-catalyzed reactions
are considerably less important. With a Co/Mn-cata-
lyzed autoxidation experiment, No. 10, catalyst failure
occurs at about 70 min when the rate decreases, the
carbon dioxide selectivity greatly increases, and the
color turns from colorless to green. The conversion of p-
tolualdehyde is 75% at this point. When hydrobromic
acid was added at 125 min, experiment 11, the rate of
oxygen uptake immediately increases from 2.5 to
7
.5 mL/min, the carbon dioxide selectively decreases
from 22 to 7% and the color changes from green back to
colorless.
Mechanism of Benzylic Bromide Formation and
Disappearance
Figure 12. Effect of hydrobromic acid addition on the Co/
Mn-catalyzed aerobic oxidation of p-tolualdehyde.
There are different views on how benzylic bromides
[7,8]
form during autoxidation.
Our view is based on
evidence that they form directly in the coordination
[
23]
sphere of the metal. Cobalt(II) metal dimers react
failure should result in an immediate increase in the rate
of the reaction and a color change from brown back to
pale blue. In the toluene experiment shown in Figure 11,
addition of hydrobromic acid at 743 minutes results in
immediate surge in the rate of reaction and a color
change from dark brown to light pink as expected. The
reaction was allowed to continue and eventually an
preferentially with the two-electron oxidant, peroxy
[24]
acids quickly and selectively, see reaction 5. In water/
acetic acid mixtures, the bromide is a mixture of directly
bonded bromide to the metal and ion-pairs in the
[22]
coordination sphere.
Hence incorporation of the
methylaromatic compound into the cobalt coordination
sphere and subsequent oxidation of the methylaromatic
compound can lead directly to the benzylic bromide:
8
3 mol % yield of benzoic acid was achieved with a mass
accountability of 98.8% with the other compounds
being acetophenone and benzylic bromide.
ꢀ
9†
One can initially start with a Co/Mn catalyst and
demonstrate the same features of −catalyst failure× and
its rectification with an ionic bromide source using
benzaldehydes as reagent. Figures 12 and 13 describe
the more complex autoxidation of p-tolualdehyde. The
oxidizability of benzaldehydes, 290, are much higher
than those of toluenes, 0.014 [oxidizability is defined as
Since the formation of the benzylic bromide is a two-
electron oxidation, this mechanism is particularly effi-
cient as it occurs in one step.
The rate of its disappearance of the benzylic bromides
can occur by hydrolysis:
1
/2
k /(2k ) ; k ˆ propagation rate constant; k ˆ termina-
p
t
p
t
[
21]
tion rate constant].
Benzaldehydes spontaneously
ꢀ
10†
and vigorously autoxidize without any addition of
Adv. Synth. Catal. 2004, 346, 297 ± 306
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
303

FULL PAPERS
W. Partenheimer
and by oxidation. In the case of peroxy radicals one has
the initial reaction:
ꢀ
11†
The products of benzylic bromide oxidation are report-
ed to be benzal bromides, benzaldehydes, benzoic acid
[
7]
and molecular bromide.
Evidence for Hydrolysis and Oxidation of Benzylic
Bromides
Figure 14. Hammett plot for the rate of solvolysis of selected
benzylic bromides at 958C in 10 weight % water/acetic acid.
Points from left to right are for substitutents 4-CH , 3-CH ,
3
3
In experiment 12, both HBr and a-Br-p-toluic acid were H, 4-COOH. Equation of the line is y ˆÀ 1.925x ± 0.0396 with
initially added using the highly reactive 1,2,4-trimethyl- R2 ˆ 0.955 and rho ˆÀ 1.93(0.29).
benzene as the reagent. A vigorous reaction ensued in
which all of the first methyl group was consumed in 24 h.
During this time the a-Br-p-toluic acid has been reactivity of previously reported compounds [1] in 10
potentially exposed to 0.083 moles of peroxy radicals weight % water. Also included in the table are the
since the conversion of the 1,2,4-trimethylbenzene was catalyst failures or lack of catalyst failure observed in
1
00% during this time. From experiment 13 we know this work, and in the authors experience with other
that BPTA does not form under these conditions, i.e., reagents, at 0 weight % water. Assuming the trends
k(formation) ˆ 0.0. Thus the BPTA will disappear by observed are the same at 0% water as in 10% water, one
oxidation and by hydrolysis:
observes that electron-withdrawing substituents result
in a slower rate of benzylic bromide solvolysis as well as
a slower rate of autoxidation. Since benzylic bromides
react both by solvolysis and oxidation there will be an
increased tendency toward catalyst failure as one moves
ꢀ
12†
The observed rate of disappearance of BPTA was down in Table 2. However, there will be a decreased
À5 À1
1
.44 e s . The independent measurement of k(solvol- tendency for the rate of benzylic bromide formation as
À6 À1
ysis) is 8.01e s . One then calculates the rate of one proceeds down the table as well so it is not possible
oxidation of BPTA, k(oxidation), as 6.39 e
Eq. (12). Therefore, the contribution of the overall rate
À6 À1
s
using to predict when catalyst failure will occur.
Empirically, one observes an increased tendency
of BPTA of disappearance due to oxidation is 44% and towards catalyst failure with more electron-withdrawing
that due solvolysis is 56%. Both mechanisms of oxida- substituents.
tion and solvolysis compete at approximately the same
degree under these experimental conditions.
Benzylic Bromide Formation during Nitrotoluene
Oxidation
When would one Expect Catalyst Failure?
It is known that 2- and 4-nitrotoluenes are difficult to
[
1]
The rates of solvolysis of selected benzylic bromides (X- oxidize in high yield. This may be so because of their
PhCH Br; X ˆ 3-CH , 4-CH , H, 4-COOH) were meas- exceptionally low reactivity and high stability of the
2
3
3
ured in 10 weight % water at 958C. The Hammett plot of benzylic bromides, see Table 2. The low yields to the
these data is shown in Figure 14. The rho value is carboxylic acids that have been observed may be due to
À 1.93(0.29) which says that electron-withdrawing sub- the type of catalyst failure described here. Also con-
stitutents on the benzene nucleus result in decreased sistent with this analysis, is that the co-oxidation of p-
rates of solvolysis. The previously published rho value xylene with o-nitrotoluene does produce high yields of
for the Co/Mn/Br-catalyzed oxidation of substituted o-nitrobenzoic acid. This is expected since the peroxides
[
15]
toluenes in 10 weight % water at 808C is À 0.95(.16).
from the oxidation of the more reactive p-xylene can
Assuming the minor difference in temperature is not oxidize the benzylic bromide of o-nitrotoluene so that a
significant, changing substitutents on the ring increases sufficient amount of HBr is always present.
the rate of solvolysis of the benzylic bromide nearly
twice as fast as the rate of oxidation, i.e., 1.93/0.97 ˆ 2.1.
Table 2 gives the predicted half-lives of the disappear-
ance of the benzylic bromides and their relative
304
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 297 ± 306

Synergy of Water in the Metal/Bromide Autoxidation of Hydrocarbons
FULL PAPERS
Benzylic Bromide Formation during the Oxidation of
Polymethylbenzenes
± decreased activity ± undesirable condition
± increased propensity of manganese(IV) formation ±
undesirable condition
The methyl groups in polymethylbenzenes are approx- b) As benzylic bromide formation increases
imately consecutively oxidized to carboxyl groups. The ± the Br/(Co ‡ Mn) ratio decreases (Br is active ionic
reasons for this is that when the first methyl group reacts,
bromide)
the intermediate alcohols, benzaldehydes and carbox- ± results in decreased activity ± undesirable condition
ylic acids are electron-withdrawing groups which deac- ± results in increased propensity for Mn(IV) formation-
tivate the other methyl groups on the ring. Thus for
durene the approximate sequence is:
undesirable condition
c) Propensity for soluble manganese(IV) formation
increases
±
reduced catalytic activity due to decreased rate of
initiation reaction
±
±
±
insoluble MnO forms
Co/Mn ratio increases ± catalytic activity increases
black particles of Mn on isolated product ± undesir-
able condition
2
[20]
The relative reactivities were calculated from the
Hammett equation. It is expected from their relative ± reaction of MnO with reagent forms colored impur-
2
[
18]
reactivities that increasingly stable benzylic bromides
would form in this same order. The question arises is
whether when a given methyl group is being oxidized (in
ities.
II for example), would the more stable benzylic Summary of the Water Effect
bromides of III form and cause a catalyst failure?
In experiment 13 1,2,4-trimethylbenzene and 4-car- We have considered just one effect of water ± on how it
boxytoluene, having reactivities of 3.9 and 0.40, respec- changes the steady state concentration of benzylic
tively, were reacted with air. The LC analysis of the bromide in an oxidation reactor and the consequences
reaction mixtures throughout the experiment did not of those changes. Table 4 gives a number of other factors
detect any of the very stable BPTA forming. Thus, the that are also affected by water. What the net effect of
co-oxidation of two methylbenzenes of very different these factors is difficult to predict. There is a delicate
reactivity will not form the more stable benzylic balance of water concentration required to avoid
bromide of the less reactive compound. The rate of detrimental phenomena from occurring during an
disappearance of 1,2,4-trimethylbenzene was 26 times autoxidation reaction. The net effect of these factors is
faster than that of p-toluic acid. This is consistent with so difficult to predict that autoxidation, using metal and
[
1]
calculations using the Hammett equation.
metal/bromide catalysts, will probably always be an art.
By-Product Cascades Caused by Benzylic Bromide
Formation
Experimental Section
The glass autoclave, procedure, calculations, and GC instrument
have been previously described. The chemicals were used as
There is a number of complex inter-relationships in
[25]
metal-catalyzed autoxidations which makes predicted received. Initial cobalt and manganese catalysts were their
outcomes very difficult. Consider what can happen as metal(II) acetate tetrahydrates. The reactants, intermediates,
by-products, and products formed in the reactions were con-
firmed by GC/MS. Unless otherwise stated, all reactions were
performed with a flow of air at 100 mL/min through the reactor
one oxidizes a methylaromatic compound:
a) Water concentration increases which leads to:
±
decreased propensity of benzylic bromide formation ± with 100 g solvent at ambient atmospheric pressure. Periodically
a desirable condition liquid samples were removed during reaction and analyzed via
Table 4. The balance of factors affected by changes in water concentration during autoxidation using metal/bromide catalysts.
Favored at low water concentration
Favored at high water concentration
reactivity of peroxy radicals of aldehydes, alcohols
catalyst inhibitor conversion
solvolysis of benzylic bromide with associated catalyst failure
less yield loss due to aryl formate formation
less Mn(IV) formed ± faster reaction with bromide
less phenol formation ± less concentration of a strong antioxidant
less carboxylic acid metal precipitation
less Mn(IV) formed ± less MnO precipitation
2
reactivity of peroxy radicals due to acetic acid H-bonding
Adv. Synth. Catal. 2004, 346, 297 ± 306
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
305

FULL PAPERS
W. Partenheimer
GC. At the same time, on-line gas analysis was performed for
dioxygen, carbon monoxide, carbon dioxide, and dinitrogen.
[14] W. Partenheimer, A Chemical Model for the Amoco
−MC× Oxygenation Process to Produce Terephthalic Acid,
in: Catalysis ofOrganic Reactions , (Ed.: D. W. Black-
burn), Marcel Dekker, New York, 1990.
References
[15] W. Partenheimer, A Method for the Preparation of
Pyromellitic Acid from Durene via Liquid Phase Oxida-
tion, in: Catalysis ofOrganic Reactions , (Ed.: D. W.
Blackburn), Marcel Dekker, New York, 1994.
[16] W. Partenheimer, Nature ofthe Co-Mn-Br Catalyst in the
Methylaromatic Compounds Process: Kinetic and Ther-
modynamic Studies, in: Catalytic Selective Oxidation,
(Eds.: S. T. Oyama, J. W. Hightower), American Chem-
ical Society, Washington, 1993.
[
[
[
1] W. Partenheimer, Catal. Today 1995, 23, 69.
2] V. N. Aleksandrov, Kinetika i Kataliz 1979, 20, 357.
3] K. Ukegawa, Y. Kamiya, Bull. Chem. Soc. Jap. 1976, 49,
1
632.
[
[
4] I. V. Zakharov, Y. V. Geletii, V. A. Adamyan, Kinet.
i Kataliz 1991, 32, 31 .
5] M. P. Czytko, G. K. Pub, Ind. Engl Chem. Prod. Res. Dev.
1
961, 20, 481.
[17] P. D. Metelski, J. H. Espenson, J. Phys. Chem. 2001, 105,
5881.
[
[
6] L. J. Csanyi, K. Jaky, J. Catal. 1993, 41, 721.
7] G. M. Dugmore, G. J. Powels, B. Zeelie, J. Mol. Catal.
[18] W. Partenheimer, J. Mol. Catal. 2003, 206, 131.
[19] G. H. Jones, J. Chem. Soc. Chem. Commun. 1979, 536.
[20] W. Partenheimer, J. Mol. Catal. 2003, 206, 105.
[21] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations
ofOrganic Compounds , Academic Press, New York,
1981, chapter 2.
[22] W. Partenheimer, J. Mol. Catalysis 2001, 174, 29.
[23] T. V. Bukharkina, N. D. Gavrilenko, N. G. Digurov, N. A.
Knyazeva, Kinet. Kataliz 1978, 19, 506.
1
995, 99, 1 .
[
[
8] P. D. Metelski, A. Adamian, J. H. Espenson, Inorganic
Chem. 2000, 39, 2434.
9] W. Partenheimer, Novel Catalytic Characteristics ofthe
Co/Mn/Cl/Br Liquid Phase Oxidation Catalyst, in: Cat-
alysis ofOrganic Reactions , (Ed.: F. E. Herkes), Marcel
Decker, New York, 1998.
[
10] A. S. Hay, US Patent 3,139,452, 1964.
[
11] D. E. Burney, G. H. Weisemann, N. Fragen, Petroleum
Refiner 1959, 38, 186.
[24] G. H. Jones, J. Chem. Research (M) 1981, 2801 .
[25] W. Partenheimer, V. V. Grushin, Adv. Synth. Catal. 2001,
343, 101.
[
[
12] A. Saur, N. Seine, P. Simonnin, US Patent 3,255,243, 1966.
13] A. S. Hay, H. S. Blanchard, Can. J. Chem. 1965,43, 1306.
306
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 297 ± 306