The influence of pH in metal/bromide catalyzed homogeneous aerobic oxidation. Part 1. Effect on activity and selectivity of p-xylene oxidation

Article abstract of DOI:10.1016/j.apcata.2013.11.037

By variation of the pH i.e. the acetate concentration, during oxidation of p-xylene in acetic acid/water mixtures and catalyzed by Co/Mn/Br, a maximum activity and selectivity is observed. The same is observed for the autoxidation of acetic acid by Co(II) acetate. If one operates at a lower or higher than the optimum pH, the rate of the formation of the desired product decreases and the rate of undesirable by-products increases. At sufficiently low pH the autoxidation reaction does not initiate. These observations can be rationalized by noting how the structures of the catalyst metal compounds change resulting in an increase in redox potentials of the metals as the acetate concentration decreases. Thus a thermodynamic balance of the required oxidation and reduction reactions of the metals is required. This explains why metal acetates i.e. basic compounds, are used as the initial catalytic species. Often autoxidation reactions are optimized by changing the catalyst concentration and it should be noted that this changes the pH as well. The pH dependence of the reactions that form undesirable by-products is summarized. The concept of thermodynamic and kinetic balance can also rationalize why metal/bromide catalysts are active while metal/fluoride, chloride, and iodide are either non-active or very weak.

Full text of DOI:10.1016/j.apcata.2013.11.037

Accepted Manuscript
Title: The Influence of pH in metal/bromide catalyzed
homogeneous aerobic oxidation. Part 1. Effect on activity and
selectivity of p-xylene oxidation
Author: Walt Partenheimer
PII:
S0926-860X(13)00727-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2013.11.037
APCATA 14592
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
25-7-2013
18-11-2013
24-11-2013
Please cite this article as: W. Partenheimer, The Influence of pH in metal/bromide
catalyzed homogeneous aerobic oxidation. Part 1. Effect on activity and
selectivity of p-xylene oxidation, Applied Catalysis A, General (2013),
http://dx.doi.org/10.1016/j.apcata.2013.11.037
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

1
The Influence of pH in metal/bromide catalyzed homogeneous aerobic oxidation.
Part 1. Effect on activity and selectivity of p-xylene oxidation
Walt Partenheimer, Central Research and Development, E.I. DuPont de Nemours & Co.,
Inc., Experimental Station, Wilmington, Delaware, 19880-0328 USA (retired). Currently
a consultant, 7253 SW 26th Ave, Portland, OR, e-mail:wpartenheimer@msn.com
Abstract
By variation of the pH i.e. the acetate concentration, during oxidation of p-xylene in
acetic acid/water mixtures and catalyzed by Co/Mn/Br, a maximum activity and
selectivity is observed. The same is observed for the autoxidation of acetic acid by Co(II)
acetate. If one operates at a lower or higher than the optimum pH, the rate of the
formation of the desired product decreases and the rate of undesirable by-products
increases. At sufficiently low pH the autoxidation reaction does not initiate. These
observations can be rationalized by noting how the structures of the catalyst metal
compounds change resulting in an increase in redox potentials of the metals as the acetate
concentration decreases. Thus a thermodynamic balance of the required oxidation and
reduction reactions of the metals is required. This explains why metal acetates i.e. basic
compounds, are used as the initial catalytic species. Often autoxidation reactions are
optimized by changing the catalyst concentration and it should be noted that this changes
the pH as well. The pH dependence of the reactions that form undesirable by-products is
summarized. The concept of thermodynamic and kinetic balance can also rationalize why
metal/bromide catalysts are active while metal/fluoride, chloride, and iodide are either
non-active or very weak.
Keywords
Page 1 of 30

2
autoxidation, metal/bromide catalysts, p-xylene, cobalt, manganese
1
Introduction
The discovery of metal/bromide catalysts in 1954 has developed into a major industrial
method for producing aromatic carboxylic acids from alkylaromatics. These are
homogeneous oxidations where dioxygen is the primary oxidant and acetic acid is the
solvent. The most common catalyst is a mixture of Co/Mn/Br in which the Co(II) and
Mn(II) acetates are the initial compounds added. Over 251 hydrocarbons have been thus
oxygenated, often with yields of 80-95%. Four major commercial products are produced
using this method- terephthalic acid, isophthalic acid, trimellitic acid, and 2,6-
dicarboxynaphthalene [1]. Optimization of these multi-billion pound/yr process for
maximum activity and selectivity is obviously important. In this, and the subsequent
paper, we give experimental results and rationalize why the acetate anion can be used to
optimize the activity and selectivity in these reactions as well as prevent the precipitation
of the catalyst metals as metal carboxylate complexes. The ability of acetate to improve
these autoxidations has been patented for the oxidation p-xylene [2] and 2,6-
dimethylnaphthalene [3].
While employed at Amoco Corporation, the author discovered the effect of the acetate
concentration serendipitously. Traditionally, the initial catalyst components of
metal/bromide catalyzed autoxidations use Co(II) and Mn(II) acetates along with a source
of ionic bromide. Thinking that other salts of Co and Mn with similar weakly
coordinating ligands may be more active, we evaluated the tetrafluoroborate salt of Co
along with hydrogen bromide when oxidizing 1,2,4-trimethylbenzene. We were
Page 2 of 30

3
surprised by the fact that no oxidation occurred – particularly since the original patent [4]
states that a metal with bromide is all that is required for an active catalyst. After some
thought, we realized that acetate is the anion which is a base in acetic acid, just as the
hydroxide anion is the characteristic base in water. The experiment was repeated and then
an increment of ammonium acetate was added. The reaction immediately initiated! This
led to the work described in the patents and in these two papers.
There are reports which indicate that activity, yield of the major product, as well as by-
product formation are affected by the acetate concentration. However, often these have
not been attributed to changes in pH. Thus addition of metal acetates have resulted in
shorter induction periods for the Co/Br catalyzed oxidation of various methylaromatic
compounds [5,6]. The initial activity of the Co/Br catalyzed oxidation of 1,2,5,6-
tetramethylbenzene is one-half order with respect to the acetate concentration [6]. The
addition of amines to metal/bromide catalysts have been reported [7,8]. The amine can
react with the acetic acid to form the acetate hence amine addition may be partially a pH
effect. The acetate concentration affects by-product formation since the addition of
sodium acetate to a cobalt catalyzed autoxidation of butane to acetic acid reduces the
formation of carbon dioxide and carbon monoxide [9]. The presence of acetate has been
reported to enhance the yield during the oxidation of 4,4’dimethylbiphenyl to
4
,4’dicarboxybiphenyl [10,11].
In this paper we extend the above observations to 1) metal/bromide catalysts in general
there are 29 of them [1]) 2)show that an optimum activity and selectivity is obtained at a
(
Page 3 of 30

4
given acetate concentration and 3) illustrate, in part 2, that precipitation of metal
carboxylate complexes are dependent upon the acetate concentration. The results are
understandable using the solvent system definition of acids and bases and the effect of the
acetate upon the redox potentials of the metals.
2
Experimental
The rates of dioxygen uptake and of formation of carbon monoxide and carbon dioxide
were measured by passing air (50 ml/min) at ambient pressure through a glass fritted
cylindrical reactor containing 100 ml of solvent (acetic acid or water/acetic acid mixture),
oxidant and catalyst. The flow rate into and exiting the reactor are monitored. All color
changes and precipitation phenomena were directly observable since the reactor is
constructed of glass. The vent dioxygen concentration was constantly monitored by a
Beckman 715 oxygen meter whose output is continuously monitored with a Scientific
Instrument recorder. The vent gases were periodically analyzed for dioxygen, nitrogen,
methane, carbon monoxide carbon dioxide via GC as previously described [12]. Rates of
dioxygen uptake and carbon monoxide and carbon dioxide formation are calculated from
the GC data knowing the flow rate of gases through the reactor and the composition of
pure air. Rates of dioxygen uptake are given in ml O /min and carbon dioxide selectivity
2
is given by the rate of carbon dioxide generated divided by the rate of dioxygen uptake.
Percent completion was calculated by numerical integration of the total oxygen uptake
and is based on all the methyl groups in the reagent being converted to carboxylic acids.
The experimental procedure was to add the methylaromatic compound to the acetic acid
with the air flowing through the reactor. The catalyst components were than added. The
Page 4 of 30

5
vent dioxygen, carbon monoxide, and carbon dioxide concentrations were measured ½ hr
later via GC. The water concentration in the reactor was changed via addition of the
appropriate amount of water and the GC measurement again made.
All metal salts were obtained from Alfa Inorganics and used as received. The
methylaromatic compounds were of the highest purity obtainable and used as received.
3
Results
3
.1 Effect of acetate concentration on the autoxidation of acetic acid by cobalt(II). The
formation of CO (x=1,2) gases is undesirable during methylaromatic oxidation and its
x
reduction is one of the main economic drivers in the process. During optimization of
these processes by variation of the cobalt concentration, it becomes obvious that cobalt is
x
one of the main causes of CO . Addition of cobalt(II) acetate tetrahydrate to an
oxygenated solution of acetic acid results in a well-characterized reaction [12]:
2 2 3 3 2 2
Co(II)(OAc) + 3 HOAc + O ===Î Co(III)(OAc) + CH OH + H O + CO (1)
The reaction is initiated with adventitious peroxide followed by a catalyst modified free
radical chain reaction. If cobalt(II) perchlorate is used in place of cobalt(II) acetate no
2
reaction occurs i.e. the vent O level remains at 20.9%, see figure 1. Figure 1 gives the
rate of reaction, as measured by the oxygen uptake, as a function of acetate
concentration. The later was increased by addition of ammonium acetate. A maximum
oxygen uptake occurs around a acetate/metal molar ratio of ~2.0 The maximum carbon
Page 5 of 30

6
dioxide and methane formation occur at the same ratio, see figure 2. Figures 1,2 combine
experimental points from two independent experiments.
3
.2 Effect of acetate concentration on metal/bromide autoxidations of p-xylene.- The
activity of Co/Mn/Br catalyzed oxidations of p-xylene as a function of acetate metal ratio
is given in figure 3. The initial salts used are cobalt(II) perchlorate, manganese(II)
tetrafluoroborate, and sodium bromide. It is not known whether acetate is generated via
solvolysis when these salts are placed in acetic acid. Each curve is generated at constant
conversion of p-xylene. No reaction occurs without addition of acetate. A maximum
activity is generated at an acetate/metal ratio of ~1.0. The activity decreases when the
water concentration in acetic acid is increased as has been previously reported [1,13].
The data is presented as a function of water concentration because 1) often the initial
acetic acid used contains a fixed amount of water and 2) water is generated during the
oxidation of alkylaromatic compounds. Figures 4,5 gives the selectivity to COx (x=1,2)
generated in the same experiments during the oxidation of p-xylene. These undesirable
by-products form at much higher rates at acetate/metal ratio below ~2.0. They then
increase as the acetate concentration increases. It is known by isotope labeling studies
that COx forms from both the methylaromatic compound and the acetic acid [14].
3
.3 Comparison of using NaBr and HBr as bromide sources. When performing
methylaromatic aerobic oxidations with metal/bromide catalysts, one has a choice of
using either NaBr or HBr as the bromide source. Thus the acetate concentration, i.e. the
pH, will be somewhat different depending upon the bromide source. The acidity of HBr
in acetic acid has been measured and compared to other acids such as HClO and H SO
4
2
4
[
15]. In general the activity (oxygen uptake) and COx selectivity is different when HBr
Page 6 of 30

7
or NaBr are used for Co/Br, Mn/Br and Co/Mn/Br catalysts, see table 1. All the
comparisons in table 1 are made at low conversions of the methyaromatic compound.
This is necessary since the formation of COx changes as the conversion increases
because a significant increase in concentration of intermediates and the product occur
during the reaction. The activity and selectivity advantages of using either HBr or NaBr
will depend upon the product yield and the by-products produced when the reaction is
complete i.e. 100% conversion. It is anticipated that pH will not effect the reaction of
different feedstocks in an identical fashion.
3
.4 Effect on the rate of reaction at low pH. We have also observed that addition of
methanesulfonic acid to a Co(II) acetate/ NaBr catalyzed oxidation of pseudocumene will
terminate the reaction as the rate of oxidation was reduced from 6.5 ml O min to 0.0 at
2
/
an acid/metal ratio of 3.0 mol/mol. Similarly, with a Mn(II)acetate/NaBr catalyst with an
activity of 3.9 cc O /min, addition of methanesulfonic acid so that the acid/metals ratio
2
was 3.8, the reaction rate was reduced to zero. Thus, when the acetate is protonated with
methanesulfonic acid, the pH is decreased and eventually the reaction ceases. This is
similar to when non-acetate Co(II) and Mn(II) salts, such as perchlorate and
tetrafluoroborate where no activity was observed, see figure 3.
4
Discussion
4
.1 The solvent system definition of acids and bases in acetic acid-water mixtures. The
solvent system definition of acids and bases is useful in rationalizing the effects of
varying the acetate concentration on the rates, selectivities, and metal precipitation
described in part 2. The definition of an acid is a substance which produces (i.e.,
increases the concentration of) the cation characteristic of the solvent. The definition of a
Page 7 of 30

8
base is a substance that increases the concentration of the anion characteristic of solvent
[
16]. For acetic acid we have [15]:
+
-
-13.60
2
HOAc <====> H OAc + OAc
K=10
(2)
2
Of the common mineral acids, only perchloric acid is a strong acid in acetic acid, with
hydrobromic acid being the next strongest [15]. In acetic acid/water mixtures, the
+
hydroxide ion and protonated acetic acid (H OAc ) are stronger bases and acids than the
2
+
acetate anion and H O respectfully. They will be subsequently "leveled" to the
3
concentration of acetate and hydronium ion. Thus we then have in acetic acid/water
mixtures:
+
-
HOAc + H
2
O Í====Î H
3
O + OAc
(3)
Thus an acid in acetic acid/water mixtures is a substance that produces the hydronium
cation and a base is one that increases the acetate concentration. The dissociation
constant of equation 3, increases with increasing water concentration. The pK
from 13.9 to 8.1 to 6.65 to 5.89 as the water concentration increases from 0 to 10 to 20 to
0 wt %. [17] We have found above that Co(II),Mn(II) perchlorates and tetrafluoroates
d
increases
3
salts mixed with HBr or NaBr are not active catalysts. It is known that Co(II) and Mn(II)
acetate along with HBr or NaBr are active catalysts, as long as the HBr is not added in
excess (HBr/metal ratio less than 2.0 mol/mol) [1]. Co(II) acetate does dissociate to
-7.53
acetate in anhydrous acetic acid since the measured dissociation constant is 10
[15].
We conclude that metal/bromide catalysts are only active when a source of acetate is
added as part of the catalyst mixture. We found above that addition of an excess of
methanesulfonic acid to an active Co(II) acetate/NaBr catalyst immediately resulted in
Page 8 of 30

9
cessation of the reaction. This can be explained by the methanesulfonic acid reacting
with the acetate to produce too low of a acetate concentration for the mixture to be active.
4
.2 An explanation of the acetate effect in autoxidations
In catalytic systems, it is required that the metals be oxidized and then reduced back to
their original oxidation state. In homogeneous autoxidation using a free radical chain
mechanism, the metals undergo one electron changes [18]. Thus one has:
M(N) + (O) ==Î M(N+1)
(4)
(5)
M(N+1) + S ===Î M(N) + P
_
__________________________________________
S + (O) ====Î P (6)
ligands omitted for simplicity, M is a metal, N is the oxidation state, O is an oxidant, S
(
is a substrate, and P the product)
In rxn 4, the oxidant may be organic, such as peroxides and peroxyacids. Or it can be
another metal such as in Co/Mn/Br [19] and Co/Ce/Br [20] catalysts where we have:
Co(III) + M(N) =======Î Co(II) + M(N+1)
(7)
Where M(N) is Mn(II) or Ce(III).
The substrate, S, can be an organic specy such as toluene being oxidized to the benzylic
radical or an inorganic specy such as bromide which becomes oxidized to a bromine(O)
specy, such as occurs in metal/bromide catalysis [1].
Page 9 of 30

1
0
For an efficient catalytic system both rxns 4 and 5 must be thermodynamically and
kinetically feasible. If the oxidation potential of the metal is too high, then rxn 4 may be
thermodynamically prohibitive while rxn 5 may be thermodynamically prohibitive if the
oxidation potential is too low. There must be a some intermediate oxidation potential of
the metal to make both rxns 4,5 thermodynamically and kinetically feasible. We call this
achieving thermodynamic or kinetic balance in the chemistry. Redox potentials are
dependent upon the pH of medium [21] since the structure i.e. ligand field of metal
complexes vary with pH. In aqueous systems, the redox potential and structure of metals
as a function of pH have been described via “E-pH” diagrams also called “Pourpaix
diagrams” [22]. The energy required to oxidize a metal decreases as the pH is increased.
The result is that the spontaneity of rxn 4 increases and that of rxn 5 decreases with an
increase in pH. This results in a minimum in Gibbs energy and therefore a maximum in
activity as illustrated in figure 6. Indeed, these type of curves are experimentally
observed in figures 1-5.
There is only limited data on E-pH effects in acetic acid but the effect is similar to that in
water. For example, for Mn(II)/Mn(III) acetate, increasing the acetate concentration i.e.
increasing the pH, decreases the redox potential [21]. The formal redox potential of
Mn(II)/Mn(III) acetate in 10% wgt. water/acetic acid is 0.801. Addition of 0.45M
perchloric acid raises the potential to 1.45 V while addition of 1.7M trichloroacetic acid
raises the redox potential to 1.39V. Addition of water decreases the redox potential from
0
.801V in 10% water to 0.791 in 30 wgt. percent water [10]. These changes reflect the
structural changes of Mn(II)/Mn(III) acetate in water/acetic acid solutions. The structure
Page 10 of 30

1
1
of these complexes, based on available information in the literature, has been reported
[
(
30]. The structure of Mn(II) acetate is approximately [Mn(II)(HOAc) (H O)
x 2 y
2
OAc) ](x+y=4) where the [ ] brackets indicate the ligands in the coordination sphere.
Polynuclear compounds exist in acetic acid and their presence will probably increase with
metal concentration. Increasing the water concentration increases the water in the
coordination sphere i.e. decreases the x/y ratio. Addition of acids will neutralize acetate
acid and produce cationic species such as {[Mn(II)(HOAc) (H O) ](ClO ) } (x+y=6).
x
2
y
4 2
The { } brackets indicate that the perchlorate anion is ion-paired with the metal complex.
The redox potential in 10% water/acetic acid solution for Co(III)/Co(II) is 1.2V [15].
Co(III) in water is one of the strongest known oxidants at low pH and its behavior is
dramatically different as a function of pH. For example, in water Co(III) forms
spontaneously from cobalt(II) in the presence of dioxygen at pH of 8.0 but does not at a
pH of 1 since the redox potential of Co(III)/Co(II) changes from 0.17 to 1.84V. The
redox potential changes because the structure of Co in aqueous solution changes from
monomeric aquo complexes to complex, polynuclear aquo-hydroxo-oxo complexes as the
+
2
pH increases. The analogous change of Co(II) in acetic acid is from [Co(II)(HOAc)
to [Co(II)(HOAc) (OAc) ] [30].
6
]
4
2
The proposed initiation sequence for rxn 1 is [12]:
*
Co(OAc)
2
+ ROOH (trace) + HOAc ==Î Co(OAc)
3
+ RO + H
2
O (8)
*
Co(OAc) ===Î Co(OAc) + CH + CO
(9)
3
2
3
2
Page 11 of 30

1
2
The methyl radical abstracts a hydrogen atom from acetic acid in rxn 9 to produce the
observed methane gas. We propose that the peroxide, ROOH, does not react with
cobalt(II) perchlorate but does with Co(II) acetate due to the differences in their redox
potentials. This is feasible because the redox potential of Co(IIl)/Co(II) of 1.82 V in
water while hydrogen peroxide has a value somewhat less of 1.7 V [15].
Similarly in Co/Br or Mn/Br catalyzed oxidations we have:
*
M(II)(OAc) + PhCH OOH + HOAc =Î M(III)OAc) + PhCH O + H O (10)
2
2
3
2
2
-
*
M(III)(OAc)
3
+ Br ===Î Mn(II) + Br
(11)
(12)
*
*
Br + PhCH
3
===Î HBr + PhCH
2
The dependence of these reactions on Co(III) and Mn(III) formation make all of them
prohibitive at low pH because they cannot be oxidized from their metal(II) oxidation
states. Thus, Co/Br and Mn/Br are not active at low pH.
It is important to note that the autoxidation catalysts use metal acetates, most often Co(II)
and Mn(II). Thus we are adding basic species to the acetic acid/water solvent. Changing
the catalyst concentration is also changing the acetate concentration. When comparing
the results of different experiments one must compare them at the same catalyst
concentration otherwise one will be comparing them at different pH values and one will
be at different points in figure 6. This is also important to consider when optimizing
autoxidation reactions. One optimization procedure is to vary the catalyst concentration
where one normally sees an increase in rate of reaction and then a decrease in the rate
with increasing catalyst concentration [8,32,33].
Page 12 of 30

1
3
There are various reports on the affect of addition of acids and bases to autoxidation
reactions of different methylaromatic compounds. Both increases [42,44] and decreases
[
43] in rate were observed during the autoxidation of p-xylene by addition of acids.
Similarly, addition of acetate either had no effect [43] or a retarding effect [45]. Thus,
the initial catalyst to the reaction may be improved or detrimentally affected by the
change in pH. The initial pH in most of these reports are different because of the different
catalyst concentrations used.
One of the reasons that an intermediate value of pH is important is that by-products from
the decarboxylation of the acetic acid and aromatic acids is acid catalyzed hence they
would be predominant at low pH. By-product formation is economically undesirable
because they often constitute a yield loss and have to be separated and disposed of.
One of the origins of carbon dioxide and methane is from the decarboxylation of acetate
with cobalt(III), see rxn 9. The selectivity of carbon dioxide formation is poorest at the
lowest acetate concentrations, see figure 4, consistent with the expected highest rate at
lowest pH. The complex chemistry involved when Co(III),Mn(III) and Ce(IV)
decarboxylate fatty acid solvents has been studied. It is found that decreases in pH
enhance the rate of decarboxylation [16-19].
Not only does decarboxylation form undesirable by-products but it deactivates the
catalyst as well. This has been demonstrated by ‘spiking’ selected aromatic acids into
Page 13 of 30

1
4
autoxidation [23]. The reason is that the desirable reaction of Co(III) is with ionic
bromide or Mn(II) which is in competition with Co(III) decarboxylating the aromatic
acid. When the aromatic acid is decarboxyated it initiates a chain of undesirable
reactions. The phenyl radical is oxidized to a phenol which is strong anti-oxidant, the
phenol is oxidized to quinones which are highly colored, and the quinone is oxidized to
carbon dioxide and monoxide [23].
Many other undesirable reactions during autoxidation reactions that produce unwanted
by-products are pH dependent. Thus the thermal decarboxylation of trimellitic acid is
acid catalyzed [24]. The thermal decomposition of p-methylbenzyl hydroperoxide is acid
catalyzed to give the yield reducing, antioxidant cresol in 55% yield [25]. The preferred
reaction is the dehydration of the peroxide by metals of high Lewis acidity [26,27]. The
thermal decomposition of the peroxyacids have their maximum rate of thermal
decomposition at their pK value [28]. The desired reaction of the peroxyacid is with
dimeric Co(II) which is much faster and more selective than thermal decomposition [19].
In metal/bromide catalyzed autoxidation the bromide/metals ratio is usually around 1.0
mol/mol. When HBr is used as the bromide source, the metal(II) diacetates reacts with
and neutralizes the HBr and thus decreases the importance of such things as thermal
decomposition of the primary peroxides and the corrosivity of the system. Sometimes
this not realized [29]. During autoxidation with metal/bromide the active bromide can
react with the methylaromatic compound to form benzylic bromides (PhCH Br) which
2
are an inactive form of bromide which can reduce the activity and selectivity of the
reaction [30]. The rate of solvolysis of these detrimental benzylic bromides can be
Page 14 of 30

1
5
enhanced by the presence of sodium acetate [31,32] and nitrogen containing bases in
solution [33] hence is effected by pH.
A pertinent illustration of the effect of pH on the selectivity of methylaromatic oxidation
is the reaction of the hydroxyl radical with the methylaromatic compound to give the
cyclohexyldienyl radical, see figure 7 [23]. The highly reactive hydroxyl radical is
generated via the thermal decomposition of the peroxides in solution. At higher pH, the
desired reaction to the benzylic radical occurs which eventually is oxidized to the
carboxylic acid. At lower pH, the formation of cresol is favored. The latter is
undesirable because it constitutes a carboxylic acid yield loss and the cresol is a very
strong anti-oxidant which in very small quantities significantly decreases the rate of
reaction [30].
When CoBr is used as the initial catalyst one observes quite unusual behavior. We
2
suggest the reason is that one is initially adding a non-acetate compound, similar to
Co(II) tetrafluoroborate, which results a very low pH. The metal/bromide catalysts used
to oxidize over 251 compounds almost invariably use a catalyst composed of metal
acetates along with a source of bromide. Thus one adds a basic compound, the metal
acetate, with a source of bromide. These autoxidation reactions, in contrast with those
without bromide, do not have induction periods. Interestingly, CoBr
2
as a catalyst is
distinctly different than a Co(OAc) /HBr catalyst. The CoBr catalyst 1) has an induction
2
2
period, 2) is much less active and 3) and addition of a metal acetate to the CoBr2
eliminates or reduces the induction period and enhances the reactivity of the catalyst
Page 15 of 30

1
6
[
5,6,31,35]. The rate law for the oxidation of durene is first order with respect to CoBr
2
and one-half order with respect to the acetate concentration [6]. This clearly indicates
the importance of the proper pH balance i.e. acetate concentration for making an active
and selective catalyst. It has also been reported that autoxidation catalysts in which an
amine (pyridine, triethanolamine) has been added to CoBr
7,33]. We have performed an UV-VIS study showing that amines do not coordinate to
the metal in acetic acid which suggests that the reaction is: amine + CoBr + HOAc
2
make the catalyst more active
[
2
+1
-1
=
=Î CoBr(OAc) + Hamine + Br [32]. It has been reported that quanidine is
completely protonated in acetic acid [8]. So, the amine generated acetate which then
activated the reaction.
The fact that high acidity prevents the autoxidation from proceeding is consistent with the
inability to oxidize 4-methylbenzenesulfonic acid to 4-carboxybenzenesulfonic acid. The
sulfonic acid moiety must be first ‘masked’ by making it a methyl ester, or sodium salt,
or sulfonamide derivative before oxidation[1].
4
.3 Thermodynamic and kinetic balance in metal/halogen catalysts. Co/F, Co/Cl, and
Co/I autoxidation catalysts are either non-active or much less active than Co/Br catalysts
1]. These observations can be rationalized by assuming a mechanism of initiation of the
[
free radical chain sequence as:
-
*
Co(III) + X ===Î Co(II) + X
(13)
(14)
*
*
X + PhCH ====Î PhCH + HX
3
2
Where X is a halogen. The spontaneity (assuming that Gibbs energy changes parallel the
enthalpy changes) of rxn 13 is I>Br>Cl>I, see table 2. Reaction 14 is opposite i.e.
Page 16 of 30

1
7
F>Cl>Br>I. These values are only approximate since they are based on the redox
*
*
potentials in water and the solvation energies of X , PhCH , and PhCH are ignored.
3
2
Both rxns 13 and 14 must occur for the initiation reaction to occur. Only for the bromide
is the enthalpy change moderately positive for both reactions, while presumably the
others are inhibited by either rxn 13 or 14. Perhaps more pertinent, however is the rate of
these reactions. Rxn 13 is not observed to occur for fluoride, is quite slow for chloride,
much faster for bromide, with iodide being the fastest. The Co/I catalyst does not
function well because the iodine atom is unable to oxidize toluene at a sufficient rate.
5
Conclusions
The variation of pH i.e. the acetate concentration, results in changes in the structure and
redox potential of the catalyst metals. This in turn results in an optimum pH where
maximum activity and selectivity are observed during autoxidation reactions. This has
been demonstrated for the aerobic oxidation of p-xylene, which produces the largest
volume of the commodity chemical, terephthalic acid. And it is also demonstrated for
probably the most important side-reaction, the oxidation of Co(II) and its subsequent
decomposition of acetic acid. Too low a pH results in the termination of the oxidation
reactions.
References
[
[
[
1] W. Partenheimer, Catal. Today, 23, (1995) 69-158
2] W. Partenheimer, D.J. Graziano, US Patent 5,081,290 (1991) to Amoco Corporation
3] P.A. Sanchez, D.A. Young, G.E. Kuhlmann, W. Partenheimer, W. P. Schammel,US
patent 4,950,786 (1990) to Amoco Corporation
4] A. Saffer, R.S. Barker US Patent 2,833,778 (1958)
[
Page 17 of 30

1
8
[
[
[
[
[
[
[
[
[
[
5] Y. Kamiya, Tetrahedron, 22, (1966) 2029
6] F.F. Scherbina and T.V. Lysukho, Kinetika I Kataliz, 19 (1978) 1076.
7] G. R. Steinmetz, C. E. Sumner. Journal of Catalysis 100 (1986) 549-551
8] Y. Cheng, X. Li, Q. Wang, L. Wang, Ind. Eng. Chem.. Res. 44 (2005) 7756-7760
9]M.W. Prengle, Jr. and N. Barona, Hydrocarbon Processing (1970) 106
10] G.S. Lee, K.A. Burdett, and J.J. Maj, US Patent 4,996,353 (1991)
11] T. Yamada, K. Yokota, and I. Ito, Japan Patent, 91,225,049 (1991)
12] W. Partenheimer, J. Mol. Catal., 67(1991) 35-46.
13] W. Partenheimer, Adv. Synth. Catal. 347 (2005)580-590
14] J. Dermietzel, C. Wienhold, H. Grundmann, A. Staschok,
J. Koch, E. Bordes. Chem.Tech., Vol. 35 (1983) 2932
[
[
[
15] I.M. Kolthoff and A. Willman, J.Amer. Chem. Soc.56, (1934)1007
rd
16] J. E. Hueey, Inorganic Chemistry, 3 edit. Harper and Row (1983) 288.
17] A. I. Popov, in The Chemistry of Nonaqueous Solvents III, JJ.
Lagowski, ed. Academic Press, NY, (1970) 214
18] R. A. Sheldon, and J.K.Kochi, Metal Catalyzed Oxidations of Organic
Compounds, Academic Press, New York, N.Y. 1981, 7
19] W. Partenheimer and R. K. Gipe, Catalytic Selective Oxidation, in S.T. Oyama, J.W.
Hightower (eds) American Chemical Society, (1993) 81-88
20] R.K. Gipe and W. Partenheimer in Studies in Surface Science and Catalysis, Eds. R.
K. Grasselli, S.T. Oyama, A.M. Gaffney, J.E. Lyons (eds) Elsevier. 1997, 1117
[
[
[
[
21] U.G.K. Raju, V.V.Rao, B. Sethuram and T.N. Rao, J. Electroanal. Chem. 133
1982) 322.
(
Page 18 of 30

1
9
[
22] M. Pourbaix, “Atlas of electrochemical equilibria in aqueous solutions”, National
Association of Corrosion Engineers, Houston, Texas, 1974.
23] W. Partenheimer and J.A. Kaduk, Proceedings from the 4th International
[
Symposium on Activation of Dioxygen and Homogeneous Catalytic Oxidation, Elsevier,
New York, N.Y., 1991, 613-521
[24] J.B.Dunn, M.L. Burns, S.E. Hunter, P.E. Savage, J. of Supercritical Fluids (2003),
2
63-274.
[
[
[
[
[
[
[
[
[
(
[
(
[
25] E.J. Lorand and E.I. Edwards, J. Am. Chem. Soc., 77 (1955) 4035
26] Partenheimer, J. Mol. Catal., 206 (2003) 105-119
27] Partenheimer, J. Mol. Catal. 206 (2003) 131-144
28]. J. F. Goodman, P. Robson, J. Chem. Soc., (1963) 2871
29] L. Kuo-Tseng; L., Shih-Wei , Appl Catal A, 340 (2008) 271-277
30] W. Partenheimer, Adv. Synth. Catal. 346 (2004) 297-306
31] A. S. Hay, H. S. Blanchard. Can. J. of Chem. 43 (1965) 1306(1965)
32] W. Partenheimer, observation in our laboratory.
33] G. Valehrachova, M. Hronec, and V. Vesely, Coll. Czech. Chem. Commun. 51
1986) 2574
34] J. K. Okunowski, H. E. van Dam, H. van Bekkum, Recl. Trav. Chim Pays-Bas, 109
1990)103-106
35]Y. Kamiya, T. Nakajima, K. Sakoda. Bull. Chem. Soc. Japan, 39 (1966) 2211-2215
Page 19 of 30

2
0
Figure 1. Rate of reaction when ammonium acetate additions are made to cobalt(II)
O
perchlorate ( 0.040M) in oxygenated acetic acid at 114 C
Figure 2. Vent carbon dioxide and methane formation when ammonium acetate additions
O
are made to cobalt(II) perchlorate ( 0.040M) in acetic acid at 114 C
Figure 3.
o
Rate of oxidation of p-xylene at 95 C as a function of water concentration.
[
4 2 4 2
Co(II)(ClO ) ] = [Mn(BF ) ]= 0.0200M. [NaBr]=0.0400M. Acetate source was
NH OAc
4
Figure 4.
Carbon dioxide selectivity during oxidation of p-xylene as a function of water. Initial
conditions the same as those on figure 3.
Figure 5.
Carbon monoxide selectivity during oxidation of p-xylene as a function of water.
Initial conditions the same as those on figure 3.
Figure 6. The effect of variation of the pH on the Gibbs energy of rxns 4,5 (given in text).
The pH and free energy are given in arbitrary units to illustrate the overall effect of rxns
4
,5.
Figure 7. Effect of pH on the selectivity of the cyclohexyldienyl radical
Page 20 of 30

2
1
Figure 1
0.2
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
0
5
10
15
acetate/Co, mol/mol
Page 21 of 30

2
2
Figure 2.
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Vent CO2, %
CH4, %, x 10
0
5
10
15
acetate/(Co+Mn), mol/mol
Page 22 of 30

2
3
Figure 3
1
0
9
8
7
6
5
4
3
2
1
0
5
1
1
1
% w ater
0% w ater
3% w ater
7% w ater
0
1
2
3
4
5
acetate/(Co+Mn), mol/mol
Page 23 of 30

2
4
Figure 4
1
1
2
0
8
6
4
2
0
5
1
1
1
% water
0% water
3% water
7% water
0
1
2
3
4
5
acetate/(Co+Mn) ratio
Page 24 of 30

2
5
Figure 5
3
2
1
0
.5
2
5
1
1
1
% w ater
0% w ater
3% w ater
7% w ater
.5
1
.5
0
0
1
2
3
4
5
acetate/(Co+Mn), mol/mol
Page 25 of 30

2
6
Figure 6
1
2
0
8
6
4
2
0
1
sum of rxn 4 and 5
rxn 4
rxn 5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
pOAc (arbitrary units)
Page 26 of 30

2
7
Figure 7
CH3
[
5]
OH
CO2
CH3
lower pH
(2)
CH3
HO.
cresol
CH2.
+
.
(
1)
(
3)
H
OH
+
+
H , -H O
O2
4)
2
higher pH
desired products
(
benzyl radical
Page 27 of 30

2
8
Graphical abstract
8
7
6
5
4
3
2
1
0
1
0% water in HOAc
1
3% water in HOAc
0
0.5
1
1.5
2
2.5
3
3.5
4
acetate concentration
Page 28 of 30

2
9
Table 1. Rates and COx selectivity for p-xylene oxidation using HBr and NaBr at
different water concentrations
catalyst
catalyst], mM
Co/Mn/Br Co/Mn/Br Co/Br
2.0/2.0/4.0 2.0/2.0/4.0 4.0/4.0 4.0/4.0 4.0/4.0 4.0/4.0
Co/Br Mn/Br Mn/Br
[
H O,
2
%
0.3
5.0
10.0
13.0
0.3
NaBr
5.40
HBr
5.24
NaBr
5.76
4.61
4.68
4.03
0.023
0.025
0.026
0.037
0.0021 0.023
0.0037 0.0038
0.0039 0.0045
0.0056 0.0056
2.7
6.1
9.6
12
HBr
5.30
4.72
3.77
NaBr
1.34
0.61
0.57
0.46
0.14
0.16
0.20
---
---
---
---
---
5.2
5.8
6.2
6.5
HBr
1.03
0.96
1.23
1.11
---
---
---
---
---
---
---
---
4.8
5.6
6.2
7
Rate,cc O
2
/min
rate,cc O
rate,cc O
2
/min
/min
4.46
4.09
2.26
4.10
3.62
3.91
2
rate,cc O /min
3.62
2
CO
2
select.
0.016
0.018
0.021
0.034
0.0014
0.0017
0.0026
---
5.4
4.5
8.6
10.2
0.015
0.017
0.020
0.038
---
0.0022
0.0028
0.0051
3.1
0.021
0.020
0.019
0.034
CO select.
5.0
2
CO select.
10.0
13.0
0.3
2
CO
2
select.
CO select.
CO select
CO select
CO select
Completion,%
Completion,%
Completion,%
Completion,%
5.0
10.0
13.0
0.3
5.0
10.0
13
---
3
8.8
10.9
6.8
8.8
10.2
Table 2 Selected Thermodynamic and Kinetic Values Related to the Initiation of
Hydrocarbon Oxidation
-
-
-
-
X=F X=Cl X=Br X=I
(
1)
Free energy change for rxn 13, kcal/mol
Relative rate, rxn 13
Free energy change for rxn 14, kcal/mol
43.6 18.9
0.00 0.003 1.00
-43.0 -11.8 4.0
6.75
-9.40
>>1
20
(2)
(
3)
1
. These are calculated from the redox potential of cobalt(II)(III) of 41.1 kcal/mol measured
in water; from the values of the oxidation potentials of the halogens in water (66.2,
3
1.3, 25.1, and 14.1 kcal/mol for F,Cl,Br,I respectively); and from the gas phase
dissociation energies of 37.5, 57.8, 46.1, and 35.8 of F , Cl , Br , and I .
2
2
2
2
o
2
3
. Measured in the authors laboratory in 80% acetic acid-20% water at 25 C
. These were calculated from from the bond dissociation energies of 128, 96.8, 81.0, and
6
5.0kcal/mol for HF, HCl,HBr, and HI respectively; and for toluene which is 85
kcal/mol.
Page 29 of 30