Prevention of manganese precipitation during the continuous selective partial oxidation of methyl aromatics with molecular oxygen in supercritical water

Article abstract of DOI:10.1002/adsc.200900143

Presented here is an investigation of the activity and recovery of the homogeneous manganese/bromide catalyst during the continuous flow oxidation of o-xylene, as model substrate, with molecular oxygen (O₂) in supercritical water (scH₂O). Two strategies are discussed for preventing catalyst precipitation, mainly in the form of oxides such as manganese(IV) oxide, The first strategy involves varying the manganese:bromide ratio using either manganese(II) acetate or manganese (II) bromide in the presence of hydrobromic and other acids. The results show that the effect of acidity and bromide concentration plays an important role in preventing the manganese/bromide catalyst from precipitating. The second strategy uses aromatic carboxylic acids in combination with the manganese/bromide catalyst, particularly benzoic acid, which improves the catalyst recovery dramatically over a certain range of acid concentrations. Our studies show how the presence of an organic acid and/or its precursors is important in stabilising the catalyst. Our results are rationalised on the basis of a tentative reaction mechanism.

Full text of DOI:10.1002/adsc.200900143

FULL PAPERS
DOI: 10.1002/adsc.200900143
Prevention of Manganese Precipitation during the Continuous
Selective Partial Oxidation of Methyl Aromatics with
Molecular Oxygen in Supercritical Water
a,c
a,e
a
Joan Fraga-Dubreuil, Eduardo Garcia-Verdugo, Paul A. Hamley,
a
b
b
b
Eduardo Perez, Ian Pearson, W. Barry Thomas, Duncan Housley,
Walt Partenheimer, * and Martyn Poliakoff *
d,
a,
a
University of Nottingham, Nottingham NG7 2RD, U.K.
Fax : (+49)-115-951-3058; phone: (+49)-115-951-3520; e-mail: Martyn.Poliakoff@nottingham.ac.uk
INVISTA Performance Technologies, Wilton, Cleveland, TS10 4RF, U.K.
INVISTA Textiles (U.K.) Limited, Wilton, Cleveland, TS10 4RF, U.K.
Current address: 7253 SW 26th Ave., Portland, OR 97219, USA
b
c
d
E-mail: wpartenheimer@msn.com
University Jaume I, Department of Organic and Inorganic Chemistry, Campus Ctra. Riu Sec., 12071 Castellꢀn, Spain
e
Received: March 5, 2009; Revised: June 1, 2009; Published online: July 27, 2009
Abstract: Presented here is an investigation of the manganese/bromide catalyst from precipitating. The
activity and recovery of the homogeneous manga- second strategy uses aromatic carboxylic acids in
nese/bromide catalyst during the continuous flow ox- combination with the manganese/bromide catalyst,
idation of o-xylene, as model substrate, with molecu- particularly benzoic acid, which improves the catalyst
lar oxygen (O ) in supercritical water (scH O). Two recovery dramatically over a certain range of acid
2
2
strategies are discussed for preventing catalyst pre- concentrations. Our studies show how the presence
cipitation, mainly in the form of oxides such as man- of an organic acid and/or its precursors is important
ganese(IV) oxide, The first strategy involves varying in stabilising the catalyst. Our results are rationalised
the manganese:bromide ratio using either mangane- on the basis of a tentative reaction mechanism.
se(II) acetate or manganese(II) bromide in the pres-
ence of hydrobromic and other acids. The results
show that the effect of acidity and bromide concen- Keywords: aromatic acids; catalyst stability; manga-
tration plays an important role in preventing the nese bromide; supercritical water oxidation; xylenes
Introduction
of p-xylene to terephthalic acid due to its importance
in the chemical industry.
[
9–12]
Research activity in the field of supercritical water,
Our studies using continuous flow systems
have
scH O (T =3748C, p =221 bar) or more generally in shown that scH O can provide: 1) an environmental
2
c
c
2
high-temperature water has been growing for more advantage by eliminating the use of acetic acid as a
than 20 years because of the unusual properties of solvent and reducing consequent waste due to un-
[
1]
water under these conditions. A wide range of wanted by-products formation (mainly carbon diox-
chemical reactions have been studied where scH O ide, carbon monoxide, methyl bromide and organic
2
[
16]
plays the role of a solvent and/or reactants and/or cat- acetates); 2) an economic advantage resulting from
[2–5]
alyst.
The scH O technology is particularly well- faster reactions as well as operating at higher temper-
2
known for efficient destruction of organic wastes by ature which allows for more efficient energy recovery
[6]
total oxidation. By contrast, the homogeneously cat- from these highly exothermic reactions; 3) process in-
alysed oxidation of methyl aromatics in ꢁhotꢂ and tensification by reducing the size of the reactor and
scH O has also shown considerable potential for the the residence time compared to the conventional
2
selective production of carboxylic acids, provided that acetic acid process.
[
7–16,17b]
the reaction conditions are well controlled.
of the most studied reactions is the selective oxidation have been aerobically oxidized using around 30 dif-
ferent combinations of soluble metal/bromide cata-
One
Since the 1960s, at least 251 different substrates
AHCTUNGTRENNUNG
1866
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876

Prevention of Manganese Precipitation during the Continuous Selective Partial Oxidation
FULL PAPERS
[
17a]
lysts.
The solvent typically used is acetic acid and the addition of bromide to the metal-catalysed system
[
15]
the most common catalyst comprises a mixture of co- increases the activity and yield; and 4) increasing
ACHTUNGTRENNUNGb alt(II) and manganese(II) acetates with hydrobromic the catalyst concentration gives higher yield and se-
acid. One of the drawbacks, in both scH O and the lectivity.
[
7,8,15]
2
conventional acetic acid process, is the fact that ho-
A simplified free radical chain mechanism for a
mogeneous metal catalysts, especially manganese, can manganese/bromide catalyst consists of initiation,
[
20,21]
hydrolyse to form insoluble metal oxides. The work- propagation, and termination steps.
ing oxidation states for manganese, Mn(II)–Mn(III), step forms the benzylic radical [Eq. (1)] where X*
have been identified by UV-Vis studies in acetic can be a oxidized bromide species, indicated by Br(0),
The initiation
AHCTUNGTRENNUNG
[18]
acid.
or other radicals formed in the reaction such as the
In a supercritical oxidation process for the produc- peroxy radical, ArCH OO*.
2
tion of aromatic carboxylic acids, the precipitation of
manganese oxide(s) would reduce or prevent the ef-
fective and economically essential recycle of catalyst
as well as reduce its activity and selectivity. In addi-
tion, the precipitation could hinder flow in a tubular
reactor, which would need to be cleaned or unblocked
The propagation step [Eq. (2)] is the rapid reaction
in order to continue operation. This paper demon- of the benzylic radical with molecular oxygen fol-
strates how the formulation of the catalyst can be im- lowed by the hydrogen atom abstraction of the
proved in order to maximise metal/bromide stability methyl group with the peroxy radical [Eq. (3)].
by largely eliminating precipitation of manganese
oxides in the scH O reaction system.
2
Reaction Mechanism and Possible Cause
of Catalyst Precipitation
Without a metal present the peroxide will rapidly
The interpretation of the experimental data in this dissociate above 808C to the undesirable, highly ener-
paper is very complex since a homogeneous aerobic getic, unselective hydroxyl radical (OH*) [Eq. (4)].
oxidation will be occurring, along with precipitation
of the catalyst metals. The precipitated catalyst metals
themselves can then potentially act as heterogeneous
oxidation catalysts. Our goal is to eliminate catalyst
precipitation so that only homogeneous oxidation
One of the purposes of the catalyst, Mn(II), is to
occurs. In our attempts to do so, we will make react selectively with the peroxides and peroxya-
[
22]
changes that will in most cases change the pH as well. cids, generating Mn AHCUTNGRNENUG( III), thus avoiding the forma-
In this section of the paper we will 1) give evidence tion of the hydroxyl radical [Eq. (5)].
that a free radical chain mechanism is occurring, 2)
give a simplified catalyst modified chain mechanism,
3) discuss how pH affects the solubility of the metals
and the free radical chain mechanism, 4) give a model
which rationalises the increased solubility of the cata-
lysts, and 5) give strategies for elimination of manga- MnAHCTUNGTRENNUNG
In manganese/bromide-catalysed chemistry, the
(III) then rapidly reacts with bromide to generate
nese precipitation.
an oxidised bromide species [Eq. (6)].
In acetic acid, there is considerable evidence that
the oxidation proceeds through a catalyst modified
free radical chain mechanism. There are a number of
phenomena in sub- and scH O that are the same as
2
those observed in acetic acid suggesting a similar
This species continues the reaction via hydrogen
mechanism. Some of the similarities are: 1) the se- abstraction reactions such as in the initiation step, Eq.
quence of oxidation products, i.e., p-xylene to 4-meth- (1). The Br(0), as well as the peroxy radical [Eq. (3)],
ylbenzaldehyde, to p-toluic acid, to 4-carboxybenzal- very selectivity reacts with the CÀH bonds of the
dehyde and finally to terephthalic acid, is the methyl group rather than those of the aromatic
[7,8,12]
[21]
same;
2) the uncatalysed oxidation in supercriti- ring.
cal water gives low terephthalic acid yields and is very
unselective producing large amounts of toluene; 3) oxidation in supercritical water. Firstly, the
Two types of catalyst precipitation are pertinent to
[
19]
Adv. Synth. Catal. 2009, 351, 1866 – 1876
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1867

Joan Fraga-Dubreuil et al.
FULL PAPERS
manganese ACHTUNGTRENNUNG( III) disproportionates to the highly insolu- soluble as the oxidation state of the metal increases,
ble, manganese(IV) dioxide. In the commercial pro- the degree of metal oxide/hydroxide precipitation
cess over-oxidation of manganese to manganese(IV) also increases. 4) As the pH increases the dispropor-
dioxide occurs when the water concentration in acetic tionation of manganese
acid is too high or the bromide/metals ratio is too
ACHTUNGTNERNUNG( III) will increase [Eq. (7)].
[23,24]
low.
Black manganese(IV) dioxide co-precipitates
with the terephthalic acid, not only contaminating the
solids with manganese, but also changing the colour
of the solids from white to grey. Secondly, a mixture
5) Maximum yield and selectivity to a given aro-
of manganese(II) bromide and manganese(II) acetate matic acid occur at an intermediate pH, i.e., not too
in supercritical water at 4008C precipitates as manga- low or not too high. This has been demonstrated for
[
17a,22]
nese(II) oxide at concentrations similar to those when oxidation of p-xylene in acetic acid.
For the
these manganese compounds are used as a homogene- metal to act as a catalyst, a) it must be both oxidised
[25]
ous catalyst.
Dissolving manganese(II) acetate in and reduced at equal rates and b) it must fulfil its
water and heating it to form supercritical water, is a function as a catalyst. The function of manganese is
method for generating nanoparticles of manganese to selectively react with the hydroperoxides and oxy
[26,27]
oxides.
radicals (denoted by [O]) [Eq. (8)] and subsequently
Current evidence suggests that catalyst metal pre- be reduced by the bromide anion [Eq. (9)].
cipitation is undesirable because: 1) in acetic acid,
manganese(IV) dioxide precipitation increases the
rate of carbon dioxide and methyl bromide forma-
[28]
tion, 2) it is known that manganese(IV) dioxide is
an excellent heterogeneous catalyst in supercritical
water for converting organic compounds to carbon di-
[
7]
oxide, i.e., unselective oxidation, and 3) the concen-
tration of the homogeneous metal/bromide catalyst
If the pH is too low, then the redox potential of the
will decrease causing a decrease in the rate of reac- manganese(II)–(III) couple will be too high and Eq.
tion as well as a decrease in selectivity since thermal, (8) will not occur. If the pH is too high, then
rather than metal-catalysed, decomposition of the hy- manganese ACHUTNGRENNUG( III) is too weak an oxidant for Eq. (9) to
droperoxides increases. Thermal decomposition leads occur. Obviously, an intermediate pH where both re-
to such undesirable products as toluene, benzoic acid, actions occur as rapidly as possible will provide the
[28]
and cresols.
The pH of the solvent plays a critical role in the
best catalysis.
To summarise, an increase in pH will increase the
successful oxidation of methyl aromatic compounds tendency for precipitation of metal oxides, will in-
and the understanding of this effect is necessary to ra- crease the propensity for dioxygen to oxidise the
tionalise the experimental data in this paper. pH is manganese to more insoluble oxides, will increase the
critical for the following reasons. 1) The strength of tendency towards disproportionation to the highly in-
manganese ACHTUNGTRENNUNG( III) as an oxidant decreases with an in- soluble manganese(IV) dioxide and, if the pH is too
crease in pH. For manganese(II)–(III) the redox po- low, will decrease the rate of formation of the desired
tential is 1.54 V at pH 1 but decreases to 0.15 V at a product of the oxidation reaction, i.e., the aromatic
[23]
pH 8. In ambient water, the reason for the change acid. The optimal state is operating at a pH to give
in redox potential is because the structure of the man- maximum aromatic yield but which is not too high so
ganese(II) changes. The structure at pH 1 is [Mn(II)- that significant amounts of the catalyst metal precipi-
2+
ACHTUNGTRENNUNG( H O) ] but with increasing hydroxide anion con- tates.
2 6
+
centration it changes to [Mn(II)
Mn(II)(H O) (OH) ]. (This is somewhat over-simpli- tions of water at ambient temperature. It is well es-
fied for the sake of argument; various polynuclear tablished that the physical properties of water greatly
A
H
U
G
R
N
U
G
The above discussion has been based on observa-
2
5
[
ACHTUNGTRENNUNG
2
4
2
[
29]
compounds are known to form ). 2) The propensity change when the temperature is increased to super-
for metal precipitation as their oxide/hydroxides will critical temperatures, >3748C. For example, the di-
increase with increase in pH. [Mn(II)
A
H
U
G
R
N
U
G
ACHTUNGTRENNUeGN lectric constant decreases from 80 to 4. As a conse-
2
4
2
which occurs at high pH, being a neutral compound quence, coordination compounds do not ionise readi-
and containing the hydroxide anion is the precursor ly, i.e., they are charge-neutral, their coordination
[
29]
to metal oxide/hydroxide precipitation.
3) As the number decreases from 6 to 4, and the geometry
pH increases, the propensity increases for O to oxi- changes from octahedral to tetrahedral. This has been
2
dise manganese(II) spontaneously. It is known that di- demonstrated for MnBr which exists predominately
2
2+
oxygen can oxidise magnanese(II) to manganese
A
H
U
G
R
N
N
ACHTUNGTRNENU(G H O) ] in ambient water, but changes to tet-
2 6
[25]
[30]
AHCTUNGTRENNG(U H O) Br ] in scH O. Furthermore, for
2 2 2 2
1868
asc.wiley-vch.de
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876

Prevention of Manganese Precipitation during the Continuous Selective Partial Oxidation
FULL PAPERS
Ni, Cr and Fe, metal oxides form at lower pHs as the added, the driving force is the reaction of the aromat-
temperature is increased; the pH shifts from 8 in am- ic acid with the hydroxyl ligand to produce water and
[
32]
bient water to 5 at ca. 3008C. Furthermore, the the aromatic acid salt of manganese(II). This reac-
strength of acids changes as the temperature of water tion will avoid formation of manganese(IV) dioxide
increases. For example, hydrochloric acid is a strong since the concentration of the MnÀOH precursor has
acid in ambient water but becomes a weak acid at been decreased.
À4
3
2
808C with a dissociation constant of 2.6ꢄ10 (at
There are a number of strategies for preventing
[39]
7 MPa).
The dissociation constant of acetic acid precipitation of manganese(IV) dioxide in scH O. The
decreases from 1.8ꢄ10 at ambient water to 3.0ꢄ first involves designing a flow reactor for the best pos-
at 3508C. This is the probable reason why man- sible mixing between the methyl aromatic, dioxygen
ganese(II) acetate has been observed to precipitate as and the catalyst.
2
À5
À8
1
0
[11,12]
One can also apply one or more
manganese(II) oxide in scH O. The acetate anion is a of the methods used currently in acetic acid/water sol-
2
very weak base in ambient water, but is stronger in vents. 1) By increasing the bromide concentration by
scH O so it hydrolyses to Mn(II)(OH) and dehy- addition of a bromide source to manganese, i.e., in-
2
2
drates to Mn(II)O.
creasing the ratio Br/Mn so that the rate of reduction
(III) to Mn(II) by the bromide ions [Eq. (9)] is
with the discussion above as well as the experimental increased relative to the rate of disproportionation of
data that will be subsequently presented. All of the Mn(III) to Mn(II) and Mn(IV), Eq. (7). 2) By finding
In Figure 1, we present a model which is consistent of MnACHTUNGTRENNUNG
AHCTUNGTRENNUNG
reactions shown are balanced reactions. The coordina- an optimum pH, see discussion above. 3) By adding
tion compounds are four-coordinate tetrahedral com- metals which have a higher Lewis acidity than manga-
pounds consistent to what has been found in nese(IV). In the presence of a strong Lewis acid
[
25]
[19,23]
scH O.
The initial catalyst, [Mn
A
H
U
G
R
N
U
G
the concentration of
2
2
2
2
oxidised, as described above, to a manganese
cies, Eq. (13). If the pH of the solution is increased of the Mn salt and precursor of Mn oxides, is re-
and hydrolysis occurs, one will lose hydrobromic acid duced:
(III) spe- manganese hydroxides MnÀOH, formed by hydrolysis
and Eq. (10) occurs. Oxidation of the product of Eq.
(
10), Eq. (14), will produce a dihydroxy manganese-
ACHTUNGTRENNUNG( III) compound. Oxidation of manganese(II) bromide,
Eq. (13), necessitates that a monohydroxy species
must form. Disproportionation occurs via Eq. (16) Results and Discussion
and Eq. (17) via species that have manganeseÀOH
bonds which are precursors to oxide formation. The Since several of the products from p-xylene oxidation,
model predicts that addition of excess hydrobromic particularly terephthalic acid, are poorly soluble in
acid to manganese(II) bromide will avoid mangane- cold water, it is necessary to quench the hot reaction
se(IV) dioxide formation either by Eq. (11) or by re- mixture with an NaOH aqueous solution at the end
[31]
[9–12]
versing Eq. (17). The model also predicts that Eq. of our flow reactor.
The sodium hydroxide reacts
(
12) will occur, if an acid, such as an aromatic acid, is with the aromatic acids to form water-soluble sodium
Figure 1. A model that rationalises the observed effects of bromide and aromatic acid addition to a Mn/Br catalyzed oxida-
tion of o-xylene. ꢁ[O]ꢂ represents an oxidant such as a peroxy radical or a peroxide. ArCOOH is an aromatic carboxylic acid
(
not an aliphatic carboxylic acid such as acetic acid).
Adv. Synth. Catal. 2009, 351, 1866 – 1876
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1869

Joan Fraga-Dubreuil et al.
FULL PAPERS
salts. However, this quench also causes any dissolved
Benzoic acid is the product of both Eq. (18) and
catalyst to precipitate as metal hydroxide/oxides in Eq. (19). Benzoic acid, a product which is always seen
the heat exchanger at the end of the reactor. To avoid in these oxidations, is also produced via the non-cata-
[
28]
this problem, it was decided to study the continuous lysed thermal decomposition of the hydroperoxides.
oxidation of o-xylene to o-phthalic acid (OPA),
[
31,32]
To enhance our understanding of the factors that
because water-soluble products can be obtained under influence the stability, i.e., solubility of the manga-
similar oxidation conditions without using sodium hy- nese, we first present data in the absence of o-xylene,
droxide. This switch from the para to the ortho see Table 1. One is observing the intrinsic solubility of
isomer allowed a direct analysis of the metal content manganese(II) bromide in the absence of the reaction
in the collected homogeneous samples by atomic ab- intermediates and products shown in Figure 2. Benzo-
[34]
sorption. Figure 2 gives the observed intermediates ic acid has limited solubility at room temperature,
and products from o-xylene oxidation. o-Xylene is dif- so the feed mixture of manganese(II) bromide and
ferent from p-xylene in that the phthalide (PTL, the benzoic acid was heated to 808C to obtain a homoge-
internal ester) can form, which can be either oxidised neous solution prior to its entry into the reactor.
to o-phthalic acid directly or hydrolysed and oxidised From Table 1 the following points are clear. 1) With-
to 2-CBA. o-Xylene is also different from p-xylene in out dioxygen present, 12% of the manganese precipi-
[
33]
that the rates of thermal decarboxylation, Eq. (18),
tates, presumably as manganese(II) oxide/hydroxide
[
27]
and catalytic decarboxylation are both faster for o- see entry 1. So the manganese(II) bromide itself is
phthalic acid than terephthalic acid.
slightly insoluble under these conditions. 2) When the
Figure 2. Oxidation pathway of o-xylene to o-phthalic acid (OPA)
Table 1. The effect of benzoic and hydrobromic acid on the stability of manganese(II) bromide under SCWO conditions in
the absence of o-xylene. Initial concentration of Mn in reactor is 10.3 mM.
[b]
[c]
[d]
Entry [BA], mM O (Y/N)
BA/Mn [mol/mol] Br/M [mol/mol] BA recovery [%]
Mn recovery [%]
Colour
2
1
2
3
4
5
6
7
8
9
1
0
0
N
N
N
N
N
Y
Y
Y
Y
Y
0.00
0.00
0.80
1.32
2.65
0.00
0.00
0.40
1.32
2.65
2.00
2.40
2.00
2.00
2.00
2.00
2.40
2.40
2.00
2.00
–
–
99
99
98
–
88
99
84
99
99
48
66
60
85
90
colourless
colourless
colourless
colourless
colourless
yellow
yellow
yellow
pink
pink
[
a]
0.0246
0.0409
0.0819
0
[a]
0
–
[a]
0.0123
0.0409
0.0819
99
98
95
0
[
[
[
[
a]
Addition of 4.17 mM HBr.
Oxygen feed, Y=yes, N=no.
Determined by HPLC analysis.
Recovered manganese as determined by atomic absorption.
b]
c]
d]
1870
asc.wiley-vch.de
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876

Prevention of Manganese Precipitation during the Continuous Selective Partial Oxidation
FULL PAPERS
experiment is repeated with O present, entry 6, then benzoic acid is not occurring under these experimen-
2
5
4
2% of the manganese precipitates. So about 52À12= tal conditions.
0% of manganese has been oxidised to either In Table 1, it was demonstrated that the addition of
(III) or manganese(IV). We conclude that benzoic acid to manganese(II) bromide enhanced the
manganese
ACHTUNGTRENNUNG
[32]
the pH is sufficiently high that spontaneous oxidation manganese recovery and hence its solubility.
In
of manganese(II) can occur. Evidence that oxidation Table 2, the same experiment is performed but now
has occurred is also indicated by the change in colour during the oxidation of o-xylene. The same phenom-
of the effluent from colourless to yellow. This colour enon is seen – manganese recovery increases with in-
could be from the oxidation of manganese or bromide creasing benzoic acid concentration from 76% with-
or both. The dioxygen to manganese molar ratio is 3 out benzoic acid monotonically increasing to 97%
under these conditions. 3) Addition of hydrobromic with benzoic acid present, see entries 1–4 in Table 2.
acid to the manganese(II) bromide, in a molar ratio Table 2 also illustrates that a number of different aro-
of 0.4 HBr:MnBr , significantly increases the recovery matics have an identical effect on solubility – o-
2
of the manganese(II) bromide, from 88% to 99% in phthalic acid (OPA) (entries 5 and 6), trimellitic acid
the absence of dioxygen (entries 1 and 2), and from (TMA) (entry 7), mellitic acid (MLA) (entry 8). The
48% to 66% under dioxygen (entries 6 and 7). So a non-aromatic acid, acetic acid (AA) (entry 9), de-
decrease in pH does indeed enhance the solubility creases the solubility. Along with the increase in man-
suggesting the formation of Mn(II) hydroxides does ganese solubility, a modest increase in o-phthalic acid
occur, see Eq. (11) of Figure 1. It should be noted (OPA) yield and selectivity is observed when an aro-
that the acid strength of hydrobromic acid is probably matic acid is present, while the presence of acetic acid
substantially diminished at these high temperatures significantly decreases the yield.
similar to what has been observed for hydrochloric
One would anticipate that the solubility of manga-
[39]
acid. 4) Similarly, addition of benzoic acid either in nese during o-xylene oxidation to be higher because
N or O , enhances the solubility of the MnBr , com- o-toluic acid and o-phthalic acid are forming as one
2
2
2
pare entries 1 with 3, 4 and 5 and entry 6 with 9 and proceeds down the tubular reactor. As confirmation
0. This change in recovery suggests the incorporation of this, the solubility of manganese is 76% during the
1
of the benzoic acid into the coordination sphere of oxidation of o-xylene (entry 1, Table 2) while it is
manganese(II), see Eq. (12) of Figure 1. There will be only 48% in its absence (entry 6, Table 1).
a increase in pH as well since aromatic acids still have
Addition of benzoic acid and the other aromatic
measurable acid character at high water tempera- acids further enhanced the solubility because the Mn-
[39]
tures.
Under dioxygen, the colour also changes aromatic acid interaction occurs when the reagents
from yellow to pink with increasing benzoic acid con- are mixed together at the beginning of the reactor
tent consistent with less oxidation of manganese(II) before oxidation has formed significant amounts of
to manganese ACHTUNGTRENNUNG( III). 5) The recovery of the benzoic aromatic acids. The enhanced yield to o-phthalic acid
acid after the experiment is close to quantitative. This seen in Table 2 may be due to three effects – the de-
suggests that thermal or catalytic decarboxylation of creased amount of catalyst precipitation will provide
Table 2. The effect of selected carboxylic acids on the oxidation of o-xylene and manganese recovery in scH O. Catalyst was
2
manganese(II) bromide at an initial concentration of 10.3 mM.
[a]
[b]
Selectivity [%]
ACHTUNGTRENNUNG[ Acid], mM Acid/Mn [mol/mol] OPA Yield [%] OPA 2-CBA PTL BA o-TolA TALD
Mn [%]
[c]
Entry Acid
1
2
3
4
5
6
7
8
9
None 0.0
0.0
0.8
1.3
2.7
1.4
2.3
1.3
0.2
5.4
62
73
70
72
74
78
73
70
43
91.8 1.0
92.0 1.0
94.0 0.8
92.4 0.5
93.7 0.6
96.5 0.5
94.3 0.3
94.4 0.4
95.8 0.4
0.8
2.1
1.3
0.8
3.4
1.3
1.1
1.0
2.7
6.2
–
0.0
0.0
0.2
0.0
0.0
0.0
0.0
0.0
0.0
76
78
94
97
88
91
94
92
64
BA
8.2
3.4 1.3
2.9 0.7
6.2 0.0
3.4 0.8
1.2 0.4
3.8 0.5
3.7 0.3
1.0 1.0
BA
BA
OPA
OPA
13.6
27.3
14.1
23.3
TMA 13.7
MLA 2.3
AA
55.6
[
a]
Selectivity for the identified aromatic compound calculated as the molar concentration of that compound relative to the
total molar concentration of all identified aromatic compounds.
Recovered manganese as determined by atomic absorption.
BA =benzoic, OPA=o-phthalic, TMA=trimellitic (1,2,4-tricarboxybenzene), MLA=hexacarboxybenzene, AA=acetic
acid.
[
[
b]
c]
Adv. Synth. Catal. 2009, 351, 1866 – 1876
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1871

Joan Fraga-Dubreuil et al.
FULL PAPERS
a more active catalyst (due to its higher concentra-
The data in Table 2 are consistent with an insignifi-
tion), decreased unselective oxidation due to the pre- cant amount of thermal decarboxylation occurring
cipitated Mn oxide solids, and a change in structure during the reaction. The relative degrees of thermal
of the catalyst that could be more active and/or selec- decarboxylation are trimellitic acid@o-phthalic acid>
[
33]
tive. In the model given in Figure 1, this could occur benzoic acid. The product of thermal decarboxyla-
via Eq. (12). There are other possibilities such as the tion of o-phthalic acid is benzoic acid. Addition of o-
reaction given in Figure 3 where hydrobromic acid is phthalic acid to the oxidation of o-xylene did not in-
released which could further enhance the solubility of crease the selectivity to benzoic acid, as can be seen
the Mn.
by comparing entry 1 with entries 5 and 6 in Table 2.
The data are also consistent with the fact that catalyt-
ic decarboxylation, although it is almost certainly oc-
curring, is not a dominant reaction. Catalytic decar-
boxylation causes a decrease in rate of reaction be-
cause Eq. (9), which propagates the reaction, is in
competition with the manganese AHCUTNGRNENUG( III) in the catalytic
decarboxylation, see Eq. (19). It is established that,
during the Co/Mn/Br-catalysed oxidation in acetic
acid, the addition of aromatic acids decreases the rel-
ative rate of reaction in the order: benzoic acid
Figure 3. The possible reaction of benzoic acid with the
n+
Mn centre in scH O.
2
(
1.0)<o-phthalic acid (2.2)<trimellitic acid (3.8)<
hemimellitic acid (1,2,3-tricarboxybenzene; 16). Thus
The equilibrium constants for the formation of one might expect to see lower yields in the presence
metal-carboxylate complexes in ambient water are of o-phthalic acid, trimellitic acid, and mellitic than
small but are much higher in non-aqueous solvents when benzoic acid is present. This was not observed.
[35]
such as acetone. This suggests that the formation of The effect of addition of aromatic carboxylic acids,
metal-carboxylate complexes in scH O could be quite ArCO H, has also recently proved to be beneficial for
2
2
favourable.
improving the activity of the Mn/Co-catalysed air oxi-
[
36]
Significantly higher manganese precipitation and dation of cyclohexane to adipic acid.
lower yields are observed in the presence of acetic In Table 1, we found that both benzoic acid and hy-
acid. It is not obvious to us why this should be so, drobromic acid increased the solubility of mangane-
since the opposite effect is seen with aromatic acids. se(II) bromide in the presence of air even without o-
À5
À5
The acidity of benzoic acid (K =6.5ꢄ10 ) is some- xylene present. In Table 3, we confirmed that addition
d
what higher than acetic acid (K =1.8ꢄ10 ) in ambi- of hydrobromic acid also increased its solubility
d
[
35]
ent water, but both are weak acids. The acidity of during the oxidation of o-xylene with a resulting in-
acetic acid is reduced as the temperature of water in- crease in o-phthalic acid yield and selectivity, compare
creases hence may be substantially different from that entry 1 with entries 2 and 3. The yield is lower at a
of benzoic acid at the temperatures used in these ex- Br:Mn ratio of 4.0 than at 2.4 (entries 3 and 2, respec-
periments. More subtle structural differences in the tively) possibly because the pH is too low and too far
coordination chemistry, for example, the steric bulk of from its optimum value (see discussion above). Other
the aromatic ring, which affect the precipitation of interpretations are also possible. The recation of ther-
[39]
[
33]
the metal, may also be present.
mal decarboxylation is acid catalysed, so this rate
Table 3. Oxidation of o-xylene with different catalytic mixtures. Initial concentration of manganese in reactor is 10.3 mM.
[a]
[b]
Entry Catalyst
Br/Mn [mol/mol] Yield [%]
Selectivity [%]
OPA 2-CBA PLT BA o-TALD
Mn [%]
1
2
3
4
5
6
7
8
MnBr2
2.0
2.4
4.0
2.0
62.4
66.0
64.5
63.9
58.3
8.0
91.8
93.4
91.5
94.4
73.3
21.3
28.2
80.4
1.0
0.2
0.7
0.5
0.8
5.7
14.4
0.2
0.8
0.8
0.7
0.4
3.9
61.2 7.6
53.4 1.8
1.0
6.2
5.6
7.1
4.6
0.0
0.0
0.0
0.0
76
88
99
85
61
14
49
89
MnBr +HBr (4.17 mM HBr)
2
MnBr +HBr (20.8 mM HBr)
2
MnBr +HCl (20.8 mM HCl)
2
MnBr +NaBr (4.17 mM NaBr) 2.4
21.9 0.0
2
Mn
Mn
Mn
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(OAc)
0.0
0.0
2.0
4.1
2.2
2
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(OAc) +HCl (20.8 mM)
16.8
63.4
2
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(OAc) +HBr (20.8 mM)
18.3 0.0
2
[
a]
Selectivity for the identified aromatic compound calculated as the molar concentration of that compound relative to the
total molar concentration of all identified aromatic compounds.
Recovered manganese as determined by atomic absorption.
[
b]
1872
asc.wiley-vch.de
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876

Prevention of Manganese Precipitation during the Continuous Selective Partial Oxidation
FULL PAPERS
could be increasing, which is consistent with the assuming that the precipitated catalyst does not have
higher observed benzoic acid selectivity. Also the significant activity). Thus the relative activities of cat-
HPLC analysis suggests that brominated by-products alysts in acetic acid and scH O are in the same order.
2
may have also formed, accounting for the lower yield. However in acetic acid, one is comparing the intrinsic
A non-bromide-containing mineral acid, hydrochloric activities of the homogeneous catalysts because they
acid, also increases the manganese recovery and in- are 100% soluble. But in scH O they are not com-
2
creases the yield, compare entries 1 and 4. Sulfuric pletely soluble and the relative activities of the cata-
acid, H SO , was also tried, but the reactor tubing lysts and their relative solubilities are the same. So
2
4
rapidly become plugged due to formation of insoluble one cannot deduce the intrinsic activities of the cata-
manganese(II) sulfate. Addition of sulfuric acid to a lysts in scH O.
2
Co/Mn/Br catalyst in acetic acid also causes precipita-
Another example can be seen in entries 6–8 in
tion of the metals as their sulfates. The replacement Table 3 where one compares the addition of hydro-
of hydrobromic acid by sodium bromide results in a chloric acid and hydrobromic acid to a manganese(II)
dramatic decrease in Mn solubility as well as lower o- acetate catalyst. The relative activity of the catalysts,
phthalic acid yield and selectivity, see entries 2 and 5. based either on the yields or selectivities, is Mn/Br>
This clearly demonstrates that there is a pH effect Mn/Cl>Mn. This is the same as is found in acetic
[
17a,37]
present in these reactions. At 3808C, sodium chloride acid.
But a confounding factor in scH O is that
2
increases the pH of water because one is adding chlo- solubility of the catalyst occurs in the same order:
[39]
ride, which is the anion of a weak acid, hydrochlo- Mn/Br>Mn/Cl>Mn. So we cannot assign an intrinsic
ric acid. A similar increase in pH is probably also oc- activity to these catalysts in scH O because of the
2
curring when sodium bromide is being added in these concentration differences of the homogeneous cata-
experiments. The unusually high selectivity to benzoic lyst. In acetic acid, the Mn/Cl catalyst is less active
acid in entry 5 may be due to the large decrease in than the Mn/Br because the reaction of manganese-
catalytic activity which leads to the unselective forma-
tion of toluene, and its subsequent oxidation to ben- and because the Cl(0) species is much less selective
ACHTUNGTRENNU(GN III) with chloride is much slower than with bromide
[
28]
[17a,37]
zoic acid.
In acetic acid, the manganese(II) acetate catalyst
has very weak activity, while manganese(II) bromide acetate/hydrobromic acid mixtures, increasing the
than Br(0).
Table 4 demonstrates that, using manganese(II)
[
17a,37]
is much more active.
The same is true in scH O, Br:Mn ratio up to a value of 1.6 increases the yield
2
see entries 1 and 6 in Table 3. However, in acetic acid, but the selectivity continues to increase up to a
both manganese(II) acetate and bromide are com- Br:Mn value of 2.4. The solubility of the Mn also in-
pletely soluble at these concentrations and one is ob- creases in this order. Changing the bromide:metal
serving the intrinsic activity of these catalysts. This is ratio is a common method for optimising metal/bro-
NOT true in scH O because the solubility of the cata- mide catalysts and many studies have been reported
2
[
17]
lyst is increasing as the yield increases, i.e., in entry 1 for different methyl aromatic compounds. General-
we have a yield of 62% with a manganese recovery of ly one finds that increasing the bromide:metal ratio
7
1
6% and in entry 6 we have a yield of 8% with only increases the activity and selectivity up to some maxi-
4% manganese recovery. Since yield generally in- mum value and beyond that maximum the activity
creases with catalyst concentration, one cannot tell either remains constant or decreases. This is also what
whether this reflects the intrinsic catalytic activity or is seen in Table 4, but again we have the problem that
merely the change in catalyst concentration. (We are the solubility of the catalysts is in the same order as
Table 4. Oxidation of o-xylene with catalytic mixtures of manganese(II) acetate and hydrobromic acid.
[a]
[b]
Entry % Org.w/w [Br] mM [Mn] mM Br/Mn [mol/mol] OPA Yield [%]
Selectivity [%]
OPA 2-CBA PTL BA o-TALD
Mn [%]
1
2
3
4
5
6
7
8
0.44
0.44
0.44
0.44
0.44
0.44
0.44
0.26
0.0
4.2
10.3
10.3
10.3
10.3
10.3
10.3
7.3
0.0
0.4
1.2
1.6
2.0
2.4
2.8
2.8
8.0
21.3 5.7
54.5 4.5
64.6 2.7
73.6 1.6
80.4 0.2
87.8 <0.1
73.3 2.4
84.2 0.5
61.2 7.6 4.1
27.2 13.8 0.0
17.2 15.4 0.1
7.0 17.8 0.0
1.0 18.4 0.0
0.2 11.9 0.0
7.9 16.4 0.0
0.0 13.1 2.2
14
11
64
70
89
87
93
69
39.1
60.0
64.9
63.4
63.6
56.8
57.9
12.5
16.7
20.9
25.0
20.9
20.9
7.3
[
a]
Selectivity for the identified aromatic compound calculated as the molar concentration of that compound relative to the
total molar concentration of all identified aromatic compounds.
Recovered manganese as determined by atomic absorption.
[
b]
Adv. Synth. Catal. 2009, 351, 1866 – 1876
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1873

Joan Fraga-Dubreuil et al.
FULL PAPERS
the observed activity in acetic acid, making any defin- Table 5. Effect of switching the o-xylene feed on or off on
the manganese recovery. The initial concentration of manga-
itive conclusion regarding the intrinsic activity of the
nese in reactor is 10.3 mM.
catalyst impossible.
One can say that increasing the bromide concentra-
Entry Catalyst
o-Xylene stream
On/Off)
Mn
[%]
tion in the coordination sphere of the Mn does in-
crease the solubility of the catalyst. As increasing
amounts of hydrobromic acid are added to manga-
[a]
(
1
2
MnBr2
On
Off
On
75
18
90
ACHTUNGTRENNUNGn ese(II) acetate, the reactions shown in Figure 4 prob-
MnBr +HBr
2
ably occur. So one is dealing with at least three differ-
ent catalyst species - with each probably having a dif-
ferent intrinsic activity/selectivity. Apparently one un-
desirable characteristic of starting with a mangane-
se(II) acetate catalyst is that, at high conversion of
the o-xylene, the yield and selectivity to benzoic acid
is unusually high. This can be seen from entries 2–8 in
Table 4; an even clearer example is entry 1 in Table 3
where one initially starts with a manganese(II) bro-
mide catalyst and obtains an o-phthalic acid yield of
(
4.17 mM)
Off
On
52
84
3
MnBr +HCl
(
2
4.17 mM)
Off
Off
On
Off
47
20
75
25
4
5
MnBr2
MnBr
4.17 mM)
+HBr
2
(
On
92
[a]
62% with a benzoic acid selectivity of 6% while, in
Recovered manganese as determined by atomic absorp-
tion. After 20 min equilibration time, 5 samples were
taken, one every 10 min, and the results averaged.
entry 5 in Table 4 with a Mn(OAc) /2HBr mixture,
AHCTUNGTRENNUNG
2
one obtains an o-phthalic acid yield of 63% but the a
much higher benzoic acid selectivity at 18%. As pre-
viously discussed, benzoic acid can arise from three
sources, see Eqs. (18) and (19). Why the rates of one constantly being reduced to manganese(II) by their
or more of these reactions should be changed when reaction with bromide, see Eq. (9). With metal/bro-
one starts with manganese(II) acetate awaits further mide catalysts in acetic acid, the steady-state concen-
understanding of the chemistry and catalytic behav- tration of the metals in the higher oxidation states is
iour of the different catalytic species.
virtually zero; during Mn/Br-catalysed autoxidation of
A further run was conducted using sodium bromide methyl aromatics in acetic acid, the solution is colour-
as co-catalyst with manganese(II) acetate but the re- less, i.e., the steady state concentration of manganese
actor blocked very quickly due to the precipitation of is essentially 100% manganese(II). Without bromide,
the catalyst and also perhaps because of the insolubil- a manganese(II)-catalysed autoxidation reaction ap-
ity of sodium bromide in scH O, demonstrating that pears brown, i.e., contains a mixture of mangane-
2
both acidity and a high bromide content are required se(II), manganese
ACHTUNGTRNENUNG( III), and manganese(IV). The same
to avoid precipitation of the catalyst. is probably occurring in scH O, which accounts for
2
Table 5, entries 1–3 illustrate the effect on manga- the increased solubility of the manganese when o-
nese solubility when o-xylene, dioxygen and catalyst xylene is being oxidized.
are passed into the continuous reactor, the reaction is
allowed to come to steady state, but then the o-xylene
feed is turned off. An abrupt decrease in the manga- Conclusions
nese recovery, i.e., an increase in precipitation of the
manganese, is observed in the absence of o-xylene. This study has shown that the activity and the recov-
Comparison of entries 1 with 4 or 2 with 5 illustrates ery of a manganese(II) bromide catalyst can be im-
that the effect is reversible. Based on the discussion proved by addition of aromatic acids, by mineral acids
above, the results are unsurprising since aromatic (HCl, HBr), and by increasing the Br:Mn ratio during
acids are being generated in increasing concentration the Mn/Br-catalysed oxidation of o-xylene in scH O.
2
as one moves down the tubular reactor and we have This result can probably be extrapolated to different
already shown that the aromatic acids help to keep kinds of metal/bromide-catalysed oxidation of methyl
the manganese in solution, see Table 2. Another aromatics in scH O, particularly, the commercially
2
factor which helps to keep the manganese soluble is more interesting oxidation of p-xylene to terephthalic
that the manganese ACHUTNGRNENUG( III)–(IV) species in solution are acid.
Figure 4. Proposed reactions occurring when increasing the amount of HBr added to Mn/Br catalyst.
1874
asc.wiley-vch.de
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876

Prevention of Manganese Precipitation during the Continuous Selective Partial Oxidation
FULL PAPERS
In this study, hydrobromic acid proved to be the cally for acid concentrations greater than 1500 ppm in the
catalyst feed solution). A 100 W band heater was placed
most effective acidic co-catalyst because it is a source
around the catalyst feed pipe (1/8 inch), controlled by a
VARIACꢅ (100 V needed to reach 808C). Hot water
of both bromide and acidity. The bromide provides
activity and selectivity to the reaction; the acidity
(
808C) was circulated around the catalyst pump head to pre-
plays a more complex role because there is an optimal
pH for maximum catalyst activity and selectivity, and
because a low pH increases the solubility of the cata-
lyst.
Aromatic acids also increase the solubility of the
manganese catalyst not only because they are acids
vent any crystallisation of the acids used.
Before each run, the apparatus was hydrostatically pres-
sure-tested when cold, and was then heated with a flow of
À1
pure water (5–10 mLmin ). Once the operating tempera-
ture had been reached, H O /H O was fed into the reactor
2
2
2
and the pumps for o-xylene and acid/catalyst were started,
but also because they probably enter the coordination typically in that order. The residence time for each run re-
sphere of the metal. Benzoic acid is particularly fav- mained constant, typically up to 1 min, but in most cases 5–
oured because it is produced in the reaction as a by- 10 seconds.
product and eventually must be separated from the
terephthalic acid.
Precipitated species (mainly in the form of manganese
oxides) were continuously filtered during the reaction via an
[11,12]
on-line filter placed after the cooler.
The resulting aque-
This discovery offers the opportunity of using ben-
zoic acid to limit the use of high bromide concentra-
tions that could accelerate the rate of corrosion of the
ous solution of products exiting the reactor via the back
pressure regulator was then sampled for periods of 2 min
every 10 min and analysed by HPLC and atomic absorption.
[
38]
reactor tubing in scH O. Three examples are given
2
HPLC analysis was carried out on a Hewlett Packard
050 Series HPLC apparatus from Agilent Technologies UK
Ltd. equipped with an XTerraꢅ RP18 3.5 mm, 30ꢄ150 mm
column and an oven at a temperature of 408C; detection
was at 254 and 230 nm. Every compound (acid and inter-
mediates) was calibrated by HPLC prior to analysis of the
where the activity of different catalysts in scH O are
2
1
in the same order as those in acetic acid. One impor-
tant difference however, is that in acetic acid the cata-
lysts are 100% soluble, while in scH O the trend of in-
2
creased activity has the same trend as increased cata-
lyst solubility. So the trend of increased activity in su- samples. The actual yield was calculated by HPLC as a per-
percritical water coincides with an increased concen- centage of the expected yield based on the measured
tration of homogeneous catalyst as well.
amount of organic and water pumped into the apparatus.
Selectivity is defined as the molar amounts of each aromatic
compound as measured by HPLC given on a percentage
basis.
Experimental Section
For the atomic absorption analysis, the aqueous samples
were filtered using disposable membrane syringe filters and
appropriately diluted with water to suit the operating condi-
tions of the atomic absorption apparatus, which was calibrat-
ed with freshly prepared Mn standards (5, 10, 15 ppm Mn in
water) from commercially available Mn aqueous solutions.
Mn concentrations were then measured based on this cali-
bration. The yields of Mn were calculated from the flow-out
of Mn divided by the flow-in of Mn where the concentration
of Mn was measured by atomic absorption. The average
yields and selectivities of the samples collected during each
run are quoted in the tables.
Caution: High pressure oxidation experiments are potential-
ly hazardous. It is the responsibility of individual researchers
to ensure that their equipment is suitable and safe to oper-
ate under these conditions.
All chemicals (Sigma–Aldrich UK Ltd.) were used with-
out further purification. All the experiments were conducted
using a tubular continuous flow reactor, similar to the one
[11,12]
that has been previously described.
Experimental work
was carried out on the continuous oxidation of o-xylene by
O in scH O at about 3808C and 230 to 240 bar with a com-
2
2
bined catalyst and acid solution. In a typical experiment,
6 mL of 100 volume H O solution (>30% w/v) commer-
5
2
2
cially available from Fisher Scientific UK Ltd. was mixed
with 760 mL deionised H O. The H O solution was pumped
into the system and decomposed in a pre-heater under the
required experimental conditions to generate a homogene-
2
2
2
Acknowledgements
We thank INVISTA and the EU Marie Curie Programme for
funding. We are grateful to Messrs P. Fields, R. Wilson and
M. Guyler for their technical support.
ous mixture of O and scH O. This corresponds to an O
2
2
2
concentration of 0.276 mol/L (1.5 molar equivalents of the
stoichiometry required for oxidation of the o-xylene to o-
phthalic acid, i.e., 3 O /o-xylene). The full decomposition of
2
H O to O was demonstrated by measuring the volume of
2
2
2
gas and liquid exiting the reactor. Individual flow rates References
À1
were: H O (or H O ), 8 mLmin ; Mn/Br solution (contain-
2
2
2
À1
ing either mineral or organic acids), 4 mLmin ; pure organ-
ic, 0.06 mLmin which gave 0.5% v/v organic/H O at the
[1] H. Weingꢆrtner, E. U. Franck, Angew. Chem. 2005, 117,
À1
2730–2752; Angew. Chem. Int. Ed. 2005, 44, 2672–
2692.
[2] A. Kruse, E. Dinjus, J. Supercrit. Fluids 2007, 39, 362–
2
mixing point. The total volumetric flow rate (total FR) after
mixing was about 12 mLmin . A hot-catalyst feed system
À1
was used when aromatic carboxylic acids were used (typi-
380.
Adv. Synth. Catal. 2009, 351, 1866 – 1876
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1875

Joan Fraga-Dubreuil et al.
FULL PAPERS
[
3] J. Fraga-Dubreuil, M. Poliakoff, Pure Appl. Chem.
[19] S. H. Jhung, Y-S. Park, Bull. Korean Chem. Soc. 2002,
23, 369.
2
006, 78, 1971–1982.
[
[
4] P. E. Savage, Chem. Rev. 1999, 99, 603–621.
5] D. Brçll, C. Kaul, A. Krꢆmer, P. Krammer, T. Richter,
[20] G. W. Parshall, S. D. Ittel, Homogeneous Catalysis, 2nd
edn., John Willey and Sons, Inc., New York, 1992,
Chapter 10.
M. Jung, H. Vogel, P. Zehner, Angew. Chem. 1999, 111,
3
3
180–3196; Angew. Chem. Int. Ed. 1999, 38, 2998–
014.
[21] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations
of Organic Compounds, Academic Press, New York,
N. Y. 1981, chapter 2.
[
[
[
[
6] M. D. Bermejo, M. J. Cocero, AIChE J. 2006, 52, 3933–
951.
7] P. E. Savage, J. B. Dunn, J. Yu, Combust. Sci. Tech.
006, 178, 443–465.
8] J. B. Dunn, P. E. Savage, Environ. Sci. Technol. 2006,
9, 5427–5436.
[
22] W. Partenheimer, The influence of pH in metal/bromide
catalyzed homogeneous aerobic autooxidation. to be
submitted for publication.
3
2
[
[
[
[
23] W. Partenheimer, J. Mol. Cat. A: Chem. 2003, 206,
1
31–144.
3
24] J. L. Broeker, W. Partenheimer, B. I. Rosen, U. S.
Patent 5,453,538, 1995.
25] Y. Chen, J. L. Fulton, W. Partenheimer, J. Am. Chem.
9] J. Fraga-Dubreuil, E. Garcia-Verdugo, P. A. Hamley,
E. M. Vaquero, L. M. Dudd, I. Pearson, D. Housley, W.
Partenheimer, W. B. Thomas, K. Whiston, M. Poliakoff,
Green Chem. 2007, 9, 1238–1245.
Soc. 2005, 127, 14085–14093.
26] E. Lester, P. Blood, J. Denyer, D. Giddings, B. Azzopar-
di, M. Poliakoff, J. Supercrit. Fluids 2006, 37, 209–214.
27] J. Otsu, Y. Oshima, J. Supercrit. Fluids 2005, 33, 61–67.
28] W. Partenheimer, Adv. Synth. Catal. 2005, 347, 580–
[
[
[
10] E. Garcia-Verdugo, J. Fraga-Dubreuil, P. A. Hamley,
W. B. Thomas, K. Whiston, Green Chem. 2005, 7, 294–
3
[
[
00.
11] E. Garcia-Verdugo, E. Vernardou, W. B. Thomas, K.
Whiston, W. Partenheimer, P. A. Hamley, M. Poliakoff,
Adv. Synth. Catal. 2004, 346, 307–316.
12] P. A. Hamley, T. Ilkenhans, J. M. Webster, E. Garcia-
Verdugo, E. Vernardou, M. J. Clarke, R. Auerbach,
W. B. Thomas, K. Whiston, M. Poliakoff, Green Chem.
5
90.
[
29] C. F. Baes, R. E. Mesmer, The Hydrolysis of Cations,
Krieger Publishing Company, Malabar, FL, USA, 1986,
p 219.
[
30] N. N. Greenwood, A. Earnshaw, Chemistry of the ele-
ments, Pergamon Press, Oxford, 1984, Chapter 24,
p 1211.
2
002, 4, 235–238.
13] J. B. Dunn, P. E. Savage, Ind. Eng. Chem. Res. 2002, 41,
460–4465.
[
[
[
[
[
[
[
[
[
[
[
31] J. Fraga-Dubreuil, S. D. Housley, W. Partenheimer, M.
Poliakoff, World Patent WO2008070365, 2008.
32] J. Fraga-Dubreuil, S. D. Housley, W. Partenheimer, M.
Poliakoff, World Patent WO2008057924, 2008.
33] J. B. Dunn, M. L. Burns, S. E. Hunter, P. E. Savage, J.
Supercrit. Fluids 2003, 27, 263–274.
4
14] J. B. Dunn, D. I. Urquhart, P. E. Savage, Adv. Synth.
Catal. 2002, 344, 385–392.
15] R. L. Holliday, B. Y M. Jong, J. W. Kolis, J. Supercrit.
Fluids 1998, 12, 255–260.
16] J. B. Dunn, P. E. Savage, Green Chem. 2003, 5, 649–
34] A. Apelblat, E. Manzurola, N. Abo Balal, J. Chem.
6
55.
Thermodyn. 2006, 38, 565–571.
35] H. Bartnicka, I. Bojanowska, M. K. Kalinowski, Aust.
J. Chem. 1991, 44, 1077–1084.
17] a) W. Partenheimer, Cat. Today 1995, 23, 69–158;
b) W. Partenheimer, M. Poliakoff, The Aerobic Oxida-
tion of p-Xylene to Terephthalic Acid: a Classic Case of
Green Chemistry in Action, in: Handbook of Green
Chemistry – Green Catalysis: Homogeneous Catalysis,
36] D. Bonnet, T. Ireland, E. Fache, J. Simonato, Green
Chem. 2006, 8, 556–559.
[37] W. Partenheimer, in: Catalysis of Organic Reactions,
(Ed.: F. E. Herkes), Marcel Dekker, Inc., New York,
1998, pp 357–368.
[38] P. Kritzer, J. Supercrit. Fluids 2004, 29, 1–29.
[39] K. P. Johnston, J. B. Chlistunoff, J. Supercrit. Fluids
1998, 12, 155–164.
(Eds.: P. T. Anastas, R. H. Crabtree), Wiley-VCH,
Weinhein, 2009, Volume 1, Chapter 12, pp 375–397.
[
18] W. Partenheimer, in: Catalysis of Organic Reactions,
(Ed.: D. W. Blackburn), Marcel Dekker, Inc., New
York, 1990, chapter 20.
1876
asc.wiley-vch.de
ꢃ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 1866 – 1876