Selective aerobic oxidation of para-xylene in sub- and supercritical water. Part 2. the discovery of better catalysts

Article abstract of DOI:10.1039/c1gc15138j

An extensive and systematic study has been carried out on the catalytic effect of more than 20 elements on the aerobic oxidation of p-xylene to terephthalic acid in super- and subcritical water. Reactions have been performed in a continuous reactor under

Full text of DOI:10.1039/c1gc15138j

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2011, 13, 2397
www.rsc.org/greenchem
PAPER
Selective aerobic oxidation of para-xylene in sub- and supercritical water.
Part 2. The discovery of better catalysts†
Eduardo Pe´rez,*a Joan Fraga-Dubreuil,a,b Eduardo Garc´ıa-Verdugo,a Paul A. Hamley,a
Morgan L. Thomas,a Chong Yan,a W. Barry Thomas,b Duncan Housley,b Walt Partenheimerc and
Martyn Poliakoff*a
Received 3rd February 2011, Accepted 18th May 2011
DOI: 10.1039/c1gc15138j
An extensive and systematic study has been carried out on the catalytic effect of more than 20
elements on the aerobic oxidation of p-xylene to terephthalic acid in super- and subcritical water.
Reactions have been performed in a continuous reactor under catalyst unsaturated conditions.
Reaction product, by-products and intermediates have been quantified as well as the burn (the
amount of CO2 originating from total oxidation of p-xylene). CuBr2 has been found to be a
superior catalyst to MnBr2, which has been widely used in the literature for this reaction in water
at high temperatures. At catalyst unsaturated conditions (i.e. with low concentrations of catalyst),
MnBr2 gives a terephthalic acid yield of 36.1% whereas CuBr2 enhances this value to 55.6%. A
strong synergistic effect has been found between CuBr2 and other metals and sources of bromide.
Indeed, we show that Cu/Co/Br, Cu/Co/NH4/Br and other mixtures give better results than
CuBr2 reaching a terephthalic acid yield of 70.5% for the four component catalyst. The
compositions of the catalyst as well as the reactor temperature have been optimized and their
effects on the analyzed compounds are discussed. A substantial amount of additional data is
included in the electronic supplementary information.
be used as a reaction medium for a range of other chemical
Introduction
transformations.17–21
Terephthalic acid (TA) is an important commodity chemical. It
This paper is the second of two describing our research on
is an intermediate of polyethylene terephthalate, which is used
the aerobic, homogeneously catalyzed autoxidation of p-xylene
for making plastics, textiles and other polymeric products. Most
in sub- and super-critical water. The first paper22 contrasted
of the TA produced in the world is made by catalytic aerobic
oxidation of p-xylene (pX) using acetic acid as the solvent. Over
the different behaviour of para- and ortho-xylene under uncat-
alyzed, bromide-catalyzed, and Co/Br and Mn/Br-catalyzed
the past few years, attempts have been made by our group1–5 and
conditions and we described the impact that water has on the
others6–14 to substitute acetic acid by supercritical water (SCW)
catalyst activity and yields of the different products and by-
and high temperature water (HTW) because of the potential
products of this reaction (especially CO2 and benzoic acid). In
economic and environmental advantages that water offers.15 In
this paper we describe improved catalysts for HTW and SCW
addition, the unusual properties of water (Tc = 374 ◦C, pc =
over those that have already been reported.
Commercial manufacture of TA started in 1951 using Co
221 bar) at near- and supercritical conditions16 allows it to
acetate and Co/Mn mixtures in acetic acid. Since then, a huge
effort has been made to improve the efficiency of this process.23
The order of catalyst activity for the formation of p-toluic acid
aUniversity of Nottingham, Nottingham, UK, NG7 2RD. E-mail:
(PTA) is Co(II) ꢀ Mn(II) > Ni(II) with the remaining first
martyn.poliakoff@nottingham.ac.uk, eduardo.velilla@nottingham.ac.uk;
row transition elements having no activity.24 In 1954 a major
Fax: +44 115 951 3058; Tel: +44 115 951 3520
breakthrough occurred with the discovery of metal/bromide
bINVISTA Textiles UK Limited, Wilton, Cleveland, UK, TS10 4RF
cDuPont (retired), Consultant, Current address 7253 SW 26th Ave.,
Portland, OR, 97219, USA. E-mail: wpartenheimer@msn.com
catalysts especially the combination of Co and Mn acetates
with bromide.25 The metal/bromide catalysts greatly improve
the selectivity and yield for converting pX to TA. The order of
catalyst activity for the formation of p-toluic acid of the metallic
bromides is Co(II) > Mn ꢀ Ni.24 The combination Co/Mn/Br
† Electronic supplementary information (ESI) available: Full experimen-
tal details, including the configuration of the reactor, Fig. S1; additional
experimental data, Tables S1–S18; and an analysis of the colour of some
samples, Fig. S2 and Table S19–S20. See DOI: 10.1039/c1gc15138j
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2397
©

View Online
+
Fig. 1 Summary of all of the elements plus NH4 used in this study. Most metals were used as the halide salts. Non-redox active metal ions, e.g.
Na+ or Ca2+, were used to vary the concentration of Br- relative to the redox active metals. W, Ti, V were evaluated but, due to the unavailability or
instability of their bromide salts in water, NaWO4, titanium(IV) bis(ammonium lactate) dihydroxide and NH4VO3 were used instead. Combinations
of these W, Ti and V salts with different sources of bromide are reported in the ESI;† however the results were poor. Some metals such Bi and Sb
could not been evaluated because their bromide salts are not soluble in water. SnBr2 is soluble in water only at low concentrations. BiBr3 and SbBr3
hydrolysed in water to BiOBr or SbOBr and HBr.32
performs much better than its parts separately. This synergistic
effect gives a scientific explanation of the superior activity of this
catalyst for the TA manufacture. All these discoveries have led
to the current technology for converting pX to TA.
see Fig. 1. Particularly significant is the fact that copper in
metal/bromide catalyst is generally strongly antagonistic, i.e.
it strongly inhibits the rate of reaction in acetic acid.23 Here we
report that Cu is one of the best catalysts in HTW.
Although the Co/Mn/Br combination in acetic acid is the
most extensively investigated catalyst, many other combina-
tions of metal with bromide have been reported.23 Ni can be
substituted for Co; Ni/Mn/Br is as active as Co/Mn/Br26
and Ce can be substituted for Mn although it is less active.27
Combinations of more than two metals have also been attempted
incorporating Zr(IV), Hf(IV), Zn(II) and other metals in the
catalyst configuration.23 Mixed halogen catalysts have also been
reported28–30 as well as metal/NHPI combinations31 (NHPI =
N-hydroxyphthalamide).
This paper summarizes the key points from several person
years of research undertaken at the University of Nottingham.
Additional data are given in the ESI.†
Results
Oxidation of TA in acetic acid follows a sequence of
partially oxidized intermediates: para-xylene (pX)
→
4-
methylbenzaldehyde (p-tolualdehyde, pTOL) → p-toluic acid
(PTA) → 4-carboxybenzaldehyde (4CBA) → TA; see Scheme 1.
Reaction 1a and 3a to the benzylic alcohols are minor reactions.
Small amounts of benzoic acid (BA) are also detected, which
probably comes from decarboxylation of TA or one of its
precursors.22
In 1998, Holliday et al.6 demonstrated that selective batch
oxidation of methylaromatics is possible on a small scale in
HTW. Since then, there has been a number of papers by different
research groups reporting oxidation of pX using water as a
solvent.1–14 However, very little research has been carried out
to optimize the nature of the catalysts. Dunn et al.9,10 established
that MnBr2 was more effective in HTW than Co/Mn/Br, the
most commonly used catalyst in acetic acid. Interestingly, it was
also observed to be more active than Mn/Zr/Br, Co/Mn/Hf/Br
or Mn/Ni/Zr/Br. Most of the subsequent research conducted
on this reaction in water has used MnBr2 as catalyst.
In this paper, we present a systematic study of the effect
of different inorganic salts used as catalysts in the continuous
oxidation of pX to TA in SCW and HTW. The primary objective
has been to find the most active catalyst for this reaction and
we have done this by investigating a wide range of elements,
Scheme 1 Sequence of intermediates of the oxidation of pX to TA in
acetic acid. Reaction 5 is only one possible route to BA.15
2398 | Green Chem., 2011, 13, 2397–2407
This journal is
The Royal Society of Chemistry 2011
©

View Online
All the experiments were conducted using tubular continuous
flow reactors, the details of which are reported elsewhere.3 A
simplified diagram is shown in Fig. 2. An aqueous solution of
hydrogen peroxide was passed through a pre-heater where it
decomposed to a mixture O2/SCW before entering the reactor.
The pX and a cold aqueous solution of catalyst are continuously
pumped into the reactor where they contact the O2/SCW. At the
reactor outlet, the mixture was quenched with aqueous NaOH
to prevent precipitation of TA. The products were analyzed by
HPLC and the CO2 yield was measured by acid titration of
the carbonate in solution. Unless otherwise stated, the reactions
were carried out at a pressure of 230 bar and at temperature of
380 ◦C. The inlet pipes of the pX and catalyst feeds protruded
into the reactor unless specified otherwise. More details of the
equipment and procedures can be found in the ESI.†
quantified by HPLC. The selectivity for each compound was
calculated as the concentration of that compound divided by
the sum of those of the six aromatic compounds. The yield of
TA, YTA , was calculated by eqn (1)
[TA]
Y TA =100
(1)
[pX]
where [pX] is the molar concentration of pX in the sample if
the conversion had been 0. i.e. allowing for dilution. [pX] =
(FpXrpX)/(FtotalMpX), where FpX, Ftotal are the flowrates of pX
and the total flow after quench; rpX and MpX are the density
and molar mass of pX respectively. The CO2 yield originating
from total combustion was calculated from the carbonate
concentration obtained from titration by the eqn (2), with [BA]
being subtracted to avoid counting the CO2 molecules arising
from decarboxylation. There are eight carbon atoms in p-xylene
hence the factor of 8 in eqn (2).
[CO32−]−[BA]
(2)
Y CO =100
2
8[pX]
Finally, we define “burn” as shown in eqn (3):
(3)
Burn = YCO /YTA
2
i.e. the amount of CO2 relative to the yield of the main product,
TA. The burn is a measure of the amount of pX that is lost
through combustion.
Fig. 2 Schematic of the continuous reactor used in this work; PH is the
preheater.
1. Search for the best redox metal(s) as catalyst
Reproducibility of yields in runs performed within a few days
of each other was better than 1%. However, the work presented
in this paper is the result of a project that lasted for several
years and with many experimenters. Over this time, the CO2
yield was found to vary by several percentage units. This does
not change the conclusions of this work because the relative
trends were maintained in any set of experiments done within a
given period. The reproducibility of the TA yield and selectivity
remained good over this time. Generally the error is 1% for TA
selectivity and 2% for TA yield. BA yield reproducibility was
lower: 5%.
When evaluating the relative activity of various catalysts
in homogeneous oxidation, it is important that the solution
should be not “catalyst saturated”; i.e. the catalyst concentration
should be lower than that needed to give the maximum
yield. Seeking improved catalysts when the system is catalyst
saturated often gives negative results because the high catalyst
concentration “masks” the intrinsic catalytic activity.39 In this
study we have deliberately tested the relative activity of different
catalysts under ‘catalyst unsaturated’ conditions with lower
than optimum yields. Thus, the typical catalyst concentrations
in the reactor were 2.6 mM, ca. 25% of those we have used
previously.1–5 In our discussion, each catalyst is designated by
the sum of its cations and anions and the concentration of
each species is expressed as multiples of this concentration. For
example, if equal amounts of CoBr2 and MnBr2 were added,
the catalyst formulation would be Co/Mn/Br with the relative
amounts: 0.5/0.5/2.0 and concentrations of the species would
be 1.3/1.3/5.2 mM. Apart from pX and p-tolualcohol, the
concentrations of the six other species shown in Scheme 1 were
Scope of this paper. Redox catalysts usually contain metals
with variable valence, i.e. the transition metals. With homo-
geneous oxidation catalysts the two valencies normally differ
by one unit, e.g. Co(II) and Co(III). A wide range of different
metallic bromide solutions have been used as catalysts in our
experiments. Fig. 3 illustrates the catalytic activity of different
metallic bromides for this reaction in supercritical water.
Pure metal bromides. Table 1 summarizes the data quanti-
tatively for the pure metal bromides. The following points are
clear from Fig. 3 and Table 1. (a) The CuBr2 catalyst gives the
highest TA yield and selectivity of the simple metal bromides
investigated. It also gives the lowest 4CBA and PTA selectivities
and no detectable amounts of HMBA and pTOL. (b) Under
unsaturated conditions, the order of activity of metals according
to the highest TA yield and selectivity is Cu(II) > Fe(III) > Mn(II)
> Co(II) >other metals. These results are consistent with the
observation by Savage et al.9 and our group22 that MnBr2 is
better catalyst in SCW than CoBr2. (c) MnBr2 is not the optimal
catalyst since FeBr3 and CuBr2 perform better. (d) Of the metal
bromides, Co/Br is the best catalyst in acetic acid but, in SCW,
TA yield and selectivities are significantly lower than several
other metals. (e) The CO2 yield for CuBr2, 24.4%, is higher
than for MnBr2, which only gives 18.1%. Even so, the burn is
lower with Cu because the CO2 yield is related to the amount
of TA produced. (f) Taking this into account, CuBr2 is also the
best catalyst from the point of view of burn. (g) ZrBr4 gives a
lower TA yield than MnBr2 but the selectivity is better. ZrBr4
is believed to hydrolyze to ZrO2 and HBr and to precipitate in
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2399
©

View Online
Fig. 3 TA and CO2 yields for selected catalysts evaluated at the same unsaturated concentration. The arrows highlight the most important catalysts.
It can be seen that the metal has a more dramatic effect on the TA than on the CO2 yields. All reactions carried out at 230 bar and 380 ◦C.
Table 1 Catalytic activity of different metallic bromides in the oxidation of pX to TA in supercritical water
Yield,
mol (%)
Selectivity
Br : metals
mol/mol
Entry
Catalyst
[catalyst]/2.6 mM
CO2
TA
Burn
TA
PTA
4-CBA
HMBA
pTOL
BA
1
2
3
none
—
—
14.3
10.7
13.4
22.8
26.4
18.5
27.1
25.1
14.8
18.1
19.8
24.4
0.6
1.9
2.5
4.8
17.1
5.0
6.0
6.5
6.3
23.8
5.63
5.36
4.75
1.54
3.70
4.52
3.86
2.35
0.50
0.50
0.44
8.4
73.3
47.0
61.7
48.9
30.8
52.4
59.9
57.2
56.4
34.9
9.9
4.7
1.8
2.3
2.5
2.7
1.5
2.2
2.1
3.2
2.6
1.4
0.0
0.0
9.9
6.6
9.6
17.0
0.6
3.7
3.1
4.2
3.0
0.0
4.8
0.0
1.9
4.7
2.7
3.1
8.0
3.6
3.4
3.2
2.8
6.0
6.8
6.5
Zn/Br
La/Br
Eu/Br
Zr/Br
Ni/Br
Ce/Br
Gd/Br
Co/Br
Mn/Br
Fe/Br
Cu/Br
1.0/2.0
0.66/2
1.0/3.0
1.0/4.0
1.0/2.0
1.0/3.0
1.0/3.0
1.0/2.0
1.0/2.0
1.0/3.0
1.0/2.0
2.0
3.0
3.0
4.0
2.0
3.0
3.0
2.0
2.0
3.0
2.0
14.5
12.3
13.4
56.2
18.9
22.7
19.2
19.4
51.3
75.3
86.4
25.0
11.2
14.9
2.8
19.2
8.9
13.0
15.8
6.3
4
5a
6
7
8
9
10
11b
12
36.1
39.3
55.6
3.2
1.2
5.9
a ZrBr4 is likely to hydrolyze to ZrO2 + 4HBr. Catalytic activity is probably due only to the 10.4mM HBr released (see Table 1 in the first paper of
this series22). b FeBr3 caused blockages in the pipes so it was tried using a different reactor configuration: pX and catalyst streams were premixed in a
T piece before entering the reactor.
the reactor, as happens in acetic acid.33 The rest of the bromides
evaluated gave poor TA yields and selectivities.
bromide of the redox metal. In the oxidation of o-xylene, Fraga-
Dubreuil et al.5 have already demonstrated that addition of HBr
to MnBr2 enhances the solubility of Mn and hence increases
its activity. However, adding bromide as a mixture of HBr and
NaBr was much less effective (Table S2 in the ESI†). Attempts
to reduce the amount of bromide by substituting chloride or
iodide also gave lower activity (Table S4).
However the real surprise came when HBr was used to
keep the absolute concentration of bromide constant while the
concentration of Cu(II) was decreased. Entry 1 of Table 2 shows
Search for the best source of bromide. In acetic acid, the
presence of bromide greatly boosts the activity and selectivity to
TA in the oxidation of pX.24 The source of bromide in acetic acid
results in only small differences in TA yield or selectivity. In SCW,
the high sensitivity of the reaction to the bromide concentration
has been reported in several cases.1,5,9,22 However, until now,
the source of bromide used in all SCW reactions has been the
2400 | Green Chem., 2011, 13, 2397–2407
This journal is
The Royal Society of Chemistry 2011
©

View Online
Table 2 Catalytic activity using different combinations of CuBr2 + MBrn as catalysta
Yield,
mol (%)
Selectivityc
Br : metals
Entry
Catalystb
[catalyst]/2.6 mM
mol/mol
CO2
TA
Burn
TA
PTA
4-CBA
HMBA
BA
1
2
3
4
5
6
7
8
9
Cu/H/Br
0.15/1.7/2
0.15/1.7/2
0.15/0.85/2
0.15/0.85/2
0.15/0.57/2.1
0.15/0.85/2
0.15/0.85/2
0.1/0.9/2.9
0.1/0.9/2.9
13.3
13.3
2.0
2.0
2.9
2.0
2.0
2.9
2.9
19.9
18.7
23.5
15.3
19.9
19.6
17.5
23.4
21.6
53.9
58.0
55.2
61.7
41.2
59.0
46.8
57.5
56.2
0.37
0.32
0.43
0.25
0.48
0.33
0.37
0.41
0.38
90.8
92.1
91.3
92.1
72.3
89.7
70.3
90.8
90.4
1.4
0.0
0.6
0.0
15.4
0.7
13.8
0.0
0.0
0.2
0.0
0.0
0.0
3.3
0.0
2.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
1.1
0.0
0.0
7.6
7.9
8.1
7.9
9.0
9.5
12.3
9.2
9.6
Cu/NH4/Br
Cu/Mn/Br
Cu/Co/Br
Cu/Fe/Br
Cu/Ni/Br
Cu/Zn/Br
Cu/La/Br
Cu/Eu/Br
a Data for additional Cu/M/Br, M = Na, K, Ca, are given in Table S3 in the ESI. b Concentration of metals is 2.6 mM in every case except in entries
1 and 2 that is 0.39 mM and in entry 5 that is 1.9 mM. c Selectivity for pTOL was 0 in all cases.
that more than a six-fold reduction in concentration of Cu gave
a better catalytic performance than CuBr2 itself. Furthermore
replacing HBr by NH4Br gave even better results, see Fig. 3 and
Table 2, entry 2. BA is the only aromatic by-product present in
significant quantities and no 4CBA was detected. The burn is
also low, giving so far the smallest amount of CO2 per unit of
TA produced. NH4Br is therefore an interesting candidate as
source of bromide. Additional results are summarized in Table
S3 in the ESI†.
Unlike in acetic acid, the effect on TA yields of adding non-
redox metal bromides to CuBr2 was very large in HTW ranging
from 21 to 58% following the order NH4Br > HBr > ZnBr2 >
CaBr2 > NaBr > KBr; see Tables 2 and S3.
Synergy occurs when the resultant effect of combining the
components is greater than the sum of its parts. Examination
of Tables 1 and 2 suggests synergism between Cu and other
components. A more detailed study reveals synergistic effect
between Cu and Co, Mn, Ni, Zn, HBr and NH4Br. Details can
be found in the ESI (Table S10).†
The experiments described above have demonstrated the high
activity of Cu for the oxidation of pX to TA when this metal is
in solution, i.e. it acts as a homogenous catalyst. An additional
study using solid metallic Cu plus HBr as the catalyst was
also carried out. The results showed that the reaction could
be catalysed by Cu(II) formed in situ by dissolution of copper
metal by HBr. Details of these experiments are also given in the
ESI, Table S18.†
Synergistic effects between Cu and other metals. From the
results in the previous sections, Cu appears to be an especially
active catalyst, possibly playing the same role as Co in the
acetic acid-based reaction. The next step was to search for
synergistic effects between Cu and other metals in SCW. Table
2 shows the effects of combining CuBr2 with other metallic
bromides. Catalyst configurations are Cu/M/Br where the
bromide concentration was kept constant at 2 ¥ 2.6 mM and
Cu concentration is 0.15 ¥ 2.6 mM, except in entries 5, 8 and
9. When M was either Co or Ni, a better TA selectivity and
yield were obtained than for CuBr2 alone. La and Eu also gave
better TA yields and selectivities but the Br concentration is also
higher because of the stoichiometry of MBr3. No 4CBA was
detected for the best combinations (M = Mn, Co, Ni, Eu or La).
Cu/Co/Br gave the highest TA yields obtained so far under
catalyst unsaturated conditions, as well as high selectivity and
low burn. We also investigated whether adding a third metal
to Cu/Co/Br would give additional benefits. Table S6 in the
ESI† shows the results for a range of Cu/M/M¢/Br, where M,
M¢ = Co, Ni, Mn, Fe, La or Zn. None of them gave as good a
performance as Cu/Co/Br in terms of TA yield, TA selectivity
and low burn together. Some Ni-containing catalysts gave a
higher TA yield but also higher burn. Cu/Co/Zn/Br gave low
burn and might be worth investigating further. Some Fe-based
compositions (Table S7) and a number of binary and ternary
mixtures which did not contain Cu were also studied, see Table
S8. These included Mn, Fe, Ni, Zn and Ce. None came close to
the performance of the Cu-based catalysts.
Co-oxidation of p-xylene with selected substrates. One of
the former commercial scale methods for TA manufacture
catalyzed by cobalt(II) acetate, was conducted by co-oxidizing
pX with acetaldehyde, which is oxidized readily to peroxyacetic
acid.34 This acid is known to oxidize Co(II) rapidly to Co(III).15
Without the co-oxidant, the pX is oxidized chiefly to p-toluic
acid. The formation of the additional Co(III) in the presence
of acetaldehyde drives the reaction to high TA yields. Since
co-oxidation has been used commercially in acetic acid,34 we
evaluated this approach to see whether it would be beneficial
for the oxidation of pX in SCW. Benzaldehyde can also serve
as an effective co-oxidant since it will form peroxybenzoic
acid quickly and selectively.35 However we found that the co-
oxidation of pX with benzaldehyde did not improve the TA yield
with either Cu/Br or Cu/Co/Br catalysts. GC analysis after the
experiments confirmed that most of the benzaldehyde had been
consumed. Similarly the use of other co-oxidants such as phenol,
isopropanol, methanol or toluene was also unsuccessful. Details
of these experiments are given in the ESI, Table S11.†
2. Catalyst optimization
In the first part of this paper, we described the discovery of
Cu/Co/Br as a new homogeneous catalyst for the oxidation of
pX to TA in SCW. After the discovery of such a catalyst, it can
be optimized further in the following ways:
1. Catalyst Concentration. The activity of the oxidation
system normally increases with increased catalyst concentration
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2401
©

View Online
until the system is ‘catalyst saturated’ whereupon the activity
either remains constant or decreases.39 Often the yield with
metal/bromide catalysts also increases with catalyst concen-
tration because the system becomes more selective towards
the desired aromatic acid, i.e. mechanisms leading to the total
destruction of the aromatic ring to CO2 are decreased.23
2. M : Br ratio. In acetic acid, the activity of the system
usually increases as this ratio increases, reaches a maximum and
then decreases. There are reports of the selectivity of the system
changing as well.23
3. M¢ : M¢¢ ratio. Variation of this ratio in Co/Mn/Br
and Co/Mn/Zr/Br catalysts has a strong influence on activity
during pX oxidation in acetic acid.24 It can also affect amount
of ring degradation using a Co/Mn/Br catalyst in acetic acid.36
Changing the Cu : Co ratio in the Cu/Co/Br catalyst in acetic
acid changed the relative rates of pX and PTA acid oxidation37
4. Temperature. Temperature will affect to the reaction rate
but it can also affect the extent of decarboxylation38 and ring
degradation.22
Here we describe the first stages in optimizing the new
Cu/Co/Br catalyst using these approaches. We also discuss the
influence of temperature on the reaction.
Effect of catalyst concentration. Without any catalyst
present and assuming no leaching of the metals from the reactor
walls, the TA yield is only 0.6%. Cu is an especially active
catalyst, by adding just 0.39 mM of CuBr2, the TA yield jumps
from 0.6 to 12.2% (by a factor of 20), see Table S3. Increasing
the concentration to 2.57 mM, the yield increases to 55.6%, see
Table 1. When only 0.18 mM of CuBr2 is combined with 1.09 mM
of CoBr2 one obtains 60% TA yield and 88.5% selectivity.
However, four times more catalyst only increases the TA yield to
65.1%, indicating that the catalyst is close to saturation, see
Table S12. More active catalysts such as the Cu-based ones
become saturated at lower concentrations. This fact has a crucial
importance on the optimum design of a viable process using such
catalysts.
Fig. 4 Effect of varying the Cu2+ content of the Cu/Co/Br catalyst on
the % yields of TA, BA, PTA and CO2 and the % TA selectivity. During
the experiment the Br- was held constant at 5.2 mM. x(CuBr2) is the
molar fraction of CuBr2: x(CuBr2) = nCuBr /(nCuBr + nCoBr ).
2
2
2
Effect of temperature. The effect of temperature on the
performance of the Cu/Co/Br catalyst is shown in Fig. 5.
Concentration is (0.1/0.9/2.0) ¥ 2.6 mM which was shown to be
the optimum ratio in the previous section. There are maxima in
TA yield and selectivity at 330 ◦C. However, yields of CO2 and
BA increase monotonically with temperature. There is, therefore,
a compromise between highest TA yield and lowest CO2 and
BA yields, with the optimum point at subcritical temperatures
(330 ◦C).
Several further examples comparing catalyst performance at
330 and 380 ◦C can be found in the ESI.† In general, yields
for Cu-containing catalysts are always higher at 330 ◦C with
correspondingly lower yields of CO2 and BA. However, if NH4Br
is added to increase Br : metals ratio, the opposite effect is
observed; see Table S14.
Effect of Cu : Co ratio. Fig. 4 shows the effect on several
reaction products of varying the Cu : Co ratio in the Cu/Co/Br
catalyst. The combined concentration of the metals and of the
bromide were kept at 2.6 and 5.2 mM respectively. The figure
again reveals the high activity of Cu. At a concentration of
2.6 mM CoBr2 gives only a 8.8% TA yield. Addition of an
amount of Cu so that the Cu : Co ratio is only 0.010 mol/mol
(i.e. the Cu concentration is only 0.026 mM) boosts the TA
yield to 55%. Thus the TA and CO2 yields are sensitive to very
small changes in the Cu:Co ratio at low Cu proportions. The
TA yield reaches a maximum value of 65% at a Cu : Co ratio
of 0.1 : 0.9 and then quickly decreases as the proportion of Cu
is raised further. Also the CO2 yield reaches a minimum at this
point. This change in activity can be also seen by observing the
changes in the selectivity to the intermediates, such as p-toluic
acid. On the other hand, the BA yield remains constant over the
whole range.
As expected, a less active catalyst, like EuBr3, usually shows
better activity at the highest temperature but the loss to BA is
always lower at 330 than at 380 ◦C.
A similar trend is observed for other catalyst combinations
such as Cu/Mn/Br. The yield is higher when there is an excess
of Mn over Cu and the CO2 yield is lower. Details of all these
experiments can be found in the ESI, Table S13.†
Variation of CoBr2 & NH4Br in the Cu/Co/NH4/Br catalyst.
We have seen that the variation of the Cu : Co ratio, Fig. 4,
and the presence of NH4Br, Table 2, both have a positive effect
on the TA yield and decrease the formation of CO2. A study
2402 | Green Chem., 2011, 13, 2397–2407
This journal is
The Royal Society of Chemistry 2011
©

View Online
The shape of the curves is different at the two temperatures.
At 380 ◦C a clear maximum in TA yield is obtained for
[Co] = 0.43 mM with a similar CO2 yield as for the Cu/Co/Br
catalyst without NH4Br. The TA selectivity does not change
significantly. At 330 ◦C no local maxima are observed. However
it is noticeable how even a low proportion of CoBr2 is able to
enhance greatly the performance of the Cu/NH4/Br catalyst,
which is not very active at this temperature. The initial TA
yield of 44% rose to 62% upon substitution of just the 25%
of NH4Br by CoBr2, with the CO2 yield decreasing from 21% to
19%. The BA selectivity does not vary significantly for any of
the temperatures studied. These results demonstrate a beneficial
effect on combining CoBr2 and NH4Br in the presence of CuBr2.
Fig. 5 Effect of reactor temperature on the yields TA, CO2 and BA and
the TA selectivity using a Cu/Co/Br catalyst in SCW. These experiments
were carried out using Cu/Co/Br = (0.1/0.9/2.0) and a slightly different
reactor configuration: inlet catalyst pipe is 1/8¢¢ instead of 1/16¢¢. (ꢀ)
TA selectivity, (ꢁ) TA yield, (᭡) CO2 yield, (᭜) BA yield.
Effect of Br : metals ratio. Table 3 shows the effect of adding
extra bromide as HBr or NH4Br to selected catalysts at 380
and 330 ◦C. In entries 1 and 2, HBr is added to MnBr2, the
“traditional” catalyst, and to Cu/Co/Br, the newly discovered
catalyst. In entries 3 and 4 NH4Br is added to the formulation
Cu/Co/NH4/Br at the proportion where the maximum in TA
yield was found in the previous section.
Several trends were observed: (1) apart from entry 4, the TA
yield decreases when small amounts of Br are added and then
slightly increases if more Br is added (probably due to the activity
of the Br catalyst itself); (2) the CO2 yield generally increases
upon Br- addition; and (3) the BA selectivity decreases upon Br
addition.
Entries 1 to 3 show that Br addition has negative effect on
catalyst performance. The higher TA yield observed when 8 ¥
2.6 mM of Br is added to MnBr2 is probably due to the activity
of HBr itself. Entry 2 shows that the optimum Br : metals ratio is
2. The same tendency is observed when NH4Br is used as source
of Br instead.
was subsequently performed on the incremental substitution
of CoBr2 by NH4Br in a Cu/Co/NH4/Br catalyst. It would
be beneficial to reduce the amount of CoBr2 in the catalyst
formulation in favour of NH4Br for the following reasons: (1)
NH4Br is less expensive; (2) CoBr2 is more prone to precipitation
in SCW due to hydrolysis.
The Cu and Br concentrations were held constant at 0.385
and 5.13mM respectively. The CoBr2:NH4Br ratio was varied in
the way that Co concentration is 25, 50 and 75% of the original
value in the Cu/Co/Br catalyst; that is 0.21, 0.43 and 0.64 mM
respectively. Two temperatures were studied: 330 and 380 ◦C.
Fig. 6 shows the results obtained. Details of all these experiments
can be found in the ESI, Table S15.†
Fig. 6 Effect of varying the relative concentration of CoBr2 and NH4Br in the Cu/Co/NH4/Br catalyst at (a) 380 and (b) 330 ◦C. (ꢀ) TA selectivity,
(ꢁ) TA yield, (᭡) CO yield, (᭜) BA yield, ( ) PTA yield.
᭹
2
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2403
©

View Online
Table 3 Effect of increasing Br : metals ratio for some selected catalysts at 380 ◦
C
Yield,
mol (%)
Selectivity
Entry
1
Catalysta
Br addeda
Br : metals mol/mol
CO2
TA
Burn
TA
BA
Mn/Br (1.0/2.0)
—
2
3
10
2
3
18.1
14.2
25.2
22.0
24.4
23.6
17.9
23.3
23.6
17.0
17.8
17.4
36.1
22.7
46.9
60.0
53.7
55.9
66.8
60.1
63.0
64.9
70.5
70.2
0.50
0.63
0.54
0.37
0.45
0.42
0.27
0.39
0.37
0.26
0.25
0.25
51.3
39.8
96.2
88.5
91.1
93.9
93.1
91.6
92.4
85.2
91.6
92.6
6.0
3.7
3.0
10.2
7.2
6.1
6.9
8.4
7.6
3.7
4.2
4.0
HBr (1.0)
HBr (8.0)
—
HBr (0.5)
HBr (1.0)
—
NH4Br (1.0)
NH4Br (2.0)
—
NH4Br (1.0)
NH4Br (2.0)
2
Cu/Co/Br
(0.07/0.42/1.0)
4
3
Cu/Co/NH4/Br
(0.15/0.43/0.85/2.0)
3.4
5.2
6.9
3.4
5.2
6.9
4b
Cu/Co/NH4/Br
(0.15/0.43/0.85/2.0)
a The numbers between brackets indicate the concentration of every species relative to 2.6 mM. b Temperature was 330 ◦C
Entry 4 is an exception to the general trend. Addition
of 2.6 mM NH4Br increases TA yield to the highest values
presented in this series of papers. Further addition of bromide
does not affect greatly the performance. Notice that the CO2
yield and BA selectivity remain almost constant in every case.
is the activation of the methyl group, i.e. reactions 1, 1a, 3, 3a.
Very small amounts of 4HMBA are usually detected, suggesting
that the sequence goes mainly through reactions 1 and 3 rather
than through 1a and 3a.
In acetic acid, the oxidation of pX to TA proceeds through a
catalyst-modified free radical chain mechanism. The reaction is
initiated by a oxidized bromide specie, Br∑ (reaction 6, Scheme
2). The propagation step is the rapid reaction of the benzylic
radical with molecular oxygen followed by the hydrogen atom
abstraction of the methyl group with the peroxy radical (reaction
8). The metal can react with the hydroperoxide in two different
ways: either by a redox reaction to form a benzylic alcohol
(reaction 9) or acting as a Lewis acid to generate the aldehyde by
dehydration (reaction 10), thereby by-passing reaction 9 and the
formation of the alcohol.39–41 Very low amount of such alcohols
are usually found, so reaction 10 dominates reaction 9 in Scheme
2. The aldehyde is then oxidized to the acid following a set of
analogous reactions but in addition the peroxyacid thus formed
rapidly oxidizes the catalyst metals.27 Finally, the oxidized metal
is reduced by reacting with Br- to restart the catalytic cycle.
When CoBr2 is used by itself as a catalyst, it acts by linking the
redox pair Co(III)/Co(II) with Br(0)/Br-. The same mechanism
has been proposed for the SCW process with Mn playing the
same role in SCW as Co does in acetic acid, see Scheme 3.
Discussion
Some of the major conclusions from this work are:
1. At a concentration of 2.6 mM, CuBr2 is a significantly better
catalyst than MnBr2 in terms of high TA yield and low burn,
even though Cu is a poor catalyst in acetic acid.23
2. If the Br- concentration is kept constant, some of the CuBr2
can be substituted by HBr or NH4Br to give Cu/H/Br and
Cu/NH4/Br catalysts which have higher activity than CuBr2
alone. We have found that both HBr and NH4Br are synergistic
with CuBr2.
3. There is a strong synergy between Cu and Co even though
CoBr2 alone is a poor catalyst in SCW.
4. The synergy of Cu and Co is not unique and several other
metals, e.g. Ni2+, also enhance the performance of CuBr2 in SCW
(see Tables S8 and S10).
5. The best Cu/Co/Br catalyst so far performs better in terms
◦
of TA selectivity and burn in sub-critical water at 330 C than
in SCW at 380 ◦C at the same catalyst loading and flowrates.
Mechanistically, the aerobic oxidation of pX is a complex
multi-step process with potentially competing pathways at
several points. Despite this complexity, it is possible to present
a plausible rationalisation which could explain how these
conclusions arise.
The basis of this rationalisation is that the mechanism of the
reaction in SCW is similar to that in acetic acid. This seems to be
a reasonable assumption because the sequence of intermediates
is the same1,7,8 and the TA selectivity and yield are favoured
by increasing the concentration of catalyst.6–8 In addition, the
relative yields of TA and the reaction intermediates give an
idea of the facility which the reactions 1 to 5 in Scheme 1
take place. In both acetic acid and SCW, we generally observe
higher concentrations of TA and PTA than for the rest of
compounds, indicating that the “difficult step” of the sequence
Why is CuBr2 an active catalyst in SCW?
Presumably, CuBr2 acts as a catalyst in SCW via a similar
mechanism to Scheme 3. This is slightly surprising because
Cu does not have a higher consecutive oxidation state in
aqueous solution at ambient temperatures. However, its redox
behaviour will change at the temperature is increased. The
redox behaviour of the different metals in aqueous solution
can be represented in E–pH diagrams (Pourbaix diagrams)
that show the thermodynamically stable chemical form of the
metal at different redox potential and pH. Such diagrams have
been available for many years for metals in water at room and
moderate temperatures.43
More limited studies have been carried out at high tempera-
tures. Beverskog and Puigdomenech reported E–pH diagrams
◦
for Cu, Fe, Ni, Zn at temperatures up to 300 C.44–47 Fig. 7
2404 | Green Chem., 2011, 13, 2397–2407
This journal is
The Royal Society of Chemistry 2011
©

View Online
Scheme 2 Reaction mechanism of the catalyzed oxidation of pX to TA in acetic acid.
Scheme 3 Suggested redox cascade for metal bromide catalyzed
oxidation of pX in water.42
shows that at increasing temperature, an increasingly broad
region of stability for Cu+ appears. Indeed, the potential for
the pair Cu+/Cu2+ increases with temperature up to 0.5 V at
◦
300 C. Furthermore, the existence of Cu+ in presence of Br-
has been experimentally observed by EXAFS structural studies
at 325 ◦C.48 At this temperature Cu2+ was found to be a highly
oxidizing species, since it was able to oxidize Pt(0) from the
spectroscopic cell. Thus Cu(II) in SCW would have a sufficiently
high redox potential to oxidize bromide and even pX itself. In
addition, due to its stability in high temperature water, Cu(I)
would remain in the solution thus being oxidized by the peroxide
to close the redox cycle.
Why are HBr and NH4Br effective in partial replacement of
CuBr2 in the catalyst?
Scheme 3 shows that the metal and Br- play different roles in the
catalytic cycle and there is no a priori reason why the 1 : 2 ratio
defined by the formula CuBr2 should be the optimal ratio for
highest activity. Additional experiments showed that addition of
several different bromide sources enhanced the activity of CuBr2
but that HBr and NH4Br were particularly effective (see Table
S3 in the ESI†).
In acetic acid, there is little difference in Co/Br catalyzed
oxidation when using NaBr or NH4Br as source of bromide.49
Since NH4Br and NaBr, at ambient conditions, are simple
bromide salts, one might expect their catalytic activity also to
be similar in SCW. Instead we find that NaBr is a very inferior
catalyst to NH4Br when combined with CuBr2. When Cu/Br
and NH4Br or HBr are added together they are not acting
independently, i.e. there is a synergistic interaction. NH4Br has
a similar catalytic effect to HBr both when used alone or in
combination with copper. In ambient water, the ammonium
Fig. 7 E–pH diagrams for copper in water at (a) 25 ◦C and (b) 300 ◦C.
Note the greatly enlarged coexistence area for Cu+ at 300 ◦C. We believe
that the existence of Cu(I) in solution is a key factor in making Cu a
highly active catalyst for the oxidation of pX. The arrows indicates the
value of neutral pH. The dashed sloped lines are the stability limits of
water under these conditions. The diagrams have been redrawn from
reference 44, which also includes data at higher pH. Reproduced by
permission of ECS – The Electrochemical Society.
+
cation is a weak acid: NH4 Br- ꢀ NH3 + HBr; Ka = 1.8 ¥ 10-5.
NH4OH become weaker, i.e. less dissociated;50 that would cause
the equilibrium to be displaced towards NH3 + HBr and hence
there would be an increase in the amount of HBr in the medium.
In SCW the equilibrium could be quite different. Due to the low
dielectric constant, some of the non-ionic acids such as HCl and
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2405
©

View Online
The lower activity of Cu/NH4/Br catalyst at 330 ◦C may be due
to a shift in this equilibrium at different temperature. The NH3
might also coordinate with either Cu(I) or Cu(II) and change
their catalytic behaviour.
Why does CoBr2 enhance the activity of CuBr2 in SCW?
The discovery of the synergy between Co and Mn gave scientific
explanation to the superior activity of this combination in the
development of the acetic acid process.28 This combination of
two metals enables the oxidation to proceed through a set of
linked redox reactions thus lowering the activation energy very
substantially, Scheme 4.15,23,51
Fig. 8 Correlation between (ꢁ) TA yield and (ꢀ) TA selectivity with
Lewis acidity of M in the catalyst Cu/M/Br. Fe3+ is shown rather than
Fe2+ because our experiments used FeBr3 as catalyst.
Scheme 4 Redox cascade for the Co/Mn/Br catalyzed oxidation of pX
in acetic acid.51
and precipitate during oxidation.33 And as mentioned before this
is likely to happen in water.
The experiments summarised in Table 2 and in Table S10
show that CuBr2 also experiences a synergistic enhancement of
its activity with a range of metal bromides, particularly Co but
also Fe, Ni and Zn. It is possible that some of these metals acted
via linked redox cycles similar to those in Scheme 4. However,
this clearly cannot apply to all metals because Zn, for example,
is not redox active.
There is, however, a second role for a metal catalyst, namely
as a Lewis acid to promote the conversion of the peroxide
intermediate to the aldehyde, reaction 10 in Scheme 2. Increasing
the rate of this transformation willnecessarilydecreasethe extent
of the competing generation of OH∑ which is formed by the
thermal dissociation of the peroxide. The hydroxyl radical is
highly energetic, with a redox potential of 2.9 V and unselective.
It attacks the aromatic ring and hence destroys the aromatic
moiety.22
Metals such as Zr, Fe, V, Ti, Hf etc. activate Co and Co/Mn
catalyst in oxidation of methylaromatics in acetic acid.39,40 The
intermediates observed suggest that they favour reaction 10 and
there is a correlation between the degree of activation and Lewis
acidity.40 The same may be true for synergies with CuBr2 in SCW.
If we take Z/r2 as a measure of Lewis acidity, where Z is the
formal charge of the cation and r its ionic radius,52 there is a
reasonably good correlation between Z2/r of a particular metal
and the TA yield and selectivity when that metal is added to
CuBr2, see Fig. 8. Particularly encouraging is the fact that there
is monotonic increase in TA yield in the presence of the non-
transition metals which can only act as Lewis acids i.e. Zn(II) >
Ca(II) > Na(I) > K(I).
However, the correlation in Fig. 8 is only positive up to a
maximum at Z2/r £ 6.2 (Co) and negative for higher values. This
suggests two opposing effects related to Lewis acidity. Thus, as
already explained, a Lewis acid can catalyze the dehydration
of the peroxide (Scheme 2, reaction 10) but a too strong Lewis
acid may have its activity limited by hydrolysis and eventual
precipitation of the metal oxide, MBrx + yH2O → MBrx-y(OH)y
+ yHBr. Indeed, it is known in acetic acid that Zr, which is a very
strong Lewis acid, in a Co/Mn/Zr/Br catalyst does hydrolyze
Hydrolysis of metal salts and formation of metal oxide
nanoparticles in SCW is a significant research area in its own
right.53–55 It is therefore quite interesting that Co2+ salts are
known to give rise to oxide particles in SCW much more readily
than Cu2+.54 This agrees with some of the observations we have
already made. Furthermore, we have noticed that Cu-based
catalysts give rise to less fouling of the reactors by precipitation
than MnBr2. The presence of HBr, will lower the pH and further
discourage precipitation.
Why does Cu/Co/Br work better at 330 ◦C than 380 ◦C?
Generally the rate of a given reaction increases with temperature.
However we have a complex set of reactions and a continuous
process so it not easy to evaluate the effect of temperature in
the reaction performance. Temperature can affect the process
by (a) modifying the reaction rate; (b) changing undesirable
side reactions such as decarboxylation, which deactivates the
catalyst3 or (c) changing the residence time as a consequence of
variation of density with temperature.
The shape of the TA selectivity and yield vs. temperature in
the Fig. 5 suggests the existence of opposing effects that are
optimized at 330 ◦C. T◦he large difference in TA selectivity and
yield from 310 to 330 C as well as the dramatic reduction of
intermediates (see Table S14) clearly indicates an enhancement
of the reaction rate with temperature. At higher temperatures,
however, only the by-products (BA and CO2) increase signifi-
cantly; hence effect (b) is making the difference. The fact that
◦
few intermediates are seen at 330 or 380 C also indicates that
residence time is sufficient to complete the reaction.
In some cases temperature can have a major impact on the
activity of a catalyst. The clearest example, as shown above is
the Cu/NH4/Br catalyst. A possible explanation is a shift in the
+
equilibrium NH4 Br- ꢀ NH3 + HBr. That would be consistent
with the fact that, in Table 3, further addition of NH4Br does
improve the performance of the Cu/Co/NH4/Br catalyst at
2406 | Green Chem., 2011, 13, 2397–2407
This journal is
The Royal Society of Chemistry 2011
©

View Online
330 ◦C while at 380 ◦C it does not. This also explains the shape
of the curves in Fig. 6.
12 Y.-L. Kim, J.-D. Kim, J. S. Lim, Y.-W. Lee and S.-C. Yi, Ind. Eng.
Chem. Res., 2002, 41, 5576–5583.
13 M. Osada and P. E. Savage, AIChE J., 2009, 55, 710–716.
14 J. B. Dunn and P. E. Savage, Green Chem., 2003, 5, 649–655.
15 W. Partenheimer and M. Poliakoff in Handbook of Green Chemistry
– Green Catalysis: Homogeneous Catalysis, ed. P. T. Anastas and R.
H. Crabtree, Wiley-VCH, Weinhein, 2009, vol 1, ch. 12, pp. 375–
397.
Summary and Conclusions
A detailed study of the catalytic effect of different substances
on the oxidation of pX to TA as catalyst has been carried out.
CuBr2 itself, or combined with other metallic bromides, has
been shown to be the best in terms of TA selectivity, of TA
yield and of burn. Low concentrations of copper are enough
to give good performances provided that the overall amount of
bromide is high enough. Br can be added in several different
ways (as metallic bromide, HBr or NH4Br), all giving high TA
yields and selectivities.
The role of the different species in the reaction mechanism
has been discussed: Cu is the most active redox catalyst, Mn
and Fe are weaker redox catalysts. Many metallic bromides can
activate CuBr2 acting as redox metals, Lewis acids or sources
of HBr by hydrolysis. There is a strong synergy between Cu
and Co and probably the best composition identified so far is
Cu/Co/NH4/Br because the concentrations of both metals are
minimized, the TA yield is high and the burn is low.
16 H. Weinga¨rtner and E. U. Franck, Angew. Chem., Int. Ed., 2005, 44,
2672–2692.
17 A. Kruse and E. Dinjus, J. Supercrit. Fluids, 2007, 39, 362–380.
18 J. Fraga-Dubreuil and M. Poliakoff, Pure Appl. Chem., 2006, 78,
1971–1982.
19 P. E. Savage, Chem. Rev., 1999, 99, 603–621.
20 D. Bro¨ll, C. Kaul, A. Kra¨mer, P. Krammer, T. Richter, M. Jung, H.
Vogel and P. Zehner, Angew. Chem., Int. Ed., 1999, 38, 2998–3014.
21 M. D. Bermejo and M. J. Cocero, AIChE J., 2006, 52, 3933–3951.
22 E. Pe´rez, J. Fraga-Dubreuil, L. Dudd, E. Garcia-Verdugo, P. A. Ham-
ley, M. L. Thomas, C. Yan, W. B. Thomas, D. Housley, W. Parten-
heimer and M. Poliakoff, Selective aerobic oxidation of para-xylene
in sub- and supercritical water. Part 1. Comparison with ortho-xylene
and role of the catalyst, Green Chem., DOI: 10.1039/c1gc15137a.
23 W. Partenheimer, Catal. Today, 1995, 23, 69–158.
24 W. Partenheimer, J. Mol. Catal., 1991, 67, 35–46.
25 A. Saffer and R. S. Barker, Chem Eng. Prog., 1968, 64, 20.
26 US Patent 4 786 753, 1988.
27 R. K. Gipe and W. Partenheimer, Stud. Surf. Sci. Catal., 1997, 110,
1117–1127.
Overall, this paper has described the discovery of new and
substantially different catalysts for the aerobic oxidation of pX
to TA in high temperature and supercritical water. Given this
new direction, there is almost certainly opportunity for further
improvement and optimization of this fascinating reaction.
28 W. Partenheimer in Catalysis of Organic Reactions, ed. F. E. Herkes,
Marcel Dekker Inc., New York, NY, 1998.
29 H. D. Holtz, J. Org. Chem., 1972, 37, 2069–2074.
30 H. D. Holtz, Chem Commun., 1971, 19, 1166.
31 F. Recupero and C. Punta, Chem. Rev., 2007, 107, 3800–3842.
32 N. V. Sidwick, The chemical elements and their compounds, Oxford-
Clarendon Press, 1950, vol 1.
33 US patent 5 112 992, 1954.
34 US patent, 2 673 217, 1957.
Acknowledgements
35 W. Partenheimer and R. K. Gipe in Catalytic Selective Oxidation, ed
S. T. Oyama and J. W. Hightower, American Chemical Society, 1993,
pp. 81–88.
We thank EPRSC, the University of Nottingham and INVISTA
for financial support. We thank Graham Aird, Stuart Coote, Lu-
cinda Dudd, Annette Matthews, Ian Pearson, Tony Richardson,
John Runnacles, Peter Fields, Richard Wilson and Mark Guyler
for their help.
36 US Patent 5 183 933, 1993.
37 V. N. Aleksandrov, Kinet. Katal., 1978, 19, 1057–1060.
38 J. B. Dunn, M. L. burns, S. E. Hunter and P. E. Savage, J. Supercrit.
Fluids, 2003, 27, 263–274.
39 W. Partenheimer, J. Mol. Catal. A: Chem., 2003, 206, 105–119.
40 W. Partenheimer, J. Mol. Catal. A: Chem., 2003, 206, 131–144.
41 E. J. Y. Scott, J. Phys. Chem., 1970, 74, 1174–1182.
42 Y. Chen, J. L. Fulton and W. Partenheimer, J. Am. Chem. Soc., 2005,
127, 14085–14093.
References
1 P. A. Hamley, T. Ilkenhans, J. M. Webster, E. Garc´ıa-Verdugo, E.
Vernardou, M. J. Clarke, R. Auerbach, W. B. Thomas, K. Whiston
and M. Poliakoff, Green Chem., 2002, 4, 235–238.
2 J. Fraga-Dubreuil, E. Garc´ıa-Verdugo, P. A. Hamley, E. M. Vaquero,
L. M. Dudd, I. Pearson, D. Housley, W. Partenheimer, W. B. Thomas,
K. Whiston and M. Poliakoff, Green Chem., 2007, 9, 1238–1245.
3 E. Garc´ıa-Verdugo, E. Vernardou, W. B. Thomas, K. Whiston, W.
Partenheimer, P. A. Hamley and M. Poliakoff, Adv. Synth. Catal.,
2004, 346, 307–316.
43 M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous Solu-
tions, Pergamon Press, 1966.
44 B. Beverskog and I. Puigdomenech, J. Electrochem. Soc., 1997, 144,
3476–3486.
45 B. Beverskog and I. Puigdomenech, Corros. Sci., 1996, 38, 2121–
2135.
46 B. Beverskog and I. Puigdomenech, Corros. Sci., 1997, 39, 969–
980.
47 B. Beverskog and I. Puigdomenech, Corros. Sci., 1997, 39, 107–114.
48 J. L. Fulton, M. M. Hoffman, J. G. Darab and B. J. Palmer, J. Phys.
Chem. A, 2000, 104, 11651–11663.
4 E. Garcia-Verdugo, J. Fraga-Dubreuil, P. A. Hamley, W. B. Thomas,
K. Whiston and M. Poliakoff, Green Chem., 2005, 7, 294–300.
5 J. Fraga-Dubreuil, E. Garc´ıa-Verdugo, P. A. Hamley, E. Pe´rez,
I. Pearson, W. B. Thomas, D. Housley, W. Partenheimer and M.
Poliakoff, Adv. Synth. Catal., 2009, 351, 1866–1876.
6 R. L. Holliday, B. Y. M. Jong and J. W. Kolis, J. Supercrit. Fluids,
1998, 12, 255–260.
49 Y. Kamiya, T. Nakajima and K. Sakoda, Bull. Chem. Soc. Jpn., 1966,
39, 2211–2215.
50 K. P. Jonhston and J. B. Chlistunoff, J. Supercrit. Fluids, 1998, 12,
155–164.
51 W. Partenheimer, in Catalytic Selective Oxidation, ed. S. T. Oyama
and J. W. Hightower, American Chemical Society, 1993, ch. 7.
52 R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor.
Gen. Crystallogr., 1976, 32, 751–767.
7 P. E. Savage, J. B. Dunn and J. Yu, Combust. Sci. Technol., 2006, 178,
443–365.
8 J. B. Dunn and P. E. Savage, Environ. Sci. Technol., 2005, 39, 5427–
5436.
53 T. Adschiri, Y. Hakuta, K. Sue and K. Arai, J. Nanopart. Res., 2001,
3, 227–235.
9 J. B. Dunn and P. E. Savage, Ind. Eng. Chem. Res., 2002, 41, 4460–
4465.
54 K. Byrappa, T. Adschiri and Progr, Prog. Cryst. Growth Charact.
Mater., 2007, 53, 117–166.
55 E. Lester, P. Blood, J. Denyer, D. Giddings, B. Azzopardi and M.
Poliakoff, J. Supercrit. Fluids, 2006, 37, 209–214.
10 J. B. Dunn, D. I. Urquhart and P. E. Savage, Adv. Synth. Catal., 2002,
344, 385–394.
11 Y.-L. Kim, S. J. Chung, J.-D. Kim, J. S. Lim, Y.-W. Lee and S.-C. Yi,
React. Kinet. Catal. Lett., 2002, 77, 35–43.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2397–2407 | 2407
©