Standard Article

You have free access to this content

The Aerobic Oxidation of p-Xylene to Terephthalic acid: A Classic Case of Green Chemistry in Action

Part 1. Homogeneous Catalysis

  1. Walt Partenheimer1,
  2. Martyn Poliakoff2

Published Online: 15 MAR 2010

DOI: 10.1002/9783527628698.hgc012

Handbook of Green Chemistry

Handbook of Green Chemistry

How to Cite

Partenheimer, W. and Poliakoff, M. 2010. The Aerobic Oxidation of p-Xylene to Terephthalic acid: A Classic Case of Green Chemistry in Action. Handbook of Green Chemistry. 1:12:375–397.

Author Information

  1. 1

    E.I. DuPont de Nemours & Co., Inc., Central Research and Development, Wilmington, DE, USA

  2. 2

    University of Nottingham, School of Chemistry University Park, Nottingham, UK

Publication History

  1. Published Online: 15 MAR 2010

Chemistry Terms

Choose one or more boxes to highlight terms.


  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

The sections in this article are

  • Introduction
  • Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  • Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  • Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  • Potential Processes Using Water as a Solvent
  • Summary and Final Comments

Although the ultimate aim of most green chemistry research is industrial implementation, the research is often perceived as being primarily an academic pursuit. This chapter is different. It focuses on the continuing development and improvement of a major commercial bulk chemical process, the oxidation of p-xylene. The two authors, W.P., a US industrial chemist with more than 30 years' experience of oxidation processes, and M.P., a UK academic green chemist, were first brought together by the oxidation of p-xylene in supercritical water, which forms the final section of this chapter. Their aim is to bring this story of this process to a wider chemical public. The chapter shows how apparently simple reactions can have unexpected mechanistic complexity and how this complexity offers fascinating opportunities for major innovation even in well-established processes. A key objective is to show that, despite the fact that the drivers continue to be largely commercial, the developments in this process have all the appearance of being guided by the Principles of Green Chemistry. Indeed, the oxidation of p-xylene may well be one of the best demonstrations that greener processes are more profitable.

12.1 Introduction

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

The commercial methods for the manufacture of terephthalic acid are based on dissolving p-xylene and the components of the catalyst into a given solvent and passing a source of molecular oxygen (O2), usually air, through it (Scheme 12.1).

original image

Scheme 12.1.

Most of the manufactured terephthalic acid is used to produce poly(ethylene terephthalate) (PET) by its reaction with ethylene glycol (Scheme 12.2).

original image

Scheme 12.2.

One of the most visible forms of PET is in the manufacture of water bottles, although it has many other forms, including fibers, films and engineered plastics. Over 6 million tons per year were produced in the 2002 (1).

The structure of this chapter is based on the historical methods of preparation of terephthalic acid, demonstrating how the discovery of each method allowed for a large increase in greenness in the manufacturing process (Table 12.1). Each advance is based on the invention of new catalysts; small variants in these processes have been discussed elsewhere (2). After a brief description of each process, the greenness of the process will be outlined and then the catalytic mechanisms responsible for the green aspects will be highlighted, ending with a discussion of the weaknesses in the process. Our discussion of greenness will be based on the 12 Principles of Green Chemistry and Engineering (3); those that are relevant to this chapter are given in Table 12.2. Engineering aspects of these processes are beyond the scope of this chapter and have been discussed elsewhere (1, 2, 4-7).

Table 12.1. Past, present and future methods for manufacturing terephthalic acid from p-xylene
  1. a

    Autoxidation is defined as products obtained by reaction of a hydrocarbon using dioxygen as the primary oxidant.

PastNone – stoichiometric
PresentCo(OAc)2Acetic acidWitten process. Autoxidation followed by esterification
PresentCo(OAc)2Acetic acidEastman process. Autoxidation with an organic co-oxidant
PresentCo(OAc)2[BOND]Mn(OAc)2[BOND]HBrAcetic acidMid-Century process. Autoxidation
Future (?)MnBr2 (?)WaterAutoxidation
Table 12.2. Selected Principles of Green Chemistry and Engineering used in this chapter (adapted and modified from (3))
  1. a

    Note that the numbering of these principles has been changed from those used in (3).

1It is better to prevent waste than to treat or clean up waste
2Atom economy; maximize all materials used into the desired product
3Safer solvents; should be innocuous and made unnecessary if possible
4Design for energy efficiency and simplicity
5Use renewable feedstocks
6Use catalysts, as selective as possible, rather than stoichiometric reagents
7Avoid derivatization or masking of the reagents
8Less hazardous chemical syntheses; methods should be designed to generate substances that have little or no toxicity.

The earliest methods for preparing terephthalic acid used stoichiometric reagents such as permanganate and dichromate salts. The first commercial manufacturing methods were initiated with the invention of the Co(II) acetate catalyst using acetic acid as the solvent in 1938 (8). There are two variants of this method based on aerobic oxidation with the products being separated and purified by esterification (Witten process) and air oxidation using a co-oxidant such as acetaldehyde (Eastman process). A giant step forward in terephthalic acid manufacture was the invention of metal–bromide catalysts, the most prominent being the combination of Co(II) and Mn(II) acetates with bromide salts in 1954 (4, 9, 10). Finally, we discuss a possible future process based on changing the solvent from acetic acid to ‘hot’ water (250–374 °C) or supercritical water (Tc = 374 °C, Pc = 221 bar).

12.2 Methods of Making Terephthalic Acid Using Stoichiometric Reagents

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

Introductory textbooks in organic chemistry and also advanced treatises on oxidation chemistry teach the oxidation of methylaromatic compounds using permanganate and dichromate salts (11), e.g. Scheme 12.3.

original image

Scheme 12.3.

The atom economy of the reaction in Scheme 12.1 to give terephthalic acid (Green Principle 2, Table 12.2), based on the moles of product and by-products produced (12, 13), is only 38% Current terephthalic acid plants often have a capacity of at least 1 billion lb per year. It is difficult to conceive of a process using potassium permanganate and generating multi-million pounds of waste in the form of manganese(II) oxide and alkali metal hydroxide.

12.3 Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

We begin by discussing the large increase in greenness in going from a stoichiometric preparation of terephthalic acid to a catalytic process using O2 as the oxidant. Then we describe the catalytic mechanisms responsible for the increasing greenness. Finally, we explain the disadvantages of the Co–acetic acid-based chemical processes so that one can appreciate the value of the next advance in greenness – the invention of the Co[BOND]Mn[BOND]Br catalyst.

Principle 2 – Atom Economy The Co catalyst allows for the direct incorporation of dioxygen into p-xylene in which the atom economy is increased from 38% in Scheme 12.3 to 82% in Scheme 12.1. The number of by-products is reduced from 3 in the stoichiometric oxidation to just simply water in Scheme 12.1. Although water is normally considered a green by-product, it is not ‘green’ in this instance since it deactivates the catalyst (14); hence the energy-intensive separation of the water from the acetic acid eventually becomes necessary. This is one of the key driving forces for developing a process that uses water as a solvent.

Principle 5 – Renewable Feedstocks Air, which contains 20.9% of O2, is usually used as the source of O2 in these processes. Air is a renewable resource since O2 is generated biologically via photosynthesis. p-Xylene, however, is not renewable since it comes from petroleum refining.

Principle 4 – Design for Simplicity and Energy Efficiency One outstanding green property of the acetic acid solvent is that terephthalic acid is highly insoluble in it, whereas all the other intermediates (shown in Scheme 12.4) and by-products produced are much more soluble. Thus, as p-xylene is being oxidized in acetic acid, the terephthalic acid preferentially precipitates. The product is thus separated from the by-products and catalyst as it is being produced. The solubility of terephthalic acid at 150 °C, the approximate operating temperature of these processes, is only 0.38 g per 100 g of acetic acid. Hence the one disadvantage of homogeneously catalyzed processes – that of separation of the product from the solvent and catalyst – is largely absent when acetic acid is used.

original image

Scheme 12.4. Intermediates present during the autoxidation of p-xylene to terephthalic acid. For a more detailed description and the relative reactivities of the intermediates, see (21).

One of the great green advantages of this aerobic reaction is that the enthalpy change of the reaction in Scheme 12.1 is −336 kcal mol−1. This enormous amount of heat can be captured by conversion of water to steam in heat exchangers and then used to provide energy in subsequent chemical steps – such as for the purification of the terephthalic acid or for conversion of terephthalic acid to PET. For example, one method used to purify terephthalic acid is by recrystallization in water at around ∼300 °C. The energy required to heat 1 mol of terephthalic acid from room temperature (25 °C) to 300 °C is 275 kcal mol−1, which could be supplied by the heat of reaction.

Principle 8 – Less Hazardous Chemical Synthesis Producing Substances with Little or No Toxicity Both terephthalic acid and dimethyl terephthalate have low toxicity and cause only mild and reversible irritation to the skin, eyes and the respiratory system (16).

Principles 1 and 6 – Use Selective Catalysts to Reduce By-products Co(II) acetate dissolved in acetic acid strongly activates molecular oxygen and also reduces undesirable by-products. Activation of oxygen is necessary because it exists in a triplet electronic spin state inline image and p-xylene is in a singlet spin state. This dissimilarity in spin produces a large activation energy barrier, which causes the reaction to be very slow (17). Experimentally, this can be shown by heating toluene in acetic acid to 205 °C at 34 bar pressure for 25 min and obtaining only a trace of benzaldehyde (17, 18). However, addition of Co(II) acetate immediately initiates a reaction and benzyl alcohol, benzaldehyde and benzoic acid are obtained. The increase in activity and selectivity that Co(II) imparts to this free radical chain mechanism has been demonstrated by oxidation in acetic acid of 4-methylbenzaldehyde (p-tolualdehyde) one of the intermediates in p-xylene oxidation (15). The oxidizability of 4-methylbenzaldehyde is 340 times greater than that of p-xylene; consequently, its oxidation can be observed even in the absence of a catalyst. As shown in Scheme 12.5, addition of the cobalt catalyst increases the rate of reaction by a factor of 2.5 and dramatically decreases the yields of the by-products. The yield of toluene is reduced by a factor of 10, that of p-tolyl formate is decreased by a factor of 300 and p-cresol falls to undetectable amounts. The amount of CO2 is reduced by a factor of 7.4! p-Cresol is a very strong antioxidant, the concentration of which increases with time in the uncatalyzed reaction and eventually causes the reaction to terminate at a conversion of only 43%. In the presence of cobalt catalyst, the reaction continues to 100%, giving high yields of p-toluic acid (15).

original image

Scheme 12.5. p-Toluic acid and by-products from the oxidation of p-tolualdehyde (4-methylbenzaldehyde) in 5% water–acetic acid with and without the presence of Co(II) acetate (at 10 mM).

The mechanistic rationale for the dramatic increase in reaction rate and selectivity is given in Scheme 12.6 (Reactions 12.1–12.16). The initiation, propagation and termination steps of the free radical chain mechanism are given in Reactions 12.1–12.4. The product of these reactions is 4-methylperoxybenzoic acid. The bond strengths of peroxides are weak (∼25–30 kcal mol−1) and, above 80 °C, the thermal dissociation of these bonds can occur to give the carboxyl radical and hydroxyl radicals, Reaction 12.5. The carboxyl radical readily undergoes decarboxylation, Reaction 12.9, to give the 4-methylphenyl radical. Hydrogen atom abstraction of this radical generates the observed by-product, toluene. The 4-methylphenyl radical will also react at a diffusion-controlled rate with O2 and go through a series of reactions to generate the 4-methylcresol, Reaction 12.13. The hydroxyl radical, generated in Reaction 12.5, is always undesirable in selective oxidations, for two reasons. (i) It is highly energetic and reacts exothermically with virtually any available C[BOND]H bond; it reacts preferentially with the aromatic ring, with an enthalpy change of −15 kcal mol−1, to form the cyclohexyldienyl radical, which subsequently rearranges to 2-hydroxyl-4-methylbenzaldehyde, Reaction 12.6. (ii) The reaction rates of the hydroxyl radical with organic substrates is often close to diffusion controlled; it reacts with p-xylene at a rate of 7.0 × 109 L mol−1 s−1) (19). The observed by-product, p-tolyl formate, forms via the Baeyer–Villiger rearrangement of the reaction of the 4-methylbenzaldehyde with the 4-methylperoxybenzoic acid, Reaction 12.11. The greenhouse gas CO2 forms from the oxidation of the phenols, Reactions 12.8 and 12.15, and from the decarboxylation of the 4-methylbenzylcarboxyl radical, Reaction 12.9. The cresols oxidize to the quinones, which are highly colored and hence undesirable, since one of the key specifications for commercial terephthalic acid is a high degree of whiteness.

original image

Scheme 12.6. The mechanism for the formation of toluene, p-cresol, p-tolyl formate and carbon dioxide from the autoxidation of 4-methylbenzaldehyde (p-tolualdehyde).

It is well established that Co(II) reacts with peroxyaromatic acids at a very high rate (15). In acetic acid, the reaction of 3-chloroperoxybenzoic acid with Co(II) is 400 000 times faster than its thermal dissociation (Scheme 12.7). The reaction with Co(II) is also much more selective, giving higher yields to the carboxylic acid with significant decreases in the 4-chlorobenzene by-product (the equivalent of toluene formation in Scheme 12.6) and CO2. Thus in Scheme 12.6, the reaction of Co(II) with the 4-methylperoxyacid, Reaction 12.16, is much faster than both the thermal dissociation, Reaction 12.5, and the Baeyer–Villiger reaction, Reaction 12.11. The direct formation of the aromatic acid via Reaction 12.16 is therefore highly favored when a cobalt catalyst is added and is consistent with the large observed reduction in yield of the by-products, toluene, 4-methylcresol, p-tolyl formate and CO2.

original image

Scheme 12.7. A comparison of the thermal decomposition and stoichiometric oxidation of 3-chloroperbenzoic acid with Co(II) acetate in acetic acid at 60 °C in 10% water–acetic acid (20).

Co(II) acetate appears to be unique in its ability to catalyze autoxidations of this type. One of the fundamental problems with free radical chain mechanisms is how the initiation step occurs (18). It has been shown that Co(II) acetate reacts with trace amounts of peroxides in acetic acid to generate small amounts of Co(III) and of products which are derived from the formation of the methyl radical (methyl acetate, CO2 and CH4). Both the methyl radical, CH3[FREE RADICAL], and Co(III), are known to generate the benzylic radical (CH3PhCH2[FREE RADICAL]) and hence initiate the reaction (Scheme 12.8a). Of all the first-row transition metals (Ti, V, Cr, Mn, Fe, Co, Ni, Cu and Zn), only Co has this ability. Similarly, Co is unique in its ability to catalyze the oxidation of p-toluic acid, one of the intermediates in p-xylene oxidation (18).

original image

Scheme 12.8. The initiation of the oxidation of p-xylene with (a) Co, (b) Co[BOND]Br and (c) Co[BOND]Mn[BOND]Br catalysts. The t1/2 data are from experiments at 60 °C in 10% water–acetic acid. Vertical distances between Co(II)[BOND]Co(III), Mn(II)[BOND]Mn(III) and Br(–1)–Br(0) are proportional to their redox potentials. Co(III)a and Co(III)i are different forms of Co(III) with the activity of Co(III)a being greater than that of Co(III)i. Redrawn from (55).

Not only is the Co(III) generated by trace amounts of peroxide but also, once the free radical chain mechanism is operational, Co(III) is generated from the peroxy acid, Reaction 12.16 (Scheme 12.6), and by reaction with the highly energetic hydroxyl radical. As discussed above, the reduction of the steady-state concentration of the hydroxyl radical in these reactions is highly desirable because of the lack of selectivity and high reactivity of OH[FREE RADICAL] toward virtually any organic substrate (19). Co(II) and most other metals used in autoxidation react very quickly with the hydroxyl radical (Table 12.3), thereby lowering its concentration and hence increasing the selectivity of the process. The increase in selectivity imparts ‘greenness’ by decreasing waste.

Table 12.3. Rates of reaction of the hydroxyl radical with selected compounds in aqueous solution (19)
Species reacting with OH radicalProduct of reactionpHRate constant (L mol−1 s−1)
Bromide ionBrOH11.1 × 1010
Co(II)Co(III)OH78 × 105
Mn(II)Mn(III)OH6.72.9 × 107

This free radical mechanism explains how the conservation of electronic spin, which prevents the direct reaction of inline image with hydrocarbons, can be overcome. In the propagation step of the free radical mechanism, the hydrocarbon is itself a radical species in a doublet spin state and reacts essentially at diffusion-controlled rates with O2 to give the peroxy radical:

mathml alt image(12.1)

We have seen how the aerobic autoxidation of p-xylene is superior in greenness to stoichiometric oxidation and how the cobalt catalyst imparts great selectivity in reducing many unwanted by-products. The problem with this chemistry, however, is that one is limited to low terephthalic acid yields of ∼15% with large amounts of partially oxidized intermediates shown in Scheme 12.4. The reasons for this are complex but include the following:

  • The reaction is sluggish even at high catalyst concentrations since Co(III) reacts slowly with p-xylene (Scheme 12.8a).

  • Studies using the Hammett equation show that the transition state of Co(III) responds strongly to electron-withdrawing substituents on the ring, such as a carboxyl group, and hence p-toluic acid is 76 times less reactive than p-xylene (15, 21, 22).

  • The steady-state concentration of Co(III) is high during the reaction, often 50–70% of the metal present (17, 23). Co(III) is a very strong oxidant in acetic acid [Co(II)[BOND]Co(III) reduction potential = 1.8 V] and is known readily to decarboxylate aromatic and aliphatic acids (24).

The structure of the coordination compounds in acetic acid–water solutions has been discussed in detail (25). The rates of ligand exchange of Co(II) are essentially instantaneous at room temperature and the coordination sphere contains acetic acid, acetate, water and oxygenated products such as benzaldehydes and carboxylic acids. Both acetic acid and p-toluic acid are decarboxylated by Co(III), generating the plethora of by-products shown in Scheme 12.9. The reaction is also strongly deactivated by the decarboxylation reaction because (i) the p-cresol formed is a strong antioxidant and (ii) Co(III) is in competition with the desired reaction, Reaction 12.21, which produces the benzylic radical as opposed to the undesired decarboxylations, Reactions 12.22 and 12.23 (Scheme 12.9).

original image

Scheme 12.9. Structure of possible coordination compounds of cobalt in water–acetic acid solutions, incorporating aromatic acids in the coordination sphere of Co(II), and some of the by-products from decarboxylation.

Why can the rate of reaction not be increased by increasing the temperature so that the oxidation can be driven to completion giving terephthalic acid in high yield? The activation energy for the Co(III) decarboxylation is very high (42 kcal mol−1 in anhydrous acetic acid, 33 kcal mol−1 in 5% water–acetic acid and 28 kcal mol−1 in 10% water–acetic acid (26)). As the temperature increases, the importance of decarboxylation, Reactions 12.22 and 12.23 become dominant at ∼120 °C and the relative rates of conversion to the desired aromatic acid decrease above this temperature (17, 23, 26).

Two strategies have been used to overcome the low terephthalic acid yields in the cobalt-catalyzed reaction. Both of these strategies use co-oxidation processes to drive the reaction of p-xylene to terephthalic acid to high yield. The Eastman process adds acetaldehyde to the oxidation reaction along with the p-xylene. The acetaldehyde is oxidized to peroxyacetic acid and this peroxy acid effectively oxidizes Co(II) to Co(III), which drives the reaction to completion. Large amounts of acetaldehyde are required (about 2 mol of acetaldehyde–p-xylene (27)), making the reaction stoichiometric in acetaldehyde. The process is a net producer of acetic acid. Similar processes have used paraldehyde (26) and 2-butanone (23) as the co-oxidants. In the Witten process, the initial oxidation mixture of p-toluic acid and terephthalic acid is esterified with methanol. The dimethyl terephthalate is separated from the monomethyl-p-toluic acid and the latter is recirculated back to the oxidation reactor. Then one has a co-oxidation of the more active p-xylene with the less reactive monomethyl-p-toluic acid to produce monomethyl terephthalate. Continuous oxidation followed by esterification and separation of the esters eventually produces dimethyl terephthalate with about 90% selectivity.

12.4 Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

A huge step forward in greening of this reaction occurred in 1954 with the discovery of metal–bromide catalysts by Landau, Saffer and Barker (10, 28). Their initial discovery was with an MnBr2 catalyst with 1,4-diisopropylbenzene as the substrate. Although the Co[BOND]Mn[BOND]Br catalyst is much more active than the Mn[BOND]Br catalyst and the diisopropylbenzene substrate is significantly more difficult to oxidize to terephthalic acid than is p-xylene, a large boost in activity and yield was apparent and they immediately recognized the commercial significance of the discovery. Some of the more important characteristics of this momentous discovery are given in Table 12.4. The metal–bromide catalysts give a yield boost in terephthalic acid from 15% to >95%! Furthermore, even at typical operating temperatures of around 200 °C, the solubility of terephthalic acid is only 1.7 g per 100 g of acetic acid and the hence a large majority of it precipitates as it is formed in the reactor. The composition of the solids is typically >99% terephthalic acid.

Table 12.4. Comparison of uncatalyzed with Co-, Co[BOND]Br- and Co[BOND]Mn[BOND]Br-catalyzed oxidation of p-xylene in acetic acid. Data from (4, 15, 17)
DescriptionNo catalystCo catalystCo[BOND]Br catalystCo[BOND]Mn[BOND]Br catalyst
Terephthalic acid yield (%), on a per pass basis (17)∼15∼95>95
Relative reactivity of catalyst (15, 17)
CO2, relative, substrate = 4-methylbenzaldehyde (p-tolualdehyde) (15)172.21
CO2, relative, substrate = p-xylene (17)5.91.0
Toluene, yield (%) (15)0.0780.0710.0
Tolyl formate, yield (%) (15)1.50.0050.0
Reactivity, p-xylene/p-toluic acid762221
Steady-state [Co(III)] (%) (17)11<0.6<0.6
Typical catalyst concentration (M)∼0.10.02/0.020.01/0.01/0.02
Range of operating temperatures (°C)70–15030–22530–225

Some of the enhancement in greenness includes the following.

Principles 1 and 6 – Prevent Waste. Use More Selective Catalysts As seen in Table 12.4, addition of bromide to cobalt not only quadruples the rate of reaction, but simultaneously make it much more selective and also the rate of CO2 generation decreases by a factor of 5.9. Addition of Mn to the Co[BOND]Br catalyst further increases it activity and selectivity (Table 12.4).

Principles 4 and 7 – Design for Energy Efficiency and Simplicity. Avoid Derivatization or Masking of the Reagents In the Witten process, esterification, a masking of the carboxylic acid group, is used to separate and purify the terephthalic acid. Esterifications are slow, equilibrium-controlled reactions and the products are separated by distillation. Since the Witten process makes dimethyl terephthalate (DMT), MeOH is released during the manufacture of PET and it must be captured and returned to the manufacturer of the DMT. All these extra energy-intensive steps can be eliminated with the use of the Co[BOND]Mn[BOND]Br catalyst. Similarly, the Eastman process generates an excess of acetic acid, which is avoided with the Co[BOND]Mn[BOND]Br catalyst.

Principle 4 – Design for Energy Efficiency and Simplicity The fact that the product separates itself as it is being produced avoids the problem usually associated with homogeneously catalyzed processes, i.e. the separation of the product from the solvent and catalyst. Also, the process is sufficiently selective that a significant amount of the solvent containing the catalyst (mother liquor) can be directly recycled back to the oxidation reactor without purification. Various schemes have been developed to recycle the remaining catalyst metals. The metals can be precipitated from acetic acid as their oxalate (29) or carbonate salts (30). The solid salts can be returned to the oxidation reactor since their dissolution in acetic acid and subsequent oxidation simply release CO2. Also, methods have been developed to separate the cobalt and manganese from the corrosion metals, such as iron and chromium (31). Since the Co[BOND]Mn[BOND]Br catalyst is much more active, much less catalyst needs to be used than for the Co catalyst alone; see Table 12.4 for typical catalyst concentrations. Finally, the capture of the energy released during the oxidation reaction is improved because the reactors can be operated at higher temperatures. For the cobalt catalyst the temperatures are 120–150 °C whereas for the Co[BOND]Mn[BOND]Br catalyst typical operating temperatures will range from 170 to 225 °C.

Principle 3 – Less Hazardous to the Environment The Co[BOND]Br and Co[BOND]Mn–Br catalysts generate less CO2 and toxic carbon monoxide than the Co catalyst (Table 12.4). Methods have been developed to destroy effectively the carbon monoxide and other organic species, in the vent gas streams. Also, the action of bacteria is often used in the plants to destroy all of the organic waste (acetic acid, terephthalic acid, intermediates and by-products). The resulting sludge can be dried and burned or spread on land (16, 32).

One key observation to explain the mechanism of the increased selectivity of the metal–bromide catalysts is the steady-state concentration of Co(III) in the reactor. In one example in Table 12.4, the concentration was reduced from 11% to <0.6% of the total cobalt in solution. This is readily observed when performing these oxidations in glass reactors. When the Co(II) acetate is initially dissolved with the p-xylene in acetic acid the color is light blue. After initiation of the reaction, the color is changed to deep green by the Co(III) that is formed. If one now adds an equimolar amount of sodium bromide to the cobalt in the flask, the color immediately reverts to light blue as nearly all of the Co(III) disappears. At the same time, the rate of reaction increases and the rates of CO2 and toluene generation decrease significantly (Table 12.4). The reduction in by-product formation is expected because they are the result of Co(III) decarboxylation and now much less Co(III) exists in solution. There is a rapid reaction of Co(III) with Br to generate a bromine(0) species (Scheme 12.8b), which effectively lowers the concentration of Co(III) in solution. The bromine(0) rapidly abstracts a hydrogen atom from p-xylene to generate the benzylic radical, which propagates the free radical chain.

Also known are the rates of reaction when a peroxy acid is added to a mixture of Co(II), Mn(II) and bromide in acetic acid (Scheme 12.8c). The reaction of Co(III) with Mn(II) is faster than that of Br, hence Mn(II) is oxidized before Br. The very fast reaction of Mn(II) with Co(III) lowers the concentration of Co(III) even further than in the Co[BOND]Br catalyst, resulting in an even more selective system. Addition of Mn(II) to a Co[BOND]Br catalyst does not significantly change its color. Since Mn(II) is nearly colorless whereas Mn(III) is intense brown, both the steady-state concentration of Co and Mn are in their +II state. Finally, the Co, Mn and Br can lower the steady-state concentration of the unselective OH radicals which react spontaneously and very rapidly with OH (Table 12.3), to form catalytically active species, i.e. Co(III), Mn(III) and Br(0).

As discussed above, the maximum temperature of a Co catalyst is limited because the decarboxylation of the carboxylic acids via Co(III) becomes the predominant reaction at temperatures >120 °C. The lower steady-state concentrations of Co(III) and Mn(III) in a Co[BOND]Mn[BOND]Br catalyst allow one to operate at higher temperatures. As a consequence, p-xylene can be completely oxidized to terephthalic acid in a reasonably short, industrially acceptable, residence time.

There are, however, four drawbacks associated with the Co[BOND]Mn[BOND]Br-catalyzed autoxidation of p-xylene, three of which are associated with the acetic acid solvent. (i) Significant quantities of acetic acid are oxidatively destroyed; the exact amounts are proprietary to the commercial manufacturers of terephthalic acid but, in 1995, about 12% of the acetic acid (6.8 × 105 tons) sold worldwide was used to replenish the terephthalic acid plants that use acetic acid as a solvent (33). (ii) Methyl bromide, a severe ozone-depleting chemical, is found in the vent gases; it is probably formed from the methyl radicals generated by acetic acid decarboxylation, Reaction 12.22 (Scheme 12.9), with the reduced Br(0) species in solution. (iii) Water is a product of the oxidation of p-xylene(2 mol per mole of terephthalic acid) and this water must be separated from the acetic acid by a energy-intensive distillation step. (iv) Even though the product precipitates from the reaction in >99% purity, it contains ∼0.5% of 4-CBA, which is a chain stopper in the esterification needed for PET manufacture; the concentration of 4-CBA has to be lowered by further purification either via hydrogenation in water or by further oxidation at higher temperatures which causes a significant increase in solvent oxidative degradation (6).

12.5 Potential Processes Using Water as a Solvent

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

At this point, we switch to discussing a possible future development in xylene oxidation, namely switching from acetic acid to H2O as the process solvent. These developments are not mere speculation; they based on results published in the open literature, particularly by Savage's group in Michigan and Poliakoff's group in Nottingham. Although xylene oxidation in high-temperature water is still at an early stage, there is a long history of commercial wet air oxidation and total oxidation in supercritical water has also been carried out commercially (58), albeit on a scale very modest compared with that of current terephthalic acid plants.

Three of the four problems described in the previous section are associated with the acetic acid solvent. Therefore, it is not surprising that attempts have been made in the past by terephthalic acid manufacturers to eliminate the acetic acid or to use water as a solvent. As early as 1961 there was a report by McIntyre and Ravens of ICI (34). This was followed by a series of reports by Hronec and Ilavsky in 1982–83 (35, 36) and from Amoco in 1990 (37).

In 1996, work began at Nottingham on continuous selective partial oxidation in supercritical H2O. Interest in high-temperature water was further stimulated by a paper by Holliday et al., who demonstrated that selective batch oxidation could be performed on a small scale in sub-critical water at 300–355 °C (38). This was followed by reports on continuous oxidation from the Nottingham team in collaboration with INVISTA (39-42) and from Savage's group in Michigan largely on batch reactions (43-46).

Partenheimer was introduced to this area by the 2002 report from Nottingham of the continuous oxidation of p-xylene using MnBr2 as a catalyst in supercritical water to produce terephthalic acid in high yield (>80%) with no detectable 4-CBA in the product (39). This is particularly exciting because there is a possibility of simultaneously eliminating all four of the problems discussed above! More specifically, if water were successfully employed as a solvent, one would have the following advancements in green chemistry.

Principles 1 and 8 – Prevent Waste and Generation of Hazardous Substances The elimination of the products of the oxidative combustion of acetic acid (CO – toxic; CO2 – greenhouse gas; and CH3Br – ozone depleting).

Principle 3 – Safer Solvents Acetic acid is susceptible to explosions whereas water is non-combustible.

Principle 4 – Design for Energy Efficiency and Simplicity There would no need to separate water from the acetic acid. With no 4-CBA in the product, downstream purification steps, such as hydrogenation, may become unnecessary. If hydrogenation in water is necessary, then there is solvent compatibility and simplicity, since oxidation, hydrogenation and crystallization can be performed in the same solvent. Also, there is the advantage of operating at higher temperatures (300–400 °C) than either the Co- or Co[BOND]Mn[BOND]Br-catalyzed processes (∼150 and ∼190 °C, respectively), so that energy recovery would be more efficient.

Phenomenologically, the mechanism of terephthalic acid in sub- and supercritical water appears to be much the same as that in acetic acid, namely a catalyst-modified free radical chain mechanism. Similarities between the reactions in the two solvents include the following:

  • The same product sequence as shown in Scheme 12.4 is reported in sub- and supercritical water (39, 45, 46).

  • At these high temperatures, one can observe the uncatalyzed oxidation of p-xylene with the expected products (i.e. large amounts of toluene and cresols from the thermal decomposition of the hydroperoxides; see Scheme 12.6); the cresols will prevent high yields of the aromatic acids. The uncatalyzed oxidation of p-xylene at 240–500 °C and 200–300 bar gives a maximum yield of only 22% terephthalic acid and up to a 64% yield of toluene (45, 46).

  • Benzylic alcohols are always found in low yields (∼1–5%) in metal-catalyzed oxidation in acetic acid because the catalyst metals not only operate as redox catalysts but also as Lewis acids to dehydrate the benzyl hydroperoxides to aldehydes very rapidly, thereby by-passing the formation of the benzylic alcohol (49, 50):

    mathml alt image(12.2)

    Hence one would expect relatively high amounts of the benzylic alcohols to be formed in uncatalyzed oxidation of p-xylene in water. This indeed has been reported where p-xylene was found to give a ∼40% yield at 380 °C in water (46).

  • Addition of the appropriate catalyst metal and bromide enhances the yield of aromatic acid and increasing the catalyst concentration increases the yield further, all of which suggests that oxidation in water is operating with the same mechanisms as in acetic acid (4, 38, 44, 45).

It is well established that the properties of water change greatly with increasing temperature. Gases and organic substrates are very soluble in supercritical water since its physical properties become similar to those acetone or anhydrous acetic acid. For example, the dielectric constant of water decreases from 80 at ambient temperature to 4 at the critical point. By comparison, at room temperature, the dielectric constant of acetone and anhydrous acetic acid are 1.3 and 6.2, respectively.

The structure of MnBr2 in water has been determined at 25, 325 and 400 °C using EXAFS and XANES (51, 52). As can be seen from Scheme 12.10, very little bromide is in the coordination sphere at room temperature, whereas at 325 °C, a mixture of tetrahedral and octahedral mono- and dibromo compounds exist with most of the bromide directly bonded to Mn. At 400 °C in supercritical water, only the tetrahedral [MnIIBr2(H2O)2] is detected. This is consistent with the decreasing dielectric constant of the solvent since ionic species are becoming less favored. This is supported by the fact that the structure of CoBr2 in acetone, [CoBr2 (acetone)2] (53), is essentially the same as that of MnBr2 in supercritical water [MnBr2(H2O)2].

original image

Scheme 12.10. Structure of MnBr2 in water as a function of temperature as determined by Br and Mn EXAFS and XANES measurements (51, 52).

With water as a solvent, the catalyst is significantly less active at 200 °C than at 400 °C. At 200 °C, a 37 times higher concentration of the Co[BOND]Mn[BOND]Br catalyst was needed to obtain a yield of terephthalic acid similar to that obtained in acetic acid (37). Much lower catalyst concentrations are required in water at 400 °C. The dielectric constant of water changes from 37 at 200 °C to 4 at 400 °C; hence there appears to be a correlation between the degree of metal–bromide bonds and catalytic activity. A similar relationship exists in acetic acid–water solutions between the degree of metal–bromide bonds with dielectric constant and catalytic reactivity (Table 12.5). A simple model has been devised to rationalize these relationships (Scheme 12.11) (52).

Table 12.5. Relationship of catalytic activity of the Co[BOND]Mn[BOND]Br catalyst to the amount of metal–bromide bonds and dielectric constant in acetic acid–water mixtures. Data from (25, 56, 57)
SolventDielectric constantMetal as metal–bromide species (%)aRate of p-toluic acid with Co–Mn–Br catalyst (cm3 O2 min−1)
  1. a

    The percentage of the Co and Mn metals in solution that have a metal–bromide bond as determined using a bromide-selective electrode.

10% H2O in HOAc1731.2
5% H2O in HOAc13142.1
1% H2O in HOAc7.667
original image

Scheme 12.11. A model to rationalize the activity of metal–bromide catalysts in acetic acid and water.

In autoxidation, the energetically difficult step is the initiation step where the benzylic radical CH3PhCH2[FREE RADICAL] is generated from p-xylene,. When the dielectric constant is high, the value of the equilibrium constant in Scheme 12.11 is small. The active catalyst is then generated via Reactions 12.25–12.28. Thus, Mn(II) is initially oxidized to Mn(III), ligand exchange occurs and bromide is incorporated into the Mn(III) coordination sphere. This coordination is followed by the intramolecular electron transfer, Reaction 12.27, to generate the active catalyst species which in turn generates the benzylic radical, Reaction 12.28. By contrast, when the dielectric constant is low the equilibrium constant is high and most of the manganese exists with an Mn(II)[BOND]Br bond. The active catalytic species is then generated via Reactions 12.30 and 12.27. There are two reasons for expecting the route via Reactions 12.25–12.27 to be slower than that via Reactions 12.30 + 12.27 because three reactions are needed to generate the active species via Reactions 12.25–12.27, whereas only two reactions are needed via Reactions 12.30 and 12.27 and the rate of incorporation of the bromide anion into the coordination sphere will be much slower for Mn(III) than that for Mn(II) (54). The rate of generation of the active species is expected to be faster when the dielectric constant is low, in agreement with the experimental data in both water and acetic acid.

So far, discussion in this chapter has been entirely restricted to the oxidation of p-xylene. This is because current processes involve the upstream separation of the refinery C8 aromatic stream, consisting of a mixture of the three isomeric xylenes with ethylbenzene, into its constituent compounds. Such separation is energy intensive because the physical properties of the C8 components are rather similar. After separation, the isomers are oxidized in separate processes (Figure 12.1), because their rates of oxidation are substantially different. An interesting aspect of oxidation in supercritical water is that the rates of oxidation of the isomeric xylenes are very similar (41) and it is possible to co-oxidize them in a way which cannot easily be done in acetic acid (Table 12.6 and Figure 12.2). This opens up the exciting possibility of directly co-oxidizing the three xylenes without separating them; this is potentially important from a green chemistry perspective because the corresponding carboxylic acids have more widely differing properties and are more easily separated than the xylene precursors.

original image

Figure 12.1. Schematic of current processes for separating the refinery C8 aromatic stream into its components and oxidizing the different compounds to their corresponding acids. Reproduced with permission from (41), © The Royal Society of Chemistry.

Table 12.6. Yieldsa observed in the continuous Mn–Br oxidation of mixed xylenes in supercritical H2O. Adapted from (41)
o-:m-:p-Xylenea (w/Wt)Di-CA: total yield (%)o-Di-CA (%)bm-Di-CA(%)bp-Di-CA(%)b
  1. a

    All calculated by HPLC (di-CA = dicarboxylic acid).

  2. b

    Feed composition of the xylene mixture by weight. All the reactions were carried out under the same conditions; see (41) for details.

33.3 : 33.3 : 33.359476269
66 : 17 : 1760586567
17 : 66 : 1752345847
17 : 17 : 6647395048
100 : 0 : 05252--
0 : 100 : 066-66-
0 : 0 : 10061--61
original image

Figure 12.2. Schematic of a possible simultaneous oxidation process in supercritical H2O for a mixture of all three isomeric xylenes to their corresponding acids. Reproduced with permission from (41), © The Royal Society of Chemistry.

12.6 Summary and Final Comments

  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References

We have described how the preparation of terephthalic acid has been greened from the highly wasteful stoichiometric oxidation, to the Co(II) acetate catalyst in acetic acid which a renewable source of oxidant, O2, can be used and which is able to drive the oxidation of p-xylene down selective pathways to terephthalic acid. Further greening was seen with the simple addition of Mn and Br to the Co catalyst to give a great boost in activity and selectivity, so that >95% yields could be obtained with >99% purity of the isolated terephthalic acid. These advances have enabled major process simplification to be implemented. In the future, a change in solvent from acetic acid to water could eliminate the wasteful oxidation of the acetic acid and possibly reduce the number of processing steps.

The use of water as an oxidation solvent is in infancy and many questions still need to be addressed. For example, the specifications for the terephthalic acid to be used in PET are very stringent. The terephthalic acid must be very white and not contain trace organic and inorganic impurities that may impart undesirable properties to the PET during its preparation. ‘Hot’ water can be very corrosive, so materials of construction will be important. Elimination of the oxidative destruction of acetic acid will eliminate one of the primary variable costs of the current process. However, one does not yet know how much of the p-xylene, its intermediates and terephthalic acid product will undergo this same oxidative destruction. One might expect that the ‘burn’ of p-xylene oxidation in water might be higher than in acetic acid because, for example, the hydroxyl radicals, which were destroying the acetic acid, will now attack the p-xylene, thereby increasing its rate of destruction. The use of water will have the challenging task of competing against a very mature terephthalic acid manufacturing industry which benefits from empirical learnings gathered over the last 50 years and also from good scientific and technological understanding.


  1. Top of page
  2. Introduction
  3. Methods of Making Terephthalic Acid Using Stoichiometric Reagents
  4. Methods for Preparing Terephthalic Acid Using Cobalt Acetate and Dioxygen in Acetic Acid
  5. Adding Bromide to Improve Terephthalic Acid Production Using Cobalt and Manganese Acetates in Acetic Acid
  6. Potential Processes Using Water as a Solvent
  7. Summary and Final Comments
  8. References
  • 1
    Cornils, B. and Herrmann, W.A. (2002) Synthesis of dimethyl terephthalate/terephthalic acid and poly (ethylene terephthalate), in Applied Homogeneous Catalysis with Organometallic Compounds, 2nd edn (ed. Schiraldi, D.A.), Wiley-VCH Verlag GmbH, Weinheim, pp. 544552.
  • 2
    Raghavendrachar, P. and Ramachandran, S. (1992) Liquid-phase catalytic oxidation of p-xylene. Industrial & Engineering Chemistry Research, 31, 453462.
  • 3
    (a) Anastas, P.T. and Warner, J.C. (1998) Green Chemistry: Theory and Practice, Oxford University Press, New York, pp. 30; (b) Anastas, P.T. and Zimmerman, J.B. (2003) Environmental Science & Technology, 37, 94.
  • 4
    Partenheimer, W. (1995) The methodology and scope of metal/bromide autoxidation of hydrocarbons. Catalysis Today, 23, 69158.
  • 5
    Cincotti, A., Orru, R. and Cao, G. (1999) Kinetic and related engineering aspects of catalytic liquid-phase oxidation of p-xylene to terephthalic acid. Catalysis Today, 52, 331347.
  • 6
    Suresh, A.K., Sharma, M.M. and Sridhar, T. (2000) Engineering aspects of industrial liquid-phase air oxidation of hydrocarbons. Industrial & Engineering Chemistry Research, 39, 39583997.
  • 7
    Carra, S. and Santacesaria, E. (1980) Engineering aspects of gas–liquid catalytic reactions. Catalysis Reviews: Science and Engineering, 22, 75.
  • 8
    Loder, D.J. (1938) (DuPont), US Patent 2 245 528.
  • 9
    Burney, D.E., Weisemann, G.H. and Fragen, N. (1959) Now – dibasic acids by direct oxidation. Petroleum Refiner, 38, 186.
  • 10
    Landau, R. and Saffer, A. (1968) Development of the MC process. Chemical Engineering Progress, 64, 20.
  • 11
    Hudlicky, M. (1990) Oxidations in Organic Chemistry, ACS Monograph No. 186, American Chemical Society, Washington, DC.
  • 12
    Sheldon, R.A. (1992) Organic synthesis – past, present and future. Chemistry & Industry (London), 903.
  • 13
    Sheldon, R.A. (1994) Consider the environmental quotient. Chemotherapy, 38.
  • 14
    Partenheimer, W. (2004) The complex synergy of water in the metal/bromide autoxidation of hydrocarbons caused by benzylic bromide formation. Advanced Synthesis and Catalysis, 246, 297306.
  • 15
    Partenheimer, W. (2005) The complex synergy of water in metal/bromide autoxidations. Part II. Effect of water and catalyst on the aerobic oxidation of benzaldehydes and the effect of water on the elementary pathways. Advanced Synthesis and Catalysis, 347, 580590.
  • 16
    Sheehan, R.J. (1995) Terephthalic acid, dimethylterephthalate and isophthalic acid, in Ullmann's Encyclopedia of Industrial Chemistry, (eds. Ehers, B., Hawkins, S. and Russey, W.), Wiley-VCH, Weinheim. Vol A26, 193226.
  • 17
    Partenheimer, W. (1990) A chemical model for the Amoco ‘MC’ oxygenation process to produce terephthalic acid, in Catalysis of Organic Reactions, (ed. Blackburn, D.W.), Marcel Dekker, New York, Chapter 20.
  • 18
    Partenheimer, W. (1991) Characterization of the reaction of Co(II) acetate, dioxygen and acetic acid and its significance in autoxidation reactions. Journal of Molecular Catalysis, 67, 3546.
  • 19
    Buxton, G.V., Greenstock, C.L., Helman, W.P. and Ross, A.B. (1988) Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals ([FREE RADICAL]OH/[FREE RADICAL]O) in aqueous solution. Journal of Physical and Chemical Reference Data, 17, 513886.
  • 20
    Partenheimer, W., and Gipe, R.K. (1993) Nature of the Co–Mn–Br catalyst in the methylaromatic compounds process: kinetic and thermodynamic studies, in Catalytic Selective Oxidation, (eds. Oyama, S.T. and Hightower, J.W.), American Chemical Society, Washington, DC, 8188.
  • 21
    Partenheimer, W. (2003) The effect of zirconium in metal/bromide catalysts on the autoxidation of p-xylene. Part I. Activation and changes in benzaldehyde intermediate formation. Journal of Molecular Catalysis, 206, 105119.
  • 22
    Partenheimer, W. (2004) The unusual characteristics of the aerobic oxidation of 3,4-dimethoxytoluene with metal/bromide catalysts. Advanced Synthesis and Catalysis, 347, 14951500.
  • 23
    Brill, W.F. (1960) Terephthalic acid by single stage oxidation. Industrial & Engineering Chemistry, 52, 837.
  • 24
    Lande, S.S. and Kochi, J.K. (1968) Formation and oxidation of alkyl radicals by Co(III) complexes. Journal of the American Chemical Society, 90, 5196.
  • 25
    Partenheimer, W. (2001) The structure of metal/bromide catalysts in acetic acid/water mixtures and its significance in autoxidation. Journal of Molecular Catalysis, 174, 2933.
  • 26
    Nakaoka, K., Miyama, Y., Matsuhisa, S. and Wakamatsu, S. (1973) Preparation of terephthalic acid using paradehyde promoter. Industrial & Engineering Chemistry Product Research and Development, 12, 150.
  • 27
    Hull, D.C. (1957) (Eastman Kodak), US Patent 2 673 217.
  • 28
    Saffer, A. and Barker, R.S. (1958) (Mid-Century), US Patent 2 833 816.
  • 29
    Hobbs, C.C., Trebellas, J.C. and Huguet, J.L. (1974) (Celanese), US Patent 3 840 469; Feld, M. (1984) (Dynamit Nobel), US Patent 4 490 298; Brown, A.P. and Hundley, J.G. (1986) (Amoco), US Patent 4 487 355; Nowicki, N.R. and Thomka, D.E. (1988) (Amoco), US Patent 4 769 488; Holzhauer, J.K., Schammel, W.P. and Nowicki, N. R. (1988) (Amoco), US Patent 4 786 752; Lowry, J.D. Nowicki, N.R. and Partenheimer, W. (1989) (Amoco), US Patent 4 816 601.
  • 30
    Jeffery, I., and Whiston, K., (1998) (INVISTA), Patent WO98/08605, ICI; Trevillan, A.E. (1972) (Atlantic Richfield), US Patent 3,673,154.
  • 31
    (a) Trevillan, A.E. (1972) (Atlantic Richfield), US Patent 3 673 154; (b) Partenheimer, W. (1980) (Amoco), US Patent 4 228 091; (c) Partenheimer, W. (1980) (Amoco), US Patent 4 242 128.
  • 32
    Shelley, S. (1991) Chemical Engineering, 98, 9093.
  • 33
    Weissermel, K. and Arpe, H.J. (1997) Industrial Organic Chemistry, 3rd edn, Verlag Chemie, Weinheim, p. 178.
  • 34
    McIntyre, J.E. and Ravens, D.A.S. (1961) The oxidation of alkylaromatic compounds in aqueus hydrogen bromide. Journal of the Chemical Society, 4082.
  • 35
    Hronec, M. and Ilavsky, J. (1982) Oxidation of polyalkylaromatic hydrocarbons. XII. Technological aspects of p-xylene oxidation to terephthalic acid in water. Industrial & Engineering Chemistry Product Research and Development, 21, 455.
  • 36
    Hronec, M. and Ilavsky, J. (1983) Effect of pyridine on cobalt bromide catalyzed oxidation of p-xylene in various solvents. Oxidation Communications, 3, 303.
  • 37
    Nowicki, N.R. and Lowry, J.D. (1990) (Amoco), US Patent 4 892 970.
  • 38
    Holliday, R.L., Jong, B.W.M. and Kolis, J.W. (1998) Organic synthesis in subcritical water. Oxidation of alkylaromatics, Journal of Supercritical Fluids, 12 1255.
  • 39
    Hamley, P.A., Ilkenhans, T., Webster, J.M., Garcia-Vedugo, E., Vernardou, E., Clarke, M.J., Auerbach, R., Thomas, W.B., Whiston, K. and Poliakoff, M. (2002) Selective partial oxidation in supercritical water: the continuous generation of terephthalic acid from para-xylene in high yield. Green Chemistry, 4, 235238.
  • 40
    Garcia-Verdugo, E., Vernardou, E., Thomas, W.B., Whiston, K., Partenheimer, W., Hamley, P.A. and Poliakoff, M. (2004) Is it possible to achieve highly selective oxidations in supercritical water? Aerobic oxidation of methylaromatic compounds. Advanced Synthesis and Catalysis, 346, 307.
  • 41
    Garcia-Verdugo, E., Fraga-Dubreuil, J., Hamley, P.A., Thomas, W.B., Whiston, K. and Poliakoff, M. (2005) Simultaneous continuous partial oxidation of mixed xylenes in supercritical water. Green Chemistry, 7, 294300.
  • 42
    Fraga-Dubreuil, J. and Poliakoff, M. (2006) Organic reactions in high-temperature and supercritical water. Pure and Applied Chemistry, 78, 17911982.
  • 43
    Dunn, J.B. and Savage, P.E. (2002) Terephthalic acid synthesis in high-temperature liquid water. Industrial & Engineering Chemistry Research, 41, 4460.
  • 44
    Dunn, J.B. Urquhart, D.I. and Savage, P.E. (2002) Terephthalic acid synthesis in supercritical water. Advanced Synthesis and Catalysis, 341, 385392.
  • 45
    Dunn, J.R. and Savage, P.E. (2005) High-temperature liquid water: a viable medium for terephthalic acid synthesis. Environmental Science & Technology, 39, 5427.
  • 46
    Dunn, J., Savage, P. and Yu, J. (2006) Combustion Science and Technology, Recent advances in catalytic oxidation in supercritical water, 178, 443465.
  • 47
    Kim, Y.-L., Chung, S.J., Kim, J.-D., Lim, J.S., Lee, Y.-W. and Yi, S.-C. (2002) Uncatalyzed partial oxidation of p-xylene and supercritical water. Reaction Kinetics and Catalysis Letters, 77, 3543.
  • 48
    Kim, Y.L., Kim, J.D., Lim, J.S., Lee, Y.W. and Yi, S.C. (2002) Reaction pathway and kinetics for uncatalyzed partial oxidation of p-xylene in sub- and supercritical water. Industrial & Engineering Chemistry Research, 71, 55765583.
  • 49
    Partenheimer, W. (2003) The effect of zirconium in metal/bromide catalysts on the autoxidation of p-xylene. Part I. Activation and changes in benzaldehyde intermediate formation Journal of Molecular Catalysis, 206, 105119.
  • 50
    Partenheimer, W. (2003) The effect of zirconium in metal/bromide catalysts on the autoxidation of p-xylene. Part II. The effect in Co/metal/Br catalysts on manganese(IV) formation and precipitation with pyromellitic acid. Journal of Molecular Catalysis, 206, 131144.
  • 51
    Chen, Y., Fulton, J.L. and Partenheimer, W. (2005) XAFS study of hydration and ion pairing in ambient aqueous MnBr2 solutions. Journal of Solution Chemistry, 4, 993.
  • 52
    Chen, Y., Fulton, J.L. and Partenheimer, W. (2005) The structure of the homogeneous oxidation catalyst, Mn(II)[BOND]Br([BOND]I)x, in supercritical water: An X-ray absorption fine structure study. Journal of the American Chemical Society, 127, 14085.
  • 53
    Sawada, K., Onoda, T. and Suzuki, T. (1983) Bromo-complexes of Co(II) in acetone. Journal of The Chemical Society-Dalton Transactions, 1565.
  • 54
    Gray, H.B. and Langford, C.H. (1968) Chemical & Engineering News, 46, 68.
  • 55
    Jones, G.H. (1979) p-Xylene autoxidation studies. Oxidation Co(II) and manganese(II) acetates by peracids. Journal of the Chemical Society. Chemical Communications, 536.
  • 56
    Popov, A.I. (1970) in The Chemistry of Nonaqueous Solvents III (ed. Lagowski, J.J.), Academic Press, New York. p. 214.
  • 57
    Sawada, K. and Tanaka, M. (1973) Journal of Inorganic & Nuclear Chemistry, 35, 2455.
  • 58
    Bermejo, M.D. and Cocero, M.J. (2006) Supercritical water oxidation: a technical review. AIChE Journal, 52, 39333951.