Oxidation

J. HENRIQUE TELES, BASF SE, Ludwigshafen, Germany

IVE HERMANS, University of Wisconsin, Madison, USA
GERHARD FRANZ, Consultant, Marl, Germany
ROGER A. SHELDON, Technische Universiteit Delft, Delft, The Netherlands

1. Introduction . . . . . . . . . . . . . . . . 1 5. Homolytic Oxidation. . . . . . . . . . 51

2. Overview of Industrial Oxidation 5.1. Historical Development; Reaction
Processes . . . . . . . . . . . . . . . . . . 2 Mechanism; Radical Reactivity . . . 51
2.1. Inorganic Chemicals Produced by 5.2. Kinetics of Autoxidation . . . . . . . 58
Oxidation . . . . . . . . . . . . . . . . . 5 5.3. Differences between Liquid- and
2.1.1. Sulfur Compounds . . . . . . . . . . . 5 Gas-Phase Autoxidations . . . . . . . 61
2.1.2. Carbon Derivatives. . . . . . . . . . . 7 5.4. Homolytic Oxidation in the Liquid
2.1.3. Nitrogen Compounds . . . . . . . . . 7 Phase . . . . . . . . . . . . . . . . . . . . 62
2.1.4. Chlorine Derivatives . . . . . . . . . . 9 5.4.1. Secondary Reactions of Radicals,
2.2. Organic Chemicals Produced by Peroxides, and other Intermediates 62
Oxidation . . . . . . . . . . . . . . . . . 10 5.4.2. Metal Catalysis in Homolytic
2.2.1. Olefins and Alkynes . . . . . . . . . . 10 Liquid-Phase Oxidation . . . . . . . 68
2.2.2. Epoxides . . . . . . . . . . . . . . . . . . 11 5.4.3. Inhibitors of Homolytic Liquid-
2.2.3. Hydroperoxides . . . . . . . . . . . . . 17 Phase Oxidation . . . . . . . . . . . . . 71
2.2.4. Alcohols . . . . . . . . . . . . . . . . . . 20 5.4.4. Special Cases of Homolytic
2.2.5. Aldehydes . . . . . . . . . . . . . . . . . 21 Oxidation . . . . . . . . . . . . . . . . . 72
2.2.6. Ketones . . . . . . . . . . . . . . . . . . . 24 5.4.5. Process Technology of Homolytic
2.2.7. Phenols and Derivatives . . . . . . . 26 Liquid-Phase Oxidation . . . . . . . 73
2.2.8. Carboxylic Acids, Saturated . . . . 29 5.5. Homolytic Gas-Phase Oxidation. . 77
2.2.9. Carboxylic Acids and Anhydrides, 5.6. Gas Explosions and Safety Data. . 78
Unsaturated. . . . . . . . . . . . . . . . 35 6. Heterolytic Oxidation . . . . . . . . . 83
2.2.10. Carboxylic Acids and Anhydrides, 6.1. Heterolytic Oxidation in the Liquid
Aromatic . . . . . . . . . . . . . . . . . . 37 Phase . . . . . . . . . . . . . . . . . . . . 83
2.2.11. Esters, Lactones, and Carbonates 41 6.2. Heterogeneously Catalyzed Gas-
2.2.12. Nitriles . . . . . . . . . . . . . . . . . . . 43 Phase Oxidation . . . . . . . . . . . . . 88
2.2.13. Compounds Containing NO 6.2.1. Oxidation Catalysts . . . . . . . . . . 88
Bonds . . . . . . . . . . . . . . . . . . . . 44 6.2.2. Process Technology of
2.2.14. Chlorinated Compounds . . . . . . . 47 Heterogeneously Catalyzed
2.2.15. Organosulfur Compounds . . . . . . 49 Oxidations . . . . . . . . . . . . . . . . . 93
3. Thermodynamics; Stability References . . . . . . . . . . . . . . . . . 96
of Hydrocarbons . . . . . . . . . . . . . 50
4. Reaction Types . . . . . . . . . . . . . . 50

1. Introduction used to prepare it makes use of a typical oxidant

such as O2. Processes like oxidative chlorina-
The chemical term oxidation is used here both tions with HCl/O2 are thus included here, but
in the sense of the take-up of oxygen by organic simple chlorinations are not. Biological oxida-
and inorganic materials and also in cases when tions, although of considerable technical impor-
the product contains no oxygen, but the process tance, are not included in this article. The same

# 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

10.1002/14356007.a18_261.pub2
2 Oxidation

is true for the combustion of organic material to not only the nature of the oxidant, but also the
obtain energy, and undesired oxidations like the workup of the product, the recycling of solvents
corrosion of metals and the aging of polymers. and catalysts, and the disposal of byproducts
All of these are outside the scope of this article. and wastewater are considered, has very seldom
Atmospheric oxygen is by far the most been done in an academic context. In an indus-
important and also the cheapest oxidizing trial context, all these aspects are taken into
agent, since all others, for example, HNO3, consideration, because they all have an impact
H2SO4 and SO3/H2SO4 (oleum), Cl2, MnO2, on the cost of production. Also, although this
CrO3, KMnO4, as well as H2O2 and the organic may sound politically incorrect, the economics
hydroperoxides, must first be produced by of a process are a good indicator of its sustain-
using oxygen or electric current. However, ability and, especially for low-volume special-
oxidations with molecular oxygen usually ties or pharmaceuticals, old-fashioned oxidants
require much more elaborate catalyst and pro- may still be the best choice. For commodities,
cess development to achieve the high selectivi- in which the raw materials often make up 80%
ties required in industrial oxidation processes of the production costs, it is no surprise that the
than, for example, those with Cl2 or HNO3. cheapest oxidant (air) has become the most
They generally also require a much higher important one.
investment, which only becomes economically
worthwhile upon exceeding a certain minimum
capacity for a production plant. This currently 2. Overview of Industrial Oxidation
lies between 104 and 105 t/a, depending on the Processes
product, although exceptions are not rare.
Before World War II, organic materials were In this chapter the main classes of industrially
mostly oxidized with inorganic oxidizing relevant products which are obtained by an
agents and only rarely directly with O2. The oxidation reaction are briefly reviewed, but
most important products of industrial organic older reviews can still be recommended, at least
chemistry at that time were pharmaceuticals from a historical point of view [2, 3]. The
and dyes, for which only relatively small quan- individual processes and reactions are not dis-
tities of organic intermediates were required. cussed in detail, but reference is made to appro-
Exceptions included phthalic anhydride, for priate comprehensive reviews, and current
which the heterogeneously catalyzed gas-phase trends and challenges are pointed out. For
oxidation of naphthalene was developed to an most of the reactions, specialized keywords
industrial level in Germany in 1916 (BASF) and are available and appropriate reference is
in the USA (Bureau of Chemistry), and maleic made to these.
anhydride, which was first obtained in 1928 by In some cases an arbitrary line had to be
the Barret Co. in an analogous process involv- drawn as to what is still considered an oxidation
ing the oxidation of benzene. or not. For instance, the reaction of tert-amyl
The age of plastics, with the mass produc- alcohol with H2O2 to produce tert-amyl hydro-
tion of polyamides, polyesters, polyurethanes peroxide, the reaction of acetic acid with H2O2
etc., and the precursors necessary for them, to produce equilibrium peracetic acid solution,
only began after World War II. and the reaction of cyclic ketones with H2O2 to
Since about the mid-1970s, ecological argu- produce perketals are all industrially important
ments have become more frequent, especially and could at least formally also be seen as
in academic literature, not only to explain the oxidation reactions. However, as there is no
decreasing importance of inorganic oxidizing change in the oxidation state of the organic
agents, but also to promote their substitution (or residue, these reactions are better classified as
even outright ban) by allegedly greener or nucleophilic substitutions or additions and are
more sustainable oxidants [1]. However, this not considered here. Dehydrogenations, for-
discussion has often been limited to the nature mally also oxidations, are also not covered here.
and cost of the oxidant itself, and very often it All known industrial processes which use
has even been misused to simply justify oxidation and have a capacity of at least 1000 t/a
research. A truly holistic discussion, in which are included in this overview. Although some
 Oxidation 3

agrochemicals also include an oxidation step in in the USA varied from year to year between 200
their multistep synthesis, and some of these can t in 2010 and 2200 t in 1994 [4].
indeed have capacities of well over 1000 t/a, The pharmaceutical industry also uses a
these are not included here. The main reason is plethora of oxidation reactions, but these are
because the production amounts of an agro- mostly scaled-up laboratory methods. This sub-
chemical can vary widely from year to year ject has been thoroughly reviewed [5, 6].
and capacities or synthesis methods are usually An overview of all organic products indus-
not openly available. For instance, in the case of trially obtained by oxidation are summarized in
Propargite, a widely used acaricide, for which Tables 1a1d. The products are arbitrarily
cyclohexene oxide, made by an oxidation divided into commodities (Table 1a, products
reaction, is a building block, the amounts used with a capacity of more than 106 t/a), large

Table 1a. Commodity organic chemicals produced by oxidation (capacity >1
106 t/a)

Product or intermediate Capacity, 106 t/a Oxidanta Phaseb Catalystc

Terephthalic acidd 60 O2 liq homo

Ethylene dichloride 30 O2 gas hetero
Ethylene oxide 26 O2 gas hetero
Formaldehyde 20 O2 gas hetero
Cumene hydroperoxidee 17 O2 liq none
Propylene oxidef 10 Cl2 liq none
ROOH liq homo/hetero
H2O2 liq hetero
Acetylene 7 O2 gas none
Acrylonitrileg 7 O2 gas hetero
Ethylbenzene hydroperoxideh 6 O2 liq none
Acrylic acid 6 O2 gas hetero
Vinyl acetatei 5.5 O2 gas hetero
Hydrogen peroxide 5.5 O2 liq none
Phthalic anhydride 5.5 O2 gas hetero
Acrolein 5 O2 gas hetero
Cyclohexanolj 4.5 O2 liq homo
Adipic acid 4 HNO3 liq homo
Methyl nitrite 2.5 O2 liq/gas none
Epichlorohydrin 2.3 Cl2 liq none
tert-Butylhydroperoxidek 2 O2 liq none
Cyclohexanonej 2 O2 liq homo
Hydrogen cyanidel 2 O2 gas hetero
Acetaldehyde 2 O2 liq homo
gasm hetero
Maleic anhydride 2 O2 gas hetero
Methacrolein 1.2 O2 gas hetero
Methacrylic acidn 1.1 O2 gas hetero
Isophthalic acid 1 O2 liq homo
a
O2 stands for both pure oxygen and air.
b
liq: reaction in liquid phase; gas: reaction in gas phase.
c
homo: homogeneous catalyst, hetero: heterogeneous catalyst.
d
Includes ca. 3
106 t/a of dimethyl terephthalate.
e
First step in the Hock process for phenol and in the CUPO process for propylene oxide.
f
The split among the technologies is approximately 4.3
103 t/a for chlorohydrin, 4.8
103 t/a for cooxidation processes and 0.9
103 t/a for
hydrogen peroxide based processes.
g
200
103 t/a are based on propane and the remaining is based on propene.
h
First step in the SMPO process to propylene oxide.
i
The total capacity for vinyl acetate is around 7
106 t/a, but only 5.5
106 t/a are produced by oxidative acetoxylation of ethylene.
j
Combined capacity of cyclohexanol and cyclohexanone is around 7
106 t/a.
k
First step of the MTBE/PO process for propylene oxide.
l
Only oxidative processes. The total world capacity is ca. 2.5
106 t/a.
m
The gas-phase oxidation of ethanol has only a negligible contribution to the overall acetaldehyde capacity.
n
This is only the capacity share of methacrylic acid and methyl methacrylate that includes an oxidation step. The total capacity is around
3.4
106 t/a.
4 Oxidation

Table 1b. Intermediates produced by oxidation (capacity between 100 thousand and 1
106 t/a)

Product or intermediate Capacity, 103 t/a Oxidanta Phaseb Catalystc

Acetic acidd
Acetaldehyde oxidation 350 O2 liq homo
Ethane oxidation 30 O2 gas hetero
Ethylene oxidation 170 O2 gas hetero
Allyl acetate 600 O2 gas hetero
Cyclohexyl hydroperoxidee 500 O2 liq none/homo
Propionic acid and other non-a-branched acids 470 O2 liq none
10 O3 liq none
Benzoic acid 400 O2 liq homo
Butadiene 300 O2 gas hetero
Oxalic acidf 300 HNO3 liq homo
Glyoxal 250 HNO3 liq none
O2 gas hetero
2-Ethylhexanoic acid and other a-branched acids 225 O2 liq none
Alfols 200 O2 liq none
Poly(phenylene ether) 200 O2 liq homo
Epoxidized soybean and linseed oils 150 H2O2 liq homo
Trimellitic acid 150 O2 liq homo
Dimethyl carbonateg 150 O2 liq homo
But-2-ene-1,4-diol diacetateh 100 O2 liq heter
Cymene hydroperoxides 100 O2 liq none
Diisopropylbenzene dihydroperoxidei 100 O2 liq none
Cyclododecanol 100 O2 liq homo
Methyl methacrylatej 100 O2 liq hetero
a
O2 stands for both pure oxygen and air.
b
liq: reaction in liquid phase; gas: reaction in gas phase.
c
homo: homogeneous catalyst, heter: heterogeneous catalyst.
d
Only oxidation processes. Total world capacity for acetic acid is around 15
106 t/a.
e
First step in the KA-oil (cyclohexanone/cyclohexanol) process.
f
Only oxidation processes. Total world capacity for oxalic acid is around 600
103 t/a.
g
The capacity given is just for plants using the Eni oxidative carbonylation technology.
h
Intermediate in the Mitsubishi BDO/THF process.
i
Combined capacity for both meta and para isomers. First step in the Hock process for hydroquinone/resorcinol.
j
Asahi process for oxidative esterification of methacrolein. Total world capacity for methyl methacrylate is ca. 4
106 t/a.

intermediates (Table 1b, products with a less than 2% are intermediates, less than 0.5%
capacity between 105 and 106 t/a), specialities are specialties, and less than 0.1% are small-
(Table 1c, products with a capacity of scale specialties. Overall, almost 93% of the
(20100)
103 t/a), and small specialties capacity uses air or O2 as the oxidant, which is
(Table 1d, products with a capacity of less easy to understand because O2 is the cheapest
than 20
103 t/a). In addition to the estimated available oxidant. The next most important
capacities, which are mostly obtained from oxidants are HNO3 and nitrogen oxides (3.6%
company intelligence, the tables also include of the total capacity), Cl2 (2.8% of the total
data on the oxidant used, on the phase in which capacity), alkyl hydroperoxide (2% of the total
the reaction takes place (liquid or gas), and the capacity), and H2O2 (0.6% of the total capacity).
type of catalyst used (homogeneous, heteroge- Approximately equal amounts of organic
neous, or no catalyst at all). chemicals are produced by oxidation processes
The data in Tables 1a1d already make some in the liquid phase and in the gas phase. This
interesting conclusions possible. The total might at first sight seem surprising, but quite a
capacity for organic chemicals produced by few large commodities are produced by oxida-
an oxidation process is close to 250
106 t/a tion in the liquid phase, and above all the largest
or ca. 35 kg per human per year. Of this amount, one, terephthalic acid, uses one such process.
almost 98% of the capacity are commodities, Smaller products tend to use liquid-phase
 Oxidation 5

Table 1c. Specialties produced by oxidation (capacity of (20100)
103 t/a)

Product or intermediate Capacity, 103 t/a Oxidanta Phaseb Catalystc

Dimethyl sulfoxide 95 NO2/O2 liq none

Secondary alkane sulfonates 50 O2/hn liq none
35 Cl2/hn
Glyoxylic acid 80 HNO3 liq none
Acetonitriled 80 O2 gas heter
N,N-Dimethyl fatty amine N-oxide 75 H2O2 liq none
Benzaldehydee 70 O2 liq homo
Cyclohexanone oximef 67 H2O2 liq heter
Hydroquinone 60 H2O2 liq homo/hetero
O2 liq none
MnO2 liq none
Caprolactone 60 H2O2 liq none
m-Phthalodinitrile 60 O2 gas heter
Methanesulfonic acid 30 HNO3 liq none
25 Cl2 liq none
Oxidized polyethylene waxes 50 O2 liq none/homo
Nitroalkanes (C1-C3)g 50 HNO3 gas none
1,2-Butylene oxide 50 Cl2 liq none
1,6-Hexanediolh 50 O2 liq homo
Nicotinic acid 40 HNO3 liq none
2,6-Naphthalene dicarboxylic acid 30 O2 liq homo
Cyclododeca-4,8-diene-1-onei 30 N2O liq none
Dodecanedicarboxylic acid 25 HNO3 liq homo
Catechol 20 H2O2 liq homo/hetero
Isoprenol 20 O2 gas hetero
3-Cyanopyridine 20 O2 gas hetero
a
O2 stands for both pure oxygen and air.
b
liq: reaction in liquid phase; gas: reaction in the gas phase.
c
homo: homogeneous catalyst, hetero: heterogeneous catalyst.
d
Capacity for purified acetonitrile formed as byproduct in acrylonitrile synthesis.
e
As a byproduct in benzoic acid synthesis.
f
Only includes the capacity for the ammoximation process.
g
Total capacity from propane nitration.
h
Ex 6-hydroxycaproic acid and adipic acid byproducts formed in the cyclohexane oxidation.
i
Intermediate in the BASF N2O-based cyclododecanone process.

processes (7585% of the capacity), which produced by homogeneous catalysis, far larger
reflects the high investment costs and the than oxo products or acetic acid.
high costs for developing the required hetero-
geneous catalysts that are associated with pro-
cesses in the gas phase. 2.1. Inorganic Chemicals Produced by
Almost half of the capacity uses heteroge-
neous catalysts, one-third uses homogeneous
Oxidation
catalysts, and the remaining 20% uses no cata-
2.1.1. Sulfur Compounds
lyst at all. This might also seem counter-
intuitive, because large processes are often Sulfur Dioxide.
associated with heterogeneous catalysis and
when asked for examples of homogeneously S O2 ! SO2 1
catalyzed processes, usually only the hydro-
formylation of olefins or the carbonylation of Sulfur dioxide, SO2, is one of the most basic
methanol to acetic acid is mentioned. Actually, inorganic chemicals, located right at the base of
many of the large commodities produced by the sulfur value chain (! Sulfur Dioxide). It is
oxidation also use homogeneous catalysis, and formed in many combustion processes and in
here again terephthalic acid is the most promi- ore processing as a secondary product, so it is
nent example, being the largest product quite difficult to assess the world capacity. It is
6 Oxidation

Table 1d. Small-scale specialties produced by oxidation (capacity <20
103 t/a )

Product or intermediate Capacity, 103 t/a Oxidanta Phaseb Catalystc

Methyl ethyl ketone oxime 15 H2O2 liq heter

4-tert-Butyl benzaldehyde 15 e liq none
MnO2 liq none
O2 liq homo
Vanillin/ethylvanillin 10 O2 liq homo
Trimethylbenzoquinone 10 O2 liq homo
Azelaic acid 10 O3/O2 liq none
Dimethylol propionic acid 10 H2O2 liq none
Pyromellitic dianhydride 10 O2 gas heter
Cyclopentanoned 10 N2O liq none
o-Phthalodinitrile 10 O2 gas hetero
Benzonitrile 10 O2 gas hetero
Hexafluoropropylene oxide 10 O2 liq/gas none
Triethylamine N-oxide 8 H2O2 liq none
Dimethyl sulfone 8 H2O2 liq none
Cocamidopropyl N-oxides 5 H2O2 liq none
1,8-Naphthalenedicarboxylic anhydride 5 O2 gas hetero
f
1-Pentene oxide 5
Cyclododecanediols 5 O2 liq homo
Trichloroacetaldehyde hydrate 5 Cl2 liq none
Trimethyl adipic acid 4 HNO3 liq none
4-Methoxybenzaldehyde (anisaldehyde) 35 e liq none
MnO2 liq none
Crotonic acid 3 O2 liq homo
o-Chlorobenzonitrile 3 O2 gas hetero
Salicylaldehydee 3 O2 liq hetero
N-Methylmorpholine N-oxide 3 H2O2 liq none
Cyclohexene oxide 2 H2O2 liq none
Menadione 1 CrVI liq none
Biphenyl tetracarboxylic dianhydride 1 O2 liq homo
Methoxyacetic acid 1 O2 liq hetero
Furaneol 0.5 O3 liq none
a
O2 stands for both pure oxygen and air.
b
liq: reaction in liquid phase; gas: reaction in the gas phase.
c
homo: homogeneous catalyst, hetero: heterogeneous catalyst.
d
Total capacity for cyclopentanone. The fraction produced by N2O oxidation has not been disclosed.
e
Only part of it is produced by an oxidation reaction.
f
Both Cl2/water, H2O2, and peracids seem to be used commercially as oxidants by different producers.

also produced on purpose, mainly by burning a catalyst this is a very slow reaction. The oldest
sulfur with air. Its main use is in the production industrial processes for the production of sul-
of sulfuric acid. furic acid (lead chamber process and tower
process) involved the oxidation of SO2 in the
Sulfur Trioxide and Sulfuric Acid. aqueous phase with nitrogen oxides as cata-
lysts. However, this process has the dis-
SO2 1=2 O2 ! SO3 2 advantage that the concentration of the
produced sulfuric acid is limited to a maximum
To obtain sulfuric acid, SO2 must be oxidized of 78%. For this reason, the process is not
further to sulfur trioxide (SO3) (Eq. 2) before it economically competitive with the newer con-
is absorbed in concentrated sulfuric acid to first tact process. In the contact process a heteroge-
obtain oleum, a solution of SO3 in sulfuric acid neous catalyst based on vanadium oxide and
(! Sulfuric Acid and Sulfur Trioxide), which is alkali metal sulfate (usually sodium and potas-
then diluted with water to obtain concentrated sium) as the promoter on a silica support is used
sulfuric acid. Although the oxidation of SO2 almost exclusively. The alkali metal sulfate
with O2 to SO3 is exothermic, in the absence of forms a melt that contains the active vanadium
 Oxidation 7

complex and remains trapped in the pores of the 2.1.2. Carbon Derivatives
support, so this is actually a supported liquid-
phase-type catalyst. Standard catalysts are Synthesis Gas.
active only above 400 C, but at this temperature
Cx Hy Oz O2 ! CO H2 CO2 4
the equilibrium conversion of SO2 is already
below 99%. This limitation can be overcome by
technical measures like the intermediate Synthesis gas, or syngas, is the common name
absorption of SO3 or by stabilizing the melt used for mixtures of CO, CO2, and H2 of
by adding cesium salts. The development of a varying composition. These mixtures can be
catalyst that would be active at significantly obtained from almost any carbon-containing
lower temperature is still a major challenge. feedstock. Usual feedstocks are natural gas,
Considering the huge worldwide capacity petroleum, and coal (! Methanol, Section
for sulfuric acid, which is estimated at ca. 5.1), but recently the production of synthesis
350
106 t/a (as 100% sulfuric acid) a break- gas from renewable feedstocks has become a
through in low-temperature catalysts for the focus of active research [9, 10]. The largest
oxidation of SO2 would have a major impact [7]. source of syngas is the steam reforming of
methane, which is not an oxidation reaction
but actually a disproportionation. However,
Sulfur, Elemental. Sulfur is one of the basic syngas is also formed in partial oxidation pro-
raw materials of chemical industry. (! Sulfur) cesses like the oxidative gasification of coal or
It occurs naturally mainly as gypsum, calcium hydrocarbons (! Gas Production, 2. Processes)
sulfate, but this is of no relevance for the or as a byproduct in acetylene synthesis. Syngas
industrial production of sulfur. It also occurs is a pivotal raw material in industrial organic
as elemental sulfur, as H2S in natural gas, and as chemistry and has many uses. It is used as such
organic sulfur compounds in oil and coal. H2S for the production of methanol (estimated
must be removed from natural gas, because it is world capacity in 2014: 98
106 t/a) and as a
both corrosive and highly toxic. Organic sulfur source of H2 and CO for hydroformylation and
compounds must also be removed from the production of oxo chemicals (estimated
naphtha and fuels, mostly by hydrogenative world capacity for oxo chemicals in 2014:
desulfurization, which also produces H2S. 14
106 t/a). Syngas can be subjected to a water
Essentially all of the H2S is oxidized to gas shift to convert CO and water to H2 and CO2
elemental sulfur in the Claus process (Eqs. 3), for the production of H2, which is then used for
many variations of which are in commercial use. the synthesis of ammonia and for all kinds of
hydrogenation reactions. Syngas can also be
H2 S 3=2O2 ! SO2 H2 O separated into its components (Eq. 4), mostly to
3
produce pure CO, which is then used as a raw
SO2 2 H2 S ! 3 S 2 H2 O
material for the production of methyl formate
Both steps can be conducted in one stage or and pivalic acid, among others.
separately in two stages. In a first step H2S is
burned with air to produce SO2, no catalyst 2.1.3. Nitrogen Compounds
being required. In a second step the SO2 reacts
with H2S to form elemental sulfur and water, NO and Nitric Acid.
which can be viewed as an oxidation in which
SO2 plays the role of the oxidant. This step 4 NH3 5 O2 ! 4 NO 6 H2 O 5
requires a catalyst, and many materials have
been tested for this purpose, although vanadium Industrially, NO is produced by catalytic com-
oxides are by far the most active ones. This bustion of ammonia over a suitable catalyst,
subject has been thoroughly reviewed [8]. usually platinum gauze, at temperatures in the
The total production capacity for elemental range of 850950 C (Eq. 5). (! Nitric Acid,
sulfur is estimated at around 60
106 t/a, but no Nitrous Acid, and Nitrogen Oxides, Sections
breakdown between mining output and Claus- 1.3.1 and 3.2). This is a highly efficient process
derived sulfur production is available. with very high selectivity and conversion at
8 Oxidation

very low contact times. The only drawback of No reliable data exist on the world capacity
the process is the slow loss of platinum from for hydroxylamine, so this can only be roughly
the catalyst at the high reaction temperatures. estimated from the capacity of caprolactam,
However, the use of 510% rhodium can con- which is by far the largest single use for
siderably stabilize the catalyst and decrease hydroxylamine. In 2008 the estimated world
platinum losses. capacity for hydroxylamine should thus be ca.
Most of the NO produced worldwide is used 1.5
106 t/a (calculated as 100% NH2OH).
to manufacture nitric acid. In this process mix-
Hydrazine. Historically, hydrazine was first
tures of NO and air are adsorbed in water,
industrially produced by the Raschig process.
usually at pressures between 1 and 12 bar.
This is a two-step process involving the oxida-
This is a multistep reaction that requires no
tion of ammonia with sodium hypochlorite to
catalyst. First, NO is oxidized by O2 to yield
chloramine and subsequent reaction of the
NO2, which then reacts with water to a mixture
chloramine with a large excess of ammonia
of nitric and nitrous acid. The latter dispropor-
(Eqs. 9) (! Hydrazine, Chap. 4).
tionates to nitric acid and NO. The overall
stoichiometry is given by the following equation. i NaOCl NH3 ! NH2 Cl NaOH 9

4 NO 2 H2 O 3 O2 ! 4 HNO3 6 ii NH2 Cl NH3 NaOH ! NH2 NH2 NaCl H2 O

An improvement of the original Raschig

The estimated world capacity for nitric acid process was achieved by performing the
in 2011 was estimated to be around 72
106 t/a reaction in the presence of acetone (Bayer
(based on 100% HNO3) and has been stagnant Ketazine process, now operated by Lanxess)
since the late 1980s. or methyl ethyl ketone (Fisons process, which
Hydroxylamine. is no longer in use). In these processes, the
primary product is not hydrazine but a ketazine.
TS-1 In a second step the ketazine is hydrolyzed to
NH3 H2 O2 
! NH2 OH H2 O 7
produce free hydrazine and recover the ketone.
A ketazine is also the primary product of the
Hydroxylamine in the form of hydroxyl-
state-of-the-art technology. This process, devel-
ammonium sulfate is mostly produced by
oped originally by Pechiney-Ugine-Kuhlmann,
hydrogenating NO. However, the discovery
involves the reaction of H2O2 with NH3 in the
that titanium silicalite (TS-1) oxidizes NH3 to
presence of methyl ethyl ketone and a catalyst
hydroxylamine, which can be trapped in situ by
to form methyl ethyl ketazine. This ketazine is
ketones to form oximes (Eq. 7). has led to the
then hydrolyzed in a similar way as in the
development of several industrial processes for
Bayer Ketazine process to produce hydrazine
the production of oximes. These are discussed
(Eq. 10).
in more detail in Section 2.2.13.
The old Raschig process, involving the oxi-
dation of ammonium carbonate with NO and
NO2 to ammonium nitrite, reaction with SO2 to
hydroxylamine disulfonate, and subsequent
hydrolysis to hydroxylammonium sulfate, is
still in use (Eqs. 8) (! Hydroxylamine, Chap. 4).

10

Although the Pechiney-Ugine-Kuhlmann

process is already a fairly efficient process
and uses a less problematic oxidant, it is still
a two-step process with considerable complex-
ity. In view of the large number of applications
8 for hydrazine, much thought has been given to
 Oxidation 9

potential single-step processes. The hydrogen- Cathode (ODC) Technology for HCl electrol-
ation of N2 and the dehydrogenative coupling ysis [13] together with ThyssenKrupp and
of ammonia are so thermodynamically UHDENORA [14] that is claimed to provide
unfavorable that they can not be considered a energy savings up to 600700 kWh/t Cl2.
reasonable reaction to target (DGf(25 C) UHDENORA recently announced a 80
103 t/a
159.4 and 192.2 kJ/mol respectively). The sodium hydroxide plant in Binzhou City,
direct oxidation of ammonia with H2O2 is China using this technology [15]. Bayer
thermodynamically feasible (DGf(25 C) announced in 2006 that it intends to build an
159.4 kJ/mol), but hydrazine is a much stron- ODC HCl electrolysis plant at its integrated
ger reducing agent than NH3 and so it should be site in Shanghai to recycle HCl from isocyanate
very difficult to avoid overoxidation. Another (MDI) production and also offers this technology
thermodynamically feasible route would be the for licensing [16]. More recent developments,
hydrogenation of N2O (DGf(25 C) 172.9 for instance, from Sumitomo Chemical, use
kJ/mol), but this route seems to have never been the Deacon process, involving the catalytic
taken into consideration and should also be very oxidation of HCl with oxygen in the gas
challenging due to the kinetic inertness of N2O. phase on supported Ru catalysts [17]. Bayer
Nevertheless, hydrazine has a wide range of announced in 2008 that it will build a plant
applications, and the world capacity for hydra- to recycle HCl from toluene diisocyanate
zine is estimated to be at least 60
103 t/a, so the synthesis by using Sumitomos technology at
development of a one-step process for the its Shanghai site [18].
production of hydrazine is an attractive but
challenging target. Chlorate and Perchlorate. For the production
of chemicals containing chlorine in a higher
2.1.4. Chlorine Derivatives oxidation state than elemental chlorine, see !
Chlorine Oxides and Chlorine Oxygen Acids.
Chlorine. Chlorine is one of the largest-scale Two of them are discussed here, because they
inorganic chemicals produced by oxidation. are produced by dedicated oxidation processes
The world capacity for the production of chlo- and also because of their importance.
rine is estimated at 77
106 t/a with several Sodium chlorate (NaClO3) is produced on a
hundred producers and production sites world- large scale by the electrolysis of sodium chlo-
wide. Almost the entire amount of chlorine is ride. This is a complex reaction involving the
produced by the electrolysis of sodium chloride electrochemical oxidation of chloride to hypo-
solutions in mercury, diaphragm, or membrane chlorite, followed by an autoxidation of hypo-
cells (Eq. 11) (! Chlorine, Chaps. 5, 6, 7). chlorite to chlorate. The simplified pathway is
shown in Equations (12).
e
2 NaCl 2 H2 O ! Cl2 2 NaOH H2 11
2 Cl ! Cl2 2 e
Cl2 H2 O HClO H Cl 12
The fixed ratio of chlorine and NaOH pro-

duced in the above processes has the drawback 3 HClO ! ClO
3 3 H 2 Cl

that they have little flexibility to respond to

fluctuations in the demand for these two chem- The world capacity in 2012 was estimated to
icals. For this reason, and since HCl is also exceed 4
106 t/a. The major use of sodium
available as a byproduct in several processes chlorate is in the on-site production of chlorine
using Cl2, such as phosgene-based isocyanate dioxide for the elemental chlorine-free bleach
production, processes have also been developed in the pulp and paper industry. The second
to oxidize HCl back to chlorine [11]. However, largest use of sodium chlorate is for the pro-
such processes amount to at most 5% of the duction of sodium perchlorate by electrochem-
total world capacity for chlorine. Older pro- ical oxidation (Eq. 13).
cesses like the Kel-Chlor process used nitric ClO3  H2 O ! ClO4  2 H 2 e 13
acid as the oxidant, but this process is not used
commercially anymore [12]. Bayer Material- The major use of sodium perchlorate is in the
Science developed an Oxygen Depolarized production of ammonium perchlorate, which is
10 Oxidation

one of the major components in solid rocket technology for the on-purpose production of
fuels for civil and military use. For this reason butadiene [21]. The basic technology itself is
there are no reliable data on production capaci- quite old but has been considerably improved
ties, but it can be estimated that it is in the range [22, 23]. Currently, this technology is used in a
of (1020)
103 t/a. few plants in China with a combined capacity of
350
103 t/a. Both Mitsubishi and BASF/Linde
have announced the development and licensing
2.2. Organic Chemicals Produced by of on-purpose butadiene technology by oxida-
Oxidation tion of n-butenes using Bi/Mo catalysts [24]. It
is expected that the butadiene capacity based on
2.2.1. Olefins and Alkynes such oxidative dehydrogenative technology
Olefins by Oxidative Dehydrogenation. will grow rapidly.
Despite considerable research efforts [19], the Another olefin for which a process involving
oxidative dehydrogenation of alkanes to pro- an oxidation is available is styrene. Most of the
duce lower olefins such as ethylene, propene, styrene is produced by dehydrogenation, but
butenes, and isobutene is of no commercial since this reaction is highly endothermic it
relevance. Although pure dehydrogenation is must be performed in multistage reactors with
an endothermic reaction and thus requires con- intermediate heating. As an alternative, UOP and
siderable energy input, neither the pure oxidative Lummus offer the Styrene Monomer Advanced
dehydrogenation nor the autothermal oxidative Reheat Technology (SMART) for licensing. In
dehydrogenation, in which only part of the this process, after the first dehydrogenation
hydrogen is oxidized in order to make the overall stage, the hydrogen formed is selectively burned
reaction slightly exergonic, have been able to with air in the presence of steam on an alumina-
supersede dehydrogenation (Eqs. 14). supported platinumtin catalyst [25]. It is not
known if this process is already being commer-
cially used.
A similar situation is found in the oxidative
coupling of methane (Eq. 15). In spite of inten-
sive research efforts and the fact that only
modest yields of about 20% are deemed neces-
sary to achieve economic feasibility, this tech-
nology is still not used industrially, and the
target yield has still not been achieved [26].
14 2 CH4 O2 ! H2 C CH2 2 H2 O 15

Oxidative dehydrogenation can also be used It is expected that research on this subject
for the production of butadiene from linear will continue in spite of the fact that the avail-
butenes. Although in the past several plants ability of cheap ethane from shale gas and the
existed to produce butadiene by oxidative conversion of a large number of naphtha crack-
dehydrogenation of n-butane and linear butenes ers to light feed (essentially ethane) has made
(! Butadiene, Section 4.1), the surplus of oxidative methane coupling less attractive.
butadiene on the market has led to a shutdown
of almost all plants using oxidative dehy- Acetylene. In contrast to olefins, acetylene is
drogenation technology (using either the Petro- produced industrially by the coupling of meth-
Tex Oxo-D or the Phillips O-X-D process). ane. The reaction is strongly endothermic, but
Butadiene is now almost exclusively isolated due to the favorable entropy it becomes exer-
from butadiene-containing streams obtained gonic at temperatures above 1230 C (Eq. 16).
from the steam cracking of hydrocarbons, but DH298 376 kJ=mol
recent price increases and supply constraints 2 CH4 ! HC  CH 3 H2 r
16
DGr < 0 at T > 1230 C
may lead to a renaissance of on-purpose buta-
diene processes [20]. UOP has recently To provide the required energy for the
announced that it will license the OXO-D reaction, part of the methane used is oxidized
 Oxidation 11

in situ and thus effectively the reaction is an development of processes that directly lead to
uncatalyzed oxidative methane coupling. The ethylene glycol. This is especially true for the
many different process variations are described C1-based process based on the oxidative car-
in detail elsewhere (! Acetylene). Although bonylation of methyl nitrite to dimethyl oxalate
these processes account for most of the acety- und subsequent hydrogenation to ethylene
lene produced in Europe and the USA, globally glycol. This process is discussed under
they only make up less than 10% of the global oxalic acid (Section 2.2.8) and methyl nitrite
capacity of over 7
106 t/a. Most of the capacity (Section 2.2.13).
is located in China and uses calcium carbide as Propylene Oxide. Propylene oxide is, after
the source for acetylene. ethylene oxide, the second most important
epoxide with an estimated worldwide capacity
2.2.2. Epoxides
in 2012 of close to 10
106 t/a. In contrast to
Ethylene Oxide. Ethylene oxide is now pro- ethylene oxide, all propylene oxide produced is
duced exclusively through the oxidation of isolated as pure product and only a small part of
ethylene with O2 in the gas phase using a sil- it, around 20%, is transformed into propylene
ver-based catalyst (Eq. 17) (! Ethylene Oxide). glycol, the major use being the production of
polyetherols as raw materials for polyurethanes
(ca. 60%). In strong contrast to ethylene oxide,
17 propylene oxide cannot be economically pro-
duced by direct oxidation of propene with O2
(Eq. 18). This is due to the high reactivity of
Ethylene oxide is the second largest scale the allylic hydrogen atoms in the methyl group
organic chemical produced by oxidation, of propene leading to a high percentage of
with an estimated installed capacity of over propene being oxidized by an uncatalyzed
20
106 t/a in 2012. However, more than 75% and unselective background reaction.
of the world capacity is not isolated as pure
ethylene oxide; instead, the crude product
mixture is immediately treated with water to 18
produce ethylene glycol. A typical example
of a modern EO/EG plant design is the
Dow/Union Carbide METOR technology Nevertheless, in the early 1970s a process
[27]. for the direct oxidation of propene to propylene
Much research has been devoted to the oxide in the liquid phase using benzene as the
optimization of the catalyst. Modern catalysts solvent was developed up to the pilot-plant
use silver supported on alumina and doped with scale (10 t per month). The reaction was per-
small amounts of other metals. Depending on formed at 200 C and 55 bar. The yield of
the nature of the dopants, catalysts can be propylene oxide based on propene was 45%
divided into two types. The first type has a and the spacetime yield was 700 kg m3 h1
somewhat lower selectivity (8082%) at high [30, 31]. However, due to the skyrocketing price
productivity (>200 kg m3 h1), works at of oil in the mid-1970s these developments were
lower temperatures (210230 C), and has a abandoned. Still, these values are interesting as a
longer lifetime of up to four years. The second benchmark for later works on the direct oxida-
type has significantly higher selectivity tion of propene that are discussed below. The
(8890%) but a lower productivity (<180 kg selective catalytic epoxidation of propene with
m3 h1), works at temperatures above 230 C, O2 still remains one of the Holy Grails of indus-
and has a shorter lifetime of at most two trial research on oxidation. Many efforts have
years. This subject has been thoroughly been devoted to the development of catalysts for
reviewed [28, 29]. this oxidation in the gas phase, mostly based on
Since most of the ethylene oxide is not used Ag as the active metal, but the problem is still far
as such, but is immediately transformed to from being solved. For example, one of the best
ethylene glycol, the future of ethylene oxide catalysts known contains Ag doped with Mo, K,
technology is very dependent on the current Cl, and N and supported on CaCO3 [32]. The
12 Oxidation

selectivity to propylene oxide is 64% at a per- also called TBA/PO process) which is shown
pass conversion of propene of only 1.5% at in Equation (20).
250 C and requires the addition of NO, CO2,
and ethyl chloride to the reactor feed. Strikingly,
the catalyzed reaction in the gas phase actually
runs at a higher temperature than the uncatalyzed
20
reaction in the liquid phase. The catalyzed
reaction achieves an almost 20% higher selec-
tivity, but at the price of a very low per-pass
conversion, and most of all of a very low space
time yield of only 4.5 kg m3 h1. This is more In the first step, isobutane is oxidized with
than two orders of magnitude lower than that of O2 to produce tert-butyl hydroperoxide (see
the uncatalyzed reaction. also Section 2.2.3). The hydroperoxide is
Since the reaction with O2 is not selective then used to epoxidize propene, wheras the
enough, it was necessary to develop processes hydroperoxide is reduced to tert-butyl alcohol
based on other terminal oxidants [33] (! Pro- (TBA). Since the market for TBA is rather
pylene Oxide). The oldest process for the prep- limited it was mostly transformed into methyl
aration of propylene oxide is the chlorohydrin tert-butyl ether (MTBE), which used to find
process or PCH process. In this process, shown wide application as an oxygenate additive for
in Equation (19), propene reacts with chlorine fuels. The phasing out of MTBE in the USA,
in the presence of a large excess of water to the development of more efficient cooxidation
produce a dilute aqueous solution of propylene processes and the process-inherent problems,
chlorohydrins (mostly 2-chloro-1-propanol). such as the low selectivity in the first step and
This solution is then treated with a base to the safety problems associated with the oxida-
generate propylene oxide, which is then iso- tion of isobutane, quickly led to a loss of
lated by distillation. attractiveness of this process and its capacity
share never went much above 10% of the total
installed capacity. Recently, two new plants
19
have been announced in China which use this
technology, because in China there is still an
attractive and growing market for MTBE as a
This process is still in use and currently fuel additive.
accounts for about 40% of the world capacity, A few years after the MTBE/PO process, a
but its importance is quickly decreasing new cooxidation process took over the techno-
because this process is not being used in new logical lead, the so-called SM/PO process [34],
plants. This is both due to the environmental which is summarized in Equation (21).
problems associated with it, such as the large
amount of wastewater and the chlorinated
byproducts, and also to the high specific invest-
ment, because parallel to a propylene oxide
plant, a chlorine plant is also necessary to 21

deliver both the Cl2 as well as the NaOH for

the second step. Although large PCH plants in
western countries have been upgraded to cope
with waste streams in an environmentally
acceptable manner, these additional measures Two main factors contributed to the success
just added to the already high specific invest- of the SM/PO process, of which several varia-
ment for plants using this technology. tions exist using homogeneous MoVI or hetero-
To overcome the problems associated with geneous Ti-on-silica catalysts: first of all, the
the PCH process a whole family of cooxidation good selectivity in the oxidation of ethylben-
processes was developed. The oldest of these is zene with O2 to the corresponding hydro-
the so-called MTBE/PO process (sometimes peroxide, and second the large available
 Oxidation 13

market for the coupled product styrene. In fact, More recently, the PO technology has been
the SM/PO process actually produces more styr- revolutionized by the introduction of the so-
ene than propylene oxide (about 2.3 t of styrene called HPPO processes, in which hydrogen
per tonne of propylene oxide). In times when the peroxide (HP) is used as the oxidant.
styrene market is short and the prices are high,
the SM/PO process makes it possible to produce
23
propylene oxide at very attractive costs. How-
ever, when the styrene market is long the picture
can reverse completely. In spite of these draw-
backs, the technology was for a long time best- The discovery of titanium silicalite-1 (TS-1)
in-class, and until recently about one-third of the in 1979 at Enichem [35] and the subsequent
world capacity for propylene oxide used this demonstration that it is an efficient heteroge-
technology. neous catalyst for the epoxidation of propene
To try to mitigate the dependence on the using methanol as the solvent (Eq. 23) [36]
coupled product, Sumitomo developed a coox- sparked a frenzy of work both in industry and
idation process that is free of coupled products academia [37]. However, it took almost 20
and started the first plant in 2003. This is usually years for the commoditization of hydrogen
called the CUPO or CMHP process (Eq. 22). peroxide to bring its costs of production into
a region where its use as an oxidant in the
propylene oxide production became economi-
cally attractive. The first two commercial plants
went into service almost simultaneously. Both
plants, from a BASF/Dow joint venture and
22
Solvay as the hydrogen peroxide supplier
(300
103 t/a in Antwerp, Belgium) and from
SKC with a technology license from Evonik/
KruppUhde and Evonik as the hydrogen perox-
ide supplier (100
103 t/a in Ulsan, South
The process starts from cumene, which is Korea), started in late 2008. Since then, other
first oxidized to cumene hydroperoxide (see plants have either started production, are under
also Section 2.2.3). The cumene hydroperoxide construction, or are in the planning phase. Both
is then used to epoxidize propene to propylene the processes of BASF/Dow [38] and of Evonik/
oxide by using a proprietary Ti-on-silica cata- KruppUhde [39], as well as the technology in
lyst, and is itself reduced to cumyl alcohol. The general [40, 41], have been recently reviewed.
cumyl alcohol is then hydrogenated back to Besides the technologies which are used
cumene. Overall the process consumes 1 mol of commercially, many others have been devel-
O2 and 1 mol of H2 per mole of propylene oxide oped, several of them up to the pilot-plant stage.
formed. Although two plants using this tech- The ultimate dream would be to be able to
nology are currently in operation (200
103 t/a epoxidize propene directly with O2. As men-
from Sumitomo in Japan and 200
103 t/a from tioned above, BASF piloted the oxidation of
Rabigh Refining & Petrochemical in Saudi propene with O2 in the liquid phase in
Arabia), this cooxidation process is not consid- the presence of a cobalt salt as catalyst. The
ered to be the most economically attractive. In reaction had a good spacetime yield, but the
spite of the many improvements possible by selectivity to propylene oxide was only 45%. In
using the easily oxidizable cumene, in particu- addition, its purification was a major challenge,
lar the decreased amount of unconverted alkane due to the plethora of byproducts formed. Since
that must be recovered and recycled, this pro- then many efforts have been made to improve
cess still suffers from high complexity that on this reaction. One of the directions taken was
leads to a high specific investment. It does to modify Ag catalysts to improve the selectiv-
not suffer the burden of a coupled product, ity in the propene epoxidation. For instance,
but it also does not have the advantage that Arco has developed Ag catalysts doped with
SM/PO enjoys when the styrene market booms. Mo, K, Cl, and N and supported on calcium
14 Oxidation

carbonate [32]. By using these catalysts at Besides the oxidation with O2, the oxidation
250 C and cofeeding NO, ethyl chloride, and with H2O2 already became a focus of research
CO2 as promoters, a selectivity to propylene long before hydrogen peroxide was available in
oxide of 64% at a conversion of 1.5% was the amount and at the price that made it attract-
achieved. However, the spacetime yield of ive for large-scale production of propylene
4.5 kg of propylene oxide per cubic meter of oxide. In the 1970s Bayer/Degussa [50, 51]
reactor and hour is far too small for a technical and Interox [52] had already independently
application. developed propylene oxide processes using
Recently, Sumitomo reopened this line of perpropionic acid, generated from propionic
development by using copper-based catalysts acid and H2O2. The process was later described
doped with Ru and Na supported on SiO2 and in detail in the literature [53]. In addition to the
also cofeeding a chlorinated compound. At cost for H2O2, this process has serious problems
250 C and propene conversions of 59% due to corrosion and additional costs due to the
selectivities to propylene oxide of 4547% decomposition of perpropionic acid to CO2 and
were obtained [42]. Note, however, that the ethanol. The work was stopped in the mid-
uncatalyzed oxidation of propene with O2 in 1970s.
the gas-phase was thoroughly investigated and Another possible alternative to the above-
optimized, and under optimized conditions mentioned cooxidation processes is the so-
(290 C, 5 bar) a selectivity to propylene oxide called anolon/PO process. Here cyclohexane
of 62.8% at a propene conversion of 15.1% can is oxidized with O2 in the absence of a catalyst
be achieved [43]. This means that the uncata- to produce a mixture of cyclohexyl hydro-
lyzed reaction is still the most selective one. peroxide, cyclohexanol, and cyclohexanone.
Addition of a radical starter like acetaldehyde The cyclohexyl hydroperoxide can then be
increases the reaction rate and thus decreases used to epoxidize propene. BASF operated a
the reaction temperature. At 230 C the selec- 5 t/a pilot plant in the late 1980s using homo-
tivity to propylene oxide increases to 61%, but geneous Mo catalysts similar to those used in
this is still too low for a practicable technical the MTBE/PO process. Later, in the mid-2000s,
process [44]. ABB Lummus developed a similar process with
Another variation of the direct oxidation of Ti/Si oxide catalysts similar to the ones used in
propene, that is, one that uses O2 as oxidant and the Shell version of the SMPO process. [54].
does not require a coreductant, is the oxidative Compared to the other cooxidation processes,
acetoxylation of propene. It was originally the anolon/PO process has the disadvantage
developed by Halcon in the early 1970s and that the coupled product, cyclohexanone and
then further pursued by ChemSystems and cyclohexanol, has a much smaller market than
BASF in the late 1970s. Halcon developed a MTBE or styrene, and so the commercial
catalytic system based on iodine or HBr and opportunity for the realization of such a process
TeO2 [45], which was able to acetoxylate pro- is much more limited.
pene with O2 and acetic acid to propylene In the early 1990s BASF also developed a
glycol diacetate. BASF developed a catalyst nonoxidative process for propylene oxide based
based on basic titanium acetate [46], but later on the trimerization of formaldehyde to dihy-
joined ChemSystems to jointly develop the droxyacetone with a thiazolium catalyst [55]
TeO2/I2-based catalyst system [47, 48]. This and subsequent hydrogenation to propanediol
process had severe corrosion problems in the [56]. The development was stopped in the mid-
oxidative acetoxylation step. After the synthe- 1990s due to insufficient catalyst lifetime of the
sis of the diacetate it had to be hydrolyzed, and thiazolium salts.
the propylene glycol transformed into propyl- Another active field of research is the syn-
ene oxide [49]. Unfortunately, the catalysts for thesis of propylene oxide from propene, H2, and
the ring closure of propylene glycol to propyl- O2. This is also sometimes called a direct
ene oxide proved to be only moderately selec- oxidation, but this often leads to confusion
tive and to have only a limited lifetime, so the with the reaction in which only O2 is used.
process development was stopped in the early Sometimes it is referred to as the in situ H2O2
1980s. process, but this is also misleading, since it is
 Oxidation 15

not at all clear that H2O2 is in all cases the capacity [60]. It is also produced by BASF in
intermediate. Such a process would be particu- its propylene oxide plant in Ludwigshafen,
larly interesting for a propylene oxide producer Germany. Both companies use the chlorohydrin
that has no access to H2O2 technology, and so technology (Eq. 25).
this route has been intensely investigated by
many companies. The first report of such a
process came already in 1992 from Tosoh
[57]. It became even more attractive after the 25
discovery that gold catalysts were also active
for this reaction and very high selectivities of There is little information available on the
propylene oxide based on propene were end uses of butylene oxide, but known appli-
obtained. However, the major problem is still cations are as a raw material in the production
the low H2 efficiency. Between 70 and 90% of of base oils for lubricants and for fuel additives.
the hydrogen used reacts unproductively to The world capacity can be roughly estimated as
produce water, and in spite of extensive 50
103 t/a.
research efforts there has been little progress
in coping with this problem. The work in this Other Epoxides from Chlorohydrins. Due to
field has been reviewed [58]. the constraints of the technology, the produc-
tion of chlorohydrins by addition of chlorine
Epichlorohydrin. After ethylene and propyl- and water to an olefin is not a general method
ene oxide, epichlorohydrin is the third largest for producing epoxides. The process can only
scale epoxide commercially produced. The be used when two conditions are met: 1) the
world capacity for epichlorohydrin is estimated chlorohydrin has some solubility in water,
to be around 2.3
106 t/a. In contrast to propyl- because otherwise an organic phase is formed
ene oxide, more than 95% of the capacity for in which olefin gathers, leading to extensive
epichlorohydrin uses the chlorohydrin process, formation of chlorinated byproducts, and 2) the
a process similar to the PCH process used for epoxide must have a boiling point below that of
propene but with allyl chloride as the starting water or form a low-boiling azeotrope with
material (Eq. 24) (! Epoxides). water.
Due to these limitations, the number of
epoxides that can be produced by the chloro-
24
hydrin process is rather limited. Besides 1-
butene oxide (bp 63 C), these are 2-butene
Although a lot of work was put into the oxide (bp 5460 C), isobutene oxide (bp
development of a hydrogen peroxide-based 51 C), 1-pentene oxide (bp 90 C), cyclopen-
process, mostly in the early 2000s by Solvay tene oxide (bp 102 C, forms a low-boiling
and more recently by Dow, this technology has azeotrope with water), and cyclohexene oxide
not yet been developed to a commercial pro- (bp 129 C, forms a low-boiling azeotrope with
cess. In the meantime, with the development by water (Eq. 26)).
Solvay of the Epicerol technology, the focus
has shifted to a process which uses glycerol as
feedstock and does not involve an oxidation.
The first plant with a capacity of 100
103 t/a
and located in Map Ta Phut, Thailand started 26
production in 2012 [59]. This is a remarkable
development, especially if one considers that
less than a decade ago epichlorohydrin was still
used to produce synthetic glycerol.
But of these, only 1-pentene oxide and
Butylene Oxide. Dow has produced 1-butene cyclohexene oxide are produced in sizeable
oxide at least since the 1950s in a dedicated amounts. 1-Pentene oxide is used in the pro-
plant in Freeport, Texas of undisclosed duction of 1,2-pentanediol, a building block for
16 Oxidation

the fungicide Propiconazole [61] and also as a water; otherwise, the epoxides will be hydro-
solvent in cosmetics. Although many different lyzed to the corresponding diols.
processes have been proposed for the synthesis The method is used commercially in the
of this building block [62], the hydrolysis of the production of epoxidized oils, such as epoxi-
epoxide made by using the chlorohydrin pro- dized soybean and linseed oils, used as plasti-
cess is the dominant one. From the market size cizers and stabilizers for PVC. This is by far the
for Propiconazole it can be estimated that the largest application for this technology, with a
capacity for 1-pentene oxide should be around world capacity for epoxidized oils of around
5
103 t/a. Cyclohexene oxide is also produced 150
103 t/a.
commercially, mostly as a building block The same technology is also used to produce
for the synthesis of the acaricide Propargite a wide range of specialty aliphatic and cyclo-
[63, 64]. Although many processes for the aliphatic epoxides used, for instance, as
preparation of cyclohexene oxide have been reactive diluents in epoxy resins. (! Epoxy
described, the method most likely used indus- Resins, Chap. 11).
trially is an indirect synthesis of the chlorohy- Other specialty epoxides produced using per-
drin by treating HCl with aqueous hydrogen acids formed in situ include epoxidized a-olefins
peroxide in the presence of cyclohexene. With with 1020 carbon atoms, epoxidized b-oligoi-
this method a steady low concentration of Cl2 is sobutene, 3,4-epoxycyclohexyl methyl-3,4-
generated and the formation of excessive epoxycyclohexylcarboxylate, both mono- and
amounts of chlorinated products can be avoided bis-epoxidized limonene and a-pinene oxide,
[65]. Isobutylene oxide and 2-butylene oxide but probably none of these products reaches a
are also produced commercially by using the capacity above 103 t/a.
chlorohydrin process, but with capacities well The use of performic acid formed in situ is
below the 103 t/a threshold. advantageous from a process point of view,
The limitations of this technology can be since the reactions can be performed in multi-
overcome if the chlorohydrins are produced in a purpose batch reactors. However, from an eco-
different way, for instance, by OH/Cl exchange nomic point of view it suffers from two main
as in the synthesis of epichlorohydrin and gly- disadvantages. First, it requires the use of con-
cidol from glycerol, or by addition of a nucleo- centrated hydrogen peroxide solutions, usually
phile to epichlorohydrin. These are important 70 wt%, to drive the equilibrium of performic
processes to produce epoxides, but since these acid formation to the right. The second and
are not oxidations, they are not discussed here. most serious disadvantage is that the formic
acid is usually discarded. The corrosiveness of
Other Epoxides from Nonchlorohydrin Pro- formic acidwater mixtures and the fact that
cesses. Many other epoxides are produced as formic acid forms an azeotrope with water,
specialties using oxidation technologies other make the workup of aqueous formic acid eco-
than the chlorohydrin process. The most tech- nomically unattractive compared to discarding
nically relevant is the so-called performic acid it, especially if it still contains hydrogen
technology. In this case epoxidation is per- peroxide.
formed by performic acid formed in situ Hydrogen peroxide is also the terminal oxi-
from formic acid and aqueous hydrogen perox- dant in the Degussa process for the synthesis of
ide (Eqs. 27). glycidol from allylic alcohol (Eq. 28) [6669].

27 28

The process is carried out in water as the

The method can only be successfully applied solvent with wolframate as the catalyst, and a
in the synthesis of epoxides when the epoxide plant with a capacity of 3000 t/a was brought
to be produced has a wide miscibility gap with into operation in 1980 in Germany [70], but a
 Oxidation 17

few years later it was partly destroyed by a a second step, the anthrahydroquinone is oxi-
severe explosion [71]. Glycidol is still produced dized with O2 to form H2O2 and regenerate the
commercially from 1-chloro-2,3-propanediol, anthraquinone, see Equation (30).
but this process does not involve an oxidation.
A similar process, but on a much smaller scale,
was used for the commercial production of cis-
2,3-epoxy-1,4-butanediol as a raw material for
the production of 1,2,4-butanetriol [72].
Another commercially important epoxide is
hexafluoropropene oxide (! Fluorine Com-
pounds, Organic, Section 6.1.2). It is prepared
by oxidation of hexafluoropropene, either in the
gas phase or in solution in the absence of a
30
catalyst (Eq. 29) [7375].
The oxidation of the anthrahydroquinone
with O2 does not require the use of a catalyst.
29
In the oldest processes the alkyl side chain in
the anthraquinone is simply ethyl, but in mod-
ern processes anthraquinones with a longer side
There are no reliable data on the world chain are used to increase their solubility in the
capacity for hexafluoropropene oxide, but con- working solution. One of the big advantages of
sidering that there are at least two major pro- this process, besides the safety aspects of not
ducers (DuPont and 3M) and a large number of having to bring H2 and O2 together in one step,
applications, ranging from polymeric perfluori- is the fact that in the extraction step a rather
nated ethers, small building blocks such as concentrated (usually up to 40 wt% H2O2) is
perfluoro(methyl vinyl ether), hexafluoroace- directly obtained without need for costly con-
tone, and hexafluoroisopropyl alcohol, and centration steps. This crude hydrogen peroxide
widely used anesthetics like Sevoflurane and solution is, for example, directly usable in the
Enflurane, the capacity probably exceeds HPPO processes.
10
103 t/a.
tert-Butyl Hydroperoxide. tert-Butyl hydro-
peroxide is produced as an intermediate in
2.2.3. Hydroperoxides the MTBE/PO process (see Section 2.2.2) by
oxidizing isobutane with O2. The reaction is a
Hydrogen Peroxide. In the last 20 years hydro- radical chain oxidation performed without the
gen peroxide has grown from an expensive use of catalysts in order to maximize the selec-
specialty to a large-scale commodity. In the tivity to the hydroperoxide (Eq. 31). The
mid-1990s, the world capacity for hydrogen reaction is typically conducted at 140 C and
peroxide (based on 100% H2O2) was only a pressure of 3.5 MPa with long residence times
around 1.5
106 t/a and a world-scale plant of around 10 h. The reaction is carried out to
had a capacity of around (2040)
103 t/a. less than 40% conversion, but even then the
With the increasing use of hydrogen peroxide selectivity to tert-butyl hydroperoxide is only in
in pulp bleaching and after 2008 in the HPPO the range of 5060%, the remainder being
process (see Section 2.2.2) world capacity has mostly tert-butanol [77].
now reached almost 5.5
103 t/a, with the larg-
est plant having a capacity of 390
103 t/a [76]
Although in the past other processes have 31
been used industrially, currently hydrogen per-
oxide is produced exclusively by the anthraqui-
none process. (! Hydrogen Peroxide) In this From the installed capacities for the MTBE/
process a 2-alkylanthraquinone is first hydro- PO process, a capacity of around 2
106 t/a for
genated to the corresponding hydroquinone. In tert-butyl hydroperoxide can be inferred.
18 Oxidation

Although the oxidation is not very selective,

the fact that it uses cheap isobutane as raw
material, requires no catalyst, and that tert-
butanol is also a valuable coproduct, have
established this process as the major source
of tert-butyl hydroperoxide. Part of the tert-
butyl hydroperoxide produced is worked-up 32
for sale, since it finds use as polymerization
starter or as a raw material for the production The oxidation of cyclohexane is one of the
of other peroxides, such as tert-butyl most intensively studied reactions with more
peresters and dialkyl peroxides (! Peroxy than 1200 patents and scientific papers, but also
Compounds, Organic). one of the fields in which unsubstantiated
claims of improved selectivities abound.
Cyclohexyl Hydroperoxide. Cyclohexyl Because the reaction must be stopped at low
hydroperoxide is produced as an intermediate conversions, it is not easy to determine the
or byproduct, depending on the process variant cyclohexane conversion and the product selec-
used, in the oxidation of cyclohexane with O2, tivities, so the usual procedure is to try to detect
but is not isolated and not available commer- all possible reaction products and to calculate
cially (! Cyclohexanol and Cyclohexanone). the selectivity of the reaction from the sum of
Still, it plays a major role and it is estimated all observed reaction products. Unfortunately,
that around 500
103 t/a is contained in the the oxidation of cyclohexane leads, besides the
product streams of cyclohexane oxidation major products cyclohexanol, cyclohexanone,
reactors. Most of the processes for oxidation and cyclohexyl hydroperoxide, to a plethora of
of cyclohexane use a soluble cobalt(II) carbox- byproducts [78]. Many of these byproducts,
ylate as a catalyst in minute amounts (less than such as hydroperoxides and carboxylic acids,
100 ppb of metal). The reaction is carried out are difficult to detect, and with such a large
between 145 and 160 C and only allowed to number of byproducts, even if each one of them
proceed to a conversion in the range of 34% in is formed in very small amounts, they add up to
order to keep selectivity high. The main prod- non-negligible amounts. So, most of the claims
ucts of the reaction are cyclohexanol and cyclo- of improved selectivities have proven on closer
hexanone with only small amounts (<10%) of scrutiny to be incorrect. Even some of the
cyclohexyl hydroperoxide. In a subsequent reaction pathways that have been accepted
deperoxidation step the cyclohexyl hydro- for a long time have now been proven to be
peroxide is decomposed to form more cyclo- incorrect [79, 80]. The picture becomes even
hexanol and cyclohexanone. The molar ratio of more complicated when a catalyst is involved,
cyclohexanol to cyclohexanone in the final and only very recently has light been shed on the
product, which is usually known as KA oil, pathways for the cobalt-catalyzed reaction [81].
is around 2. Ethylbenzene Hydroperoxide. Ethylbenzene
The oxidation can also be carried out hydroperoxide, or more correctly 1-phenylethyl
without the use of a metal catalyst. In this hydroperoxide, is an important intermediate in
case slightly higher temperatures, up to the so-called SMPO processes for the synthesis
180 C, are used to compensate for the slower of propylene oxide (see Section 2.2.2). Ethyl-
reaction. The conversion is also restricted to benzene hydroperoxide is obtained as a dilute
34%, but in this case the main product solution by oxidizing ethylbenzene with air at
formed is cyclohexyl hydroperoxide. This is 140150 C and a pressure of 34 bar without a
then decomposed in a deperoxidation step catalyst (Eq. 33) [82].
using a homogeneous cobalt carboxylate to
yield a product mixture almost identical to that
obtained when a catalyst is used in the oxida-
tion step. In both cases the overall selectivity 33
to cyclohexanol plus cyclohexanone is about
85% (Eq. 32).
 Oxidation 19

To keep the selectivity high, the reaction is hydroperoxide is used captively. A relatively
stopped at 78% conversion, but even at this small amount is used in the production of cumyl
low conversion the selectivity to ethylbenzene peroxyesters and dicumyl peroxide, which are
is below 85%, with 1-phenylethanol and ace- important radical initiators.
tophenone as the major byproducts. The reactor An analogous technology is also used,
effluent is then flashed to concentrate the mainly by Japanese companies, to produce
hydroperoxide to around 1518 wt%. This cresols from cymenes. In this case mixtures
solution is then used on site for the epoxidation containing both the tertiary and the primary
of propene. The world capacity for ethylben- hydroperoxide are obtained and the primary
zene hydroperoxide is estimated at around hydroperoxide has to be selectively decom-
6
106 t/a, but the product is only consumed posed before performing the Hock cleavage
captively and is not commercially available. [83]. The capacity for cresols using this tech-
nology is, however, quite modest, and estimated
Cumene Hydroperoxide. Cumene hydro- to be only around (100150)
103 t/a.
peroxide, or more correctly 1-methyl-1-phenyl-
ethyl hydroperoxide, is an important Diisopropylbenzene Dihydroperoxide. p-Dii-
intermediate both in the synthesis of phenol sopropylbenzene dihydroperoxide is an inter-
by the Hock process and in the CUPO process mediate in the Hock-type process for producing
developed by Sumitomo for the production of hydroquinone (see also Section 2.2.7). Diiso-
propylene oxide (see Section 2.2.2). Cumene propylbenzene is oxidized with air in the pres-
hydroperoxide is obtained by oxidizing cumene ence of small amounts of base to obtain the
with air, usually in a cascade of stirred-tank dihydroperoxide, which is isolated and
reactors or bubble columns at temperatures in then treated with acid to induce the cleavage
the range of 100140 C and a pressure of 67 (Eq. 35) (! Hydroquinone, Section 4.1).
bar and usually with small amounts of a buffer
to prevent acids from building up (Eq. 34).
Since cumene hydroperoxide, as a tertiary alkyl
hydroperoxide, is much more stable than ethyl-
benzene hydroperoxide, the oxidation can be
taken to a higher conversion with still reason-
able selectivity. Usually the conversion is lim-
ited to around 20%, leading to selectivities to
cumene hydroperoxide in the range of 9095%. 35

An analogous process starting with m-diiso-

propylbenzene is used to produce resorcinol (!
34 Resorcinol, Section 4.2). The combined capac-
ity for both isomers can only be roughly esti-
mated, but should be in the range of
(100150)
103 t/a, since some of the plants
In the CUPO process the reactor effluent is can be used to produce both products.
further concentrated to obtain a 35 wt% solu-
tion of cumene hydroperoxide, which is then Other Hydroperoxides. Not many other hydro-
used in the epoxidation of propene. In the Hock peroxides made by using oxidation reactions
phenol process the cumene hydroperoxide is have commercial significance. The synthesis of
further concentrated to at least 80 wt% before tert-amyl hydroperoxide from tert-amyl alco-
being sent to the acid-catalyzed cleavage hol and H2O2 is considered to be a nucleophilic
(! Phenol). substitution and not an oxidation.
The worldwide capacity for the production 7-Hydroperoxy-3,7-dimethyloct-5-ene is an
of cumene hydroperoxide is estimated to be intermediate in two different commercial syn-
around 17
106 t/a. In contrast to ethylbenzene theses of rose oxide. It is generated by treating
hydroperoxide, not all of the cumene b-citronellol with singlet oxygen, which,
20 Oxidation

depending on the process, can be photo- traces of a metal as catalyst or as a Bashkirov

chemically [84] or chemically [85] generated. reaction in the presence of stoichiometric
The hydroperoxide is then reduced and cyclized amounts of boric acid, were old processes
to obtain rose oxide (Eq. 36). used to prepare mixtures of aliphatic secondary
alcohols. The technology is not in use anymore
and the last plants were closed decades ago.
Cyclohexanol. The oxidation of cyclohexane
to cyclohexyl hydroperoxide, cyclohexanol,
and cyclohexanone is discussed in Section
2.2.3.
36
A modification of the process that leads
almost exclusively to the formation of cyclo-
Percarboxylic Acids. Currently, percarboxylic hexanol, the so-called Bashkirov oxidation,
acids, mainly peracetic acid, are produced by uses boric acid in the oxidation step and the
acid-catalyzed equilibration of the correspond- main product obtained is cyclohexyl borate
ing carboxylic acid with concentrated H2O2, (mainly tricyclohexyloxyboroxin). In a second
but this reaction is better viewed as a nucleo- step, the cyclohexyl borate is hydrolyzed to
philic substitution and not an oxidation produce cyclohexanol. The boric acid is
reaction. Until recently peracetic acid was still recovered and recycled to the oxidation step
produced by oxidation of acetaldehyde with O2, (Eq. 38).
but the last plant using this technology, Dow
Chemicals Peroxymeric Chemicals unit, was
closed down in October 2006 [86]. This
reaction is now of no commercial relevance. 38

2.2.4. Alcohols
Linear Even-Numbered Alcohols. Linear
This process has the advantage that the
even-numbered primary alcohols with a wide
oxidation can be carried out to a higher conver-
distribution of chain lengths, so-called Alfols,
sion (up to 20%, but usually the reaction is
can be produced by a Ziegler-type process in
stopped at 15% conversion) while still achiev-
which ethylene is oligomerized with triethyla-
ing good selectivities of around 90%. However,
luminum. The resulting trialkyl aluminum
most of the advantage in conversion and selec-
compounds are then oxidized with dry air to
tivity is lost because of the additional steps of
produce aluminum trialkoxides, which are then
borate hydrolysis and boric acid recovery [87,
hydrolyzed to alcohols (Eq. 37) (! Fatty Alco-
88]. For this reason the Bashkirov oxidation is
hols, Section 2.4.1, ! Alcohols, Aliphatic,
usually viewed as less advantageous from an
Section 2.3.5, ! Ziegler Processes, Chap. 4).
economic point of view, and it accounts for only
about 3% of the total installed capacity.
37 The world capacity for KA oil or anolone
(the technical mixture of cyclohexanol and
cyclohexanone) is around 7
106 t/a. Of this
This process has been steadily losing
total capacity, approximately 5% is made by
ground, especially because it always leads to
hydrogenation of phenol and not by oxidation
statistical mixtures of alcohols. Currently only
of cyclohexane. This means that the total
Sasol is using this technology to produce linear
amount of cyclohexanol produced directly in
alcohols with an estimated capacity of around
the oxidation of cyclohexane is around
200
103 t/a.
4.5
106 t/a.
Alcohols from Paraffins. The oxidation of From the byproducts formed in the cyclo-
paraffinic hydrocarbons with oxygen, either hexane oxidation in the absence of boric acid,
as an unselective radical chain reaction with and more exactly from the acid washing water
 Oxidation 21

that contains adipic acid and 6-hydroxycaproic as dehydrogenation processes [95]. The
acid as the main components, a considerable dehydrogenation of methanol to formaldehyde
amount of 1,6-hexanediol can be obtained. is an endothermic process, which severely lim-
For this purpose the byproducts are either its the thermodynamically achievable methanol
esterified and hydrogenated [89] or directly conversion. To overcome this limitation air is
hydrogenated [90]. The capacity for 1,6- also fed into the reactor, albeit in substoichio-
hexanediol from this source has been recently metric amounts. Under these reaction condi-
increased to 50
103 t/a [91]. tions, the O2 reacts with H2 formed in the
dehydrogenation, thus providing the thermo-
Cyclododecanol. While the Bashkirov process dynamic driving force to achieve high methanol
is only of minor importance in the oxidation of conversions. Oxygen is also required to gener-
cyclohexane, in the oxidation of cyclododecane ate (sub)surface oxygen species on the Ag,
it is the method of choice used by all the which are thought to be essential for the first
producers (! Cyclododecanol, Cyclododeca- step of the catalytic cycle, which is the depro-
none, and Laurolactam). The reason for this lies tonation of methanol (Eq. 39) [96, 97].
in the higher selectivity and higher conversion
CH3 OH x =2 O2 ! H2 C O x H2 O 1  xH2 39
achievable by using the Bashkirov process. In
fact, these advantages more than compensate
for the disadvantages associated with the use of In the conventional silver-catalyzed process,
boric acid, with the necessity of hydrolyzing the usually referred to as the BASF or full-conver-
primarily formed cyclododecyl borates, with sion process, O2 is used in less than 60% of the
the recycling of boric acid, and most of all stoichiometric amount, but conversions of
with the problems associated with handling methanol in excess of 97% are achieved. The
the wastewater containing borates [92]. overall yield of formaldehyde is around 90%.
The total world capacity for cyclododecanol In the wet gas recycling silver process, also
is estimated to be around 100
103 t/a, almost often referred to as the incomplete-conversion
equally divided between two plants of Invista process, even less air is used, with O2 only at
(Victoria, Texas) and Evonik (Marl, Germany). around 40% of the stoichiometric amount. The
During the oxidation of cyclododecane in conversion of methanol is only around 80% and
the presence of boric acid, considerable unconverted methanol must be separated and
amounts, up to 10%, of cyclododecanediols, recycled. Although the selectivity in the
are formed as byproducts [93]. These diols can reaction is slightly higher than in the full-con-
be isolated by distillation [94] and the technical version process, the losses in methanol during
mixture containing 3540% of 1,4-, 1012% of workup lead to very similar overall formalde-
1,5-, and 4550% of 1,6-cyclododecanediol is hyde yields. In both of these silver-catalyzed
offered by Invista under the trade name C12 processes the reaction is performed in an adia-
CD. The capacity for this product is estimated batic manner in a short-bed reactor with very
to be around 5
103 t/a. short contact times and reaction temperatures
around 600 C. In both variants the reactor
effluent contains considerable amounts of H2.
2.2.5. Aldehydes Finally, the third process, often referred to as
the FORMOX process, is a real oxidation pro-
Formaldehyde. Formaldehyde is the largest- cess [98100]. In this process an iron molybdate
scale aldehyde produced commercially, with catalyst is used and the reactor feed contains O2
an estimated worldwide capacity of around in approximately twice the stoichiometric
20
106 t/a (as 100% formaldehyde). In general amount. The reaction can be performed at a
there are three distinct types of processes for much lower temperature of around 300 C, but
the production of formaldehyde, all of which a tube-bundle reactor must be used to remove the
use methanol as the feedstock and air as oxidant heat of reaction. The conversion of methanol is
(! Formaldehyde). essentially complete and the selectivity to form-
Two of these processes use silver as the aldehyde is 24% higher than in the silver-cata-
catalyst and are more appropriately described lyzed processes. However, the high costs for the
22 Oxidation

reactor put this process at a disadvantage from can be achieved. Usually, small amounts of
the point of view of capital expenditure. acrylic acid are also formed, but since the
acrolein produced is usually not isolated and
Acetaldehyde. Currently two different oxida-
the crude stream is simply directed to a second
tive processes are in use for the production of
reactor and oxidized to acrylic acid, this is
acetaldehyde (! Acetaldehyde). The oldest
usually not a drawback.
process, but probably the one with the brightest
The world capacity for acrolein is estimated
future, is the oxidation of ethanol (Eq. 40),
to be around 5
106 t/a. Most of the acrolein is
while the Wacker oxidation of ethylene is
not isolated but is instead used as a crude
quickly losing ground.
gaseous stream for the synthesis of acrylic
CH3 CH2 OH 1=2 O2 ! CH3 CHO H2 O 40 acid. Probably less than 100
103 t/a of acrolein
are isolated and used in other processes, such as
The development of efficient processes the synthesis of glutaric aldehyde, methionine,
for the synthesis of acetic acid through carbon- 1,3-propanediol or tetrahydrobenzaldehyde.
ylation of methanol made the synthesis of Although a considerable amount of effort
acetic acid from acetaldehyde economically has been devoted to the development of a
unattractive, and many plants using the Wacker catalyst capable of directly oxidizing propane
oxidation technology were forced to close. to acrolein, this transformation still remains
Presently the worldwide capacity for acetalde- elusive. The best catalysts known have a selec-
hyde is still around 2
106 t/a, but this capacity tivity below 30% at a propane conversion below
is seriously underutilized. In fact the amount of 20% [103, 104].
acetaldehyde produced worldwide has dropped
by almost 50% since the mid 1990s and is Methacrolein. Methacrolein (! Acrolein and
currently estimated to be only around Methacrolein) only has relevance as an inter-
1.5
106 t/a. mediate in the production of methyl methacrylate
The increased availability of bioethanol is and methacrylic acid (Eq. 42). Aroundthree quar-
leading to a resurgence of ethanol oxidation, ters of theworld capacityfor methyl methacrylate
especially at times when oil and naphtha prices is produced from acetone cyanohydrin, but espe-
are high. This process has the added advantage cially in Asia, the oxidation of isobutene or tert-
of being quite simple and requiring only modest butanol is an important route for making the
capital investments, making it attractive for methacrolein used in the production of methyl
small- to medium-scale plants in developing methacrylate [105].
countries having access to cheap bioethanol.
The Wacker oxidation technology, on the other
hand is burdened with severe corrosion prob-
lems and problematic waste streams. These
problems lead to high investment costs and 42
for this reason, only very large scale plants
are economically viable. It is thus not likely An estimated 1.2
106 t/a of methacrolein is
that any new plants for acetaldehyde will be produced in Asia in this way. As in the case of
built using this technology in the next decades. acrolein, quite a lot of work has been devoted to
Acrolein. Acrolein (! Acrolein and Metha- the development of a direct oxidation of iso-
crolein) is exclusively produced by oxidation of butane to methacrolein, but also in this case,
propene with air in the gas phase over a bismuth known catalysts are still far from achieving
iron molybdate catalyst [101, 102], the so- economically interesting selectivities [106].
called SOHIO process (Eq. 41).
Isoprenal (3-Methyl-3-butenal). Isoprenal is
41 produced industrially as an intermediate in
BASFs citral synthesis. It is made by oxidizing
The reaction is run at propene conversions isoprenol (3-methyl-3-buten-1-ol) with O2 on a
above 95% and, depending on the exact catalyst copper or silver [107109] catalyst in the gas
being used, a selectivity to acrolein above 90% phase (Eq. 43).
 Oxidation 23

43
tert-Butylbenzaldehyde is also produced
commercially by oxidizing 4-tert-butyltoluene
with manganese dioxide [115] or with air in the
BASF has published a capacity for its citral presence of a Co salt and bromide as catalysts in
plant in Ludwigshafen [110], and from this the acetic acid [116]. These two processes lead to
capacity for isoprenal can be estimated to be the formation of mixtures of the aldehyde and
approximately 20
103 t/a. the corresponding carboxylic acid, but at least
in the case of the oxidation of 4-tert-butylto-
Benzaldehyde. Although many processes have luene this is less of a problem, because the 4-
been proposed for the direct or indirect oxida- tert-butylbenzoic acid is itself a commercial
tion of toluene to benzaldehyde, only two product. The oxidation with MnO2 is also
of these have any industrial relevance [111] used commercially for the oxidation of 4-
(! Benzaldehyde). The older process, based methoxytoluene to 4-methoxybenzaldehyde,
on the hydrolysis of benzal chloride, is not also known as anisic aldehyde.
considered here. The second process, the There are no disclosed capacities, neither for
liquid-phase oxidation of toluene with O2 the electrochemical [117] nor for the chemical
(Eq. 44), is the dominant process, contributing oxidation processes. The fact that the same
most of the estimated 70
103 t/a of benzalde- plant can be used to produce both aldehydes
hyde produced worldwide. This process is actu- makes estimates even more uncertain. How-
ally performed to obtain benzoic acid as the ever, it is estimated that the total capacity for
valuable product, but the reaction also yields 4-tert-butylbenzaldehyde is approximately
considerable amounts of benzaldehyde as a 15
103 t/a, while the capacity for 4-methoxy
byproduct. benzaldehyde is considerably smaller and in the
range of (35)
103 t/a.

44 Glyoxal. Glyoxal is produced commercially by

two different oxidation processes (! Glyoxal).
The oldest process, now used only by Clariant
in France, is based on the oxidation of acetal-
In fact, since benzaldehyde is actually more dehyde with nitric acid (Eq. 46). The reaction
valuable than benzoic acid, efforts have been takes place in the liquid phase at 40 C and the
made to improve the yield of the byproduct yield is only around 70%. The major drawback
benzaldehyde [112]. This process is discussed of this process is the consumption of nitric acid,
in more detail in Section 2.2.10. which is used as a stoichiometric oxidant, and
Specialty para-substituted benzaldehydes the production of an off-gas containing consid-
like 4-tert-butyl and 4-methoxy benzaldehydes erable amounts of N2O.
are industrially produced by the hydrolysis of
their dimethyl acetals, which are in turn pro-
duced by electrochemical oxidation of the cor-
responding toluenes (Eq. 45) [113, 114].
46

In this process 375 kg of N2O is generated

per tonne of glyoxal produced [118]. Even
after a catalytic treatment of the off-gas, 65 kg
of N2O still remains unconverted and is emit-
ted into the atmosphere [119]. This process
45
accounts for a little under 20% of world
A typical example is the synthesis of 4-tert- capacity for glyoxal, which is estimated to
butyl benzaldehyde dimethyl acetal, a building be around 250
103 t/a. Most of the capacity,
block for the Lily of the Valley scent 2-methyl- both in western countries and in Asia, uses the
3-(4-tert-butyl-phenyl)propanal, by electro- catalytic oxidation of ethylene glycol with air
chemical oxidation of 4-tert-butyltoluene. (Eq. 47).
24 Oxidation

Salicylic Aldehyde. Salicylic aldehyde (!

47
Benzaldehyde, Section 8.3) is usually produced
by the ReimerTiemann reaction of phenol,
which involves no oxidation step. Rh^one-Pou-
The reaction has very high requirements for lenc (now Rhodia-Solvay) developed a differ-
catalyst performance, because there are only ent process based on the oxidation of saligenin
limited possibilities to purify the crude aqueous with O2 using a Bi-modified Pt catalyst in the
glyoxal solution obtained [120]. Industrially, presence of boric acid (Eq. 49) [128, 129].
both metallic Ag [121, 122] and Cu [123] are
used as catalysts and both show similar per-
formances. In practice, it is necessary to
achieve almost complete conversion of ethyl- 49
ene glycol (> 98%) and only a low selectivity to
glycolaldehyde (less than 0.5 wt%) can be
tolerated. The main side product is CO2, There are no secure data on the world
together with small amounts of formaldehyde capacity for salicylaldehyde, but it is esti-
and formic acid which must be removed by mated to be only around 3
103 t/a. It is
stripping and ion exchange. The price to pay for also not known how much of it is produced
the high conversion required is a quite low by the oxidation route.
selectivity to glyoxal, usually in the range of
6371%, obviously leaving a lot of room for
improvement [124]. 2.2.6. Ketones
Vanillin and Ethylvanillin. The synthesis of Acetone. Acetone is currently produced almost
vanillin and ethylvanillin, with a combined exclusively as a byproduct in Hock-type pro-
capacity of around 10
103 t/a among the larg- cesses from the acid-catalyzed decomposition
est flavors/fragrances, also encompasses an of isopropylbenzene hydroperoxide (! Ace-
oxidation step. Vanillin is made by reaction tone). The synthesis of the hydroperoxide pre-
of guaiacol with glyoxylic acid to selectively cursor is discussed in Section 2.2.3. On a much
form 4-hydroxy-3-methoxy mandelic acid. The smaller scale, acetone is also produced from
mandelic acid is subsequently oxidatively propan-2-ol. However, the fraction of acetone
decarboxylated to vanillin in the presence of made from propan-2-ol is less than 5% of
a homogeneous metal catalyst and oxygen as the overall capacity for acetone and even this
the terminal oxidant (Eq. 48) (! Flavors and small fraction of the overall capacity uses a
Fragrances, Section 2.6.3). This synthetic route dehydrogenation process. The uncatalyzed oxi-
has now almost completely replaced the older dation of propan-2-ol with O2 to give a mixture
process involving isolation from black sulfite of acetone and H2O2 was used until the mid-
pulp liquor [125]. 1980s as a process for H2O2 production (the so-
called Shell process) but this has been super-
seded by the anthraquinone process and is not
in use anymore (Eq. 50).
48

50

However, the oxidation of propan-2-ol was

Classically the reaction was performed with reconsidered by ARCO (now LyondellBasell)
a CuII salt, usually copper sulfate, as the cata- as a method for generating H2O2 in situ for the
lyst [126], but recently claims have been made epoxidation of propene [130, 131]. However,
that better results can be achieved by using a the additional complexity introduced by the
combination of two metals like CuII and CoII presence of unconverted propan-2-ol in the
[127]. H2O2 solution made this process more
 Oxidation 25

complicated and less efficient than the HPPO

process using crude H2O2 from an anthraqui-
none plant. The oxidation of propan-2-ol to
acetone has thus only historical relevance. 51
The Wacker oxidation of propene to acetone
is also no longer in use. Compared to the classical process (hydro-
genation of cyclododecatriene, oxidation to
Cyclohexanone. The oxidation of cyclohexane cyclododecanol and dehydrogenation to cyclo-
to cyclohexyl hydroperoxide, cyclohexanol, dodecanone), this route is shorter and requires
and cyclohexanone is discussed in Section only one catalyst instead of three. The process
2.2.3. The uncatalyzed or cobalt-catalyzed oxi- has also a higher overall yield and higher
dation of cyclohexane leads to the formation of product purity.
a mixture of cyclohexanone and cyclohexanol The world capacity for cyclododecanone is
in a 0.5:1 ratio. However, for the production of estimated to be around 100
103 t/a, but only
caprolactam pure cyclohexanone is required one-third is produced using the N2O-based
and considerable effort has been devoted to process.
search for catalysts with a higher selectivity
to cyclohexanone. Chromium catalysts are Cyclopentanone. In contrast to cyclohexanone
claimed to provide a higher selectivity to cyclo- and cyclododecanone, industrial processes for
hexanone [132], but the rapid deactivation of the synthesis of cyclopentanone do not involve
the catalyst has thus far prevented its industrial an oxidation. The classical process for cyclo-
use. Recently, the combination of cobalt and pentanone production involves a Dieckmann
chromium has been claimed to lead to an condensation of adipic acid or its methyl ester,
increased ratio of cyclohexanone to cyclohex- and is a simple and high-yield process but with
anol, up to 1.4:1, but without the deactivation a very low atom economy. The oxidation of
problems observed when chromium is used as cyclopentane has never been used for the pro-
the sole catalyst [133]. duction of cyclopentanol and cyclopentanone,
The world capacity for KA oil or anolone although cyclopentane is available as a starting
(the technical mixture of cyclohexanol and material. This is due to the fact that at similar
cyclohexanone) is around 7
106 t/a. Of this conversions, the selectivity to alcohol and
total capacity, approximately 5% is made by ketone is 20-25% lower in the case of cyclo-
hydrogenation of phenol and not by oxidation pentane than it is in the case of cyclohexane.
of cyclohexane. This means that total amount of Recently, BASF developed and built a new
cyclohexanone produced directly by the oxida- production plant for cyclopentanone based on
tion of cyclohexane is around 2
106 t/a. The the uncatalyzed oxidation of cyclopentene with
additional demand is produced by the gas- N2O (Eq. 52) [134, 136].
phase dehydrogenation of cyclohexanol to
cyclohexanone.
52
Cyclododecanone. Cyclododecanone is
mostly produced by dehydrogenation of cyclo-
dodecanol. Recently BASF developed a new The world capacity for cyclopentanone is
process for the production of cyclododecanone estimated to be around 10
103 t/a. The capac-
[134] based on the uncatalyzed oxidation of cis, ity of BASFs plant, the only one using an
trans, trans-1,5,9-cyclododecatriene with N2O oxidation route, has not been disclosed.
to produce a mixture of cis, trans- and trans,
cis-cyclododeca-4,8-diene-1-one in approxi- Furaneol. Another flavor and aroma chemical
mately 1:1 ratio [135]. In a second step, the whose synthesis also includes an oxidation step
cyclododecadienones are then hydrogenated to is furaneol, an important pineapple and straw-
produce the desired cyclododecanone (Eq. 51). berry aroma. It is synthesized by the reaction of
The first plant using this process was started up hex-3-yne-2,5-diol with ozone and subsequent
in 2009. reductive work-up (Eq. 53) [137].
26 Oxidation

53

The hexane-2,5-diol-3,4-dione formed in the The process was in use until 2010, when
oxidation step is then cyclized in a separate step DSM sold its benzoic acid plant to Emerald
to form furaneol. There are no published data and stopped operations at the oxidative
on the world capacity for furaneol, but it is decarboxylation plant in the Netherlands
estimated to be approximately 500 t/a. [141]. Although toluene is a cheaper raw
material than benzene (the starting material
2.2.7. Phenols and Derivatives in the Hock cumene hydroperoxide process),
the toluene-based process was not competitive
Phenol. The major process for the production for the production of phenol.
of phenol, the Hock process, also involves an Another phenol process that was developed
oxidation step, namely, the oxidation of cumene up to pilot-plant scale and was close to being
with O2 to cumene hydroperoxide. This commercialized was the so-called AlphOx pro-
reaction has already been discussed in Section cess. This was developed by Solutia in cooper-
2.2.3. This process accounts for the vast major- ation with the Boreskov Institute of Catalysis
ity of the worldwide capacity of approximately [142, 143]. The process involves the oxidation
11
106 t/a for phenol (! Phenol). of benzene with N2O in the gas phase on a Fe-
Until recently, the Dow/California Research doped ZSM-5 zeolite (Eq. 55).
process starting from toluene was still used
industrially. In the first step, toluene is oxidized
in the presence of a homogeneous cobalt cata-
lyst to produce benzoic acid. This step is still in
use for the production of benzoic acid, benzal-
dehyde, and benzyl alcohol and is described in 55

more detail in Sections 2.2.5 and 2.2.10. In the

second step, benzoic acid is oxidatively decar- The N2O required as oxidant would be
boxylated using CuII salts as catalysts to pro- obtained from the off-gas of the downstream
duce phenol (Eq. 54) [138]. adipic acid plant, or from a dedicated catalytic
oxidation of NH3 with O2. Solutia announced
plans to build a 150
103 t/a plant by 2003, but
the project was stopped in 2001 [144].
54 Recently, ExxonMobil took up the idea, first
put forward in 1954 by Rhone-Poulenc [145],
of using cyclohexylbenzene instead of cumene.
The idea was further developed by many dif-
The oxidative decarboxylation of benzoic
ferent companies. Cyclohexylbenzene, which
acid is a very interesting reaction, because
can be produced by self-hydroalkylation of
the OH group is actually introduced at the
benzene, is oxidized by air to give the tertiary
carbon atom next to the one to which the
hydroperoxide as the main product. This can be
carboxyl group was attached [139]. The last
cleaved with an acid catalyst to produce a
step actually consists of two steps that occur
mixture of phenol and cyclohexanone (Eq. 56).
sequentially in the same reactor. The primary
product of the oxidative decarboxylation is
phenyl benzoate, which is hydrolyzed in situ
to phenol and benzoic acid. For this reason,
magnesium salts were also commonly used as
co-catalysts to accelerate the hydrolysis step
[140]. 56
 Oxidation 27

However, the selectivity in the oxidation of

cyclohexylbenzene without a catalyst is much
58
lower than the selectivity obtained in the oxida-
tion of cumene, mostly because the secondary
CH groups in cyclohexylbenzene are more
reactive than the primary CH3 groups in cumene.
This makes the use of a catalyst such as N- The most widely used method is the so-
hydroxyphthalimide necessary to achieve high called Rhodia process, which uses a mixture
selectivities [146]. It remains to be seen whether of phosphoric and perchloric acids as catalyst
this process will achieve commercial status. and 70 wt% aqueous hydrogen peroxide as the
Although the direct oxidation of benzene to oxidant, the reaction being effectively an elec-
phenol without production of acetone as a co- trophilic aromatic substitution (Eq. 58) [148].
product is still considered as a Holy Grail in The reaction is performed in a large excess of
oxidation, the example of the Hock phenol phenol. The phenol conversion is only around
process (and also some processes for propylene 5%, but the hydrogen peroxide is completely
oxide) show that a coupled production can be converted. The combined yield of catechol and
quite successful if there is a sufficiently large hydroquinone is around 80% based on H2O2
market for both products. and around 90% based on phenol. The main
strength of this process is its simplicity and the
Hydroquinone and Catechol. Three processes fact that it is easily performed in a continuous
are commercially used for the production of manner. The ratio of catechol to hydroquinone
hydroquinone and catechol, all of which involve can be to a certain extent tuned, but is usually in
oxidations (! Hydroquinone). The oldest pro- the range of 1.5 to 2. This process accounts for
cess, which now accounts for less than 10% of about one-third of the combined world capacity
the world capacity and is only used by Chinese for catechol and hydroquinone.
producers, is based on the oxidation of aniline In the past, a different hydroxylation of
with manganese dioxide to give benzoquinone phenol used an FeII salt as the catalyst and
(! Benzoquinone). The benzoquinone is proceeded by a radical mechanism. However,
then reduced in a separate step with iron powder the process that had been operated by Brichi-
to produce hydroquinone (Eq. 57) [147]. mica, later Enichem, in Italy had low yields and
led to the formation of substantial amounts of
tar [149]. This plant was closed and replaced in
1986 by a new plant, with a capacity of 10
103
57 t/a, using the newly developed TS-1-catalyzed
hydroxylation of phenol. In this process phenol
is oxidized in the presence of titanium silicalite
TS-1 as the catalyst. The catalyst allows for a
Although the yield of hydroquinone is quite conversion of phenol as high as 30% and full
good, the costs for manganese dioxide and iron conversion of H2O2 while retaining good prod-
powder, the costs of disposal of the inorganic uct selectivities. The selectivities are claimed to
byproducts, and the fact that this process can be around 94% based on phenol and 84% based
only be performed in a batch manner makes it on H2O2, and catechol and hydroquinone are
less attractive. However, this process offers the produced in a ratio of approximately 1:1 [150,
possibility of accessing benzoquinone, which is 151]. In spite of these excellent values, the
a small-scale but quite high priced specialty technology has not been able to firmly establish
chemical. itself. In 1994 Enichem sold the diphenols plant
The second oldest process involves the and the catalyst production plant to Borregaard,
hydroxylation of phenol with concentrated but in 2010 Borregaard was forced to cease
hydrogen peroxide. This process, of which sev- production on account of economic problems.
eral variations exist, always leads to the forma- It finally sold the business in 2011 to the Indian
tion of a mixture of hydroquinone and catechol, company Camlin Fine Sciences, which has now
with catechol usually being the major product. restarted operations [152]. Interestingly, new
28 Oxidation

findings from Ube seem to indicate that Ti is not

necessarily required as an active metal in the
59
zeolite. Actually, H-BEA or H-BEA that has
been partly exchanged with Ca can lead, in the
presence of ketones as a cocatalyst, to similar or
even superior selectivities [153, 154]. The pres- From the point of view of capacity, the
ence of the ketone seems to be crucial, and largest quinone is trimethylbenzoquinone.
indeed Ube uses a variation of the Rhodia This is one of the key building blocks for
process in which a ketone is added as a catalyst synthetic vitamin E, with an estimated capacity
in its plant in Japan [155]. of around 10
103 t/a. The trimethyl benzo-
The above processes of phenol hydroxyl- quinone is obtained by oxidizing 2,3,6-trime-
ation all have the problem that they always thylphenol with O2 in the presence of a catalyst.
produce a mixture of catechol and hydro- Older processes used a cobalt(II) salen complex
quinone. Companies operating the phenol as the catalyst and dimethyl formamide as
hydroxylation are mainly interested in the cat- the solvent [156]. Although this process was
echol, like Rhodia, which uses it as a starting simple and gave high yields, it had the dis-
material in the production of vanillin and ethyl advantage that the catalyst could only be used
vanillin. However, since the market demand for once, since it decomposes during the reaction.
hydroquinone is much higher than for catechol, The process is not used commercially anymore
a modern process for the production of hydro- and has been superseded by a copper(II)-
quinone without the coproduction of catechol catalyzed oxidation. The oxidation is per-
had to be developed. This process involves the formed in a biphasic system consisting of water
alkylation of benzene or cumene to p-diisopro- and an aliphatic alcohol that has a low misci-
pylbenzene, its oxidation to the dihydroperox- bility with water, with O2 as the oxidant. The
ide, and subsequent acid-catalyzed cleavage. catalyst consists of a mixture of CuCl2 and a
This process is discussed in Section 2.2.3. chloride of an alkali or alkaline-earth metal,
The overall world capacity for hydroquinone usually lithium chloride or magnesium chloride,
and catechol is estimated to be around 80
103 and remains dissolved in the aqueous phase.
t/a, about three quarters of which is The trimethylphenol and the product quinone
hydroquinone. remain in the organic phase (Eq. 60) [157].
Resorcinol. Two processes are currently in use
for the production of resorcinol. The older
process, which is not an oxidation, uses ben- 60
zene sulfonation and alkaline melt of the meta-
benzene disulfonate. The process still accounts
for almost half of the world capacity for cate-
chol, which is estimated to be around 80
103 t/ Although the copper catalyst must be used in
a. Due to environmental problems associated high concentration, its recycling is very simple:
with the disposal of byproducts, newer plants the aqueous phase containing the copper(II)
use the Hock cleavage of the bis-hydroperoxide chloride and the other chloride salts is just
derived from meta-diisopropylbenzene. This separated and recycled. The losses of copper
reaction is described in Section 2.2.3. in the organic phase are minimal and the pro-
cess is truly catalytic. In fact, the total turnovers
Quinones. Only a few quinones are produced that can be achieved based on copper can be
commercially by oxidation. The parent com- extremely high, and are only limited by the
pound, benzoquinone, is obtained as an inter- physical losses of copper(II) chloride during
mediate in the oxidation of aniline with workup. The relevant parameter for an indus-
manganese dioxide (Eq. 59). The benzo- trial process is not the catalyst concentration. If
quinone is mostly used captively to produce a catalyst is easily recyclable, having to use
hydroquinone, but some is also commercially high concentrations is not a real drawback.
available. Unfortunately, this fact has not yet been widely
 Oxidation 29

recognized in academia, not even in very recent

reviews [158].
62
Another quinone to which a lot of attention
has been dedicated is 2-methylnaphthoquinone
(! Naphthoquinones), also known as menadi-
one or vitamin K3. Although vitamin K3 does
not occur naturally, it can be used by higher The mechanism of the reaction has been
organisms to produce the biologically active 3- studied and reviewed [161]. Poly(phenylene
phytyl derivatives known as vitamins K1 and ether) and its blends with other polymers are
K2. Menadione is produced by oxidizing 2- an important class of engineering plastics with
methylnaphthalene with chromium trioxide in a wide range of applications; world capacity is
sulfuric acid or sodium dichromate in acetic estimated to be 200
103 t/a.
acid (Eq. 61) [159]. This reaction has often
been quoted as being a prominent example of 2.2.8. Carboxylic Acids, Saturated
nonsustainable chemistry which should be
replaced by greener oxidation methods. Formic Acid. Formic acid is now exclusively
produced by hydrolysis of methyl formate or by
acidification of sodium formate, but in the past
formic acid was also produced as one of the
61 products in the oxidative degradation of n-
butane (! Formic Acid, Section 4.2.1). The
last plant using this technology was a Celanese
plant in Pampa, Texas which went out of
The use of stoichiometric amounts of CrVI operation in January 2009 [162]. The Celanese
compounds as the oxidant has not yet been process oxidized n-butane with oxygen in the
displaced by other oxidation methods for a liquid phase at 170180 C and ca. 50 bar with a
number of reasons. First of all, the oxidation homogeneous CoII salt as the catalyst. The
with CrVI proceeds very selectively and leads oxidation is quite unselective and produces a
only to very small amounts of the undesired 5- mixture of acetic acid, formic acid, propionic
methylnaphthoquinone, which is very difficult acid, butyric acid, methyl ethyl ketone, ace-
to separate from the desired menadione. After tone, acetaldehyde, acetone, and methanol.
the reaction, the mixture is treated with an Most of these products were isolated and
excess of sodium bisulfite to completely reduce sold, with the exception of acetaldehyde, which
any remaining CrVI to CrIII. The CrIII basic was recycled to the oxidation, where it served
sulfate formed as a byproduct is not emitted as a catalyst and as an additional source of
into the environment or even considered as acetic acid. The technology is now only of
waste, but is actually a valuable byproduct historical interest.
sold as a tanning agent for chrome leather.
Additionally, the world capacity for menadione Acetic Acid. The state-of-the-art technology
is quite small, at most 1000 t/a, and so the for the production of acetic acid is the carbon-
overall impact on the environment is only very ylation of methanol, which accounts for more
small, even when compared with other sources than 95% of the world capacity of about
of CrIII in wastewaters, such as the corrosion of 15
106 t/a (Eq. 63). In the past, however,
stainless steel equipment. oxidation processes played a major role (!
Acetic Acid). As described in the previous
Poly(Phenylene Ether). Poly(phenylene ether) section, acetic acid was obtained as the main
is a polymer obtained by oxidative polymeriza- product in the oxidative degradation of n-
tion of 2,6-xylenol (! Poly(Phenylene butane in the liquid phase, but this technology
Oxides)). The oxidation uses an homogeneous is not in use any longer. However, SABIC has
basic CuII chloride and an amine as ligand, for developed a technology for the catalytic oxida-
example, N,N,N0 ,N0 -tetramethylethylenedi- tion of ethane to acetic acid and is using it in its
amine (TMEDA) (Eq. 62) [160]. 30
103 t/a plant in Yanbu, Saudi Arabia.
30 Oxidation

63 Section 2.2.10). Indeed, the required make-up

of solvent due to its oxidative degradation has
become one of the major uses for acetic acid
Depending on the exact catalyst used and the and has been studied in detail [167].
reaction conditions, only CO2 or ethylene and
CO2 are formed as byproducts, and selectivities Propionic Acid and Other Carboxylic Acids
to acetic acid of up to 70% are claimed [163 without a-Branches. In contrast to acetic acid,
165]. The process is probably only economical for which the carbonylation of methanol is the
where cheap ethane is available as a feedstock. most important technology, the oxidation of
Showa Denko has also developed a one-step propionaldehyde is the major technology
oxidation of ethylene to acetic acid using an for the production of propionic acid (Eq. 65)
heterogeneous Pd catalyst (Eq. 64) [166]. How- (! Propionic Acid and Derivatives, ! Car-
ever, due to higher costs of ethylene as raw boxylic Acids, Aliphatic). The Ni-catalyzed
material, as compared to methanol or carbonylation of ethylene is used only in one
ethane, this technology has not found wide large plant of BASF. The oxidation of propio-
acceptance and only one plant with a capacity naldehyde accounts for around 70% of the
of 130
103 t/a has been built. world capacity of propionic acid which is esti-
Before the development of methanol carbon- mated to be around 350
103 t/a.
ylation, the oxidation of acetaldehyde was the
major route for the production of acetic acid.
The reaction is a radical chain oxidation per- 65
formed in the liquid phase using small amounts
of Mn and often also Cu salts as homogeneous The reaction is performed in the liquid phase
catalysts. and does not require a catalyst or an additive. The
same is true for all aliphatic aldehydes without
64
an a branch. They are all oxidized in excellent
yields without the need for a catalyst. The
reaction mechanism has been studied and the
Currently, after the 120
103 t/a plant of reaction kinetics has been shown to be dependent
Wacker in Germany was idled in September on the mechanism of initiation (thermal or pho-
2012, the only plants still using this technology tochemical). For the thermally induced oxida-
are located in Asia, mainly in China, but they tion, the only relevant one from an industrial
account for less than 5% of the world capacity point of view, the reaction has been shown to be
of acetic acid. of order 1.5 in aldehyde and 0.5 in O2 [168, 169].
This technology may regain some impor- The global production of commercially rele-
tance with the increased availability of ethanol vant non-a-branched carboxylic acids other
from renewable resources at attractive prices. than propionic acid by oxidation of the corres-
Indeed, SEKAB in Sweden runs a 23
103 t/a ponding aldehydes with O2 is estimated at ca.
plant for acetic acid based on ethanol. They use 120
103 t/a, the major products being, in
either cellulosic ethanol from its own pilot plant decreasing order of production amounts:
or ethanol imported from Brazil. 3,5,5-trimethyl hexanoic acid (isononanoic
Although only a modest share of acetic acid acid), n-heptanoic acid, n-valeric acid, n-
is produced by oxidation reactions, it plays a butyric acid, and n-nonanoic acid.
very important role as the solvent in the largest n-Nonanoic acid is additionally co-produced
oxidation process currently operated: the oxi- with azelaic acid from the ozonolysis of oleic
dation of p-xylene to terephthalic acid (see acid (Eq. 66).

66
 Oxidation 31

This technology is only used by Emery Adipic Acid. Adipic acid, with a worldwide
Oleochemicals and an estimated 20 thousand capacity of ca. 4
106 t/a, is the major aliphatic
tons of oleic acid per year are cleaved in this dicarboxylic acid and an important building
way. block for polyamides (! Adipic Acid). Indus-
trially, adipic acid is produced exclusively by
2-Ethylhexanoic Acid and Other a-Branched the oxidation of pure cyclohexanol or mixtures
Carboxylic Acids. Although non-a-branched of cyclohexanol and cyclohexanone with nitric
aldehydes can be oxidized with O2 in excellent acid (Eq. 68). These mixtures of cyclohexanol
yields to the corresponding non-a-branched and cyclohexanone known as KA oil are
carboxylic acids without the use of any cata- obtained by the oxidation of cyclohexane as
lysts or additives, this is no longer true of described in Section 2.2.3. Pure cyclohexanone
a-branched aldehydes. In the oxidation of can also be used as a starting material if sodium
a-branched aldehydes, it is necessary to use nitrite or NOx is added as starter, but since
an additive, usually a sodium or a potassium cyclohexanone is usually the more expensive
salt like a hydroxide or the salt of the acid to be raw material it is only used when cyclohexanol
produced, although potassium salts are usually is not available, for instance, when cyclohexa-
preferred due to their better solubility [170, none is obtained by hydrogenation of phenol.
171]. Other alkali metals, magnesium, zinc, The oxidation is mechanistically quite a
and cadmium [172] also show a positive effect complicated reaction, but it has nevertheless
on selectivity, albeit less pronounced. These an excellent yield, well above 95%. The only
additives are not catalysts, since they do not organic byproduct formed in sizeable amounts
increase the rate of the oxidation and in fact is glutaric acid which is usually also separated
even slow the reaction to a certain extent, but and sold, either as pure glutaric acid or as a
they do have a pronounced effect on the selec- mixture of adipic and glutaric acids.
tivity. In the absence of potassium the selectiv-
ity for the oxidation of 2-ethylhexanal to
2-ethylhexanoic acid is only around 80%,
68
with 2-heptyl formate being formed as the
main byproduct, whereas in the presence of
potassium the selectivity to 2-heptyl formate
decreases to less than 3% and the selectivity to The reaction uses a mixture of CuII nitrate
2-ethylhexanoic acid increases to more than and ammonium vanadate as homogeneous
95%. The same effect is also observed in the catalysts. During workup the catalysts are
oxidation of other a-branched aldehydes. recovered and reused, and only very small

67

2-Ethylhexanoic acid is by far the largest amounts have to be replenished. The use of
scale a-branched carboxylic acid with a nitric acid as the oxidant and the fact that
world capacity of around 200
103 t/a, the vast approximately 1 mol of NO2 and 1 mol of
majority of which is produced by oxidation N2O per mole of adipic acid produced are
of 2-ethylhexanal with O2 (Eq. 67). Other formed is often pointed out as a major environ-
a-branched carboxylic acids, such as isobutyric mental drawback of this process, especially in
acid and 2-methylbutyric acid, are much smaller academic papers [173]. However, this is a very
in scale and make up, all together, less than hind-sighted view of reality. In fact, the off-gas
25
103 t/a. from an adipic acid plant is not emitted into the
32 Oxidation

atmosphere. The off-gas is usually returned to The technology is now almost exclusively
the nitric acid plant, where the NO2 is reoxi- used in China and India with only one remain-
dized by air to nitric acid. The remaining N2O is ing producer in Europe, Oxiquim SA in Spain.
inert and must be catalytically decomposed. Although the yield, which is only around 70%,
Meanwhile, almost all adipic acid plants are is not particularly good, the process is simple
equipped with an off-gas treatment reactor in and can even use agricultural waste material as
which the N2O is decomposed to N2 and O2 to feedstock.
minimize N2O emissions [174]. According to The second most important oxidation pro-
the EPA, N2O emissions from adipic acid plants cess involves the oxidation of propene with
in the USA in 2011 add up to less than 3% of nitric acid and O2 (Eq. 70) [181].
anthropogenic emissions of N2O [175] and less
than 1% of the combined natural and anthropo- 70
genic emissions of N2O [176].
Many alternative processes have been pro-
posed for the production of adipic acid, but none The process has a high yield but has lost
of these has yet achieved better economics than most of its earlier importance because of the
the nitric acid oxidation. One of the most prom- disadvantage in raw material cost. It is pres-
ising developments was the one-step oxidation ently only used in one plant from Clariant in
of cyclohexane in acetic acid developed by France, and contributes to less than 2% of the
Rhodia [177]. The process used a mixture of worldwide oxalic acid capacity.
CrIII and CoII salts as homogeneous catalysts in Small amounts of oxalic acid are also pro-
acetic acid as solvent at 105 C and 100 bar, and duced as a byproduct in the synthesis of glyox-
achieved a selectivity to adipic acid of 66% at a ylic acid by oxidation of glyoxal with nitric acid
conversion of 11%. However, the low conversion (see Glyoxylic Acid).
per pass and the high corrosivity of acetic acid Ube also produces oxalic acid by a three-
containing water led to a very high investment, step process in which the first step involves the
making the process economically unattractive. oxidation of an alcohol, either methanol or n-
In recent years the Japanese company Daicel butanol, with NO and O2 to the corresponding
developed an alternative technology for the pro- alkyl nitrite. This is subsequently carbonylated
duction of adipic acid using N-hydroxyphthali- with CO on a Pd catalyst to dialkyl oxalate and
mide as the catalyst [178]. According to a finally hydrolyzed to produce oxalic acid and
company report [179], a pilot plant with a capac- regenerate the alcohol, which can then be
ity of 30 tons per year has been built, but to date reused. The alkyl nitrite synthesis is described
no commercial plant is using this technology. in more detail in Section 2.2.13.
Oxalic Acid. Oxalic acid, with a worldwide
Dodecanedicarboxylic Acid. Dodecanedicar-
capacity of ca. 600
103 t/a, is the second
boxylic acid (! Dicarboxylic Acids, Aliphatic)
largest scale aliphatic dicarboxylic acid. Only
is the third largest scale aliphatic dicarboxylic
about half of the total capacity is produced by
acid, with an estimated world capacity of ca.
oxidation. The remaining capacity utilizes the
25
103 t/a [182]. Two different processes, both
dehydrogenation of sodium formate to sodium
involving an oxidation, are used for its produc-
oxalate [180].
tion. The most important one is the oxidation of
Several different oxidation processes are still
cyclododecanol (see Section 71) with nitric
in use, but they all have in common that they use
acid (Eq. 71).
nitric acid as the oxidant (! Oxalic Acid). The
most important process, accounting for ca. 45%
of the capacity, is the oxidation of carbohydrates,
such as sugars, hydrolyzed starches, and 71
molasses, with nitric acid (Eq. 69).

69 The technology is, in principle, very similar

to that used for the production of adipic acid
 Oxidation 33

from KA oil, including the catalyst used, but the above-mentioned long-chain dicarboxylic
with a few minor differences. Whereas in the acids from paraffins [185].
oxidation of KA oil the adipic acid remains in At first glance the technology may look
solution during the reaction, dodecanedicar- quite attractive, but it has some serious draw-
boxylic acid, due to its lower solubility in water, backs. First, the chemical yield is rather low
already precipitates in the reactor. For this because the bacteria use some of the paraffins
reason the reactor design for the production for growth. Although no exact numbers have
of dodecanedicarboxylic acid is quite different been published, the chemical yield is estimated
from that of the adipic acid process. to be only in the neighborhood of 50%. Another
As in the case of adipic acid, the main problem arises because the individual paraffins
byproducts are shorter-chain dicarboxylic required as raw material are not commercially
acids, mainly undecanedicarboxylic acid, available. Producers have to use the commer-
together with smaller amounts of shorter-chain cially available paraffin mixtures as the starting
acids. The undecanedicarboxylic acid is not material and first separate them by distillation,
isolated as a pure product, but the mixture of which can be quite costly. Furthermore, the
undecane- and dodecanedicarboxylic acids economics of the process also critically depend
obtained from the mother liquor of the dodec- on not wasting any of the paraffins contained in
anedicarboxylic acid crystallization is sold as the light wax oil. This means that the process is
such as a corrosion inhibitor. Invista, the not attractive if only one of the long-chain
worlds largest producer of dodecanedicarbox- diacids is desired. However, the process will
ylic acid sells a mixture containing ca. 40% probably continue to be used because it offers a
each of undecane and dodecanedicarboxylic way to access long-chain dicarboxylic acids
acids, ca. 8% of sebacic acid, with the remain- that are not easily available by other routes.
der being other dicarboxylic acids with less
than ten carbon atoms, under the trade name Azelaic Acid. Azelaic acid is co-produced with
Corfree M1 [183]. n-nonanoic acid (pelargonic acid) by cleavage of
In addition to the above technology, dodec- oleic acid with ozone. The only producer is
anedicarboxylic acid is also produced by oxi- Emery Oleochemicals, and an estimated
dation of paraffins isolated from light wax oil. 10
103 t/a of azelaic acid are produced by
The light wax oil is first separated by distilla- this process. Alternative technologies for oxida-
tion to obtain the pure paraffins with chain tively cleaving oleic acid have been reviewed,
lengths between C11 and C16. The pure paraf- but in conclusion the use of ozone is still the best
fins are then oxidized by aerobic fermentation method. This is particularly true if the ozonolysis
with a selected strain of Candida tropicalis to is conducted in pelargonic acid as the solvent,
produce undecanedioic, dodecanedioic, bras- because in this way the workup of the product
sylic (C13), tetradecanedioic, pentadecanedioic, mixture is considerably simplified [186].
or hexadecanedioic acid (Eq. 72). Trimethyladipic Acid. Trimethyladipic acid is
a mixture containing ca. 40% 2,2,4- and 60%
2,4,4-trimethyladipic acid (! Dicarboxylic
Acids, Aliphatic, Section 2.4.4) and is obtained
by oxidizing 3,5,5-trimethylcyclohexanol with
72
nitric acid without the use of a catalyst (Eq. 73)
[187]. The 3,5,5-trimethylcyclohexanol itself is
Although the technology was originally
obtained by hydrogenating isophorone.
developed by Cognis for the conversion of
stearic acid or other saturated fatty acids to
long-chain dicarboxylic acids, it was never
used commercially for this purpose [184].
Rather, the technology was further developed
by two Chinese companies, Cathay Industrial
Biotech Ltd. and Zibo Guangtong Chemical
Co., which are now using it to produce all 73
34 Oxidation

The trimethyladipic acid itself is not a com-

mercial product, but is used as a starting mate-
rial for trimethylhexanediamine, a cross-
linking amine sold by Evonik under the trade
75
name Vestamin TMD. The world capacity for
trimethyladipic acid is estimated to be around
4
103 t/a. The primary product after ozonolysis and
reductive workup was methyl glyoxylate
methyl hemiacetal, which was itself a commer-
Glyoxylic Acid. Glyoxylic acid (! Glyoxylic cial product. The purified methyl glyoxylate
Acid) is a versatile intermediate used mainly as methyl hemiacetal was then hydrolyzed to pro-
a building block in the synthesis of vanillin (see duce the commercial 50% aqueous solution of
Section 74), allantoin, and several antibiotics methyl glyoxal. Because the hemiacetal can be
like amoxicillin, ampicillin, and the fungicide purified by distillation, the glyoxylic acid
azoxystrobin. The world capacity for glyoxylic obtained by this route had the highest purity on
acid is estimated to be around 80
103 t/a. It is the market. However, after the production plant
now exclusively produced by the oxidation of was destroyed twice, the first time in August
glyoxal with nitric acid (Eq. 74). 2003 and again after reconstruction in August
2004, the technology was abandoned [194].
74
The cathodic reduction of oxalic acid using
lead electrodes that used to be widespread
among Chinese producers is not in use any-
The yield of glyoxylic acid is only around more, because the glyoxylic acid produced by
75% and considerable amounts of oxalic acid this process was always contaminated with
are formed as over-oxidation product [188 lead.
190]. However, the oxalic acid can be easily Methoxyacetic Acid. Methoxyacetic acid is an
separated by crystallization and sold, so over- important intermediate for the synthesis of two
oxidation does not represent a big economic major fungicides: Mefenoxam [70630-17-0 and
penalty. The reaction also produces considera- 69516-34-3] and Oxadixyl [77732-09-3]. The
ble amounts of N2O, approximately 200 kg worlds largest producer, WeylChem, uses the
N2O per tonne of glyoxylic acid as a byproduct oxidation of aqueous 2-methoxyethanol with
[191]. Plants in developed countries are O2 in the presence of a Pt catalyst to synthesize
required to install off-gas treatment units to methoxyacetic acid (Eq. 76) [195].
reduce N2O emissions. The reaction is acid-
catalyzed and thus the nitric acid plays a dual
role as both oxidant and catalyst. By using 76
aqueous hydrogen chloride as the catalyst,
the reaction can even be performed with cata-
lytic amounts of nitric acid and sodium nitrite Another process using nitric acid as the
by using O2 as the terminal oxidant [192]. oxidant has been developed, but is not used
Although this technology has definite advan- commercially [196]. It cannot be excluded that
tages because the emissions of nitrogen oxides some of the methoxyacetic acid offered com-
are considerably reduced and O2 is a cheaper mercially is derived from chloroacetic acid, but
oxidant than nitric acid, the high corrosivity of the largest part is certainly made using the
the reaction mixture and the associated high above-mentioned oxidation process.
investment costs for the production plant have, There are no reliable data on the world
up to now, prevented this technology from capacity for methoxyacetic acid, but it is esti-
becoming commercial. mated to be in the range of 500 to 1000 t/a.
In the past, glyoxylic acid was also synthe-
sized by DSM Austria (formerly Chemie Linz) Dimethylol Propionic Acid. Dimethylol pro-
by the ozonolysis of dimethyl maleate in meth- pionic acid was developed as a component for
anol (Eq. 75) [193]. waterborne dispersions. It is made by oxidation
 Oxidation 35

of dimethylol propionic aldehyde with hydro- industrially available and, if the resulting
gen peroxide without any catalyst [197]. The MnO2 is recycled, can even have a good atom
aldehyde is made by aldol condensation of economy [200], other oxidation systems were
propionic aldehyde with formaldehyde (Eq. 77) developed using cheaper oxidants. The most
successful chemical oxidation system that was
used industrially was sodium hypochlorite in
combination with nickel peroxide as the catalyst,
77
which gives yields well above 90% [201]. How-
ever, the chemical oxidation route has now been
almost completely replaced by a fermentation
The oxidation leads to the formation of route.
dimethylol propionic acid in approximately Oxidized Polyethylene Waxes. Specialty
70% yield together with a plethora of byprod- waxes are produced commercially by oxidizing
ucts. However, since the desired product can be polyethylene with O2 or air (sometimes con-
isolated by crystallization and no catalyst needs taining small amounts of ozone) in the presence
to be separated or recycled, the overall process [202] or absence [203, 204] of radical starters,
is quite simple. The major challenge of the metal catalysts [205], or carboxylic acids [206]
process is actually safety. The oxidation of at elevated temperatures, in the solid phase or in
aldehydes with H2O2 leads to the formation the polymer melt (! Waxes, Section 6.1.5).
of H2 as a byproduct [198]. Since some decom- The oxidation leads to a reduction of the aver-
position of H2O2 to oxygen also takes place at age molecular mass of the polymer, to a broad-
reaction temperature, the off-gas from the oxi- ening of the molecular mass distribution, and
dation contains both H2 and O2 and thus to the introduction of functionality, such as
requires special safety measures. The world carboxyl groups.
capacity for dimethylol propionic acid is esti- There are no published data on the world
mated to be around 10
103 t/a. An analogous capacity for oxidized polyethylene waxes, but it
reaction was also used for the preparation of the is estimated to be around 50
103 t/a.
higher homologue dimethylol butyric acid, but
production was stopped in 2009 because of low 2.2.9. Carboxylic Acids and Anhydrides,
demand [199]. Unsaturated

Ascorbic Acid. The classical synthesis Acrylic Acid. Acrylic acid is currently pro-
of ascorbic acid (! Vitamins, 7. Vitamin C duced exclusively by the two-step oxidation
(L-Ascorbic Acid)) commonly known as the of propene to acrolein and subsequently to
Reichstein synthesis, also involves a chemical acrylic acid (! Acrylic Acid and Derivatives).
oxidation step. The synthesis starts from Both oxidations are performed in the gas phase
D-glucose, which is first hydrogenated to with heterogeneous catalysts. The first step, the
D-sorbitol, which is in turn oxidized to L-sor- oxidation of propene to acrolein, is described in
bose in a fermentation with Acetobacter xyli- Section 2.2.5.
nium or Acetobacter suboxidans. The L-sorbose
79
is then protected as the diacetonide and the
diacetone-L-sorbose is chemically oxidized to
diacetone-2-keto-L-gulonic acid (Eq. 78). The acrolein produced in the first step is
usually not isolated or purified but is instead
immediately converted in a second reactor on a
different catalyst to acrylic acid (Eq. 79). The
state of the art for the processes used by the
technology leaders has been recently reviewed
78 [207].
An active research topic in the last two
In the classical Reichstein process, KMnO4 is decades has been the one-step catalytic oxida-
used as the oxidant. Although KMnO4 is tion of propane to acrylic acid (Eq. 80).
36 Oxidation

80 methacrylate in one step. The reaction is per-

formed in the liquid phase with an intermetallic
compound of palladium and lead, Pd3Pb, as the
Despite the huge research effort, there has catalyst (Eq. 82) [209211].
been little progress in developing more selec-
tive catalysts. Although this transformation
would economically be very interesting, over- 82
coming the hurdle of having a yield per pass of
acrylic acid of at least 50% still remains an
elusive goal [208].
The direct oxidation of isobutane or isobuty-
The world capacity for acrylic acid as of 2013
ric aldehyde in one step to methacrylic acid has
is estimated to be around 6
106 t/a, but capacity
also been studied, but the selectivities and con-
is growing rapidly and new plants with an addi-
versions obtained were not sufficiently attractive
tional 1.5
106 t/a of capacity are planned to
for the development of a commercial process.
come on stream before the end of 2015.
Methacrylic Acid and Methyl Methacrylate. Maleic Anhydride. Maleic anhydride (!
Although methyl methacrylate is a large-scale Maleic and Fumaric Acid) is currently pro-
commodity chemical with an estimated world duced only by the oxidation of benzene or
capacity of 3.6
106 t/a, most of the commer- n-butane in the gas phase on heterogeneous
cial routes, accounting for three quarters of the catalysts (Eq. 83). The process based on ben-
world capacity for methyl methacrylate, do not zene is considered outdated and no new plants
involve an oxidation and do not have meth- have been built in the recent past. In fact,
acrylic acid as an intermediate. Examples actually a few small plants have even converted
include the acetone cyanohydrin process and from benzene to n-butane feed.
the Lucite process based on methyl propionate
and formaldehyde. Methacrylic acid (! Meth-
acrylic Acid and Derivatives) itself is a rather
small-scale specialty with a market size in the 83
lower single-digit thousands of tonnes per year, Nevertheless, the process based on benzene
but it is an important intermediate in one of the still accounts for approximately half of the world
processes for the production of methyl meth- capacity. Modern plants use n-butane as a feed-
acrylate. The remaining quarter of the world stock, and although both fluidized-bed and
capacity is produced by the oxidation of meth- fixed-bed processes have been developed, the
acrolein. Methacrolein itself is made mostly by fixed-bed process is currently considered to be
oxidation (see Section 2.2.5). The different the state of the art [212]. The catalyst used for the
potential routes to methacrylic acid by partial oxidation of butane is based on vanadyl pyro-
oxidation have been reviewed [109]. phosphate. Since this is one of the few catalysts
The most widespread process for the oxida- that are known to activate alkanes, a lot of effort
tion of methacrolein in the gas phase uses a has been devoted to understanding its mode of
heteropoly compound as the catalyst (Eq. 81). operation, but many points have still not been
This technology accounts for an approximate fully understood [213]. The development of this
world capacity of around 1.1
106 t/a of meth- class of catalysts and of the process itself have
acrylic acid. been reviewed [214].
Small amounts of maleic anhydride are also
81 recovered as a byproduct in the production of
phthalic anhydride, but these account for less
than 1% of the world capacity of maleic anhy-
A different approach has been taken by dride. This is estimated to be ca. 2
106 t/a.
Asahi. In the Asahi process, for which a
100
103 t/a plant was built in Japan in 1998, Crotonic Acid. Crotonic acid (! Crotonalde-
methacrolein is oxidatively esterified with hyde and Crotonic Acid) is, when compared to
methanol to directly produce methyl acrylic and methacrylic acids, only a small-scale
 Oxidation 37

specialty. The successful use of many different sodium benzoate, or plasticizers. The world
catalysts, both homogeneous and heterogeneous capacity for benzoic acid is approximately
metal catalysts, strong acids, and even uncata- 400
103 t/a.
lyzed oxidations, has been claimed in the litera- The oxidation itself is a radical chain
ture. It is not known with certainty which method reaction. Its mechanism and the differences
is actually used industrially. However, an unca- in mechanism between the radical chain oxida-
talyzed oxidation, with or without the addition of tion of toluene and the radical chain oxidation
a pyridine derivative as moderator, seems to be of cyclohexane have recently been studied in
the most likely process (Eq. 84) [215, 216]. detail [217].
The toluic acids have a certain importance as
starting materials mostly for agrochemicals but
84
the overall worldwide production is probably in
the single-digit thousands of tonnes per year.
There are no reliable estimates available There are no dedicated plants to produce toluic
of the world capacity of crotonic acid, because acids and the plants can probably produce all
it can also be produced in multipurpose three isomers. Both oxidations with O2, analo-
plants. It is unlikely that the world production gous to the oxidation of toluene, [218] and with
is >3000 t/a. WeylChem, together with nitric acid [219] have been described for the
Eastman one of the major producers of crotonic preparation of the toluic acids, and both are
acid, has announced in 2000 the construction of suitable for the commercial production of the
a 1000 t/a plant in Frankfurt, Germany, that was required amounts of these products.
due to come on stream in early 2001. This plant
Terephthalic Acid. Terephthalic acid (! Ter-
is probably the largest plant worldwide for
ephthalic Acid, Dimethyl Terephthalate, and
crotonic acid.
Isophthalic Acid) is by now the largest organic
product produced by an oxidation reaction and
2.2.10. Carboxylic Acids and Anhydrides,
also the largest-scale process using homoge-
Aromatic
neous catalysis.
Benzoic Acid. Benzoic acid (! Benzoic Acid The first process used commercially for the
and Derivatives) is commercially produced by production of terephthalic acid was the oxidation
oxidation of toluene with oxygen in the liquid of p-xylene with nitric acid, but this process has
phase, usually in the presence of small amounts long since been discontinued. The oldest process
of cobalt salts as a catalyst (Eq. 85). Besides still in use is the so-called Witten or Witten
benzoic acid, benzaldehyde is also formed as Katzschmann process, which was developed
the major byproduct (see Section 2.2.5), and around 1953. In this multistage process p-xylene
benzoic acid was used in earlier days in an is partially oxidized with O2 to p-toluic acid by
oxidative decarboxylation reaction to produce using cobalt as an homogeneous catalyst. The p-
phenol (see Section 2.2.7). toluic acid is then isolated and esterified with
methanol. The methyl p-toluate is recycled to the
oxidation reactor, where it is oxidized to
monomethyl terephthalate. The monomethyl
85 terephthalate is then esterified to dimethyl ter-
ephthalate and the diester is isolated and purified
(Eq. 86). A detailed description including a flow
Currently, benzoic acid is produced only as a diagram of the most modern version of this
starting material for food preservatives, such as process has been published [220, 221].

86
38 Oxidation

In spite of the several process steps and the

additional costs for the methanol, the process
87
is still in use because it offers some advan-
tages compared to the state-of-the-art Amoco
process discussed below. First, since neither
acetic acid nor bromide is required, the oxi-
dation reactor and the workup section can be The process is one of the fastest growing,
built from standard stainless steel and do not with a total installed capacity worldwide of ca.
require expensive, highly corrosion resistant 60
106 t/a.
alloys. The second advantage is the easy puri- The chemistry of the process has been exten-
fication of the dimethyl terephthalate by dis- sively reviewed [222226]. The development of
tillation. The major use of terephthalic acid the MnCoBr catalyst system and its exten-
and dimethyl terephthalate is in the production sion as CoMnBrZr can be seen as a major
of polyesters like PET (with ethylene glycol) breakthrough in oxidation chemistry. During
or PBT (with 1,4-butanediol) by condensation. the oxidation, some of the acetic acid solvent
Since the presence of impurities with just one is also lost by oxidation, and at present, solvent
carboxyl group is the limiting factor for the make-up in terephthalic acid synthesis accounts
achievable molecular mass of the polymer, for around one-third of the worldwide demand
very stringent specifications must be met as for acetic acid.
to the purity of the terephthalic acid, and these Phthalic Anhydride. Phthalic anhydride (!
are easier to achieve by distillation of the Phthalic Acid and Derivatives) is commercially
diester. For special applications, the diester produced by the oxidation of o-xylene or naph-
also offers some advantages in the poly- thalene with air in the gas phase over VTi
condensation reaction. These advantages jus- mixed-oxide catalysts, often containing addi-
tify the survival of a considerable number of tional metals as promoters (Eq. 88) [227].
plants using this technology, which still Naphthalene, which had lost much of its impor-
accounts for ca. 3
106 t/a in capacity. Still, tance as a feedstock, is now regaining it due to
no new plants using this technology are the increased availability of cheap naphthalene
expected to be built in the near future, and from coal, especially in China.
the technology will eventually be completely
replaced by the Amoco process.
The Amoco process, earlier also known as
the Mid-Century process, was also developed
in the 1950s, but the first plant was only built
in 1965. The process uses a homogeneous
catalyst system consisting of a mixture of 88
cobalt, manganese, and bromide in acetic
acid as solvent. More recently, the catalyst The reaction is usually performed in tube-
system was extended to include zirconium as bundle reactors with molten salts as the heat-
the third metal, leading to even higher cata- transfer medium. The o-xylene conversion per
lytic activities (Eq. 87). In contrast to the pass is essentially complete and no o-xylene
Witten process, the Amoco process can recycle is required. A critical parameter for the
achieve the conversion of p-xylene to tereph- conversion is the formation of phthalide as a
thalic acid in a single step with very high byproduct. Since phthalide is very difficult to
yields, but it requires highly corrosion resist- separate from the phthalic anhydride, it is
ant construction materials. In addition, the essential that the catalysts form only tolerable
purification of the product to fiber grade is amounts of this intermediate. To achieve
much more complicated and usually involves this goal, modern processes use multilayer
a hydrogenation step to allow the separation catalyst beds with 25 layers of different cata-
of the 4-formylbenzoic acid formed as a lysts [228, 229]. The yield of phthalic anhy-
byproduct. dride based on o-xylene is usually only around
 Oxidation 39

80%. CO2 is formed as the major byproduct The installed capacity for isophthalic acid is
together with 24% of maleic anhydride, which ca. 1
106 t/a, but the plants have been under-
can also be isolated as a second valuable prod- utilized.
uct. One of the most important features of the
process is that the considerable heat of reaction Trimellitic Anhydride. Trimellitic anhydride
is generated at a temperature of ca. 350 C and (! Carboxylic Acids, Aromatic, Section 4.1)
can thus be used to produce 100 bar steam. If is produced by oxidation of pseudocumene
this steam can be used on site, it contributes a (1,2,4-trimethylbenzene) with O2 in acetic
significant margin to the economics of the acid and with the same catalyst system
process. This is at least one of the reasons described above for the production of tereph-
why the gas-phase process is still the most thalic and isophthalic acids (Eq. 90). Since
economical, even if an Amoco oxidation would pseudocumene is more difficult to oxidize
lead to the formation of phthalic acid with a and tends to form terephthalic and isophthalic
yield almost 15% higher than the gas-phase acid as undesired byproducts, it is preferable to
oxidation. conduct the oxidation stagewise [231]. The
The world capacity for phthalic anhydride is primary product of this process is the tri-
ca. 5.5
106 t/a. carboxylic acid, which is then transformed
into the monoanhydride in a subsequent dehy-
Isophthalic Acid. Isophthalic acid (! Tereph- dration step.
thalic Acid, Dimethyl Terephthalate, and Iso-
phthalic Acid) is commercially produced by the
oxidation of m-xylene with O2 in acetic acid, by
using a mixture of homogeneous cobalt, man- 90
ganese, and bromide as homogeneous catalysts
(Eq. 89). The process is essentially identical to
that used for terephthalic acid and some plants
can even be used as switch plants that can The worldwide capacity for trimellitic anhy-
produce either product. dride is estimated to be ca. 150
103 t/a.
Pyromellitic Dianhydride. Several processes
have been proposed for the production of pyro-
89 mellitic dianhydride (! Carboxylic Acids,
Aromatic, Section 4.2) but little information
is available to determine which process is actu-
A similar process which does not require the ally commercially used [232]. The oldest pro-
use of bromide and thus does not need highly cess involving the nitric acid oxidation of
corrosion resistant construction materials was durene (1,2,4,5-tetramethylbenzene) is obsolete
developed in the 1970s by the Societa Italiana and has been superseded by oxidation in the gas
Serie Acetica Sintetica (SISAS). Instead of phase with catalysts similar to those used for the
bromide, this process uses acetaldehyde as production of phthalic anhydride [233]. Instead
the oxidation booster, which is consumed in of durene, other 1,2,4,5-tetraalkyl-substituted
the process [230]. Due to the additional costs benzenes, such as tetraethylbenzene and 2,4,5-
for the acetaldehyde this process was not com- trimethylisopropylbenzene, have also been pro-
petitive and was never used commercially. posed as starting materials (Eq. 91) [234].

91
40 Oxidation

The oxidation in the gas phase has the Nicotinic Acid. Nicotinic acid (niacin) and its
advantage that it leads directly to the dianhy- amide (nicotinamide, niacinamide) are physio-
dride, which is the desired product, but the yield logically equivalent and both fall under the
is only around 60%. A better yield in the gas- general name vitamin B3. However, the acid
phase reaction, about 70%, can be achieved if and the amide are produced by different
2,4,5-trimethylbenzaldehyde is used as the routes (! Pyridine and Pyridine Derivatives,
starting material instead of durene [235]. Since Section 3.7). The amide is produced by hydrol-
both durene and 2,4,5-trimethylbenzaldehyde ysis of 3-cyanopyridine, which is itself made
are derived from pseudocumene, the former by by ammonoxidation of b-picoline (see
FriedelCrafts alkylation and the latter by acid- Section 93). The free acid is made by oxidation
catalyzed carbonylation, the higher yield in the of 2-methyl-5-ethylpyridine with nitric acid in
oxidation step makes 2,4,5-trimethylbenzalde- the so-called Lonza process [238, 239]. Nico-
hyde the most attractive feedstock. tinic acid can also be made by oxidation of
Although a considerable amount of work has b-picoline with nitric acid. BASF built a plant
been devoted to the preparation of pyromellitic in the early 1990s which used this process, but
anhydride, it is still a surprisingly low volume it was closed soon afterwards due to
specialty, with an estimated worldwide capacity unfavorable economics and technical problems
of at most 10
103 t/a. (Eq. 93).

93

It is also possible to oxidize b-picoline with

Biphenyl Tetracarboxylic Dianhydride. O2 in the gas phase by using a mixed-oxide
Biphenyl tetracarboxylic dianhydride is pro- catalyst [240] similar to that used in phthalic
duced by oxidatively coupling dimethyl phthal- anhydride synthesis, but the technology has
ate by using air and a homogeneous catalyst never been used commercially.
containing Pd, Cu, and acetylacetone to obtain The world capacity for nicotinic acid from
biphenyl tetracarboxylic acid dimethyl ester oxidation of alkyl pyridines with nitric acid is
[236], which is subsequently hydrolyzed and estimated to be ca. 40
103 t/a.
then dehydrated to obtain the desired dianhy-
dride (Eq. 92). Naphthalene-2,6-dicarboxylic Acid. Naphtha-
lene-2,6-dicarboxylic acid is produced by the
oxidation of 2,6-dimethyl-naphthalene with O2
in acetic acid as solvent with a mixture of
Mn, Co, and bromide as catalysts in a process
similar to that used for terephthalic acid
92 (Eq. 94) [241].
Although both isomers are formed, the
asymmetric one shown above predominates.
The major producer is Ube Industries in Japan,
94
which uses the product for the production of
specialty polyimides. Although no data on Historically, naphthalene-2,6-dicarboxylic
capacity have been published, it is estimated acid was also produced by oxidation of 2,6-
to be at least 1000 t/a. diisopropylnaphthalene, which was derived
The symmetrical isomer is also produced from the alkylation of naphthalene with
commercially, but on a much smaller scale, propene. Since the development of a viable
by reductive coupling of 4-chlorophthalic synthesis of 2,6-dimethylnaphthalene from o-
acid dimethyl ester [237]. xylene and butadiene by Amoco, all
 Oxidation 41

naphthalene-2,6-dicarboxylic acid is now pro- capacity, mostly located in China, uses the
duced by oxidation of 2,6-dimethylnaphtha- addition of acetic acid to acetylene.
lene [242].
In the early 1990s there was hope for a large 96
growth for naphthalene-2,6-dicarboxylic acid
because of the good barrier properties of poly
(ethylene naphthalate), but other improvements The reaction is performed in the gas phase at
to the barrier properties of the standard poly moderate pressures (<8 bar) and temperatures
(ethylene terephthalate) turned out to be more (<200 C), uses a heterogeneous catalyst, and
economical. For this reason the capacity growth gives excellent selectivities, usually greater
remained well below expectation and the world than 96% [249]. Interestingly, although Pd
capacity for naphthalene-2,6-dicarboxylic acid alone can catalyze the reaction, the catalysts
is estimated to be only about 30
103 t/a. used commercially also contain Au. The addi-
Naphthalene-1,8-dicarboxylic Anhydride. tion of Au leads to a very significant increase in
Naphthalene-1,8-dicarboxylic acid anhydride the catalytic activity [250]. This was the first
is produced by oxidation of acenaphtene with and remains the most important use of a gold-
O2 in the gas phase by using vanadium oxide containing catalyst in industry [251].
based catalysts (Eq. 95) [243245]. The estimated world capacity for vinyl ace-
tate is ca. 7
106 t/a, of which about 5.5
106 t/a
is produced by oxidative acetoxylation of
ethylene.
95

Allyl Acetate. Allyl acetate (! Allyl Com-

pounds, Section 3.2.1) is produced by oxidative
There are no exact data on the world capacity acetoxylation of propene with acetic acid in the
for naphthalene-1,8-dicarboxylic acid anhy- gas phase by using Pd catalysts similar to those
dride, but the largest producer in China disclosed used for vinyl acetate synthesis (Eq. 97) [252].
a capacity of 4500 t/a for its dedicated produc- In fact, some of the commercial plants used for
tion plant. However, it is estimated that the world the production of allyl acetate are former vinyl
capacity is actually at most 5000 t/a [246]. acetate plants which have been converted to
Interestingly one of the main uses for naph- produce allyl acetate instead.
thalene-1,8-dicarboxylic anhydride is in the pro-
duction of perylene dyes, and these also involve
an oxidation of the corresponding imide. This is
usually performed in two steps: first, a base-
97
catalyzed dimerization to the leuco dye, fol-
lowed by oxidation with O2 and/or H2O2 to Compared to vinyl acetate, allyl acetate is
form the fully aromatic perylene system [247]. only a relatively small scale product, with an
estimated world capacity of ca. 600
103 t/a.
2.2.11. Esters, Lactones, and Carbonates
Vinyl Acetate. Vinyl acetate (! Vinyl Esters, But-2-ene-1,4-diol Diacetate. A similar oxida-
Section 2.2.3) is currently produced by two tive acetoxylation process was also developed
different processes [248]. The major process, by Mitsubishi for the synthesis of but-2-ene-
which accounts for approximately 80% of the 1,4-diol diacetate, but in this case the reaction is
world capacity, involves the oxidative acetox- performed in the liquid phase at ca. 100 C and
ylation of ethylene (Eq. 96). The remaining 60 bar (Eq. 96) [253].

98
42 Oxidation

The process uses a Pd catalyst doped with Recently, efforts have been made to develop
Te. The selectivity to the desired but-2-ene-1,4- a caprolactone process based on a catalytic
diol diacetate based on butadiene is only ca. oxidation using H2O2 as the oxidant and with-
88%, with the isomeric but-1-ene-3,4-diol diac- out the need to make peracetic acid, but no
etate as the major byproduct. commercial process has yet evolved from these
This technology has come under consider- efforts [255].
able pressure recently due to the high costs The world capacity for caprolactone by oxi-
for the butadiene raw material, especially in dation is estimated to be ca. 60
103 t/a.
Asia. This has even led to the closing of
BASFs plant in Ulsan, South Korea in 2008.
Currently it is estimated that the world capac- Dimethyl Carbonate. Dimethyl carbonate (!
ity for but-2-ene-1,4-diol diacetate has fallen Carbonic Esters, Section 3.3) is an important
to only 100
103 t/a. intermediate in the non-phosgene process for
producing polycarbonate resins (! Polycar-
bonates). In such a phosgene-free process,
Caprolactone. Caprolactone is mostly pro- dimethyl carbonate can be generated by one
duced industrially by the BaeyerVilliger oxi- of several routes, of which only one involves
dation of cyclohexanone with equilibrium an oxidation reaction [256]. It is formed as a
peracetic acid. The process was originally valuable byproduct in the carbonylation of
developed by Daicel using 60 wt% aqueous methyl nitrite to dimethyl oxalate (see Section
hydrogen peroxide as the oxidant, and shifting 100) or on purpose by transesterification of
the equilibrium of the peracetic acid formation ethylene carbonate, itself made from ethylene
by separating the water formed by using an oxide and CO2, with methanol. The only
entrainer like ethyl acetate [254]. To circum- process for producing dimethyl carbonate
vent the necessity of thermally stressing the which involves an oxidation is the oxidative
unstable peracetic acid to separate water, Sol- carbonylation of methanol using copper
vay developed a similar process using 87 wt% chloride as the catalyst, developed by Eni-
aqueous hydrogen peroxide as the oxidant. Chem and now owned by Versalis (Eq. 100)
Using the highly concentrated hydrogen perox- [257].
ide shifts the equilibrium towards peracetic
acetic and the thus-obtained solution can be
used directly in the BaeyerVilliger oxidation
of cyclohexanone (Eq. 99).
100

The reaction is performed in the liquid

phase with CuCl2 as the catalyst in suspension
at 120 C and ca. 25 bar [258]. The reaction is
99 very selective, but the reaction section must be
built of highly corrosion resistant materials
like tantalum, because the combined effects of
copper ions, chloride, HCl, and oxygen leads
to severe pitting corrosion. This leads to high
costs for investment and maintenance.
Solvays caprolactone plant in the UK was For many years Eni operated a 12
103 t/a
sold in 2007 to Perstorp. After Dows decision plant using this process, but the plant was
to leave the caprolactone market, Perstorp and closed in 2005 due to poor profitability. Versalis
Daicel are the only two producers using the awarded licenses for two plants using their
BaeyerVilliger oxidation technology. BASF technology, with a combined capacity of
also produces caprolactone, but by using 6- approximately 150
103 t/a. The largest one
hydroxyhexanoic acid contained in the byprod- in Spain from GE Plastics was still operating
uct streams of the cyclohexane oxidation as the in 2002, but it is not clear whether it is currently
raw material. still in operation.
 Oxidation 43

2.2.12. Nitriles the SOHIO process, an ammoxidation of

propene with ammonia and O2 in the gas phase
Hydrogen Cyanide. Hydrogen cyanide (! with a heterogeneous catalyst in fluidized bed
Cyano Compounds, Inorganic) is both pro- (Eq. 102).
duced on purpose and obtained as a byproduct
in the ammoxidation of propene to produce
acrylonitrile. In this section only the on-pur-
pose processes are discussed. The most impor- 102
tant process for the direct synthesis of hydrogen
cyanide is the Andrussow process, which A huge amount of work has been devoted to
amounts to a catalytic ammoxidation of meth- development of more active and more selective
ane (Eq. 101) [259, 260]. catalysts. The newest-generation catalysts are
based on complex multiphase, multicomponent
Pt cat:
CH4 NH3 1:5 O2 ! HCN 3 H2 O 101 molybdates (MCM) or antimonates (MCA).
With such catalysts, yields of acrylonitrile based
The reaction is performed in the gas phase on propene can be as high as 80% [261, 262].
above 1000 C and near atmospheric pressure. It Approximately 100 kg of hydrogen cyanide is
uses a Pt/Rh gauze as the catalyst at very short formed per tonne of acrylonitrile as a valuable
contact times on the order of milliseconds. The byproduct. Another important byproduct, which
Andrussow process accounts for approximately can also make a substantial economic contribu-
half of the world capacity for hydrogen cyanide. tion to the process, is acetonitrile (see below).
Hydrogen cyanide is also formed as a A lot of development work has also been
byproduct in the acrylonitrile synthesis. The dedicated to the development of an acrylonitrile
amount of hydrogen cyanide formed can be synthesis based on propane instead of propene.
tuned to a certain extent, but it is usually at This is economically attractive due to the lower
least 10% of the acrylonitrile output or even cost of the propane feedstock. The intensive
considerably higher in propane-based acryloni- research work has led to the development of
trile processes. The amount of hydrogen cya- MoVNb(Te,Sb)O catalysts that can give acry-
nide can be increased either by adjusting the lonitrile selectivities of up to 70% based on
reaction conditions, or by co-feeding methane. propane (Eq. 103) [263].
The recovery and use of hydrogen cyanide is an
important economic factor for an acrylonitrile
producer, and often HCN consumers, like for
103
example methyl methacrylate plants, are built
close to large acrylonitrile plants. In such cases The first plant using propane as the feed-
it can make sense to deliberately increase the stock, a 200
103 t/a plant from PTT Asahi
amount of hydrogen cyanide formed in order to Chemical in Thailand, was started in 2012
provide the required capacity for the down- but took over one year to achieve stable opera-
stream plant using it. tion. It is said to be only economically viable
The hydrogen cyanide formed as a byprod- due to strongly subsidized propane prices [264].
uct in the acrylonitrile accounts for approxi- The world capacity for acrylonitrile is
mately one third of the world capacity. The increasing rapidly and will soon exceed
remaining capacity uses processes which do not 7
106 t/a, of which only 200
103 t/a is based
involve an oxidation, like the BMA process, the on propane as feedstock.
Shawinigan or Fluohmic process, or the cleav-
age of formamide. Acetonitrile. Presently no dedicated process
The total world capacity for hydrogen cya- exists for the production of acetonitrile. The
nide production by oxidation processes is esti- entire world consumption is satisfied by isolat-
mated to be ca. 2
106 t/a. ing the acetonitrile formed as a byproduct of
acrylonitrile. In the conventional SOHIO pro-
Acrylonitrile. Acrylonitrile (! Nitriles) is cess using propene as the feedstock, approxi-
currently almost exclusively produced by mately 2530 kg acetonitrile is produced per
44 Oxidation

tonne of acrylonitrile. The process based on Benzonitriles. The parent compound, benzoni-
propane as the feedstock produces approxi- trile, is produced industrially by the ammox-
mately twice that amount. Especially for the idation of toluene, but compared to other
process based on propane, the isolation of aromatic nitriles it is only a very small scale
purified acetonitrile makes an important con- product, with an estimated world capacity of
tribution to the profitability. Hence, Asahi only ca. 5
103 t/a [272].
Kasei, shortly after starting its 200
103 t/a 3-Cyanopyridine or nicotinonitrile is also
acrylonitrile plant in Thailand, announced a produced by ammoximation of b-picoline and
new plant for the isolation of purified used in the production of niacin. The estimated
acetonitrile in South Korea with a capacity of capacity for nicotinonitrile is around 20
103 t/a.
11
103 t/a [265]. Other substituted benzonitriles, such as 2-
During the economic crisis in 2009, the chloro and 2,6-dichlorobenzonitrile are also
acrylonitrile plants had to be shut down or produced by ammoxidation of the correspond-
operated at reduced capacities, leading to an ing substituted toluenes, but only 2-chloroben-
extreme shortage of acetonitrile on the market. zonitrile, with an estimated world capacity of
In the meantime, the problem can be tackled by 3
103 t/a, is of industrial importance [273].
co-feeding ethanol to the ammoxidation reactor
in the SOHIO process, thus allowing an 2.2.13. Compounds Containing NO Bonds
increase in the production of acetonitrile with-
out having to increase the output of acrylonitrile Cyclohexanone Oxime. Conventionally, cyclo-
[266, 267]. hexanone oxime is synthesized industrially on a
The current world capacity for the isolation very large scale (ca. 5
106 t/a) by the reaction
of purified acetonitrile is estimated to be around of cyclohexanone with hydroxylammonium sul-
80
103 t/a, which is less than half of the fate. However, this reaction leads to the produc-
theoretical amount of acetonitrile formed as a tion of considerable amounts of ammonium
byproduct. sulfate as byproduct, because the sulfate must
be trapped by neutralization with ammonia.
Phthalodinitriles. Both o-phthalodinitrile and This limitation could finally be overcome in
m-phthalodinitrile are produced commercially a process developed by Enichem, commonly
by ammoxidation of the corresponding xylenes. known as the ammoximation process. In this
process hydroxylamine is generated in situ by
oxidizing ammonia with hydrogen peroxide in
104
the presence of a titanium silicalite (TS-1) as
the catalyst (Eq. 105). The hydroxylamine
formed is immediately trapped by the cyclo-
The reaction is performed in the gas phase hexanone present in solution thus preventing its
over a supported vanadium catalyst at temper- further oxidation [274].
atures between 400 and 500 C in a fluidized-
bed reactor or a fixed-bed reactor at ambient
pressure (Eq. 104) [268, 269]. Although both
phthalodinitriles are produced by using essen-
tially the same process, their uses are very
different. While o-phthalodinitrile is almost 105
exclusively used for the synthesis of phthalo-
cyanine dyes (! Phthalocyanines), m-phthalo- The reaction is performed in wet tert-buta-
dinitrile is almost exclusively hydrogenated to nol as the solvent with a suspended catalyst.
m-xylylenediamine, which is mainly used for The process is characterized by high conver-
the production of Nylon-MXD6 [270]. sions (>98% for cyclohexanone and essentially
The estimated world capacity for o-phthalo- complete for hydrogen peroxide) and high
dinitrile is estimated to be only ca. 10
103 t/a, selectivities to the oxime (99.6% based on
while the capacity for the m somer is consider- cyclohexanone and 89% based on hydrogen
ably larger at ca. 60
103 t/a [271]. peroxide). No information is given on the
 Oxidation 45

selectivity based on ammonia, but this is Nitromethane, Nitroethane, and Nitropro-

expected to be higher than the selectivity based panes. A large proportion of the commercially
on hydrogen peroxide, because the main side produced nitromethane (! Nitro Compounds,
reactions consuming hydrogen peroxide are the Aliphatic) is obtained from the reaction of
overoxidation of ammonia to ammonium nitrate dimethyl sulfate with sodium nitrite and does
(2 NH3 4H2O2 ! (NH4)NO3 5 H2O) and its not involve an oxidation reaction [283]. It is
decomposition to oxygen (2 H2O2 ! 2 H2O also possible to obtain nitromethane from the
O2). reaction of methane with nitric acid in the gas
Enichem built a 12
103 t/a semi-commer- phase (Eq. 107). In contrast to the nitration of
cial plant in Porto Marghera, Italy [275] and aromatics, which are electrophilic aromatic
sold the technology to Sumitomo in 2000 [276]. substitutions, the nitration of alkanes is a radi-
Sumitomo closed down the plant in Italy and cal chain reaction [284], and either nitric acid or
built a full-scale plant with a capacity of mixtures of NO2 and O2 can be used as the
67
103 t/a in Japan using this technology oxidant [285]. It is not clear whether the nitra-
[277]. This plant combines the ammoximation tion of methane is currently used commercially,
technology of Enichem with Sumitomos own but most probably it is not.
technology for the gas-phase Beckmann A mixture of all possible C1C3 nitroalkanes
rearrangement to achieve, for the first time, a is produced commercially by Angus Chemicals
caprolactam process without any ammonium using the gas-phase nitration of propane [286].
sulfate as byproduct.
Degussa has developed a similar process for
the ammoximation of cyclododecanone to
cyclododecanone oxime [278], but this process
has not yet been commercialized.
Methyl Ethyl Ketone Oxime. Next to cyclo-
107
hexanone oxime and cyclododecanone oxime,
methyl ethyl ketone oxime is the only oxime
The reaction is performed in the gas phase at
with significant commercial relevance. It is
390440 C and leads to a mixture of all possi-
used as an antiskinning agent in paints [279].
ble nitroalkanes. By using O2 or a halogen as
Recently, a plant for the production of
homogeneous catalyst the product distribution
methyl ethyl ketone oxime with a capacity of
can be changed within broad limits. The pro-
15
103 t/a using the ammoximation technol-
cess is very demanding from the safety point of
ogy was built in China. However, the capacity
view, and the Angus plant in Sterlington, Loui-
indicated has to be taken with some caution,
siana was almost completely destroyed in an
because the entire market for methyl ethyl
accident in May 1991 [287], but it was rebuilt
ketone oxime has recently been estimated to
and is still in operation. A second Angus plant is
be only 6
103 t/a [280] and furthermore
in operation at its site in Germany. W. R. Grace
because methyl ethyl ketone oxime is under
also operated a plant for nitroalkanes in Deer
pressure because of its high volatility and its
Park, Texas using a mixed feed of propane and
unfavorable toxicological profile, which have
ethane, but production ceased in 1992.
led to efforts to find substitutes [281].

106

In principle, the technology is similar to the Recent development work has been dedi-
Enichem/Sumitomo technology described in cated to increasing the selectivity to 2-nitro-
the previous section, but it uses a different propane, which is the fastest growing and most
titanium zeolite as the catalyst: Ti-MWW valuable product, mainly used in the production
instead of TS-1 (Eq. 106) [282]. of 2-methyl-2-aminopropan-1-ol. By using
46 Oxidation

lower temperatures, the selectivity to 2-nitro- and is similar to that developed by Ube for the
propane can be pushed to above 80% [288]. synthesis of oxalates. It uses methyl nitrite as the
There are no published data on the capacity key intermediate [295]. The methyl nitrite is
of the nitroalkane plants, but this is estimated to produced by the reaction of methanol, NO and
be ca. 50
103 t/a. O2 as shown below in Equation (109).
Nitrocyclohexane. DuPont developed and
2 MeOH 2 NO 1=2 O2 ! 2 MeONO H2 O 109
operated a process for the synthesis of nitro-
cyclohexane from cyclohexane, the so-called This reaction is very fast, very selective,
Nixan process, as an intermediate in a capro- does not require a catalyst and can be per-
lactam process [289]. The reaction is usually formed both in the gas and in the liquid phase.
formulated as a nitration, but it is actually a Mechanistic work shows that the key step is a
radical chain reaction, and nitric acid [290], concerted reaction between methanol and
butyl nitrite [291], or nitric oxide [292] can be N2O3 that, depending on the N2O3 isomer,
used as oxidants (Eq. 108). is computed to have an activation energy as
low as 20 kJ/mol [296].
108
Six plants for ethylene glycol, each with a
capacity of 200
103 t/a have come on stream
since 2009 and more are either in planning or
A nitrocyclohexane plant with a capacity of already under construction. Since, per mole of
25
103 t/a using 60% nitric acid as the oxidant ethylene glycol 2 mol of methyl nitrite are
was put into operation by DuPont in Beaumont, required, the existing plants already add to a
Texas in 1963, but it was closed shortly after in total methyl nitrite capacity of ca. 2.5
106 t/a.
1967 due to poor profitability. Since the closing Amine N-Oxide Surfactants. N,N-Dimethyl
of the DuPont plant nitrocyclohexane has lost fatty amine N-oxides are important nonionic
commercial significance and is not produced on surfactants which are especially valued for their
a large scale anymore. foam-stabilizing properties and widely used in
Methyl Nitrite. Methyl nitrite is a versatile dishwashing formulations (! Surfactants,
intermediate, but due to its instability it is seldom Section 7.5). They are prepared by oxidizing
used outside of an industrial setting and even N,N-dimethyl fatty amines with aqueous hydro-
there, it is never isolated in pure form. In the past gen peroxide (Eq. 110). Both pure amines like N,
it has been used on a small scale (<103 t/a) for N-dimethyl lauryl amine and mixtures with
the oxidation of acetone to methylglyoxal chain lengths between C12 and C18 are used
dimethyl acetal [293]. Its main use is, however, commercially to produce a wide range of prod-
in the indirect oxidative carbonylation of alco- ucts.
hols, whereby, depending on the reaction condi-
tions and catalyst used, dimethyl carbonate or 110
dimethyl oxalate can be obtained as the major
product. Although the carbonylation of methyl
nitrite to dimethyl carbonate has been exten- Usually no catalyst is required to perform
sively studied, no commercial process ever the reaction [297], but chelating agents such as
materialized and the work essentially stopped the sodium salt of ethylenediamine tetraacetic
in the mid-1990s [294]. The carbonylation of acid can be added to suppress H2O2 decompo-
alkyl nitrite to dialkyl oxalate has been devel- sition [298], and other additives can be used to
oped by Ube, which has used this technology reduce the formation of nitrosamines as
since 1978 in a 6000 t/a plant for oxalic acid. byproducts [299, 300].
Recently, this technology is experiencing tre- The world capacity for N,N-dimethyl fatty
mendous growth, especially in China, since it amines is estimated to be ca. 75
103 t/a.
allows for the production of ethylene glycol from Another different class of amine N-oxide
coal-derived CO. The technology used in China surfactants are the cocamidopropyl amine
was developed at the Fujian Institute of Matter oxides [301]. These are prepared by treating
 Oxidation 47

a fatty acid, usually a mixture of fatty acids with N-Methylmorpholine N-Oxide. N-Methyl-
chain lengths between C12 and C18 carbon morpholine N-oxide is a good solvent for cel-
atoms, with N,N-dimethyl-1,3-propanediamine lulose [307, 308] and is used in the production
to the corresponding amide, and subsequently of cellulose fibers such as Courtaulds Tencel.
oxidizing the tertiary amino group to the N- The N-methylmorpholine N-oxide, which is
oxide with H2O2 (Eq. 111). commercially available as a 50% solution in

111

The world capacity for cocamidopropyl water, is prepared by the uncatalyzed oxidation
amine oxides is estimated to be ca. 5
103 t/a. of N-methylmorpholine with aqueous hydrogen
peroxide (Eq. 113) [309, 310].
Triethylamine N-Oxide. Triethylamine N-
oxide is an intermediate in the synthesis of
N,N-diethylhydroxylamine. It is obtained by
oxidizing triethylamine with aqueous hydrogen 113
peroxide in the presence of wolframate as
catalyst [302]. The triethylamine N-oxide
formed is then pyrolyzed to induce a Cope
It has been claimed that performing the
elimination and form the desired N,N-diethyl
oxidation under a CO2 atmosphere leads to a
hydroxylamine and ethylene (Eq. 112).

112

Although there would be a considerable considerable reduction of the N-nitrosomor-

potential for cost reduction if diethylamine pholine content of the product.
could be directly oxidized to the hydroxyl- The world market for N-methylmorpholine-
amine, this turns out to be quite a difficult N-oxide is estimated to be ca. 3
103 t/a.
reaction. The oxidation of diethylamine with
H2O2 in the presence of a homogeneous cata- 2.2.14. Chlorinated Compounds
lyst leads mainly to the formation of the corre- Ethylene Dichloride. Ethylene dichloride or
sponding nitrone [303]. Heterogeneous 1,2-dichloroethane is an intermediate in the
catalysts like TS-1 can give much better results, production of vinyl chloride, which consumes
but acceptable selectivities can only be roughly 99% of all ethylene dichloride produced
achieved at partial conversion of diethylamine (! Chloroethanes and Chloroethylenes, Section
and some nitrone is always formed as a byprod- 2.3) Two routes are currently used to produce
uct [304306]. For this reason, the classical ethylene dichloride, for which several process
two-step synthesis of diethyl hydroxylamine variants exist. The first one is the Lewis acid-
is still the method used, not only due to its catalyzed addition of chlorine to ethylene, and
simplicity, but also because it gives the purest the second one is the oxychlorination of ethylene
product with only minimal work-up. with HCl and O2 [311]. Only the second route
The world capacity for diethylhydroxyl- falls under the scope of this keyword (Eq. 114).
amine is estimated to be ca. 6
103 t/a, which
would put the capacity for triethylamine N-
oxide at ca. 8
103 t/a. 114
48 Oxidation

The direct chlorination of ethylene with Cl2 performed in the gas phase by using copper(II)
is typically conducted in the liquid phase with chloride supported on alumina as the catalyst
iron(III) chloride catalysts and ethylene (Eq. 115) [315317], but many other variations
dichloride as solvent. In the oxychlorination of the process, both in the liquid and the gas
reaction, catalysts based on copper(II) chloride phase, have been claimed [318].
supported on alumina are used. The reaction is
conducted in the gas phase in a fixed-bed or
fluidized-bed reactor [312]. Although at first 115
glance these appear to be unrelated processes,
they are in practice usually operated jointly to
create what is called a balanced process.
Indeed, when ethylene is chlorinated with chlo- At the peak of its commercial use in the
rine and subsequently cleaved to produce vinyl 1960s the capacity for benzene oxychlorination
chloride, 1 mol of HCl is also produced. In a reached almost 200
103 t/a, but the develop-
balanced process, this HCl is used for the oxy- ment of the cumene-based phenol process and
chlorination of ethylene and to produce more the decay in the demand for chlorobenzene led
ethylene dichloride. With this strategy one can to the closure of all plants using this process.
avoid the production of waste HCl, which has
Chloroacetyl Chloride. Chloroacetyl chloride
little or no value. Theoretically, in a perfectly
(! Chloroacetic Acids, Section 2.6.2), an
balanced plant, half of the ethylene dichloride
important building block for agrochemicals,
output would stem from chlorination and the
is mostly produced by chlorination of chloro-
other half from oxychlorination. In practice this
acetic acid with phosgene, thionyl chloride, or
ratio is usually not 1:1, mainly because chlorine
PCl3. However, Dow at least still uses the
is more expensive than HCl, and waste HCl is
photooxidation of 1,2-dichloroethylene with
often available from other processes, for exam-
O2 in the presence of radical initiators such
ple from isocyanate plants. The use of the waste
HCl from other processes minimizes the con- as Cl2 or Br2 (Eq. 116) [319].
sumption of chlorine and contributes to the
overall economics of the process.
116
The worldwide capacity for ethylene
dichloride is estimated to be ca. 46
106 t/a,
of which only approximately half is produced There are no reliable data on the capacity of
by oxychlorination and the other half by chlo- the chloroacetyl chloride plant that uses the
rine addition. oxidation process, but it is probably not more
The growth rates for this technology are than 1000 t/a. In the long run and with increas-
expected to be low because vinyl chloride ing availability of cheap and pure chloroacetic
production will shift from ethylene to acety- acid, the oxidation process will probably not be
lene, especially in China, where cheap acety- competitive with the chlorination route.
lene from coal is available. This has been made
possible by the development of new mercury- Trichloroacetaldehyde Hydrate. Trichloroa-
free catalysts based on gold for the addition of cetaldehyde hydrate (! Chloroacetaldehydes,
HCl to acetylene [313, 314]. Chap. 4), often known as chloral hydrate, is
produced commercially either by the chlorina-
Chlorobenzene. Chlorobenzene (! Chlorin- tion of acetaldehyde [320] or by the oxidation
ated Benzenes and Other Nucleus-Chlorinated of ethanol with chlorine (Eq. 117) [321].
Aromatic Hydrocarbons, Section 2.2.2) is now Although in the past, acetaldehyde was the
exclusively produced by electrophilic aromatic preferred starting material, the ever-increasing
chlorination of benzene, but historically the price of acetaldehyde and its lower availability
oxychlorination of benzene with HCl and O2 have caused a shift to ethanol as the preferred
was also used commercially as the first step in feedstock. All the processes for the production
the RaschigHooker phenol process. The orig- of chloral hydrate have recently been reviewed
inal oxychlorination in the Raschig process was [322].
 Oxidation 49

amine
117 2 CH3 SH S! CH3 SSCH3 H2 S
CH3 SSCH3 4 HNO3 ! 2 CH3 SO3 H 3 NO NO2 H2 O

The main uses of chloral hydrate are in the 5

3 NO NO2 2 H2 O O2 ! 4 HNO3 119
2
synthesis of the insecticide DDT, as a pharma-
ceutical drug, and as an additive for Ni electro- The nitrogen oxides formed are then reoxi-
plating. The use of DDT has significantly dized in a separate step with air to regenerate
decreased since the 1960s, when its environ- nitric acid. Since in this process no N2O is
mental persistence was recognized and its use formed, the nitrogen oxides can be fully reoxi-
drastically reduced. However, DDT is still the dized to nitric acid. The nitric acid can thus be
insecticide of choice for emergency use in the seen as a carrier for O2 which is effectively the
control of disease vectors like the Anopheles terminal oxidant in this process.
mosquito, which transmits malaria. There are BASF has recently announced a capacity
no reliable figures available for the current expansion for the nitric acid-based process to
capacity for chloral hydrate, but it is estimated 30
103 t/a [327]. There are no published capac-
to be about 5000 t/a. ities for the chlorine-based methanesulfonic acid
process, but the capacity is estimated to be ca.
2.2.15. Organosulfur Compounds 25
103 t/a.
Paraffin Sulfonic Acids. Paraffin sulfonic
Methanesulfonic Acid. Methanesulfonic acid
acids, or more precisely their sodium salts, also
(! Sulfonic Acids, Aliphatic, Section 2.6.2)
known as secondary alkane sulfonates (SASs)
has traditionally been synthesized by the
are a class of easily biodegradable anionic sur-
hydrolysis of methanesulfonyl chloride. The
factants (! Surfactants, Section 6.2.3). Two
oldest process for the synthesis of methanesul-
different processes are commercially used for
fonyl chloride was the photochemical chloro-
their production, both starting from paraffin
sulfonation of methane with a mixture of SO2
mixtures with chain lengths between C12 and
and Cl2 [323, 324]. Due to the high energy
C18. The first process is the photochemical
consumption, severe corrosion problems, and
sulfochlorination of paraffins or Reid reaction,
the problems associated with the formation of
with subsequent hydrolysis of the primarily
considerable amounts of chloromethanes as
formed sulfonyl chlorides (Eq. 120) [328, 329]
byproducts, the process is no longer in use,
the last plant having been closed in 1993. The
photochemical sulfochlorination was super- 120
seded by the oxidation of methyl mercaptan
with Cl2 in the presence of aqueous HCl
(Eq. 118) [325]. The second process, the photochemical sul-
foxidation of paraffins, is from a mechanistic
CH3 SH 3 Cl2 2 H2 O ! CH3 SO2 Cl 5 HCl 118 point of view closely related to the Reid reaction,
This process is still used commercially and since both are photochemically initiated radical
is still the preferred process for the production chain reactions and, as such, both must be
of methanesulfonyl chloride in spite of the large stopped at low conversion of the paraffin, usually
amounts of waste HCl produced as a byproduct. less than 5%, to avoid the formation of disulfonic
Recently, a new process that does not require acids (Eq. 121). However, compared to the sul-
the use of chlorine as the oxidant has been fochlorination, the sulfoxidation has a number of
developed and commercialized by BASF. advantages: there are no serious corrosion prob-
This process also uses methyl mercaptan as lems, the reaction leads in one step to the final
the starting material, which is oxidized in a product and it delivers surfactants that are free of
first step with elemental sulfur and an amine chlorinated impurities [330].
catalyst to dimethyl disulfide. The dimethyl
disulfide is then oxidized in a second
step with nitric acid to methane sulfonic acid 121
(Eq. 119) [326].
50 Oxidation

Capacities for both processes have been 123

published [331], but since then several plants
have ceased production. Currently, the capacity
for the sulfoxidation process is only 50
103 t/a, The capacity for dimethyl sulfone is much
while the capacity for the sulfochlorination is lower than the capacity for dimethyl sulfoxide
somewhat smaller at approximately 35
103 t/a. and it is estimated to be only ca. 8
103 t/a.
Methyl ethyl sulfone and ethyl isopropyl
Dimethyl Sulfoxide and Dimethyl Sulfone.
sulfones are also produced commercially, for
Dimethyl sulfoxide is presently, next to aceto-
instance by Sumitomo Seika, probably through
nitrile, one of the major specialty solvents,
the oxidation of the corresponding sulfides and
mainly due to its exceptional properties as a
find use as electrolyte solvents, but these are
polar aprotic solvent with a low volatility and
probably only small-scale specialties.
low toxicity (! Sulfones and Sulfoxides). All
the processes proposed in the literature for the
synthesis of dimethyl sulfoxide use the oxida- 3. Thermodynamics; Stability
tion of dimethyl sulfide, and many different
oxidants ranging from nitric acid to ozone,
of Hydrocarbons
oxygen, hydrogen peroxide, and even electro-
Under normal conditions in the presence of O2,
chemical oxidation, can be used. However,
CO2 is the only thermodynamically stable car-
most probably only the oxidation with NO2
bon compound; at higher temperatures CO also
is really used industrially (Eq. 122) [332, 333].
becomes increasingly stable. Thus, in the pres-
ence of free O2, all organic materials (living
organisms, wood, coal, oil, synthetic polymers,
122 solvents, etc.) are only metastable intermedi-
ates on the way to CO2. The stability of all
organic compounds in the presence of O2
and hence the yield of desired products in
The NO formed can be reoxidized with O2 oxidations therefore depends only on the
either in a separate step, in which case NO2 is kinetics of the reaction involved, i.e., on the
the oxidant, or in situ, in which case NO2 can be prevailing conditions (temperature, effect of
used in catalytic amounts and O2 is the terminal high-energy radiation, nature and concentration
oxidant. of catalysts or stabilizers (inhibitors) present,
More recently, the oxidation of dimethyl presence of other reaction partners, impurities,
sulfide with hydrogen peroxide by using tita- etc.).
nium silicalite (TS-1) as the catalyst has Table 2 lists roughly estimated rates of
been claimed. The reaction has excellent yields, typical reactions, and the linearly extrapolated
but does not seem to be in commercial use time required for complete conversion.
(Eq. 123) [334].
The production capacity for dimethyl sulf-
oxide has been growing rapidly, especially in 4. Reaction Types
China, and it is now believed to be ca. 95
103 t/
a, although the output is probably considerably The oxidation of organic materials can be
lower [335, 336]. roughly divided into two groups:
A small fraction of the dimethyl sulfoxide
produced is further oxidized to produce Homolytic oxidation reactions are chain
dimethyl sulfone. Although there is very little reactions of organic and inorganic compounds
reliable information about the process used with oxygen involving radicals formed by
commercially for the production of dimethyl homolytic cleavage of interatomic bonds.
sulfone, it appears most likely that hydrogen Explosions and combustion reactions, ageing
peroxide is used as the oxidant, since the of polymers and oils, and many industrial oxi-
reaction is clean and easy to perform, not dation reactions in the homogeneous liquid and
even requiring a catalyst [337]. gas phases belong to this group. All types of
 Oxidation 51

Table 2. Rates u of various oxidation reactions and linearly extrapolated time t required for complete formula conversion

u a, u b, tc Examples of reactions
molecules cm3  s1 mol L1 s1

1031010 10181011 1010103 a initial rate of uncatalyzed hydrocarbon oxidations

10131014 108107 2 a10 weeks degradation of unstabilized plastics under irradiation
1016 105 ca. 20 h industrial radical reactions
(e.g., liquid-phase oxidation and polymerization)
1018 103 ca. 10 min heterolytic reactions
(e.g., epoxidation with ROOH/Mo complex)
1020 101 ca. 6 s heterolytic reactions
heterogeneously catalyzed hydrocarbon oxidations
10 22
10 ca. 102 s heterogeneously catalyzed hydrocarbon oxidations
1024 103 ca. 104 s explosions;
ion reactions
a
Pure hydrocarbons contain ca. 6
1021 molecules per cubic centimeter.
b
1 mol/L corresponds to ca. 6
1020 molecules per cubic centimeter.
c
1 a  8760 h  3.15
107 s.

organic compounds can be subjected to radical or O2) or a metal ion in its high valence state (e.
chain oxidation. Because many of these g., Pd2 or V5) oxidizes the starting material
reactions proceed spontaneously even at low in a two-electron transfer reaction. Normally, a
temperatures they are also called autoxidations. stoichiometric amount of the oxidizing com-
Radicals recombine readily on surfaces. pound is used and a catalyst, for instance a
Therefore surface enlargement is a proven tool complex of Mo, V, Ti has to be used.
to stop radical reactions like explosions and The catalyst can be dissolved homoge-
flames. neously in a liquid or be present in solid
Metal ions that can transfer only one electron, form. The reaction can take place in the liquid
such as Co2/Co3, Mn2/Mn3, Fe2/Fe3, phase or in the heterogeneous gas phase.
and Cu/Cu2, act as catalyst in the liquid or Examples of heterolytic oxidations in the
in the solid phase. The main function of the metal homogeneous liquid phase are the epoxidation
ions is splitting of the hydroperoxides. Sub- of olefins with hydroperoxides, catalyzed by
stances which readily form radicals, such as Mo complexes, and the oxidation of ethylene to
phenols, many halogen compounds etc., can acetaldehyde by Pd2 complexes.
act either as catalysts or inhibitors, depending Examples of heterogeneous gas-phase
on their nature, concentration, and temperature. reactions are the oxidation of benzene or
Because radicals isomerize easily, and the butane/butene to maleic anhydride; and the oxi-
resulting intermediates are usually more readily dation of propene with mixed-oxide catalysts
oxidizable than the starting hydrocarbons, containing Mo to give acrylonitrile (or acrolein
homolytic oxidations generally have low selec- and acrylic acid) and with Pd catalysts to give
tivity. Only in several special cases can homo- allyl acetate.
lytic oxidation be performed with high Unlike radical autoxidations, heterolytic
selectivity. Examples include oxidation of oxidations proceed via defined complexes
cumene to the corresponding hydroperoxide, and are therefore, in principle, very selective.
of alkyl hydroanthraquinone to alkyl anthraqui- For an introduction and review see [339] and
none and H2O2, of p-xylene to dimethyl tereph- [342]; further literature: [343349].
thalate (DMT) or terephthalic acid (TPA), of
aldehydes to acids and some other processes. 5. Homolytic Oxidation
As an introduction [338341] are
recommended. 5.1. Historical Development; Reaction
Mechanism; Radical Reactivity
Heterolytic Oxidation Reactions. In hetero-
lytic oxidation reactions an active oxygen An extensive study of the history of homolytic
compound (e.g., a peroxyacid, hydroperoxide, oxidation is still missing. Short reviews of the
52 Oxidation

historical development of knowledge of homo-

lytic oxidation can be found, for example, in
[338, pp. 114], [340], and, with emphasis on
the development of homogeneous gas-phase
oxidation, in a publication by SHTERN [350].
The development of concepts of oxidation
began with the introduction of the phlogiston
theory in the 17th century by BECHER and STAHL
and its disproving in 1773 by LAVOISIER. The
development of theories for the oxidation of
organic materials started in 1897 with ENGLERs
proposal that the first step in the oxidation of an
olefin consists in the addition of one O2 molecule
to a CC double bond forming a four membered
ring (see Fig. 1a). Later on this theory was called
moloxide theory by STAUDINGER. Despite the Figure 2. Typical course of a slow hydrocarbon oxidation
fact that a growing number of arguments could (oxidation of a 1:1 benzene/O2 mixture in a glass flask; T
be brought forward against the moloxide theory, 470 C, initial pressure p 600 hPa)
it was accepted until 1939. In this year CRIEGEE
and coworkers could unambiguously show, that hydrocarbons (see Fig. 2). Since the few radi-
in the first stage of the oxidation of an olefin a cals known to the late 1920s were very reactive
hydroperoxide is formed (see Fig. 1b). Already and therefore short lived, nearly nobody could
in 1900 BAEYER and VILLIGER had isolated per- imagine that radicals could significantly partic-
benzoic acid as an intermediate in the oxidation ipate in the slow oxidation of hydrocarbons.
of benzaldehyde to benzoic acid. CHRISTIANSEN had therefore developed in 1924
Another problem that was difficult to under- the concept of an energy chain reaction with
stand for a long time was the S-shaped slope of the inclusion of a hot molecule.
the velocity curve of the oxidation of In 1927 B opposed this idea with
ACKSTROM
the concept of a radical chain reaction, on the
grounds of the high quantum yield in photo-
chemical oxidations.
A mathematical solution to a similar problem
the reaction of H2/O2 mixtures was found
by SEMENOV in 1926 who based his theory on the
work of BODENSTEIN and NERNST. The latter
developed the theory of the unbranched chain
reaction taking the experimental results of
BODENSTEIN, who studied the photochemically
induced Cl2/H2 reaction in 1913 (Scheme 1). In
1926 SEMENOV first proposed the theory of the
branched chain reaction describing mathemati-
cally the pressure rise in the H2/O2 explosion
(Scheme 2, see [350353]).
In the early 1930s SEMENOV was able to prove
mathematically that the pressure rise in the
autoxidation of hydrocarbons could also be
explained using the model of a branched chain
reaction. Because of the long induction period
and the slow rise of the pressure, compared with
Figure 1. Concepts of the structure of peroxides
the very fast H2/O2-reaction, he called this
a) Moloxide (ENGLER, MOUREU et al.); b) Hydroperoxide reaction degenerate branched chain reaction
[338] (see Fig. 3 and [353, pp. 317ff and pp. 438ff]).
 Oxidation 53

Initiation Initiation

Cl2 hn ! 2 Cl (124) RH O2 ! R OOH (139)

Propagation RH O2 HR ! R HOOH R (140)

Cl H2 ! HCl H (125) RR hn ! R R (141)

H Cl2 ! HCl Cl (126) Propagation

Termination R O2 ! ROO (142)

2 H (wall) ! H2 (wall) (127) RH ROO ! R ROOH (143)

2 Cl (wall) ! Cl2 (wall) (128) ROO CC ! ROOCR2C(R2) (144)

H Cl (wall) ! HCl (wall) (129) Branching

ROOH (ROOR) ! RO OH ( OR) (145)

Scheme 1. Unbranched radical chain reaction
ROOH ROOH ! ROO H2O RO (146)

ROOH ROH ! RO H2O RO (147)

Inititiation
Termination
H2 O2 ! 2 HO (130)
R R ! RR (148)
Propagation
RO R ! ROR (149)
HO H2 ! H2O H (131)
ROO ROO ! ROOR O2 (150)
HO OH ! HOO H (132)
ROO R ! ROOR (151)
O H2 ! OH H (133)

Branching Scheme 3. Hydrocarbon oxidation (degenerate branched

radical chain reaction)
H O2 ! HO O (134)

HOO ! HO O (135) 2. The chains branch by decomposition of the

Termination alkylhydroperoxide molecule. However,
molecules react orders of magnitude more
H OH (wall) ! H2O (wall) (136)
slowly than radicals (therefore degenerate
H H (wall) ! H2 (wall) (137) chain branching). Only at high temperatures

O O (wall) ! O2 (wall) (138)
branching of chains occurs via peroxy radi-
cals in hydrocarbon oxidations. But even
Scheme 2. Branched radical chain reaction then these oxidations are not as fast as the
H2/O2-reaction (see Fig. 3 and Section 4.5).
Confirming the findings of CRIEGEE, FARMER
in 19411942 was the first to propose the chem-
ical radical chain mechanism of the homolytic
oxidation of hydrocarbons (Scheme 3).
The essential differences between the rela-
tively slow hydrocarbon oxidation and the very
fast oxygenhydrogen reaction (see Fig. 3) and
similar branched radical chain reactions are:

1. In the hydrocarbon oxidation, hydroperoxide

formation occurs prior to the branched chain
reaction. Hydroperoxide formation takes
place by an unbranched radical chain reaction Figure 3. Course of a methane explosion and a hydrogen
according to Reactions (142) and 3(143) explosion in a sealed vessel (stoichiometric mixture with
(Propagation in Scheme 3). O2) [358, p. 10]
54 Oxidation

3. In large hydrocarbon radicals the energy

absorbed on dissociation is dissipated rap-
idly over the entire radical by atomic vibra-
tions. The radical thus loses a considerable
portion of reactivity. This behavior is par-
ticularly marked for radicals from aromatic
compounds and olefins in which the
unpaired electron (odd electron) is addition-
ally stabilized by the p-electron system.
4. Because of their lower reactivity these radi-
cals have a longer lifetime than the more
reactive radicals.
Figure 4. Relationship between Ea, E, Eo, and DH for
5. Because of the longer lifetime and the ability exothermic (Ea! ) and endothermic (Ea ) reactions
to rapidly dissipate energy, the large radicals
can readily recombine in binary collisions
(chain termination). The recombination of
small radicals (e.g., H H or H OH) energy of the attacking radical. The
occurs only in the much rarer three-body reactivity of CH bonds increases with
collisions or at the wall because of the decreasing bond energy:
need to remove the bond energy liberated. 2. Vinyl  phenyl < primary < secondary <
tertiary < allyl < cumyl < diallyl
Bond Strength and Radical Reactivity. Radi- 3. The first members of homologous series
cal chain reactions essentially occur through (e.g., methane, ethylene, benzene) have par-
the attack of a radical at a covalent bond such ticularly high CH bond energies and are
that an existing bond is broken and a new one correspondingly stable towards attack by
formed (Eq. 152): radicals. In alkanes the energy of the pri-
mary CH bonds decreases with increasing
R3 C  H X ! R3 C H  X  DH 152
chain length. Alkenes are either attacked at
the allylic CH bond or the radicals add to
In such homolytic substitution reactions, a
the CC double bond.
radical is more reactive the greater the value of
DH, i.e., the greater the difference between 4. The hydroxy, alkoxy, and acyloxy radicals
the bond energies according to Equation (153): can abstract a H atom from all CH bonds.
The peroxy radicals, in contrast, exhibit a
DH HX  DH CH DH kJ=mol 153 similarly high reactivity only towards CH

According to the rule of POLANYI and EVANS,

more precisely formulated by SEMENOV [353, p.
26 ff], from Equation (154):

Ea 48 0:25DH kJ=mol 154

the activation energy Ea of an exothermic

reaction is lower and the reaction faster the
greater the value of DH (see Figs. 4 and 5).
In Figure 6 dissociation energies for various
types of bond are shown. The following is
noted:
Figure 5. Dependence of the activation energy Ea on the
enthalpy of reaction DH for exothermic radical exchange
1. The attack of a radical at a CH bond takes reactions of type A  B C ! A B  C [351353]
place more easily and faster the lower the The reactions investigated by POLANYI and EVANS lie in the
energy of the bond and the greater the shaded area
 Oxidation 55

Figure 6. Dissociation energies of various bonds [353, p. 16 ff.]

bonds of hydrocarbons, alcohols, and alde- hydrocarbons these data can be applied to
hydes that lie below the tertiary CH bond the reactivity in the liquid phase without
of isobutane in the energy scale. With CH complications because in the liquid phase
bonds of compounds which lie above iso- hydrocarbon molecules interact with one
butane, these radicals react much more another only weakly and are not altered
slowly, but much more selectively (see
Section 4.4.1, Tables 9, 10).
5. Through the cleavage of the comparatively Table 3. Oxidizability, absolute chain propagation rate kp, and
chain propagation rate relative to cumene (kp)rel for various organic
weak OO bond in peroxides the highly compounds at 30 C [339, p. 19], [340, p. 35]
active HO and RO radicals are formed. p
Compound kp = 2kt  103 , kp, L (kp)rel
0.5 0.5
mol s mol1 s1
According to Figure 6 the CH bond of the n-Nonane 0.19
aldehyde group is less reactive than the allylic n-Dodecane 0.32
CH bond in olefins or the tertiary CH bond Cyclooctane 0.39
Cyclododecane 0.55
of cumene. However, in reality aldehydes are 1-Octene 0.062 1.0
much more reactive than these hydrocarbons Cyclohexene 2.3 5.4 10
because they can be oxidized at much lower 2,3-Dimethyl-2- 3.2 2.6 20
temperatures than e.g., cumene or butene (see butene
Toluene 0.014 0.24 0.28
Table 3, last line). There are three reasons for p-Xylene 0.049 0.84 0.73
this confusing phenomenon: Cumene 1.50 0.18 1.00
Tetralin 2.30 6.4 4.0
1. The data in Figure 6 originate from mea- Benzyl alcohol 0.89 2.4
Benzaldehyde 290 33 000
surements in the gas phase. In the case of
56 Oxidation

chemically until they are oxidized. But in of radical recombination take place without
the case of aldehydes these data are mis- activation energy and are therefore very fast.
leading, because aldehydes form addition This mechanism is the cause of the ageing of
products with polar compounds like water, organic materials in the presence of oxygen
alcohols, and acids in the liquid phase. In the even at low temperatures. It is also the mecha-
addition product the remaining CH bond nism of initiation of autoxidation.
of the aldehyde group is strongly activated Although radical chains are continuously
by either two OH groups or one OH and one initiated even at low temperatures, the reaction
OR group at the same C-atom. does not exceed a uniform low rate as long as
2. Peroxyacyl radicals are much more reactive more radicals are destroyed than are formed.
than peroxyalkyl radicals in cleaving a CH This is the cause of the induction period, which
bond to gain a H-atom in order to stabilize to is always exhibited by uncatalyzed autoxida-
the corresponding peroxy acids. tion reactions. In particular hydroperoxides
3. Peroxy acids decompose much faster to accumulate during this induction period as
radicals than alkyl hydroperoxides (see active intermediates, which are formed in an
Section 4.4.1). unbranched chain reaction according to
Reactions (142) and (143).
Dangerous peroxide build-up can occur in
Initiation; Induction Period; Peroxide Forma- olefin-containing liquids, since peroxy radicals
tion. Small molecules such as H2, CH4, and also add easily to double bonds according to
NH3 normally exist as gases. Because of the Reaction (144). Because oxygen generally
stable bonds, which necessitate high dissocia- reacts much faster with the alkylperoxyalkyl
tion energies, and because of the weak inter- radicals formed than the olefin, copolymeric
actions between the gas molecules, these types peroxides are formed in this way. Because of
of molecule only react with oxygen at high the missing OH group, these are considerably
temperatures, if no catalysts or initiators are more stable than hydroperoxides and can there-
present. At low temperatures their mixtures fore accumulate in large quantities. At higher
with oxygen are metastable in the absence of temperatures they can decompose explosively.
catalysts. Because the common reagents for peroxide
In the autoxidation of higher hydrocarbons detection are mostly sensitive to hydro-
the situation is quite different, especially in the peroxide, these polyperoxides often remain
liquid and solid phase. Because of inter- undetected. For more details, see [354357].
molecular interactions and the relatively low Hydroperoxides also accumulate in paraffins,
bond energies of many CC and CH bonds, glycol ethers, and other nonvolatile organic com-
radicals are formed relatively rarely, but con- pounds on exposure to air. These can initiate
tinually, even at low temperatures. This may rapidly accelerating oxidation reactions even at
occur by reaction with O2 (Reactions (139) or moderate temperatures. Cleaning cloths soaked
(140)) or by concentration of the thermal vibra- with linseed oil can even ignite at room temper-
tion energy into one bond (cf., the lowering of ature. The self-ignition temperature of paraffins,
the ignition temperature for gas explosions surfactants, glycol ethers, and similar substances
from ca. 600 C for methane to ca. 200 C for drops below 100 C if, for example, insulating
decane, see Table 15). The rate of radical material with a large surface area, such as rock
formation is increased drastically by high- wool which has been moistened with them has
energy radiation (sunlight) and heat. In most been lying for a few days in the air.
cases these radicals recombineabove all in The induction period can be shortened or
the liquid or solid phase, where the mobility of completely eliminated by the addition of
the molecules and their fragments is severely peroxides or other radical-forming reagents,
limited as a result of the forces of interaction or through irradiation.
(solvent cage). However, if oxygen is present,
it competes with the simple radical Termination; Inhibition. Radical chain reac-
recombination (Reaction 148) because of its tions can be slowed down or stopped by reduc-
diradical structure (Reaction 142). These types ing the concentration of active intermediates,
 Oxidation 57

the overall rate of radical formation by peroxide

decomposition (Reactions 145 to 147) begins to
exceed that of radical destruction (Reactions
148 to 151). From this point, the overall
reaction rate increases exponentially to a maxi-
mum, which is reached at an early stage of the
overall reaction.
Afterwards the reaction rate slows down to
zero. In the oxidation of hydrocarbons the
reaction normally ceases long before the start-
ing material is completely consumed. One rea-
son for this behavior is the dilution of the
reaction mixture by the end products formed,
mainly water and lower carboxylic acids such
Figure 7. Typical autoxidation curves in the absence and as acetic acid, the latter being very resistant
presence of antioxidants [338, p. 4] against oxidation. In the presence of metallic
a) Absence of antioxidant; b) x parts antioxidant; c) 2 x parts
antioxidant ions as catalysts, water and carboxylic acids act
as catalyst poisons because they are stronger
complexing agents than hydroperoxides (see
for instance by enhancing radical recombina- Section 4.4.2). Raising the temperature can
tion reactions. revive the reaction for a more or less short time.
In homolytic gas- and liquid-phase oxida- High starting material conversions are, how-
tions, solid surfaces are the most effective ever, only rarely desired in homolytic oxidations
radical scavengers. Fillings or packings made because the product selectivity decreases rapidly
from thin tubes (capillaries) or Raschig rings with increasing conversion due to secondary
are used overall in industry as flame and explo- reactions. Frequently, starting material conver-
sion barriers, and certain inorganic dusts (e.g., sions of 1030%, sometimes even only 1%, are
ammonium phosphate) are among the most aimed for. (For more details see Section 4.2.)
effective fireextinguishing agents.
Organic antioxidants (e.g., phenols, thiols, Hydroperoxide Decomposition. The decom-
amines) act in different ways. Phenols, for position of hydroperoxides is the main source
instance, are transformed by active radicals of radicals during the entire reaction (Scheme 3,
into low energy radicals, which subsequently Reactions 145 to 147).
add other active radicals forming stable prod- The thermal decomposition of peroxides has
ucts. Thiols are oxidized by hydroperoxides to activation energies of between 120 and 164 kJ/
sulfoxides whereby the hydroperoxides are mol. These are the highest activation energies
reduced to alcohols. In these reactions the of the reactions shown in Scheme 3.
organic antioxidants are, however, gradually The hydroperoxide decomposition therefore
consumed. Once they have been consumed, determines the temperature range for the oxi-
the autoxidation continues at the rate it would dation of organic compounds. Temperature
have in the unstabilized substance (see Fig. 7); ranges for liquid-phase oxidations are as fol-
for more details, see Section 4.4.3. An intro- lows: aldehydes: up to 80 C; olefins and
duction to organic stabilizers and antioxidants cumene: up to 130 C; aliphatics, alkylaro-
can be found in [338]; fire and explosion pro- matics, alcohols, ketones, esters: 80 to 200 C.
tection are described in [358367]. Since hydroperoxides, and of course also
peroxides, are also used commercially as starters
Rise in the Reaction Rate; Final Conversion. for radical chain polymerizations, a large body of
As soon as the peroxide concentration has data exists on their rates of thermal decay.
exceeded a critical value which depends Reviews on decomposition rates and activation
on the nature of the compound to be oxidized, energies for a wide range of peroxides can be
the state of aggregation, the presence or found both in the literature [368] and in bro-
absence of catalysts, and the temperature chures from peroxide producers [369371].
58 Oxidation

5.2. Kinetics of Autoxidation Since a kinetic equilibrium (steady state) is

established very rapidly after initiation of the
Kinetic measurements, in particular with radical formation and radical scavenging
respect to individual elementary reactions, reactions, and because only Reaction (150)
are associated with particular difficulties and plays a role as a chain termination reaction at
uncertainties in radical chain reactions because sufficient O2 supply, Equation (156) is valid:
of the great number of parallel and secondary
reactions and the normally very low concentra- ui ut 2kt ROO 2 156
tion of the reacting compounds. Nevertheless,
many insights into the course of the reaction where ui rate of chain initiation and ut rate
could be obtained, for example, by of chain termination. Inserting Equation (156)
into (155) gives:
1. Following the pressure rise in gas-phase
oxidations dO2  kp p
 p RH  ut 157
2. Investigation of the peroxide decomposition dt 2 kt
in hydrocarbons
3. Simultaneous oxidation (cooxidation) of Equation (157) is only valid for low conver-
different compounds sions as the secondary reaction products formed
at higher conversions are generally more read-
4. Oxidation in the presence of a tertiary
ily oxidized thanpthe
starting hydrocarbon. The
hydroperoxide
expression kp = 2kt is referred to as the
5. By using specific analytical methods: ESR, oxidizability, since it indicates how much
chemiluminescence, and UV absorption for faster one organic compound is oxidized than
radicals; liquid chromatography for hydro- another. Table 3 lists values for the oxidizability
peroxides and nonvolatile compounds; gas as well as kp and relative kp (based on cumene)
chromatography for the stable, volatile values for various hydrocarbons. Under mild
products, etc., and many other methods. oxidation conditions, the value of kp increases
in the following order: n-alkanes < isoalkanes
The main problem is still, however, the < alkylaromatics < alkenes < alkynes  styr-
quantitative determination of all components ene derivatives
aldehydes.
present in a reaction mixture. For example, If, however, the oxidation rates of alkanes
active intermediates react further after the sam- are compared under more severe conditions
ple has been taken, so that the composition of (temperature > 120 C) where the hydro-
the sample changes rapidly. For low-conversion peroxide decomposition determines the
reactions, which are required for the precise reaction rate, then the order of overall oxidation
determination of the kinetics, the sensitivity rate is as given in Table 4 and Figure 8.
limits of all analytical methods are quickly
It can be seen that the overall oxidation rate
reached. For a review of experimental methods
increases sharply both with increasing chain
for the determination of the kinetics of radical
length and with decreasing branching. The
chain oxidations see [341], and for analytical
reasons for this are:
methods for peroxide determination see [372].
Under mild conditions (temperature
< 100 C and sufficient oxygen, i.e., for a
liquid-phase reaction: O2 partial pressure Table 4. Relative oxidation rate of various n-alkanes [373, p. 94]
> 10 kPa, for a gas-phase oxidation: O2 excess) n-Alkane Relative oxidation rate [mol O2
the rate of oxygen conversion is only deter- (mol alkane)1 time1]
mined by Reaction (143) in the case of the Ethane 0.001
oxidation of hydrocarbons, since Reaction Propane 0.1
(142) is very fast and the concentration of R Butane 0.5
Pentane 1.0
is very small. Equation (155) is therefore valid: Hexane 7.5
Octane 200
dO2 
 kp ROO RH 155 Decane 1380
dt
 Oxidation 59

Figure 8. Oxidation rates of hexane isomers [mol O2 (mol alkane)1 time1] relative to 2,3-dimethylbutane [373, p. 95]

1. In the series primary ! secondary ! ter- [340, 341, 351354] (Because autoxidation
tiary CH bonds, the rate of formation of is generally initiated by the decomposition
the hydroperoxides increases dramatically, of added peroxides, peroxide decomposition
but their rate of decomposition decreases frequently is designated initiation in the
equally drastically. The secondary hydro- literature, which is incorrect and may lead
peroxides are formed more slowly than ter- to confusion.)
tiary hydroperoxides but decompose The following points may be made:
considerably faster, so that the lower rate
of formation is more than compensated for. 1. According to Equation (157) the rate of O2
Therefore, under these conditions, hydro- take-up in the initial phase of an oxidation is
carbons with secondary CH bonds are only proportional to the hydrocarbon con-
oxidized faster than hydrocarbons having centration and is independent of the O2
tertiary CH bonds. concentration under mild reaction
2. The number of secondary H atoms per conditions.
molecule increases with decreasing branch- 2. The k value of Reaction (142) is estimated
ing. All secondary CH bonds in an alkane to be 107 to > 109 L mol1  s1 (corre-
are attacked equally rapidly. sponding to 1014 to > 1012 cm3 mole-
cule1  s1) at Ea  0 kJ/mol [339]. It is
orders of magnitude faster than Reaction
Tables 58 list some literature values for (143), which therefore determines the propa-
the collision factor A, activation energy Ea and gation rate. Because of its high rate, Reaction
rate constant k for the initiation of the ele- (142) is generally controlled in the liquid
mentary reaction, peroxide decomposition phase under industrial conditions by O2
(branching), propagation, and termination diffusion.

Table 5. Kinetic data for the thermal decomposition of some initiators at 80 C [340, p. 23], [341, p. 21 ff], [354, p. 95]

Initiator Solvent T,  C kd, s1 Ad, s1 Ea,d, kJ/mol

0 3 15
2,2 -Azodiisobutyronitrile benzene 7.59
10 1.6
10 129
Cyclohexyl hydroperoxide cyclohexane 150 3.1
105 1.2
1013 142
tert-Butyl hydroperoxide n-octane 150 0.90
105 1.2
1015 163
water 4.0
1012 142
Di-tert-butyl peroxide benzene 9.55
106 1.2
1014 142
Cumene hydroperoxide cumene 150 1.24
105 7.2
108 111
mineral oil 150 1.35
104 1.3
1011 121
tert-Butyl peroxybenzoate ethylbenzene 9.55
105 3.8
1014 139
n-undecane 2.34
106 7.1
1016 157
Benzoyl peroxide benzene 79.8 3.3
105 139
Dibenzoyl peroxide benzene 5.75
104 1.2
1013 120
n-octane 2.63
103 3.5
1014 121
Tetralyl hydroperoxide mineral oil 150 1.35
104 1.3
1011 121
9-Decalyl hydroperoxide decalin 150 9.9
105 8.5
1013 134

The values are only valid for highly dilute solutions. Only under these conditions is the decomposition unimolecular.
60 Oxidation

Table 6. Preexponential factors and activation energies for the initiation reactions 3(139) and 3(140) a [340, 341, 351353]

Compound Reaction mechanisma T,  C Ai,1, L mol1  s1 Ea,i,1, kJ/mol Ai,2, L2 mol2  s1 Ea,i,2, kJ/mol
3
Tetralin 140 130 3.4
10 86.7
Cyclohexane 139 8
1012 167.5
Cyclohexanol 140 8.3 67.0
Cyclohexanone 140 2.1
102 73.3
Cumene 139 120 3.5
109 113.0
o-Xylene 139 2
108 129.8
Ethanol 139 6
107 125.6
Benzaldehyde 140 5 2.5
103 48.5
a

(139): RH O2 ! R OOH ui,139, ki,139, etc.

(140): RH O2 HR ! R HOOH R ui,140, ki,140, etc.
ui,139 ki,139  [RH] [O2]; ui,140 ki,140  [RH]2  [O2];
ki,139 Ai,139  exp (Ea,i,139/R T); ki,140 Ai,140  exp (Ea,i,140/R T).

Equation (140) is probably an oversimplification because there is experimental evidence that this reaction proceeds via two or more steps
involving a complex of the type [RH . . . O2] [341, p. 20 ff.].
a
If RH consumption is proportional to the reaction rate ui, the reaction mechanism (139) is valid. With a quadratic relationship between RH
consumption and ui, the reaction mechanism (140) is valid. For more details, see [341, p. 18 ff.].

Table 7. Rate constants, preexponential factors, and activation energies for the chain-propagation reaction (143)a

Compound T,  C kp, L mol1 s1 Ap, L mol1 s1 Ea,p, kJ/mol

b
Tetralin 25 13.3 2.5
104 18.8
Tetralin 30 1.6b
Cyclohexane 60 0.53
Cyclohexane 1.5
106 35.6
Cyclohexane 75 0.34 57.8
Cyclohexanol 40 0.70
Cumene 30 0.18
Cumene 65 0.50 1.1
104 28.1
Benzaldehyde 5 1900 4.8
104 7.5
Benzaldehyde 30 33 000
a
(143): RH ROO ! R ROOH; up kp  [ROO ]  [RH]; kp Ap  exp (Ea,p/R T).
b
Data from different authors sometimes differ by more than one order of magnitude.

Table 8. Rate constants, preexponential factors and activation energies for the chain termination reaction (150)a

Compound Solvent T,  C 2kt, L mol1 s1 2 At, L mol1 s1 Ea,t, kJ/mol

HOO 30 7.60
105
Primary RCH2OO 30 107
Tetralinb Tetralin 60 2.14
107 4
107 1.7
Cyclohexaneb Cyclohexane 60 5.25
106 5.9
107 6.7
Ethylbenzeneb Ethylbenzene 60 1.90
107 1.9
107 0.0
Cumeneb Cumene 60 1.12
105 1.3
1010 32.2
Isobutaneb Isobutane 60 1.23
104 1.6
109 32.6

a
(150): ROO ROO ! ROOR O2; ut kt  [ROO ]2; kt At  exp (Ea,t/R T).
b
The compound name stands for the secondary or tertiary peroxy radical, respectively.

3. The number of repetitions of Reactions follows that:

(142) and (143) is referred to as the kinetic up kp
n pp  RH 158
chain length n. It is given by the ratio of the ut 2kt ui
propagation rate to the termination rate. With increasing rate of radical formation
Because of the stationary state ui ut, it and radical concentration, i.e., with
 Oxidation 61

intensification of the reaction conditions starting material is completely

(e.g., by raising the temperature and/or consumed.
initiator concentration), the chain length 6. Aldehydes are particularly readily oxidized
becomes shorter. Under mild reaction con- (see Sections 4.1 and 4.4.1).
ditions n can exceed 1000. Under industrial
conditions it is usually ca. 10.
4. The hydroperoxide decomposition has the 5.3. Differences between Liquid- and
highest activation energy. The values given Gas-Phase Autoxidations
are extrapolated to infinite dilution and are
therefore valid for the uncatalyzed, thermal In principle, autoxidations proceed with the
decomposition. Under normal oxidation same mechanism in the liquid and gas phases.
conditions (no metal catalyst, but typical However, to achieve comparable reaction rates,
concentrations of radicals, hydroperoxides, the temperature in the gas phase must be
alcohols, acids, and other substances with an 100200 C higher than in the liquid phase
accelerating effect), the activation energies because of the weaker interactions between
for the hydroperoxide decomposition are the molecules (in the gas phase nearly no
considerably lower, especially in the liquid hydrogen bonds are formed, no metallic cata-
phase. Nevertheless, they are still higher lysts are present, etc.).
than those of the other reactions. The hydro- Above 250 C paraffins increasingly react
peroxide decomposition therefore deter- with oxygen in a first step to olefins and
mines the temperature range of the hydrogen peroxide. Hydrogen peroxide is
oxidation. far more stable than alkyl hydroperoxides
against splitting of the OO bond. Addition-
5. With progressing oxidation an increasing
ally, above 250 C peroxy radicals increasingly
number of intermediates are formed (alde-
do not stabilize to form hydroperoxides, but
hydes, alcohols, ketones) that are more read-
isomerize and/or split directly to products
ily oxidized than the starting hydrocarbon.
such as aldehydes plus RO . But these
These products form, on the one hand, other
reactions require higher activation energies
hydroperoxides and peroxy radicals, some
than decomposition of hydroperoxides. There-
more labile, some more stable, and on the
fore, in the region 350450 C, depending on
other hand some of these products and their
the hydrocarbon, the reaction velocity
radicals catalyze the decomposition of the
decreases with increasing temperature (region
hydroperoxides.
of negative temperature coefficient). At higher
This has several consequences:
temperatures the reaction rate rises again with
a. The reaction rate increases until it
temperature.
reaches its maximum with the appear-
The effect of material and heat transport is of
ance of aldehydes.
considerable importance for the different path-
b. The O2 uptake is accelerated to a ways followed by homolytic oxidations in the
greater extent than the consumption liquid or gas phase.
of the starting material owing to for- In the liquid phase, heat transfer is good and
mation of the more readily oxidizable the oxidation rate is mostly limited by oxygen
intermediates. transport from the gas phase. In the gas phase
c. The products formed from these inter- mass transfer does not limit the oxidation rate
mediates principally lower carbox- but heat transfer is poor due to the low mass of
ylic acids are more stable to the gas. The oxidation reaction can therefore
oxidation than the starting hydrocar- readily turn into a thermal explosion due to
bons and, in the case of metal-cata- insufficient heat removal.
lyzed autoxidations inhibit the More details on gas phase oxidation are
peroxide decomposition with increas- given in Section 4.5. A detailed account of
ing concentration. The rate of radical the differences between gas- and liquid-phase
formation thus decreases gradually to oxidations can be found in, for example [341,
zero and the reaction stops before the pp. 442495].
62 Oxidation

5.4. Homolytic Oxidation in the Liquid Because Reaction (142) is so rapid, even a
Phase low O2 concentration (O2 partial pressure in the
gas phase 25 kPa) is sufficient to make the
5.4.1. Secondary Reactions of Radicals, concentration of hydrocarbon radicals so small
Peroxides, and other Intermediates that they can practically no longer participate in
recombination reactions.
It is only possible here to go briefly into the In industry, however, oxidation reactions are
secondary reactions of the intermediates and mostly run with O2 off-gas concentrations near
radicals that are important for the understand- to zero. The upper part of the bubble column is
ing of radical chain oxidations by means of thus deficient in O2, so that the concentration of
selected examples. For further information see hydrocarbon radicals increases, and radical
[372, 374]. recombination reactions can occur in this region.
Thus in practically all industrial autoxidations
radical trapping products formed according to
Reactions of Alkyl and Aralkyl Radicals.
Reactions (148) to (151) are found.
The most important reactions of these radicals
The addition of alkyl radicals to CC dou-
are:
ble bonds occurs in olefin oxidations and acts as
an initiation reaction for radical polymerization
1. Addition of O2 (Reaction 142).
in the oxygen-deficient regions of the bubble
2. Recombination with other radicals (termi- column.
nation Reactions 148, 149, 151). Radicals that are stabilized by p-electron
3. Addition to CC double bonds. systems isomerize very readily; examples
4. Isomerizations (Reactions 159 to 162). include Reactions (159) to (162):

159

Yield 100%;
no polymeric peroxide

160

Yield 100% polyperoxide;

no hydroperoxide

161

Yield 100% hydroperoxide

162
 Oxidation 63

Table 9. Relative rate of Reaction (143) as a function of the by the chain carrying peroxy radicals. Taking
nature of the attacked CH bond cyclohexane autoxidation as a case study,
RH CH prim. sec. tert. peroxy radicals react 50 times faster with
the hydroperoxide (i.e., the primary propaga-
2-Methylpentane urel 1 30 300
tion product) than with the parent hydrocarbon
[79]. The abstraction product (i.e., CyaH
OOH) is not stable and promptly dissociates
to cyclohexanone plus HO [375]. The
Reaction (162) involves a doubly activated hydroxyl radical rapidly goes on to abstract
CH2 group as is present in, for example, lino- an H atom from the cyclohexane substrate,
lenic acid derivatives. Double bond systems of producing water and an alkyl radical. Both
this type isomerize almost completely during steps combined are exothermic by 209 kJ/mol,
oxidation to give conjugated double bonds. causing a rapid increase of the local tempera-
With an adequate O2 supply they reactlike ture by several hundred degrees. This local-
all olefinsto form copolymeric peroxides. ized hotspot affects the further fate of the
With a deficiency of O2 they react to give solvent encapsulated radicals. Indeed, either
homopolymers (principle of self-drying oil the species just diffuse away from each other,
paints). or the R radical reacts with the nascent
CyOOH to form alcohol and an alkoxy radical.
Reactions of Peroxy Radicals. The most Although the latter reaction features an appre-
important reaction of the peroxy radical is H ciable barrier compared to the diffusive sepa-
abstraction (Reaction 143). The relative rates of ration, the local high temperature makes this
attack of ROO on primary, secondary, and OH abstraction possible. As such, the consec-
tertiary CH bonds and on CH bonds which utive copropagation of the primary hydro-
have another substituent (OH, OOH) on the peroxide produces both ketone and alcohol.
same C atom are given in Tables 9 and 10. For This mechanism was verified for a number of
example, alkylperoxy radicals react with the different substrates and readily explains the
primary, secondary, and tertiary CH bonds in experimental trend in alcohol-to-ketone ratio:
2-methylpentane in the ratio 1/30/300. substrates that form (resonance) stabilized
The OOH substituent in the hydroperoxide alkyl radicals feature a lower alcohol-to-
activates the aH atom towards H abstraction ketone ratio as the barrier of the R plus
ROOH reaction increases [79, 216, 376].
The OH substituent in alcohols also acti-
vates the a-H atom towards H-abstraction by
Table 10. Relative rate of Reaction 3(143)  as a function of the peroxy radicals. Although slower than for the
substituent X at the same carbon atom bearing the attacked CH corresponding hydroperoxde, this copropaga-
bond
tion leads to the formation of a-hydroxylper-
XRH X: H OOH OH oxy radicals after the addition of O2. These
urel: 1 9.5 12 radicals can eliminate HO2 radicals in an
equilibrated reaction [377] that have an inhib-
iting effect on autoxidations, as they terminate
diffusion controlled with peroxy radicals
[378].
Because of their comparatively low
Ethylbenzene-X [(1-X,1-phenyl) reactivity the ROO radicals can accumulate
ethane] in detectable concentrations (unlike, for exam-
ple, HO , R , RO ). In the oxidation of cumene,
urel: 1 13 6.3

radical concentrations of up to 8
105 mol/L,

corresponding to 5
1016 molecules/cm3, were
detected by ESR. Peroxy radicals therefore
n-decyl-X
frequently recombine, for example by

Reaction (143) R  H ROO ! R ROOH Reactions (163, Russel reaction) and (164).
64 Oxidation

higher recombination rates of primary and sec-

ondary ROO radicals.
In contrast, hydrocarbons with allylic CH
bonds such as ethylbenzene and tetralin are
163 oxidized faster than those with tertiary CH
bonds. Nevertheless, because of the higher
recombination rates of allylic ROO radicals,
addition of the corresponding hydrocarbons
does not enhance, but strongly hinders the
autoxidation of tertiary CH bonds under
mild conditions. For example, the rate of oxi-
dation of cumene is reduced by two thirds by
164 the addition of 3% tetralin because the rapidly
formed tetralyl peroxy radicals capture the
cumyl peroxy radicals, which are formed
more slowly. Ethylbenzene has an analogous
Lacking H atoms at the C atom that binds the effect.
peroxy group, tertiary ROO radicals cannot be Peroxy radicals also add readily to CC
stabilized in the same simple way as primary double bonds, forming either epoxides and
and secondary peroxy radicals (see Reaction alkoxy radicals (Twigg mechanism) [379
163). They generally recombine to give tetr- 381], or polyperoxides and their decomposition
oxides, which subsequently decompose form- products (epoxides, aldehydes, alcohols, and
ing very stable di-tert-alkyl peroxides ketones) (Mayo mechanism).
(Reaction 164) or tert-alkoxy radicals. The
latter may either form tertiary alcohols or
decompose further with cleavage of the carbon
framework to form ketones, aldehydes, alkanes,
or other decomposition products:

165

Because reactions of type (165) have high

activation energies, the overall reaction of the
166
recombination of tertiary ROO radicals pro-
ceeds in general with high activation energy
(see Table 8).
Rate constants and activation energies for
the recombination of primary, secondary, ter-
tiary, and allylic ROO radicals are listed in Reactions of Hydroperoxides. The stability of
Table 8. The oxidation rate is inversely propor- hydroperoxides depends on (1) the structure,
tional to these chain termination rate constants (2) the surrounding medium, and (3) the active
(see Eq. 157). Therefore, under mild conditions components in this medium.
(ca. 100 C) autoxidations of primary and sec- For aliphatic hydroperoxides, the stability
ondary CH bonds are slower than those of decreases in the order: tertiary > secondary
tertiary CH bonds, not only because of their > primary. Electron-withdrawing groups
higher bond energies but also because of the (CN, COOR) have a destabilizing effect,
 Oxidation 65

and methyl groups a stabilizing one. Hydro- 170

peroxides in which the HOO-group is in a
position a to a double bond or an aromatic ring
(resonance stabilization) are fairly stable. Normally, Reaction (169) is the preferred
Because CH bonds in such a-positions are one. But if there are not sufficient molecules
weakened due to the resonance stabilization of with active CH bonds available, that can be
the incipient radical, these types of hydro- attacked by the acyloxy radical in order to
peroxide are formed at lower temperatures stabilize the radical as a carboxylic acid,
than others, and can be obtained in high yields Reaction (170) becomes important (formation
(e.g., cumene hydroperoxide). of CO2). Reaction (170) seems to be important
Table 5 lists kinetic data for the mono- also in the case of aromatic acyloxy radicals.
molecular decomposition of peroxide in highly There the odd electron is attracted by the
dilute solution (see Reaction 145). p-electron system of the aromatic ring which
In the presence of substrates with weak leads to radical isomerization, and subse-
CH bonds, the peroxide decomposition can quently splitting into a phenyl radical and CO2.
be enhanced as shown in Reaction (167).
Cleavage by Mineral Acids. Mineral acids
protonate peroxides and thus induce their
ROOH R0 -H ! RO H2 O R0 167
decomposition. The best known example is
the Hock rearrangement of cumene hydro-
This has important consequences for autoxi- peroxide to give phenol and acetone (Eq. 171):
dation reactions that yield products with signif-
icantly weaker CH bonds, such as
cyclohexanone in the case of cyclohexane
autoxidation [80]. Indeed, this enhanced
bimolecular initiation with a secondary reaction
product leads to an exponential increase in the 171
apparent initiation rate constant and explains
the autocatalytic nature of autoxidations.
Because of their acidic character, peroxya-
cids catalyze their own decomposition much
more strongly than the alkyl hydroperoxides
(Eq. 168):

168

Another industrially important example is

the oxidation of p-xylene to terephthalic acid,
The acyloxy radicals formed are more catalyzed by CoBr2 in acetic acid. Oxidations
reactive than the alkyl peroxy radicals. They of alkylaromatics in which mineral acids such
can react in different ways (Eqs. 169 and 170): as HBr can be formed only proceed at a satis-
factory rate if it is ensured that the concentra-
tion of the mineral acid remains low. Otherwise,
phenol, which acts as an oxidation inhibitor,
169 may be formed by acidic hydroperoxide
cleavage.
66 Oxidation

Reactions with Bases. Bases react with pri- Reactions with Ketones. In polar solvents,
mary and secondary hydroperoxides as follows hydroperoxides readily add to ketones. The
(PRITZKOW believes that these findings are arte- resulting perketals are labile and decompose
facts, caused by very small impurities of metal readily, forming a variety of products
salts in the bases) (Eq. 172): (Eq. 178):

172
ROOH B ! ROO B  H

Chain reaction (Eqs. 173 and 174):

ROO ROOH ! RO O2 ROH 173

RO ROOH ! ROH ROO 174

178
In secondary dialkyl peroxides, the CH
bond at the same atom as the peroxy group is
attacked (Eq. 175):

Peroxy acids react with ketones with CC

bond cleavage to give esters or lactones
(BaeyerVilliger reaction) (Eq. 179).
175

Bases have practically no effect on tertiary

dialkylperoxides.
179
Reactions with Olefins. In the absence of epox-
idation catalysts (e.g., Mo complexes) hydro-
peroxides predominantly react with olefins
according to (Eq. 176): Reactions with Aldehydes. Peroxy acids add to
aldehydes more readily than to ketones, form-
ing the corresponding aldehyde monoperesters.
176 Cleavage of these compounds gives the corre-
sponding acids. This cleavage is selectively
catalyzed by Mn2/Mn3 ions (see Section
whereas peroxy acids mostly give epoxides 4.4.2) In the oxidation of aldehydes to acids
very selectively in a nonradical reaction (Eq. almost the entire acid is formed via this path-
177) (Prilezhaev reaction; see Section 5.1) way (Eq. 180):

177

180
 Oxidation 67

Reactions of Alkoxy Radicals. The alkoxy proceeds via intermediate aldehydes and ace-
radicals formed in the decomposition of hydro- tals to give the corresponding acids and esters.
peroxides are very reactive. However, only The autoxidation of secondary alcohols pro-
primary alkoxy radicals appear to react pre- ceeds through a-hydroxyperoxy radicals
dominantly by abstraction of hydrogen atoms to (Reaction 178) that yield the corresponding
form alcohols. Secondary and, above all, ter- ketone upon elimination of HO2 (Eq. 185).
tiary alkoxy radicals react predominantly by
CC cleavage to give aldehydes or ketones,
together with small amounts of alcohols
(Eqs. 181183).

181

185

The ketones formed are further oxidized via

the a-hydroperoxyketones (Eq. 186):

182

183

186

The CC cleavage reactions (Reactions 182

and 183) are very important in autoxidation
because the carbon chain degradation proceeds
mostly by these mechanisms. In case of
Aldehydes form addition products with
cyclohexane oxidation, significant amounts
polar compounds in the liquid phase. In the
of cyclohexoxy radicals are formed in the
addition product the CH bond of the aldehyde
co-propagation of the hydroperoxide. A frac-
group becomes strongly activated by two polar
tion of those radicals decompose to v-formyl
groups (see Section 4.1, and Table 10, where
radicals (i.e., CH2(CH2)4CHO). These species
the activation of the CH bond by one polar
are the most significant cause of ring-opened
substituent is shown) (Eqs. 187190):
byproducts in this industrially important
process [382].
In olefin oxidations, alkoxy radicals also add
to CC bonds (Eq. 184) [380]:

187

184

Oxidation of Alcohols, Aldehydes, and 188

Ketones. The oxidation of primary alcohols
68 Oxidation

189 5.4.2. Metal Catalysis in Homolytic Liquid-

Phase Oxidation

Homolytic liquid-phase oxidation is influenced

by many metal ions, but only metals that can
190 occur in two valence states of approximately
equal stability in the oxidation medium can act
catalytically. These are above all Co2/Co3
and Mn2/Mn3. Ionic pairs such as Cu/Cu2
The acyl radicals formed initially can and Fe2/Fe3 are important only in special
decompose forming an alkyl radical and CO cases.
if there is an oxygen deficiency (Reaction 60) or The catalytic activity of these metals is
can also recombine to give diketones (Reaction based on the acceleration of peroxide decom-
190: synthesis of diacetyl from acetaldehyde). position by lowering the activation energy
Carbon monoxide and carbon dioxide are (Eqs. 192194)
formed in autoxidation almost exclusively by Metal catalysis (Ea 4050 kJ/mol):
decomposition of acyl and acyloxy radicals.
The presence of CO in the reaction waste gas
(at an order of magnitude of  0.1 vol%) is thus 192
a sensitive indicator of oxygen deficiency in
autoxidations.
Cooxidation Cooxidation is the simultaneous 193
oxidation of several substances. The cooxida-
tion of aldehydes and olefins to produce
epoxides (and acids) is particularly well known
(Eq. 191):
194

Because alkyl hydroperoxides are strong

oxidants, Reaction (193) is only important in
the decomposition of pure hydroperoxides, i.e.,
at low conversions. At higher conversions the
metal ions in the higher valence state are pref-
191 erably reduced by aldehydes, because
aldehydes are strongly reducing compounds
(Eqs. 195 and 196):

195

as is the cooxidation of p-xylene and par-

aldehyde or methyl ethyl ketone in acetic
acid to give terephthalic acid and acetic acid
(Toray, Olin-Matheson). 196
As is apparent from the preceding sections,
practically every autoxidation becomes a coox- Mn2/Mn3 catalyze the decomposition of
idation soon after initiation. Planned cooxida- hydroperoxides in the same way as cobalt ions.
tion has achieved great importance, e.g., for Although Mn3 is considerably less active than
elucidating reaction mechanisms and determin- Co3, manganese is of great importance in the
ing the relative reactivity of homologous homolytic oxidation of hydrocarbons to acids,
species. because it selectively catalyzes the splitting of
 Oxidation 69

the addition product of peracids and aldehydes In the oxidation of p-xylene to TPA the
(the aldehyde monoperester) into two mole- peroxy anion can alternatively react to form
cules of acid. the corresponding aldehyde (Eq. 200):
Prior to decomposition, the hydroperoxides
have to form complex compounds with the
metal ions. Therefore, the rates of these redox
reactions depend strongly on the medium.
In the series: alkyl hydroperoxide < alcohol
< water  acid, the hydroperoxides are the
weakest complexing agents. Therefore, alco-
hols, but in particular water and acids, act as
catalyst poisons by competing with the hydro-
peroxides for the metal ion complex sphere. If
these compounds are present in appreciable
amountsthat is the case in nearly every nor-
mal oxidationCo2/3 has to be added in
fairly high concentrations, e.g., in polar media 200
0.051 wt% and in nonpolar media 1500 ppm
(ca. 104105 mol/L).
Because water and acids act as catalyst (AcO acetate anion)
poisons, water and acids formed during the
reaction should be stripped off to as low con- The decomposition of dialkyl peroxides is
centrations as possible. Acids, which cannot be accelerated by Fe3, Cu, and complex com-
stripped off, will slow down the oxidation rate pounds of Mo6, W6, V6 and Ti4 (see
long before the hydrocarbon starting material is Section 5.1). Co2/3 and Mn2/3 decompose
consumed. In the special case of oxidation of p- dialkyl peroxides only under conditions which
xylene to DMT, the acid intermediate is esteri- are not relevant to autoxidation.
fied in order to achieve conversion of the Peroxy acids are decomposed analogously
second methyl group as. to alkyl hydroperoxides (Eq. 201):
The precise molecular mechanism of per-
oxide activation by cobalt complexes is not yet
fully understood. Recent experimental and
theoretical work suggests, at least in nonpolar
hydrocarbon environment, a slightly modified 201
mechanism (Reactions 197199) in which
Reaction 199 would be rate-determining.
Reaction 197 and 198 are computationally
predicated to be both fast due to spin catalysis, Because peroxy acids are stronger oxidizing
i.e., two-state-reactivity is suggested to signif- agents and much weaker reducing agents than
icantly lower the barrier of both steps [81]. At hydroperoxides, reduction of Co3 ions only
higher catalyst concentrations, the CoIII- occurs with aldehydes according to Reactions
OOR intermediate could alternatively form (195 and 196).
an inert dimeric complex with free CoII
species (i.e., termination), explaining a
Effect of the Catalyst and the Medium. The
remarkable catalystinhibitor transition,
metal ions have a major influence on the course
observed at higher cobalt concentrations.
of the reaction, as the following examples show.

CoII ROOH ! CoIII-OH RO (197)

The oxidation of acetaldehyde can be con-
CoIII-OH ROOH ! CoIII-OOR H2O (198) trolled to give considerable amounts of diacetyl
or acetic anhydride besides acetic acid, depend-
CoIII-OOR ! CoII ROO (199) ing on the nature and concentration of the
70 Oxidation

catalyst. In both cases the production of CO, esterification of the p-methylbenzoic acid (hin-
CO2, and CH4 is enhanced, too. drance by the acid formed) (Eqs. 206 and 207):
Oxidation in acetic acid with catalytic quan-
tities of Mn gives acetic acid.
Addition of catalytic quantities of copper
acetate to the Co2/3 catalyst leads to forma-
tion of acetic anhydride in addition to acetic 206
acid (Eqs. 201 and 202):

202

207

Single-stage oxidation of both methyl

203
groups succeeds in acetic acid in the presence
of 5001000 ppm Co and Mn acetate and either
a peroxy acid-forming cooxidizing agent such
as paraldehyde or acetaldehyde (Toray, East-
man-Kodak), or with Br ions as radical sources
(Mid-Century/Amoco).
The Cu cations seem to stabilize the acyl
Whereas the peroxy acid formed in the
cations and anions and the peroxyacyl radicals.
Toray or Eastman-Kodak process overcomes
Oxidation in the presence of large quantities
the low oxidizability of the second methyl
of manganese(III) acetate gives diacetyl, acetic
group, which is due to the carboxyl group, in
acid (and CO, CO2, CH4) (Eqs. 204 and 205):
the Mid-Century/Amoco process Co3 and the
Br radical act as abstractors for protons or H
atoms, respectively (Eq. 208):

204

205

Oxidation of p-xylene: pure p-xylene is mainly 208

oxidized to p-methylbenzoic acid in the pres-
ence of 100200 ppm Co and up to 100 ppm Mn
acetate in a normal radical chain reaction at
140180 C and 10 bar. The greater part of the The intermediate 4-carboxybenzaldehyde is
second methyl group can only be oxidized after rapidly oxidized further to give the acid.
 Oxidation 71

Table 11. Relative activity of various metal ions in the The effect of phenols and aromatic amines is
decarboxylation of carboxylic acids particularly well known. Because of their weak
Metal ion Relative activity OH and NH bonds, they readily transfer H
4
atoms to radicals. The resulting phenoxy or
Ce 1
Ag2 2
arylamino radicals are extremely unreactive
Pb4 20 due to resonance stabilization and, in the
Co3 100 case of 2,6-disubstituted phenols, because of
Mn3 500 the strong steric hindrance. The radical chain
reaction is thus interrupted, as shown in
Reaction (210):

Nevertheless, the removal of residual aldehyde

is the major problem in the purification of the
crude terephthalic acid (TPA), which crystal-
lizes out rapidly due to its insolubility in the
reaction medium and occludes small amounts
of aldehyde, which are thus protected from
further oxidation. Amoco solved this problem
by hydrogenating the carboxybenzaldehyde
with palladium as catalyst.
In the cooxidation process, only ca. 0.2 t
acetic acid per ton TPA are formed although the
cooxidant must be present in at least a stoichi-
ometric ratio.
210
In the Mid-Century/Amoco process 0.05 t
acetic acid are consumed per ton TPA.
One reason for the low yield and high loss of X Alkyl; not alkoxy
acetic acid is its metal-ion-catalyzed An inhibitor molecule can thus capture two
decarboxylation (Reaction 209). The methyl radicals. The relative efficiency of the inhibi-
radical formed is transformed to methanol tors depends strongly on the substituents
and finally to methyl acetate, which is the (Table 12).
largest by product in this process.
The decarboxylating activities of some
metal ions are listed in Table 11. The reaction Table 12. Relative inhibitor effect of trialkylphenolsa on the
(209) is fast above 140 C. oxidation of crude oil [338, p. 50]

R1 R2 R3 Relative inhibitor effect

H H Me 10
tert-Bu tert-Bu tert-Bu 36
209 Me Me tert-Bu 46
tert-Bu tert-Bu sec-Bu 80
tert-Bu tert-Bu Me 100
Me Me Me 118
tert-Bu tert-Bu Ethyl 125
tert-Bu tert-Bu n-Bu 140
Me tert-Bu Me 170
5.4.3. Inhibitors of Homolytic Liquid-Phase
Oxidation

Inhibitors are substances that hinder the forma-

tion of the active intermediates (radicals,
hydroperoxides, catalysts) or lower their con-
centration by complexing, decomposing, or
trapping them. 
2,6-Di-tert-butyl-4-methylphenol 100
72 Oxidation

Whereas phenols and amines mainly capture with the radical chain oxidation has apparently
radicals, organic sulfides and thiols usually not yet been investigated.
react with hydroperoxides as follows (Eq. 211): For an introduction to autoxidation and the
effect of inhibitors and stabilizers, in particular
with reference to the protection of polymers
from oxidative aging, see [338].

211 5.4.4. Special Cases of Homolytic Oxidation

Boric-Acid-Modified Alkane Oxidation (!
Alcohols, Aliphatic). The oxidation of hydro-
carbons in the presence of suspended metaboric
acid, known since 1928, appears to be a border-
For this reason mixtures of organic sulfides line case between homolytic and heterolytic
and phenols or amines have a stronger inhibit- oxidation. Secondary alcohols are preferably
ing effect than the individual components formed with unchanged carbon frameworks
(inhibitor synergism). (selectivity > 80%), together with ketones
The inhibiting effect of an excessively high and diols. In the presence of olefins, epoxides
metal concentration has already been men- are also formed in considerable yields.
tioned (see Section 4.4.2). Compounds that The way of action of the boric acid is not
form metal complexes can also act as inhibitors clear yet. Until the 1970s it was believed that
if the number of donor centers corresponds to boric acid reacts only with the alcohol to form
the coordination number of the metal. For boric ester, which are very resistant to further
example, disalicylidene-1,2-propylenediamine oxidation. Later on epoxides were found, that
is added to gasoline in concentrations of are formed during the oxidation of olefins in the
108 mol/L because it forms stable complexes presence of boric acid. This led to the assump-
with metals, thus preventing them from cata- tion that the boric acid already catalyzes the
lyzing autoxidation. decomposition of the hydroperoxides. But,
because in the oxidation of paraffins the ratio
of secondary alcohols to ketones is very high
(nearly no Russell reaction), it seems very
likely that the boric acid reacts directly with
the peroxy radicals to the corresponding boric
esters and hydroperoxy radicals.
After oxidation, the alcohols must be liber-
Disalicylidine-1,2-propylenediamine ated from the boric esters by hydrolysis. Only
It is noteworthy that phosphates, especially one OH group of the boric acid reacts with the
di- and triphosphates, appear to inhibit radical alcohol to give the ester (Eqs. 212214).
chain oxidation. Pyrophosphates and in partic-
ular dihydrogenpyrophosphates are used as sta-
bilizers for aqueous hydrogen peroxide,
whereby they inhibit its radical chain decom- 212
position, mostly by effectively chelating tran-
sition metals that otherwise would catalyze the
reaction. They are not only the most effective
fire-extinguishing powders (in the form of
ammonium hydrogen phosphate), and, as pyro-
phosphate (P2 O2 213
7 ), a component of many het-
erogeneous gas-phase oxidation catalysts, but
are also components of important building
blocks of all living cells in the nucleotides 214
and nucleic acids. In what way they interfere
 Oxidation 73

In Reactions (213) and (214) the protons of the Chain propagation

boric acid depicted in parentheses do not
behave like normal protons. They do not take
part in esterification.
Alkyl aromatics cannot be oxidized in this
way because boric acid also catalyzes the Hock
cleavage to give phenols, which inhibits further
oxidation shortly after initiation.

Paraffin Sulfonation (! Surfactants). A fur- 216

ther special case is the paraffin oxidation pro-
cess in the presence of SO2. Of several
variations, two are carried out industrially.

Sulfochlorination (Reed Process). Sulfur diox-

ide and chlorine are introduced into the paraffin
at 2535 C in the presence of UV radiation.
The reaction proceeds by a radical chain mech- Decomposition of peracid
anism with alkylsulfonyl chloride (i.e., chlor-
osulfonic acid) as the final product.
The kinetic chain length is 40005000. The
conversion is limited to 30% to minimize
further oxidation to di- and polysulfonic
acids. 217

Sulfoxidation (LightWater Process). If SO2

and O2 (in a molar ratio of 2:1 to 10:1) are
passed through an n-alkane in the presence of
UV radiation at 2040 C, alkyl persulfonic acid Remarkably, SO2 adds faster to the alkyl
is initially formed in a radical chain reaction. In radical than O2. Excess of SO2 is therefore
pure paraffin, the persulfonic acid reacts further important. The reactivity of the alkylpersul-
to give dark colored products. If water is added fonyl radical formed by subsequent addition
to the alkane, the reaction is slower and for each of oxygen is similar to that of the alkylperoxy
mole alkyl sulfonic acid an equivalent quantity radical. The sulfonic acid group is thus situated
of sulfuric acid is formed. However, the product predominantly at carbon atoms in the middle of
remains colorless because the water extracts the the chain.
alkylpersulfonic acid and catalyzes its reaction Aromatics, olefins, and branched-chain par-
with SO2 to give alkylsulfonic acid and sulfuric affins inhibit the reaction.
acid. The alkane conversion is limited to ca.
1% to keep the undesired polysulfonation low
5.4.5. Process Technology of Homolytic Liq-
(Eqs. 215217).
uid-Phase Oxidation
Initiation
The various starting materials (liquefied
petroleum gas, low-boiling aldehydes to
high-boiling, long-chain paraffins, alkylhy-
droanthraquinone etc.) and the widely differ-
215 ing products (from H2O2 to terephthalic acid)
also require differing conditions (temperature,
pressure, solvents, reactor material). Some
74 Oxidation

general rules may be given, which are valid for Apart from the O2 partial pressure, the rate
most radical chain oxidations. of O2 transport from the gas to the liquid phase
depends only on the area of the phase boundary
Reaction Parameters, Reactor Geometry, surface, i.e., the bubble size. Therefore, the gas
Operation. The aim of industrial reactions is must be divided into the smallest bubbles pos-
to produce as much product as possible per unit sible on entry into the reactor and these must be
time and reactor volume. Because the mass prevented from coalescing into larger bubbles
flow of the educts, reaction temperature, and while they rise up the liquid column.
catalyst concentration can be chosen freely If turbulence is too severe and there is
within limits, these parameters are normally consequent coalescence, large gas bubbles are
maximized to give high throughput and produc- formed and the oxygen cannot pass into the
tion numbers. Consequently, O2 transport from liquid phase quickly enough. Oxygen then
the gas to the liquid phase becomes the rate- appears in the waste gas, although the oxidation
determining step, often at the expense of lower suffers from an O2 deficiency.
product selectivity. This is even worsened by The gas velocity based on the reactor cross-
the fact that for safety reasons the O2 content of section should thus clearly lie below 0.02 m/s.
the off-gas is usually kept near to 0 vol%, At higher velocities turbulence arises in the
resulting in O2 deficiency in the upper part bubble column that promotes coalescence of
of the bubble column. the gas bubbles.
To test whether an oxidation is controlled by The height of the liquid phase in the reactor
O2 transport, it is sufficient to increase the O2 above the air inlet must be adapted to the rate of
pressure and to see if the reaction rate subse- O2 take-up. It can be determined in pilot experi-
quently rises. According to Equation (157) if ments (see below). Generally 46 m should be
there is an adequate O2 supply the reaction rate sufficient. The volume required for the neces-
is independent of the O2 concentration and thus sary throughput should be obtained by increas-
of the O2 transport. The solubility of O2 in ing the cross-section of the reactor (tanks
organic solvents is only ca. 102 mol/L (see instead of columns). In industry, for historical
Table 13). reasons, most oxidation reactors in operation
are much higher (up to 30 m and more with
liquid heights of 8 to more than 20 m). The
Table 13. Solubility of oxygen in organic compounds at 1 bar reaction pressure should be chosen in a range
[385, p. 1 ff] where boiling of liquids is avoided because the
Solvent T,  C O2 concentration, a vapor formed would enlarge the gas bubbles
103 mol/L and lower the O2 concentration in the bubbles,
Benzene 20 9.1 0.204
thus hinder oxygen transport.
Benzene 60 9.6 0.216 Reactors for radical chain reactions must
Toluene 20 5.3 0.119 have the smallest possible surface to volume
p-Xylene 26 8.3 0.118 ratio because of the undesired radical
p-Xylene, tech. 23 7.2 0.162
p-Xylene, tech. 80 8.0 0.179
recombination that always occurs at the walls.
p-Xylene, tech. 100 8.6 0.192 Apart from necessary built-in devices, like
Isooctane 20 13.9 0.312 sieve plates for gas redispersion and cooling
Isooctane 50 14.6 0.327 tubes, packings and devices with large surface
Dodecane 25 8.25 0.185
Methanol 20 10.5 0.235
areas should not be used in the reaction section.
Ethanol 20 6.4 0.143 However, in the gas space above the bubble
Acetone 20 9.6 0.215 column packings with large surface areas are
Cyclohexanol 116 9.5 0.213 recommended to suppress gas explosions.
Cyclohexanone 100 10 0.224
Benzaldehyde 5 4 0.090
Oxidation can be operated as a batch or
Decanal 5 4 0.090 continuous process. In a continuous process

the residence time spectrum is broadened, as
a cm3 O2 (STP) cm3 bar1.

The determination of O2 solubility requires a long time for
part of the fresh starting material is carried
equilibrium to be reached. Hence, the values for readily oxidizable straight out again because of the generally
substances such as aldehydes are probably too low. strong back-mixing. Accordingly, part of the
 Oxidation 75

product remains longer in the reactor and is Copper can inhibit many oxidations even in
overoxidized. The selectivity is thus lower than extremely low concentrations. Because copper
in batch processes. For this reason many large- is contained in fine dispersions in some lubri-
scale oxidations are still performed batchwise cating agents and slip additives, particular
even today. The selectivity in continuous oxi- attention should be paid to such substances in
dations can generally be improved by using the case of initiation problems.
baffles (sieve plates, etc.), which convert the In all parts of a plant in which O2 is absent,
reactor into a bubble column cascade and lower i.e., where reducing conditions prevail, special
back-mixing. care must be taken to ensure that the reactor
material and the temperature are mutually com-
Reactor Materials; Corrosion. The choice of patible. For example Material no. 4571 (V4A E;
reactor materials is determined by the reaction US specification: SS 316 Ti) is rapidly attacked
medium and conditions (temperature, pressure, by acetic acid in the absence of O2 above
etc.). Most base metals and alloys are only 140 C. In the presence of O2 it is stable towards
protected from chemical attack by a dense, the acid up to and above 160 C.
tightly adhering layer of oxide (passivation). The corrosivity of organic materials
These materials (Al alloys, Ti, stainless steel) increases with the acidity, reducing capacity,
are generally stable towards oxidizing media, and capacity for complex formation. Besides
but are readily attacked by reducing media (see formic acid, hydroxyacids such as lactic and
Fig. 9). glycolic acids are the most corrosive.
Because many heavy metal ions also have a
catalytic effect, aluminum and its alloys are Oxidation on Laboratory and Pilot Plant
recommended for the oxidation part under Scale. An apparatus is shown in Figure 10
moderate conditions of temperature and pres- which is suitable for most laboratory scale
sure. For higher temperatures and pressures, homolytic oxidations. Some aspects to which
stainless steel is recommended. Stainless steel attention should be paid are as follows:
is, however, rapidly attacked by halide ions
even in an oxidizing medium. Therefore, for
example, for the Mid-Century process reactors
coated with titanium or Hastelloy C must be
used.

Figure 9. Corrosion resistance of various materials Figure 10. Laboratory apparatus for liquid-phase oxidation
The effective corrosion resistance of each material in an of high-boiling compounds
acid environment is that portion of the chart below each bar a) Thermostats; b) Gas metering; c) Cold traps; d) Off-gas
(with kind permission of Gulf Publishing Comp., taken cooler; e) Water separator; f) Gas sampling; g) O2 analysis;
from Hydrocarbon Processing, June 1979, p. 214). h) CO/CO2 analysis; i) Gas meter
76 Oxidation

Reactor dimensions: ten times the volume of the vessel to obtain

correct values.
Laboratory Pilot plant
Diameter, cm >5 > 30 Course of a Liquid-Phase Homolytic Oxida-
Height, cm > 30 up to 800
tion. Batch and continuous oxidation reactions
can be initiated as follows. The carefully
In larger pilot reactors back-mixing can be cleaned reactor is charged with the starting
limited by the incorporation of, for example, material and the catalyst under nitrogen purg-
sieve plates. ing, the pressure is adjusted to the required
value, and the mixture heated to the initiation
Reactor Material. In the laboratory, glass temperature. Then air or O2 is introduced and, if
should be used if possible (so that the reaction necessary, an initiator added. The start of the
can be observed). With metals, oxygen-free reaction can be detected by the decrease in the
zones should be avoided because of possible O2 content of the off-gas and a rise in the
corrosion, possibly by O2 purging. reaction temperature. From this time onwards
Reaction temperature is the most important the O2 supply is increased in such a way that the
parameter; constant temperature can best be O2 content in the off-gas does not exceed a
achieved by using two heating circuits: one critical limit, for instance 5%. The acceleration
at the desired reaction temperature for heating of the reaction to maximum velocity usually
the jacket and one for inner cooling which is occurs rapidly (within 10 to 30 min). The heat
kept 1050 C below the reaction temperature. of reaction must be removed and the desired
The size and shape of the cooling circuit reaction temperature regulated. With proper
depend on the reaction and dimensions of the layout, the temperature of the cooling medium
reactor. Cooling tubes should be arranged ver- and the control function of the valve in the
tically and sufficiently far from one another and cooling circuit can be adjusted on installation of
from the wall to minimize disturbances to the the system such that the oscillations of the
gas distribution. reaction temperature are maintained below
0.1 C. For batch oxidations, the temperature
Gas distribution is the second most important of the cooling circuit must gradually approxi-
parameter. Tubes with small holes or frits made mate the reaction temperature towards the end
of glass, Teflon, or metal are the most suitable. of the reaction, to avoid excessive temperature
For metal reactors, the frits should be tested fluctuations or breakdown of the oxidation
periodically in glass vessels under conditions process. In continuous oxidations, charging
similar to those of the reaction. of starting material and removal of the product
begin after the required conditions have been
Oxygen Content of the Off-gas. At 0% oxygen established.
content the experimental results are uncertain
because the size of the O2 deficiency zone is
unknown. If higher O2 contents are permitted in Applicability of the Results from Laboratory
the off-gas, the necessary safety measures must and Pilot Plant to Full Plant Scale. Scale-up
be adhered to very precisely (see Section 4.6). problems are mainly caused by the following
factors:
A water separator is important in the oxidation
of nonpolar substances and is recommended for 1. The laboratory and pilot plant reactors often
other substances if water is formed as a separate have too small a diameter relative to their
phase. height. Because the rising gas bubbles
entrain liquid, a downward streaming in
Gas Sampling. The O2, CO, and CO2 content the rest of the liquid is always produced,
in the off-gas should be continually monitored predominantly at the wall of the reactor. The
if possible; attention should be paid to volatile resulting turbulence leads to rapid coales-
substances (e.g., H2, CH4, CH3CHO). The gas cence of the gas bubbles in reactors that are
sampling vessel should be purged with at least too narrow. Thus the exchange area becomes
 Oxidation 77

much smaller and the reaction appears to be Because the intermolecular interactions, in
slower than it is. particular through hydrogen bond formation,
2. The O2 content of the off-gas is kept near are considerably weaker in the gas phase than in
0% (for further explanation, see above). the liquid phase due to the greater inter-
molecular distances, in the absence of catalysts,
Results that are representative for full plant hydroperoxide decomposition can only occur
scale are only obtained with reactors which are thermally. Because the activation energy for
of the same height as large-scale reactors and thermal hydroperoxide decomposition is high
have a diameter such that the gas velocity (see Section 4.4.2), gas-phase oxidations only
remains well below 0.02 m/s (diameter 30 proceed at the same rate as a comparable liquid-
cm minimum). phase oxidation at considerably higher temper-
atures. However, because heat removal is poor,
an initiated gas-phase oxidation accelerates
considerably faster than a corresponding liq-
5.5. Homolytic Gas-Phase Oxidation uid-phase oxidation and can readily become a
thermal explosion.
Homolytic gas-phase oxidations are all com- Up to ca. 250 C, hydroperoxides predomi-
bustion processes and gas explosions that take nate as the first neutral intermediates. Between
place in free gas space without heterogeneous ca. 300 and 450 C, isomerization and decom-
catalysts. Examples include the combustion of position of the peroxy radicals to give alde-
heating oil and natural gas for the production of hydes and other fragments becomes faster than
heat and energy and the rapid combustion of hydroperoxide formation. Thus above ca.
gasoline and diesel in internal combustion 420450 C practically no more alkyl hydro-
engines. Thus the homolytic gas-phase oxida- peroxides are formed, and aldehydes and alco-
tion is one of the most important chemical hols together with olefins and H2O2 are the
reactions with respect to the amounts of mate- primary neutral intermediates.
rials consumed. The decomposition of peroxy radicals, usu-
ally with preceding isomerization, has a higher
Mechanism. In principle up to 250 C homo- activation energy than the hydroperoxide
lytic gas-phase oxidations proceed by the same decomposition. Therefore, in the transition
mechanism as liquid-phase oxidations (see range between 320 and 450 C, the overall
Section 4.3). reaction rate of gas-phase oxidation decreases
However, whereas oxygen transport from with increasing temperature (region of negative
the gas to the liquid phase generally limits ratetemperature coefficients), and subse-
the rate of the overall reaction in a controlled quently rises more sharply than before.
liquid-phase oxidation, no comparable restric-
tion exists in the gas phase. In the liquid phase Catalysis and Inhibition. Gas-phase oxida-
the reaction temperature can be held constant at tions are catalyzed by radical-forming sub-
low levels by simple means. The rise in tem- stances (halogens and their compounds, NO/
perature is inversely proportional to the mass of NO2, nitro compounds, volatile peroxides, phe-
the reaction mixture, and the heat of reaction is nols, and amines) and also by high-energy
quickly transported to the walls of the cooling radiation and cracking or dehydrogenation cat-
tubes by convection of the liquid. For reactions alysts on the reactor walls.
in the homogeneous gas phase, the temperature In the gas phase, radical-forming sub-
difference for adequate heat removal must be stances generally only act catalytically if
much greater because of the small mass. If the they are added in very low concentrations.
reaction is not limited by the quantity of fuel or In higher concentrations (sometimes even in
oxygen, and the reaction space is not kept very the ppm range), the radicals formed act as
small and well cooled, temperatures of over radical scavengers for the more reactive
1000 C can occur in gas-phase oxidations, hydrocarbon radicals. The mechanism
whereby heat transfer to the surroundings of action of halogen compounds as fire-
occurs mostly by radiation. extinguishing agents and as flameproofing
78 Oxidation

agents in plastics, and that of tetraethyl lead 2. Carbon black (rubber filler and strength-
used as an antiknock agent in gasoline up to ener) by combustion of oil or natural gas
the 1980sis based on this radical trapping. with a substoichiometric amount of oxygen
Chain termination by recombination of (see ! Carbon, Section 6.4).
smaller, and therefore highly energetic, active 3. Acetylene by partial combustion of hydro-
radicals (such as H , OH , CH3 ) can occur in carbons (! Acetylene, Section 4.2).
the gas phase only if a third collision partner 4. CO2 by complete combustion of natural gas.
absorbs the energy of recombination (bond
energy) liberated. Apart from the very rare The degradation of hydrocarbons and chlor-
three-body collisions, this occurs mostly ohydrocarbons in the atmosphere takes place by
through collisions with surfaces. Thus gas- radical chain reactions.
phase oxidations, like all radical reactions
occurring in the gas phase (e.g., decomposition
of acetylene or ethylene oxide), are sensitive 5.6. Gas Explosions and Safety Data
towards an increase in surface area. This is the
mechanism of action of packings as flame Modern civilization depends to a great extent
arrestors, powder fire extinguishers, and dusts on the use of hydrocarbons. However, great
to control firedamp in mining. Degenerate rad- dangers can arise from these hydrocarbons in
ical chain explosions that proceed via hydro- the presence of O2 (see Fig. 11). Consequently,
peroxides (e.g., in hydrocarbonoxygen an extensive body of safety data has been
mixtures) can be stopped if fire-extinguishing generated to keep these dangers within accept-
powder or other radical scavengers are added able limits and make the handling of hydro-
early enough. This is because that due to carbons less hazardous. Extensive literature,
indirect chain branching, the pressure rise dur- standards and regulations exist in all industrial
ing the initiation period is so slow, that suffi- countries [358367]. The flammability of gas
cient time remains for releasing the fire-
extinguishing device. But after the initiation
period temperature and pressure can rise very
sharply, changing the slow oxidation in a ther-
mal explosion. These types of thermal explo-
sion can rapidly become detonations, especially
in enclosed spaces such as pipelines.
Nondegenerate branched radical chain
reactions (such as the H2/O2 reaction), on the
other hand, accelerate so rapidly due to direct
chain-branching that no time remains for extin-
guishing measures once the reaction has started
(see Fig. 3). They can only be hindered by
preventive measures such as filling spaces
with packings or capillary bundles.

Syntheses of Intermediates by Homolytic Gas-

Phase Oxidation. Because of the extreme con-
ditions, only a few simple products are pro-
duced by oxidation in the homogeneous gas
phase. These include:

1. Synthesis gas (or hydrogen) by combustion

of hydrocarbons or coal with a substoichio-
metric amount of oxygen and in the Figure 11. Characteristic damage caused by an atmo-
presence of water (! Gas Production, Sec- spheric shock wave reflected from an object as a function
tion 3.2; ! Gas Production, Section 4.1). of the peak pressure [359, p. 60]
 Oxidation 79

mixtures [383] and the effect of inert gases limit is extended significantly by increasing
[384] have been reviewed. either the pressure or the ignition energy, the
The various safety terms are explained in lower explosion limit is lowered only to a small
[359], which also gives the corresponding stan- extent.
dards for France, Germany, Russia, the United Combustible dusts behave in a manner simi-
Kingdom, and the United States. lar to vapors [360].
In enclosed spaces (tanks, silos) the O2
Explosion Limits, Explosion Groups, Ignition concentration can be lowered by dilution
Temperatures, Temperature Classes, Flash with inert gases (e.g., N2, CO2) to such an
Points, and Hazard Classes. All hydrocarbon extent that an explosion can no longer be set
vapors, and also many other substances, form off (see Table 14).
explosive gas mixtures with oxygen within Above a certain temperature, the mixture
certain concentration limits, and these mixtures ignites without an external ignition source due
can be made to ignite if a certain, usually quite to autoxidation. This ignition temperature
small amount of energy (electrical sparks, hot decreases with increasing chain length (see
surfaces, etc.) is supplied. The lower and upper Table 15). It also decreases, in particular in the
limiting concentrations of combustible gas at case of hydrocarbons, the longer the gas mix-
which an explosion can no longer be set off are ture has been kept at an elevated temperature
called the lower and upper explosion limits. below the ignition temperature (see Section 4.5).
In [359, p. 65] the opposite and therefore In pure oxygen the ignition temperatures can be
incorrect definition of the explosion limits is up to 300 C lower than in air. Like the explosion
given. limits, the ignition temperatures are not sub-
Figure 12 shows the explosion limits of stance-dependent constants but depend to a great
propane in air as a function of the pressure extent on the measuring conditions. Table 15
of the gas mixture before ignition and the lists explosion limits and ignition temperatures
ignition energy. Whereas the upper explosion for some compounds in air at atmospheric pres-
sure and 20 C.
For safe handling of hydrocarbons being
processed or stored in a plant the surfaces of
hot apparatus, tubing, machines, etc. must be so
well insulated that no danger of ignition can
arise. For this reason combustible substances
are classified according to ignition temperature
into temperature classes T1 to T6 (see
Table 16).
Also, areas where there is a danger of explo-
sion are classified into areas according to the
probability of the occurrence of explosive gas
mixtures [359]:

Zone 0 Areas in which an explosive atmosphere is always

present or present for an extended period of time.
Zone 1 Areas in which an explosive atmosphere is
occasionally present.
Zone 2 Areas in which an explosive atmosphere is rarely
present or present for only a short time.

Figure 12. Explosion limits of propane in air as a function Whereas hydrogen is diluted very quickly in
of the pressure p prior to ignition and the ignition energy E
in a 7-L vessel
air because of its low density and high rate of
Exl lower explosion limit; Exu upper explosion limit diffusion, hydrocarbons from propane onwards
Reproduced from [358] with permission of the publisher sink to the ground because of their relatively
80 Oxidation

Table 14. Limiting values for the inerting of flammable gases and vapors with N2 or CO2 at 20 C and 1 bar [359]

Compound Minimum volume ratio Maximum O2 concentration in the FAinert gas

mixture for inerting with N2 or CO2

of inert gas to fuel (F) for of inert gas to air (A) for N2 CO2
inerting towards addition of air inerting towards addition of fuel

N2 CO2 N2 CO2 O2 (max), vol% O2 (max), vol%

Hydrogen 17 10 3.00 1.56 5.0 5.1

Carbon monoxide 4 2.2 2.13 1.13 5.4 5.4
Methane 6 3.3 0.61 0.34 12.1 14.6
Ethane 13 7.5 0.82 0.49 11.0 13.3
Ethylene 16 9 1.00 0.67 10.0 11.7
   
Ethylene oxide 17.2 15.5
Propane 15 8 0.75 0.43 11.8 14.2
Propene 14.5 8 0.75 0.43 11.5 14.1
Cyclopropane 15.5 8 0.75 0.45 11.7 13.9
Butane 17 9.5 0.70 0.39 12.1 14.5
Butadiene 19.5 13 0.89 0.51 10.4 13.0
Hexane 25 14 0.72 0.41 12.1 14.5
Benzene 21 13 0.79 0.45 11.2 13.9

Values do not exist due to ready decomposition of ethylene oxide.

high density and low rate of diffusion. It has inside the equipment is just unable to penetrate
been shown experimentally that 10 m3 of liquid to the outside.
propane evaporated within 10 min and distrib- Substances are divided into three explosion
uted itself in a layer of 0.7 m thickness over an groups corresponding to their hazard and
area of 700 m diameter. In addition the lower assigned particular MESG values (see
explosion limit of hydrogen is 4 vol%, while Table 18). The MESG is smaller than the
that of hydrocarbons lies between 1 and 2 vol%. quenching distance because, unlike the explo-
Hydrocarbons are at least as dangerous as sion barrier, the gap has a depth of only 25 mm.
hydrogen. The vapor pressure of liquids increases
Explosions cannot occur in spaces occu- with temperature. The temperature at which
pied by packing material. Explosions which the vapor pressure of organic liquids
have already started are quenched in vessels or exceeds the lower explosion limit so that
pipes filled with Raschig rings, tube bundles, the vapor/air mixture above the liquid can
or similar packing materials (explosion barri- be ignited, is defined as the flash point.
ers). The critical diameter of the capillary Because dangers can arise, particularly from
tubes at which no flame can be propagated low-boiling substances, they have been
is defined as the quenching distance. The divided into hazard classes according to flash
quenching distance depends above all on the point (Table 19). Because of the complexity
combustible gas, but also on the material, and great number of regulations, the hazard
temperature, and mass of the explosion bar- classes according to the UN Recommendation
rier. Capillaries and bead fillings are more which have arisen from the IATA and
effective flame barriers than parallel plates RID regulations are becoming increasingly
(see Table 17). important.
Closely associated with the quenching dis-
tance is the maximum experimental safe gap Self-Ignition of Solids, Fires from Liquids.
(MESG), which principally applies to electrical Similar to gases, heat removal from solids is
equipment in surroundings where there is a poor. In particular, with solids having large
danger of explosions. The safe gap is defined surface areas (coal dust, flour, hay, straw, paper
as that through which an ignition occurring waste, wood shavings etc.) oxidation can result
 Oxidation 81

Table 15. Explosion limits and ignition temperatures of some compounds (reproduced from [362] by kind permission of Deutscher
Eichverlag)

Compound Explosion limits in air (20 C, 103.3 kPa) Self-ignition

temperature,  C
Lower, vol% Upper, vol% Lower, g/m3 Upper, g/m3

H2 4.0 75.6 3.3 64 560

CO 12.5 74 145 870 605
NH3 15 28 105 200 630
H2N-NH2 4.7 100 60 1265 (270)
CH4 5.0 15 33 100 595
C2H6 3.0 12.5 37 155 515
C2H4 2.7 28.5 31 330 425
C2H2 1.5 100 16 1080 305
C3H8 2.1 9.5 39 180 470
C3H6 2.0 11.7 35 210 455
n-C4H10 1.52.0 8.5 3749 210 365
C(CH3)4 1.3 7.5 40 230 (450)
n-C5H12 1.4 7.8 41 240 285
n-C10H22 0.7 5.4 41 320 205
C6H5-C6H5 0.7 3.4 45 220 570
C6H6 1.2 8.0 39 270 555
C6H5CH3 1.2 7.0 46 270 535
CH3C6H4CH3 1.1 7.0 48 310 525
CH3OH 5.5 3144 73 410590 455
(CH3)2CHOH 2.0 12 50 300 425
n-C3H7OH 2.1 13.5 50 340 405
(C2H5)2CHOH 1.2 8 53 300 360
n-C5H11OH 1.3 10.5 47 380 300
CH2O 7.0 73 87 910 (420)
CH3CHO 4.0 57 73 1040 140
CH2CHCHO 2.8 31 65 380
(CH3)2CHCHO 1.6 11 47 320 165
CH3Cl 7.1 18.5 150 400 625
CH3Br 8.6 20 335 790 535
CH2CHCH2Cl 3.2 11.2 105 360 390
CH2CHCH2Br 4.3 7.3 215 370 (295)
n-C5H11Cl 1.41.6 8.6 6070 380 255

in a significant temperature rise, whereby igni- In the case of liquids it is also only the vapors
tion occurs (see Table 20). The ignited solid formed which burn. For this the liquids must be
only glows (reaction between O2 and solids). heated to temperatures above the combustion
Flames are formed by gas-phase oxidation of point. At the flash point only the gas mixture
the gases and vapors formed by cracking and above the liquid burns and the flame is then
evaporation of the volatile compounds. extinguished; however, when the combustion

Table 16. Criteria for temperature classes [363] Table 17. Quenching distances of various flame barriers [358, p.
58]
Ignition temperature Temperature Maximum Former
of flammable class surface ignition Fuel Parallel plate Flame barrier
material,  C temperature,  C class wmax, mm
Capillary, Bead filling
>450 T1 450 G1
dmax, mm d max, mm
>300 T2 300 G2
>200 T3 200 G3 H2 0.3 1 2
>135 T4 135 G4 CH4 1.3 4 7
>100 T5 100 G5 C3H8 0.9 3 6
>85 T6 85

Bead diameter.
82 Oxidation

Table 18. Classification of explosion groups according to extinguishers are due to the same cause: These
maximum experimental safe gap (MESG) and minimum ignition hydrocarbons form too many radicals. There-
current ratio (MIC) [363]
fore, at low temperatures, the proportion of
Classification Condition Example chain termination reactions is too high.
II A 0.9 mm < MESG propane
Particularly critical situations can arise from
0.8 < MIC fires involving containers of low-boiling
II B 0.5 mm  MESG  0.9 mm ethylene liquids, if the container is heated. This can
0.45  MIC  0.8 lead to sudden evaporation of a large quantity
II C MESG  0.5 mm H2, C2H2,
MIC  0.45 CS2
of liquid and a subsequent explosion. This
process is known as BLEVE (burning liquid
evaporating vapor explosion). A decision on the
particular procedure can only be taken on the
spot. But in all cases where the liquid can
point is reached so much liquid evaporates that neither be pumped off nor quenched by dilution
the gas mixture continues to burn. The com- with water, a BLEVE should be hindered by
bustion point is usually a few degrees above the intensive cooling of the container. The liquid
flash point. Halogenated hydrocarbons, for which could otherwise not be disposed of
which the combustion point can be up to should, however, be allowed to burn away, since
100 C above the flash point, are an exception. clouds of gas forming after complete quenching
This striking behavior and the suitability of by subsequently evaporating liquid could ignite
halogenated hydrocarbons to be used as fire at places which were still hot.

Table 19. Flash point limits of various transportation and storage regulations [359]

Regulation Applies to Class Flash point limit,  C Notes

VbF storage AI < 21 incompletely miscible

A II 2155 with water
A III 55100
B < 21 completely miscible with water
IATA (RAR) air transport < 37.8 measured in closed cup
37.893.3
IMDG Code sea transport 3.1 < 18
3.2 18 to 23
3.3 23 to 69
RID, EVO, ADR, transport by rail, road, Materials of UN 21
ADNR, etc. and canal class 3 55
100

Table 20. Smoldering temperature of deposited dust (5 mm thick layer) [359]

Material Particle size Bulk density, kg/L Smoldering temperature,  C

Majority, mm Max., mm

Coal dust 510 50 0.41 225

Charcoal 12 20 0.36 340
Flour 3050 200 0.31 325
Phosphorus, red 3050 150 0.99 305
Polystyrene 4060 150 0.23 melts
Poly(vinyl chloride) 45 10 0.55 chars
Zinc 1015 90 4.9 430

Temperature of the surface on which the dust lies. It decreases with increasing thickness of the dust layer. The maximum permissible surface
temperature must be ca. 100 C lower than the smoldering temperature.
 Oxidation 83

6. Heterolytic Oxidation 6.1. Heterolytic Oxidation in the Liquid

Phase
Both liquid- and gas-phase heterolytic oxida-
tions have long been known and are used both In heterolytic oxidations in the liquid phase
in the laboratory and in industry. other oxygen donors are frequently used instead
For example, the oxidation of naphthalene to of molecular oxygen. Some of these are listed in
phthalic anhydride was carried out by BASF Table 21, together with the byproducts formed
with sulfuric acid and K2Cr2O7 since 1872, and from them.
since 1896 by the Sapper process with sulfuric
acid and mercury as the catalyst. From 1916 Example 1: Uncatalyzed Epoxidation of Ole-
this process was replaced by air oxidation on fins by Peroxy Acids (Prilezhaev Reaction).
V2O5 catalysts. Due to the polarization of the neighboring CO
Oxidations with HNO3, SO3/H2SO4, perox- bond, the hydroperoxy group of peroxyacids can
omonosulfuric acid, MnO2, KIO4, etc., still epoxidize CC double bonds (Eq. 218). Alkyl
play a role in industry today [344347], hydroperoxides are less electrophilic and require
although, with the exception of HNO3, this is a catalyst (see Example 3 b).
limited to products with a very small volume.
A more systematic investigation and deter-
mination of these reactions began in the 1960s,
principally as a result of two developments:

1. The oxidation of ethylene to acetaldehyde

with PdCl2, developed by Wacker [386] and
Hoechst.
2. Molybdenum-catalyzed epoxidation of pro- 218

pene with tert-butyl hydroperoxide, intro-

duced by Halcon and Arco, and their joint
venture, Oxirane.
Example 2: Stoichiometric Oxidation with
Oxidation in the presence of heterogeneous Metal Oxides. Oxides such as SeO2, CrO2Cl2,
catalysts is still based almost entirely on empir- MnO 4 , OsO4, etc. have long been used as
ical principles (both in the liquid and gas oxidizing agents for organic substrates. The
phases). It profits considerably from develop- following mechanism is assumed for the oxi-
ments in homogeneous liquid-phase oxidation dation of propene by SeO2 (Eqs. 219221):
since the course of reactions on solid surfaces
cannot yet be directly investigated. It may be
assumed, however, that many reactions on het- 219
erogeneous catalysts proceed in the same or at
least similar ways with corresponding catalysts
in the homogeneous liquid phase.
Depending on the substrate, oxidizing agent,
Table 21. Oxygen donors [342, p. 5]
type of catalyst, participating complexing
agents, reaction medium, temperature, and Donor Active oxygen, % Byproduct
the other reaction conditions, heterolytic oxi- H2O2 47.0 H2O
dations proceed by various mechanisms. From tert-BuO2H 17.8 tert-BuOH
the combination of different substrates with the NaClO 21.6 NaCl
NaClO2 35.6 NaCl
different catalysts, oxygen donors, and reaction
NaBrO 13.4 NaBr
media, there is an extraordinarily large number C5H11NO2 a 13.7 C5H11NO
of possible reactions. For this reason only a KHSO5 10.5 KHSO4
short overview can be given here, based on NaIO4 29.9b NaI
[342]. Further reviews can be found in [339, a
N-Methylmorpholine-N-oxide.
343, 345348]. b
Assuming all four O atoms are utilized.
84 Oxidation

Example 3: Transition Metal Catalyzed Oxi-

dation of Olefins.
220 a. Some transition metal oxides form
inorganic peroxo acids with H2O2, analo-
gous to the reactions shown in the above
scheme. These inorganic peroxoacids,
known as Milas reagents, oxidize olefins
to epoxides. The epoxide formed is hydro-
221 lyzed in water (H2O2 is usually added as a
30% aqueous solution) (Eqs. 222 and 223):
In the presence of tert-butyl hydroperoxide
(TBHP) and small quantities of SeO2 the
reaction proceeds catalytically (see also Exam-
ple 3).
The active oxidizing agent is an oxometal
222
species in the case of metal ions that are strong
oxidizing agents in their highest oxidation state
(e.g., Se4, Cr6, V5, Ce4, Ru8), while
weakly oxidizing metal ions (e.g., Mo6,
Zr4, Ti4) react as peroxometal species.
V5 can react by both mechanisms.

223

The best known Milas reagent is OsO4.

However, it does not oxidize the olefins via
the epoxides but via the osmates (Eq. 224):

224

b. HalconOxirane Process (! Milas

reagentsPropylene Oxide). If olefins are
treated with organic hydroperoxides in
nonaqueous solvents in the presence of
It should be noticed that metal ions reacting Milas reagents as catalysts, epoxides are
via oxometal mechanism are reduced by the formed in high yields (Eq. 225):
substrate and reoxidized by the active oxidizing
agent during one cycle, whereas the metal ions
in the peroxometal complexes do not change
their oxidation state but add the active oxygen
donor to their complex sphere. In this way, 225
primary alcohols can be oxidized (with Zr
complexes) to aldehydes, secondary alcohols Homogeneous catalysis: Mo6, W6, V5,
to ketones, and olefins to epoxides or glycols Ti4 (Arco)
(see Example 3). Heterogeneous catalysis: Ti4/SiO2 (Shell)
 Oxidation 85

The following mechanism is assumed for determination of palladium. Since the cata-
(Eqs. 226228): lysts rapidly became inactive in the gas phase
because of formation of palladium aggregates,
it was decided to carry out the reaction in the
226 liquid phase (water). By combination with a
redox system (CuCl/CuCl2), a catalytic process
for acetaldehyde synthesis was developed. Con-
siderable corrosion problems arose due to the
use of chloride ions in an acidic medium (pH
0.83) at 100130 C.
227 The reaction scheme is as follows
(Eqs. 229231):

C2H4 PdCl2 H2O ! CH3CHO Pd0 2 HCl (229)

Pd0 2 CuCl2 ! PdCl2 Cu2Cl2 (230)

Cu2Cl2 2 HCl 1/2 O2 ! 2 CuCl2 H2O (231)

228

From experiments with C2D4 (very small

isotope effect), D2O (kH/kD 4.5, no deuterium
in the acetaldehyde), with varying H and Cl
Whereas the alcohols formed from the concentrations (both ions inhibit the reaction),
hydroperoxides have a strongly inhibiting with various nucleophiles (acetic acid or etha-
effect on the activity of vanadium, they inhibit nol instead of water as the solvent), and from
the action of molybdenum only in high con- the finding that trans OX substitution occurs in
centrations (e.g., when used as the solvent). ethylene (OX thus comes directly from the
For homogeneous catalysis the metals are solvent and not from the ligand sphere of the
added in the form of complexes that are readily complex) it was concluded that the mechanism
soluble in organic materials in concentrations is as follows:
of 1500 ppm. The reaction temperatures are
between 80 and 120 C. The problem of catalyst 1. Substitution of a chloride ligand by ethyl-
recycling was solved by Shell by rendering the ene, with formation of a p-complex, which
catalyst heterogeneous. is in equilibrium with the corresponding
The major problem of this process is that solvent complex by rapid ligand exchange
ca. 2 t tert-butanol (TBA) are formed per tonne (Eqs. 232 and 233):
of propylene oxide. Up to the 1980s Arco
recycled the TBA by dehydration to isobutene
and subsequent hydrogenation to isobutane.
Today, isobutene is converted to methyl-tert-
butyl ether (MTBE) for which a large market as
gasoline additive has developed (! Methyl- 232
Tert-Butyl Ether).

Example 4: Palladium-Catalyzed Oxidation of

Olefins. Wacker/Hoechst process (! Acetal-
dehyde). SMIDT et al. [386] attempted to oxidize
ethylene to ethylene oxide in the gas phase over 233
Pd/C. In contrast to their expectations acetal-
dehyde was formed. In this way a reaction
was rediscovered, which was known since 2. Formation of the p-complex lowers the
1894 and used analytically for a long time electron density on the ethylene ligand,
86 Oxidation

rendering it susceptible to nucleophilic CuCl2) is not required, unlike with Pd2.

attack. The p-complex is in equilibrium With water as solvent acrylic acid is formed
with the s-complex. The latter loses a Cl (Reaction 241).
ion in a slow reaction and decomposes after Vinyl mechanism
hydride migration from C1 to C2 via the Pd,
and two-electron transfer from C1 to Pd, to
give the final products. The Pd0 is then
reoxidized (Eqs. 234236). Alternatives to
this mechanism are discussed in [387, 388].
238

Allyl mechanism

234

239

240

236

241

Palladium(0) on activated carbon also cata-

lyzes the oxidation of the methyl group of
237
toluene under mild conditions (Reactions
242, 243):

The two major drawbacks of the Wacker

processchlorinated byproducts and severe
corrosioncan be overcome by using hetero-
polyacids (HPA) as the reoxidizing agents [389]
instead of the CuCl/CuCl2 system.
Higher olefins are oxidized to ketones.
Yields of over 95% are only obtained with
HPA as the reoxidizing agent for palladium.
242; 243
As shown above, vinyl acetate is formed in
acetic acid (and in the presence of Na acetate). Further examples of the heterolytic oxida-
Isopropenyl acetate is formed from propene tion of olefins are:
since Pd2 is electrophilic and therefore reacts
with the double bond in the manner shown in
Reactions (232237) and (238). In contrast, Pd0
244
abstracts the allylic H atom and gives a p-allyl
complex. Thus propene is oxidized to allyl
compounds on Pd0 (on activated carbon or
another carrier), whereby a cocatalyst (e.g., TBHP tert-butyl hydroperoxide
 Oxidation 87

Oxidative CC cleavage:
252

245

cat. : Pb2.63Ru1.37O6.5 or Bi2.39Ru1.61O7y

Heteropolyacids (HPA) are gaining increas-
246 ing importance because of their unique combi-
nation of properties:

Alcohols can be oxidized in a similar manner, 1. Strong Brnsted acids

as the following examples (Eqs. 247252) show: 2. Multi-electron oxidizing agents
3. Solubility in water and oxygenated organic
solvents (soluble oxides)
247 4. HPA anions are stable to oxidation and
function as multi-electron ligands that can
stabilize reactive, higher-valent oxometal
species

248 For example, the oxidative cleavage of

diols was achieved with H2O2 in tert-butanol
in the presence of the cetylpyridinium salt of
the heteropolyacid H3PW12O40 as catalyst
(Eq. 253).
249

253

acac acetylacetonate
NAFK Nafion R 511 perfluorinated ion-
cat.: (cetylpyridinium)3 PW12O40
exchange resin
To avoid the problems associated with cat-
alyst recycling, Anic/Enichem has incorpo-
rated Ti4 into a zeolite lattice and obtained
titanium-silicalite (TS-1). TS-1 has the same
elemental composition as the Shell TiSiO2
catalyst, but a considerably wider range of
applications. It is assumed that active, isolated
titanyl species are responsible for the catalytic
activity (Eq. 254):

250
254

TS-1

With TS-1 as catalyst, the following

reactions have, among others, been carried
251 out with 30% H2O2 (Eqs. 254259):
88 Oxidation

6.2. Heterogeneously Catalyzed Gas-

255 Phase Oxidation

The area of heterogeneously catalyzed gas

phase oxidations is still largely empirical,
with regard to the composition and preparation
256
of the catalysts, and the control and influencing
of the processes occurring on the catalyst sur-
face during the reaction.
Considerable progress has been made in the
investigation and characterization of surfaces.
Surface analysis methods, such as Auger elec-
257
tron spectroscopy (AES), secondary ion mass
spectroscopy (SIMS), and low energy electron
diffraction (LEED), have been applied to cata-
lysts. By using such techniques, considerable
advances in the characterization of and the
differentiation between good and poor catalysts
258
have been made. This is important both for
reproducible catalyst production and for the
explanation of poisoning and aging processes.
Most of these spectroscopic methods can,
however, only be performed under high vac-
uum. Thus for all industrially interesting cata-
259 lysts one is still far away from investigating the
processes occurring on the catalyst surface
during the reaction. Nevertheless, a large num-
For hydrogen peroxide and the inorganic ber of observations and indirect methods of
water-soluble oxygen donors, there is often a investigation (e.g., sorption measurements,
miscibility problem with the organic com- use of isotopically labeled compounds, alter-
pound. This problem can be solved in many nating oxidation and reduction) in combination
cases by using a two-phase aqueousorganic with the above-mentioned methods of physical
solvent system in conjunction with a phase analysis allows a rough picture to be built up of
transfer catalyst. what happens on the surface of the catalyst.
As long as the processes on the surface of the
Process Technology of Heterolytic Liquid- catalyst are practically inaccessible to direct
Phase Oxidation. Because of the great number investigation, all proposed mechanisms can
of possible reactions, general rules cannot be only be speculative.
drawn up. In the cases where only liquids are For a summary of the literature up to 1972,
reacted with each other, stirred tanks are suit- see [390]; up to 1979, [343, 391, 392]; (see also
able reactors for small batches. For continuous ! Heterogeneous Catalysis and Solid
production, reactors with turbulent plug flow Catalysts).
are generally the most suitable.
If a catalyst precursor is used a formation 6.2.1. Oxidation Catalysts
phase is necessary for initiation. After this
formation phase or when using preformed cat- Table 22 lists the most important oxidation
alysts, the reaction proceeds immediately with catalysts and the reactions catalyzed by them
its maximum possible rate. [393]. The catalysts are divided into three
The reactions are generally very selective. If groups:
the catalyst permits, they can be run until
complete conversion of the starting material 1. The noble metals Pt and Ag are the longest
is reached. known catalysts. Their large-scale use as
 Oxidation 89

Table 22. Types of catalyst for heterogeneous gas-phase oxidation

Catalyst Reactions and examples

Noble metals(Pt, Pd, Ag)

Pt( Rh,Au) gauze NH3 oxidation to NO/NO2
Andrussow process: CH4 NH3 32 O2 ! HCN 3 H2O
Pt( Rh)/TiO2/carrier off-gas combustion/purification
Pt( Bi,Au)/carrier oxyacetylation, e.g.:
ethylene ! vinyl acetate
propene ! allyl acetate
Ag(Cs)/carrier ethylene oxide direct synthesis
Ag (pure) oxidative alcohol dehydrogenation:
methanol ! formaldehyde
ethanol ! acetaldehyde
Cu - Fe catalysts
CuCl2 (Alkali chloride)/carrier Deacon process: 2 HCl 1/2 O2 ! Cl2 H2O
oxychlorination: C2H4 2 HCl 1/2 O2 ! ClCH2CH2Cl H2O
Cu - Mn-Cr oxide off-gas combustion/purification
Mg - Cr ferrite oxidative dehydrogenation:
butene ! butadiene
CuO - ZnO oxidative decarbonylation:
a-branched aldehydes ! ketones
(e.g., isobutyraldehyde ! acetone)
Cu2O/carrier propene ! acrolein
Catalysts based on V2O5 MoO3
V2O5( K2S2O7)/carrier contact process: SO2 1/2 O2 ! SO3
V2O5( P2O5)/carrier o-xylene ! phthalic anhydride (! phthalic acid)
V2O5( P2O5 TiO2)/carrier butane/butene ! maleic anhydride (! maleic acid)
V2O5( MoO3 additives)/carrier benzene ! maleic anhydride (! maleic acid)
V2O5( SnO2 additives)/carrier ammonoxidation: alkyl aromatics ! aromatic nitriles
MoO3 - Bi2O3 ( various additives) oxidative dehydrogenation: butene ! butadiene
ammonoxidation: propene ! acrylonitrile
propene ! acrolein
MoO3 - V2O5 ( various additives) acrolein ! acrylic acid
MoO3 - Fe2O3 Formox process: methanol ! formaldehyde

Of minor industrial importance.

oxidation catalysts was limited for a long the oxyhydrochlorination of ethylene/HCl

time to the combustion of ammonia (Pt/Rh, mixtures to give dichloroethane (precursor
Ostwald process) and the synthesis of form- for vinyl chloride).
aldehyde (Ag; BASF process). After World 3. After Pt and Ag, the vanadium oxide catalysts
War II, the direct oxidation of ethylene to are the longest used industrial catalysts for
ethylene oxide on Ag, the oxydehydrogena- oxidation processes; V2O5 was first used in
tion of ethanol on Ag to give acetaldehyde, the oxidation of SO2 to SO3, and later in the
and the Andrussow process for the synthesis oxidation of naphthalene to phthalic anhy-
of HCN on Pt catalysts became important. In dride, and benzene to maleic anhydride.
the late 1960s, the combustion of off-gases
on Pt/Rh and the acetoxylation of olefins on The discovery of the Bi2O3MoO3 catalysts
Pd were introduced. by Sohio had a similarly revolutionary effect on
2. In the 1940s copper-based catalysts became the development of mixed-oxide catalysts for
important for the oxidation of propene to gas-phase oxidation, as the discovery of the
acrolein. In the 1950s various mixed-oxide palladium catalysts for ethylene oxidation on
catalysts based on Fe/Cr oxide were devel- the development of the heterolytic liquid-phase
oped for the oxidative dehydrogenation of oxidation.
butenes to give butadiene. In the 1960s the Properties and Mechanism of Action of Het-
CuCl2/carrier catalysts, known since 1868, erogeneous Oxidation Catalysts. Most heter-
achieved renewed industrial importance for ogeneously catalyzed oxidations probably
90 Oxidation

proceed by heterolytic mechanisms. As shown

in Section 5.1 these mechanisms differ, depend-
ing on the catalyst used.

Noble Metal Catalysts. A characteristic of

noble metal catalysts is that reducing condi-
tions must generally be used for selective oxi-
dations; i.e., the hydrocarbon must be in excess
and oxygen in deficiency. With excess oxygen
the pure platinum metals in particular catalyze
only complete combustion, which begins at
200250 C on active catalysts.
The most active catalysts for total combus-
tion (off-gas purification) are based on Pt/Rh;
they catalyze the complete oxidation of hydro-
carbon compounds to CO2 and H2O at a stoi-
chiometric O2 availability. Figure 13. Development of the selectivity in the oxidation
For the acetoxylation of olefins over a Pd0 of ethylene oxide since 1950
catalyst, a mechanism was proposed in The selectivities attained in industrial plants lie in the
Reaction (97). It is assumed that this mecha- shaded region
nism is also valid for the corresponding
reactions in the gas phase. The palladium
must be modified with Bi and Au, and sodium selectivities of over 86% were achieved by
acetate must be continuously added to the additional doping of the Ag/Cs/carrier catalysts
starting material as a moderator, so that the with rhenium and sulfate ions.
reaction proceeds with the desired selectivity. Although the oxidation of ethylene on silver
has been known for 70 years and has been
Oxidation of Ethylene to Ethylene Oxide on investigated intensively, there is no clarity
Silver Catalysts (see also ! Ethylene Oxide, over the mechanism either with regard to the
Chap. 4.1., ! Ethylene Oxide, Chap. 4.2.). Up activation of the oxygen or the ethylene. Two
to now ethylene is the only olefin which can be mechanisms are principally discussed for the
selectively oxidized to the epoxide in the gas oxygen activation (see also ! Ethylene Oxide,
phase, and silver is the only catalyst on which Section 4.2):
the reaction proceeds with high selectivity. According to mechanism 1 oxygen mole-
Ethylene can be oxidized to ethylene oxide cules adsorbed end-on on the silver surface at
with an excess of oxygen on silver catalysts. certain interstitial sites are believed to oxidize
Because the selectivity decreases with increas- the ethylene diffusing to the site from the gas
ing conversion, small ethylene conversions are phase to give ethylene oxide selectively. The
used industrially. Because of the need to recycle oxygen atom bound to the silver remains on the
the ethylene, pure oxygen is generally used. surface.
The oxygen concentration is determined by the Ethylene oxide and carbon dioxide forma-
lower explosion limit with respect to oxygen. tion proceed in parallel in the industrially used
Figure 13 shows the development of the temperature range and also over wide ranges of
ethylene oxide selectivity. The first catalysts ethylene and oxygen concentration, and appear
gave selectivities of ca. 50%, which were grad- to be coupled with one another. Moreover,
ually increased to 6570% by addition of a few several studies showed that chemisorbed oxy-
ppm of a chlorohydrocarbon to the make-up gas gen atoms only totally oxidize ethylene. It was
to lower the activity, and by improvements in therefore assumed that the chemisorbed oxygen
catalyst preparation. In the 1970s Shell raised atom remaining after transfer of one oxygen
the ethylene oxide selectivity from below 70% atom to ethylene, is subsequently removed by
to ca. 75% by doping the catalyst with 100500 the other ethylene molecule with total oxida-
mg/kg cesium [394]. Later, ethylene oxide tion. Subsequently, the cycle can start again.
 Oxidation 91

In this mechanism one ethylene molecule Copper Catalysts. Whereas the oxidation of
can remove a maximum of six oxygen atoms HCl and the oxychlorination of ethane require
from the interstitial sites at which ethylene temperatures over 400 C, the oxychlorination
oxide has previously been formed. Therefore, of ethylene proceeds at ca. 230 C. An electro-
the ethylene oxide selectivity should not exceed philic attack of the ethylene double bond on a
the limit of 100
6/7 85.7%. copper hydroxy chloro complex with formation
In mechanism 2 oxygen atoms adsorbed on of an ethylene-chloronium complex and reduc-
the silver surface bring about both selective and tion of Cu2 to Cu with subsequent trans
total oxidation of ethylene. It is assumed that addition of a second chloride ion to this chloro-
promoter and moderator ions in the neighbor- nium complex is assumed. This is followed by
hood of the active oxygen atom (e.g., Cs, Cl, reoxidation of the copper (Reaction 260 and
and also O2) influence its electronegativity 261).
and thus determine whether it acts as a selective
or total oxidant (lowering the electronegativity
favors selective oxidation).
Given the available experimental evidence,
arguments can be made in support of either
mechanism [395].
If the ethylene oxide selectivity of more than
86% reported in [396] is confirmed, the second
part of mechanism 1 would be refuted. 260
There have been practically no investiga-
tions of the nature of the ethylene activation.
It only seems certain that it does not react
directly from the gas phaseas postulated in
mechanism 1but is first adsorbed on the
catalyst surface.
The reaction is zero, first, or fractional order,
depending on the ethylene/oxygen ratio in the
gas phase. Under industrial conditions (concen-
tration < 9% O2, 3040% ethylene; per pass,
40% of the O2 and only ca. 10% of the ethylene
are converted), the reaction is zero order with
respect to ethylene and first order with respect
to oxygen.
The oxidative dehydrogenation of ethanol to
give acetaldehyde (see ! Acetaldehyde, Chap.
4.1.) on silver or copper catalysts has long been
known and was formerly the method of choice
in the laboratory. Formaldehyde was also pre-
viously produced by this method, but is now
mainly produced in the Formox process by
oxidative dehydrogenation of methanol on
iron molybdate catalysts (! Formaldehyde,
Chap. 4.2.).
In the Huls process (originally developed by
BASF), ethanol is mixed with air and passed
over coarse-grained crystals of pure silver at
500650 C. There is only 5070% conversion
of ethanol, but complete consumption of oxy-
gen. The off-gas contains up to 11% unoxidized
261
hydrogen.
92 Oxidation

Mixed Oxide Catalysts. There are essentially [total selectivity (acrolein acrylic acid):
two groups of mixed oxide catalysts: 96%].
The selective further oxidation of acrolein
1. Molybdenum-based catalysts that contain can be achieved on MoV catalysts (see !
bismuth or vanadium and/or tungsten as Acrylic Acid and Derivatives). Obviously the
the second main component and which are addition of vanadium leads to stronger adsorp-
used, for instance for the oxidation of pro- tion and activation of the acrolein than on the
pene to acrolein and acrylic acid or to acry- MoBi catalyst.
lonitrile, and for oxidative dehydrogenation The oxidative dehydrogenation of butene to
of butene to give butadiene. give butadiene proceeds similarly. On MoBi
2. Catalysts based onvanadium, with, e.g., phos- catalysts the first step of the oxidation proceeds
phate, sulfate, TiO2, or MoO3 as the second as with propene. Like acrolein, the butadiene
main component, are principally used for the formed is practically completely desorbed from
oxidation of aromatics and butane/butene to the catalyst (selectivity > 90%) and is not
give acid anhydrides (benzene or butane/ further oxidized. Two mechanisms can be dis-
butene to maleic anhydride, naphthalene or cussed (Eqs. 263 and 264):
o-xylene to phthalic anhydride).

While V2O5 and MoO3 belong to the oxides

which are themselves catalytically active,
oxides such as Bi2O3, Sb2O3, SnO2, and TiO2
possess virtually no catalytic activity of their
own. However, in combination with V2O5 and
MoO3 the catalytic activity and selectivity of
the latter are greatly increased.
263
Hydrocarbon Activation on Mixed-Oxide Cat-
alysts. Experiments with 14C-labeled propene
indicate that it is chemisorbed as a p-allyl
complex on the Mo/Bi catalysts with hydrogen
abstraction (Eq. 262). It is assumed that this
p-complex reacts further with the metal oxo 264
complex in a manner analogous to the reaction
with SeO2 with oxygen transfer and metal
reduction (see Reactions 219 to 221).
On vanadium-containing catalysts (see
Table 22), however, the butadiene is rapidly
oxidized further to maleic anhydride.
At the high temperatures necessary for the
oxidation of butane (400450 C), vanadyl
phosphate catalysts, (VO)2P2O7, have proven
particularly selective. The P : V ratio in the bulk
catalyst particle is 1 : 1, corresponding to
the stoichiometry, but on the catalyst surface
262 it is 2 : 1. The phosphate apparently has two
functions: it stabilizes the vanadium in high
valence states by complexation and it inhibits
On a catalyst with the composition Mo10Bi- homogeneous gas-phase oxidation of butane
FeCo4W2K0.06Si1.35 a selectivity for acrolein and its intermediates.
of 90% was achieved at a propene conversion Methyl aromatics (toluene, xylene) and
of 98%. Only ca. 6% of the acrolein was naphthalene are oxidized with the greatest
further oxidized to acrylic acid on this catalyst selectivity on vanadium catalysts via the
 Oxidation 93

aldehydes to the acids or acid anhydrides. High increased to 9095%. It is assumed that the
yields are obtained only with those hydrocar- very high activation energy of the oxygen dis-
bons that form moderately polar, readily sociation necessary for the reoxidation of molyb-
desorbable, volatile products, such as o-xylene, denum is strongly lowered by these additives.
from which phthalic anhydride is formed. p- Today the reduction/oxidation-mechanism
Xylene initially affords terephthalic acid, which developed by MARS and VAN KREVELEN in the
is too polar and is therefore hardly desorbed. 1960sis generally accepted as the only valid
The anhydride of terephthalic acid is oligo- one. However, taking the later developed
meric and therefore insufficiently volatile. oxometalperoxometal scheme (see Section
For this reason p-xylene can be oxidized selec- 5.1) into account, with oxidation catalysts con-
tively in the gas phase only by ammoxidation to taining Mo and even V, a peroxometal mecha-
give terephthalonitrile. Normal air oxidation, in nism without reduction and reoxidation of the
contrast, results predominantly in total central metal ion should be considered too.
oxidation. The oxidation of hydrocarbons on mixed
Mixed oxide catalysts are only catalytically oxide catalysts generally follows Langmuir
active and selective in their higher oxidation Hinshelwood kinetics, i.e., the reaction rate
states. Unlike the noble metal catalysts they depends on both the concentration of the hydro-
must be used in an excess of oxygen, i.e., in an carbon and the concentration of oxygen in the
oxidizing atmosphere. gas phase.
It is generally assumed that the oxygen Because of the large number of parallel and
originates from the oxide lattice in selective consecutive processes occurring simulta-
oxidations on mixed oxide catalysts and that the neously on the catalyst surface (adsorption,
metal ions are reduced and oxidized again chemical reactions, and desorption) kinetic
alternately (redox mechanism of MARS and measurements only give apparent values for
VAN KREVELEN). Up to now there are only vague heterogeneously catalyzed oxidations. CREMER
ideas about the way in which atmospheric was the first to point out the compensation
oxygen is activated and incorporated into the effect occurring in these types of kinetic mea-
lattice. surements [397]. In the Arrhenius equation
Whereas reoxidation proceeds very rapidly (265)
on vanadium catalysts (the reaction temperature
is mainly determined by the temperature nec- k Ao  eEa =RT 265
essary to activate the hydrocarbon), the reox-
idation of the catalyst is the activity- the collision factor A decreases with decreasing
determining process for molybdenum-based activation energy and vice versa.
catalysts, which also determines the reaction
temperature. This is demonstrated, for example, 6.2.2. Process Technology of Heteroge-
by the fact that for catalysts based on V2O5 the neously Catalyzed Oxidations
active mass can be applied as a thin layer on a
carrier (egg-shell catalyst), while catalysts Here only the oxidation processes on carrier or
based on MoO3 only exhibit sufficient activity mixed oxide full-contact catalysts are
as bulk catalysts. described, whereas processes involving gauzes
The importance of the reoxidation rate in or thin layers of noble metals and waste gas
mixed oxide catalysts, is exemplified by the combustion are not considered.
development of the BiMo catalysts. When
they were introduced by Sohio in the early Catalyst Preparation and Characterization
1960s, the temperature required for oxidation [349] (see also ! Heterogeneous Catalysis
of propene was 400450 C, and the selectivity and Solid Catalysts). The surface area of oxi-
for acrolein, between 60 and 70%. By the dation catalysts should be relatively small
addition of metal ions such as FeII/III, CuII, (0.110 m2/g), i.e., the catalyst should have
CoII and NiII, which only catalyze one-electron no pores smaller than 50 nm, since these favor
transfers, the reaction temperature could be total combustion due to the poor material
lowered to 300330 C and the selectivity transport.
94 Oxidation

Reactors. Because of the large quantities of with a sharp temperature peak (hot spot), whose
heat liberated in oxidation reactions and the height is generally proportional to the hydro-
poor heat transport through the gas, either carbon loading (large reserve of catalytically
multitube reactors (with up to 40 000 tubes active centers in the region of the main reaction
of 1040 mm diameter and 112 m length) or zone). Because of the large temperature differ-
fluidized-bed reactors are used. ence between the cooling medium and the
In multitube reactors the heat is removed catalyst in the region of the temperature maxi-
either by boiling liquids (water or organic heat mum (high temperature in the main reaction
transfer media) or by salt melts (e.g., 53 wt% zone and relatively low temperature in the post-
KNO3, 40 wt% NaNO2, 7 wt% NaNO3, mp reaction zone), the hydrocarbon is generally not
150 C, usable up to 550 C; reactor material completely converted, but often only to the
preferably untempered mild steel). In fluidized- extent of 80 or 90%. The reaction is first order
bed reactors the heat is removed by cooling with respect to the hydrocarbon.
tubes containing water (high-pressure steam With a moderately active catalyst (Fig. 14 B),
production). there is only a small temperature rise with
Because of the unavoidable back-mixing in increasing hydrocarbon load, and the tempera-
fluidized-bed reactors the selectivity is always ture maximum is broadened (most active centers
lower than in multitube reactors with their plug of the catalyst in the region of the temperature
flow. For this reason fluidized-bed reactors are maximum participate in the reaction). The
only used for catalysts with particularly short reaction is first order with respect to the hydro-
lifetimes (< 1 year). In multitube reactors, cat- carbon provided that sufficient catalytically
alyst exchange always means a shutdown of active sites are free in the contact bed region
several weeks. In particular, the pressure loss below the reaction zone, which is indicated by
compensation of the freshly filled tubes is still a the decrease of the reaction temperature below
tedious manual process. In fluidized-bed the temperature maximum to the level of the
reactors the catalyst can be continuously cooling medium. The hydrocarbon conversion is
exchanged, even during operation. Typical more complete the smaller the difference
examples for the use of fluidized beds are in between the temperature maximum and the cool-
the older acrylonitrile plants, and the oxidation ing medium (mode of operation typical of an
of butane to maleic anhydride. integral reactor).
With a low-activity catalyst (Fig. 14 C) the
Temperature Profiles. In fluidized-bed reaction is zero order with respect to the hydro-
reactors, heat transport by the turbulent mixed carbon because all active centers are occupied.
solid catalyst is very good. Therefore, uniform Because of the hydrocarbon excess which is
temperature is established throughout the entire usually necessary, the reaction is run according
bed. to the permitted oxygen conversion (mode of
In fixed-bed reactors, heat transport by the operation typical of a differential reactor with
gas is poor and characteristic temperature pro- hydrocarbon recirculation).
files develop along the axis of each tube. Regular measurements of the temperature
Because the temperature differences between profile reveal changes in catalyst behavior (e.g.,
catalyst bed and bath in each spot along the due to aging or poisoning). In addition, the true
tubes are proportional to the heat evolved in the catalyst performance (kilogram product per
catalyst bed at this spot, these temperature kilogram catalyst per second) in sections of
profiles give a fairly accurate indication of the bed can be evaluated from the temperature
the overall reaction taking place at each spot profile.
along the catalyst bed. If capillaries (thermo-
wells) are mounted in the axis of some tubes, Addition of Steam. Water is adsorbed more
these temperature profiles can be measured by strongly on the catalyst surface than most
moveable thermocouples. Figure 14 shows organic products because of its high polarity.
three typical temperature profiles. By addition of steam to the make-up gas, polar
With highly active catalysts (Fig. 14 A) the substances can thus be desorbed from the cata-
main reaction occurs in a very limited region lyst surface by competing adsorption and their
 Oxidation 95

Figure 14. Temperature profiles of various catalysts

A) Temperature profile of a highly active catalyst (V2O5 type) as a function of hydrocarbon loading (DT1, low loading; DT3,
high loading).
B) Temperature profile of a moderately active catalyst (BiO3MoO3 type) as a function of hydrocarbon loading (DT1, low
loading; DT3, high loading).
C) Temperature profile of a low-activity catalyst (Pd0/Al2O3 moderator). The activity can only be increased by raising the
temperature.

yields increased. This method is used above all dehydration equilibrium lies completely on
in the oxidation of propene to acrolein and the side of the anhydride at the reaction
acrylic acid and in oxidative dehydrogenation. temperatures.
For acid anhydrides the addition of steam Addition of steam also improves heat
clearly leads to a lower yield, although the removal. Because the diffusion rate in gases
96 Oxidation

increases with decreasing molecular mass, the narrower laboratory fluidized beds are gener-
heat conducting capacity of a gas mixture ally considerably higher than on the industrial
increases with the number of small molecules. scale.
Like methane and ammonia, water is one of the
smallest stable gas molecules. Influence of Homogeneous Gas-Phase Oxida-
tions. As a rule, homogeneous gas-phase oxi-
Laboratory and Pilot Plant Experiments; dations generally play only a minor role, if any,
Applicability of the Results. The choice of in heterogeneously catalyzed gas-phase oxida-
reactor is orientated according to the planned tion, because for safety reasons the temperature
investigations: of the gas mixture above the catalyst bed must
For kinetic measurements gradientless be kept well below the (sometimes very low)
reactors must be used; details of construction ignition temperature. Owing to the short resi-
are given in [349, 398, 399]. dence times, a significant oxidation reaction
For the development of catalysts, reactors cannot normally occur in this region. In the
similar to those used on industrial scale should contact bed no significant oxidation reaction
be used. should be able to take place in the homoge-
In tube reactors catalyst testing is straight- neous gas phase because of the very small free
forward. A suitable apparatus is described in gas space. Below the catalyst bed, the gas
[400]. Smaller reactors (e.g., made of glass with mixture is normally cooled so rapidly that
an inner diameter of 20 mm and a bed height of hardly any oxidation in the homogeneous gas
100200 mm) gave identical results to pilot phase can occur.
plant reactors with bed heights of 13 m,
provided that the reactor diameter, catalyst Safety. Although under normal operating con-
diameter and height, and the residence time ditions homogeneous gas-phase oxidation does
based on the free reactor volume were in agree- not play an important role, critical states can
ment with the pilot plant-reactor; that the resi- quickly arise during start-up and shutdown and
dence time was not greater than 5 s (at longer in the case of faults. Particular attention must be
residence times it is not certain whether the paid to the make-up gas compressors, which
flow is still turbulent); and that the catalyst was drive the make-up gas through the reactor. If the
not too active. compressor stops, the flow rate of the gas about
With very active catalysts, whose tempera- to enter the reactor decreases so much that the
ture profiles are similar to Figure 14 A, scaling gas mixture in the voluminous head of the
up is often associated with loss in selectivity reactor, can be ignited by the hot catalyst.
due to the higher temperature in the hot spot or Equally critical states can occur if the catalyst
a lower conversion as a result of the lower bath falls out of one reactor tube during the reaction
temperature to compensate for the increasing or if the flow resistance in the reaction tubes
heat production with increasing starting mate- becomes very different. Particularly in pro-
rial load. cesses with high oxygen conversion, ignitable
For catalysts with temperature profiles cor- gas mixtures can enter the space downstream
responding to those in Figures 14 B and 14 C, from the catalyst bed and can ignite either
laboratory and pilot plant results are largely spontaneously or through the action of catalyst
consistent, i.e., they are also valid for the dust. Most plants should have been equipped in
industrial multitube reactor, as the pilot plant the meantime with safety devices which render
reactor behaves like a single tube from the the plant inert in critical areasindependent of
multitube reactor. service personnelor bring it to a safe state in
Fluidized beds behave similarly to liquid another way.
bubble columns; i.e., results that are valid for
a large industrial scale are only obtained in beds
with a diameter exceeding 300 mm. The expen- References
diture associated with this requirement in the 1 D. Lenoir, Angew. Chem. Int. Ed. 45 (2006) 32063210.
pilot plant is considerable. Due to lower back 2 K. Weissermel, H.-J. Arpe: Industrial Organic Chemistry, 3rd
mixing the selectivity data obtained in the ed., VCH Verlagsgesellschaft, Weinheim, Germany, 1997.
 Oxidation 97

3 M. Sittig: Combining Oxygen and Hydrocarbons for Profit, 22 TPC Group LLC, WO 2013/148908, 2013 (L. Caciula, J.G.
Gulf Publishing Company, Houston 1962. Duff, E. Ballard, M. Potter, S. Chada).
4 US Geological Survey, National Water-Quality Assessment 23 TPC Group LLC, WO 2013/148913, 2013 (L. Caciula, J.G.
(NAWQA) Program, Pesticide National Synthesis Project, Duff, S. Chada, E. Ballard).
data available online under: http://water.usgs.gov/nawqa/ 24 BASF, Linde eye on-purpose butadiene technology, Hydro-
pnsp/usage/maps/show_map.php?year2011&map carbon Processing, 06.03.2014, http://www.hydrocarbon
PROPARGITE&hiloH&dispPropargite (accessed 19th processing.com/Article/3347976/Search/BASF-Linde-eye-on-
May 2015). purpose-butadiene-technology.html?
5 S. Caron, R.W. Dugger, S.G. Ruggeri, J.A. Ragan, D.H.B. KeywordsBASFLindeon-purposebutadiene (accessed
Ripin, Chem. Rev. 106 (2006) 29432989. 19th May 2015).
6 S.G. van Ornum, R.M. Champeau, R. Pariza, Chem. Rev. 106 25 J.C. Bricker, Topics Catal. 55 (2012) 13091314.
(2006) 29903001. 26 C. Hammond, S. Conrad, I. Hermans, ChemSusChem 9 (2012)
7 P. Pfeifer, T. Zscherpe, K. Haas-Santo, R. Dittmeyer, Appl. 16681686.
Catal. A 391 (2011) 289296. 27 T. Cleveland, H. Soo, Hydrocarbon Eng. Oct (2001)
8 X. Zhang, Y. Tang, S. Qu, J. Da, Z. Hao, ACS Catal. 5 (2015) 6971.
10531067. 28 Y. Liu, J. Liu, F. Lin, Huagong Keji Shichang (Chem. Technol.
9 K. Goeransson, U. Soederlind, J. He, W. Zhang, Renewable Market) 31 (2008) 1921, 24.
Sustainable Energy Rev. 15 (2011) 482492. 29 V.I. Bukhtiyarov, A. Knop-Gericke in C. Hess, R. Schlogl
10 B. Kamm, Angew. Chem., Int. Ed. 46 (2007) 50565058. (eds.): Nanostructured Catalysts: Selective Oxidations, RSC
11 J. Perez-Ramrez, C. Mondelli, T. Schmidt, O.F.-K. Schluter, Nanoscience & Nanotechnology, No. 19, Royal Society of
A. Wolf, L. Mleczko, T. Dreier, Energy Environ. Sci. 4 (2011) Chemistry, 2011, pp. 214247.
47864799. 30 Badische Anilin und Soda Fabrik, DE2108922, 1972 (R. Platz,
12 Kel-Chlor- A Case History, C.P. van Dijk, H. Jongenburger, N. Rieber, R. Schanz, A. Scholz).
IPCOM000221598D, IP.com electronic publication, Sept. 14, 31 Badische Anilin und Soda Fabrik, DE2225657, 1973 (W.
2012. Fuchs, R. Platz, N. Rieber, A. Scholz).
13 207th Meeting of The Electrochemical Society, Abstract 32 ARCO Chem. Technology, US5780657, 1998 (B. Cooker, A.
#1259, May 1618, 2005, Quebec City, http://ma.ecsdl.org/ M. Gaffney, J.D. Jewson, W.H. Onimus).
content/MA2005-01/31/1259.full.pdfhtml (accessed 19th 33 J. Zhang, Y. Xie, Z. Niu, Huagong Keji (Sci. Technol. Chem.
May 2015). Ind.) 18 (2010) 7579.
14 Hydrochloric Acid ElectrolysisSustainable Chlorine 34 J.K.F. Buijink, J.-P. Lange, A.N.R. Bos, A.D. Horton, F.G.M.
Production, http://www.thyssenkrupp-industrial-solutions. Niele in S.T. Oyama (ed.): Mechanisms in Homogeneous and
com/fileadmin/documents/brochures/ Heterogeneous Epoxidation Catalysis, Elsevier, Amsterdam
Hydrochloric_Acid_Electrolysis.pdf (accessed 19th May 2008, pp. 355371.
2015). 35 Snamprogetti, US4410501, 1983 (M. Taramasso, G. Perego, B.
15 UHDENORA receives first order for the new energy-efficient Notari). The italian priority application dates back to Decem-
electrolysis technology, press release, UHDENORA, 20 Feb- ber 1979.
ruary 2014, http://www.uhdenora.com/details.asp?id_news 36 ANIC, EP0100119, 1984 (C. Neri, B. Anfossi, A. Esposito, F.
83 (accessed 19th May 2015). Buonomo).
16 Bayer MaterialScience constructs world-scale plant for chlo- 37 M.G. Clerici: Titanium Silicalite-1, in S.D. Jackson, J.S.J.
rine recycle - Inovative hydrochloric acid electrolysis in Hargreaves (eds.): Metal Oxide Catalysis, Wiley-VCH, Wein-
Shanghai, press release, Bayer, September 20, 2006, http:// heim 2009.
www.investor.bayer.com/en/nc/news/archive/investor-news- 38 J. Yu, Z. Li, Q. Liu, Huaxue Tuijinji Yu Gaofenzi Cailiao
2006/investor-news-2006/innovative-hydrochloric-acid- (Chem. Propellants Polym. Mater.) 9 (2011) 823.
electrolysis-in-shanghai/ (accessed 19th May 2015). 39 J. Yu, Huaxue Tuijinji Yu Gaofenzi Cailiao (Chem. Propellants
17 K. Iwanaga, K. Seki, T. Hibi, K. Issoh, T. Suzuta, M. Nakada, Polym. Mater.) 7 (2009) 1522, 30.
Y. Mori, T. Abe, Sumitomo Kagaku (Osaka, Japan) (2004), no. 40 V. Russo, R. Tesser, E. Santacesaria, M. DiSerio, Ind. Eng.
1, 412; a translation is available online at: http://www Chem. Res. 52 (2013) 11681178.
.sumitomo-chem.co.jp/english/rd/report/theses/docs/ 41 M. Ghanta, D.R. Fahey, D.H. Busch, B. Subramanian, ACS
20040100_p0e.pdf (accessed 19th May 2015). Sust. Chem. Eng. 1 (2013) 268277.
18 Bayer employing new process at chlorine recycling plant, 42 Sumitomo Chem. Co., WO2012/102918, 2012, (A. Seusbal, S.
Process Engineering, 17 November 2008, http:// Senkan, Y. Ohishi, C.G. Knapp-Bjeren).
processengineering.theengineer.co.uk/bayer-employing-new- 43 H. Orzesek, R.P Schulz, U. Dingerdissen, W.F. Maier, Chem.
process-at-chlorine-recycling-plant/309019.article (accessed Eng. Technol. 22 (1999) 691700.
19th May 2015). 44 B.T. Pennington, M.S. Stark, D.J. Waddington, Combust. Sci.
19 F. Cavani, N. Ballarini, A. Cericola, Catal. Today 127 (2007) Tech. 106 (1995) 297305.
113131. 45 Halcon International, US3907874, 1975 (R.J. Harvey, J. Kollar,
20 A. Brice, A. Greenwood: On-purpose BD could offset J.P. Schmidt)
supply constraints, ICIS Chemical Business 2012 April 46 BASF, US4122286, 1978 (J. Hartig, H.-M. Weitz).
622, 16. 47 Chem Systems Inc., US4045477, 1977 (M.B. Sherwin, I-D.
21 UOP to license TPC Group process technology for on-purpose Huang).
butadiene, Hydrocarbon Processing, 06.12.2014, http://www 48 BASF, US4219666, 1980 (R. Platz, H.-M. Weitz, J. Hartig).
.hydrocarbonprocessing.com/Article/3351350/UOP-to- 49 BASF, US4226780, 1980 (G. Fouquet, F. Merger, K. Baer).
license-TPC-Group-process-technology-for-on-purpose- 50 Bayer AG and Deutsche Gold- und Silber-Scheideanstalt,
butadiene.html (accessed 19th May 2015). DE2519297, 1976 (G. Prescher, G. Schreyer, O. Weiberg, R.
98 Oxidation

Wirthwein, H. Waldmann, H. Seifert, W. Schwerdtel, W. 86 Dow to shut down PXC operations in US, ICIS News, 07
Swodenk). September 2006, http://www.icis.com/Articles/2006/09/07/
51 Bayer AG and Deutsche Gold- und Silber-Scheideanstalt, 1089707/dow-to-shut-down-pxc-operations-in-us.html
DE2519298, 1976 (G. Prescher, O. Weiberg, R. Wirthwein, (accessed 19th May 2015).
H. Waldmann, H. Seifert, W. Schwerdtel, W. Swodenk). 87 Halcon Int. Inc., US3796761, 1974 (R.L. Marcell, T.Y. Chang).
52 Interox Chemicals, DE2602776, 1976 (A.M. Hildon, P.F. 88 Bayer Antwerpen N.V., DE2123181, 1972 (P. Bamelis, J. van
Greenhalgh). Esbroeck, P. Lejeune, K.H. Meyer, R. Thiel).
53 O.Y. Guzechak, B.Y. Stul, V.N. Sarancha, V.I. Emelyanov, I. 89 BASF SE, WO2012/038848, 2012 (O. Abillard, D. Breuninger,
M. Romanyuk, O.V. Filatova, B.B. Chesnokov, Neftekhimiya E. Kretzschmar, R. Pinkos).
26 (1986) 272276. 90 BASF SE, US8383866, 2013 (R. Pinkos, A. Paul, R.-H.
54 ABB Lummus Global, US7091365, 2006 (Z. Shan, C.Y. Yeh, P. Fischer, R. Dostalek).
J. Angevine, F.M. Dautzenberg, J.C. Jansen). 91 BASF press release P100/13, 2013, https://www.basf.com/en/
55 BASF, EP0568872, 1993 (E. Gehrer, W. Harder, H. Vogel, B. company/news-and-media/news-releases/2013/01/p-13-100.
Knuth, K. Ebel, C. Groening). html (accessed 19th May 2015).
56 BASF, EP0544157, 1992 (E. Gehrer, W. Harder). 92 F. Broich, H. Grasemann, Erd ol und Kohle 18 (1965) 360364.
57 Tosoh, JP 04-352771, 1992 (A. Sato, T. Miyake). 93 V.V. Kamzolkin, A.N. Bashkirov, K.M. Sokova, T.I. Andrejeva,
58 J. Huang, M. Haruta, Res. Chem. Intermed. 38 (2012) 124. G.A. Zelenaja, Neftekhimija 4 (1964) 599602.
59 Solvay Commercializes New Route to Epichlorohydrin, 94 E. I. DuPont de Nemours and Co., WO2003/027052, 2003 (S.
Chem. Eng. News (2012) March 5, 15. Lane, E. Moran, J.J. Ostermaier).
60 Other Intermediates Ind. Eng. Chem. 46 (1954) 62. 95 Y. Yuan, Y. Zhao, J. Lang, Huaxue Gongchengshi (Chemical
61 Propiconazole, Merck Index, 14th ed., 2006, 07818. Engineer) 25 (2011) 5557.
62 Y. Luo, B. Xu, G. Tang, K. Chen, Nongyao (Agrochemicals) 49 96 Q. Sun, B. Shen, K. Fan, J. Deng, Chem. Phys. Lett. 322 (2000)
(2010) 161163. 18.
63 Propargite, Merck Index, 14th ed., 2006, 07808. 97 L. Lefferts, J.G. van Ommen, J.R.H. Ross, Appl. Catal. 31
64 Y. Jia, Y. Xu, K. Yu, J. Gao, Huagong Shikan (Chem. Ind. (1987) 291308.
Times) 16 (2002) 1115. 98 F. Li, C. Zhu, Huafei Gongye (Chemical Fertilizer Industry) 32
65 Uniroyal, US3733313, 1973 (A.N. Johnson, W. Nudenberg). (2005) 3941, 43.
66 W. Weigert, A. Kleemann, G. Schreier, Chem.-Ztg. 99 (1975) 99 N. Pernicone in M.F. Portela, F.G. Freire, M.P. Castillo (eds.):
1925. Actas do Encontro Nacional de Cat alise Basica e Aplicada
67 Degussa, DE2123930, 1972 (A. Kleemann, G. Schreyer, O. (Industrial e Ambiental), 3rd ed., Lisbon, May, 1995, pp.
Weiberg). 195206.
68 Degussa, DE2408948, 1975 (W. Heim, A. Kleemann, H. Kolb, 100 H. Yilmaz, F.S. Atalay, S. Atalay, NATO Science Series, II:
G. Schreyer). Mathematics, Physics and Chemistry 191 (2005) 345349.
69 Degussa, DE2454630, 1976 (H. Friedrich, W. Heim, A. Klee- 101 T.A. Hanna, Coord. Chem Rev. 248 (2004) 429440.
mann, H. Kolb, G. Schreyer). 102 R.K. Grasselli: Surface Properties and Catalysis by Non-
70 A. Kleemann, R.S. Nygren, R.M. Wagner, Chem.-Ztg. 104 Metals, J.P. Bonnelle et al. (eds.), NATO ASI Series,
(1980) 283286. Series C: Mathematical and Physical Sciences 105 (1983)
71 Berufsgenossenschaft der Chemischen Industrie, Jahresbericht 289304.
82, 1982, pp. 6768. 103 G. Fu, X. Yi, C. Huang, X. Xu, W. Weng, W. Xia, H.-L. Wan,
72 Shell Develop. Corp., US2833788, 1958 (J.R. Skinner, C.H. Surf. Rev. Lett. 14 (2007) 645656.
Wilcoxen, G.C. Carlson). 104 X. Zhang, H. Wan, W. Wenig, Appl. Catal A 353 (2009) 2431.
73 Du Pont de Nemours and Co., US3536733, 1970 (D.P. 105 L. Wang, Z. Li, S. Zhang, X. Zhang, Guocheng Gongcheng
Carlson). Xuebao (Chin. J. Process Eng.) 7 (2007) 202208.
74 G.G. Furin, Chem. Sust. Dev. 14 (2006) 97110. 106 S.A. Schunk, N. Brem in C. Hess, R. Schlogl (eds.):
75 D. Lokhat, D. Ramjugernath, M. Starzak, Ind. Eng. Chem. Res. Nanostructured Catalysts, Selective Oxidations, RSC Nano-
51 (2012) 1396113972. science & Nanotechnology 19 (2011) 169189.
76 Thailands Siam Cement, Dow to start up HPPO plant next 107 BASF, WO2008/037693, 2008 (G. Wegner, G. Kaibel, J.
week, ICIS News, 07 Sept. 2011. Therre, W. Aquila, H. Fuchs).
77 Huntsman Specialty Chemicals Corp., WO 99/47499, 1999 108 BASF, DE2715209, 1978 (B. Meissner, W. Fliege, O. Woerz,
(W.K. Culbreth, M.E. Taylor, K.L. Preston, M.E. Mueller). H. Diem).
78 E. Duynstee, J. Hennekens, Rec. Trav. Chim. Pays-Bas 89 109 BASF, EP0055354, 1982 (W. Sauer, W. Aquila, W. Hoffmann,
(1970) 769780. K. Brenner, K. Halbritter).
79 I. Hermans, T.L. Nguyen, P.A. Jacobs, J. Peeters, ChemPhy- 110 BASF: BASF Wissenschaft popular: Zitrus; Rose und Laven-
sChem 6 (2005) 637645. delnaturidentische Dufte aus dem Chemielabor, http://www
80 I. Hermans, P.A. Jacobs, J. Peeters, Chem. Eur. J. 12 (2006) .basf.com/group/corporate/de_DE/news-and-media-relations/
42294240. science-around-us/citral/story (accessed 19th May 2015).
81 E. Spier, U. Neuenschwander, I. Hermans, Angew. Chem. Int. 111 L. Nian, Guangdong Huagong (Guandong Chemical Industry)
Ed. Engl. 52 (2013) 15811585. 39 (2012) 9596.
82 Halcon Int. Inc., US4066706, 1978 (J.P. Schmidt). 112 J.A.A. Hoorn, P.L. Alsters, G.F. Versteeg, Int. J. Chem. Reactor
83 Sumitomo Chemical Co., US5166451, 1992 (A. Takeshita, S. Engineering 3 (2005) Article A6.
Masaki, O. Maruyama). 113 BASF, WO2012/034930, 2012 (U. Griesbach, J. Botzem, F.
84 W. Rojahn, H.U. Warnecke, Dragoco Report 27 (1980) 159. Stecker).
85 P.L. Alsters, W. Jary, V. Nardello-Rataj, J.-M. Aubry, Org. 114 BASF, DE3322399, 1985 (D. Degner, H. Hannebaum, H.
Proc. Res. Dev. 14 (2010) 259262. Siegel, W. Gramlich).
 Oxidation 99

115 S. Wang, Q. Gu, X. Chen, T. Zhao, N. Zhao, Y. Zhang, Huaxue 145 Societe des Usines Chimiques de Rhone-Poulenc, GB712264,
Shiji (Chemistry and Adhesion) 32 (2010) 835839. 1954.
116 Nanjing University, China, CN103145527, 2013 (Y. Zhao, T. 146 ExxonMobil, WO2013/109348, 2013 (B.A. Patel, J.M.
Ma, L. Wang). Dakka).
117 D. Hoormann, J. J orissen, H. Putter, Chem. Ing. Technik 77 147 Y. Jiang, Y. Dong, B. Wang F. Zhang, Huagong Shengchan Yu
(2005) 13631376. Jishu (Chemical Production and Technology) 18 (2001) 5054.
118 C. Hamon, P. Janssens in NOXCONF, Conf erence Internatio- 148 Rhodia, WO2012/156381, 2012 (L. Garel, S. Normand, P.
nale La Pollution Atmosph erique dOrigine Industrielle, Mait- Pitiot, J.-C. Bigouraux).
rise des Rejets dOxydes dAzote NOx-N2O : Panorama des 149 Brichimica, US3914323, 1975 (P. Maggioni).
Techniques Disponibles, Paris, 21 et 22 mars 2001, pp. 18. 150 F. Maspero, Chimica e LIndustria (Milano) 75 (1993)
119 Ministere de lEcologie et du Developement Durable, Projet de 292296.
Plan dAffectation des Quotas dEmission de Gaz a Effet de 151 B. Notari, Adv. Catal. 41 (1996) 253334.
Serre (PNAQ II) (Periode: 2008 a 2012), 6.2, page 24, http:// 152 Camlin Fine Chemicals buys Italys Borregaard, Business
ec.europa.eu/clima/policies/ets/pre2013/nap/docs/nap_france Standard, March 9, 2011, http://www.business-standard.com/
2008_2012_en.pdf (accessed 19th May 2015). article/companies/camlin-fine-chemicals-buys-italy-s-
120 X. Li, Z. Zhang, D. Zhao, Hebei Keji Daxue Xuebao (J. Hebei borregaard-111030900240_1.html (accessed 19th May 2015).
Univ. Sci. Technol.) 27 (2006) 4548. 153 T. Atoguchi, T. Kanougi, T. Yamamoto, S. Yao, J. Mol. Catal.
121 Mitsui Toatsu Chemicals Inc., EP0597454, 1994 (K. Sudoh, K. A: Chem. 220 (2004) 183187.
Wakimura, M. Tanaka, T. Miura, H. Inoue, M. Yamazaki). 154 T. Atoguchi, T. Kanougi, J. Mol. Catal. A: Chem. 222 (2004)
122 Mitsui Toatsu Chemicals Inc., DE19519426, 1995 (K. Waki- 253257.
mura, K. Sudoh, M. Tanaka, H. Inoue, N. Iwane). 155 Ube Industries, DE2514742, 1975 (S. Umemura, N. Taka-
123 BASF, DE102008018480, 2008 (B. Rumpf, J.R. von Watzdorf, mitsu, T. Hamamoto, N. Kuroda).
T. Johann, G. Koppenh ofer, T. Krug, M. Schreiber). 156 Badische Anilin & Soda Fabrik AG, DE1793183, 1972 (L.
124 Y. Li, M. Li, Jingxi Shiyou Huagong Jinzhan (Advances in Fine Schuster, H. Pommer).
Petrochemicals) 12 (2011) 4652. 157 Mitsubishi Gas Chem. Co., EP0127888, 1984 (T. Isshiki, T.
125 S.M.V. Pacheco, A.F. Junior, A.F. Morgado, Tch^ e Qumica 4 Yui, H. Uno, M. Abe).
(2007) 4554. 158 K. Moller, G. Wienhofer, F. Westerhaus, K. Junge, M. Beller,
126 J. D. Riedel - E. deHaen AG, DE617763, 1935 (F. Boedecker, Catal. Today 173 (2011) 6875.
H. Volk). 159 Memorandum from Irelands Environmental Protection
127 Rhodia, WO2008/148760, 2008 (C. Malivernay, J.-C. Bigour- Agency from March 26th, 1999, http://www.epa.ie/licences/
aux, L. Garel). lic_eDMS/090151b280042130.pdf (accessed 19th May 2015).
128 Rh^ one-Poulenc Industries, US4026950, 1977 (J. LeLudec). 160 A.S. Hay, Polym. Eng. Sci. 16 (1976) 110.
129 Rh^ one-Poulenc Chimie, EP0667331, 1995 (H. Lefranc). 161 S.E. Allen, R.R. Walvoord, R. Padilla-Salinas, M.C. Kozlow-
130 ARCO Chemical Technology, EP0677518, 1995 (J.G. Zajacek, ski, Chem. Rev. 113 (2013) 62346458, see section 7.3.2.
G.L. Crocco, J.J. Jubin Jr.). 162 Chemical Profile: Formic Acid, ICIS Chemical Business
131 ARCO Chemical Technology, EP0732327, 1999 (G.L. Crocco, 2006 26 July 2006, http://www.icis.com/resources/news/
J.J. Jubin Jr., J.G. Zajacek). 2006/07/26/2015258/chemical-profile-formic-acid/ (accessed
132 V.N. Chichagov, V.M. Solyanikov, T.G. Bondarenko, G.A. 19th May 2015).
Tkacheva, A.P. Zhavoronkov, E.T. Denisov, Neftekhimiya 20 163 Saudi Basic Ind. Corp., US6087297, 2000 (M.H.K. Karim, A.
(1980) 55967. K. Al-Hazmi).
133 DSM, WO2005/030690, 2005 (C. Daguenet, R. Parton, J.T. 164 Saudi Basic Ind. Corp., US6030920, 2000 (K. Karim, E.
Tinge). Mamedov, M.H. Al-Hazmi, A.H. Fakeeha, M.A. Soliman, Y.
134 K. Jopp, PROCESS 17 (2010) 4041. S. Al-Zeghayer, A.S. Al-Fatish, A.A. Al-Arify).
135 BASF, WO2005/030690, 2005 (J.H. Teles, B. Rossler, R. 165 Saudi Basic Ind. Corp., WO99/13980, 1999 (K. Karim, M.H.
Pinkos, T. Genger, T. Preiss). Al-Hazmi, E. Mamedov).
136 BASF, WO2010/023211, 2010 (J.H. Teles, W. Ruppel, U. 166 W. Chu, Y. Ooka, Y. Kamiya, T. Okuhara, H. Hattori, Chem.
Wegerle, A. Meier, T. Genger, M. Schelper, P. Resch). Lett. 34 (2005) 642643.
137 Firmenich, CH474500, 1969 (L. Re, G. Ohloff). 167 W. Partenheimer, Appl. Catal. A 409410 (2011) 4854.
138 W.W. Kaeding, Hydrocarbon Processing 43 (1964) 168 H.R. Cooper, H.W. Melville, J. Chem. Soc. 1951 19842002.
173176. 169 J.R. McNesby, C.A. Heller Jr., Chem. Rev. 54 (1954) 325346.
139 W. Buijs, J. Mol. Catal. A: Chem. 146 (1999) 237246. 170 BASF, US5504229, 1996 (K. Fischer, G. Fritz).
140 L. van Dierendonck, P. de Jong, J. von den Hoff, H. Vonken, R. 171 Mitsubishi Chem. Ind. Co., JP53105413, 1978 (T. Maki).
Vermeijs, Adv. Chem. Ser. 133 (1974) 432448. 172 Farbwerke Hoechst, DE1518847, 1969 (O. Probst, G. Mau, L.
141 Emerald: Emerald Performance Materials completes acquisi- Hornig).
tion of DSM Special Products, BV, press release, Dec. 16th, 173 R. Raja, S. Lee, M. Sanchez-Sanchez, G. Sankar, K.D.M.
2010, http://www.emeraldmaterials.com/cms/epm/page.html? Harris, B.F.G. Johnson, J.M. Thomas, Top. Catal. 20 (2002)
p_name2010 (accessed 19th May 2015). 8588.
142 G.I. Panov, K.A. Dubkov, A.S. Kharitonov in N. Mizuno (ed.): 174 R.A. Reimer, C.S. Slaten, M. Seapan, T.A. Koch, V.G. Triner
Modern Heterogeneous Oxidation Catalysis, Wiley-VCH, in J. van Ham et al. (eds.): Implementation of Technologies
Weinheim 2009, pp. 223230. for Abatement of N2O Emissions Associated with Adipic
143 V.N. Parmon, G.I. Panov, A. Uriarte, A.S. Noskov, Catal. Acid Manufacture, 2nd Symposium on Non-CO2 Green-
Today 100 (2005) 115131. house Gases (NCGG-2), Noordwijkerhout, The Netherlands,
144 JLM, Solutia call off phenol agreement, Chemical Week, 810 Sept. 1999, Kluwer Academic Publishers, Dordrecht,
August 8, 2001. pp. 347358.
100 Oxidation

175 U.S. Environmental Protection Agency: Inventory of 205 VEB Leuna-Werke Walter Ulbricht, DD128507, 1977 (W.
U.S. Greenhouse Gas Emissions and Sinks: 19902011, Eifert, G. Kotte, I. Rangott, K. Ritter).
April 12, 2013, http://www.epa.gov/climatechange/ 206 Clariant, EP0890583, 1998 (G. Hohner).
emissions/usinventoryreport.html (accessed 4th June 2015). 207 Z. Liyan, D. Wie, Shiyou Huagong (Petrochemical Technology)
176 U.S. Environmental Protection Agency: Methane and Nitrous 39 (2010) 818822.
Oxide Emissions from Natural Sources, April 2010, http:// 208 A. Trunschke: Nanostructured Catalysts: Selective
www.epa.gov/methane/sources.html (accessed 4th June 2015). Oxidations in C. Hess, R. Schlogl (eds.): RSC Nanoscience
177 Rhodia Fiber and Resin Intermediates, WO2000/59858, 2000 & Nanotechnology, No. 19, Royal Society of Chemistry, 2011,
(E. Fache, M. Constantini). Chap. 4, pp. 5695.
178 Y. Ishii, S. Sakaguchi, Catal. Today 117 (2006) 105113. 209 T. Yamaguchi, K. Yokota, S. Yamamatsu, O. Nagano, M.
179 Daicel Chemical Industries, Ltd.: Environmental, Safety and Tyouno, A. Aoshima, Kagaku to Kogyo (Osaka, Japan) 86
Social Report 2008, pp. 2627. (2012) 1724.
180 S. Gong, J. Chen, Guangdong Huagong (Chemical Industry in 210 S. Yamamatsu, Kagaku Kogaku 66 (2002) 358360.
Guangdong) 38 (2011) 7576. 211 S. Yamamatsu, Shokubai 43 (2001) 549554.
181 Societe des Usines Chimiques de Rh^one-Poulenc, FR1501725, 212 Z. Chen, L. Zhang, Y. Li, F. Yang, X. Wang, Y. Qi, Y. Li,
1967 (A. Charamel, J. Descroix, J. Duroux, S. Siquet). Huaxue Gongcheng (Chemical Engineering (China)) 39
182 Bundeskartellamt, 3. Beschlussabteilung, B3-136/05, http:// (2011) 9298.
www.bundeskartellamt.de/SharedDocs/Entscheidung/DE/ 213 N. Ballarini, F. Cavani, C. Cortelli, S. Ligi, F. Pierelli, F. Trifiro,
Entscheidungen/Fusionskontrolle/2006/B3-136-05.pdf? C. Fumagalli, G. Mazzoni, T. Monti, Topics Catal. 38 (2006)
__blobpublicationFile&v3 (accessed 19th May 2015). 147156.
183 Invista, http://c12.invista.com/doc/files/314/Corfree%20M1% 214 Y.S. Bhat: Catalysts in Petroleum Refining and Pet-
20Technical%20and%20Corrosion%20Inhibitor% rochemicals, 10th Saudi-Japanese Symposium, Dhahran,
20Application.pdf (accessed 19th May 2015). Saudi Arabia, Nov. 1920, 2000, pp. 232237.
184 Henkel, EP0296506, 1988 (J. Schindler, U. Hanggi, H. 215 Consortium fur Elektrochemische Industrie, DE870846, 1953
Waldhoff). (W.O. Herrmann, W. Haehnel).
185 Ailin Zhang, China, CN102115768, 2011 (W. Cao, Y. Chen, M. 216 Eastman Kodak Co., US2945058, 1960 (R.M. Watson, G.K.
Caoxun). Finch).
186 A. K ockritz, A. Martin, Eur. J. Lipid Technol. 113 (2011) 217 I. Hermans, J. Peeters, L. Vereecken, P.A. Jacobs, ChemPhy-
8391. sChem 8 (2007) 26782688.
187 Hibernia, DE1418 074, 1968 (E. Rindtorff, K. Schmitt, H. 218 Dow Chemical Co., GB816420, 1959 (W.H. Taplin, G.H.
Heumann). Harris).
188 D. Zhao, L. Dong, J. Yang, Y. Zhang, Xiandai Huagong 219 H.E. Zaugg, R.T. Rapala, Org. Synth., Coll. Vol. 3 (1955)
(Modern Chemical Industry) 30 (2010) 2831. 84.
189 Y. Liu, Hebei Huagong (Hebei Chemical Industry) 35 (2012) 220 M. Stepanski, A. R utti, Sulzer Tech. Rev. 2 (2003) 46.
3536. 221 E. Schaefer, M. Stepanski, K. Fiedler, Hydrocarbon Eng. 7
190 K. Song, W. Cai, Y. Liu, Huagong ke-ji (Science & Technology (2002) 6466, 68.
in Chemical Industry) 15 (2007) 7377. 222 W. Partenheimer, Catal. Today 23 (1995) 69158.
191 European Commission: Support for the Development and 223 Y. Yin, Y. Xiang, L. Diao, H. Zhang, Huaxue Gongye Yu
Adoption of Monitoring and Reporting Guidelines and Gongcheng Jishu (J. Chem. Ind. & Eng.) 32 (2011) 3338.
Harmonised Benchmarks for N2O Activities for Unilateral 224 R.A.F. Tomas, J.C.M. Bordado, J.F.P. Gomes, Chem. Rev. 113
Inclusion in the EU ETS for 2008-12, Final Report, February (2013) 74217469.
2008, Entec UK Limited, Chap. 5, pp. 4547, http://ec.europa. 225 W. Partenheimer, Appl. Catal. A 481 (2014) 183189.
eu/clima/policies/ets/cap/allocation/docs/entec_study_2008_ 226 W. Partenheimer, Appl. Catal. A 481 (2014) 190195.
en.pdf (accessed 19th May 2015). 227 F. Cavani, A. Caldarelli, S. Luciani, C. Cortelli, F. Cruzzolin,
192 Clariant Specialty Fine Chemicals, WO2009/092734, 2009 Catalysis 24 (2012) 204222.
(F. Bleger, O. Simon, A. Schouteeten). 228 BASF, WO2004/103561, 2004 (S. Storck, J. Zuehlke, S. Neto,
193 Chemie Linz, AT376197, 1984 (A. Sajtos, M. Wechsberger, E. F. Rosowski).
Roithner). 229 BASF, WO2011/061132, 2011 (S. Altwasser, J. Zuehlke, F.
194 DSM Fine Chemicals: Explosion in Glyoxylic acid plant: Rosowski, C.K. Dobner).
current status of investigation, Press Release DSM Fine Chem- 230 S.I.S.A.S. Societa Italiana Serie Acetica Sintetica,
icals Austria, November 18th, 2004. DE2935461, 1980 (F. Codignola, F. Celeste).
195 Hoechst, DE2936123, 1981 (E.I. Leupold, W. Blau, H. Baltes). 231 Polynt, WO2006/119939, 2006 (G. Castiglioni, R. Pirola, C.
196 BASF, EP0048396, 1982 (W. Disteldorf, W. Eisfeld). Fumagalli).
197 Trojan Powder Co., US3312736, 1967 (R.J. Ruhf). 232 D. Jianfei, P. Weichun, W. Yibiao, C. Zhongping, L. Dongchun,
198 W.R. Thiel, Eur. J. Org. Chem. (2004) 31083112. Yun Zhi, S. Rong, Huagong Xingxing Cailiao (New Chemical
199 Perstorp Discontinues Production of the Specialty Polyol Materials) 40 (2012) 1517.
DMBA, Chem. Week 2009, Sept. 23rd. 233 Kawasaki Steel Corp., US4925957, 1990 (N. Enomoto, Y.
200 N. Singh, D.G. Lee, Org. Proc. Res. Dev. 5 (2001) 599603. Naruse).
201 K. Nakagawa, R. Konaka, T. Nakata, J. Org. Chem. 27 (1962) 234 Nippon Steel Chem. Co., US5225572, 1993 (T. Hara, N. Daito,
15971601. M. Takeuchi, K. Wada).
202 BASF, US5064908, 1991 (L. Schuster, A. Hettche, W. Liedy, S. 235 Nippon Shokubai Co., US6452021, 2002 (T. Takahashi, Y.
Weiss, L. Ehemann). Emoto, E. Matsunami).
203 BASF, EP0853089, 1998 (A. Deckers, W. Weber, D. Ebisch). 236 UBE Industries, US8541615, 2011 (M. Nishio, H. Imajima, T.
204 Honeywell, US3322711, 1967 (R.W. Bush, C.L. Kehr). Nakamura Y. Yamauchi).
 Oxidation 101

237 W. Wu, M. Guo, Huaxue Yu Nianhe (Chemistry and Adhesion) 270 A. Ammala: Nylon-MXD6 resins for food packaging, Mul-
3 (2002) 173175. tifunctional and Nanoreinforced Polymers for Food Packaging,
238 R.A. Bergamin, Chimia 44 (1990) 255257. Woodhead Publishing, Cambridge 2011, pp. 243260.
239 R. Chuck, Appl. Catal. A 280 (2005) 7582. 271 Mitsubishi Gas Chemical Co., Increasing Capacity at MXDA
240 BASF, WO2005/118545, 2005 (J. Petzoldt, H. Wilmer, F. Facilities, Press release, February 22nd, 2011.
Rosowski). 272 R.G. Rizayev, E.A. Mamedov, V.P. Vislovskii, V.E. Sheinin,
241 C. Zhang, B. Zhou, Huaxue Yu Nianhe (Chemistry and Adhe- Appl. Catal. A 83 (1992) 103140.
sion) 34 (2012) 5457. 273 J. Chen, Jingxi Huagong (Fine Chemicals) 16 (1999) 69.
242 Mitsubishi Gas Chemical to switch process for 2,6-NDCA, 274 Enichem Anic, EP0564040, 1993 (M.A. Mantegazza, G. Pet-
New Materials Japan, Jan 1st, 1996. rini, A. Cesana).
243 Liaoning Liangang Pigment and Dyestuff Chemicals Co., 275 M.G. Clerici, Chimica e lIndustria (Milan, Italy) 85 (2003)
CN101601996, 2009 (W. Zhang, Y. Shen, S. Xu, Z. Zhang, 5556.
P. Huang, Y. Wang). 276 E. VanSavage: Sumitomo, EniChem to Build a Plant Featur-
244 Nippon Steel, DE3812423, 1988 (T. Hara, H. Sugawara). ing New Caprolactam Process, ICIS Chem. Bus., October
245 BASF, DE-OS2029029, 1971 (O. Kratzer, H. Suter, F. Wirth). 16th, 2000.
246 Liaoning Liangang Pigment and Dyestuff Chemicals Co., 277 Japans Sumitomo launches new caprolactam, PO lines, ICIS
http://www.lyliangangdyes.com/about_en.html (accessed News, April 17th, 2003.
19th May 2015). 278 Degussa, EP1433777, 2004 (S. Leininger, J. Herwig, M. Roos,
247 BASF, DE2837731, 1980 (H. Hoch, H. Hiller). G. Oenbrink).
248 C.E. Sumner, J.R. Zoeller, Chem. Ind. (Dekker) 49 (1993) 279 J. Bieleman: Antiskinning Agents Additives for Coatings
22540. Wiley-VCH, Weinheim 2000.
249 Q. Li, M. Tao, M. Zhang, Fenzi Cuihua (J. Mol. Catal. China) 280 D. Edelmann, M. Gawol, G. Link, Farbe Lack 107 (2001)
26 (2012) 478485. 5466.
250 D. Kumar, M.S. Chen, D.W. Goodman, Catal. Today 123 281 N.M. Martyak, D. Alford, R. Picker, C. Dowling, JCT Coat-
(2007) 7785. ingsTech 2 (2005) 3641.
251 C. Della Pina, E. Falletta, L. Prati, M. Rossi, Chem. Soc. Rev. 282 P. Wu, Y. Kubota, T. Yokoi, ACS Catal. 4 (2014) 2330.
37 (2008) 20772095. 283 J. Han, J. Dun, Jingxi Yu Zhuanyong Huaxuepin (Fine and
252 H.J. Schmidt, G. Roscher, CompendiumDeutsche Gessel- Specialty Chemicals) 13 (2005) 2730.
schaft fur Mineral olwissenschaft und Kohlechemie 7576 284 G. Olah, R. Malhotra, S.C. Narang: Nitration, Methods and
(1975) 318326. Mechanisms, VCH Publishers, New York 1989, pp. 219228.
253 Y. Tanabe, Hydrocarbon Process. Int. Ed. 60 (1981) 187190. 285 Societe Chimique de la Grande Paroisse, Azote et Produits
254 Daicel Chemical Ind., EP0974581, 1998 (H. Matsuda, E. Chimiques, US4260838, 1981 (P. Lhonore, J. Quibel, B.
Takemoto). Jacquinot).
255 M. Guo, R. Qiu, S. Yin, Shiyou Huagong (Petrochemical 286 P.W. Sherwood, Chem. Prod. Chem. News 19 (1956) 192193.
Technology) 41 (2012) 354363. 287 A. Richter, Ind. Fire World 22 (2007) 1011.
256 M.A. Pacheco, C.L. Marshall, Energy & Fuels 11 (1997) 229. 288 Angus Chem. Co., WO2009/129099, 2009 (D.M. Trauth, R.L.
257 Eni, Polimeri Europa, DimethylcarbonateProprietary process James).
technology, http://www.eni.com/it_IT/attachments/azienda/ 289 K. Weissermel, H.-J. Arpe: Industrielle Organische Chemie
attivita-strategie/petrolchimica/licensing/Dimethylcarbonate- Bedeutende Vor- und Zwischenprodukte, 3rd ed., VCH, Wein-
A4-lug09.pdf (accessed 19th May 2015). heim 1988, p. 271.
258 Snamproggetti, US5194656, 1993 (F. Ancillotti, E. 290 E. I. DuPont de Nemours and Co., US3255261, 1966 (V.E.
Pescarollo). Melllo, T.W. Wilson, A.B. Wester).
259 W.N. De Graaf, S. Afr. Chem. Process. 3 (1968) 11318. 291 E. I. DuPont de Nemours and Co., US2883433, 1959 (C.P.
260 R. Horn, R. Schloegl, Catal. Lett. 145 (2015) 2329. Spaeth).
261 R.K. Grasselli, RSC Nanosci. Nanotechnol. 19 (2011) 96140. 292 E. I. DuPont de Nemours and Co., US2883432, 1959 (C.P.
262 R.K. Grasselli, Top. Catal. 21 (2002) 7988. Spaeth).
263 R.K. Grasselli, C.G. Lugmair, A.F. Volpe Jr., Top. Catal. 54 293 F. Hoffmann-La Roche, DE2635958, 1977 (K. Meyer).
(2011) 595604. 294 Bayer, EP0634387, 1995 (J.-D. Jentsch, H. Landscheidt, B.
264 A. Parmar, E. Burridge: Chemical Profile: Acrylonitrile, Lenders, B. Pennemann, E. Wolters, E. Zirngiebl).
ICIS Chemical Business 2012, Feb. 27th. 295 J. Cui, Mei Huagong (Coal Chemical Industry) 39 (2011) 912.
265 Asahi Kasei Chemicals Corp.: Opening Ceremony for 296 I. Hermans, H. Teles, R. Dehn, T. Bruggemann, R. Send, P.N.
New Acetonitrile Plant in Korea, Press release, March 28th, Plessow, Top. Catal. 57 (2014) 12561264.
2014. 297 E. I. DuPont de Nemours and Co., US3215741, 1965 (A.F.
266 INEOS: INEOS Nitriles breakthrough secures global supply Chadwick).
of acetonitrile, press release, October 13th, 2009, http://www 298 Millmaster Onyx Corp., US3432555, 1969 (H. Mahnken).
.ineos.com/businesses/ineos-nitriles/news/ineos-nitriles- 299 Ethyl Corp., US4889954, 1989 (K.S. Laurenzo, D.P. Bauer).
breakthough-secures-global-supply-ofacetonitrile/ (accessed 300 Ethyl Corp., EP0320694, 1989 (D.P. Bauer, T.K. Summerford).
19th May 2015). 301 K. Suzuki, Fragrance J. 10 (1997) 8791.
267 INEOS, WO2011/119207, 2011 (C.J. Besecker, J.F. Brazdil, 302 Pennsalt Chemicals Corp., US3274252, 1966 (H.E. Albert, P.
M.A. Toft, M.J. Seely, R. Gustaferro). G. Haines).
268 BASF, DE3540517, 1987 (H. Engelbach, R. Krokoszinski, W. 303 S. Murahashi, H. Mitsui, T. Shiota, T. Tsuda, S. Watanabe, J.
Franzischka, M. Decker). Org. Chem. 55 (1990) 17361744.
269 Mitsubishi Gas Chem. Co., EP2671864, 2013 (T. Kumano, Y. 304 Enichem, EP0314147, 1988 (S. Tonti, P. Roffia, A. Cesana, M.
Ibi, K. Arasuna, S. Nagao). A. Mantegazza, M. Padovan).
102 Oxidation

305 W.R. Grace & Co., US4918194, 1988 (S. Marans, S.G. Harsy, 341 N.M. Emanuel, G.E. Zaikov, Z.K. Maizus: Oxidation of
H.S. Harris). Organic Compounds (Medium Effects in Radical Reactions),
306 Atofina, EP0835688, 1997 (R. Teissier, E. Jorda). Pergamon Press, Oxford 1984.
307 A. Turbak in C. Woodings (ed.): Regenerated Cellulose Fibres, 342 R.A. Sheldon: Catalytic Oxidations in the Manufacture of
Woodhead Publishing, Cambridge 2001, pp. 174177. Fine Chemicals, in G. Centi, F. Trifiro (eds.): New Develop-
308 A.F. Turbak, R.B. Hammer, R.E. Davies, H.L. Hergert, CHEM- ments in Selective Oxidation, Elsevier Science Publ., Amster-
TECH 10 (1980) 5157. dam 1990.
309 Eastman Kodak Co., US3447939, 1969 (D.L. Johnson). 343 Catalytica Associates, Selective Catalytic Oxidation of Hydro-
310 BASF Corp., US5502188, 1996 (K. Massonne, G. Konrad, M. carbons: A Critical Analysis, Study no. 1077, Santa Clara,
D. Sandison, G.E. Moffitt, L.E. James, J. Banger). Calif., Oct. 1977.
311 K. German, Chemik 59 (2006) 242247. 344 Ullmanns Encyklop adie der Technischen Chemie, 3rd ed., vol.
312 N.B. Muddada, T. Fuglerud, C. Lamberti, U. Olsbye, Top. 13, Urban und Schwarzenberg, Munchen 19511970,
Catal. 57 (2014) 741756. pp. 95109.
313 M. Conte, C.J. Davies, D.J. Morgan, T.E. Davies, D.J. Elias, A. 345 Houben-Weyl Methoden der organischen Chemie, 4th ed., vol.
F. Carley, P. Johnston, G.J. Hutchings, J. Catal. 297 (2013) 4, Georg Thieme, Stuttgart, part 1 A and 1 B.
128136. 346 Organic Chemistry, A Series of Monographs, Academic Press,
314 E. I. du Pont de Nemours, US2389628, 1945 (W.A. Lazier). New York, K.B. Wiberg (ed.): Oxidation in Organic Chem-
315 W. Mathes, Angew. Chem. 52 (1939) 591592. istry, vol. 5 a, 1965; W.S. Trahanovsky (ed.): Oxidation in
316 R.M. Crawford, Chem. Eng. Progr. 46 (1950) 483485. Organic Chemistry, vol. 5 b, 1973.
317 P.W. Sherwood, Petrol. Process. (1953) 15431545. 347 R.L. Augustine (ed.): Technics and Application in Orgnanic
318 N.A. Bhatt, Chem. Eng. World 15 (1980) 6574. Chemistry, Marcel Dekker, New York, Oxidation, vol. I,
319 The Dow Chem. Co., US3674664, 1972 (E.R. Larsen, A.K. 1969; Oxidation,, vol. II, 1971.
Keller, R.A. Plepys). 348 Catalytica Associates, Advances in Multimetallic Catalysis,
320 W.T. Cave, Ind. Eng. Chem. 45 (1953) 18531857. Study no. 4155, Santa Clara, Calif., May 1982.
321 Shandong Dacheng Pesticide Co., CN102924248, 2013 (S. 349 Catalytica Associates, A Practical Guide to Catalyst Testing,
Xiao, S. Li, Y. Li, H. Liu, L. Chen, X. Li, X. Wang, S. Qiao). Study no. 4186 TE, Mountain View, Calif., 1987.
322 Y. Liu, H. Zhang, Tianjin Huagong (Tianjin Chemical Industry) 350 V.Ya. Shtern: The Gas-Phase Oxidation of Hydrocarbons,
27 (2013) 912. Pergamon Press, Oxford 1964.
323 Societe Nationale Elf Aquitaine, EP0194931, 1986 (J. Ollivier, 351 N.N. Semenov: O nekoterikh problemakh khimicheskei kinetiki
H. Baptiste, C.H. Lagaude, M. Larrouy). i reaktsionnoi sposobnosti, Publ. Akad. Nauk SSSR, Moscow
324 Elf Atochem, US6045664, 2000 (J. Ollivier). 1958.
325 Elf Atochem, EP0675107, 1995 (S.G. Schon, G.A Wheaton, T. 352 N.N. Semenov: Some Problems of Chemical Kinetics and
H. Overgaard). Reactivity, Pergamon Press, London 1959.
326 BASF, US6531629, 2003 (M. Eiermann, C. Tragut, K. Ebel). 353 N.N. Semenov: Einige Probleme der chemischen Kinetik und
327 BASF: BASF increases production capacity for methane sul- Reaktionsf ahigkeit, (Freie Radikale und Kettenreaktionen),
fonic acid in Ludwigshafen, BASF press release P398/12e, Akademie-Verlag, Berlin 1961.
September 6th, 2012. 354 O.L. Magelli, Ch.S. Sheppard in [37], vol. I, pp. 1104.
328 K. Rasheed in K.R. Langer (ed.): Surfactants, A Practical 355 R.D. Mair, R.T. Hall, in [37], vol. II, pp. 535636.
Handbook, Hanser Gardner Publications Inc., Cincinnati, 356 R. Hiatt in [37], vol. II, pp. 1151.
1999, pp. 103106. 357 Houben-Weyl Methoden der organischen Chemie, 4th ed., vol.
329 J.P. Canselier in U. Zoller (ed.): Handbook of Detergents, Part 8, Georg Thieme, Stuttgart, pp. 374.
F: Production, CRC Press, Boca Raton 2008, pp. 142151. 358 W. Bartknecht: ExplosionenAblauf und Schutzmanahmen,
330 H. Ramloch, G. Tauber, Chem. unsere Zeit 13 (1979) 157162. Springer-Verlag, Berlin 1980.
331 H.G. Hauthal, B. Makowka, M. Pieroth: 5th World Surfactants 359 W. Berthold, U. Loffler: Lexikon sicherheitstechnischer
Congress, Firenze, Italy, May 29June 2, 2000, pp. 295904. Begriffe in der Chemie, Verlag Chemie, Weinheim, Germany,
332 Z. Yi, C. Ren, Z. Hu, Jingxi Huagong Zhongjianti (Fine 1981.
Chemical Intermediates) 33 (2003) 1112, 18. 360 W. Bartknecht: StaubexplosionAblauf und Schutzmanah-
333 A.Kh. Sharipov, Chem. Technol. Fuels Oils 37 (2001) 6272. men, Springer-Verlag, Berlin 1987.
334 China Petroleum & Chemical Corp. and Research Inst. of 361 H. Freytag: Handbuch der Raumexplosionen, Verlag Chemie,
Petroleum Processing, Sinopec, WO2014/067235, 2014 (C. Weinheim, Germany, 1965.
Shi, M. Lin, X. Shu, X. Mu, B. Zhu). 362 K. Nabert, G. Schon: Sicherheitstechnische Kennzahlen brenn-
335 Development Direction of the Dimethyl Sulf oxide Sector, barer Gase und D ampfe, 2nd ed., Deutscher Eichverlag,
China Chemical Reporter 2004, Sept 6th, 21. Braunschweig 1978.
336 25 000 t/a DMSO Project Activated in Guizhou Province, 363 K. Nabert, G. Schon: Sicherheitstechnische Kennzahlen brenn-
China Chemical Reporter 2012, Aug 6th, 13. barer Gase und D ampfe, 5th suppl., Deutscher Eichverlag,
337 Sloss Industries Corp., US6552231, 2003 (C.A. Jones, J.D. Braunschweig 1980.
Friedrich, J.C. Paschall, W.K. Alwis). 364 D.R. Stull: Fundamentals of Fire and Explosion, AICHE
338 G. Scott: Atmospheric Oxidation and Antioxidants, Elsevier Monogr. Ser. 73 (1977) no. 10.
Publishing Comp., Amsterdam 1965. 365 L. Bretherick: Handbook of Reactive Chemical Hazards, But-
339 R.A. Sheldon, J.K. Kochi: Metal-Catalyzed Oxidations of terworth, London 1975.
Organic Compounds, Academic Press, New York 1981. 366 H.F. Coward, G.W. Jones: Limits of Flammability of Gases
340 Autoxidation von Kohlenwasserstoffen. Technische For- and Vapors, Bull U.S. Bur. Mines 1952, no. 503.
tschrittsberichte, vol. 65, Deutscher Verlag fur Grundstoffin- 367 K.N. Palmer: Dust Explosions and Fires, Chapman and Hall,
dustrie, Leipzig 1981. London 1973.
 Oxidation 103

368 E.T. Denisov, T.G. Denisova: Handbook of Antioxidants, 2nd 383 V. Schroder, M. Molnarne, J. Haz. Mat. A121 (2005) 3744.
ed., CRC Press, Boca Raton 2000, pp. 1764. 384 M. Molnarne, P. Miszey, V. Schroder, J. Haz. Mat. A121 (2005)
369 Akzo Nobel, Initiators and Reactor Additives for Thermo- 4549.
plastics, 2010, https://www.akzonobel.com/polymer/system/ 385 Landolt-B ornstein, 6th ed., vol. 2, Springer, Heidelberg 1971,
images/AkzoNobel_Initiators_and_Reactor_Additives_for_ part 2 b, 1 ff.
Thermoplastics_Low-res_protected_July%202010_tcm96- 386 J. Smidt et al., Angew. Chem. 71 (1959) 176.
39468.pdf (accessed 19th May 2015). 387 R. Jira in B. Cornils, W.A. Hermann (eds.): Applied Homoge-
370 Pergan, Organische Peroxide fur die Polymerisation, 2013, neous Catalysis with Organometallic Compounds, VCH Ver-
http://www.pergan.com/Downloads/ lagsgesellschaft, Weinheim, Germany, 1996, p. 374393.
Organische_Peroxide_fuer_die_Polymerisation.pdf (accessed 388 P.A. Jacobs in G. Ertl, H. Knozinger (eds.): Handbook of
19th May 2015). Heterogeneous Catalysis, VCH Verlagsgesellschaft, Weinheim
371 United Initiators, Peroxides in action, http://www.united- 1997, p. 22842295.
initiators.com/files/media/Product-Range.pdf (accessed 19th 389 Catalysis Institute Nowosibirsk, DEOS2516244, 1975.
May 2015). 390 D.J. Hucknall: Selective Oxidation of Hydrocarbons, Aca-
372 D. Swern (ed.): Organic Peroxides, Wiley-Interscience, New demic Press, London 1974.
York, vol. I, 1970; vol. II, 1971; vol. III, 1972. 391 R. Higgins, P. Hayden in: Catalysis, vol. I, Chemical Society,
373 R. Schoellner: Die Oxidation organischer Verbindungen mit London 1977.
Sauerstoff, Akademie-Verlag, Berlin 1964. 392 W.D. Mross, J. Jira in J. Falbe, U. Hasserodt (eds.): Kataly-
374 Wataru Ando (ed.): Organic Peroxides, J. Wiley & Sons, New satoren, Tenside und Mineral oladditive, G. Thieme, Stuttgart
York 1992. 1978.
375 L. Vereecken, T.L. Nguyen, I. Hermans, J. Peeters, Chem. 393 M.V. Twigg (ed.): Catalyst Handbook, Wolfe Publ. Co., Lon-
Phys. Lett. 393 (2004) 432436. don 1989.
376 I. Hermans, J. Peeters, P.A. Jacobs, J. Org. Chem. 72 (2007) 394 Shell, US 4012425, 1972 (R.P. Nielsen, J.H. LaRochelle).
30573064. 395 W.M.H. Sachtler, C. Backx, R.A. van Santen, Catal. Rev. Sci.
377 I. Hermans, J.F. Muller, T.L. Nguyen, P.A. Jacobs, J. Peeters, J. Eng. 23 (1981) 127149.
Phys. Chem. A 109 (2005) 43034311. 396 Shell, EP0266015, 1988 (A.M. Lauritzen).
378 A.A. Boyd, P.M. Flaud, N. Daugey, R. Lesclaux, J. Phys. 397 E. Cremer: Advances in Catalysis, vol. 7, Academic Press, New
Chem. A 107 (2003) 818821. York 1955.
379 U. Neuenschwander, F. Guignard, I. Hermans, ChemSusChem 398 L.K. Doraiswamy, D.G. Tajbl, Catal. Rev. Sci. Eng. 10 (1974)
3 (2010) 7584. 177219.
380 U. Neuenschwander, I. Hermans, Phys. Chem. Chem. Phys. 12 399 J. Nelles et al., Chem. Techn. (Leipzig) 30 (1978) 441446,
(2010) 1054210549. 555559.
381 U. Neuenschwander, E. Meier, I. Hermans, ChemSusChem 4 400 H. Pichler, F. Obenaus, Brennst. Chem. 45 (1964) 97103; 46
(2011) 16131621. (1965) 258264.
382 I. Hermans, P. Jacobs, J. Peeters, Chem. Eur. J. 13 (2007)
754761.