The high yield synthesis of benzaldehydes from benzylic alcohols using homogeneously catalyzed aerobic oxidation in acetic acid

Article abstract of DOI:10.1002/adsc.200505394

The kinetics and yields for the formation of benzaldehydes from benzylic alcohols via aerobic oxidation using homogeneous catalysts in acetic acid are reported. For compounds of the type X-C₆H₄CH₂OH where X = H, 4-CH₃, 4-OCH₃, 4-Cl, 4-COOH the observed yields were 84, 78, 57, 88 and 52 mol %, respectively. These yields are comparable to those reported for the stoichiometric Co(III)/Br oxidation of these same benzylic alcohols. These results are limited to benzylic alcohols. The high yields are possible for two reasons: 1) the rate of reaction exhibits two maxima in which the benzaldehyde is predominately being generated in the first phase and the carboxylic acid in the second and 2) the competing esterification of the benzylic alcohol with the acetic acid solvent is an order of magnitude slower. A variety of other catalysts were also evaluated which give lower yields mainly because they are less active and the competing esterification reaction becomes more important. Water addition to the acetic acid increased the rate of acetylation of the alcohol, contrary to expectation. The oxidation of the benzaldehyde appears to be inhibited by the presence of the benzylic alcohol. Even though the rate of oxidation of the benzaldehyde, by itself, is equally reactive as the benzylic alcohol, the benzaldehyde does not significantly react until the benzylic alcohol concentration is very low. Two possible reasons, using a free-radical chain mechanism, are suggested to explain this behavior. A mechanism demonstrating why the stoichiometric and catalytic reactions give similar yields is given.

Full text of DOI:10.1002/adsc.200505394

FULL PAPERS
DOI: 10.1002/adsc.200505394
The High Yield Synthesis of Benzaldehydes from Benzylic
Alcohols using Homogeneously Catalyzed Aerobic Oxidation in
Acetic Acid
Walt Partenheimera, b
a
Central Research and Development, E. I. DuPont de Nemours & Co., Inc., Experimental Station, Wilmington, Delaware
19880-0328, USA
b
Current address: 16 Clermont Rd., Wilmington, Delaware 19803, USA
E-mail: delparts@att.net
Received: October 8, 2005; Accepted: December 16, 2005
Abstract: The kinetics and yields for the formation of mainly because they are less active and the competing
benzaldehydes from benzylic alcohols via aerobic ox- esterification reaction becomes more important. Wa-
idation using homogeneous catalysts in acetic acid are ter addition to the acetic acid increased the rate of
reported. For compounds of the type X-C6H4CH2OH acetylation of the alcohol, contrary to expectation.
where X¼H, 4-CH3, 4-OCH3, 4-Cl, 4-COOH the ob- The oxidation of the benzaldehyde appears to be in-
served yields were 84, 78, 57, 88 and 52 mol %, re- hibited by the presence of the benzylic alcohol.
spectively. These yields are comparable to those re- Even though the rate of oxidation of the benzalde-
ported for the stoichiometric Co(III)/Br oxidation of hyde, by itself, is equally reactive as the benzylic alco-
these same benzylic alcohols. These results are limited hol, the benzaldehyde does not significantly react un-
to benzylic alcohols. The high yields are possible for til the benzylic alcohol concentration is very low. Two
two reasons: 1) the rate of reaction exhibits two max- possible reasons, using a free-radical chain mecha-
ima in which the benzaldehyde is predominately be- nism, are suggested to explain this behavior. A mech-
ing generated in the first phase and the carboxylic anism demonstrating why the stoichiometric and cat-
acid in the second and 2) the competing esterification alytic reactions give similar yields is given.
of the benzylic alcohol with the acetic acid solvent is
an order of magnitude slower. A variety of other cat- Keywords: autoxidation; benzylic alcohols; bromide;
alysts were also evaluated which give lower yields cobalt; manganese
Introduction
Autoxidation is defined as the reaction of organic sub-
strates with molecular oxygen using a free radical chain
The conversion of alcohols to carbonyl compounds is mechanism. The oxidizability of a substrate is defined by
one of the most basic organic chemical transforma- the expression kp/(2kt)1/2 where kp and kt are the propa-
tions.[1] The use of molecular oxygen, present in the air gation and termination rate constants, respectively. In
at 20.9%, as the primary oxidant is especially economi- the non-catalyzed autoxidation of methylaromatics the
cally useful and there are many previous reports using reactivities of the methyl group and intermediate ben-
this approach.[2] We report here the high yield synthesis zylic alcohols and benzaldehydes, are 0.05, 0.85 and
of benzaldehydes from benzyl alcohols using homoge- 290, respectively.[4] Thus, since the reactivity of the ben-
neous catalysts based on cobalt(II) and the highly active zaldehyde is far higher than the alcohol, high yields of
of metal/bromide catalysts[3] such as the Co/Mn/Br mix- benzaldehydes from benzylic alcohols are not expected.
ture. The Co/Mn/Br-catalyzed aerobic oxidation of p- In catalyzed autoxidation, these reactivities can change.
xylene to terephthalic acid, used in the production of During modeling studies in the Co/Mn/Br-catalyzed ox-
polyesters, is a well-established, large-scale industrial idation of toluene the intermediate concentrations of
process. The problems associated with metal/bromide the alcohols and aldehydes remain low, consistent with
catalysts, such as their corrosive characteristics, have their oxidizability, but the alcohol is found to be 3 times
been largely eliminated through 55 years of optimiza- more reactive than the aldehyde.[5] Also with the Co/
tion of these processes in the polyester industry.
Mn/Br oxidation of hydroxymethylfurfural the alcohol
group of this compound is preferentially oxidized to
Adv. Synth. Catal. 2006, 348, 559 – 568
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
559

Walt Partenheimer
FULL PAPERS
the aldehyde so that high yields ofthe dialdehyde are ob- Results and Discussion
served.[5] For these reasons, we subsequently investigat-
ed the metal/bromide-catalyzed autoxidation of benzyl-
Reaction Sequence
ic alcohols (X-Ph-CH2OH). Some of this work has been
preliminarily published.[5]
The kinetics of the Co/Br-catalyzed autoxidation of The intermediates and products that form when benzyl-
benzyl alcohol in acetic acid has been previously report- ic alcohols were oxidized were established via GC/MS.
ed but no yields are stated.[6] The stoichiometric oxida- Monitoring the reaction with GC and subsequent kinet-
tion of benzyl alcohol has been kinetically studied with ic analysis are consistent with the sequence of reactions
Co(III) and Mn(III) with and without bromide.[7] There shown in Figure 1.
is a report that benzyl alcohol inhibits the radical-chain
The author did briefly investigate the metal/bromide
aerobic oxidation of 3-pentene-2-ol and 1-octanol.
Both substances reacted but the yields to their corre-
sponding aldehydes were low. Thus, so far, the results re-
ported here are limited to benzylic alcohols.
oxidation of benzaldehyde.[8]
Characterization of the Co/Mn/Zr/Br-Catalyzed
Oxidation of Benzyl Alcohol
Since these reactions are performed at ambient atmos-
pheric pressure, samples can be removed directly from
the reactor and their compositions accurately deter-
mined at any given point in time during the reaction
via subsequent GC and LC analysis. A typical result in
given in Figure 2 for the Co/Mn/Zr/Br-catalyzed autox-
idation of benzylic alcohol.
Up to 120 minutes, the principle reaction is reac-
tion (1) – the reaction of benzyl alcohol to benzalde-
hyde. The competing reaction to give the benzyl acetate,
reaction (3), is much slower as is reaction (2) to the ben-
zoic acid. Subsequently the maximum benzaldehyde is
84%. After 120 minutes, the rate of reaction (2) sudden-
ly increases and nearly quantitative yields of the benzoic
acid are obtained. The benzyl acetate concentration also
decreases after 120 min due to its oxidation and possibly
hydrolysis.
Figure 1. Reaction sequence observed during the aerobic ox-
idation of benzylic alcohols.
Figure 2. Distribution of products during the aerobic oxidation of benzyl alcohol; data from example 1 of Table 1.
560
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 559 – 568

Synthesis of Benzaldehydes from Benzylic Alcohols
FULL PAPERS
Table 1. Initial conditions for the aerobic oxidation of benzylic alcohols and related reactions.
Example
Type of alcohol
Type of catalyst
[catalyst] M
[H2O] (wt %)
T [ 8C]
[substrate] M
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
H
H
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/Zr/NaBr
Co/Mn/HBr
0.013/0.013/0.026/0.00050
0.013/0.013/0.026/0.00050
0.01/0.01/0.0049/0.06
0.013/0.013/0.026/0.00049
0.013/0.013/0.026/0.00047
0.01/0.01/0.0049/0.07
0.01/0.01/0.0049/0.02
0.013/0.013/0.026/0.00047
0.013/0.013/0.026/0.00049
0.01/0.01/0.02
0
0
0
0
0
0
0
0
3
0
5
0
5
0
0
75
75
80
75
75
80
80
95
75
95
95
95
95
75
75
0.790
0.790
0.150
0.790
0.791
0.150
0.944
0.794
0.790
0.919
0.926
0.917
0.921
0.790
0.79
4-CH3
4-OMe
4-Cl
4-COOH
H
H
H
H
H
Co/Mn/HBr
Co/Mn/HBr
Co/Mn/HBr
Co/Zr
0.01/0.01/0.02
0.01/0.01/0.02
0.01/0.01/0.02
12.1/2.4
4-Cl
4-Cl
H
H
Co/Mn/NaBr
0.026/0.026/0.052
Table 2. Composition of the mixture at maximum benzaldehyde yield. See Table 1 for Initial conditions. Example 2 is the aver-
age value for five independent experiments.
Example
Type of
alcohol
Time [h]
Conversion [%]
Benzaldehyde
Benzoic
acid
Benzylic
acetate
Selectivity to
benzaldehyde
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
H
H
2.1
2.4
6.5
–
3.5
3.1
2.5
1.6
15
1.8
3.9
1.8
8.3
–
89.1
88.6 (2.4)
89.2
94.0
91.9
75.0
93.7
88.8
100.0
86.1
83.6
77.8 (4.0)
77.7
57.0
88.4
52.0
71.5
59.7
78.0
66.0
54.8
75.6
62.7
60.0
72.0
9.42
6.03 (3.2)
–
–
acid insoluble
–
17
6.52
8.90
5.74
6.23
3.35
2.31
0.00
4.80
5.54
5.3 (0.9)
8.25
22.00
4.77
–
93.8
87.9 (5.5)
87.1
60.6
96.2
69.3
76.3
67.2
78.0
76.6
58.4
81.7
67.2
75.0
80.0
4-CH3
4-OMe
4-Cl
4-COOH
H
H
H
H
H
6.7
6.52
22.0
10.8
24.2
12.1
27.6
31.0
13.0
93.8
92.5
93.3
80.0
4-Cl
4-Cl
H
H
–
90.0
Experimental details of this reaction are given in ex- dizes at a rate constant of 43.0Â105 sÀ1. This is faster
amples 1 of Tables 1–3. Table 1 gives the initial experi- than the oxidation of benzylic alcohol (33.6Â105 sÀ1).
mental conditions, Table 2 the composition of the mix- This suggests that benzyl alcohol is inhibiting the reac-
ture at the maximum observed benzaldehyde yield, tion. 2) The rate of molecular oxygen uptake is biphasic
and Table 3 the kinetic results assuming first-order dis- with the second maximum occurring during the rapid
appearance of the reagents. When the benzaldehyde oxidation of the benzaldehyde. 3) When the benzalde-
yield is at its maximum value of 84% the acetate yield hyde begins to rapidly oxidize the color of the solution
is 5.5% (Table 2, example 1). This is consistent with changes from light blue to green. Each of these phenom-
the rate constants of 33.6 (Â105 sÀ1) for the oxidation ena are discussed further below.
of benzyl alcohol and 1.35 (Â105 sÀ1) for the esterifica-
tion of the benzyl alcohol by the acetic acid solvent (Ta-
ble 3, example 1). Reaction (1) is 24 times faster than re- Highest Yields to Benzaldehydes
action (3).
There are three unusual phenomena that occur in The reported benzaldehyde yields contain a considera-
these reactions. 1) After the maximum benzaldehyde ble amount of error since they are being obtained
yield of 84% is reached, the benzaldehyde rapidly oxi- when its concentration is rapidly changing during reac-
Adv. Synth. Catal. 2006, 348, 559 – 568
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
561

Walt Partenheimer
FULL PAPERS
Table 3. Kinetic analysis for the oxidation of benzylic alcohols and related reactions. See Table 1 for initial conditions. Stan-
dard deviation in parenthesis. Example 2 is the average value for five independent experiments.
Example
Type of alcohol
k1 (Â105)[a]
k2 (Â105)[b]
k3 (Â105)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
H
H
33.6 (0.8)
29.4 (3.5)
10.7 (0.5)
–
43.0 (2.8)
32.7 (9.2)
–
–
13.2 (5.8)
–
63 (13)
101 (8.3)
1.35 (0.05)
1.37 (0.09)
2.19 (0.10)
–
4-CH3
4-OMe
4-Cl
4-COOH
H
H
H
H
H
20.0 (3.8)
22.7 (1.9)
28.1 (2.6)
44.4 (3.3)
4.78 (0.22)
38.0 (3.3)
18.1 (0.83)
39.9 (4.2)
8.47 (0.87)
–
0.99 (0.19)
5.1 (1.0)
1.37 (0.41)
5.73 (0.18)
1.39 (0.08)
5.91 (0.18)
5.61 (0.29)
5.17 (0.30)
3.15 (0.26)
–
–
70 (15)
–
–
–
–
–
4-Cl
4-Cl
H
H
–
–
[a]
The rate constant k1 and k2 were obtained assuming first-order formation of the benzaldehyde and acetate from the benzyl
alcohol.
[b]
Obtained from the rate of disappearance of the benzaldehyde.
tion. Since GC samples are taken at roughly 15 min in- Influence of Experimental Variables
tervals, the measured yield may not represent the high-
est yield that actually occurred in the reaction. Except Temperature: For benzyl alcohol, increasing tempera-
for benzyl alcohol, no attempt was made to optimize tures from 75 to 80 to 958C gave increasingly lower
the yields.
yields, see examples 1, 7, 8.
The reaction for benzyl alcohol discussed above ap-
Water Concentration: It was initially thought that in-
pears to be generalfor benzylicalcohols. For compounds creasing the initial water concentration would increase
of the type X-C6H4CH2OH where X¼H, 4-CH3, 4- the benzaldehyde yields because the water would sup-
OCH3, 4-Cl, 4-COOH the maximum observed yields press the rate of acetylation of the alcohol, reaction (6).
were 84, 78, 57, 88 and 52 mol % , see examples 1–6 of
Tables 1–3. These substituents were chosen because
they represent a very large difference in reactivity of
ð6Þ
the corresponding substituted toluenes. The Hammett
substitutent values (sigmaþ) range from À0.78 to Experimentally just the opposite was observed. Invaria-
0.42 for the methoxy to chloro groups, respectively,[9] bly significantly higher concentrations of benzyl acetate
and give a difference in rates by a factor of 14.[3]
were present at the maximum benzaldehyde yield which
Rather striking, is that the yields reported here (84, 78, resulted in decreased yields. Compare examples 1 and 9
57, 88% for X¼H, 4-CH3, 4-OCH3, 4-Cl) are quite sim- for benzyl alcohol with a Co/Mn/Zr/Br catalyst at 758C,
ilar to those reported for the stoichiometric, non-aero- examples 10, 11 for benzyl alcohol with a Co/Mn/Br cat-
bic, Co(III)/Br oxidation of the same alcohols, reac- alyst at 958C, and examples 12, 13 for 4-chlorobenzyl al-
tion (5), which gave reported yields of 86, 81, 88, and cohol with a Co/Mn/Br catalyst at 958C. There are at
85%, respectively.[10] The reason for this similarity is dis- least two reasons for the increased acetate concentra-
cussed below.
tion:
1) Water is well-known to suppress the rate of oxida-
tion and indeed this was found when comparing the
rate constant for the oxidation of benzyl alcohol
(k1) in Table 3. The ratio of k1 is 33.5/4.78¼7 in ex-
amples 1, 9, 2.1 in examples 10, 11 and 4.7 in exam-
ples 12, 13. Thus, a longer time occurs before the
maximum yield is achieved at higher water concen-
tration and the alcohol has a longer time to react
with the acetic acid solvent.
2) Addition of water to acetic acid strongly affects the
dissociation constant, pKa.
ð5Þ
562
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 559 – 568

Synthesis of Benzaldehydes from Benzylic Alcohols
FULL PAPERS
ments in yield could be obtained. Also changes in their
composition (molar ratios of the metals and bromide)
and total catalyst concentration were evaluated. In gen-
ð7Þ
Water is a stronger nucleophile than acetic acid and its eral, but not invariably, as the activity of the catalyst in-
addition to acetic acid increases both the acetate and hy- creased, the yield increased. Thus for Co/Zr, Co/Mn/Br,
dronium ion concentrations. The pKa of 100% acetic and Co/Mn/Zr/Br catalysts at 758C, their increased ac-
acid is 13.92, with 2.5% water it is 10.95, and with 5% wa- tivity (in the order given see k1 in Table 3) gave higher
ter it is 9.64.[12] This corresponds to an increase in the hy- yields, compare examples 1, 14, 15. As the activity in-
dronium and acetate concentrations from 1.1Â10À7
M
creased from 1.7 to 8.0 to 33 (Â105 sÀ1), the yields in-
to 3.3Â10À6 M to 1.5Â10À5 M, respectively. One does creased from 60 to 72 to 84%. No yield improvements
not see an increase in acetylation due to the increase were observed with increases in catalyst concentration
in acetate concentration because the increase in acetate for X¼H at 758C with a Co/Mn/Zr/Br catalyst, for
concentration is too small. This was demonstrated by X¼H at 958C with a Co/Ce/Br catalyst, and X¼Cl at
measuring the rate of acetylation of benzyl alcohol 958C for a Co/Mn/Br catalyst. Different types of cata-
with and without the presence of Co(II) acetate. The lysts were evaluated at 958C. Observed maximum ben-
rate constants were, within experimental error, the zaldehyde yields were (yields in parenthesis) for X¼
same (3.03 and 2.71Â10À5 sÀ1, respectively) at 758C in H, Mn/Br (15), Co (65), Co/Br (61), Co/Ce/Br (67),
5% water/acetic acid. It is therefore suggested that it is Co/Mn/Br (66 at a Co/Mn/Br molar ratio of 1/1/2), Co/
the increase in acid concentration, by a factor of 138 Mn/Br (61 at a Co/Mn/Br molar ratio of 1.8/0.2/2.0),
on addition of 5 wt % water (2.78 M), that catalyzes Co/Mn/Br (63 at a Co/Mn/Br ratio of 1/1/1), and Co/
the rate of acetylation accounting for the higher benzylic Mn/Zr/Br (60). There is no obvious improvement in
acetate formation. These arguments assume that the benzaldehyde yield with these different catalysts at
trends observed for acetic acid/water mixtures at room 958C.
temperature, for which the pKas were measured, are
the same as at 758C and 958C where the oxidation ex-
periments were performed.
Biphasic Uptake of Molecular Oxygen
In light of the above, the complexity of these reactions
can be appreciated by noting that the water concentra- The concentration of the oxygen in the vent gas emanat-
tion is constantly increasingly during the oxidation:
ing from the reactor is continuously measured hence the
rate of reaction of dioxygen as a function of time can be
calculated. The rate of oxygen uptake as a function of
time for benzyl alcohol catalyzed by a Co/Mn/Zr/Br
mixture at 758C is given in Figure 3.
ð8Þ
Effect of Catalyst: A large number of different metal/
The rate of oxygen uptake slowly declines until
bromide catalysts were evaluated to discern if improve- 100 min. Then, unexpectedly, it again increases to amax-
Figure 3. Rate of oxygen uptake with benzyl alcohol using a Co/Mn/Zr/Br catalyst. From example 1 in Table 1.
Adv. Synth. Catal. 2006, 348, 559 – 568
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
563

Walt Partenheimer
FULL PAPERS
imum and subsequently declines in rate to nearly zero at even though in the same experiment one finds that the
around 250 min. This behavior has been observed 14 reactivities of the benzyl alcohol and benzaldehyde are
times in our laboratory with compounds of the type X- the same, but also that the yields to benzaldehyde are
C6H4CH2OH where X¼H, 4-CH3, 4-OCH3 with Co/ very high. We suggest that, while the benzyl alcohol is
Mn/Zr/Br and Co/Zr catalysts at 758C. Biphasic behav- present, it is inhibiting the oxidation of the benzalde-
ior is not observed at 958C, however. The minimum in hyde. It is only when the benzyl alcohol concentration
the rate of oxygen uptake at about 100 minutes occurs falls to quite lowvalues that this inhibition is sufficiently
at the maximum concentration of benzaldehyde in solu- weak so that a rapid oxidation of the benzaldehyde oc-
tion, see Figure 2. So the subsequent increase in the rate curs. This interpretation is also consistent with the bi-
of dioxygen uptake occurs due primarily to the oxida- phasic molecular oxygen uptake since, in the absence
tion of the benzaldehyde to the benzoic acid. The two of inhibition, simple monophasic behavior would be ob-
maxima in oxygen uptake suggest that the reactivities served.
of the benzylic alcohol and benzaldehyde are about
Carbon and hydrogen NMR was used to investigate if
the same since both substrates react with the same stoi- there were any interactions of benzyl alcohol with ben-
chiometric amount of dioxygen:
zaldehyde. No changes were observed when a mixture
of benzylic alcohol and benzaldehyde was measured as
a function of temperature in deuteoacetic acid at 40,
60, 80, 958C. The NMR spectra of benzaldehyde and
benzyl alcohol were examined in deuteochloroform
and dimethyl sulfoxide. In DMSO the chemical shift of
the H atom in the CHO group was 10.2 tau and the C
ð9Þ
ð10Þ
This is in agreement with the measured rates of oxida- atom was at 192 tau. However, there is a shift in these
tion of benzyl alcohol (k1) and benzaldehyde (k2), see resonances in deuteoacetic acid to 9.7 and 195 tau, re-
example 2 of Table 3.
spectively. This may suggest a hydrogen bonding inter-
Such biphasic behavior has not been observed in our action of the acetic acid with the aldehyde moiety of
laboratory during the last nine years during which time the benzaldehyde, see Figure 4a. This could conceivably
numerous different alkylaromatic substrates have stabilize the benzaldehyde relative to the benzyl alco-
been similarly evaluated – except for the case of 3,4-di- hol.
methoxytoluene – which has been previously reported
There are at least two other possibilities. It is known
in detail.[13] Other non-equilibrium behavior has been that acetic acid, water, and alcohols deactivate alkylper-
reported for the oscillating metal/bromide catalyzed oxyl radicals with hydroxyl and carbonyl alpha-substitu-
aerobic oxidation of benzaldehyde[14] in acetic acid.
ents, see Figure 4 b and c.[15] It could be that the deacti-
vation is much stronger for an alcohol hydrogen-bond-
ing to the carbonyl substitutent than to the hydroxyl sub-
stitutent, hence, in essence, stabilizing the benzaldehyde
when the benzyl alcohol is present. Another possibility
Evidence for Inhibition by the Benzylic Alcohols
The high yield to the benzaldehyde from the benzylic al- is that, during the co-oxidation of the alcohol and alde-
cohol is possible because: 1) the competing reaction to hyde, the termination step of the radical-chain mecha-
the acetate is slowand 2) the benzaldehyde does not sig-
nificantly react until nearly all the benzyl alcohol is ex-
hausted.
Example 2 on Table 3 gives average kinetic values
over five experiments. From example 2 one finds that
the competing reaction to the acetate (k3) is 21 times
slower than benzyl alcohol. [An independently meas-
ured value of the oxidation of benzyl acetate to benzal-
dehyde is also relatively slow, k4 , is 2.95 (0.11)Â10À5
sÀ1.]
The rate of oxidation of benzyl alcohol (k1) and that of
benzaldehyde (k2) are, within experimental error, the
same (example 2, Table 3). Assuming a set of elementa-
ry reactions where A!B ! C whose rates are the same
would predict a maximum yield to the benzaldehyde of
Figure 4. Possible hydrogen bonding Interactions during the
aerobic oxidation of benzylic alcohols with benzaldehydes:
(a) acetic acid with benzaldehyde; (b) alcohol with peroxy
radical of alcohol; (c) alcohol with peroxy radical of aldehyde
34% rather than the experimentally observed 84%.
Also the benzoic acid would immediately form and its
predicted increase would be at a much faster rate than
experimentally seen. Thus one has the conundrum that (R¼CH2C6H5).
564
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 559 – 568

Synthesis of Benzaldehydes from Benzylic Alcohols
FULL PAPERS
nism associated with benzylic alcohol is significantly This was measured because if it occurred during oxida-
higher than that of benzaldehyde. This explanation is tion of the benzylic alcohol so that 100% of it is lost,
the same as that given previously for the strong reduc- then the solution would experience a rapid formation
tion in reaction rate of cumene due to small amounts of Co(III), accounting for the green color. The bromine
of tetralin.[16]
.
loss was measured six times during complete benzylic al-
cohol oxidation at 758C and averaged 23.6% with a
standard deviation of 7.6%. Therefore, this cannot ac-
count for the color change. The bromine loss measured
during the oxidation of benzaldehyde, under the identi-
cal conditions, was 47%. It therefore appears that nearly
all the bromine loss in our benzylic alcohol reactions oc-
curs during the oxidation of benzaldehyde. Mechanisti-
cally, the greater loss of bromine during benzaldehyde
versus benzyl alcohol oxidation suggests that the rate
of bromine atom formation (Br·) is significantly faster
during the oxidation of benzaldehyde than that of ben-
zyl alcohol increasing the probability of reaction (12).
Samples were removed from the reactor before and
after the color change during a Co/Mn/Zr/Br-catalyzed
benzyl alcohol oxidation at 758C. The UV-VIS spectrum
before the color change occurred consists of three ab-
sorptions at 530, 640, and 670 nm. This is consistent
with the steady state oxidation states during reaction be-
ing Co(II), Mn(II), Zr(IV). The color is due to Co(II)
since Mn(II) and Zr(IV) are transparent in the visible
region. The crystal field analysis for cobalt(II) com-
pounds assigns the first absorbance to an octahedral
compound and the latter two to a tetrahedral species.[18]
A detailed analysis of the UV-VIS spectra of cobalt/bro-
mide mixtures in acetic acid is available.[19] Addition of
bromide to cobalt(II) perchlorate at a 1:1 Co/Br mol ra-
tio gives predominately monobromo octahedral and tet-
rahedral compounds:
Overoxidation to Carbon Dioxide and Carbon
Monoxide
The formation of carbon dioxide and carbon monoxide
is one primary characteristic of aerobic oxidations.
However, in these experiments their formation is not
important. The observed carbon dioxide increase is
only 0.01% above the normal content of carbon dioxide
in air. No carbon monoxide was observed hence no
measurable rate of decarbonylation expected from the
acyl radical is occurring. Since no significant oxygen up-
take occurs due to carbon oxide formation, the rates of
oxygen uptake reflect only the oxidation of the benzylic
alcohol, benzaldehyde, and benzyl acetate.
The Cause for the Color Change
The solution during the oxidation of the benzyl alcohols
changes from light blue to intense green soon after the
maximum in benzaldehyde concentration is observed.
The green color is associated with the oxidation of ben-
zaldehyde. A similar color change from pale blue to
green is also seen during the oxidation of benzaldehyde
alone, under identical conditions, when the conversion
of the latter is approximately 50%.
It is known that autoxidation mixtures using cobalt
catalysts without bromine are intense green in color
due to a significant amount (10–50%) of the cobalt in
solution existing as Co(III).[17] The colors of autoxida-
tions catalyzed by Co/Br are pale blue because the
steady state concentration of Co(III) is not detectable.
The pale blue color is that of Co(II) which has two
weak crystal field bands at 450 nm with a molar absorp-
tivity of around 15. A rapid reduction of Co(III) with
bromide occurs in Co/Br-catalyzed autoxidations and
is thought to be responsible for the steady state concen-
tration of cobalt being predominately Co(II) (ligands
omitted for simplicity):
ð13Þ
Since the perchlorate and acetate are comparatively
weak crystal field ligands, one can assume that the re-
ported molar absorptivities also apply to the measure-
ments reported here. The measured absorbance of the
solution taken during the benzyl alcohol oxidation is
consistent with 9% of the cobalt(II) existing as a
Co(II)Br tetrahedral species with the remaining being
octahedral Co(II). No Co(III) is present as evidenced
by the lack of a broad intense band near 350 nm.
When the solution turns to green, the observed absor-
bances can be assigned to 9% Co(II) tetrahedral species
and 22% Co(III) and the remainder to Co(II) octahe-
dral species.
ð11Þ
The bromide loss that occurs during these experiment
was measured. This presumably occurs via bromine for-
mation:
Currently we do not understand why Co(III) forms
during a Co/Mn/Zr/Br-catalyzed oxidation of benzalde-
ð12Þ
Adv. Synth. Catal. 2006, 348, 559 – 568
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
565

Walt Partenheimer
FULL PAPERS
hyde and is not present when methylbenzenes are simi-
larly oxidized. Co(III) formation does occur during the
metal/bromide-catalyzed autoxidation of methylben-
zenes when all of the ionic bromide forms benzylic bro-
mides (PhCH2Br), hence completely depleting the ionic
bromide. Similarly, one could propose the formation of
¼
ð14Þ
benzoyl bromide, PhCBr( O), during the oxidation of An internal redox reaction follows in which the metal
benzaldehyde. Benzoyl bromide has never been detect- oxidizes the bromide anion to the bromine atom, indi-
ed in the GC/MS studies, however. Benzylic bromides cated by Br(0) [oxidation states are indicated within
are quite stable during autoxidation but benzoyl bro- brackets ()]:
mide is expected to be hydrolytically unstable due to
the presence of water. So currently one cannot explain
why Co(III) formation occurs during benzaldehyde ox-
idations.
ð15Þ
A Mechanism Demonstrating why Stoichiometric and
Aerobic Oxidation of Benzylic Alcohols give Similar
Yields
The product ofreaction (15), aCo(II)-Br(O) species, is a
dimer hence it is a two-electron oxidant which oxidizes
The metal salts used in metal/bromide oxidations in ace- the benzyl alcohol in one step to the gem-diol:
tic acid have been shown to exhibit both redox and
Lewis acid chemistry.[20,21] We shall use the notation de-
veloped previously in which much of the chemistry oc-
curs in the coordination sphere of the metals.[20,21] Only
ligands necessary for the mechanism are placed in the
first coordination sphere of the metals.[20,21]
ð16Þ
The stoichiometric oxidation is given in reac-
tions (14)–(17). Since the bromide catalyzes the stoi-
chiometric synthesis,[10] bromide is first incorporated The water in the coordination sphere provides the oxy-
into the coordination sphere, reaction (14), and the re- gen atom for the gem-diol and the two protons neutral-
duced bromine species is subsequently used to oxidize ize the acetate in the coordination sphere to give two
the benzyl alcohol, reaction (16). A ligand replacement moles of acetic acid. The conversion of the gem-diol
reaction occurs as the bromide enters the coordination may be Lewis acid-catalyzed by Co(II) where formation
sphere of the metal with the protons of the hydrobromic of the cobalt(II)-H2O bond is the thermodynamic driv-
acid neutralizing two acetate ligands:
ing force:
Figure 5. Proposed mechanism for the aerobic oxidation of benzyl alcohol to benzaldehyde.
566
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 559 – 568

Synthesis of Benzaldehydes from Benzylic Alcohols
FULL PAPERS
metal-catalyzed steps is that the observed rate of carbon
dioxide formation (discussed above) is very lowsince a
high carbon dioxide formation rate accompanies non-
catalyzed benzaldehyde oxidations.[11] The Co/Mn/Zr/
ð17Þ
The aerobic oxidation is given in Figure 5. The free rad- Br catalyst is more active than without Zr, i.e., the Co/
ical chain mechanism is initiated in reaction (18) and Mn/Br catalyst.[20] This may be partially because
propagated in reactions (19) and (20). This generates Zr(IV) is a much stronger Lewis acid than Co(II) hence
the peroxide of the benzylic alcohol. Generation of the the dehydration of the gem-diol [reaction (26)] is greatly
benzaldehyde using non-catalyzed reactions proceeds enhanced. Evidence that Co(II)-Zr(IV) complexes exist
by thermal dissociation of the peroxide [reaction (21)], in acetic acid has been summarized.[20]. Reactions (24)
hydrogen atom abstraction [reaction (22)] and dehydra- and (26) may therefore occur in a concerted manner in
tion of the gem-diol [reaction (23)]. The metal-catalyzed the coordination sphere of the metals.[20]
steps are the initiation step [reaction (18)] and reac-
tions (24), (25), and (26). It will be assumed that the per-
oxide of the alcohol reacts with Co(II) in an identical
fashion as it does with peroxy acids – a well-character-
ized, very rapid and selective reaction.[11,22].
Conclusions
High yields of benzaldehydes are not expected from
benzylic alcohols in the non-catalyzed autoxidation
since the aldehydes are much more reactive than the al-
cohols. Consistent with past work, however, the metal/
bromide-catalyzed autoxidation opens up a number of
new pathways which supercede those of a free radical
chain mechanism. Indeed, the yields to benzaldehyde
ð27Þ observed in this work are comparable to those observed
in stoichiometric Co(III)/Br oxidation. Making an as-
sumption that the peroxide of benzyl alcohol reacting
The gem-diol is directly produced from the benzyl alco-
hol peroxide, reaction (24). But importantly, the reac-
with Co(II) is similar to peroxy acids, a mechanism is giv-
en explaining why stoichiometric and catalytic aerobic
tion produces as a by-product the same dimeric
oxidations give similar benzaldehyde yields. The unusu-
Co(II)-Br(O) oxidant species as in the stoichiometric re-
al biphasic behavior, observable in both the high yields
action – and this obviously will react with another mole
to the benzaldehyde and benzoic acid as well as the
rate of oxygen uptake, is difficult to rationalize without
the development of a highly sophisticated, complex ki-
netic model.
of benzyl alcohol to give the gem-diol [reaction (25)].
Thus two moles of the gem-diol are produced in the
same catalytic cycle and 50% ofthe gem-diol produced
is stoichiometric. [We do not suggest that the coordina-
tion compounds are identical in the stoichiometric and
catalytic aerobic oxidations but only that the same
Co(II)-Br(O) compound forms.] If we assume that reac-
tion (24) is similar to that of peroxy acids, then this reac-
tion will be much faster than the thermal dissociation of
the peroxide [reaction (21)] and the non-catalyzed reac-
tions become unimportant as has been previously dem-
onstrated.[11] Hence the metal-catalyzed steps will domi-
nate. The Co(II)-Br(O) compound produced can also be
used in the initiation step, reaction (18). But, since nor-
mally a large number of propagation steps occur for a
single initiation reaction, the Co(II)-Br(O) species will
be primarily used for the stoichiometric reaction. One
can conclude that only slightly less than 50% of the ox-
idation of benzyl alcohol is stoichiometric and that the
other 50% is a very rapid, selective reduction of the per-
oxide of benzyl alcohol. Therefore one has a mechanism
that would predict similar benzaldehyde yields when a
benzyl alcohol is oxidized stoichiometrically and well
Experimental Section
The glass autoclave, procedure, calculations, and GC and LC
instruments has been previously described.[6] The chemicals
were used as received. Initial cobalt and manganese catalysts
were their metal(II) acetate tetrahydrates and Zr(IV) as its
acetate. The reactants, intermediates, by-products, and prod-
ucts formed in the reactions were confirmed by GC/MS. Yields
were calculated from the GC data and occasionally checked by
LC. Unless otherwise stated, all reactions were performed with
a flowof air at 100 mL/min through the reactor with 100 g sol-
vent at ambient atmospheric pressure. Periodically liquid sam-
ples were removed during reaction and analyzed via GC. At
the same time, on-line gas analysis was performed for dioxy-
gen, dinitrogen, carbon monoxide and carbon dioxide. Select-
ed samples were analyzed for Co, Mn, Br via ICP (inductively
coupled plasma). UV-VIS spectra were obtained on an HP in-
as catalytically. Also consistent with the domination of strument with diode array detector.
Adv. Synth. Catal. 2006, 348, 559 – 568
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
567

Walt Partenheimer
FULL PAPERS
References
[7] N. G. Digurov, N. D. Gavrilenko, T. V. Bukharkina, E. S.
Enyukova, Kinetika i Kataliz 1978, 19, 136.
[8] N. A. Batygina, T. V. Bukharkina, N. G. Digurov, Nefte-
khimiya 1984, 24, 679 [abstract only].
[9] J. March, Advanced Organic Chemistry, McGrawHill,
NewYork, 2nd edn, 1977, p 253.
[1] J. I. Kroschwitz, Kirk Othmer Encyclopedia ofChemical
Technology, vol. 4, 4th edn., Wiley-Interscience Publica-
tions, NewYork, 1992, pp 64–72.
[2] Some recent examples of aerobic transformations of al-
cohols to carbonyl compounds are: a) S. Paavola, K .Zet-
terberg, T. Privalov, I. Csoregh, C. Moberg, Adv. Synth.
Catal. 2004, 346, 237; b) F. Minisci, F. Recupero, A. Cec-
chetto, C. Gambarotti, C. Punta, R. Faletti, R. Paganelli,
G. F. Pedulli, Eur. J. Org. Chem. 2004, 109; c) R. Liu, X.
Liang, C. Dong, X. Hu, J. Am. Chem. Soc. 2004, 126,
4112; d) I. E. Marko, A. Gautier, R. Dumeunier, K.
Doda, P. Kanae, F. Phillippart, S. M. Brown, C. J. Urch,
Angew. Chem. Int. Ed. 2004, 43, 1588; e) S. R. Reddy,
S. Das, T. Punniyamurthy, Tetrahedron Lett. 2004, 45,
3561; f) V. R. Choudhary, P. A. Chaudhari, V. S. Nar-
khede, Catal. Commun. 2003, 4, 171; g) R. Ciriminna,
S. Campestrini, M. Pagliaro, Adv. Synth. Catal. 2004,
346, 231; h) P. J. Figiel, J. M. Sobczak, J. J. Ziolkowski,
Chem. Commun. 2004, 244; i) C. Annunziatini, M. F.
Gerini, O. Lanzalunga, M. Lucarini, J. Org. Chem.
2004, 69, 3431; j) R. A. Sheldon, I. W. C. E. Arends, G.-
J. ten Brink, A. Dijksman, Acc. Chem. Res. 2002, 35,
774; k) J. Muzart, Tetrahedron, 2003, 59, 5789; l) S. E.
Martin, D. F. Suarez, Tetrahedron Lett. 2002, 43, 4475;
m) J. Muldoon, S. N. Brown, Org. Lett. 2002, 4, 1043;
n) M. Hasan, M. Musawir, P. N. Davey, I. V. Kozhevni-
kov, J. Mol. Catal. A: Chemical 2002, 180, 77; o) A. Wolf-
son, S. Wuyts, D. E. de Vos, I. F. J. Vankelecom, P. A. Ja-
cobs, Tetrahedron Lett. 2002, 43, 8107; p) R. Anderson,
K. Griffin, P. Johnston, P. I. Alsters, Adv. Synth Catal.
2003, 345, 517.
[10] M. Hirano, T. Morimoto, K. Itoh, Bull. Chem. Soc. Jpn.
1988, 61, 3749.
[11] W. Partenheimer, Adv. Synth. Catal, 2005, 347, 580–590.
[12] A. I. Popov, in: The Chemistry ofNonaqueous Solvents
III, (Ed.: J. J. Lagowski), Academic Press, New York,
1970, p 213.
[13] W. Partenheimer, Adv. Synth. Catal. 2004, 346, 1495–
1500.
[14] a) J. D. Druliner, E. Wasserman, J. Am. Chem. Soc. 1988,
110, 5270; b) M. G. Roelofs, E. Wasserman, J. H. Jensen,
A. E. Nader. J.Am.Chem.Soc. 1983, 105, 6329–6330;
c) M. G. Roelofs, E. Wasserman, J. H. Jensen, A. E. Nad-
er, J.Am.Chem.Soc. 1987, 109, 4207–4217.
[15] K. Ukegawa, Y. Kamiya, Bull. Chem. Soc. Jpn. 1976, 49,
1632.
[16] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations
ofOrganic Compounds , Academic Press, NewYork,
1981, p 27.
[17] W. Partenheimer, A Chemical Model for the Amoco
ꢀMCꢁ Oxygenation Process to Produce Terephthalic
acid, in: Catalysis ofOrganic Reactions , (Ed.: D. W.
Blackburn), Marcel Dekker, Inc., NewYork, 1990.
[18] A. B. P. Lever, Inorganic Electronic Spectroscopy, Elsev-
ier Publishing Company, NewYork, 1968, p 481.
[19] K. Sawada, M. Tanaka. J. Inorg. Nucl. Chem. 1977, 39,
339.
[20] W. Partenheimer, J. Mol. Catal. 2003, 206, 105–119.
[21] W. Partenheimer, J. Mol. Catal. 2003, 206, 131–144.
[22] W. Partenheimer, R. K. Gipe, Nature ofthe Co-Mn-Br
Catalyst in the Methylaromatic Compounds Process: Ki-
netic and Thermodynamic Studies, in: Catalytic Selective
Oxidation, (Eds.: S. T. Oyama, J. W. Hightower), Ameri-
can Chemical Society, Washington, 1993.
[3] W. Partenheimer, Catal Today 1995, 23, 69.
[4] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations
ofOrganic Compounds , Academic Press, NewYork,
1981, p 19.
[5] W. Partenheimer, V. V. Grushin, Adv. Synth. Catal. 2001,
343, 102–111.
[6] F. F. Shcherbina, N. P. Belous, Kinetika i Kataliz 1983, 24,
489.
568
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 559 – 568