Nitroarene catalyzed oxidation with sodium percarbonate or sodium perborate as the terminal oxidant

Article abstract of DOI:10.1016/j.tetlet.2004.09.156

A new catalytic oxidation method for the preparation of aromatic carboxylic acids from methyl aryl ketones is reported. The method is an alternative to the haloform reaction; it is benign and affords the desired product without production of any harmful side products. The catalytic cycle is based on the use of an electron-deficient nitroarene as catalyst with either of the two cheap and green oxidants sodium percarbonate or sodium perborate. The method gives a good yield (87%) and shows excellent selectivity when the model substrate (acetophenone) is oxidized. A series of benzoic acids of industrial interest were prepared by means of this method.

Full text of DOI:10.1016/j.tetlet.2004.09.156

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 8615–8620
Nitroarene catalyzed oxidation with sodium percarbonate or sodium
perborate as the terminal oxidant
*
Hans-Rene Bjørsvik, Jose Angel Vedia Merinero and Lucia Liguori
´
´
´
Department of Chemistry, University of Bergen, Allegaten 41, N-5007 Bergen, Norway
Received 18 August 2004; revised 20 September 2004; accepted 27 September 2004
Abstract—A new catalytic oxidation method for the preparation of aromatic carboxylic acids from methyl aryl ketones is reported.
The method is an alternative to the haloform reaction; it is benign and affords the desired product without production of any harm-
ful side products. The catalytic cycle is based on the use of an electron-deficient nitroarene as catalyst with either of the two cheap
and green oxidants sodium percarbonate or sodium perborate. The method gives a good yield (87%) and shows excellent selectivity
when the model substrate (acetophenone) is oxidized. A series of benzoic acids of industrial interest were prepared by means of this
method.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Electron-deficient nitroarenes can be utilized as stoichio-
metric oxidants1 or as catalysts with molecular oxygen
as terminal oxidant.2,3 Owing to several advantages
from both research and industrial standpoints, we
wanted to investigate the possibility of replacing oxygen
with sodium perborate, NaBO3Æ4H2O (SPB) or sodium
percarbonate, Na2CO3Æ1.5H2O2 (SPC) as the terminal
oxidant. SPB and SPC are both cheap, large-scale indus-
trial chemicals that are nontoxic and relatively stable
solids and are thus simple to handle on a large scale
and permit accurate dosage. A few facets of the very
complex nitroarene oxidation are outlined in Scheme
1. The nitro group was expected to possess a twofold
role as the oxidizing species in the process, namely (1)
as the electron acceptor in the single electron transfer
process and (2) to act as the oxygen transfer reagent.
dinitrobenzene 3 is regenerated from the reduction
products 1-nitro-3-nitrosobenzene 7 and N-(3-nitro-
phenyl)-hydroxylamine 8 by treatment with base and
by oxidation with SPB or SPC.4
Russell et al.5–7 demonstrated that when a mixture of
nitrosobenzene and N-hydroxyaniline is treated with
base, two types of nitrosobenzene radical anions are
formed, a reaction which corresponds to 7+8 ! 9. It is
likely that both 7 and 8 may be formed during the
oxidation process, since nitroarenes can be reduced over
several steps (see Eq. 1).
ð1Þ
The enolate ion 2 (Scheme 1) is produced when aceto-
phenone 1 is treated with a base. In a single electron
transfer (SET) process the C-centred radical 5 and the
radical anion 4 are formed, and then combined to form
the labile anion 6. Decomposition of 6 affords the par-
tially oxidized intermediate anion 10 that is oxidized
further to yield the target product 14 via a similar
mechanism to that already described. The catalyst 1,3-
The investigations by Russell and co-workers revealed
moreover that if the reaction is performed in the pres-
ence of molecular oxygen, the radical anions are oxi-
dized to yield two nitrobenzene radical anions that
upon reaction with the nitroso group (e.g., compound
7) afford the dinitrobenzene 3. Here, the oxidation of
the radical anion species 9 proceeds via the operation
of either sodium perborate or sodium percarbonate.
Initially, the proposed oxidation process, Scheme 2, was
investigated by utilizing a small series of nitroarenes
as catalysts: 1,8-dinitronaphthalene 15, 2,4-difluoro-
nitrobenzene 16, 1,2,3-trifluoro-4-nitrobenzene 17,
Keywords: Electron transfer; Green chemistry; Oxidations; Radicals;
Radical reactions; Industrial chemistry.
*
Corresponding author. Tel.: +47 55 58 34 52; fax: +47 55 58 94 90;
e-mail: hans.bjorsvik@kj.uib.no
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.09.156

8616
H.-R. Bjørsvik et al. / Tetrahedron Letters 45 (2004) 8615–8620
HO
NH
NO2
8
O
O
N
N
NO2
H2O2
Base
NO2
O
NO2
NO2
O2N
3
9
7
O
N
H2C
O
O
O
O
CH3
H2C
O
H2C
O
NO2
Base
O
SET
N
O
1
2
4
6
5
H H
O
O
OH
O
O
O
C6H4NO2
O
O
N
O
H
SET
H2C
O
O
O
14
13
12
11
10
N
NO2
NO2
N
7
HO
O
4
O
NH
N
NO2
H2O2
Base
NO2
NO2
O2N
8
3
9
Scheme 1.
O
O
tion temperature as has been previously determined to
be beneficial for the satisfactory operation of nitroarene
oxidation.2,3
Catalyst
Solvent, PTC, Base
[ O ], 55-80 oC, 5 h
CH3
OH
1
14
All of the investigated nitroarenes (3 and 15–18) operated
as catalysts for the oxidation process. The major part of
the investigation was, however, carried out with 1,3-dini-
trobenzene 3 as catalyst, mainly for comparison purposes
(with the aerobic oxidation process3), but also due to the
low cost of the substance that will be highly advantageous
for any eventual industrial application. Furthermore,
most of the oxidation experiments were performed with
SPC, since this appeared to give slightly higher yields
than for experiments utilizing SPB as the oxidant.
Solvent: t-BuOH, EtOH, BTF.
Base: t-BuOK, EtONa, KOH
[ O ] : NaBO3 4H2O or Na2CO3 1.5H2O2
Catalysts:
NO2 NO2
NO2
NO2
F
NO2
3
15
16
F
NO2
O
F
F
NO2
O2N
Solvent and base screening were performed, which dem-
onstrated that the solvents a,a,a,-trifluorotoluene (BTF)
and t-BuOH and the bases NaOH, KOH and t-BuOK
can be utilized (see Table 1). Furthermore, the use of
the phase transfer catalyst tetrabutylammonium bro-
mide (TBAB) and the addition of the terminal oxidant
in several portions were investigated.
18
17
F
Scheme 2.
2,7-dinitro-9-fluorenone 18 and 1,3-dinitrobenzene 3.
The investigation comprised two reaction temperature
ranges: 55–60ꢁC was used when SPB was present as
the terminal oxidant, and 80ꢁC was used when SPC
was utilized. This avoided potential decomposition of
SPB and SPC while maintaining a slightly elevated reac-
The rather large variations and moderate yield of the
screening experiments (a few results are reported in Table
1), prompted us to continue the investigation by means
of statistical experimental design and multivariate mod-

H.-R. Bjørsvik et al. / Tetrahedron Letters 45 (2004) 8615–8620
Table 1. Oxidation experiments with SPB or SPC and different nitroarenes as catalyst
8617
#
Experimental conditionsa
Resultsb
mmol T (ꢁC) Conv. (%) Yield (%) Selec. (%)
SPC 3.00 80 58 6 2 45
Subst. (mmol) Cat. TBAB (mmol) Solvent mL Base
mmol [O]
1c 2.4
.422
15
15
16
17
3
0.24
0.24
0.24
0.24
0.12BTF
—
t-BuOH 12KOH
8.0
BTF
BTF
BTF
12
12
12
KOH
KOH
KOH
5.0
5.0
5.0
SPB 2.40
SPB 2.40
SPB 2.40
SPC 1.20
SPC 2.40
SPC 6.87
SPB 6.87
SPC 1.20
SPB 1.20
SPB 2.40
SPB 2.40
55
55
55
80
80
80
60
80
55
55
55
16
16
16
57
37
96
94
70
32
72
54
15
10
14
31
28
18
31
28
26
38
29
92
59
89
54
76
19
34
39
82
52
54
3
2
.4
.4
4
5
6
2
1.2
2
6
t-BuOK 2.4
t-BuOK 4.8
t-BuOK 6.0
t-BuOK 6.0
.4
3
BTF
12
7d 2.4
8d 2.4
18
18
3
—
—
t-BuOH 12
t-BuOH 12
9
1.2
0.12
0.12
0.24
0.24
BTF
BTF
6
6
NaOH
KOH
2.5
2.5
10 1.2
11 2.4
1.242
3
3
t-BuOH 12
t-BuOH 12
t-BuOK 4.8
NaOH 5.0
3
a Procedure: Acetophenone was used as substrate (Subst.) with 10mol% of a nitroarene as catalyst. The nitroarenes were 1,3-dintrobenzene (DNB),
1,8-dinitronaphthalene (DNN), 2,4-difluoronitrobenzene (DFNB) or 1,2,3-trifluoro-4-nitrobenzene (TFNB). TBAB was used as phase transfer
catalyst. The terminal oxidant was either SPC or SPB. The reaction time was 5h performed at 55–80ꢁC under a nitrogen atmosphere.
b Conversion, yield and selectivity are based on isolated product and recovered substrate. Isolated products were corrected for impurities by GC.
Moreover, isolated products were identified by GC–MS and NMR.
c The oxidant was added in two equal portions, 1.2mmol, at the start and after 2h.
d The oxidant was added in three portions, 3.12mmol at the start, then 2.4mmol after 1h and 1.35mmol after 2h.
elling. For this investigation, t-BuOH was selected as sol-
vent and t-BuOK as base. Furthermore, we decided not
to use the phase transfer catalyst as no significant
improvements were observed with TBAB as an additive.
The oxidant SPC was added portion-wise, at the begin-
ning of the reaction and then after 1 and 2h of reaction
time, respectively. The necessity for the somewhat large
excess of oxidant may be due to partial thermal decom-
position during heating of the reaction mixture.
the quantity of t-BuOK [mmol] (x1) and the quantity of
SPC [mmol] (x2).
The data from Table 2 was submitted to multivariate
data analysis by means of multiple linear regression
(MLR)9–11 to establish an empirical mathematical pre-
dictive model, y = f(x1,x2). In order to obtain a better
fit of the response y to the model matrix [1 x1 x2
x1 · x2 x21 x22], the transformation of the response yield
(y) was performed (Eq. 2). The estimated regression
coefficients for the final model12 with the yield as the
response variable are given in Eq. 3.
Even though numerous experimental variables may
influence the oxidation process, we have only studied
two of the experimental variables, thoroughly, in terms
of a statistical experimental design Table 2,8 namely
pffiffiffi
g ¼ y ¼ f ðx1; x2Þ
ð2Þ
Table 2. Statistical experimental design exploring the influence of the base (t-BuOK) and the oxidant (SPC)
Experimental conditionsa
Measured responsesb
#
x1 (Base) x2 (SPC)
% 1
% 2
C%
S%
1
27.50
4.50
3.75
27
42
73
58
3.75
5
57
95
60
94
9294
62
3
4.50
7.50
6.00
3.00
9.00
6.00
6.00
6.00
6.00
5.00
5.00
4.38
4.38
4.38
3.129
3.1216
5.63
5.00
35
8
6265
86
4
5
1255
55
5
88
6
36
76
46
95
79
79
7
8
87
83
91
83
96
9
10
100
89
11
15
68
70
77
83
11
85
126.00
13
14
3.75
10
71
90
79
4.50
7.50
4.38
4.38
21
6
55
75
79
94
70
79
a Reaction conditions: x1: t-BuOK (4.5–7.5mmol), x2: Na2CO3Æ1.5H2O2 at the start of the reaction (3.75–5.00mmol) were mixed in t-BuOH (12mL)
under a N2 atmosphere. Following the addition of acetophenone (2.4mmol, 0.28mL) and 1,3-dinitrobenzene (0.24mmol, 40.3mg) as catalyst, the
resulting solution was heated for 5h at 80ꢁC. After 1 and 2h reaction time, additional quantities of the oxidant Na2CO3Æ1.5H2O2 were added
(2.5mmol after 1h and 1.5mmol after 2h).
b Samples were analyzed by GC and the mol% estimated by means of 1-chloro-2,4-dinitrobenzene as internal standard. C% = percent conversion,
and S% = percent selectivity.

8618
H.-R. Bjørsvik et al. / Tetrahedron Letters 45 (2004) 8615–8620
g ¼ f ðx1; x2Þ ¼ 7:862 þ 0:6526 ꢀ x1
NO2
ꢁ 0:1296 ꢀ x21 þ 0:2545 ꢀ x22
ð3Þ
O
CH3
R1
(10 mol%)
NO2
O
OH
R1
y ¼ g2
tert-BuOH (12 mL)
R2
R4
R4
R2
tert-BuOK
Na2CO3 1.5H2O2
R3
8a-13a 2.4 mmol
R3
8b-13b
The model (Eq. 3) was used to create the response sur-
face projected as an iso-contour map depicted in Figure
1. This map reveals that it is mandatory to apply a con-
comitant high quantity of the base t-BuOK and the oxi-
dant SPC, or low quantities of oxidant but still with a
high quantity of base. However, additional experiments
under such conditions afforded no further improvement
compared to the best results already included in the
experimental design matrix portrayed in Table 2 (entries
8 and 9).
80 oC, 5 h
R1
H
H
H
F
R2
OCH3 OCH3
R3
R4
H
H
H
H
H
F
Conv.% Yield% Selec.%
82
64
50
58
38
23
73
35
61
91
42
23
74
34
8
9
10
11
12
13
H
H
H
H
H
OCH3
F
F
Cl
Cl
91
>99
98
H
Cl
98
Scheme 3.
Application of the novel oxidation process. Functional-
ized aromatic carboxylic acids are highly valuable fine
chemicals having a variety of applications. For example,
veratric acid 8b is used in the synthesis of the pharma-
ceuticals mebeverine and vesnarinone.13 Various halo-
genated benzoic acids 10b–13b are used in the
synthesis of antibacterial agents,14 cholesterol biosyn-
thesis inhibitors15 and herbicides.16
benzophenones17 using among others both sodium per-
carbonate and sodium perborate.
As an alternative to the above method17 we investigated
the potential of our new catalytic oxidation method with
a series of substituted fluoroacetophenones (see Scheme
3). The method operates with high levels of conversion
when utilizing reaction conditions developed for aceto-
phenone, however, only moderate yields and selectivity
were obtained, except for the experiment where 4-chlo-
roacetophenone 12a was oxidized to 4-chlorobenzoic
acid 12b in 73% yield and 74% selectivity.
Thus, industrial interest in the production of such pre-
cursors has resulted in disclosures of various processes.
For example, Hoechst have described a process for the
preparation of halogenated benzoic acids by means of
Baeyer–Villiger oxidation of appropriately halogenated
We have previously disclosed an optimized industrial
process for the preparation of veratric acid 8b, a process
that was based on the oxidation of 3,4-dimethoxyaceto-
phenone 8a by means of the haloform reaction.18 In
view of the advantages of using a cheap solid and non-
toxic substance as oxidant for fine chemical preparation,
we have also tried to prepare veratric acid 8b by
means of the novel oxidation process using SPC. The
conversion 8a!8b, and 4-methoxyacetophenone 9a to
4-methoxybenzoic acid 9b demonstrated comparable
results to our earlier disclosed aerobic oxidation
process.3
The oxidation process optimized for the utilization of
SPC was also attempted with SPB as the terminal oxi-
dant. Since SPB decomposes at a reaction temperature
above 60ꢁC, an extra trial was performed at this tempera-
ture. All the other experimental variables were kept at
similar levels as for the optimized conditions (Table 2,
entries 8 and 9).
Figure 1. The iso-contour plot of the response surface shows the
predicted yield as a function of two experimental variables, the
quantity of the base t-BuOK (x1), and the quantity of the oxidant
sodium percarbonate (x2). In the present plot, the contour lines show
the yield of benzoic acid as a function of the two experimental
variables x1 and x2 in the ranges 3.0–9.0mmol and 3.12–5.63mmol,
respectively. The reactions were conducted by mixing the base t-BuOK
(3.0–9.0mmol) and the oxidant Na2CO3Æ1.5H2O2 (3.12–5.63mmol) in
t-BuOH (12mL) under a nitrogen atmosphere followed by adding
acetophenone (2.4mmol, 0.28mL) and 1,3-dinitrobenzene (0.24mmol,
40.3mg) as catalyst. The resulting solution was heated for 5h at 80ꢁC.
After 1 and 2h reaction time, additional quantities of the oxidant
Na2CO3Æ1.5H2O2 were added (2.5mmol after 1h and 1.5mmol after
2h). The black filled circles are the experiments included in the
underlying model with their corresponding experimental values.
Inferior results were achieved with SPB compared to the
results obtained with SPC. The experiment performed at
60ꢁC resulted in a yield of 35% with a selectivity of 69%,
while the experiment performed at 80ꢁC, resulted in
28% yield with a selectivity of 52%. Nevertheless, the
experiments show that SPB may indeed be utilized as
the terminal oxidant with 1,3-dinitrobenzene as catalyst.
An improved yield and selectivity with SPB as the termi-
nal oxidant may be accomplished if the process is

H.-R. Bjørsvik et al. / Tetrahedron Letters 45 (2004) 8615–8620
8619
submitted for a detailed study using statistical experi-
mental multivariate modelling.
fully acknowledged. Professor George Francis is
acknowledged for linguistic assistance.
Conclusion. In summary, the oxidation method pre-
sented here, requires basic conditions and SPC as the
terminal oxidant. SPB can be utilized as the terminal
oxidant, but when operated under the conditions optim-
ized for SPC, inferior results were obtained. However,
we believe that comparable results may be accom-
plished if an optimizing study is performed for this oxi-
dant as well. The oxidation method operates with the
best selectivity at moderately elevated temperature
(75–80ꢁC).
References and notes
1. Bjørsvik, H.-R.; Liguori, L.; Minisci, F. Org. Process Res.
Dev. 2001, 5 136–140.
´
´
2. Bjørsvik, H.-R.; Liguori, L.; Rodrıguez Gonzalez, R.;
Vedia Merinero, J. A. Tetrahedron Lett. 2002, 43, 4985–
4987.
3. Bjørsvik, H.-R.; Liguori, L.; Vedia Merinero, J. A. J. Org.
Chem. 2002, 67, 7493–7500.
4. (a) McKillop, A.; Tarbin, J. A. Tetrahedron 1987,
43, 1753–1758; (b) McKillop, A.; Kemp, D. Tetra-
hedron 1989, 45, 3299–3306; (c) McKillop, A.; Sander-
son, W. R. Tetrahedron 1995, 51, 6145–6166; (d) Muzart,
J. Synthesis 1995, 1325–1347; (e) McKillop, A.; Sander-
son, W. R. J. Chem. Soc., Perkin Trans. 1 2000, 471–476.
5. Russell, G. A.; Geels, E. J.; Smentowski, F. J.; Chang,
K.-Y.; Reynolds, J.; Kaupp, G. J. Am. Chem. Soc. 1967,
89, 3821–3828.
It is reasonable to believe that the oxidation operates
according to a mechanism similar to that previously dis-
closed by us.2 In the present case, however, the reduc-
tion products from the nitroarene are oxidized by
means of SPC or SPB.
In contrast to the aerobic oxidation method2,3 the pre-
sent method is not specific for benzaldehydes and benz-
alcohols, since such functional groups are oxidized by
both SPB and SPC. The investigations performed for
the preparation of benzoic acids 8b–13b revealed vary-
ing results with respect to the yield, but the method
may be further improved for each of the substrates
8a–13a as well as for other acetophenone derivatives in
order to perform processes which are of industrial
interest.
6. Russell, G. A.; Geels, E. J. J. Am. Chem. Soc. 1965, 87,
122–123.
7. Geels, E. J.; Konaka, R.; Russell, G. A. Chem. Commun.
1965, 23, 13.
8. The statistical experimental design outlined in Table 2 is a
composite design that allows the estimation of the main
effects, the interaction effects and the quadratic effects for
each of the experimental variables.
9. Draper, N. R.; Smith, H. Applied Regression Analysis, 3rd
ed.; Wiley: New York, 1998.
10. Montgomery, D. C.; Peck, E. A. Introduction to Linear
Regression Analysis; Wiley: New York, 1982.
11. (a) The multivariate calculations and graphical represen-
tations were performed by means of in-house developed
General oxidation procedure. Starting materials and rea-
gents were purchased commercially and used without
further purification. Acetophenone or derivative
(2.4mmol) was dissolved in a solution of potassium
tert-butoxide (6.0mmol) in tert-butanol (12.0mL). The
oxidant sodium percarbonate (3.12mmol) was added
at the start of the reaction followed by the catalyst
1,3-dinitrobenzene (0.24mmol). A further two portions
of SPC were added after 1h (2.5mmol) and after 2h
(1.5mmol). The reaction mixture was stirred using a
magnetic stirrer bar and heated at 80ꢁC for a total reac-
tion time of 5h under a nitrogen atmosphere in order to
avoid any interaction of molecular oxygen. The reaction
was quenched by dilution with water (30mL) and
extracted with ethyl acetate (3 · 30mL). The aqueous
phase was acidified using concd sulfuric acid to pH1–
2, then extracted with ethyl acetate (3 · 30mL). After
diluting with further ethyl acetate to 250mL, a sample
(10mL) was mixed with internal standard (1-chloro-
2,4-dinitrobenzene) solution (10mL) and analyzed by GC.
ꢂ
procedures for MATLAB version 6.1, which previously
have been validated by comparison of the calculated
results with computational results obtained from several
commercial computer programs for statistics and mathe-
matical model building. (b) Using Matlab, Version 6, The
MathWorks Inc., Natick, MA, USA. (c) Using Matlab
Graphics Version 6, The MathWorks Inc., Natick, MA,
USA.
12. The product statistic of the model indicates an acceptable
predictive capacity; R2 = 0.711, Q2 = 0.547, RMSEP =
0.521 and RSD = 0.562. Statistical analyses indicated that
two of the experiments, entries 1–2of Table 2 were both
determined somewhat aberrantly, that is with high resi-
duals (ymeasuredꢁypredicted). If these were left out before the
estimation of the regression coefficients, the product
statistics indicated a slightly improved model, namely
R2 = 0.813, Q2 = 0.622, RMSEP = 0.431 and RSD =
0.469. However, only minor variations of coefficients
values were observed.
13. Bjørsvik, H.-R.; Liguori, L. Org. Process Res. Dev. 2002,
6, 279–290.
All of the reaction products are known compounds. Iso-
lated products were analyzed by GC19 for impurities and
14. Culbertson, T. P.; Mich, T. F.; Domagala, J. M.; Nichols,
J. B. (Warner-Lambert Co., USA) Patent 14 EP 153163
Chem. Abstr. 1986, 104, 34013.
15. Granzer, E.; Hammann, P.; Kirsch, R. (Hoechst A.-G.,
Germany). Patent EP 431480, Chem. Abstr. 1991, 115,
232218.
1
samples were analyzed by GC–MS, H NMR,20 and by
comparison with authentic samples.
16. Wheeler, T. N.; Craig, T. A.; Morland, R. B.; Ray, J. A.
Synthesis 1987, 10, 883–887.
Acknowledgements
17. Pfirmann, R. D. (Hoechst A.-G., Germany) Patent EP
600317, Chem. Abstr. 1994, 121, 133701.
18. Bjørsvik, H.-R.; Norman, K. Org. Process Res. Dev. 1999,
3, 341–346.
Economic support from the Research Council of
Norway (L.L.), Optimum Accipe and the Norwegian–
Spanish cultural exchange program (J.A.V.M.) is grate-

8620
H.-R. Bjørsvik et al. / Tetrahedron Letters 45 (2004) 8615–8620
19. GLC analyses were performed on
a
capillary gas
20. 4-Methoxybenzoic acid [100-09-4], 3,4-dimethoxyben-
zoic acid [93-07-2], 4-fluorobenzoic acid [456-22-4],
2,4-difluorobenzoic acid [1583-58-0], 4-chlorobenzoic
acid [74-11-3], 2,4-dichloro-5-fluorobenzoic acid [86522-
89-6].
chromatograph equipped with a fused silica column (L
25m, 0.20mm i.d., 0.33m film thickness) at a helium
pressure of 200kPa, split less/split injector and flame
ionization detector.