Solvent and temperature effects in the free radical aerobic oxidation of alkyl and acyl aromatics catalysed by transition metal salts and N-hydroxyphthalimide: New processes for the synthesis of p-hydroxybenzoic acid, diphenols, and dienes for liquid crystals and cross-linked polymers

Article abstract of DOI:10.1021/op034137w

The aerobic oxidation of 4,4′-diisopropyldiphenyl and 2,6-diisopropylnaphthalene, catalysed by N-hydroxyphthalimide and Co(II) salts, leads to the corresponding tertiary benzyl alcohols with high conversion and selectivity under mild conditions (temperature 30-60°C and atmospheric pressure). Solvent and temperature effects, as resulting from the pioneering work of C. Walling, and more recently from the conclusive resolution of K. U. Ingold and co-workers on a quantitative kinetic basis, strongly affect the selectivity of the aerobic oxidation. This is related to the ratio between the rate of β-scission of the alkoxyl radical, which leads to acetophenone derivatives, and the rate of hydrogen atom abstraction, leading to tertiary benzyl alcohols. These latter are efficiently converted either to diphenols for the production of liquid crystals, by reaction with H₂O₂, or to dienes, useful as cross-linking agents, by dehydration. The aerobic oxidation of p-hydroxyacetophenone catalysed by Mn(NO₃)₂ and Co(NO₃)₂ leads with high selectivity to p-hydroxybenzoic acid, a useful monomer for liquid crystals.

Full text of DOI:10.1021/op034137w

Organic Process Research & Development 2004, 8, 163−168
Solvent and Temperature Effects in the Free Radical Aerobic Oxidation of Alkyl
and Acyl Aromatics Catalysed by Transition Metal Salts and
N-Hydroxyphthalimide: New Processes for the Synthesis of p-Hydroxybenzoic
Acid, Diphenols, and Dienes for Liquid Crystals and Cross-Linked Polymers
Francesco Minisci,* Francesco Recupero, Andrea Cecchetto, Cristian Gambarotti, Carlo Punta, and Roberto Paganelli
Dipartimento di Chimica, Materiali e Ingegneria Chimica G. Natta, Politecnico di Milano, Via Mancinelli 7,
20131 Milano, Italy
Gian Franco Pedulli
Dipartimento di Chimica Organica A.Mangini, UniVersit a` degli Studi di Bologna, Via S. Donato 15, 40127 Bologna, Italy
Francesca Fontana
Dipartimento di Ingegneria Industriale, UniVersit a` di Bergamo, Viale Marconi 5, 24044 Dalmine BG, Italy
Abstract:
peroxides, whose separation is more difficult. Thus, the
benzene sulfonation process is still competitive with the
peroxidation process, and the world’s major producers of
resorcinol utilise both the alkaline fusion of m-benzenedis-
ulfonic acid (INDSPEC Chemical Corp. in the U.S.A.) and
the peroxidation of m-diisopropylbenzene (Sumitomo Chemi-
cal Company and Mitsui Petrochemical Industries in Japan).
The benzene sulfonation process forms 2.47 kg of sodium
sulfite per kg of resorcinol, whereas the peroxidation process
requires 2.08 kg of m-diisopropylbenzene per kg of resor-
cinol. The cost for the disposal of the byproduct sodium
sulfite is a major component of the total cost of the
sulfonation process, while the price of the m-diisopropyl-
benzene feedstock is a major component of the total cost of
the peroxidation process. A change in the price of either
material can, therefore, alter the cost ratio of the two
The aerobic oxidation of 4,4′-diisopropyldiphenyl and 2,6-
diisopropylnaphthalene, catalysed by N-hydroxyphthalimide
and Co(II) salts, leads to the corresponding tertiary benzyl
alcohols with high conversion and selectivity under mild
conditions (temperature 30-60 °C and atmospheric pressure).
Solvent and temperature effects, as resulting from the pioneer-
ing work of C. Walling, and more recently from the conclusive
resolution of K. U. Ingold and co-workers on a quantitative
kinetic basis, strongly affect the selectivity of the aerobic
oxidation. This is related to the ratio between the rate of
â-scission of the alkoxyl radical, which leads to acetophenone
derivatives, and the rate of hydrogen atom abstraction, leading
to tertiary benzyl alcohols. These latter are efficiently converted
either to diphenols for the production of liquid crystals, by
reaction with H
by dehydration. The aerobic oxidation of p-hydroxyacetophe-
none catalysed by Mn(NO and Co(NO leads with high
2 2
O , or to dienes, useful as cross-linking agents,
1
processes.
2
3
)
2
3
)
2
Recently, attempts to improve the industrial peroxidation
selectivity to p-hydroxybenzoic acid, a useful monomer for
liquid crystals.
of m-diisopropylbenzene by catalysis of N-hydroxyphthal-
imide (NHPI) under milder conditions did not prove suc-
cessful; a yield of 36% resorcinol and 17% m-isopropylphe-
nol was reported by decomposing the hydroperoxides with
The production of phenol and acetone from cumene
through the hydroperoxide is one of the main industrial
chemical processes, which has replaced traditional processes,
such as the alkaline fusion of benzenesulfonic acid.
Analogous processes have been developed for the produc-
tion of diphenols, such as resorcinol from m-diisopropyl-
benzene. However, compared to the phenol process, addi-
tional complexities are encountered in the making of
resorcinol. In the peroxidation of cumene a higher selectivity
is achieved by keeping conversion low, since the separation
of cumyl hydroperoxide from cumene is a relatively simple
process. On the other hand, a partial conversion of m-
diisopropylbenzene yields a mixture of mono- and dihydro-
2 4 3
H SO whereas, with the Lewis acid InCl as catalyst, yields
were 56% resorcinol and 16% m-isopropylphenol; in any
case a mixture of mono- and dihydroperoxides was formed.
With more complex diisopropylaromatics, such as diisopro-
pyldiphenyls and diisopropylnaphthalenes, these complexities
of the peroxidation process further increase.
3
Recently we have determined the absolute rate constants
for the hydrogen atom abstraction from benzylic C-H bonds
(
1) Chemical Economics Handbook; SRI International: Menlo Park, CA, (a)
June 1997; 691.700A, p 1. (b) May 1999; 580.420C, p 3. (c) February 2001;
695.4021U, p 47.
(
(
2) Fukuda, O.; Sakaguchi, S.; Ishii, Y. AdV. Synth. Catal. 2001, 343, 809.
3) Amorati, R.; Lucarini, M.; Mugnaini, V.; Pedulli, G. F.; Minisci, F.;
Recupero, F.; Fontana, F.; Astolfi, P.; Greci, L. J. Org. Chem. 2003, 68,
1747.
*
Author for correspondence. E-mail: francesco.minisci@polimi.it.
1
0.1021/op034137w CCC: $27.50 © 2004 American Chemical Society
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
163
Published on Web 01/20/2004

by phthalimido-N-oxyl (PINO) radical (eq 1)
Equation 5 can compete with hydrogen atom abstraction
from the O-H bond of NHPI (eq 6) or from C-H bonds
(eq 7).
The pioneering results of C. Walling and co-workers6,7
have shown that solvents could have a relevant effect on
the magnitude of the ratio between â-scission of the alkoxyl
and for the formation of PINO radical by hydrogen atom
abstraction from NHPI by peroxyl radicals (eqs 2, 3), thus
providing a quantitative basis to the study of aerobic
oxidation catalysed by NHPI.
radical and hydrogen atom abstraction from C-H bonds (k
5
/
k
7
in the case of cumyloxyl radical). Temperature also
influences this ratio since it affects unimolecular (eq 5) more
than bimolecular (eq 7) reactions; an increase in temperature
therefore favours â-scission, as shown in the Walling report.
8
Several years later K. U. Ingold and co-workers provided
conclusive resolution of the long-standing problem concern-
ing the solvent effects on the competitive â-scission and
hydrogen atom abstraction reactions from C-H bonds of
cumyloxyl radical by determining the respective rate con-
stants by laser flash photolysis. The results have clearly
demonstrated that hydrogen atom abstraction from C-H
bonds (eq 7) is solvent independent (i.e. for R-H )
These results are also related to the bond dissociation
enthalpy (BDE) of the OH bond in NHPI, which we have
recently determined (88.1 kcal mol ).
6
-1 -1
cyclohexane in eq 7 a rate constant of 1.2 × 10 M
s
at
3
-1
3
0 °C was determined in a variety of solvents of different
Considering the difficulties for the selective syntheses of
hydroperoxides in the biphenyl and naphthalene series from
diisopropyl derivatives as well as the industrial interest of
the corresponding diphenols, we have developed a different
strategy, based on the selective aerobic oxidation of these
diisopropyl derivatives to the corresponding tertiary benzyl
alcohols through NHPI and Co(II) salts catalysis (eq 4).
5
polarity), whereas the rate of â-scission, k , increases with
an increase of solvent “polarity”; this gives a quantitative
basis to the Walling suggestion that a decrease of the k /k
ratio with increasing solvent polarity is due to an increase
in k and that any changes in k are of only very minor
significance. The values for k at 30 °C were 2.6, 3.7, 5.5,
.8, 6.3, and 19 × 10 s in CCl , C , C Cl, Me C-
OH, MeCN, and MeCOOH, respectively. The increase in k
7
5
5
7
5
5
-1
5
4
6
H
6
H
6 5
3
5
NHPI
Me CH-Ar-CHMe + O
8 Me C(OH)-Ar-C(OH)Me2
with increasing solvent polarity is likely due to increased
stabilisation of the transition state for â-scission via increased
solvation of the incipient acetophenone product.8
2
2
2
2
Co(II)
(4)
A more pronounced solvent effect has been reported more
The aerobic oxidation of cumene, catalysed by NHPI, was
9
,10
4
,5
recently
radical from tert-butylhydroperoxide (eq 8) and phenol (eq
).
for hydrogen atom abstraction by cumyloxyl
reported in two further papers by Ishii and co-workers, in
2
addition to the above-mentioned peroxidation, and in both
9
cases acetophenone was the main reaction product. In acetic
acid solution, aerobic oxidation at 100 °C under NHPI and
k8
•
•
Co(acac)
2
catalysis provided 54% acetophenone, 10% cumyl
t-BuOOH + PhC(Me) -O
98 t-BuOO + PhC(Me) -OH (8)
2
2
alcohol, and 17% phenol; however, the conversion was low
due to inhibition by the phenol, formed during the reaction
by acid-catalyzed decomposition of the hydroperoxide. The
oxidation led to 64% acetophenone and 8% cumyl alcohol
in PhCN solution, with NHPI catalysis and in the absence
of metal salts at 100 °C.
It is well-known that the formation of acetophenone in
the oxidation of cumene arises from â-scission of the
cumyloxyl radical (eq 5).
k9
•
•
Ph-OH + PhC(Me) -O
98 Ph-O + PhC(Me) -OH (9)
2
2
spans from 2.5 × 10 M s-1 in CCl
to 6.7 ×
8
-1
Here k
8
4
6
-1 -1
10 M
s
in Me
(8.6 × 10 M
3
C-OH and the range is still larger in the
8
-1 -1
6
-1
case of k
9
s
in CCl
4
and 3.6 × 10 M
-
1
s
in Me C-OH). The absence of a solvent effect on
hydrogen atom abstraction from C-H bonds by cumyloxyl
radical necessarily implies that this dramatic change in
3
k5
(6) Walling, C.; Padwa, A. J. Am. Chem. Soc. 1963, 85, 1593.
(7) Walling, C.; Wagner, P. I. J. Am. Chem. Soc. 1964, 86, 3368.
•
•
PhCMe -O
98 Ph-COMe + Me
(5)
2
(8) Avila, D. V.; Brown, C. E.; Ingold, K. U.; Lusztyk, J. J. Am. Chem. Soc.
1993, 115, 466.
(
(
4) Ishii, Y.; Sakaguchi, S.; Iwahama, T. AdV. Synth. Catal. 2001, 343, 393.
5) Ishii, Y.; Nakayama, K.; Takeno, M.; Sakaguchi, S.; Iwahama, T.;
Nishiyama, Y. J. Org. Chem. 1995, 60, 3943.
(9) Avila, D. V.; Ingold, K. U.; Lusztyk, J. J. Am. Chem. Soc. 1995, 117, 2929.
(10) Pedrielli, P.; Pedulli, G. F. Gazz. Chim. Ital. 1997, 127, 509; De Heer, M.
I.; Korth, H.-G.; Mulder, P. J. Org. Chem. 1999, 64, 6969.
164
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

behaviour must be due to solvent effect on the substrate,
not on the cumyloxyl radical. The results suggested that the
solvent-induced decrease in k and k is associated with an
8 9
increase in the strength of complexes formed between the
O-H group and t-BuOOH or PhOH, the solvent acting as
hydrogen bond acceptor. This is also reflected in solvent
effect on the BDE values of phenol. The dramatic effect on
eqs 8 and 9 indicates, in our opinion, that an analogous
solvent effect should be likely involved in eqs 3 and 6; a
solvent acting as hydrogen bond acceptor should significantly
p-hydroxybenzoic acid and 4,4′-dihydroxybiphenyl. Acetyl-
aromatics are easily oxidised by O to aromatic carboxylic
acids under mild conditions using catalytic amounts of
2
1
1,12
Mn(NO
3
)
2
and Co(NO
3
)
2
(eq 12).
O2
Ar-COMe Mn(II), Co(II)8 Ar-COOH
(12)
The process is characterised by a free-radical chain
initiated by an electron-transfer process involving the enolic
form of the ketone and Mn(III) nitrate, which is regenerated
1
1,12
in a redox chain (eqs 13-15).
3 6
decrease the rate constants k and k .
This overall kinetic picture suggested that the temperature
and the solvent should be the basic factors that can be
exploited to achieve maximum selectivity for eq 4, consider-
ing that the previous aerobic oxidations catalysed by NHPI
4
,5
reported for cumene mainly led to acetophenone.
The synthetic strategy foresaw then the transformation
of diols to diphenols by reactions with H and acid
2 2
O
catalysis (eq 10) or to dienes (eq 11), useful industrial
products for the synthesis of liquid crystals and cross-linked
polymers.
The hydroperoxide is decomposed by Co(II), yielding the
aromatic carboxylic acid (eqs 16-18).
H
+
8
HOC(Me) -Ar-C(Me) OH + 2 H O
2
2
2
2
(
10)
11)
Ar-COCH -OOH + Co(II) f
2
f HO-Ar-OH + 2 Me-CO-Me + 2 H O
2
•
-
Ar-COCH -O + Co(III) + OH (16)
2
HOC(Me) -Ar-C(Me) OHf
2
2
(
•
Ar-COCH -O f Ar- C˙ O + CH O
(17)
(18)
f CH dC(Me)-Ar-C(Me)dCH + 2 H O
2
2
2
2
2
O2
Ar- C˙ O Co(III)8 Ar-COOH
In addition, p-hydroxybenzoic acid is another important
monomer for liquid crystal polymers. Some of our recent
results1 concerning aerobic oxidation of carbonyl deriva-
tives catalysed by Mn(NO ) and Co(NO ) and the mech-
3 2 3 2
anism of this catalysis suggested that p-hydroxybenzoic acid
could be prepared by oxidation of p-hydroxyacetophenone
This mechanism suggested that the aerobic oxidation is
not viable by this catalytic system with p-hydroxyacetophe-
none either because phenols are inhibitors of free-radical
chain processes or because metal salts with high oxidation
potential, such as Mn(III) and Co(III), rapidly oxidise the
phenolic group by an electron-transfer mechanism (eq 19).
+
1,12
2
with O .
Ar-OH + Mn(III) f Ar- O˙ + Mn(II) + H
(19)
Results and Discussion
Liquid crystal polymers (LCPs) are a form of aromatic
polyester/copolyester resins with excellent flow and me-
chanical properties. British Petroleum’s Xydar LCPs are
made from terephthalic acid, p-hydroxybenzoic acid, and
Therefore, we considered the possibility of aerobic
oxidation of p-hydroxyacetophenone in acetic acid solution
containing acetic anhydride that, by acetylating the phenolic
group, would eliminate the inhibiting effect and the process
of eq 19. Actually, in the presence of acetic anhydride the
aerobic oxidation of p-hydroxyacetophenone leads with high
yield to p-hydroxybenzoic acid (eq 20) under mild conditions.
4,4′-dihydroxybiphenyl. LCPs were initially used primarily
for dual-oven cookware; however, use in cookware declined
dramatically in 1989-1990. Electrical/electronic applications,
where exacting dimensional stability and surface-mount
soldering temperature requirements create the need for a
highly specialised material, are now the major market for
LCPs; there are also commercial applications for LCPs in
the military and in the aerospace, chemical processing,
1
3
medical, and transportation industries.
On the other hand, p-acetoxybenzoic acid was directly
Our recent results,1
1,12
concerning the mechanism of
and Co(NO and
by NHPI, suggested new processes for the production of
1b
reacted with dimethyl terephthalate to yield LCPs.
aerobic oxidations catalysed by Mn(NO
3
)
2
3 2
)
For the synthesis of 4,4′-dihydroxybiphenyl we have taken
into consideration the possibility of utilising 4,4′-diisopro-
pylbiphenyl, which is easily obtained from biphenyl and
propene (eq 21).
Attempts to obtain dihydroxybiphenyl by the classical
autoxidation procedure through dihydroperoxide gave poor
results, since mixtures of mono- and dihydroperoxides were
3
(
(
(
11) Minisci, F.; Recupero, F.; Fontana, F.; Bjørsvik, H.-R.; Liguori, L. Synlett
002, 610.
12) Minisci, F.; Recupero, F.; Pedulli, G. F.; Lucarini, M. J. Mol. Catal., A
2
2
003, 204-205, 63.
13) Chemical Economics Handbook; SRI International: Menlo Park, CA,
February 2001; 695.4021U, p 47.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
165

Table 1. Aerobic oxidation of cumene, catalysed by NHPI
time conversion
cumyl
alcohol (%)
acetophenone
(%)
solvent
T (°C) (h)
(%)
Ph-CNa
100
100
50
50
60
80
70
50
40
20
20
24
5
24
8
24
24
9
24
24
2
76
40
77
91
42
72
98
100
100
40
8
10
62
68
81
69
76
91
99
64
54
38
20
19
31
23
9
- -
- -
- -
- -
AcOHb
t-BuOH
MeCOMec
obtained. Thus, we investigated a different approach, involv-
ing aerobic oxidation of diisopropylbiphenyl to the corre-
sponding benzyl alcohol (eq 4), despite the results reported
by Ishii and co-workers4,5 for the aerobic oxidation of
cumene, catalysed by NHPI, mainly leading to acetophenone.
We took advantage of the kinetic results of Ingold and co-
workers on the temperature and solvent effects on the
â-scission (eq 5) and the hydrogen atom abstraction from
C-H bonds (eq 7) by cumyloxyl radical.
CCl
4
MeCN
MeCN
MeCN
MeCNd
Ph-Cl
50
50
100
100
100
100
1
1
,2-C
,2-C
6
H
H
4
Cl
Cl
2
2
64
15
6
4
a
b
c
Reference 5.
Reference 4, 17% of phenol is also formed.
2% of Mn(NO3)2
d
was added to the reaction mixture. 20% of NHPI was used.
Acetic acid, used by Ishii as solvent in the aerobic
oxidation of cumene, is a particularly unfavourable solvent
for our purpose of obtaining benzyl alcohol for three main
reasons: (i) it gives the highest rate of â-scission for
Co(II) is regenerated either by decomposition of the
hydroperoxide (eq 24) or by oxidation of NHPI (eq 25).
6
-1
cumyloxyl radical (1.9 × 10 s at 30 °C) among the
solvents investigated by Ingold, leading to the formation of
acetophenone; (ii) it is very likely to decrease the rate of
hydrogen atom abstraction from NHPI by cumyloxyl radical,
which leads to the benzyl alcohol (eq 6), by acting as a
3
We have recently evaluated the rate constant for eq 3
3
-1 -1
3
(k ) 7.2 × 10 M
s
at 30 °C), while we do not know
9
hydrogen bond acceptor solvent, since we have recently
the rate constant for eq 6, for which, however, we expect a
much higher rate constant on the basis of the involved
enthalpies (the BDE values of the O-H bonds are 104 and
3
observed a similar solvent effect for eq 3; (iii) it catalyses
the decomposition of the intermediate cumyl hydroperoxide
formed in eq 3, giving the formation of phenol, which inhibits
the free-radical chain of the aerobic oxidation.
-
1
88 kcal mol , respectively, for alcohols, RO-H, and
hydroperoxides, ROO-H). This is also supported by the
great reactivity difference for hydrogen atom abstraction from
Temperature is another important factor in the synthesis
of benzyl alcohol: a higher temperature favours first-order
•
•
C-H bonds by t-BuO (eq 26) and t-BuOO (eq 27) radicals.
6
-8
processes, such as â-scission of cumyloxyl radical (eq 5)
•
•
and the decomposition of the catalyst through self-decay of
R-H + t-BuO f R + t-BuO-H
(
26)
27)
3
5
6
-1 -1
PINO radical (eq 22) more than second-order hydrogen atom
k26 ≈ 10 -10 M
s
abstractions (eqs 1, 6, 7).
•
•
R-H + t-BuOO f R + t-BuOO-H
(
-
3
-1
-1 -1
k27 ≈ 10 -10
M
s
We can expect an analogous difference between eqs 3
and 6.
The results of Table 1 suggest that solvents such as
chlorobenzenes would be particularly convenient for the
synthesis of cumyl alcohol. However, the low solubility of
NHPI in these solvents does not allow high conversions; this
is a major drawback for the oxidation of diisopropylaromat-
ics, since only a complete oxidation of both isopropyl groups
allows industrial application of this process.
A high polarity of the solvent increases the solubility of
NHPI, but at the same time it reduces the selectivity in benzyl
alcohol. Therefore, a compromise has been achieved with a
solvent, acetonitrile, which grants a good solubility for the
catalyst NHPI while allowing complete conversion of the
diisopropylaromatics at a temperature low enough to favour
the selectivity in benzyl alcohol. The low temperature also
favours recovery and recycle of the catalyst because it
On the basis of the kinetic results6-8 and of the Walling-
Ingold rationalisation, the best conditions to obtain a high
selectivity in benzyl alcohol would be the use of nonpolar
solvents and the lowest possible oxidation temperature.
Cumene was chosen as a model to investigate the effect of
solvent and temperature on the selectivity between cumyl
alcohol and acetophenone in the aerobic oxidation catalysed
by NHPI. The results, reported in Table 1, essentially support
the expectations, considering that the presence of the NHPI
catalyst favours the formation of cumyl alcohol according
to eq 6, which is faster than eq 7. The mechanism for the
formation of cumyl alcohol, in fact, involves the chain
process depicted by the equation sequence 1, 2, 3, 23, 6.
decreases decomposition of the PINO radical (eq 22, k22
)
-
1
0
.1 s at 25 °C) compared to hydrogen atom abstraction
-
1
-1
from the isopropyl group (eq 1, k
1
) 3.25 M
s
at
2
5 °C).
166
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development

Yields higher than 90% were obtained for the dibenzyl
alcohol of eq 28 by working at 40 °C under atmospheric
pressure of O .
2
The autoxidation of 2,6-diisopropylnaphthalene to mix-
1
6,17
tures of mono- and dihydroperoxide
was utilised for the
production of 2-hydroxy-6-isopropylnaphthalene and of 2,6-
dihydroxynaphthalene, which are interesting intermediates
for the production of dyes, agrochemicals, drugs, liquid
crystalline materials, and heat-resistant polymers. However,
the drawbacks discussed above for other diisopropylaromat-
ics, are also present in the oxidation of 2,6-diisopropylnaph-
thalene. The formation of 2,6-dihydroxynaphthalene with
high selectivity (>90%) was also reported by Olah and co-
Similar results have been obtained from the catalytic
oxidation of 2,6-diisopropylnaphthalene (eq 29).
1
8
workers by hydroxylation of â-naphthol by H
superacid catalysis (eq 35). The practical application of this
procedure is, however, limited by the use of 90% H in a
2 2
O and
2 2
O
The 4,4′-di(hydroxyisopropyl)biphenyl obtained in eq 28
large excess of HF-BF3 at -50 °C and by the low
conversions.
14
was previously prepared by a much more expensive
procedure involving reaction of 4,4′-diacetylbiphenyl with
MeMgI, and it was used for the synthesis of 4,4′-diisopro-
penylbiphenyl by reaction with acetic anhydride (eq 30).
Thus, the approach involving 2,6-di(hydroxyisopropyl)-
naphthalene, obtained in high yields according to eq 29,
appears to be of interest. By the procedures of eqs 30 and
3
4 the corresponding diene and diphenol were obtained, but
yields are somewhat lower, likely because these molecules
are less stable than the derivatives in the biphenyl series.
No attempt was made until now to improve yields by further
investigation, for the main reason that diisopropylation of
naphthalene by propene is less selective than the correspond-
ing reaction with biphenyl; considerable amounts of 2,7-
diisopropyl and minor amounts of other disubstituted isomers
are also formed. This can affect the overall practical interest,
due to the difficulty of isomer separation.
The diene was used for the synthesis of cross-linked
polymers by copolymerisation with styrene.14
The overall synthetic approach (eqs 28 and 30) reported
in this contribution is far more convenient than the previous
1
9
14
procedure for the production of 4,4′-diisopropenylbiphenyl.
,4′-Dihydroxybiphenyl was not previously produced
from 4,4′-di(hydroxyisopropyl) biphenyl; however, phenol
can be prepared from cumyl alcohol and H by acid
4
2
O
2
1
5
Conclusions
catalysis. This latter has a two-fold purpose: it catalyses
the formation (eqs 31 and 32) and the decomposition (eq
The aerobic oxidation of p-hydroxyacetophenone, cataly-
sed by Mn(NO ) and Co(NO ) , leads with high selectivity
3 2 3 2
and under mild conditions to p-hydroxybenzoic acid, a useful
monomer for LCPs; the presence of acetic anhydride protects
the phenol as an ester group. The aerobic oxidation of 4,4′-
diisopropylbiphenyl, catalysed by NHPI and Co(OAc) ,
2
always in mild conditions (40 °C and atmospheric pressure)
in a suitable solvent, MeCN, leads in high yields to 4,4′-di-
3
3) of cumyl hydroperoxide, as in the classical process for
the production of phenol from cumene hydroperoxide.
+
+
PhCMe OH + H a PhCMe + H O
(31)
2
2
2
PhCMe2+ + H O f PhCMe OOH + H
+
(32)
(33)
2
2
2
PhCMe OOH f PhOH + MeCOMe
2
(hydroxyisopropyl)biphenyl; temperature and solvent effects
play a key role in determining the high selectivity of the
oxidation. The benzylic diol is a useful intermediate for the
synthesis of 4,4′-dihydroxybiphenyl, another important mono-
This procedure, applied to 4,4′-di(hydroxyisopropyl) bi-
phenyl in a two-phase system, has given high yields of 4,4′-
dihydroxybiphenyl (eq 34).
Thus, the combination of eqs 28 and 34 appears to be a
convenient process for the production of 4,4′-dihydroxy
biphenyl. Experiments on larger scale are in progress.
(16) Clausen, M.; Rys, P. (Ciba Geigy AG). Eur. Pat. 288.428, 1988.
(17) Tanaka, M.; Ishibashi, M.; Taniguchi, K. (Mitsui Petrochemical Industries).
Eur. Pat. O 360.729, 1990.
(18) Olah, G. A.; Keumi, T.; Lecoq, J. C.; Fung, A. P.; Olah, J. A. J. Org. Chem.
1991, 56, 6148.
(
(
14) Volyi, I.; Janssen, A. G.; Mark, H. J. Phys. Chem. 1945, 49, 461.
15) Boger, D. L.; Coleman, R. S. J. Org. Chem. 1986, 51, 5436.
(19) Katayama, A.; Toba, M.; Takeuchi, G.; Mizukami, F.; Niwa, S.; Mitamura,
S. J. Chem. Soc., Chem. Commun. 1991, 39.
Vol. 8, No. 2, 2004 / Organic Process Research & Development
•
167

CH
3
MgI, as reported14 (MS: m/z 270 (M ), 255, 334, 195,
+
mer for LCPs, and of 4,4′-di(isopropenyl)biphenyl, which
can be used for the production of cross-linked polymers.
152, 76, 59, 43).
Aerobic Oxidation of 2,6-Diisopropylnaphthalene, Ca-
talysed by NHPI. The same procedure described above for
Experimental Section
4
,4′-diisopropylbiphenyl was utilised for 2,6-diisopropyl-
4
-Hydroxyacetophenone was a commercial product (Al-
drich), 4,4′-diisopropylbiphenyl and 2,6-diisopropylnaphtha-
lene were supplied by Lonza S.p.A. Co(NO , Co(OAc)
Mn(NO , and N-hydroxyphthalimide (NHPI) were also
commercially available.
naphthalene. A 92% yield of crude 2,6-di(hydroxyisopropyl)
naphthalene was obtained by distillation of the solvent. The
3
)
2
2
,
2
crude product was crystallised from EtOH:H O (4:1 v/v) to
3 2
)
2
2
give the pure product, mp 157-158 °C, lit 156-158 °C.
1
H NMR: δ 1.61 (12 H, 4 CH
3
), 7.56 (2 H aromatic,
Synthesis of p-Hydroxybenzoic Acid by Aerobic Oxi-
dation of p-Hydroxyacetophenone. A solution of 2.72 g
positions 3 and 7), 7.76 (2 H aromatic, positions 4 and 8),
7
.88 (2 H aromatic, positions 1 and 5).
(
20 mmol) of 4-hydroxyacetophenone, 100 mg (0.4 mmol)
of Mn(NO ‚4 H O and 116 mg (0.4 mmol) of Co(NO ‚6
O in 50 mL of acetic acid and 10 mL of acetic anhydride
Synthesis of 4,4′-Diisopropenylbiphenyl. The procedure,
3
)
2
2
3 2
)
described in ref 14, was utilised; it simply consisted of
boiling 4,4′-di(hydroxyisopropyl)biphenyl in acetic anhydride
solution for 6 h; 93% yield of 4,4′-diisopropylbiphenyl was
obtained, in good agreement with the literature results.14
Synthesis of 2,6-Diisopropenylnaphthalene. The same
procedure described above for 4,4′-diisopropenylbiphenyl
was utilised, starting from 2,6-di(hydroxyisopropyl)naph-
thalene. A 87% yield of 2,6-diisopropenylnaphthalene was
obtained, determined by GLC analysis. The compound was
not known, and it was characterised by MS and NMR
H
2
was stirred for 8 h in oxygen atmosphere at 100 °C and
ambient pressure. The solvents were distilled off, and 3.7 g
of solid residue was obtained. It was dissolved in CH Cl ,
2 2
filtered from the small amounts of metal salts, and analysed
by GLC as p-acetoxymethyl benzoate after treatment with
2 2
CH N by using p-methoxymethyl benzoate as internal
standard. p-Acetoxymethyl benzoate, 3.5 g (corresponding
to 91%), was obtained. The crude p-acetoxybenzoic acid
gave, by flash chromatography on silica gel, the pure
compound (mp 191-193 °C, lit.20 191-192 °C). Acid
hydrolysis gave p-hydroxybenzoic acid (mp 215-217 °C,
+
spectra. MS: m/z 208 (M ), 193, 178, 165, 152, 139, 128,
1
1
5
7
15, 89, 76. H NMR: δ 2.27 (6 H, 2 CH
3
), 5.20 (2 H, vinyl),
.54 (2 H, vinyl), 7.66 (2 H aromatic, positions 4 and 8),
.78 (2 H aromatic, positions 3 and 7), 7.82 (2 H aromatic,
2
1
lit. 215-217 °C).
Aerobic Oxidation of Cumene, Catalysed by NHPI.
Solutions of 2.40 g (20 mmol) of cumene, 326 mg (2 mmol)
of NHPI, and 100 mg (0.4 mmol) of Co(OAc)
positions 1 and 5).
2 4
Synthesis of 4,4′-Dihydroxybiphenyl. H SO (0.15 mL,
2
2
‚4 H O were
3
mmol) was added dropwise to a mixture of 480 mg (2
mmol) of 4,4′-di(hydroxyisopropyl)biphenyl and 0.77 mL
9 mmol) of H (35%). The mixture was refluxed for 12
h and then extracted with CH Cl . The CH Cl solution was
extracted with 5% aqueous NaOH. The basic aqueous
solution was acidified with H SO and extracted by ethyl
stirred under oxygen atmosphere at ambient pressure; tem-
peratures, reaction times, and solvents (25 mL) are reported
in Table 1. The solutions were analysed by GLC: unreacted
cumene, cumyl alcohol, and acetophenone were the only
reaction products; the only exception concerns the reaction
in acetic acid, where a significant amount of phenol is also
formed. Quantitative analysis was carried out by GLC using
benzyl alcohol as internal standard. Conversions and selec-
tivities are reported in Table 1.
(
2 2
O
2
2
2
2
2
4
acetate, obtaining 0.36 g of crude 4,4′-dihydroxybiphenyl,
corresponding to 92% yield, evaluated by GLC analysis. The
pure product was obtained by flash chromatography on silica
+
gel and characterised by MS (m/z 186 (M ), 157, 139, 128,
Aerobic Oxidation of 4,4′-Diisopropylbiphenyl, Ca-
talysed by NHPI. A solution of 4.77 g (20 mmol) of 4,4′-
diisopropylbiphenyl, 652 mg (4 mmol) of NHPI, and 100
1
15, 93, 77, 63, 51) and comparison with a commercial
sample.
Synthesis of 2,6-Dihydroxynaphthalene. The same
mg (0.4 mmol) of Co(OAc)
2
2
‚4 H O in 25 mL of acetonitrile
procedure described above for 4,4′-dihydroxybiphenyl was
used, starting from 2,6-di(hydroxyisopropyl)naphthalene. 2,6-
Dihydroxynaphthalene was isolated in 48% yield and char-
was stirred at 40 °C for 12 h under oxygen atmosphere at
ambient pressure. The solvent was distilled off, and 6.4 g of
solid residue was obtained. GLC analysis using cumyl
alcohol as internal standard revealed the presence of 4.9 g
of 4,4′-di(hydroxyisopropyl)biphenyl, corresponding to 90.7%
yield. The pure product was separated by flash chromatog-
raphy on silica gel and characterised by comparison with an
authentic sample, obtained from 4,4′-diacetylbiphenyl and
1
+
acterised by MS and H NMR spectra. MS (m/z 160 (M ),
1
1
31, 115, 103, 77, 63, 51); H NMR: δ 7.07 (2H aromatic,
positions 3 and 7), 7.13 (2H aromatic, positions 1 and 5),
.56 (2H aromatic, positions 4 and 8), 8.35 (2H, OH).
7
Received for review September 23, 2003.
OP034137W
(20) Padilla, H. Tetrahedron 1962, 18, 427.
(21) Takahashi, H.; Kubota, Y.; Fang, L.; Onda, M. Heterocycles 1986, 24(4),
(22) Mazurkiewicz, R.; Stec, Z.; Zawadiac, J. Magn. Reson. Chem. 2000, 38,
1
099.
213.
168
•
Vol. 8, No. 2, 2004 / Organic Process Research & Development