Oxidation studies of phenols with cobalt(II) ion-modified silica in a liquid chromatographic reactor

Article abstract of DOI:10.1039/a801253i

The activity of cobalt(II) ion-modified silica, a new catalyst for the oxidation of phenols with dissolved oxygen in organic solvents, has been studied in liquid chromatographic reactors using several approaches. In this reactor mode, catalyst activity can be maintained over months of repeated use while phenols are oxidized to p-benzoquinones at or below 45°C. This approach eliminates problems of catalyst separations. Using the ideal liquid chromatographic reactor model with inert standard methods, the reaction is found to be first-order with respect to the phenolic reactants. Some pseudo-first-order rate constants have also been measured. The cobalt(II) ions bonded to silica probably function through reaction pathways similar to those of homogeneous cobalt complex catalysts for phenol oxidations. However, they are stable for long periods of time under the mild reaction conditions used here and can provide separation with concerted reaction.

Full text of DOI:10.1039/a801253i

View Article Online / Journal Homepage / Table of Contents for this issue
Oxidation studies of phenols with cobalt(II) ion-modified silica in a
liquid chromatographic reactor
Zengqun Deng, Gunnar R. Dieckmann and Stanley H. Langer*
Department of Chemical Engineering, University of Wisconsin, Madison, WI 53706, USA
The activity of cobalt(II) ion-modified silica, a new catalyst for the oxidation of phenols with dissolved
oxygen in organic solvents, has been studied in liquid chromatographic reactors using several approaches.
In this reactor mode, catalyst activity can be maintained over months of repeated use while phenols
are oxidized to p-benzoquinones at or below 45 ؇C. This approach eliminates problems of catalyst
separations. Using the ideal liquid chromatographic reactor model with inert standard methods, the
reaction is found to be first-order with respect to the phenolic reactants. Some pseudo-first-order rate
constants have also been measured. The cobalt(II) ions bonded to silica probably function through
reaction pathways similar to those of homogeneous cobalt complex catalysts for phenol oxidations.
However, they are stable for long periods of time under the mild reaction conditions used here and can
provide separation with concerted reaction.
Introduction
Other homogeneous catalysts have been reported. For
example cobalt() complexes with N,NЈ-disalicylindene-
Phenol oxidations are important reactions with considerable
significance in the synthesis of a variety of organic chemicals,
particularly the biogenetic syntheses of natural products.
They are important in treating hazardous wastes from petro-
chemical plants and refineries, wood processing, etc., where
phenolic compounds are major pollutants. Available oxid-
ative pathways include dehydrogenative coupling in para and
ortho positions, oxidations to free radicals, hydroxylation of the
aromatic rings, and frequently oxidations to quinones. With
such importance, fairly extensive methodologies have been
documented for phenol oxidations in both chemical and bio-
15
ethylenediamine were used for tert-butylphenol oxygenation;
3ϩ
complexes of Fe
ion with protoporphyrins known as
hematines were used for chlorinated phenol oxidations; cupric
and cobaltous phthalocyanine were used for oxidation of alkyl-
1
,2
4
16
ated phenols to p-benzoquinones. Many of these reactions can
3,4
occur at relatively low temperatures (<80 ЊC) and ambient pres-
sure. But, as is well recognized, the use of homogeneous
catalysts can result in separation problems.
Until now, major recognition of the catalytic activities of
metal ion-modified liquid chromatographic columns has been
in connection with inadvertent interference with separations
so that such activity has tended to be avoided or eliminated.
5–8
logical processes.
The present study stems from our interest in ion-modified
silicas and the finding that cobalt(II) ion-modified silica can
serve as a catalyst for selective oxidation of phenols to p-
benzoquinones with molecular oxygen dissolved in the organic
Applications of the use of catalytically active columns as liquid
17,18
chromatographic reactors have been relatively few.
Related
to the applications discussed here is the catalyzed oxidation of
hydroquinone-type compounds on normal phase metal-
9
mobile phase of an HPLC liquid chromatographic column.
18a
contaminated silica columns and triphenyl phosphite oxid-
This provides the possibility of using the column as a chrom-
atographic reactor with concerted reaction and separation
features. Iron, copper and chromium modified silicas which
demonstrate activity for hydroquinone oxidations to benzo-
quinones do not possess similar activity for phenol oxidations.
The cobalt modified silica had only slight activity for hydro-
quinone oxidation to quinones. Traditionally, the oxidation
of phenols to quinones has been carried out using oxidants
such as mercuric oxide, chromium trioxide, mercury() tri-
fluoroacetate, Fremy salt, hydrogen peroxide in the presence of
19
ation to phosphate in iron() ion-modified silica columns.
With the importance of and rising interest in reactions per-
formed on a small scale, the chromatographic reactor column
offers a means for reacting micro amounts of materials with the
concomitant benefits of conservation and separation. Losses
are minimal relative to batch reactors and unreacted dissolved
material can be recovered or recycled.
The feasibility of using liquid chromatographic reactors to
study and perform selective phenol oxidations in several ways is
demonstrated here while showing that cobalt() ion-modified
silica is an effective catalyst for oxidation to p-benzoquinones.
Some results of kinetic studies on phenol oxidations in the
cobalt() ion-modified silica column are also presented
together with speculations on the reaction mechanism.
5
2
horseradish peroxide, and more recently cobalt() acetate.
Phenol oxidations with molecular oxygen have also been
possible using certain transition metal complexes as catalysts,
the majority of which are complexes of cobalt() with Schiff
1,10
base ligands,
such as cobalt() bis[3-(salicylideneamino)-
11
propyl]methylamine, CoSMDPT, and aqua[1,2-bis(pyridine-
12
2
-carboxamido)benzene] cobalt(), [Co(bpb)H O].
latter cobalt() complexes are noteworthy for their oxygen-
bonding properties occurring through open coordination.
These
2
Experimental
1
Materials
Activation of molecular oxygen for reaction with an organic
substrate is believed to occur through partial electron transfer
Silica gel (Spherisorb S10W) obtained from Phase Separations
Inc., was used for cobalt() ion immobilization and for column
packing. The preparation of the cobalt() ion-modified silica
packing (1000 ppm cobalt) via a two stage process has been
11
from the cobalt, weakening the oxygen–oxygen bond. Often
the oxidation of phenols using an oxygen carrier chelate and
molecular oxygen in batch processes gives not only quinones
but coupled products or polymers, depending on the solvent
9
described. Two grams of Spherisorb S10W (Phase Separations)
Ϫ1
chromatographic silica were added to 75 ml of 0.04 mol
8
and catalysts. An increase in the catalytic efficiency and select-
potassium acetate solution adjusted to pH 8.4 with acetic acid.
After 4 h of stirring, followed by centrifuge separation, the
silica was washed three times with water and then with iso-
ivity of oxygen carrier complexes toward phenol oxidations has
11,13,14
been achieved through modification of ligand structures.
J. Chem. Soc., Perkin Trans. 2, 1998
1123

View Article Online
Fig. 2 Chromatograms of 1.05 m 2,6-dimethylphenol at 45.0 ЊC in
Ϫ1
(
a) cobalt() ion-modified silica column at a flow rate of 1.0 ml min ,
and (b) cobalt() ion-modified silica column at a flow rate of 0.65 ml
Ϫ1
min . For other conditions, see Fig. 1 caption. T: toluene inert. R: 2,6-
dimethylphenol reactant. P: 2,6-dimethyl-1,4-benzoquinone formed
from 2,6-dimethylphenol oxidation in the column.
Fig. 1 Chromatograms of 0.82 m phenol at 45.0 ЊC from (a) a bare
Ϫ1
silica column at a flow rate of 1.0 ml min , (b) cobalt() ion-modified
Ϫ1
silica column at a flow rate of 1.0 ml min , and (c) cobalt(II) ion-
Ϫ1
modified silica column at a flow rate of 0.5 ml min . Mobile phase: 1%
mobile phase of 1% (v/v) 2-methylpropan-2-ol in hexane. An
in-house packed bare silica column, and the cobalt() ion-
modified silica column were used under similar conditions for
comparison and identification of phenol oxidative activities.
Phenols for study were dissolved in mobile phase aliquots,
yielding initial concentrations of 0.82–1.08 m. The experi-
mental method applied for each kinetic study is listed in Table
2
-methylpropan-2-ol in hexane; sample size: 20 µl. T: toluene inert
(
unretained). R: phenol reactant. P: p-benzoquinone formed due to
phenol oxidation in the column.
propyl alcohol. The potassium exchange silica was stirred with
Ϫ1
0
.0029 mol l CoCl in isopropyl alcohol for 3 h, followed by
2
washing with isopropyl alcohol (eight times), water twice and
finally isopropyl alcohol three times to remove water. A more
active cobalt loaded catalyst (1500 ppm cobalt) was also pre-
1
. A sample loop, 20 µl, was used for introducing dissolved
reactants with solutes onto the column. Reaction chromato-
graphs result from some solutes undergoing simultaneous
reaction and separation in the column.
pared by increasing the potassium acetate and CoCl concen-
2
Ϫ1
trations to 0.10 and 0.015 mol l , respectively, and repeating
Acetophenone, inert under our experimental conditions, was
used as an inert standard, I, as well as an external standard to
correlate kinetic data (vide infra). Both ‘internal’ and ‘external’
standard methods for evaluating kinetic data from reaction
each step twice. The phenolic compounds including phenol,
chlorophenols, cresols, 2,6-dimethylphenol, 2,6-di-tert-butyl-
phenol, and corresponding p-benzoquinones were obtained
from Aldrich and used without further purification. All the
HPLC grade solvents were obtained from Burdick & Jackson.
For the chromatographic reactor column experiments, solvents
were sonicated before use in the column, but no further attempt
was made to remove dissolved oxygen which was a co-reactant.
18
chromatograms were used as described previously. Toluene
was added to all solutions except 2,6-di-tert-butylphenol for the
column dead volume Vm (and dead time, t ) measurement.
m
Three to five separate injections were used at each flow rate, and
reproducible peak areas were observed. The actual reaction
time on the silica, t , is the difference between the reactant
s
Apparatus
column residence time and the dead time for inert material.
The high-performance liquid chromatographic array used here
For the 2,6-di-tert-butylphenol oxidation study, a stopped-flow
20
21,22
is similar to that described earlier. It incorporates a Beckman
10B solvent delivery pump, a Beckman Series 210 four-port
technique
was used.
1
sample valve with a 20 µl sample loop, an LDC/Milton
Roy spectroMonitor D variable wavelength detector, mostly
adjusted to a detection wavelength of 245 nm, a Hewlett-
Packard 3396A integrator, and sometimes in parallel, a
Sargent-Welch model SRG recorder. A Perkin-Elmer LC-235
diode array detector was also used as appropriate to scan and
compare the UV spectra of standard phenols, benoquinones
and product waves from phenol oxidations.
Results and discussion
Catalytic activity
The catalytic activities of the cobalt() ion-modified silica
column for the oxidation of phenol, 2-6-dimethylphenol and
o-cresol can be seen from the reaction chromatograms in Figs.
–3. o-Chlorophenol oxidation was demonstrated earlier. The
9
1
catalytic activity for injection of phenol into the unmodified
Column (10 cm × 4.6 mm id) were slurry-packed with 10 µm
Spherisorb S10W silicas and cobalt() ion-modified Spherisorb
S10W silica prepared in our laboratory. Isopropyl alcohol was
used at about 6000 psi with an in-house slurry-packing appar-
atus. For column temperature control, a water-circulating
silica column at 45 ЊC with a mobile phase flow rate of 1.0
ml min resulted in a single narrow phenol peak (R) in the
Ϫ1
chromatogram [Fig. 1(a)], whereas injection of the same solu-
tion under the same conditions into a cobalt() ion-modified
silica column resulted in overlapping double peaks (P and R)
chromatogram [Fig. 1(b)]. With a reduced mobile phase flow
19
column jacket was used as described earlier.
Ϫ1
rate of 0.5 ml min or a longer residence time [Fig. 1(c)], the
Column reactions
The oxidation reactions were conducted at temperatures of 35
and 45 ЊC and at flow rates between 0.1 and 4.0 ml min with a
product (P) is increased while the amount of phenol reactant
(R) is decreased, consistent with the conclusion that catalytic
Ϫ1
18,19
activities originated from ion-column reactions.
Product
1
124
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
Table 1 Column reaction study method, capacity factors, kЈ, and pseudo-first-order reaction rate constants for oxidation of phenolic compounds to
II
p-benzoquinones at 45 ЊC in 1% tert-butanol in hexane on a Co 1000 ppm silica column
a
Ϫ4 Ϫ1b
Reactant
Method of study
kЈ (Reactant)
kЈ (Product)
pKa
ks/10 s
Phenol
o-Chlorophenol
Internal standard
External standard
4.9
0.66
4.2
2.5
9.99
8.48
7.9 ± 0.5
33 ± 2.3
c
1
7 ± 1.3
m-Chlorophenol
o-Cresol
m-Cresol
Internal standard
Internal standard
Internal standard
External standard
Stopped-flow
4.7
2.7
5.1
1.2
2.5
3.3
3.3
2.8
9.02
10.26
10.09
10.22
14 ± 1.2
9.3 ± 0.5
5.6 ± 0.3
5.2 ± 0.2
2
2
,6-Dimethylphenol
,6-Di-tert-butylphenol
Very low
a
b
pK values are for aqueous solutions at 25 ЊC, from Lange’s Handbook of Chemistry, 12th edn., McGraw-Hill, New York, 1979. First order rate
a
constants shown with error. Correlation coefficients for the rate constants, determined from the kinetic plots (Fig. 5) were narrowly centered around
c
0
.970 with m-chlorophenol showing a lower value of 0.958 and 2,6-dimethylphenol a higher value of 0.993. At 35 ЊC.
poor separation of the phenol and benzoquinone, caused by
using a mobile phase dried over molecular sieves. In addition to
poorer quinone yields, the phenol exhibited a larger number of
detectable products (at least five) compared with two or three
for the substituted phenols, in accord with previous experi-
23
mental results. Oxidation reactions of phenols, conducted by
other workers, are run with alkyl substituents on the aromatic
ring to help minimize side reactions and enhance quinone
8
yield.
The elution of product (P) and reactant (R) can vary (com-
pare Figs. 1 and 2). These are a result of solubility effects and
the apparent steric effect of ortho groups on solute phenol
interactions and retentions. Because of an enhanced steric
effect, the retention time of 2,6-di-tert-butylphenol, for
instance, in a silica column or in a cobalt() ion-modified silica
column is quite small, and it is not readily separated from un-
adsorbed toluene despite the phenolic hydroxy group. At a
Fig. 3 Chromatograms of 1.05 m o-cresol at 45.0 ЊC in cobalt() ion-
Ϫ1
modified silica column at a flow rate of 1.0 ml min . For other condi-
tions, see Fig. 1 caption. Dotted line represents extrapolation of signal
from reactant. Dashed line represents estimated product signal in elut-
ing reactant. T: toluene inert. I: Acetophenone inert standard. M: Per-
turbation from solvent injection. R: o-cresol reactant. P: 2-methyl-1,4-
benzoquinone formed from o-cresol oxidation in the column.
Ϫ1
mobile phase flow rate of 1.0 ml min , injection of 2,6-di-tert-
butylphenol into a cobalt() ion-modified silica column does
not show any sign of oxidation in the column [Fig. 4(a)]. How-
ever, when indicated stopped-flow intervals were utilized
emerges before reactant. Chromatograms that demonstrate the
catalytic activity of cobalt() ion-modified silica column for
m-chlorophenol oxidation to 2-chloro-1,4-benzoquinone are
similar but are not shown. Product emerges before the reactant.
21,22,24
to give a longer residence time on the stationary phase, 2,6-di-
tert-butylphenol was oxidized in the cobalt() ion-modified
silica column. Fig. 4(b) and (c) show reaction chromatograms
for stopped-flow intervals, ∆t = 10 and 25 min, respectively.
These figures show that for an increased reaction interval, the
product peak (P) is clearly increased while reactant peak (R) is
decreased, demonstrating an additional facet of chromato-
graphic reactors. UV spectra and reinjection of the effluent into
the unmodified silica column were again used for confirming
product identification. Similar stopped-flow experiments on
an unmodified silica column did not generate multiple peak
chromatograms (not shown) and no reaction was detectable.
The cobalt() ion-modified silica catalytic activity for oxidation
of 2,6-di-tert-butylphenol to 2,6-di-tert-1,4-benzoquinone was
thus verified.
(
2
See capacity factors in Table 1.) In Fig. 2, the oxidation of a
,6-dimethyl substituted phenol is demonstrated. Here, how-
ever, reactant emerges before product. Fig. 2 shows similar
features for the catalytic oxidation activities of the disubstituted
phenol, as explained for Fig. 1. In Fig. 3, a complete reaction
chromatogram for o-cresol of the type used for kinetic meas-
urements is illustrated together with I, the inert standard
‘acetophenone’.
For the phenol reactants shown in Table 1 the major reaction
product was determined to be the corresponding p-benzo-
quinone. Other products could also be detected as explained
below, but were not identified. Major reaction products were
confirmed by recording UV spectra of product waves at various
elution times and comparing these with the UV spectra of
pure p-benzoquinone and substituted p-benzoquinones under
the same experimental conditions. In addition, reinjection of
trapped effluents into the unmodified silica analytical column
further confirmed the oxidation products. When excess solvent
in trapped samples was removed by evaporation, other unidenti-
fied reaction products were more easily detected. By placing the
unmodified silica column after the cobalt() ion-modified silica
column, the product peak could be separated from the reactant
and resolved into its components, permitting quantification of
the phenol and p-benzoquinone by peak area. For phenol, m-
cresol and 2-chlorophenol reactions, studied by this method,
the p-benzoquinones accounted for approximately 50% of the
products with conversions of roughly 30, 40 and 10%, respect-
25
Compared to conventional cobalt complex catalysts, the
catalysts of this study offer advantages. Among these are the
elimination of coordinating organic ligands of the homo-
geneous catalyst, and use of a heterogeneous catalyst that is
12
effective for reaction under similarly mild conditions. For the
catalyst column described here, wherein not all cobalt ions are
equally catalytic, no significant catalyst deactivation was
observed after over 300 injections with reactive substrates (20 µl
loop, 1m concentration). The simultaneous separation feature
of products from reactants in the liquid chromatographic
reactor also tends to make phenol oxidations more selective.
The results of these experiments show that conversion of
eluting o- and m-phenols can be controlled as desired by
altering residence times. Where more complete conversion of
phenols is required for preparative purposes, additions of refer-
ence toluene and acetophenone, shown in Fig. 3, can be elimin-
ated. Then, a series of pulses with longer residence times can be
used with timed interval trapping to obtain more effective
Ϫ1
ively for each phenol at a flow rate of 1.0 ml min . The quantity
of 1,4-benzoquinone produced from the reaction of the unsub-
stituted phenol tended to vary and could be significantly
decreased (25% of products). The latter was associated with
J. Chem. Soc., Perkin Trans. 2, 1998
1125

View Article Online
Fig. 5 Pseudo-first-order kinetic plots for the oxidation of (a) o-cresol
and (b) m-cresol in cobalt() ion-modified silica column at 45.0 ЊC
k = k (t /t ) ϩ k (t /t )
(2)
a
m
m
R
s
s
R
rate constants for reaction in the mobile phase and the station-
ary phase, respectively; t and t are the reactant residence times
m
s
in the mobile phase and on the stationary phase (cobalt–silica),
and tR is the total reactant residence time in the column
(
t ϩ t ). If k is negligible relative to k , then eqn. (2) simplifies
m s m s
to give eqn. (3). Substituting eqn. (3) into eqn. (1) gives eqn. (4).
k = k (t /t )
(3)
(4)
a
s
s
R
Fig. 4 Chromatographs of 1.08 m 2,6-di-tert-butylphenol at 45.0 ЊC
in a cobalt() ion-modified silica column at a mobile phase flow rate of
ln(A /A ) = ln(A /A ) Ϫ ksts
R I R I t = 0
Ϫ1
1
.0 ml min with (a) no stopped-flow, (b) stopped-flow time ∆t = 10
min, and (c) stopped-flow time ∆t = 25 min. Mobile phase: 1%
-methylpropan-2-ol in hexane; sample size: 20 µl. R: 2,6-di-tert-
Flow rate variation experiments at controlled temperatures
similar to those shown in Figs. 1 and 2) can provide the relative
A /A data for various stationary phase residence times t .
2
(
butylphenol reactant. P: 2,6-di-tert-butyl-1,4-benzoquinone product
formed from reactant oxidation in the column.
18b,20
R
I
s
Fig. 3 is a chromatogram showing the presence of the inert
2
6,27
acetophenone. The areas of this peak is A . The reactant peak
I
reactor use of the whole column.
I9t is also possible to load
18b,19,28
(R) and product wave overlap as discussed earlier.
Hypo-
more cobalt ion-modifier on the silica.
thetical extensions are indicated by dotted lines. Means for par-
titioning areas to obtain a good approximation of AR (where
there is not excessive overlap), which were used here, have been
Separation of the products from the reactant is largely
determined by the silica, but catalytic activity is determined by
the cobalt modifier, which can alter retention times as illus-
trated in Fig. 1(a) and (b). There is a question of optimizing
conditions for separation which can involve a variety of con-
siderations, e.g. amount of modifier, eluting solvent composi-
tions, temperature. However, as one referee has noted ‘a dif-
ficulty will always be that the conditions that favor reaction will
not necessarily favor separation’. Furthermore, since reaction
takes place over the whole column in our experiments, there is
not a final separation of products from the reactant peak. How-
ever, the product wave can be trapped and separated for further
use, or a second silica column can be attached to the chromato-
graphic reactor to provide pure product or reactant for further
processing.
18b,28
discussed on a number of occasions elsewhere.
The slope
of the line for a plot of ln(A /A ) vs. t provides a value for
R
I
s
the pseudo-first-order rate constant, k , under our reactant
s
conditions.
The oxygen solubility in the organic mobile phase used here
29
is about 10 m, well in excess of the maximum phenol concen-
trations in pulses used in our experiments (1.08 m). With the
cobalt() ion-modified silica stable under these experimental
conditions, the number of catalytically active sites for the reac-
tion remains essentially constant. Thus, the oxidation of
phenols can be tested with a pseudo-first-order reaction treat-
ment, with catalyst and oxygen concentration incorporated in
the stationary phase rate constant, k , obtained through eqn.
s
(
4).
Fig. 5 illustrates the plots of ln(A /A ) versus t at 45 ЊC for
Reaction kinetics
R
I
s
Using the ideal chromatographic reactor model, the possibility
that reaction kinetic parameters might be evaluated was
investigated. The oxidation is treated as a pseudo-first-order
reaction with an excess of oxygen and constant catalyst concen-
tration, as described earlier. The microscopic mass balance
for first-order or pseudo-first-order reactions coupled with
the oxidation of o- and m-cresol. These plots are essentially
linear as were the plots for the oxidations of the other phenols
for which rate constants are reported. This supports the
pseudo-first-order reaction treatment for the phenol oxidations
to p-benzoquinones, and is in agreement with the earlier belief
that the metal catalyzed reaction is first order with respect to
the parent organic compounds. The k values obtained from
these plots are shown in Table 1.
18b
18b,20,28
appropriate initial and boundary conditions,
yields eqn.
30
s
(
1) with the inert standard (acetophenone) method, where A is
R
ln(A /A ) = ln(A /A )
Ϫ kat
(1)
R
I
R
I
t = 0
R
Reaction mechanism
Most studies on phenol oxidations using cobalt complex
11,12
the peak area of the reactant on the reactor chromatogram, AI
catalysts
agree that the reaction involves radical intermedi-
is the peak area of unreactive inert standard (either internal or
ates. It is postulated that the metal complex plays two import-
ant roles in the oxidation catalysis: (1) coordination activates an
external) under the same conditions, and k is the apparent first-
a
order or pseuso-first-order rate constant. In a more general
situation, ka contains contributions from reaction in both
mobile and stationary phases [eqn. (2)], where k and k are the
O molecule enhancing its basicity and capacity to abstract
hydrogen atoms; (2) coordination also can enhance bound
O molecule activity for free radical reactions with organic
2
2
m
s
1
126
J. Chem. Soc., Perkin Trans. 2, 1998

View Article Online
11
radicals. With cobalt() ions directly attached to silica as in
this work, a similar reaction pathway can be proposed to
account for the phenol oxidations here. Scheme 1, based on the
organic solvents used here. The controlled exposure to dis-
solved oxygen is probably also advantageous in limiting
catalyst deactivation.
It is worth noting that invoking a concerted separation prop-
erty accompanying the reaction in the liquid chromatographic
reactor, or a similar packed bed, also minimizes possibilities
for oxidative coupling to provide a more selective pathway to
product quinones. Quinones are known to react with phenols to
CoII + O2
Co – O2
(a)
(b)
O•
OH
23
yield coupled products. Thus, the separation features of
chromatographic and related reactors provide advantages in
comparison with batch and plug flow reactors where products
and intermediates both remain in relatively long term contact
with reactant. Most importantly, as demonstrated, small quan-
CoII + HO•
+
Co – O2 +
2
Ϫ8
O•
tities of materials (10 mol) may be readily reacted without the
O
problem of losses associated with other reactors. There is the
accompanying advantage of separation. Such reactors should
be of greater interest in the future to workers requiring organic
synthesis on a small scale.
+
Co – O2
(c)
H
O
+
O
Co
Chromatograms of the type shown in Fig. 3 also indicate the
possibility of further enhancing liquid chromatographic
columns as small scale preparative reactors through the use of
successive injections of reactant solute in solvents without inert
O
O
O
26,27
markers or standards.
In this manner, reacting serial sam-
CoIII OH
(d)
ples can proceed down the column reactor to exploit catalyst
sites throughout its length. We have illustrated this type of
application earlier in gas chromatography to provide formalde-
H
O
O
+
Co
26
hyde on site. Stopped flow intervals, illustrated in Fig. 4, can
also be employed to achieve greater conversions. The Co–silica
catalyst described here may well demonstrate useful oxidizing
O•
OH
31
properties with other reagents.
CoIII OH
CoII
(e)
+ H2O +
Conclusions
Cobalt() ion-modified silica which can catalyze the oxidation
of phenols to p-benzoquinones offers unique properties and
advantages compared with related homogeneous cobalt com-
plex catalysts. The use of the liquid chromatographic reactor
gives convenient access to oxidation conditions as well as
kinetic information on phenol oxidations and permits reactions
to be conducted efficiently on limited reactant quantities.
Cobalt() ions attached to silica surfaces provide a new and
alternative means for phenol oxidations to benzoquinones
under quite mild conditions. Other applications of these and
related ion modified silicas which do not require complex
organic ligand coordination at the metal ion can be expected
to emerge. Additional studies to evaluate these and related
catalysts continue in our laboratories.
Scheme
1
Proposed reaction scheme for phenol oxidations in
cobalt() ion-modified silica column
homogeneous mechanism proposed by Drago and co-
11
workers, shows a hypothetical reaction scheme to account for
these oxidations to benzoquinones. Other side products are
possibly the result of free radical reactions giving rise to coupled
products.
On the basis of the ion-exchange method used for cobalt()
9
II
ion attachment to silica and observed properties, Co can be
considered to be bonded to the silica surface in a manner
resembling cobalt complex bonding. The bound ion then can
associate with a dissolved O molecule in a similar way to form
2
1
1
an adduct [reaction (a)]. It then abstracts a hydrogen atom to
form a phenoxyl radical [reaction (b)] to initiate the oxidation.
The resulting radical can associate with other coordinated
oxygen molecules to bind to cobalt and give a semiquinone
Acknowledgements
We thank the US Army Research Office, the University of
Wisconsin and the Mobil Foundation for financial support.
We also thank Richard Miller for experimental assistance and
Jay Loppnow for help with figure presentations.
[
reaction (c)]. The bound intermediate then gives a benzo-
III
quinone and a Co intermediate which can propagate further
reaction. Initial phenol solute association with the surface
cobalt does not seem critical since poorly retained di-tert-butyl
phenol is also oxidized during stopped flow. The similarity is
supported by the order of catalytic activities as well.
References
Pseudo-first-order rate constants are listed with the pKa
values of corresponding phenols in Table 1. Although these rate
constants do not give a linear Hammett relationship with
1 R. D. Jones, D. A. Summerville and F. Basolo, Chem. Rev., 1979, 79,
139.
2
3
M. Hirano, T. Ishii and T. Morimoto, Bull. Chem. Soc. Jpn., 1991,
4, 1434.
C. W. Kenney and L. A. Uchida, USP 4 582 613/1986 (Chem. Abstr.,
986, 104, 229 978p).
6
respect to pK values [which were measured in aqueous solution
a
2
at 25 ЊC (linear regression r = 0.85)], the general trend of the
1
rate constant decreasing with increasing pK values appears to
a
4 K. L. Sublette, USP 5 004 551/1991 (Chem. Abstr., 1991, 115,
hold. This supports the hypothesis that the cobalt() ions close-
14854m).
ly joined to silica surfaces have an activity similar to that of
5 M. Hudlicky, Oxidations in Organic Chemistry, ACS monograph
186, American Chemical Society, Washington, DC, 1990, pp. 163–
11
cobalt() Schiff base ligand complexes but with a more stabil-
ized ligand structure (silica) at one face. Hence, these structures
may be viewed as heterogenized catalysts without complex
organic ligands, functioning in ways similar to the homo-
geneous catalyst, but with stabilization advantages in the
1
68.
6
7
T. Kametani and K. Fukumoto, Phenol Oxidations, Giho-do,
Tokyo, 1970.
Oxidative Coupling of Phenols, ed. W. I. Taylor and A. R. Battersby,
Marcel Dekker, New York, 1967.
J. Chem. Soc., Perkin Trans. 2, 1998
1127

View Article Online
8
9
L. Simándi, Catalytic Activation of Dioxygen by Metal Complexes,
Kluwer Academic, Dordrecht, 1992, pp. 196–243.
Z. Deng, G. R. Dieckmann and S. H. Langer, Chem. Commun.,
22 J. B. Powell and S. H. Langer, J. Catal., 1985, 94, 566.
23 M. Musso, in Oxidative Coupling of Phenols, ed. W. I. Taylor and
A. R. Battersby, Marcel Dekker, New York, 1967.
1
996, 1789.
24 J. Coca, G. Adrio, C.-Y. Jeng and S. H. Langer, in Preparative and
Production Scale Chromatography, ed. G. Ganetsos and P. E. Barker,
Chromatographic Science Series, vol. 61, Marcel Dekker, New York,
1993, p. 449.
25 A. Nishinaga, in Fundamental Research in Homogeneous Catalysis,
ed. Y. Ishi and M. Tsutsui, Plenum Press, New York, 1978.
26 J. Coca and S. H. Langer, Chemtech, 1983, 13, 682.
27 M. Sardin, D. Schwich and J. Villermaux, in Preparative and
Production Scale Chromatography, ed. G. Ganetsos and P. E. Barker,
Chromatographic Science Series, vol. 61, Marcel Dekker, New York,
1993, p. 477.
28 S. H. Langer and J. E. Patton, J. Phys. Chem., 1972, 76, 2159.
29 C.-Y. Jeng and S. H. Langer, Ind. Eng. Chem. Res., 1991, 30, 1489.
30 R. S. Willms, A. M. Balinsky, D. D. Reible, D. M. Wetzel and
D. P. Harrison, Ind. Eng. Chem. Res., 1987, 26, 148.
31 The Activation of Dioxygen and Homogenous Catalytic Oxidation,
ed. D. H. R. Barton, A. E. Martell and D. T. Sawyer, Plenum Press,
New York, 1993.
1
0 A. Gaudemer, K. Nguyen-van-Duong, N. Shahkarami, S. S. Achi,
M. Frostin-Rio and D. Pujol, Tetrahedron, 1985, 41, 4095.
1 A. Zombeck, R. S. Drago, B. B. Corden and J. H. Gaul, J. Am.
Chem. Soc., 1981, 103, 7580; B. B. Corden, R. S. Drago and R. P.
Perito, J. Am. Chem. Soc., 1985, 107, 2903.
1
1
2 P. A. Ganeshpure, A. Sudalai and S. Satish, Tetrahedron Lett., 1989,
3
0, 5929.
1
1
3 A. Sorokin, J.-L. Séris and B. Meunier, Science, 1995, 268, 1163.
4 M. Frostin-Rio, D. Pujol, C. Bied-Charreton, M. Perree-Fauvet and
A. Gaudemer, J. Chem. Soc., Perkin Trans. 1, 1984, 1971.
5 Y. Aimoto, W. Kanda, S. Meguro, Y. Miyahara, H. Okawa and
S. Kida, Bull. Chem. Soc. Jpn., 1985, 58, 646.
6 V. Kothari, USP 3 935 247/1976 (Chem. Abstr., 1976, 84, 121 468b).
7 Preparative and Production Scale Chromatography, ed. G. Ganetsos
and P. E. Barker, Chromatographic Science Series, vol. 61, Marcel
Dekker, New York, 1993.
1
1
1
1
8 (a) C. Y. Jeng and S. H. Langer, J. Chromatogr. Sci., 1989, 27, 549;
(
b) C. Y. Jeng and S. H. Langer, J. Chromatogr., 1992, 589, 1.
1
9 Z. Q. Deng, J. Z. Zhang, A. B. Ellis and S. H. Langer, J. Liq.
Chromatogr., 1993, 16, 1083.
0 A. H. T. Chu and S. H. Langer, Anal. Chem., 1985, 57, 2197.
1 K. F. Scott and C. S. G. Phillips, J. Chem. Soc., Faraday I, 1980,
2
2
Paper 8/01253I
Received 3rd November 1997
Accepted 4th February 1998
7
6, 683; C. S. G. Phillips, A. J. Hart-Davis, R. G. L. Saul and
J. Wormald, J. Gas Chromatogr., 1967, 5, 424.
1
128
J. Chem. Soc., Perkin Trans. 2, 1998