The oxidation of ethylbenzene by dioxygen catalysed by supported iron porphyrins derived from iron(III) tetrakis(pentafluorophenyl)porphyrin

Article abstract of DOI:10.1039/b007326l

Chloromethylated poly(styrene) and chloropropylated silica gel have been reacted with α,ω-diamines to give diamine modified catalyst supports which have then been covalently attached to iron(III) tetrakis(pentafluorophenyl)porphyrin (FeTF₅PP).

Full text of DOI:10.1039/b007326l

View Article Online / Journal Homepage / Table of Contents for this issue
The oxidation of ethylbenzene by dioxygen catalysed by supported
iron porphyrins derived from iron(III) tetrakis(pentafluorophenyl)-
porphyrin
2
Steven Evans and John R. Lindsay Smith*
Department of Chemistry, University of York, York, UK YO10 5DD
Received (in Cambridge, UK) 11th September 2000, Accepted 24th November 2000
First published as an Advance Article on the web 21st December 2000
Chloromethylated poly(styrene) and chloropropylated silica gel have been reacted with α,ω-diamines to give diamine
modified catalyst supports which have then been covalently attached to iron() tetrakis(pentafluorophenyl)porphyrin
(FeTF5PP). In this way the diamino groups provide a simple linker/spacer unit between the support and the catalyst.
The diamino-silica has been further modified using trimethylsilyl chloride and by acetylation or acid washing. The
polarities of the modified inorganic and organic supports have been measured using Reichardt’s dye. Spectroscopic
studies reveal that with all the supported catalysts, except that on acid washed diamino-silica where the amines are
protonated, the diamino groups reduce the iron() porphyrin to iron(). The supported iron porphyrins have been
used to catalyse the oxidation of ethylbenzene by dioxygen. These reactions give the same three products, 1-phenyl-
ethyl hydroperoxide, 1-phenylethanol and acetophenone, as the analogous homogeneous oxidation using FeTF5PP,
suggesting that they proceed by the same mechanism, however, in general they are slower. The overall product yields
are limited by the stability/activity of the iron porphyrin and these in turn are very dependent on the length of the
linker, the catalyst loading and the microenvironment provided by the support. The role of the diamino groups in
the oxidations is discussed.
iron porphyrins.3 We chose FeTF5PP as the principal catalyst
Introduction
because it is readily available, it has been reported to be effective
The work of Ellis and Lyons1 and more recently that of
Gray and Labinger and their co-workers2 has identified the
potential of metalloporphyrin-catalysed oxidations of alkanes
by dioxygen which function at relatively low temperatures in
the absence of an external reducing agent. Although the
mechanism of these reactions has been a matter of some debate
over a number of years, it is now generally agreed that oxygen-
centred free radicals bring about the oxidations in these
systems. Thus, although oxometal species are present in the
reaction mixtures, their role is in one-electron redox reactions
and free radical chemistry rather than in two-electron oxygen-
transfer processes.
for alkane oxidation by dioxygen1f,2c and it can easily be
modified to generate a selection of related catalysts. A further
advantage is that it can also, by nucleophilic substitution of
the para-fluorines on the meso-pentafluorophenyls, be grafted
by covalent bonding to solid supports.4
The mechanism of the ethylbenzene oxidation was investi-
gated over the temperature range 30–110 ЊC. The reaction pro-
files and products obtained are wholly consistent with a typical
free radical autoxidation in which the redox chemistry of the
iron porphyrin species in the system catalyses the formation of
and maintains the flux of radical chain carriers derived from
the primary autoxidation product, 1-phenylethyl hydroperoxide
(Scheme 1). The overall yields of the products are determined
by the rate of the reaction and the reactivity and stability
of the catalyst towards oxidative degradation by oxyl, and in
In an earlier paper, we described our extensive studies on
the homogeneous oxidation of ethylbenzene and some other
alkylbenzenes with dioxygen catalysed by FeTF5PP and related
Scheme 1 Reactions of iron porphyrins with 1-phenylethyl hydroperoxide.
174
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b007326l

View Article Online
Table 1 Properties of supports and supported FeTF5PP catalysts
Support surface
modificationa/
mmol gϪ1
Catalyst
loadingb/
mg gϪ1
Fe
oxidation
state
Polarity
UV–Vis
Support
(E NT
)
λmax/nmc
Silica
0.91
0.65
0.69
0.69
0.69
0.63
0.61
—
0.26
0.64
0.64
—
0.60
0.60
—
—
—
—
—
0
—
—
324, 424
336, 422
422
322, 424
422
358, 418
—
—
430
350, 412
—
—
—
2
2
2
2
2
3
3-Chloropropylated silica
DA1,6-Si
DA1,6-Si
1.38
0.54
0.54
0.54
0.51
0.54
0.54
1.6
1.10
0.92
0.92
0.64
0.45
—
10
85
172
10
10
10
—
10
10
10
1.1
0.9
—
—
—
DA1,6-Si
Trimethylsilylated DA1,6-Si
Acetylated DA1,6-Si
Acid washed DA1,6-Si
Chloromethylated poly(styrene)
DA1,3-PS
DA1,6-PS
Acid washed DA1,6-PS
DA1,9-PS
—
—
2
3
—
—
2
3
3
DA1,12-PS
—
None (FeTF5PP–hexylamine)d
None (FeTF5PP)e
None (FeTF4NMe2PP)f
322, 420
350, 410
350, 418
—
—
a mmol of surface modification per g of silica or poly(styrene). b mg of FeTF5PP per g of support. c Spectra recorded as suspensions in CH2Cl2.
d Homogeneous solution of iron() porphyrin obtained by reduction of FeTF5PP with hexylamine in CH2Cl2. e Homogeneous solution of FeTF5PP
in CH2Cl2. f Homogeneous solution of FeTF4NMe2PP in CH2Cl2.
particular, alkoxyl radicals. Electron-withdrawing substituents,
as has been reported previously,1f enhance the rate of oxidation.
However, in general the least active catalysts give the highest
overall yields. The latter advantage clearly has to be weighed
against the disadvantages of long induction periods and
reaction times.
Although supported iron porphyrins have been widely
employed as catalysts in systems, using single oxygen donors,
that mimic the oxygen-transfer reactions of cytochrome
P450 monooxygenases and electron-transfer processes of
peroxidases,5 there are few reports of their use with dioxygen
as the terminal oxidant.6 The potential advantages of using a
solid catalyst include the ease of its removal from the oxidation
mixture and subsequent reuse, and control of its reactivity
through the microenvironment created by the support. This
paper reports our studies on ethylbenzene oxidation by
dioxygen using iron porphyrin catalysts, derived from FeTF5-
PP, covalently bound to silica gel and poly(styrene) via an
α,ω-diamine linker. The results are compared with those from
our earlier study using homogeneous systems.
Scheme 2 The covalent attachment of FeTF5PP to diamine-modifed
supports.
Iron() tetrakis(pentafluorophenyl)porphyrin (FeTF5PP)
was covalently bound to the diamine-modified inorganic and
organic supports by nucleophilic substitution of the para-
fluorines on the meso-pentafluorophenyl groups (Scheme 2).
This was achieved by heating a suspension of the support in a
solution of the iron porphyrin in diglyme. The resulting silica-
based catalysts were dark brown and those on poly(styrene)
were orange coloured. Some of the supported catalysts were
subjected to a further reaction (acetylation or treatment with
acid) to modify the unreacted surface amino groups.
Results and discussion
Supports and supported catalysts
The catalysts selected for this study were all derived from
FeTF5PP and were covalently bound via a spacer to the solid
supports. Covalent binding was used to ensure that the iron
porphyrin was firmly attached to the support to minimise
problems arising from catalyst leaching. α,ω-Diamines were
chosen as spacer/linker units since they provide a synthetically
simple means of varying the distance between the catalyst
and the support. It was hoped that by increasing the length of
the spacer it might be possible to move the metalloporphyrin
away from the surface so that it behaved more like a homo-
geneous catalyst in the bulk solution.
The catalyst-supports used in this study were derived from
chloropropylated silica gel (Si) and chloromethylated poly-
(styrene) (PS). These were reacted in p-xylene with the
α,ω-diamine spacer prior to attaching the iron porphyrin
(Scheme 2). For some experiments the chloropropylated silica
gel was reacted with trimethylchlorosilane, to remove unreacted
silanol groups, prior to the reaction with the diamine. The
extent of surface modification of the supports was obtained by
elemental analysis (Table 1).
The loadings of the iron porphyrin on the supports were
determined by using UV–Vis spectroscopy to measure the total
amount of unloaded FeTF5PP in the reaction solution and
washings from the work-up procedure (Table 1).
DRIFT spectroscopy is a simple method for the detection of
key functional groups on the support and provides evidence for
the surface modification in this study. Reaction of silica gel with
(3-chloropropyl)trimethoxysilane resulted in the disappearance
of a band at 983 cmϪ1 which has been assigned to highly
reactive, isolated Si–OH7 and the appearance of three peaks
at 2867, 2950 and 700 cmϪ1. The first two new peaks arise from
the C–H stretching frequencies of the propyl chain and the
last is characteristic of the C–Cl group. 1,6-Diaminohexane-
modification of the chloropropylated silica, to give DA1,6-Si, led
to a decrease in the intensity of the 700 cmϪ1 band and the
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180
175

View Article Online
appearance of two new bands at 3300 and 3367 cmϪ1 character-
istic of an N–H group. Acetylation of DA1,6-Si resulted in
the appearance of a new band at 1625 cmϪ1 from the amide
carbonyl group.
spectroscopy. The CT λmax values were used to calculate the
standardised ETN polarities (Table 1).12 The results show that,
although the polarities of support backbones, Kieselgel 60
silica and chloromethylated poly(styrene), are widely different
(ETN, 0.91 and 0.26, respectively; cf. water, 1.0 and tetramethyl-
silane, 0.0), all the diamine-modified supports have very similar
polarities (ETN, 0.60–0.69).
13C CP-MAS NMR analysis of the chloropropylated silica
gave peaks at δ 49.7 (CH3O-Si), 46.4 (CH2CH2-Cl), 26.8
(CH2CH2CH2) and 10.3 (Si-CH2CH2). These were assigned by
comparison with data reported by Vrancken et al.,8 from other
studies at York9 and from the solution 13C NMR spectrum
of (3-chloropropyl)trimethoxysilane. Further modification
with trimethylchlorosilane led to the appearance of a peak at
δ 0.0 from the Si–CH3 groups. However, because the diamine
loadings were very low (~0.5 mmol gϪ1), no clear peaks from
1,6-diaminohexane could be detected.
The high polarity of silica gel arises from the polar silanol
groups and surface-bound water molecules. Surface modifi-
cation with relatively non-polar amines and silanes (ETN values
in the range 0.2–0.3) would, as observed, be expected to
decrease the polarity of the surface. However, the relatively
high polarities of the modified poly(styrene)s were unexpected
since the polarities of both chloromethylated poly(styrene) and
the diamine are low (ETN < 0.3). This suggests that on DA1,6-Si
the polar terminal amino groups of the hexanediamine are
probably directed back towards the polar silica surface whereas
on DA1,6-PS they are directed away from the non-polar
poly(styrene) backbone. The net result is that the measured
polarities of both are very similar. Further modification of
the supports by acetylating the amino groups and trimethyl-
silylating the surface silanols led to small decreases in polarity.
We argued that an ideal support might be expected to have
a low polarity to favour the approach of the substrate to
the supported catalyst and to encourage the diffusion of the
products into the bulk solution. Unfortunately in this study
all attempts to generate a very non-polar support were
unsuccessful.
As a further check that the diamine was bound to the silica,
the chloropropylated silica and DA1,6-Si were each reacted with
FeTF5PP in diglyme. The catalyst only loaded on to the latter.
The supported iron porphyrins were characterised by UV–
Vis spectroscopy as suspensions in dichloromethane and the
data, together with those from selected iron porphyrin solution
spectra, are recorded in Table 1. The iron() porphyrins,
Fe(TF5PP)Cl and Fe(TF4NMe2PP)Cl3 in dichloromethane
each have a strong absorption at 350 nm and their Soret
peaks are at 410 and 418 nm respectively, whereas the iron()
tetrakis(pentafluorophenyl)porphyrin bis(hexylamine) complex
has bands at 322 and 420 nm. Based on these data, all the
catalysts on diamine-modified supports are iron() porphyrins.
This is not surprising since iron() porphyrins with electron-
withdrawing groups, such as FeTF5PP, are known to be readily
reduced by amines to give low spin bis-ligated iron() com-
pounds.9a,10 The mechanism for this process involves first
forming the low spin iron() complex, followed by an electron-
transfer from the ligand to give the iron() porphyrin and an
aminium radical cation (Scheme 3).10 With the supported cat-
Ethylbenzene oxidation
The method. The oxidations were carried out at 100 ЊC,
under one atmosphere of dioxygen, with ethylbenzene acting
as both the substrate and the reaction medium. The three
major products obtained from all the oxidations were the
same as those observed in the homogeneous iron() porphyrin-
catalysed reactions, namely 1-phenylethyl hydroperoxide, 1-
phenylethanol and acetophenone. The hydroperoxide was the
main product in each oxidation except those using the more
active catalyst, acetylated DA1,6-Si-FeTF4PP, where aceto-
phenone predominated.
As described in our earlier paper,3 the hydroperoxide is
thermolysed during GC analysis to acetophenone and its yield
has to be measured indirectly making use of the quantitative
reduction of hydroperoxides to alcohols by triphenylphos-
phine.13 Thus GC analysis of the reaction mixtures before and
after PPh3 treatment allowed the yields of all three products
to be determined. A typical reaction product time profile
is illustrated in Fig. 1. This shows an induction period, as
was observed in low temperature homogeneous reactions.
The reaction comes to a halt when all the catalyst has been
destroyed. However, unlike the equivalent reaction profiles
from the homogeneous oxidations, the disappearance of the
iron porphyrin is not shown in Fig. 1. The absorbances from
the UV–Vis spectra of the catalyst suspensions are very variable
so that, although the spectra provide λmax values, they cannot
easily be used to obtain quantitative data on the concentrations
of the heterogeneous species present.
Scheme 3 The reduction of an iron() porphyrin by a primary amine.
alysts, some of the large excess of amino groups on the surface
can presumably ligate to the iron() porphyrins enabling
the electron-transfer to occur. To confirm this conclusion, the
supported catalysts were washed with acid to protonate the
surface amino groups and disrupt any ligation; this brought
about the reoxidation of the catalyst by air to give back the
iron() porphyrin. Interestingly acetylation of DA1,6-Si-FeTF4-
PP with acetic anhydride did not bring about the reoxidation
of the iron() porphyrin. It is possible that either the ligated
amino groups are not acetylated or that the amido groups
are capable of stabilising the iron() porphyrin sufficiently to
prevent reoxidation.
Optimum length of diamine linker. Four α,ω-diamine linkers,
with 3, 6, 9 and 12 carbon atoms, were examined with poly-
(styrene) as the support (Table 2). Increasing the length of the
linker led to a >50% decrease in the amount of diamine on the
support and to a ten-fold decrease in the FeTF5PP that could be
covalently attached to DA1,9-PS and DA1,12-PS. We attribute
this to an increase in back-biting, where both amino groups
react with the surface chloromethyl groups, with the long chain
linkers. As a consequence fewer diamines react with the support
and significantly fewer primary amino groups are available to
bind the catalyst.
The polarities (ETN values) of the supports were measured
using the method of Tavener et al.11 This involved adsorb-
ing 2,6-diphenyl-4-(2,4,6-triphenylpyridinium-1-yl)phenolate
(Reichardt’s dye),12 which has a polarity-dependent charge-
transfer (CT) band in its visible spectrum, on to the support
and measuring the CT λmax by diffuse reflectance UV–Vis
176
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180

View Article Online
Table 2 Influence of diamine linker and support backbone on catalyst
activity/stability in ethylbenzene oxidations at 100 ЊC
Yield
Induction
period/h
Catalyst
lifetime/h
(catalyst
Catalyst
turnovers)a
DA1,3-PS-FeTF4PP
DA1,6-PS-FeTF4PP
DA1,9-PS-FeTF4PP
DA1,12-PS-FeTF4PP
DA1,6-Si-FeTF4PP
FeTF5PP
55
5
3
1.5
60
0
20
16.5
26
23
15
1.5
20
4200
7400
250
350
4300
2900
14000
FeTF4NMe2PP
0
a Total yield of products, 1-phenylethyl hydroperoxide, 1-phenylethanol
and acetophenone, as catalyst turnovers.
Fig. 2 The effect of acid washing, trimethylsilylation and acetylation
of DA1,6-SiFeTF4PP on (a) catalyst lifetime and reaction induction
period, and (b) product turnovers, in the oxidation of ethylbenzene with
dioxygen at 100 ЊC.
Fig. 1 Product turnover versus time in the oxidation of ethylbenzene
with dioxygen catalysed by DA1,6-Si-FeTF4PP at 100 ЊC; ᭿ aceto-
phenone, ᭡ 1-phenylethyl hydroperoxide, ᭹ 1-phenylethanol.
Acetylation of the excess amino groups on DA1,6-Si-
FeTF4PP had a dramatic effect on the catalyst’s activity. The
induction period was eliminated and the rate of oxidation
increased, however, the catalyst’s lifetime and the overall
product yield (1300 turnovers) were reduced (Fig. 2). The
reaction profile became very similar to that observed in the
homogeneous oxidation of ethylbenzene using FeTF5PP,
showing that the excess amino groups are clearly responsible for
the long induction period with the catalyst on the unacylated
support. It was hoped, by analogy with the homogeneous
reactions of FeTF5PP,3 that lowering the reaction temperature
would improve the performance of the catalyst on acetylated
DA1,6-Si. However, although it resulted in the expected increase
in the induction period and catalyst lifetime, these were not
accompanied by an increase in product yield (Fig. 3).
Acid treatment of the catalyst supported on DA1,6-Si,
interestingly, had little effect on the induction period or
catalyst lifetime although it led to a significant improvement in
activity and consequently to an increase in the overall turnovers
(7500 vs. 4300) (Fig. 2).
Trimethylsilylation of the surface silanols on DA1,6-Si-
FeTF4PP reduces the polarity of the support without affecting
the surface amino groups. This capping procedure resulted in a
decrease in the induction period for the ethylbenzene oxidation
and an increase in yield of products (7500 vs. 4300 turnovers)
(Fig. 2).
The effect of changing the catalyst loading on the oxidation
of ethylbenzene was investigated using four catalyst loadings
(5, 10, 85 and 172 mg gϪ1) on DA1,6-Si. Assuming that each
iron porphyrin is attached to the surface by only one diamine
linker, these loadings correspond to 114-, 57-, 7- and 3-fold
excesses of diamine over iron porphyrin, respectively (Fig. 4).
The lowest loaded catalyst was inactive under the conditions
used, presumably because the large excess of diamine suppressed
the oxidation (see below). Increasing the loading (reducing the
excess of diamine) resulted in a decrease in both the induction
period and the catalyst’s lifetime. The overall product yield
In the oxidation of ethylbenzene, increasing the linker length
led to a large reduction in the induction period, possibly
because, as suggested by Montanari and co-workers,14 the
catalyst’s environment becomes more like the bulk solution
than that of the support. The equivalent homogeneous
reactions of FeTF5PP and FeTF4NMe2PP at 100 ЊC do not
show a detectable induction period.3 However, the highest
oxidation yields were obtained using the DA1,6-PS-FeTF4PP
(7400 turnovers) which was much more efficient than those
using the 1,9- and 1,12-diamine linkers (250 and 350 turnovers,
respectively). This may be the result of the low loading of
the catalyst (see below). DA1,3-PS-FeTF4PP gave a very long
induction period and the overall oxidation yield (4200 turn-
overs) was lower than that of DA1,6-PS-FeTF4PP. Based on
these results, the 1,6-diaminohexane linker was selected for all
the subsequent studies described below.
Comparisons of oxidations with heterogeneous and homo-
geneous catalysts. In all the reactions, the overall oxidation
yields (total catalyst turnovers) were limited by the destruction
of the iron porphyrin. As discussed previously, the catalyst
degradation is brought about by oxygen-centred radicals,
formed during the oxidation of ethylbenzene, reacting with the
porphyrin ligand.3 In general, the heterogeneous oxidations
at 100 ЊC showed moderate to long induction periods whilst,
with the homogeneous systems, these were only detectable at
temperatures <80 ЊC. Furthermore the rates of the former
systems, once initiated, tended to be lower than those of the
analogous homogeneous oxidations using iron() tetrakis-
(4-dimethylamino-2,3,5,6-fluorophenyl)porphyrin (FeTF4N-
Me2PP) or FeTF5PP.
To investigate the influence of the support in these catalysed
oxidations, we made a series of changes to the surface groups
and the support backbone and compared the performances of
the resulting catalysts.
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180
177

View Article Online
of oxidation products (4300 and 7400 turnovers, respectively)
(Table 2). We interpret these effects as arising from the involve-
ment of the tertiary benzylic C–H bonds of the poly(styrene)
backbone in these reactions.15 These C–H bonds should be
very prone to radical autoxidation, which would in turn shorten
the induction period and, by helping to maintain the flux
of radicals, would increase the yield of oxidation products.
Overall the oxidation of the poly(styrene) makes it an unsuit-
able support for these oxidations since ultimately it will result in
the degradation of the polymer backbone.
The polarities of the two supports are very comparable
(ETN: DA1,6-Si, 0.69 and DA1,6-PS, 0.64) but the organic support
was only lightly cross-linked (2% divinylbenzene) and con-
sequently it swells in ethylbenzene and is very flexible compared
with the rigid silica-based material. The support flexibility
may result in a more accessible and reactive catalyst. This
difference in flexibility of the two supports is also evident
from the brown and orange colours of DA1,6-Si-FeTF4PP and
DA1,6-PS-FeTF4PP, respectively, the latter is typical of an
iron() porphyrin diamine complex. As noted previously,
flexible organic polymers accommodate the formation of the
diamine complexes much more readily than rigid inorganic
supports.5b,16
During the early stages of the reactions the orange colour
of DA1,6-PS-FeTF4PP changes to dark brown whereas DA1,6
-
Si-FeTF4PP starts and remains brown during the course of
the oxidation. This we attribute to the formation of iron()
porphyrin from the initial bis-ligated iron() species on the
organic support. Acid washing of DA1,6-PS-FeTF4PP, which
would be expected to give the supported iron() complex
(see above), brings about an identical colour change and the
UV–Vis spectrum shows the characteristic absorptions of an
iron() porphyrin.
Fig. 3 The effect of temperature on (a) catalyst lifetime and reaction
induction period, and (b) product turnovers, in the oxidation of ethyl-
benzene with dioxygen catalysed by acetylated DA1,6-SiFeTF4PP.
Possible reasons for the differences between catalysts on
supports and in solution include:
(i) the excess amines on the support’s surface which lead to
the formation of mono- and bis-ligated iron() porphyrins; (ii)
the amines on the support surface participating in the radical
oxidation. They could either have a beneficial effect by protect-
ing the catalyst from oxidative destruction, as observed with the
NMe2 groups in homogeneous oxidations catalysed by FeTF4-
NMe2PP,3 or have a deleterious effect acting as inhibitors by
reacting with the autoxidation chain carriers; (iii) the polarity
of the catalyst’s microenvironment arising from the support’s
backbone, which in turn affects rates of diffusion of reactants,
intermediates and products to and from surface-bound
catalysts.
To investigate further the role of the amino groups in these
oxidations, the effect of adding 1,3-diaminopropane, DA1,6
-
Si and acid washed DA1,6-Si to the standard homogeneous
oxidation of ethylbenzene with FeTF5PP at 70 ЊC was studied.3
In each reaction the amount of diamine used was equivalent to
that present on DA1,6-Si-FeTF4PP (catalyst loading, 10 mg gϪ1
,
i.e. 57-fold excess of diamine over FeTF5PP). The presence
of 1,3-diaminopropane led to the rapid formation (<1 min)
of bis-ligated iron() porphyrin and the reaction was totally
inhibited. The influence of added DA1,6-Si was much less
dramatic, although it resulted in a significant decrease in
product yields, and acid treated DA1,6-Si was effectively inert
(Fig. 5). These results show that free amino groups clearly inter-
fere with the FeTF5PP-catalysed oxidation of ethylbenzene by
dioxygen; the possible causes being points (i) and (ii) above, the
inactivation of the catalyst as a bis-ligated complex and/or the
diamine acting as an autoxidation inhibitor.
The above results show that the amino groups have a large
effect on the performance of the iron porphyrin catalyst in these
oxidations and it is likely that the activity of DA1,6-Si-FeTF4PP
is markedly affected by the excess diamine groups on the
support. The induction period, rate of oxidation and catalyst
stability are all sensitive to a complex interplay of effects arising
Fig. 4 The effect of catalyst loading on (a) catalyst lifetime and
reaction induction period, and (b) product turnovers, in the oxidation
of ethylbenzene with dioxygen catalysed by DA1,6-SiFeTF4PP at 100 ЊC.
from the oxidation of ethylbenzene was also sensitive to the
catalyst loading, increasing to a maximum (4900 turnovers)
with 85 mg gϪ1 of catalyst. It is likely that, at the highest
loading, inter-porphyrin interactions are significant and these
reduce the oxidation yield by encouraging oxidative self-
destruction of the catalyst.
Changing the support from DA1,6-Si to DA1,6-PS resulted in a
shorter induction period and, surprisingly, a higher overall yield
178
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180

View Article Online
column (UGI) (Phase Sep) and screened by GC to ensure the
absence of autoxidation products.
Iron() tetrakis(pentafluorophenyl)porphyrin was prepared
as described previously.3
Support materials
3-Chloropropylated silica gel was prepared by refluxing a
suspension of silica gel (13 g, Kieselgel 60, 230–400 mesh) in
a solution of freshly distilled 3-chloropropyltrimethoxysilane
(12 cm3, 0.065 mol) in toluene (100 cm3). After 14 h, the mixture
was cooled and filtered to give a white solid that was washed
sequentially with toluene (50 cm3), acetone (24 cm3) and water
(15 cm3) and dried in a vacuum pistol (100 ЊC, 1 mmHg, 2.5 h).
DRIFT, νmax/cmϪ1 (KBr) 700, 800, 1100, 1633, 1850, 3500
(the Si–OH stretching frequency at 967 cmϪ1 of the silica gel
was absent following surface modification); 13C CP-MAS NMR
δ 10.3, 26.8, 46.4 (CH2CH2CH3), 49.7 (MeO); CH analysis
(C 6.58 and H 1.41%) gives a loading of 1.38 mmol gϪ1
.
Trimethylsilylation of 3-chloropropylsilica gel was achieved
by refluxing the modified silica (5.31 g) in trimethylchlorosilane
under argon. After 4 h the mixture was cooled and the solvent
was removed under vacuum. The white residue was washed
with deionised water until the washings were neutral and then
dried under vacuum (100 ЊC, <0.1 mmHg, 2.5 h). 13C CP-MAS
NMR δ 0.0 (SiMe3), 10.3, 26.5, 45.9 (CH2CH2CH3), 49.9
(MeO); CH analysis (C 7.55 and H 1.67%) gives a loading of
Fig. 5 The effect, on product turnovers, of adding 1,3-diamino-
propane, DA1,6-Si or acid washed DA1,6-Si to the homogeneous
FeTF5PP-catalysed oxidation of ethylbenzene by dioxygen at 70 ЊC.
SiMe3 groups of 0.25 mmol gϪ1
.
Synthesis of 1,6-diaminohexane-modified silica gel (DA1,6-Si)
was carried out by heating a suspension of chloropropylated
silica (1 g) in a solution of 1,6-diaminohexane (5.9 g, 46 mmol)
in p-xylene (30 cm3) under nitrogen at 75 ЊC. After 3.75 h,
the mixture was cooled, filtered and washed with p-xylene
(30 cm3), acetone (30 cm3) and water (until the washings were
neutral) and dried under vacuum (100 ЊC, <0.1 mmHg, 2.5 h).
DRIFT, νmax/cmϪ1 (KBr) 800, 1100, 1450, 1633, 1850, 2867,
2950, 3300, 3367; the nitrogen content from CHN analysis
(C 8.62, H 2.06 and N 1.48%) gives a diamine loading of 0.54
from the surface groups and the polarity provided by the
support’s microenvironment. Thus surface amines affect the
reactivity of the iron porphyrin by mono- and bis-ligation, by
acting as competitive substrates and by behaving as anti-
oxidants and these effects can be moderated or eliminated by
protonation or acetylation. The surface amines, ammonium
ions, amides and trimethylsilyl groups also exert their influence
on the reactions by modifying the support’s polarity: this will
affect the rates of the reactions by controlling the diffusion of
substrate, intermediates and products between the solution and
solid phases.
mmol gϪ1
.
Trimethylsilylated 1,6-diaminohexane-modified silica gel
was obtained by the method above using trimethylsilylated
3-chloropropylsilica. The nitrogen content from CHN analysis
(C 8.25, H 1.98 and N 1.44%) gives a diaminohexane loading
Experimental
Instrumental methods
of 0.51 mmol gϪ1
.
The diamine-modified poly(styrene)s were prepared from 2%
cross-linked chloromethylated poly(styrene) (5.0 g) and diamine
(46 mmol) in p-xylene by the method described above for
DA1,6-Si. After filtration, the product was washed with p-xylene
(90 cm3), methanol (90 cm3) and water (until the washings were
neutral) and dried under vacuum. The nitrogen content from
CHN analyses gave the diamine loadings (C 85.53, H 7.92 and
N 3.08%) for DA1,3-Ps, 1.10 mmol gϪ1; (C 86.43, H 7.96 and N
2.58%) for DA1,6-Ps, 0.92 mmol gϪ1; (C 84.94, H 7.87 and
N 1.80%) for DA1,9-Ps, 0.64 mmol gϪ1 and (C 86.38, H 7.93
UV–Vis spectra of solutions and suspensions were recorded
on a Hewlett Packard 8452A diode array spectrometer using
1 cm quartz cuvettes. Diffuse reflectance UV–Vis spectra were
recorded on a Perkin-Elmer Lambda 15 instrument using a
1
1 mm path length quartz cuvette. H and 19F NMR spectra of
solutions were recorded on a JEOL JNM-EX270 spectrometer
using TMS and CFCl3 as the respective internal standards.
13C CP-MAS NMR spectra of supports were obtained using a
Bruker MSL 300 MHz spectrometer using adamantane as the
external standard. Diffuse reflectance IR Fourier transform
(DRIFT) spectra were recorded at ambient temperature and
pressure on a Perkin-Elmer 1720 infrared spectrometer. Gas
chromatography was carried out with a glass column (3 m ×
2.5 mm) packed with 10% w/w carbowax 20M on GasChrom Q
(80–120 mesh) in a Pye Unicam GCD chromatograph equipped
with a flame ionisation detector. The results were analysed with
a JCL6000 for windows programme (Jones Chromatography
Ltd.) using 1,3-dichlorobenzene as the internal standard.
C,H,N Elemental analysis was performed by Butterworths
Laboratories Ltd.
and N 1.25%) for DA1,12-Ps, 0.45 mmol gϪ1
.
A typical procedure for loading the catalyst on to diamine-
modified supports involved heating a suspension of the support
(500 mg) in a solution of FeTF5PP (5 mg) in freshly distilled
diglyme (40 cm3) under nitrogen at 135 ЊC. After 3.25 h the
mixture was cooled, filtered and washed with methanol (50 cm3)
followed by dichloromethane (50 cm3). Finally the solid was
washed, for 13 h, with dichloromethane in a Soxhlet extractor
and dried (100 ЊC, <0.1 mmHg, 2.5 h).
Acetylation of excess amino groups on DA1,6-Si-FeTF4PP
required heating a suspension of the supported catalyst
(451 mg) in freshly distilled acetic anhydride (40 cm3) at 90 ЊC.
After 20.5 h the reaction was cooled and the suspension was
filtered and washed consecutively with water (until the washings
were neutral), methanol (40 cm3) and dichloromethane (40 cm3)
before it was subjected to Soxhlet extraction overnight with
Reagents
All the reagents were commercially available and used without
purification unless otherwise stated. All the alkyl aromatic
substrates were passed through a short activated alumina
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180
179

View Article Online
and H. B. Gray, J. Mol. Catal. A: Chem., 1995, 104, 119; (d) E. R.
Birnbaum, M. W. Grinstaff, W. P. Schaeffer, J. A. Labinger, M. G.
Hill and H. B. Gray, Inorg. Chem., 1995, 34, 4896; (e) A. Bottcher,
E. A. Birnbaum, M. W. Day, H. B. Gray, M. W. Grinstaff and
J. A. Labinger, J. Mol. Catal. A: Chem., 1997, 117, 229.
3 S. Evans and J. R. Lindsay Smith, J. Chem. Soc., Perkin Trans. 2,
2000, 1541.
dichloromethane. Drying under vacuum (100 ЊC, <0.1 mmHg,
2.5 h) gave a brown solid. DRIFT νmax/cmϪ1 (KBr) 800, 1100,
1625, 1633, 1850, 2867, 2950.
Measurement of support polarities
The procedure used to measure support polarities was that of
Tavener et al.11 and involved suspending the support (300–400
mg) in a solution of 2,6-diphenyl-4-(2,4,6-triphenylpyridinium-
1-yl)phenolate (Reichardt’s dye) in dichloromethane (20 cm3).
The solvent was removed under vacuum and the process was
repeated until the support was visibly coloured, at which point
the solid was dried under vacuum for 30 min to remove the
dichloromethane. A diffuse reflectance UV (DRUV) spectrum
of the solid, relative to that of a non-dye coated sample,
was recorded and the measured λmax was used to calculate the
polarity (ETN value).
4 P. Battioni, J. F. Bartolli, D. Mansuy, Y. S. Byun and T. G. Traylor,
J. Chem. Soc., Chem. Commun., 1991, 1051.
5 (a) J. R. Lindsay Smith, in Metalloporphyrins in Catalytic
Oxidations, ed. R. A. Sheldon, Marcel Dekker, New York, 1994,
ch. 11; (b) C. Gilmartin and J. R. Lindsay Smith, J. Chem. Soc.,
Perkin Trans. 2, 1995, 243; (c) P. Battioni, E. Cardin, M. Louloudi,
G. Schollhorn, G. A. Spyroullias, D. Mansuy and T. G. Traylor,
J. Chem. Soc., Chem. Commun., 1996, 2037; (d) P. Battioni, R.
Iwanejko, D. Mansuy, T. Mlodnicka, J. Poltowicz and F. Sanchez,
J. Mol. Catal. A: Chem., 1996, 109, 91; (e) H. S. Hilal, W. Jondi,
S. Khalaf, A. Keilani, M. Suleiman and A. F. Schreiner, J. Mol.
Catal. A: Chem., 1996, 113, 35; ( f ) M. A. Martinez-Lorente,
P. Battioni, W. Kleemiss, J. F. Bartoli and D. Mansuy, J. Mol. Catal.
A: Chem., 1996, 113, 343; (g) B.-Z. Zhan and X.-Y. Li, Chem.
Commun., 1998, 349; (h) M. D. Assis and J. R. Lindsay Smith,
J. Chem. Soc., Perkin Trans. 2, 1998, 2221; (i) C.-J. Liu, W.-Y. Yu,
S.-G. Li and C.-M. Che, J. Org. Chem., 1998, 63, 7364; (j) B. T.
Holland, C. Walkup and A. Stein, J. Phys. Chem., 1998, 102, 4301;
(k) P. E. F. Neys, I. F. J. Vankelecom, M. L’abbè, P. F. Parton,
E. Cenlemans, W. Dehaen, G. L’abbè and P. A. Jacobs, J. Mol.
Catal. A: Chem., 1998, 134, 209.
6 T. M. Nenoff, M. C. Showalter and K. A. Salaz, J. Mol. Catal. A:
Chem., 1997, 121, 123.
7 Spectroscopic characterisation of heterogeneous catalysts. Part A:
Methods of surface analysis, ed. J. L. G. Fiero, Elsevier, Oxford,
1990.
8 K. C. Vrancken, P. Van der Voort, I. Gillis-D’Hamers, E. F. Vansant
and P. Groget, J. Chem. Soc., Faraday Trans., 1992, 88, 3197.
9 (a) P. R. Cooke, C. Gilmartin, G. W. Gray and J. R. Lindsay Smith,
J. Chem. Soc., Perkin Trans. 2, 1995, 1573; (b) B. R. Sorrell, DPhil
Thesis, University of York, 1997.
10 K. M. Kadish, in Iron Porphyrins, Part 2, ed. A. B. P. Lever and
H. B. Gray, Addison-Wesley, 1983, p. 161.
Oxidation methods
In a typical oxidation, the supported iron porphyrin (6.6 ×
10Ϫ7 mol) was mixed with 1,3-dichlorobenzene (0.25 cm3) and
the alkyl aromatic (25 cm3) (for homogeneous reactions the iron
porphyrin was dissolved in the dichlorobenzene) and analysed
by GC prior to being heated in a slow stream of dioxygen at the
required temperature. The course of the reaction was followed
by removing small samples for GC analysis (0.4 cm3). These
were analysed twice, before and after treatment with PPh3.
The latter involved adding PPh3 (~30 mg) to the sample at
room temperature and stirring for a few minutes before GC
analysis.
Acknowledgements
S. E. thanks the EPSRC for a research studentship.
11 S. J. Tavener, J. H. Clark, G. W. Gray, P. A. Heath and D. J.
Macquarrie, Chem. Commun., 1997, 1147.
12 C. Reichardt, Chem. Rev., 1994, 94, 2319.
References
1 (a) P. E. Ellis and J. E. Lyons, J. Chem. Soc., Chem. Commun.,
1989, 1315; (b) J. E. Lyons, Coord. Chem. Rev., 1990, 105, 181;
(c) J. E. Lyons, P. E. Ellis and V. A. Durante, Stud. Surf. Sci.
Catal., 1991, 67, 99; (d) J. E. Lyons, P. E. Ellis, H. K. Myers and
R. W. Wagner, J. Catal., 1993, 141, 311; (e) H. L. Chen, P. E. Ellis,
T. Wijeskera, T. E. Hagan, S. E. Groh, J. E. Lyons and D. P. Ridge,
J. Am. Chem. Soc., 1994, 116, 1086; ( f ) J. E. Lyons, P. E. Ellis and
H. K. Myers, J. Catal., 1995, 155, 59.
13 A. G. Rowley, in Organophosphorus Reagents in Organic Chemistry,
ed. J. I. G. Cadogan, Academic Press, London, 1979, ch. 7;
R. K. Jensen, S. Korcek, L. R. Mahoney and M. Zimbo, J. Am.
Chem. Soc., 1979, 101, 7574.
14 M. Cinouini, S. Colonna, H. Molinari, F. Montanari and P. Tundo,
J. Chem. Soc., Chem. Commun., 1976, 394; H. Molinari,
F. Montanari and P. Tundo, J. Chem. Soc., Chem. Commun., 1977,
639.
2 (a) J. A. Labinger, Catal. Lett., 1994, 24, 95; (b) M. W. Grinstaff,
M. G. Hill, J. A. Labinger and H. B. Gray, Science, 1994, 264, 1311;
(c) E. R. Birnbaum, M. W. Grinstaff, J. A. Labinger, J. E. Bercaw
15 We thank one of the referees for this suggestion.
16 P. R. Cooke and J. R. Lindsay Smith, J. Chem. Soc., Perkin Trans. 2,
1994, 1913.
180
J. Chem. Soc., Perkin Trans. 2, 2001, 174–180