Catalysis by microporous phthalocyanine and porphyrin network polymers
Helen J. Mackintosh,
a
Peter M. Budd*
a
and Neil B. McKeown*
b
Received 10th October 2007, Accepted 21st November 2007
First published as an Advance Article on the web 10th December 2007
DOI: 10.1039/b715660j
Cobalt phthalocyanine and iron porphyrin network polymers of intrinsic microporosity (network-
PIMs) were prepared and their performance as heterogeneous catalysts compared with that of low
molar mass analogues. Spiro-linked Co phthalocyanine network-PIMs prepared from preformed
chlorinated phthalocyanines showed lower surface areas and lower catalytic activity than those
prepared by a phthalocyanine-forming reaction from a rigid precursor incorporating a spiro-
centre. However, all the phthalocyanine network-PIMs were much more effective catalysts than
low molar mass Co phthalocyanine for the decomposition of hydrogen peroxide, the oxidation of
cyclohexene and the oxidation of hydroquinone. An Fe porphyrin network-PIM showed a higher
surface area than any of the phthalocyanine polymers and showed higher activity for the
oxidation of hydroquinone, also outperforming a low molar mass FeCl porphyrin.
Introduction
Metalloporphyrins and related compounds facilitate many
important biological processes (e.g., chlorophyll for photo-
synthesis, hemoglobin for oxygen transport, cytochrome for
electron transport, peroxidase for oxidation, catalase for
hydrogen peroxide decomposition).
1–3
This has prompted
considerable research into porphyrins and similar aromatic
macrocycles as catalysts. Phthalocyanines are more easily
synthesized in high yield than porphyrins, and will form
complexes with more than seventy different metal ions.
4
Metallophthalocyanines are used, for example, as catalysts in
the Merox process for the industrial desulfurization of
petroleum.
5–7
Whilst the potential of phthalocyanines and
porphyrins as homogeneous catalysts has been well demon-
strated, there is particular interest in incorporating the
macrocycles into heterogeneous systems that offer easier
recovery and recycling as well as minimizing problems of
degradation and deactivation of the catalyst. Thus, they have
been encapsulated in zeolites,
8,9
immobilized in mesoporous
materials
10
and supported on polymers.
11,12
A key requirement
is to inhibit the tendency for the planar macrocycles to self-
associate. We therefore sought to construct network polymers
incorporating phthalocyanines
13
and porphyrins,
14
in which
the macrocycles were interlinked by rigid non-planar units,
forcing them to point in different directions and creating free
volume to allow access by small molecules. A spiro-centre was
utilized as a site of contortion within the linker. The products
are amorphous, glassy materials that exhibit high apparent
surface areas (500–1000 m
2
g
21
) by nitrogen adsorption at
77 K. A study of the adsorption of nitrogen and of various
organic probe molecules by a spiro-linked phthalocyanine
network polymer, and comparison with an activated carbon,
demonstrated that it behaves in many respects like a micro-
porous material as defined by IUPAC
15
(i.e., possessing pores
of dimensions , 2 nm).
16
Following on from our initial work
on phthalocyanine and porphyrin network polymers, we
developed a range of polymers of intrinsic microporosity
(PIMs), including network polymers incorporating various
catalytic centres or binding sites, and non-network, soluble
polymers that can be cast to generate membranes or be
processed into other useful forms.
17–23
This paper describes work aimed at validating the concept
that a spiro-linked phthalocyanine or porphyrin network
polymer should be more effective as a heterogeneous catalyst
than the low molar mass analogue. In our initial work,
phthalocyanine network polymers were prepared by a phtha-
locyanine-forming reaction utilizing a rigid bis(phthalonitrile)
precursor 3, formed from 5,59,6,69-tetrahydroxy-3,3,39,39-
tetramethyl-1,19-spirobisindane 1 and 4,5-dichlorophthalo-
nitrile 2 (Scheme 1).
13
Porphyrin network polymers were
prepared by a dibenzodioxane-forming reaction (a double
aromatic nucleophilic substitution) between 1 and a preformed
fluorinated porphyrin 5 (Scheme 2).
14
Although the porphyrin
network does not consist entirely of fused ring structures,
as the phthalocyanine network does, there is sufficiently
restricted rotation about the single carbon–carbon bonds at
the meso positions of the porphyrin to prohibit structural
relaxation of the network and maintain an open structure.
In the present work, phthalocyanine network polymers were
prepared both by the initial phthalocyanine-forming route
(samples CoPc-PIM-A1, CoPc-PIM-A2 and CoPc-PIM-A3)
and by dibenzodioxane-forming reactions with preformed
chlorinated phthalocyanines (Scheme 3) (samples CoPc-
PIM-B1 and CoPc-PIM-B2). Three test reactions were utilized
in catalysis studies, as outlined below.
Decomposition of hydrogen peroxide
The catalase-like activity of phthalocyanines for hydrogen
peroxide decomposition was identified as long ago as 1938.
24
a
Organic Materials Innovation Centre, School of Chemistry, University
of Manchester, Manchester, UK M13 9PL.
E-mail: Peter.Budd@manchester.ac.uk; Fax: +44 (0)161-275-4273;
Tel: +44 (0)161-275-4711
b
School of Chemistry, Cardiff University, Cardiff, UK CF10 3AT.
E-mail: mckeownnb@cardiff.ac.uk; Fax: +44 (0)2920-874030;
Tel: +44 (0)2920-875851
PAPER www.rsc.org/materials | Journal of Materials Chemistry
This journal is
ß
The Royal Society of Chemistry 2008
J. Mater. Chem.
, 2008, 18, 573–578 | 573
Downloaded by Cardiff University on 02 April 2012
Published on 10 December 2007 on http://pubs.rsc.org | doi:10.1039/B715660J
View Online
/ Journal Homepage
/ Table of Contents for this issue
Fe phthalocyanines are generally the most active, but Co
phthalocyanines also show some activity.
25
Homogeneous
phthalocyanine catalysts are liable to decompose during the
reaction, but the stability is improved when they are anchored
to a polymer chain.
11
In the present work, the performance of
low molar mass Co phthalocyanine as a heterogeneous catalyst
is compared with Co phthalocyanine network-PIMs prepared
by different routes. A gas burette was utilized to monitor O
2
evolved during H
2
O
2
decomposition.
Oxidation of cyclohexene
Phthalocyanines have been widely studied as catalysts for the
oxidation of alkenes with various oxidants.
8,9,26,27
Here,
t-butylhydroperoxide is used for oxidation of cyclohexene
and, as above, the performance of low molar mass Co
phthalocyanine as a heterogeneous catalyst is compared
with Co phthalocyanine network-PIMs prepared by different
routes. Gas chromatography-mass spectrometry (GC-MS) was
utilised to identify and quantify the reaction products.
Oxidation of hydroquinone
Metallophthalocyanines and metalloporphyrins have been
found to catalyse the oxidation of phenols to quinones by
molecular oxygen under mild conditions and use has been
made of this in multi-step electron transfer systems.
28,29
Here,
we investigate the oxidation of hydroquinone to benzoquinone
and compare the performance of low molar mass Co and Fe
phthalocyanines and a fluorinated FeCl porphyrin with a
Co phthalocyanine network-PIM (CoPc-PIM-A2) and an Fe
porphyrin network-PIM (FePorph-PIM). A gas burette
was utilized to monitor O
2
consumed during hydroquinone
oxidation.
Experimental
General methods
Apparent surface areas were determined by the multi-point
Brunauer–Emmitt–Teller (BET) method from nitrogen
adsorption measurements at 77 K using a Coulter SA3100
instrument. UV–visible spectra were recorded using a Cary 1E
Scheme 1 Preparation of spiro-linked cobalt phthalocyanine network
polymer (CoPc-PIM-A) utilizing a phthalocyanine-forming reaction.
Reagents and conditions: (i) K
2
CO
3
, DMF, 70 uC; (ii) Co(CH
3
COO)
2
,
quinoline, 220 uC.
Scheme 2 Preparation of spiro-linked iron porphyrin network poly-
mer (FePorph-PIM). Reagents and conditions: (i) K
2
CO
3
,NMP,
170 uC, 5 h; (ii) FeCl
3
, NMP, 120 uC, 24 h.
574 |
J. Mater. Chem.
, 2008, 18, 573–578 This journal is
ß
The Royal Society of Chemistry 2008
Downloaded by Cardiff University on 02 April 2012
Published on 10 December 2007 on http://pubs.rsc.org | doi:10.1039/B715660J
View Online
spectrometer. Analysis of C, H and N was carried out with a
Carlo Erba EA1101 elemental analyser, of Co and Fe with a
Fisons Horizon ICP, of Cl (following oxygen flask combus-
tion) with a Metrohm 686 titroprocessor and of F (following
oxygen flask combustion) with a Cecil Instruments CE292 UV
spectrophotometer.
Co phthalocyanine polymers
Three samples of Co phthalocyanine network-PIM (designated
CoPc-PIM-A1, CoPc-PIM-A2 and CoPc-PIM-A3) were pre-
pared by a phthalocyanine-forming reaction (Scheme 1) as
described previously.
13,16
Two samples of Co phthalocyanine
network-PIM (designated CoPc-PIM-B1 and CoPc-PIM-B2)
were prepared from a pre-formed phthalocyanine (Scheme 3)
as follows: hexadecachlorophthalocyanine 8 (1 g, 0.89 mmol),
spirobisindane 1 (0.61 g, 1.79 mmol), K
2
CO
3
(1.40 g,
10.14 mmol) and N-methylpyrrolidone (NMP, 12 cm
3
) were
stirred at 195 uC under a nitrogen atmosphere. After 24 h
an insoluble solid precipitated from solution. The reaction
mixture was cooled and a green solid was filtered off. The solid
was purified by Soxhlet extraction (methanol 48 h, tetra-
hydrofuran 48 h, acetone 24 h), refluxed in acetone (3 6 1h)
and filtered off (yield 81%). Found for CoPc-PIM-B1: C, 58.1;
H, 4.1; N, 5.0; Cl, 19.1; Co, 3.4%. Calc. for C
74
H
40
Cl
8
N
8
O
8
Co:
C, 58.8; H, 2.6; N, 7.4; Cl, 18.8; Co, 3.9%. l
max
(1-chloro-
naphthalene)/nm 678.
Fe porphyrin polymer
A metal-free porphyrin network 6 was prepared by reaction of
1 with porphyrin 5 and the network was subsequently
metallated (Scheme 2). Porphyrin 5 (0.17 g, 0.174 mmol),
spirobisindane 1 (0.12 g, 0.348 mmol), K
2
CO
3
(0.38 g,
2.75 mmol) and anhydrous NMP (4 cm
3
) were stirred at 170 uC
under a nitrogen atmosphere. After 5 h a solid precipitated
from solution. The mixture was cooled, methanol (30 cm
3
)was
added, the mixture was stirred for 30 min, then a purple solid
was filtered off. The solid was purified by Soxhlet extraction
(methanol 48 h, tetrahydrofuran 48 h, acetone 24 h), refluxed
in acetone (3 6 1 h) and filtered off to give 6 as a purple
powder (yield 94%). Found: C, 68.5; H, 3.1; F, 16.0; N 4.0%.
Calc. for C
84
H
42
F
12
N
8
O
8
: C, 69.0; H, 2.9; F, 15.6; N, 3.8%. l
max
(1-chloronaphthalene)/nm 435. BET surface area 681 m
2
g
21
.
Metallation of 6: porphyrin network 6 (0.13 g, 0.089 mmol),
Fe
III
Cl
3
(0.15 g, 0.89 mmol) and anhydrous NMP (4 cm
3
) were
stirred at 120 uC for 24 h under a nitrogen atmosphere. The
mixture was cooled, methanol (10 cm
3
) was added, and a
purple solid was filtered off. The solid was purified by Soxhlet
extraction (methanol 72 h, tetrahydrofuran 24 h, acetone 24 h),
refluxed in acetone (3 6 1 h) and filtered off to give 7 as a
purple solid (yield 65%). Found: C, 63.3; H, 3.5; F, 14.0; N,
3.7; Fe, 2.5%. Calc. for C
84
H
40
F
12
N
4
O
8
Fe: C, 66.5; H, 2.6; F,
15.0; N, 3.7; Fe, 3.6%. l
max
(1-chloronaphthalene)/nm 415.
Low molar mass catalysts
Cobalt(
II) phthalocyanine (CoPc, Aldrich) was used as
received. Iron(
II) phthalocyanine (FePc, Aldrich) was activated
as follows: FePc (2 g, 3.5 mmol) was dissolved in conc. H
2
SO
4
(50 cm
3
), the solution was filtered and the filtrate was added to
ethanol (300 cm
3
). The precipitate was separated by centrifu-
gation and the solid was washed three times with H
2
O, then
added to H
2
O (30 cm
3
) and neutralized by addition of
NH
4
OH. The active FePc was separated by centrifugation
and dried (yield 96%). 5,10,15,20-Tetrakis(pentafluorophenyl)
porphyrin iron(
III) chloride (FeCl(PhF
5
)
4
Porph, Aldrich) was
used as received.
Decomposition of hydrogen peroxide
H
2
O
2
decomposition experiments were carried out using a
simple catalysis rig comprising a glass line connected via
isolating taps to a round-bottomed flask, a gas burette, and a
port for evacuation of the line and introduction of oxygen. A
water-bath was used to control the temperature of the reaction
flask. The line was first evacuated then filled with O
2.
Water
was added to the flask and purged with O
2
for 5 min. The
catalyst was added and the mixture was stirred at 30 uCfor
12 h. H
2
O
2
was added and the reaction mixture was stirred
for 3 h. The volume of O
2
in the burette was measured
immediately after addition of H
2
O
2
, then readings of the volume
of O
2
evolved were taken at intervals. Initial rates of reaction
were determined from quadratic fits to the experimental data for
oxygen evolved as a function of time. Typical experiments were
carried out with 4.78 cm
3
H
2
O, an initial H
2
O
2
concentration of
0.75 mol dm
23
and a catalyst concentration corresponding to
1 mol% Co, relative to peroxide. To investigate the order of
reaction, experiments were also carried out with different H
2
O
2
concentrations (down to 0.19 mol dm
23
) and different catalyst
concentrations (1.9–7.5 6 10
23
mol dm
23
). For CoPc and
CoPc-PIM-A1, the temperature dependence was also investi-
gated, over the range 30–60 uC.
Scheme 3 Preparation of spiro-linked cobalt phthalocyanine network
polymer (CoPc-PIM-B) from preformed chlorinated phthalocyanine.
Reagents and conditions: (i) K
2
CO
3
, NMP, 195 uC, 24 h.
This journal is
ß
The Royal Society of Chemistry 2008
J. Mater. Chem.
, 2008, 18, 573–578 | 575
Downloaded by Cardiff University on 02 April 2012
Published on 10 December 2007 on http://pubs.rsc.org | doi:10.1039/B715660J
View Online
Oxidation of cyclohexene
For cyclohexene oxidation, catalyst (9 6 10
22
mmol Co) was
added to a mixture of cyclohexene (0.1 cm
3
, 0.99 mmol) and
dichloromethane (3 cm
3
). The reaction was stirred at room
temperature and 70% t-butylhydroperoxide solution in water
(0.14 cm
3
, 1 mmol) was added. The reaction was stirred for
4 weeks. Samples (0.1 cm
3
) were removed at intervals and
analysed by GC-MS using a Perkin-Elmer Autosystem XL
gas chromatograph with Turbomass mass spectrometer.
Chromatogram peaks were identified by comparison with
those of pure compounds.
Oxidation of hydroquinone
Hydroquinone oxidation experiments were carried out using a
rig similar to that described for H
2
O
2
decomposition, using the
gas burette in this case to measure oxygen consumed rather
than oxygen evolved. Hydroquinone (0.25 g, 2.27 mmol) was
dissolved in glacial acetic acid (5 cm
3
) in the reaction vessel.
The line was evacuated, then oxygen was added to the line
and burette. Finely ground catalyst (corresponding to 1 mol%
metal, relative to hydroquinone) was added to the reaction
vessel, the tap to the line was opened and the volume of oxygen
in the burette was immediately measured. The reaction was
stirred for 8 h and readings of the volume of O
2
consumed
were taken at intervals.
Results and discussion
Polymer catalysts
Apparent surface areas and metal contents for the polymer
catalysts utilized in this study are listed in Table 1. The
properties depend critically on the purity of monomers and the
efficacy of purification of the product, so the same procedure
can give products with a range of surface areas. However,
substantially lower surface areas were obtained for Co
phthalocyanine polymers prepared using preformed phthalo-
cyanine (CoPC-PIM-B1, B2) than for those prepared by the
phthalocyanine-forming route (CoPc-PIM-A1, A2, A3),
reflecting differences in molecular structure and the potential
for association of macrocycles to occur during synthesis. A
higher surface area was achieved for the porphyrin network-
PIM than for the phthalocyanine polymers. It should be
noted that surface area is only a crude indication of the effec-
tive porosity of polymers like these. Their properties depend
critically on how the porosity (free volume) is distributed within
the material. Techniques for properly evaluating effective
micropore distributions are the subject of ongoing research.
Decomposition of hydrogen peroxide
Results of typical H
2
O
2
decomposition experiments are shown
in Fig. 1 for each of the Co phthalocyanine network-PIMs and
for low molar mass Co phthalocyanine (CoPc). None of the
catalysts are soluble in the reaction medium and so all are
functioning as heterogeneous catalysts. With CoPc, for which
accessibility of reagents to the catalytic centre will be restricted
due to strong association of macrocycles, the reaction is slow.
All the Co phthalocyanine network-PIMs tested show a
substantial enhancement of reaction rate. This is quantified
in Table 2, which gives initial rates of reaction (average for
3–8 runs) for each catalyst. Experiments carried out with
different catalyst concentrations and different H
2
O
2
con-
centrations indicated that, over the range of conditions
investigated, the reaction is first order in catalyst concentration
[cat] and first order in peroxide concentration [H
2
O
2
], i.e.
rate = k[cat][H
2
O
2
]. Values of the second order rate constant k
are given in Table 2. The best performing network polymer
(CoPc-PIM-A3) shows two orders of magnitude increase
in rate constant compared to the low molar mass analogue.
Even the network-PIMs prepared from preformed phthalo-
cyanine (CoPc-PIM-B1, B2), which have relatively low surface
areas, show more than an order of magnitude increase in
rate constant. For CoPc and CoPc-PIM-A1, Arrhenius
plots of data obtained at various temperatures gave apparent
activation energies of 65 kJ mol
21
and 16 kJ mol
21
,
respectively.
Table 1 BET surface area, S
BET
, and metal content for polymer
catalysts
Polymer S
BET
/m
2
g
21
Metal content (%)
CoPc-PIM-A1 612 5.8
CoPc-PIM-A2 453 4.0
CoPc-PIM-A3 494 3.0
CoPc-PIM-B1 201 3.4
CoPc-PIM-B2 120 4.8
FePorph-PIM 866 3.6
Fig. 1 Typical plots of percentage reaction as a function of time for
decomposition of H
2
O
2
with (n) CoPc-PIM-A1, (m) CoPc-PIM-A2,
(¤) CoPc-PIM-A3, (%) CoPc-PIM-B1, (&) CoPc-PIM-B2 and
(6) CoPc.
Table 2 Initial rate, n, and rate constant, k, for decomposition of
H
2
O
2
with Co phthalocyanine catalysts
Catalyst n
a
/mol-O
2
mol-cat
21
h
21
k/10
23
dm
3
mol
21
s
21
CoPc 17.7 ¡ 1.3 0.033
CoPc-PIM-A1 620 ¡ 52 1.2
CoPc-PIM-A2 725 ¡ 39 1.3
CoPc-PIM-A3 1845 ¡ 163 3.4
CoPc-PIM-B1 349 ¡ 29 0.65
CoPc-PIM-B2 287 ¡ 6 0.53
a
Initial rate values quoted are the mean for at least three runs ¡
the standard error of the mean.
576 |
J. Mater. Chem.
, 2008, 18, 573–578 This journal is
ß
The Royal Society of Chemistry 2008
Downloaded by Cardiff University on 02 April 2012
Published on 10 December 2007 on http://pubs.rsc.org | doi:10.1039/B715660J
View Online
Oxidation of cyclohexene
Cyclohexene oxidation by t-butylhydroperoxide with Co
phthalocyanine catalysts yields a mixture of three products:
cyclohexene oxide, cyclohexene-2-ol and 2-cyclohexene-1-one.
The composition of the reaction mixture after 8 h and 4 weeks
is shown in Table 3 for experiments carried out with low
molar mass CoPc and with network-PIMs CoPc-PIM-A2
and CoPc-PIM-B2. In each case, the major product is
2-cyclohexene-1-one, the yield of which is shown as a function
of time in Fig. 2. In these experiments, all the catalysts are
insoluble in the reaction mixture. As was found in the
decomposition of H
2
O
2
, the Co phthalocyanine network-
PIMs show substantially enhanced catalytic activity relative to
the low molar mass analogue. After four weeks, for CoPc-
PIM-A2 the total product yield is 99%, whilst for CoPc it is
just 17%.
It is of interest to compare the results obtained here with
data reported by Sehlotho and Nyokong
26
for cyclohexene
oxidation catalysed by CoPc and iron percholorophthalocya-
nine (FeCl
16
Pc). Their experiments differ from the present
work in that a solvent mixture (DMF : dichloromethane 3 : 7)
was employed in which the catalysts were soluble, so they were
dealing with a homogeneous system. For each product, the
selectivity (yield as a percentage of the total yield for all
products) after 8 h is given in Table 4.
Oxidation of hydroquinone
Plots of oxygen consumed over time are shown in Fig. 3 for the
oxidation of hydroquinone with various catalysts. Initial rates
derived from data for the first two minutes of reaction are
listed in Table 5, along with turnover numbers (TON) after
4 h. Values are based on the assumption that all metal centres
are active, whereas in reality, of course, some may be
inaccessible because of structural defects or mass transfer
effects. Table 5 also includes results from Zsigmond et al.
29
for
CoPc, FePc and CoPc encapsulated in zeolite Y (CoPc-ZeY).
CoPc is soluble in the reaction solvent (acetic acid), creating a
homogeneous system. However, as has been observed pre-
viously,
29
CoPc has no catalytic activity, perhaps because of
dimerization and/or degradation in solution. In contrast, the
Table 3 Percentage composition of the reaction mixture at 8 h and
4 weeks for the oxidation of cyclohexene (0.33 mol dm
23
)in
dichloromethane by t-butylhydroperoxide with Co phthalocyanine
catalysts (0.03 mol Co dm
23
)
Catalyst Time
CoPc 8 h 91 1 2 6
4 weeks 83 1 5 11
CoPc-PIM-A2 8 h 43 5 7 45
4 weeks 1 9 12 78
CoPc-PIM-B2 8 h 54 5 7 34
4 weeks 2 11 15 72
Fig. 2 Percentage yield of 2-cyclohexene-1-one as a function of time
for oxidation of cyclohexene by t-butylhydroperoxide with (m) CoPc-
PIM-A2, (&) CoPc-PIM-B2 and (6) CoPc.
Table 4 Selectivity to product after 8 h for the oxidation of
cyclohexene by t-butylhydroperoxide with phthalocyanine catalysts
Catalyst Nature
Ref.
CoCl
16
Pc Homogeneous 8 20 72 26
CoPc Homogeneous 6 34 61 26
CoPc Heterogeneous 11 22 67 This work
CoPc-PIM-A2 Heterogeneous 9 12 79 This work
CoPc-PIM-B2 Heterogeneous 11 15 74 This work
Fig. 3 Oxygen consumed as a function of time for oxidation of
hydroquinone with (m) CoPc-PIM-A2, (6) CoPc, (+) FePc, ($)
FePorph-PIM and (#) FeCl(Ph
5
)
4
Porph.
Table 5 Initial rate, n, and turnover number (TON) after 4 h for the
oxidation of hydroquinone by molecular oxygen with phthalocyanine
and porphyrin catalysts
Catalyst
n/mol-O
2
mol-cat
21
h
21
TON
after 4 h Ref.
CoPc 0 — This work
0—29
FePc 7.3 12 This work
9.7 16 29
CoPc-PIM-A2 33 13 This work
CoPc-ZeY 10.4 6 29
FeCl(PhF
5
)
4
Porph 43 23 This work
FePorph-PIM 64 28 This work
This journal is
ß
The Royal Society of Chemistry 2008
J. Mater. Chem.
, 2008, 18, 573–578 | 577
Downloaded by Cardiff University on 02 April 2012
Published on 10 December 2007 on http://pubs.rsc.org | doi:10.1039/B715660J
View Online
