Phororedox and photocatalytic processes on FE(III)-porphyrin surface modified nanocrystalline TiO2

Article abstract of DOI:10.1016/S1381-1169(99)00446-X

Molecular modification of surfaces is a field of high interest in heterogeneous catalysis. Much work has been conducted on the modification of semiconductor surfaces for improving stability and charge separation. These studies have provided an impetus to investigations on using semiconductors as light-harvesting systems or as catalysts for pollution abatement. Surface derivatization of TiO₂ nanoparticles with a Fe(III)-porphyrin was performed following a novel procedure whereby the complex contains the aminopropylsilane functional group instead of the surface. Characterization of the light-transparent dispersions by laser flash photolysis, UV-vis spectroscopy, and photo-electrochemical techniques showed that the nature of the solvent is an important parameter in determining the redox processes involving the grafted porphyrin. Marked effects on the stability of the Fe(II)-porphyrin formed upon capture of the photogenerated electrons were observed. The bonded porphyrin improved the yield and formation of the monooxygenation products with respect to total degradation to CO₂ for both the examined substrates. An increase in the efficiency and selectivity with the composite photocatalytic system was claimed on this basis. In the case of cyclohexane, the iron-porphyrin complex changed the selectivity of the process, increasing the alcohol/ketone ratio.

Full text of DOI:10.1016/S1381-1169(99)00446-X

Ž
.
Journal of Molecular Catalysis A: Chemical 158 2000 521–531
www.elsevier.comrlocatermolcata
ž /
Phororedox and photocatalytic processes on Fe III –porphyrin
surface modified nanocrystalline TiO2
A. Molinari a, R. Amadelli a,), L. Antolini a, A. Maldotti a, P. Battioni b, D. Mansuy b
a
(
)
Centro di Studio su FotoreattiÕita e Catalisi CNR , Dipartimento di Chimica, UniÕersita di Ferrara, Õia L. Borsari,
`
`
46-44100 Ferrara, Italy
Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, URA 400 CNRS, UniÕersite Rene Descartes, 45 rue de Saints
b
´
´
Peres, F-75270 Paris Cedex 06, France
`
`
Received 10 September 1999; accepted 13 December 1999
Abstract
Ž
.
Surface derivatization of titanium dioxide nanoparticles with a Fe III –porphyrin has been carried out following a new
procedure whereby the complex, rather than the surface, contains the aminopropylsilane functional group. This avoids the
problems of surface deactivation by silane groups, reported in earlier investigations, on analogous systems. Characterization
of the light-transparent dispersions by laser flash photolysis, UV-vis spectroscopy and photo-electrochemical methods has
shown that the nature of the solvent is an important parameter in determining the redox processes involving the grafted
Ž .
porphyrin. In particular, one observes marked effects on the stability of the Fe II –porphyrin formed upon capture of the
photogenerated electrons. The photocatalytic activity of the composite systems was assessed in the process of monooxygena-
tion of cyclohexane and cyclohexene by molecular oxygen. The bonded porphyrin enhances the yield and the formation of
the monooxygenation products with respect to total degradation to CO2 for both the examined substrates. On this basis, we
can claim an increase in the efficiency and selectivity with the composite photocatalytic system. In the case of cyclohexane,
we observed, in addition, that the iron–porphyrin complex also changes the selectivity of the process, increasing the alcohol
to ketone ratio. q 2000 Elsevier Science B.V. All rights reserved.
Keywords: TiO2; Iron–porphyrin; Photocatalysis; Oxygen activation; Hydrocarbon oxidation
1. Introduction
face plays a significant role in a wide range of
applications, from synthesis to energy conver-
sion, from sensor applications to pollution
abatement.
Much work has been carried out on the modi-
fication of semiconductor surfaces for the im-
provement of stability and of charge separation.
These investigations have provided an impetus
to current studies on the use of semiconductors
particularly as light-harvesting systems or as
catalysts for pollution abatement. For energy
Molecular modification of surfaces is a field
of high interest in heterogeneous catalysis. A
vast literature has emphasized the importance of
the design and control of interfaces at a molecu-
lar resolution for the development of functional
w x
catalytic surfaces 1 . The structure of the sur-
)
Corresponding author. Tel.: q39-532-291-148; fax: q39-532-
240-709.
1381-1169r00r$ - see front matter q 2000 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S1381-1169 99 00446-X

(
)
522
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
conversion purposes, molecular modification of
wide band gap semiconductors with dye sensi-
tizers has long been known, and recent signifi-
recovered when the porphyrin is attached to the
silane chains. Quite interestingly, however, this
permitted the identification of the iron–
porphyrin as the active reaction site.
w x
cant achievements have been reported 2 . In the
field of photoelectrocatalysis, most of the re-
ported studies are concerned with semiconduc-
Here, we report on a new functionalization
procedure of TiO2 in which an iron–porphyrin
complex directly bears a silane chain that can be
grafted to the surface. This reduces the ratio of
silane groups per linked complex, in principle
avoiding the deactivation of the surface sites not
occupied by the bonded complex. We used a
meso-tetraaryl iron–porphyrin bearing chlorine
atoms in the meso-aryl groups able to provide a
steric protection of the porphyrin ring against
w
x
tors modified by small metallic particles 3–6 .
Less attention has seemingly been paid to
molecular functionalized semiconductor sur-
w
faces for photoelectrosynthetic applications 7–
x
22 . In this case, the aim is that of controlling
the reactivity and selectivity of the semiconduc-
tor catalyst in a given process by linking it with
non-light absorbing molecules. This strategy can
potentially provide new contributions to the field
of ‘‘green chemical processes’’ where one seeks
a high selectivity under mild and environment-
friendly conditions.
Titanium dioxide is one of the most suitable
semiconductors for practical applications; how-
ever, the highly oxidizing conditions generated
at the illuminated oxide surface place some
limitations upon its use in photoelectrosynthetic
oxidation processes since the desired products
may undergo further oxidation and in extreme
w
x
oxidative degradation 24–26 .
We discuss the results of a characterization
of the composite catalyst based on spectroscopic
and photoelectrochemical measurements as well
as on the assessment of its photocatalytic activ-
ity in the oxidation of hydrocarbons by molecu-
lar oxygen. We used colloidal TiO2 for these
studies because it is transparent enough to allow
spectroscopic investigations to be performed
and, at the same time, it can be conveniently
deposited on inert substrates, such as glass, or
on conductive electrodes.
w
x
cases even total degradation 23 .
Prior work by our group has shown that
photocatalytic systems in which porphyrins are
bonded to the surface of titanium dioxide show
a new reactivity in the monooxygenation of
2. Experimental
2.1. Synthesis and preparation of the colloidal
composite system TiO2 rFe III –porphyrin
w
x
alkanes 21,22 . We pointed out that this func-
tionalization creates an ‘‘integrated system’’
where the solid support and the molecular com-
ponent, taken separately, possess an individual
catalytic activity. In homogeneous phase, the
polyoxometallaterporphyrin integrated systems
is another example where we could achieve
significant benefits in the monooxygenation of
(
)
Colloidal TiO2 was prepared following
w
x
known procedures 27 . The precursor of the
iron–porphyrin complex bearing a trifluorosilyl
function to link to the surface of the semicon-
Ž
.
Ž
Ž
ductor was the Fe III –meso- tetrakis 2,6-di-
.
Ž
Ž
.
.
chlorophenyl porphyrin Fe III TDCPP . The
w
x
hydrocarbons 24–26 .
silanization of this complex was not possible
Ž .
directly, as already reported for Fe III –meso-
In our earlier work, the derivatization of TiO2
with porphyrins was carried out in a two-step
Ž
.
w x
tetrakis pentafluorophenyl porphyrin, 28 but it
has been necessary to modify the complex to
introduce a good leaving group on the phenyl
rings.
w
x
sequence 22 : triethoxypropylaminosilane was
first grafted to the oxide surface and then the
porphyrin was linked through the amino func-
tional groups. In the first step, the surface suf-
fers a drastic loss of activity, which is in part
All the intermediates obtained during the
preparation of the silanized Fe III –porphyrin
Ž
.

(
)
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
523
.
Ž
.
Ž
are reported in Scheme 1. The free porphyrin
phenyl , d 8.05 4H, H-p-phenyl , d y2.51 s,
( )
.
x
1 was prepared following the method sug-
2H, NH in CDCl3 . Subsequent hydrolysis
w
x
( ) Ž
gested by Lindsey et al. 29 : a 3-l one-neck
yielded 3 IR: stretching frequency of SO3H
groups at 1190 and 1050 cmy
.
1
Ž
.
flask was charged with CH2Cl2 2.2 l distilled
.
Ž
Ž
( )
from CaH2 and 2,6-dichlorobenzaldehyde 8 g,
Acros under an argon stream for 20 min. Then,
pyrrole 3.3 g, Acros and 2,2-dimethoxy-
An aqueous solution of FeCl2 4H2O Janssen
Chimica was allowed to react with 3 under
argon at 808C for 18 h. The formation of the
iron–porphyrin complex 4 was followed by
UV-vis spectrophotometry Soret band at 394
nm . After oxygenation of the sample, KOH
35% was added until pH 8–9 to precipitate
iron in excess as Fe OH . After filtration, the
.
.
Ž
.
Ž
.
( )
propane 200 ml, Aldrich were added. The
Ž
reaction mixture was purged for an hour and
Ž
.
added with three aliquots of Et2OBF3 500 ml
.
Ž
.
each, Aldrich at 30-min intervals. After 24 h of
Ž
.
rest, the formation of porphyrinogen was moni-
3
tored by UV-vis absorption spectrophotometry
reaction mixture has been neutralized with HCl
Ž
.
Ž
.
large and weak band at 512 nm . It is then
oxidized by 2,3-dichloro-5,6-dicyano-1,4-
35% and the iron–porphyrin complex was pu-
rified by column chromatography with an ionic
Ž
.
Ž
benzoquinone 13 g, Aldrich previously dis-
exchange resin Dowex 50=8-200, Janssen
Ž
.
( )
.
solved in toluene 400 ml, SDS . H2TDCPP 1
Chimica . After concentration, it was then pre-
was purified by column chromatography with
cipitated from acetone.
Ž
.
neutral activated alumina 150 mesh, Aldrich
Chlorination was performed on the complex
( ) Ž
.
Ž
.
Ž
.
and CH2Cl2 as eluent, then concentrated, pre-
cipitated from n-pentane, dried at 1008C and
characterized by UV-vis 418, 512 and 588 nm
and H NMR d 8.64 s, 8H, H-pyrrole , d
7.72 8H, H-m-phenyl , d 7.19 4H, H-p-
phenyl , d y2.54 s, 2H, NH in CDCl3 . An
aromatic electrophylic substitution reaction with
4
1 g using PCl5 3.5 g and POCl3 10 ml
( )
at 508C, for 30 min to lead to 5 . Finally, this
was a suitable complex for the nucleophilic
substitution by excess aminopropyl triethoxysi-
Ž
.
1
w
Ž
.
Ž
.
.
Ž
Ž
( ).
lane aptes, 160 mgr50 mg of 5 for 17 h
Ž
.
x
Ž
.
under argon in the presence of pyridine 30 ml
in THF 4 ml at 708C. Unfortunately, the
monomer 6 could not be directly removed
Ž
.
Ž
.
( )
( )
ClSO3H Janssen Chimica on 1 under rigor-
Ž
.
ously controlled conditions 1008C, 3 h gave
2 , which has been characterized by H NMR
from the reaction mixture because of a partial
polymerization between the monomers and the
excess nucleophilic reagent. A two-step purifi-
1
( )
w
Ž
.
Ž
d 8.62 s, 8H, H-pyrrole , d 8.60 4H, H-m-
Ž .
cation process was devised: i hydrolysis and
Ž .
Ž
polycondensation of the crude mixture; and ii
Ž
.
Ž
.
.
treatment with HF 48% rH2SO4 98% 1:1 6
mlr100 mg of porphyrin complex to produce
( )
the porphyrinotrifluorosilane monomer 7 in
CH2Cl2 solution. The excess aptes was simulta-
neously converted to aminopropyltrifluorosi-
lane, which was extracted with aqueous acid.
( )
Ž
Complex 7 exhibited a UV-vis spectrum lmax
.
s418 nm similar to that of the parent por-
( )
Ž
.
phyrin 5 and of Fe III TDCPP itself, indicat-
ing that the porphyrin rings were not modified
during the preparation.
Ž
.
( )
Scheme 1. Synthesis of the modified Fe III TDCPP. 1 MsH2 ,
( )
2
( )
3
( )
4
XsH;
MsH2 , XsSO2Cl;
MsH2 , XsSO3H;
( )
Complex 7 could be used in the functional-
3q
3q
MsFe3q, XsSO3H; 5 MsFe , XsSO2Cl; 6 MsFe
,
( )
( )
ization of the oxide surface to give the colloidal
Ž
.
Ž
.
( )
7
X s SO 2 NH CH 2 3 Si OEt
;
M s Fe3q
,
X s
3
Ž
.
Ž
.
composite system TiO2rFe III –porphyrin. In
SO2 NH CH2 3SiF3.

(
)
524
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
( )
Ž
fact, when the iron–porphyrin complex 7 was
added to an alcoholic suspension of colloidal
TiO2, the SiF3 groups were immediately con-
verted to SiOH by the water molecules present
on the oxide surface and the iron–porphyrin
frequency multiplier 355 nm, 5 ns half-width,
.
150 mJ . The eventually observed absorbance
variations after the flash have been read on a
LeCroy 9360 fast digitizing oscilloscope.
Irradiations were carried out with a medium-
pressure HgrXe lamp, with thermostated cell
Ž
complex was covalently linked to TiO2 as
schematized in Fig. 1 by one or more of its
.
Ž .
holder 22"18C . The desired wavelength in-
Ž .
aminopropyl arms. Furthermore, it is also possi-
ble that the formation of some dimers or
oligomers between the iron–porphyrin com-
plexes bonded on the semiconductor surface.
Although, given the low coverage by the com-
plex, this phenomenon should not be extensive.
terval 300-l-370 nm has been selected
Ž
.
using a NiSO4 solution 8 g in 10 ml coupled
Ž
.
with a glass cut-off T%s16 . Product analy-
ses were performed by gas chromatography us-
ing a DANI 8521-a equipped with a flame
ionization detector using columns packed with
Carbowax 20 M 5% on Chromosorb W-AW.
The reaction products were identified by com-
parison of their retention times with those of the
authentic samples. Carbon dioxide formed dur-
ing the experiments has been collected into a
2.2. Materials
The colloidal composite system TiO2r
Fe III –porphyrin used has been prepared as
described in the previous paragraph. All the
organic solvents and substrates employed
Ž
.
y
2
Ž
.
Ž
.
solution of Ba OH
1=10
M in sodium
2
y
3
Ž
.
silicate 1=10
M . Then, BaCO3 has been
Ž
.
MeOH, DMF, CH2Cl2, cyclohexane, etc. were
determined by a turbidimetric method using a
spectroscopic grade reagents and were used
without further purification. Cyclohexene was
distilled before use.
calibration curve.
The samples were prepared as follows: col-
Ž
.
loidal TiO2 5 mg was suspended in the chosen
Ž
.
medium 3 ml and an aliquot of iron–porphyrin,
5
2.3. Apparatus and methods
corresponding to a concentration of 2=10y
mol dmy3, was added. When needed, the sam-
ples were deaerated, bubbling N2 in it before
proceeding with the experiment. For the photo-
catalytic experiments, the colloidal suspension
of TiO2 added with Carbowax was stationed
Laser flash photolysis experiments were car-
ried out using an Applied Photophysics detec-
tion systems coupled with a Continuum Surelite
II-10 neodinium YAG laser, equipped with a
Ž
.
Fig. 1. Schematic structure of the composite colloidal system TiO2rFe III –porphyrin.

(
)
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
525
Ž
.
along a glass sheet 1 cm in width to form a
thin film of colloid; after solvent evaporation,
the sheet was put into the oven at 4508C for 30
min. When necessary, its immersion into a func-
tionalized iron–porphyrin solution led to a de-
posited film of the composite system. The glass
sheet so obtained was put into a suitable spec-
trophotometric cell containing cyclohexane or
Ž
.
DMFrCH2Cl2rcyclohexene 4:2:1 mixed sol-
vent to be irradiated in the presence of a satu-
rated atmosphere of O2. Electrochemical experi-
ments were performed with an EG&G potentio-
statrgalvanostat and EG&G software. Elec-
trodes were prepared by depositing films of
TiO2 onto Ti sheets, which were previously
etched in concentrated oxalic acid at 808C. When
required, the electrodes were illuminated using
a Hanau Q-400 medium-pressure mercury lamp
equipped with a filter to cut off light of wave-
length l-360 nm. The supporting electrolyte
Fig. 3. UV-vis spectra of the colloidal composite system before
Ž
.
Ž
.
the flash curve a , after the flash curve b , and after the flash and
Ž
.
oxygenation of the sample curve c .
that light was absorbed completely by TiO2,
which worked as the photochemical active com-
ponent of the colloidal composite system
3
was 0.1 mol=dmy NaClO4; a Pt sheet and
Ž
.
TiO2rFe III –porphyrin, while the analysis
Ž
.
wavelength ls423 nm corresponded to the
absorption maximum of the iron–porphyrin
complex in its reduced form. Fig. 2 reports the
main results obtained: when the composite col-
loidal system was suspended in MeOH, the
saturated calomel were used as the counter and
reference electrode, respectively.
3. Results and discussion
Ž
flash gave rise to an absorption increase curve
3.1. Photoredox properties
.
Ž
.
Ž .
a indicating the reduction of Fe III to Fe II .
Furthermore, its UV-vis spectrum after the flash
showed the typical band of reduced iron–
porphyrin Fig. 3, curve b ; oxygenation of the
sample led to the original spectrum of the oxi-
In the laser flash photolysis experiments, the
Ž
.
excitation wavelength ls355 nm was such
Ž
.
Ž
.
dized form curves a and c . Therefore, these
results point out that photoexcited TiO2 can
work as electron source for the reduction of the
Ž
.
Ž
.
iron III –porphyrin Eqs. 1 and 2 . Simultane-
ously, the alcohol can rapidly capture the va-
lence band holes or, alternatively, react with
Ž
.
surface OH radicals Eqs. 3 and 4 .
TiO2 qhn TiO ey, hq
1
2Ž
.
Ž .
cb
Fe III Porqqey Fe II Por
2
Ž .
3
Ž .
Ž
.
Ž .
Ti OH
cb
q
Pq
`
`
Ti OHqh
Fig. 2. Laser flash photolysis of the colloidal composite system
P
Pq
`
Ž
.
Ž
.
Ti OH qCH3OH TiOHq CH2OHqHq
TiO2 rFe III –porphyrin dissolved in MeOH curve a and in
Ž
.
H2O curve b , in the absence of O2. lexc s355 nm, lanal s423
4
Ž .
nm.

(
)
526
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
the following section, we discuss possible addi-
tional explanations.
3.2. Photoelectrochemical experiments
We carried out electrochemical experiments
with porphyrin-modified TiO2 photoelectrodes
to attempt a further characterization of the na-
ture of the interactions between the composite
system and the liquid medium. A comparative
behavior of the photocurrents vs. potential
curves is given in Fig. 5 for TiO2 and por-
phyrin–TiO2 electrodes in aqueous solution and
in DMFrH2O mixtures. It is clear in this figure
that only in the case of the porphyrin–
TiO2rH2O system does one observe a clearly
distinct behavior. There is, in fact, a sensible
shift of the photocurrent onset to higher positive
potentials, which indicates that electron–hole
recombination is prevailing at potentials ap-
proaching the flat band value. Accordingly, the
value of flat band potential derived from Mott–
Schottky plots was y0.3V for porphyrin–
TiO2rH2O and y0.4V for the other systems.
The data are in agreement with those discussed
above on the impossibility to reveal the pres-
Ž .
Fig. 4. Concentration of Fe II –porphyrin formed after 5 min of
Ž
.
irradiation 300-l-370 nm of the composite colloidal system
dissolved in MeOHrH2O mixed solvents with different composi-
tion.
On the other hand, when the composite col-
loidal system was suspended in water, we did
not observe any absorption increase after the
flash. Unfortunately, it has not been possible to
carry out the experiments in a shorter time scale
because of unavoidable light-scattering phenom-
ena by the colloidal suspensions. Continuous
irradiation of deaerated composite colloidal sys-
tem suspensions in various MeOHrH2O mix-
Ž
.
tures Fig. 4 showed that the concentration of
reduced iron–porphyrin, produced in the pri-
mary photochemical process decreased when
the content of water in the mixture increased.
The instability of the porphyrin reduced form in
water may be due to the fact that, in the absence
of alcohol, OH surface radicals can oxidise the
reduced porphyrin complex as in Eq. 5.
Ž .
ence of the Fe II –porphyrin in H2O.
P
q
Ti OH qFe II Por TiOHqFe III Porq
`
Ž .
Ž
.
5
Ž .
In contrast, in methanol, the capture of OH
Ž
.
radicals or valence band holes yields CH2OH
radicals with sufficient reducing power to favour
Ž .
the formation of the Fe II complex through the
following reaction:
PCH OHqFe III Porq Fe II Por
Ž
.
Ž .
2
qCH2OqHq
6
Ž .
Fig. 5. Photocurrent values vs. applied potential for colloidal TiO2
Ž
.
deposited on a Ti electrode in H2O and in H2OrDMF 30:70
This process would account for the accumula-
tion of Fe II in this particular environment. In
Ž
.
both in the absence and in the presence of the Fe III –porphyrin
complex.
Ž .

(
)
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
527
The photocurrents plot of Fig. 5 does not
indicate any significant oxidation of DMF. In-
stead, one important point to note is the ability
for example, the production of adipic acid and
Nylon 6.
It has not been possible to perform the cyclo-
hexane photooxidation using the colloidal sys-
tem itself because it could not be suspended in
any medium containing the hydrocarbon. For
this reason, we decided to use films obtained by
depositing the colloid on a glass sheet, which
has been immersed into neat cyclohexane.
Table 1 shows that the presence of the iron–
porphyrin complex significantly increases the
photocatalytic efficiency with respect to col-
loidal TiO2 alone, since the overall concentra-
tion of the oxidation products grows by about
50%. In addition, the iron–porphyrin changes
the selectivity in the formation of monooxy-
genated products, increasing the cyclohexanolr
cyclohexanone ratio.
Ž
.
of DMF shared by CH3OH to coordinate to
Ž
.
the axial positions of the Fe III –porphyrin. We
then think that a possible explanation of the
observed phenomena is due to the fact that the
coordinating organic solvent tends to change the
orientation of the porphyrin in such a way as to
decrease the degree of its interaction with the
oxide surface. In other words, in an aqueous
medium, in the absence of strongly bonded
axial ligands, the linked porphyrin would lie
closer to the surface and an interaction with
surface hydroxyl groups would be easier than in
the case of a strongly coordinating medium.
This explanation is in agreement with the mech-
anism proposed for aminopropyl surface immo-
w
x
bilized redox centers 30 in general, and por-
We observe here for the first time that sur-
face derivatization of TiO2 leads to an improve-
ment in the photocatalytic oxidation of hydro-
carbons. The new procedure adopted in the
functionalization of the surface is probably the
main reason for the comparatively higher effi-
ciency of the composite catalyst described
herein. In fact, in our previous work, we used a
two-step method where the surface was first
treated with aminopropylsilane, and then the
porphyrin complex grafted thereon. The first
step dramatically inactivates the surface, while
the photocatalytic activity is 35% regenerated
when the porphyrin is bonded to the surface
w
x
phyrins in particular 11 , where electron trans-
fer occurs via a ‘‘floppy’’ chain folding model.
3.3. Photocatalytic properties
The results so far discussed provide evidence,
at least in organic media, of redox processes
occurring between the iron–porphyrin complex
and the photoexcited TiO2. This fact arose our
interest; in particular, we wanted to verify the
behaviour of the composite system in the photo-
oxidation of cyclohexane, because it is a simple
alkane, which has already been tested with a
w
x
w
x
variety of photocatalysts 21–26 and also be-
cause its oxidation products are important inter-
mediates in many industrial processes such as,
21,22 . The procedure adopted in this work
apparently allows a better sites synergy, with
the iron–porphyrin, providing alternative path-
Table 1
Ž
.
Ž .
Photocatalytic properties 300-lexc -370 nm of TiO2 and of the composite system TiO2rFe III –porphyrin as films deposited on glass
sheets in the oxidation of cyclohexane
a
w
w
x
w
w
x
w
Monooxygenated
products r CO2
Photocatalytic system
C6 H10Oq
CO2
C6 H11OH r
a
x
x
O
x w
x
C6 H11OH
C6 H10
TiO2b
26
44
7.9
8.8
0.15
0.37
3.2
5
b
Ž
.
TiO2rFe III –porphyrin
a
4
Concentrations =10 mol dmy3 . Reported values are "5%.
Ž
.
b Medium: 3 ml of neat cyclohexane. Irradiation time 360X.

(
)
528
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
ways in the known reaction mechanism of pho-
indicates that heterogenization on TiO2 in-
creases the porphyrin stability by two orders of
magnitude with respect to the same experiments
where the complex was used in solution in the
w
x
toexcited TiO2 23 .
Concerning the selectivity of the system for
the oxidation processes under examination, it is
necessary to point out that the total degradation
of the hydrocarbons leads to CO2 formation.
When this is counted among the products, we
can conclude that in all cases, the porphyrin
derivatization of the oxide improves the overall
selectivity of the processes.
w
x
absence of TiO2 31–33 .
Scheme 2 shows some possible reaction path-
ways explaining the effect of the iron–porphyrin
on the photocatalytic efficiency of TiO2 as well
as on the observed change in the selectivity of
Ž
.
cyclohexanol formation Table 1 . Route A con-
cerns the formation of intermediates and stable
oxidation products at the semiconductor surface
independently of surface functionalization. The
Another important issue is the catalyst stabil-
ity. In this respect, we noted that during the
photocatalytic experiments, the iron–porphyrin
complex undergoes a slow degradation. It has
been possible to evaluate a turnover number
Ž
.
hydrocarbon RH is oxidized to a radical
species by the photogenerated holes in analogy
with the above discussion for methanol oxida-
tion. The reaction between RØ and surface OH
radicals leads to the formation of the cyclohex-
Ž
ratio between the moles of photooxidized cy-
clohexane and the moles of degraded iron–
.
porphyrin complex equal to 12 000. This value
Scheme 2. Schematic representation of the proposed photooxidation reaction mechanism.

(
)
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
529
anol, which can be efficiently oxidized to cyclo-
heterolytic cleavage of its peroxidic bond
whereby forming intermediate species, which
are able to hydroxylate alkanes.
w
x
hexanone before its desorption 23 . The direct
reaction of RØ with O2 would yield peroxo
species, ROOØ and ROOH, which could eventu-
ally form the ketone. Subsequent oxidation reac-
tions lead to the total degradation of the organic
substrate.
In view of our continuing interest in hydro-
Ž
.
carbon oxidation, we tested the TiO2rFe III –
porphyrin composite system also in the photo-
oxidation of cyclohexene, as model molecule of
alkenes. However, it has not been possible to
perform the oxidation experiments immersing
the glass supported films in neat cyclohexene,
because of the noteworthy thermal reactivity of
this substrate towards autooxidation processes.
For this reason, Table 2 reports the results
obtained using a mixture of solvents, where
cyclohexene was the minor component.
Although these species are formed irrespec-
tive of whether the surface is modified or not,
there could still be an influence of the iron–
porphyrin complex on how they evolve. For
example, reaction 5 of route A can be inhibited
because the peroxide species can react with the
porphyrin complex, both in its ferric and ferrous
Ž
form, leading exclusively to cyclohexanol route
.
w x
B . Based on literature data 34 , these reactions
should be particularly relevant when a halo-
genated iron–porphyrin is involved.
As noticed for cyclohexane, the presence of
the iron–porphyrin significantly increases the
photocatalytic efficiency of the system. How-
ever, we failed to observe an effect of the
porphyrin on the monooxygenation products
distribution. These results are possibly ex-
plained by the fact although the oxidation cyclo-
hexene benefits from a higher amount of dioxy-
gen activated species, it is too reactive to prefer-
ably follow some porphyrin-mediated reaction
pathways, and products are mainly formed via
autooxidation processes. It is important to note,
however, that in this case, too, we observe an
increase in selectivity with respect to unmodi-
Other possible reactions, which can explain
the high yield in alcohol among the products are
represented by Eqs. 1–4 of route C. These take
place as a consequence of the photoinduced
Ž
.
reduction of the ferric porphyrin Eq. 2 and
lead to the reductive activation of O2. On the
other hand, the reaction of the photogenerated
Ø
Ž .
R with the Fe II –porphyrin in the presence of
Ž
.
O2 can give a species of the type Fe III -OOR
32,33 . As previously proposed in the catalytic
cycle of P-450 35,36 , this species undergoes a
w
x
w
x
Table 2
Ž
.
Ž .
Photocatalytic properties 300-lexc -370 nm of TiO2 and of the composite system TiO2rFe III –porphyrin as films deposited on glass
sheets in the oxidation of cyclohexene
a
X
Ž
.
Medium: 3 ml of DMFrCH2Cl2rcyclohexene 4:2:1 . Irradiation time 180 .
b
4
Concentrations =10 mol dmy3 . Reported values are "5%.
Ž
.

(
)
530
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
fied TiO2 when the ratio of the amount of
monooxygenation products vs. CO2 is consid-
the iron–porphyrin complex also changes the
selectivity of the process for the monooxygena-
tion products formation, increasing the alcohol
to ketone ratio. The ability of the iron–porphyrin
complex to tune the high oxidation power of the
semiconductor is an important characteristic of
the composite system, especially when the inter-
est is focused on synthesis rather than on detox-
ification processes.
Ž
.
ered Table 2 .
4. Conclusions
The present work discusses the photochemi-
cal behavior of nanocrystalline TiO2, which is
Ž
.
modified by a Fe III –porphyrin linked to the
surface via a silane bond. With respect to exist-
ing literature on derivatization methods, we
adopted a new procedure whereby the complex
rather than the surface contains the silane func-
tional group.
References
w x
1
Ž
.
R. Murray Ed. , Molecular Design of Electrode Surfaces,
Wiley, New York, 1992.
w x
The nature of the solvent has an important
role in determining the accumulation of the
porphyrin reduced form in the absence of oxy-
gen. Assuming a ‘‘floppy’’ chain folding model
for the bonded porphyrin, this is explained in
terms of a coordination of solvent molecules to
the porphyrin metal center, which results in a
re-orientation of the complex so as to decrease
its interaction with the surface. The fact that the
results can be well explained on the basis of this
model, seems to exclude significant contribu-
tions of porphyrin dimerization or oligomeriza-
tion phenomena on the surface, since this would
create less flexible structures.
The photocatalytic properties of the system
were assessed in the processes of monooxygena-
tion of cyclohexane and cyclohexene with
molecular oxygen. The presence of iron–
porphyrin causes an important increase of the
photocatalytic efficiency with respect to the un-
modified nanocrystalline TiO2 alone, since the
overall concentration of the oxidation products
increases by a factor of 2 in the former case.
This is in contrast with what was observed with
previously described iron–porphyrinrTiO2
photocatalysts that were prepared by a different
procedure of functionalization of TiO2 surface.
Moreover, we find out a decrease in CO2
formation, making the composite system a bet-
ter photocatalyst for synthesis processes than
unmodified TiO2. In the case of cyclohexane,
2
P. Allongue, in: B.E. Conway, J.O’M. Bockris, R.E. White
Ž
.
Eds. , Modern Aspects of Electrochemistry vol. 23 Plenum
Press, New York, 1992.
w x
3
A.J. McEvoy, M. Gratzel, Sol. Energy Mater. Sol. Cells 32
¨
Ž
.
1994 221.
w x
4
R. Amadelli, R. Argazzi, C.A. Bignozzi, F. Scandola, J. Am.
Ž
.
Chem. Soc. 112 1990 7099.
w x
Ž .
B. O’Regan, M. Gratzel, Nature 353 1991 737.
¨
5
w x
Ž
.
6
G.J. Meyer Ed. , Molecular Level Artificial Photosynthetic
Materials, Progress in Inorganic Chemistry Series vol. 44
Wiley, New York, 1997, and references therein.
w x
7
D.F. Untereker, J.C. Lennox, L.M. Wier, P.R. Moses, R.W.
Ž
.
Murray, J. Electroanal. Chem. 81 1977 309.
w x
Ž .
8
I. Haller, J. Am. Chem. Soc. 100 1978 8050.
w x
Ž .
H.O. Flinkea, R.W. Murray, J. Phys. Chem. 83 1979 353.
9
w
w
x
Ž
.
10 M. Tomkiewicz, J. Electrochem. Soc. 127 1980 1518.
x
11 M. Fujihira, T. Kubota, T. Osa, J. Electroanal. Chem. 119
Ž
.
1981 379.
w
x
12 Kabir-Ud-Din, C. Owen, M.A. Fox, J. Phys. Chem. 85
Ž
.
1981 1679.
w
w
w
w
w
x
Ž
.
13 H.O. Flinkea, R. Vithanage, J. Phys. Chem. 86 1982 362.
x
Ž
.
14 P. Kamat, M.A. Fox, J. Phys. Chem. 87 1983 59.
x
Ž
.
15 R. Vithanage, H.O. Flinkea, J. Electrochem. Soc. 1984 799.
x
Ž
.
16 T.R. Hayes, J.F. Evans, J. Phys. Chem. 88 1984 1963.
x
17 A.P. Hong, D.W. Bahneman, M.R. Hoffmann, J. Phys. Chem.
Ž
.
91 1987 2109.
w
w
w
w
w
x
18 A.P. Hong, D.W. Bahneman, M.R. Hoffmann, J. Phys. Chem.
Ž
.
91 1987 6245.
x
19 A.P. Hong, S.D. Scott, M.R. Hoffmann, Environ. Sci. Tech-
Ž
.
nol. 23 1989 533.
x
20 W.G. Becker, M.M. Truong, C.C. Ai, N.N. Hamel, J. Phys.
Ž
.
Chem. 93 1989 4882.
x
21 R. Amadelli, M. Bregola, E. Polo, V. Carassiti, A. Maldotti,
Ž
.
J. Chem. Soc. Chem. Comm. 1992 1355.
x
22 E. Polo, R. Amadelli, V. Carassiti, A. Maldotti, in: M.
Guisnet, J. Barbier, J. Barrault, C. Bouchoule, D. Duprez, G.
Ž
.
Perot, C. Montassier Eds. , Heterogeneous Catalysis and
´
Fine Chemicals III, Elsevier, The Netherlands, 1993, p. 409.
w
x
23 P. Boarini, V. Carassiti, A. Maldotti, R. Amadelli, Langmuir
Ž
.
14 1998 2080, and references therein.

(
)
A. Molinari et al.rJournal of Molecular Catalysis A: Chemical 158 2000 521–531
531
w
w
x
w
w
w
w
w
w
w
x
Ž
.
24 A. Maldotti, A. Molinari, P. Bergamini, R. Amadelli, P.
30 P.R. Moses, R.W. Murray, J. Am. Chem. Soc. 98 1976
Ž
.
Battioni, D. Mansuy, J. Mol. Catal. A: Chemical 113 1996
7435.
x
31 P. Battioni, J.F. Bartoli, D. Mansuy, Y.S. Byun, T.G. Tray-
147.
x
Ž .
lor, J. Chem. Soc. Chem. Comm. 1992 1051.
25 A. Maldotti, A. Molinari, R. Argazzi, R. Amadelli, P. Bat-
Ž
.
x
tioni, D. Mansuy, J. Mol. Catal. A: Chemical 114 1996
32 A. Maldotti, C. Bartocci, G. Varani, A. Molinari, P. Battioni,
Ž .
D. Mansuy, Inorg. Chem. 35 1996 1126.
141.
w
w
w
x
x
26 A. Molinari, A. Maldotti, R. Amadelli, A. Sgobino, V.
33 A. Maldotti, C. Bartocci, R. Amadelli, E. Polo, P. Battioni,
Ž
.
Ž .
D. Mansuy, J. Chem. Soc. Chem. Comm. 1991 1487.
Carassiti, Inorg. Chim. Acta 272 1998 197.
x
x
27 B. O’Regan, J. Moser, M. Anderson, M. Gratzel, J. Phys.
34 M.W. Grinstaff, M.G. Hill, J.A. Labinger, H.B. Gray, Sci-
¨
Ž
.
Ž .
ence 264 1994 1311.
Chem. 94 1990 8720.
x
x
28 P. Battioni, E. Cardin, M. Louloudi, B. Schollhorn, G.A.
35 M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Chem.
Ž
.
Spyroulias, D. Mansuy, T.G. Traylor, Chem. Commun.
Rev. 96 1996 2841, and references therein.
Ž
.
x
Ž
.
1996 2037.
36 J.T. Groves, Y.Z. Han, in: P.R. Ortiz de Montellano Ed. ,
Cytochrome P450, Structure, Mechanism and Biochemistry,
2nd edn., Plenum Press, New York, 1995.
w
x
29 J.S. Lindsey, H.C. Hsu, I.C. Schreiman, Tetrahedron Lett. 27
Ž
.
1986 4969.