Oxidation of aromatic alcohols in irradiated aqueous suspensions of commercial and home-prepared rutile TiO2: A selectivity study

Article abstract of DOI:10.1002/chem.200702044

The photocatalytic oxidation of benzyl alcohol (BA) and 4-methoxybenzyl alcohol (MBA) has been performed in pure water by using commercial TiO ₂ samples (Sigma-Aldrich, Merck, Degussa P25) and rutile TiO ₂ prepared from TiCl₄ at low temperature. Particular attention has been devoted to the identification of the produced aromatic compounds along with the formed CO₂. Oxidation products such as the corresponding aromatic aldehyde and acid, as well as mono- and dihydroxylated aldehydes have been detected. The home-prepared rutile sample showed a marked selectivity towards the formation of the aromatic aldehyde (38 and 60% for BA and MBA, respectively), resulting in a three- to sevenfold improvement relative to commercial samples, with the only byproduct being CO₂. This catalyst was found to be the most selective in the formation of aldehyde in water. By using the commercial or the calcined home-prepared samples, many hydroxylated aromatic compounds were detected besides the aldehyde and the acid. This finding points to a higher selectivity performance of the home-prepared rutile relative to the commercial TiO₂ samples. Some of the home-prepared samples were also dialysed to check the influence of the presence of Cl^-1 species on catalyst reactivity and selectivity. We have attempted to explain the different reaction rate and selectivity observed for MBA and BA.

Full text of DOI:10.1002/chem.200702044

DOI: 10.1002/chem.200702044
Oxidation of Aromatic Alcohols in Irradiated Aqueous Suspensions of
Commercial and Home-Prepared Rutile TiO : A Selectivity Study
2
[
a]
[b]
[a]
[a]
Vincenzo Augugliaro,* Tullio Caronna, Vittorio Loddo, Giuseppe Marcì,
[
a]
[a]
[a, c]
Giovanni Palmisano, Leonardo Palmisano,* and Sedat Yurdakal
Abstract: The photocatalytic oxidation
of benzyl alcohol (BA) and 4-methoxy-
benzyl alcohol (MBA) has been per-
formed in pure water by using commer-
cial TiO2 samples (Sigma-Aldrich,
Merck, Degussa P25) and rutile TiO2
prepared from TiCl4 at low tempera-
ture.Particular attention has been de-
voted to the identification of the pro-
duced aromatic compounds along with
sample showed a marked selectivity to-
wards the formation of the aromatic al-
dehyde (38 and 60% for BA and
hydroxylated aromatic compounds
were detected besides the aldehyde
and the acid.This finding points to a
higher selectivity performance of the
home-prepared rutile relative to the
MBA, respectively), resulting in
a
three- to sevenfold improvement rela-
tive to commercial samples, with the
commercial TiO samples.Some of the
2
only byproduct being CO .This cata-
home-prepared samples were also dia-
lysed to check the influence of the
2
lyst was found to be the most selective
in the formation of aldehyde in water.
By using the commercial or the cal-
cined home-prepared samples, many
À
presence of Cl species on catalyst re-
activity and selectivity.We have at-
tempted to explain the different reac-
tion rate and selectivity observed for
MBA and BA.
the formed CO .Oxidation products
2
such as the corresponding aromatic al-
dehyde and acid, as well as mono- and
dihydroxylated aldehydes have been
detected.The home-prepared rutile
Keywords: alcohols · oxidation ·
photocatalysis · rutile · selectivity
Introduction
the most stable among TiO phases even in strongly acidic
2
or basic conditions and it is applied in optical devices,
energy resources, high-quality paints and cosmetics.For pho-
tocatalytic reactions, however, the most-used crystalline
phase has been anatase or anatase–rutile mixtures, whereas
pure rutile has been typically considered less (or even not)
active.Nevertheless, many preparations of active rutile have
[1]
Since the discovery of TiO properties through irradiation,
2
which resulted in the chemical mediation of redox processes,
heterogeneous photocatalysis has been mainly employed to
degrade organic and inorganic pollutants in water or air; it
[2]
has also been coupled with other technologies. Rutile is
[3]
been reported during the last two decades. Preparation of
rutile has been performed in most cases by calcining anatase
or amorphous samples at approximately 1000 K, but even
some rutile syntheses at low temperature have been report-
[
a] Prof.V.Augugliaro, Dr.V.Loddo, Dr.G.Marcì, Dr.G.Palmisano,
Prof.L.Palmisano, Dr.S.Yurdakal
[4]
“
Schiavello-Grillone” Photocatalysis Group
ed with better results in terms of activity.
Dipartimento di Ingegneria Chimica dei Processi e dei Materiali
Università degli Studi di Palermo
Viale delle Scienze, 90128 Palermo (Italy)
Fax : (+39)091-656-7280
E-mail: augugliaro@dicpm.unipa.it
palmisano@dicpm.unipa.it
Photocatalytic-selective reactions for synthetic purposes
are not as common as degradations but have been carried
[5]
out, although they have been mainly performed either in
[6]
organic solvents or in the gas phase. However, hydrocar-
[
7]
[8]
bon oxidation, aromatic hydroxylation, naphthalene oxy-
[9]
[10]
[
b] Prof.T.Caronna
genation, heterocycle functionalisation
and cyclisation
Facoltà di Ingegneria
Università di Bergamo, Via Marconi
[11]
of amino acids are some of the few studies related to the
selectivity of photocatalytic reactions in water.
5
-24044 Dalmine (Bergamo) (Italy)
The need to work on sustainable green chemical routes is
today of primary importance, and photocatalysis, especially
when carried out in organic-free solvents, offers a number of
[
c] Dr.S.Yurdakal
Kimya Bçlümü, Fen Fakültesi, Anadolu Üniversitesi
Yunus Emre Kampüsü, 26470 Eski s¸ ehir (Turkey)
4640
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4640 – 4646

FULL PAPER
advantages, the most relevant of which is the use of the sun
as a green and free energy source, even if one considers the
small portion of the solar spectrum that falls into the UV
region.Moreover, heterogeneous photocatalysis requires
neither unhealthy, dangerous heavy-metal catalysts nor
strong chemical oxidant/reducing species to carry out chemi-
cal processes at measurable rates.
(Merck) or mixtures of the two phases (Degussa P25) were
employed for comparative purposes.
The home-prepared samples were synthesised from solu-
tions of TiCl .They were prepared by hydrolysing TiCl 4 at
4
room temperature and by treating the resulting solids at dif-
ferent temperatures (up to 973 K) to attain the various sam-
ples.The catalysts are hereafter referred to as “HP” fol-
lowed by a number, which indicates the treatment tempera-
ture in kelvin.
Selective oxidation of hydroxyl groups can be considered
a key step in many organic syntheses.Hence, one of the
main goals of current chemistry has been to eliminate the
environmentally harmful conditions in which this process
Figure 1 shows the XRD patterns of all of the catalysts.
The pattern of HP298, which was prepared at room temper-
[12]
has usually been carried out, a process that generally in-
volves organic solvents at high temperature and pressure as
well as stoichiometric oxygen donors (such as chromate and
permanganate) that are not only expensive and toxic, but
also produce large amounts of dangerous waste.
Photocatalytic alcohol oxidation has been carried out
[13]
[14]
either in acetonitrile or in air, but partial oxidation of
aromatic alcohol has only recently been carried out in sus-
[15]
pensions of water and home-prepared anatase TiO2,
which was obtained by boiling a solution of TiCl .4-Meth-
4
A
C
H
T
R
E
U
N
G
oxy
A
C
H
T
R
E
U
N
G
benzyl alcohol (MBA) was partially oxidised to 4-meth-
A
C
H
T
R
E
U
N
G
oxybenzaldehyde with a selectivity of approximately 41%
ACHTREUNG
for an alcohol conversion of 65%.This paper reports a new
method of TiO preparation from aqueous solutions of TiCl
2
4
kept at low temperature (333 K) to yield a rutile phase,
which allowed a selectivity of approximately 60% to be ob-
tained under the same experimental conditions as those pre-
[15]
viously reported. It also shows how the partial oxidation
to aldehyde can be carried out with a good selectivity (ca.
2
Figure 1.XRD patterns of home-prepared and commercial TiO :
A) HP298; B) HP333; C) HP673; D) HP973; E) Sigma-Aldrich; F) De-
gussa P25; G) Merck.
38%) in the case of benzyl alcohol (BA).The obtained re-
sults have been compared with those of three commercial
samples: Sigma-Aldrich (rutile), Degussa P25 (rutile 20%,
anatase 80%) and Merck (anatase), each of which showed a
much lower selectivity, even if used in optimised amounts.
Both commercial and home-prepared samples were charac-
terised by using XRD, BET specific surface area determina-
tion and SEM.Moreover, the TiO 2 home-prepared sample
was dialysed or calcined (at 673 or 973 K) to investigate the
ature, does not show any peaks to indicate crystallinity,
probably due to the very small size of particles and to the
substantial amorphousness of this sample.HP333, prepared
by simply keeping the solution of TiCl at 333 K for two
4
days, is made of a badly crystallised rutile phase.Upon cal-
cining HP333, crystallinity increased with the temperature
treatment; the HP973 diffractogram seems quite similar to
that of the Sigma-Aldrich rutile.
Crystallite and particle agglomeration size along with spe-
cific surface area (SSA) strongly change with treatment tem-
perature (see Table 1).Primary crystallite size, determined
by using the Scherrer equation, shifts from 7 to 41 nm for
home-prepared catalysts, whereas commercial samples show
sizes ranging from 30 to 60 nm.The highest values were ob-
À
effect of residual Cl ions and heating temperature on crys-
tallinity, and hence reaction rate and selectivity.
The concentrations of aromatic alcohols and their oxida-
tion products were monitored during irradiation, and the
amount of mineralisation was recorded by means of a total
organic carbon (TOC) analyser.The aromatic molecules ob-
tained from the oxidation were identified by using HPLC
and GC–MS analyses.The different BA and MBA reaction
rates and selectivities towards aldehyde are discussed and
an explanation is provided on the basis of our results.
served for HP973 and for commercial TiO , that is, the most
2
crystalline catalysts.Particle agglomeration, measured by
using SEM images (not shown for the sake of brevity),
varies from approximately 50 to 215 nm, increasing with
preparation temperature.Specific surface area, along with
crystallite size, depends on the temperature used during the
preparation process.These qualities drastically decrease and
increase, respectively, by raising the temperature.SSA
Results and Discussion
Catalysts: In the present study the oxidation of BA and of
MBA was performed in a batch photoreactor that contained
aqueous suspensions of home-prepared rutile.Commercial
samples, consisting of rutile (Sigma-Aldrich), anatase
2
À1
values between 4 and 215 m g were obtained for home-
Chem. Eur. J. 2008, 14, 4640 – 4646
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4641

V.Augugliaro, L.Palmisano et al.
Table 1.Preparation temperature, BET specific surface area, and crystal-
lite size of the catalysts and their photocatalytic performance for the oxi-
dation of aromatic alcohols to aldehydes (catalyst amount: 0.4 gL ).
thus highlighting that even in this case selectivity does not
change considerably for all of the catalysts.
À1
From the start of irradiation two parallel reaction path-
ways take place: direct mineralisation to CO2 (occurring
through a series of reactions occurring over the catalyst sur-
face and producing intermediates not desorbing to the bulk
of the solution) and partial oxidation to aldehyde.The im-
portance of each of these pathways depends on the physico-
chemical and structural properties of the catalyst, and also
on the features of the reacting alcohols and produced alde-
hydes.
[
a]
[a]
[a]
irr.
[a]
Catalyst
SSA Crystallite
t
irr.
Selectivity
BA BA
[h] [%mol]
t
Selectivity
size
[m g] [nm]
MBA MBA
[h]
2
A
C
H
T
R
E
U
N
G
A
T
E
N
A
H
R
U
G
HP298
HP333
HP333D
HP673
HP973
Sigma-
Aldrich
Merck
Degussa
P25
215
107
107
35
–
7
7
13
41
22.0 12.1
8.4 38.2
5.3 33.5
6.0 12.2
9.4 9.9
3.8 9.2
–
–
2.36
2.18
2.20
3.81
2.15
60.0
56.2
50.1
40.0
20.9
4
2.5 52
During the BA and MBA photocatalytic oxidation, CO2
10
50
60
30
2.0 7.9
1.0 7.8
1.50
0.72
15.9
8.3
was detected and the main products formed were deter-
mined by using TOC and HPLC analyses, respectively.Only
when commercial or calcined (HP673 and HP973) catalysts
[
a] tirr. (irradiation time) and selectivity (to aldehyde) refer to an alcohol
conversion of 50%.Selectivity =(formed aldehyde)/(converted alcohol).
were used did we detect CO and the corresponding aromat-
2
ic aldehydes together with many other products, such as
mono- and dihydroxylated aldehydes.On the contrary, for
HP333 the predominant products were the corresponding al-
prepared samples, with the lowest ones being very close to
that of the Sigma-Aldrich sample.
dehydes and CO , the other products were present only as
2
traces.During BA oxidation, traces of benzoic acid were
also found no matter which catalyst was used.
Figures 2 and 3 show the trends of the reagent and main
Photoreactivity experiments: Preliminary runs indicated
that no conversion of either alcohol was observed when the
experiment was carried out in a dark environment, both in
the absence and in the presence of catalyst, even when con-
tinuously bubbling oxygen through the sample.Moreover,
oxidation of BA and MBA was carried out in irradiated
aqueous solutions, but in the absence of a catalyst under ex-
perimental conditions similar to those used during the het-
products (aldehyde and CO ) during the photocatalytic oxi-
2
dation of BA and MBA carried out with the Sigma-Aldrich
A
C
H
T
R
E
U
N
G
erogeneous experiments.BA and MBA conversion of 50%
A
C
H
T
R
E
U
N
G
required reaction times of 176 and 21 h, respectively.As this
conversion was reached in much shorter times in photocata-
lytic runs, it may be assumed that the homogeneous oxida-
tion process plays a negligible role in our heterogeneous sys-
tems.
Photocatalytic oxidation runs were carried out with
À1
amounts of HP333 in the range of 0.2 to 0.6 gL to deter-
mine the amount of catalyst capable of maximising the se-
lectivity towards aldehyde (calculated as the ratio between
the formed aldehyde and the converted alcohol).This study
showed that the selectivities are indeed very close to each
other, regardless of the amount of catalyst.Reaction rate
values are quite different; they increase when using a cata-
À1
lyst amount from 0.2 to 0.4 gL and remain constant for
higher amounts of catalyst, thus indicating that for those
amounts all of the emitted photons are absorbed by the sus-
pension.Consequently, the catalyst amount used in all of
À1
the BA and MBA oxidation runs was 0.4 gL .Based on the
above data, this quantity also determined the lowest irradia-
tion time needed to reach 50% alcohol conversion.All of
the other runs were performed with this amount of TiO to
2
compare the results obtained with HP333 with those of
other catalysts.It should be emphasised that the irradiated
surface of the samples used was very different in some
cases.Selectivity, however, was not significantly influenced
Figure 2.Experimental results of a representative oxidation run of BA
(
A) and MBA (B) with rutile TiO
2
from Sigma-Aldrich.Conversion ( *)
(~) concentra-
and selectivity (*) are scaled on the right side.The CO
2
in any case by the amount of solid present in the reacting
tion values were divided by 7 (A) or 8 (B) for normalisation purposes. &
[alcohol], ^: [aldehyde] and ~: [other species].
:
À1
medium.The catalyst amounts were decreased to 0 0. 2 gL
,
4642
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4640 – 4646

Oxidation of Aromatic Alcohols
FULL PAPER
ples, respectively.Interestingly, an increase in crystallinity
by means of thermal treatment resulted in selectivities even
lower than that of HP333.
The most selective rutile sample was HP333 (Table 1),
and the selectivity values drastically decreased from 38.2 to
9
.9%mol (for BA) by increasing the treatment temperature
at which this catalyst was subjected.The selectivity of com-
mercial TiO catalysts was four to five times lower than that
2
of HP333, of which the least selective sample was Degussa
P25.As far as MBA is concerned, the selectivity decreased
from 60%mol for HP333 to 8.3%mol for Degussa P25.
The reported selectivity clearly shows an opposing trend
with respect to catalyst activity if comparing HP333 with
commercial samples.The most active catalysts (commer-
cials) are the least selective in the formation of aldehyde.
Furthermore, by comparing home-prepared samples, one
can understand how increasing the treatment temperature
results in decreasing selectivity; this phenomenon can be as-
cribed to a drastic depletion of the surface hydroxylation of
HP673 and HP973.
HP333 powder was subjected to a dialysis treatment to
study the influence of chloride ions on selectivity, as the
samples obtained from TiCl can contain not negligible re-
4
À
siduals of Cl .The results attained with HP333D (the dia-
lysed sample) highlighted only a small decrease in selectivity
(
from 38.2 to 33.5% for BA), whereas a significant increase
Figure 3.Experimental results of a representative oxidation run of BA
in activity was evidenced: irradiation time needed to convert
50% of the initial alcohol decreased from 8.4 to 5.3 h. This
finding, which points to a higher mineralisation when using
the dialysed sample, can be explained by the well-known
detrimental effect of Cl ions owing to their good properties
as hole traps.
(
(
A) and MBA (B) with rutile TiO
*) are scaled on the right side.The CO
2
HP333.Conversion ( *) and selectivity
) concentration values were
(~
2
divided by 7 (A) or 8 (B) for normalisation purposes.
: [alcohol], ^: [al-
&
dehyde] and ~: [other species].
À
[16]
rutile and with the best-performing home-prepared rutile
Finally, the run carried out with the catalyst synthesised at
298 K, by simply adding an aqueous solution of NaOH to a
(
HP333).
The global amount of the organic carbon belonging to
clear solution of TiCl , gave worse results than that using
4
identified (but not quantified separately) and unidentified
mainly open-ring products) compounds is also reported.
The values reported in Figures 2 and 3 were evaluated from
the difference between initial TOC and the sum of residual
HP333 in terms of both activity and selectivity.The former
can be attributed to the absence of crystallinity highlighted
by XRD analysis (Figure 1).The reaction time needed to
reach 50% conversion was, however, 22 h, a much lower
figure than that found during the homogeneous reaction
(176 h).In any case, the catalyst was found to be active, and
this insight could be explained by two different hypotheses:
1) the presence of small amounts of crystallites, not detecta-
ble through XRD analysis; 2) the occurrence of a photore-
action in the adsorbed phase.On the other hand, selectivity
was considerably lower than that of HP333, highlighting
that lower activity does not always result in higher selectivi-
ty.
(
alcohol, produced aldehyde and CO (quantified from the
2
extent of mineralisation, measured by TOC analysis).The
figures also show conversion and selectivity values versus ir-
radiation time.This highlights how selectivity slowly de-
creases during reaction time.
The commercial Sigma-Aldrich, Merck and Degussa P25
samples produced a significant number of species different
from aldehyde and carbon dioxide (Figure 2), whereas
HP333 did not support the production of other species
(
Figure 3).This different behaviour can be ascribed to the
considerable presence of hydroxylated aromatic compounds
and open-ring products, which were visible in HPLC chro-
matograms at very low retention times only when commer-
cial and calcined samples were used.
The activity of the home-prepared rutile samples was
always much lower than that of commercial catalysts.HP333
needed 8.4 h to reach 50% BA conversion, against 3.8, 2.0
and 1.0 h for Sigma-Aldrich, Merck and Degussa P25 sam-
Mechanistic aspects: The obtained results clearly indicate
that, whereas home-prepared rutile (HP333) made at low
temperature gave rise only to aldehyde and CO for oxida-
2
tion of both aromatic alcohols, the commercial catalysts and
the calcined home-prepared ones yielded a very unselective
reaction, with the formation of hydroxylated aldehydes and
open-ring products detected during BA oxidation, besides
aldehyde and CO2.
Chem. Eur. J. 2008, 14, 4640 – 4646
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4643

V.Augugliaro, L.Palmisano et al.
Scheme 1.Proposed reaction schemes (it should be considered that all the aromatic compounds reported in the scheme can undergo direct mineralisatio n
to CO ).R =H, OCH .
2 3
The selectivity results of the oxidation of alcohol to alde-
hyde are surprisingly high (up to 60% for MBA; the highest
value previously reported for MBA was 41%), and they
were found by using a home-prepared anatase catalyst. It
should be considered that the reaction takes place in water
without any organic co-solvent.Under these conditions, the
solvent represents a highly unselective medium, in which
the behaviour of hydroxyl radicals is equally unselective.In-
stead the features of the synthesised catalysts allow high se-
lectivity even in this environment.
ic ring takes place in ortho and para positions when an elec-
tron-donor group is present; to the contrary, in its absence
all three of the isomers are formed.
Previously reported results for anatase showed that the
primary oxidation step of MBA involves in any case either
[15]
[15]
direct mineralisation to CO or formation of an aldehyde.
2
However, no hydroxylated 4-methoxybenzaldehyde isomers
were detected.These species were detected in our investiga-
tion in trace amounts even when using commercial and
highly crystalline (home-prepared) samples.
Table 1 shows how an increase in crystallinity causes a sig-
nificant decrease in selectivity, if we do not consider the
amorphous HP298.This could be ascribed not only to the
different crystallinity, but also to a drastic reduction of sur-
face hydroxyl groups, and hence hydrophilicity of samples,
occurring upon calcination.A current FTIR investigation
we are carrying out is showing how this could determine a
greater difficulty for the produced aldehyde to desorb from
the catalyst surface to the bulk of solution, thus undergoing
further oxidation.
Scheme 1 shows the hypothesised mechanism for the two
alcohols.Both BA and MBA can either undergo direct min-
eralisation to CO (not shown) through adsorbed intermedi-
2
[15]
ates, or be partially oxidised to other aromatic and even-
tually aliphatic compounds.
The first step (Scheme 1A) of BA and MBA partial oxida-
tion to an aldehyde is the abstraction of an electron in the
+
OH group by a hole (h ; abstraction from an aromatic ring
is indeed much more difficult).Once this rate-determining
step has taken place, the subsequent transformations involve
the formation of a ÀCHO group by means of either an elec-
Upon using highly crystalline samples, results of GC–MS
analyses have highlighted the presence of the three isomers
of hydroxybenzaldehyde, but no presence of hydroxybenzyl
alcohol was observed.This could be caused by the much
higher ease of transformation of an aromatic alcohol into an
aldehyde rather than into hydroxybenzyl alcohol isomers.
Moreover, 2,5-dihydroxybenzaldehyde was also detected,
but the presence of other isomers obtained from the en-
trance of a hydroxyl radical at the ring position activated by
the first hydroxyl radical can also be hypothesised, following
tron hole or an OH radical.
In addition, it is worth noting that O2 could play a role
À
in producing oxidant species (for instance HO radicals) by
2
reacting with H O.
2
Benzaldehyde, once formed, can be subjected to a further
attack of OH radicals on the aromatic ring (Scheme 1, step
B).This is the rate-determining step for the production of
hydroxylated aromatic compounds through the elimination
of a hydrogen atom by means of either an electron hole or
an OH radical.All of the monohydroxylated isomers are ob-
tained because an electron-withdrawing group (ÀCHO) is
[8]
a previously reported general behaviour. These recent re-
sults showed that the attack of an OH radical on an aromat-
4644
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4640 – 4646

Oxidation of Aromatic Alcohols
FULL PAPER
[8]
present (general behaviour previously observed). Analo-
gously, the second hydroxylation can take place in the posi-
tion activated by the first ÀOH group.
was shown to be significantly higher for home-prepared
rutile TiO made at low temperature (HP333) relative to
2
other home-prepared and commercial highly crystalline cat-
alysts.The obtained selectivity was three- to sevenfold
higher for both alcohols, and no byproducts were found
when working with HP333, with the exception of CO2.
Highly crystalline samples displayed, conversely, a low selec-
tivity and in the case of BA gave hydroxylated byproducts
and aromatic acid.The very different selectivities (60 and
38% for MBA and BA, respectively) were justified by con-
sidering the presence of an electron-donor group in the
former case and the hydrophobic nature of benzaldehyde.
The mild experimental conditions used, consisting of
aqueous suspensions free of any organic co-solvents, and the
absence of heavy metals, high temperatures or pressures,
allow us to propose this process as a green route for partial
oxidation of alcohols.
A different fate for benzaldehyde can be oxidation to
benzoic acid by means of hydrogen-atom abstraction from
the aldehydic carbon atom, with formation of an acyl radical
(
Scheme 1, step B).Two possibilities of how this occurs
could be 1) an electron transfer to an acyl cation followed
by the attack of a water molecule as depicted in Scheme 1,
or 2) a cross-dimerisation with an OH radical.Considering
that the concentration of these two different radicals should
be rather low, we prefer the first hypothesis.
4-Methoxybenzaldehyde, on the other hand, is only sub-
jected to direct mineralisation, because neither hydroxylated
species nor acid were formed in detectable amounts when
using catalysts prepared at low temperature.
The presence of hydroxylated aldehydes (step B), detect-
ed only from BA and with the most crystalline catalysts, sug-
gests that the higher oxidising power of these samples can
allow the hydroxylation reaction, which decreases the selec-
tivity towards aldehyde.So the mild oxidising power of
HP333 gives rise to a high selectivity, due to the greater dif-
ficulty of hydroxylating the obtained aldehyde or directly
breaking the aromatic ring.
Selectivity towards aldehyde, reaction rate and intermedi-
ates were found to be very different for BA and MBA.This
could be ascribed to the presence of an electron-donor group,
such as methoxy, in the para position of MBA.The signifi-
Experimental Section
Preparation of home-prepared TiO
obtained by slowly adding TiCl (20 mL, >97%, Fluka) dropwise into a
2 L beaker containing water (1 L) under agitation because the hydrolysis
of TiCl is a highly exothermic reaction producing significant amounts of
2
catalysts: The precursor solution was
4
4
HCl vapour.The addition had a total duration of 5 min.After magnetic
mixing (450 rpm) for 10 min, the resulting solution was kept in a closed
oven for 2 d at 333 K.Then a white suspension was obtained.Drying was
carried out at 333 K by means of a rotary evaporator (Büchi Rotavapor
M) working at 50 rpm to eventually obtain the powdered catalyst
cantly higher reaction rate of MBA can be ascribed to a
+
(
HP333).To check the influence of the degree of crystallinity, HP333 was
much easier abstraction of a hydrogen atom by a hole (
G
U
N
E
T
G
U
H
C
A
h )
N
U
N
R
E
T
N
G
T
U
H
A
C
calcined at 673 and 973 K for 6 h.The obtained samples were named
HP673 and HP973, respectively.Another sample (HP333D) was pre-
pared by dialysing HP333 by using a dialysis tubing cellulose membrane
(average flat width=76 mm, 12400 MW cut-off pores, Aldrich) for ap-
proximately 3 d, changing the distilled water every 8 h (5 L) until the
conductivity value of water was found to be negligible.An amorphous
due to the presence of a methoxy group because it increases
the delocalisation on the ring.This donor group could be ef-
fective as a hole trap because it contains an oxygen atom and
can stabilise the aromatic ring, thus increasing the selectivity
towards aldehyde.Regarding this transformation, it was
found in the literature that electron transfers in the gas phase
and in solution occur from the oxygen atom of the benzylic
TiO
2
powder (HP298) was also prepared by following the hereafter de-
(2.5 mL) was added dropwise into a 100 mL
scribed procedure.TiCl
4
beaker containing water (25 mL).During the addition, which lasted
5 min, the solution was magnetically stirred at 450 rpm.Upon closing the
beaker and mixing the solution for 12 h at room temperature, a clear so-
lution was eventually obtained.Afterwards the prepared solution was
added to 1m NaOH (90 mL) under magnetic agitation.The resulting sus-
pension was filtered and washed until the washing water was found to
have a neutral pH and a negligible conductivity.
[17]
alcohol and this transfer is easier for MBA than BA.
The presence of some amounts of benzoic acid on the one
hand, and the absence of p-methoxybenzoic acid (for con-
version <80%) on the other suggest that electron-donor
groups inhibit the formation of aromatic acid, as previously
[18]
reported.
Photocatalytic experimental procedure: A cylindrical Pyrex batch photo-
reactor with immersed lamp, containing the aqueous suspension (0.5 L),
was used to perform the reactivity experiments.The initial alcohol con-
centration was approximately 1 mm.The photoreactor was provided with
ports in its upper section for the oxygen inlet and outlet and for sam-
pling.A magnetic stirrer guaranteed a satisfactory suspension of the pho-
tocatalyst and the homogeneity of the reacting mixture.A scheme of the
reactor can be seen in ref.[8b]. A medium-pressure Hg lamp (125 W,
Helios Italquartz, Italy) axially positioned within the photoreactor was
cooled by water circulating through a Pyrex thimble; the temperature of
the suspension was about 300 K.The radiation energy impinging on the
The different amount of CO2 produced from BA and
MBA could also be attributed to the very different solubility
of benzaldehyde and 4-methoxybenzaldehyde in water (100
À1
vs.4290 mgL
at 293 K, respectively).Desorption of ben-
zaldehyde from the catalyst surface after its formation can
therefore be prevented by its hydrophobic character, so that
either a subsequent hydroxylation or mineralisation to CO2
would be favoured.
À2
suspension had an average value of 10 mWcm .It was measured by
using a UVX digital radiometer at l=360 nm.Before switching on the
lamp, oxygen was bubbled into the suspension for 30 min at room tem-
perature to reach the thermodynamic equilibrium.Adsorption of the al-
cohols in the dark was always quite low, less than 3%.Liquid samples
Conclusion
In conclusion, the oxidation of two aromatic alcohols in
water has been reported, and selectivity towards aldehyde
(
containing suspended catalyst powder) were taken at fixed time intervals
and filtered through a 0.45 mm hydrophilic membrane (HA, Millipore)
Chem. Eur. J. 2008, 14, 4640 – 4646
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4645

V.Augugliaro, L.Palmisano et al.
before being analysed.An aqueous solution of 1 m NaOH was used to
adjust the initial pH to seven for the performed runs.All of the chemicals
used were purchased from Sigma-Aldrich with a purity of >98%.
López, V.Loddo, G.Marci, R.Molinari, L.Palmisano, M.Schiavel-
lo, J. Phys. Chem. B 2004, 108, 3303–3310, and references therein.
[4] a) M.Nag, P.Basak, S V. .Manorama,
Mater. Res. Bull. 2007, 42,
1
2
1
691–1704; b) S.Yin, R.Li, Q.He, T.Sato,
Mater. Chem. Phys.
002, 75, 76–80; c) L.Li, J.Liu, Z.Jia, Mater. Lett. 2006, 60, 1753–
757.
Analytical techniques: The quantitative determination and identification
of the species present in the reacting suspension was performed by
means of a Beckman Coulter HPLC instrument (System Gold 126 sol-
vent module and 168 diode array detector) equipped with a Luna 5 mm
phenyl–hexyl column (250 mm long”2 mm i.d.) and using Sigma-Aldrich
standards.The retention times and UV spectra of the compounds were
compared with those of an authentic sample.The eluent consisted of
[
[
5] a) A.Mills, S.Le Hunte, J. Photochem. Photobiol. A 1997, 108, 1–
5; b) H.Kisch, Adv. Photochem. 2001, 26, 93–143; c) G.Palmisano,
3
V.Augugliaro, M.Pagliaro, L.Palmisano,
425–3437.
6] a) H.Yoshida, C.Murata, T.Hattori,
Chem. Commun. 2007,
3
Chem. Commun. 1999, 1551–
1
2 4
7.5% acetonitrile, 17.5% methanol and 65% 40 mm KH PO aqueous
1
552; b) K.V. Subba Rao, M. Subrahmanyam, Photochem. Photo-
solution.TOC analyses were carried out by using a 5000 A Shimadzu
TOC analyser.GC–MS analyses were performed by means of a GC–MS
Perkin–Elmer Turbomass and an Autosystem XL chromatograph; a SGE
BPX5 capillary column (25 m”0.22 mm i.d.) was used. The injected solu-
tion was prepared by following the subsequent procedure: the reaction
suspension (100 mL) was filtered and extraction was carried out by using
diethyl ether (20 mL); afterwards the organic phase was concentrated 20
biol. Sci. 2002, 1, 597–599; c) A.Maldotti, R.Amadelli, L.Samiolo,
A.Molinari, A.Penoni, S.Tollari, S.Cenini,
749–1751.
7] M.A.Gonzalez, S.G.Howell, S.K.Sikdar,
62.
Chem. Commun. 2005,
J. Catal. 1999, 183, 159–
1
[
[
1
8] a) G.Palmisano, M.Addamo, V.Augugliaro, T.Caronna, E.García-
López, V.Loddo, L.Palmisano, Chem. Commun. 2006, 1012–1014;
b) G.Palmisano, M.Addamo, V.Augugliaro, T.Caronna, A.Di Pao-
la, E.García-López, V.Loddo, G.Marcì, L.Palmisano, M.Schiavel-
lo, Catal. Today 2007, 122, 118–127.
9] F.Soana, M.Sturini, L.Cermenati, A.Albini,
Trans. 2 2000, 699–704.
10] T.Caronna, C.Gambarotti, L.Palmisano, C.Punta, F.Recupero,
Photochem. Photobiol. A 2005, 171, 237–242.
11] B.Ohtani, S.Tsuru, S- .I .Nishimoto, T.Kagiya, K.Izawa,
Chem. 1990, 55, 5551–5553.
12] a) K.Yamaguchi, N.Mizuno, Angew. Chem. 2002, 114, 4720–4724;
Angew. Chem. Int. Ed. 2002, 41, 4538–4542; b) G.J. ten Brink,
times.XRD patterns of the powders were recorded by using a Philips dif-
À1
fractometer using the CuKa radiation and a 2q scan rate of 1.288min
.
SEM images were obtained by using a microscope (Philips XL30 ESEM)
operating at 25 kV on samples sprayed on the stab and dried at room
temperature, upon which a thin layer of gold had been evaporated.BET
specific surface areas were measured by using the single-point BET
method using a Micromeritics Flow Sorb 2300 apparatus.
[
J. Chem. Soc. Perkin
[
[
[
J.
J. Org.
Acknowledgements
I.W.C.E.Arends, R.A.Sheldon,
13] O.S.Mohamed, A.E.M.Gaber, A.A.Abdel-Wahab,
Science 2000, 287, 1636–1639.
The support of Dr.M.A.Sabatino (Dipartimento di Ingegneria Chimica
dei Processi e dei Materiali, Università degli Studi di Palermo) for GC–
MS analysis is gratefully acknowledged.
[
J. Photochem.
Photobiol. A 2002, 148, 205–210.
[
[
14] U.R.Pillai, E.Sahle-Demessie, J. Catal. 2002, 211, 434–444.
15] G.Palmisano, S.Yurdakal, V.Augugliaro, V.Loddo, L.Palmisano,
Adv. Synth. Catal. 2007, 349, 964–970.
[
1] A.Fujishima, K.Honda, Nature 1972, 238, 37–38.
[
[
[
16] C.Minero, G.Mariella, V.Maurino, E.Pelizzetti,
6, 2632–2641.
17] A.J. McEvoy, M. Grꢁtzel, Sol. Energy Mater. Sol. Cells 1994, 32,
21–227.
Langmuir 2000,
[
2] a) Photocatalysis and Environment: Trends and Applications (Ed.:
M.Schiavello), Kluwer, Dordrecht, 1988; b) M.R. Hoffmann, S.T.
1
Martin, W.Choi, D .W .Bahnemann,
c) V.Augugliaro, M.Litter, L.Palmisano, J.Soria,
Photobiol. C 2006, 7, 127–144; d) L.Zang, W.Macyk, C.Lange,
W.F.Maier, C.Antonius, D.Meissner, H.Kisch, Chem. Eur. J. 2000,
, 379–384.
3] a) A.Sclafani, L.Palmisano, M.Schiavello, J. Phys. Chem. 1990, 94,
29–832; b) M.Addamo, V.Augugliaro, A.Di Paola, E.García-
Chem. Rev. 1995, 95, 69–96;
2
J. Photochem.
18] E.Baciocchi, M.Bietti, B.Chiavarino, M.E.Crestoni, S.Fornarini,
Chem. Eur. J. 2002, 8, 532–537.
6
[
Received: January 17, 2007
Published online: April 8, 2008
8
4646
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 4640 – 4646