Partial photocatalytic oxidation of glycerol in TiO2 water suspensions

Article abstract of DOI:10.1016/j.cattod.2010.01.022

The photo-oxidation of glycerol in aqueous suspensions containing commercial and home-prepared TiO₂ samples in the anatase, rutile or anatase-rutile polymorphic phases was investigated by using two different batch photoreactors, an annular photoreactor and a cylindrical photoreactor. The glycerol oxidation products detected in the aqueous phase were: 1,3-dihydroxyacetone, glyceraldehyde, formic acid and carbon dioxide. Two unknown compounds were also detected by electrospray ionization mass spectrometry as peaks at m/z 268 and m/z 176. A plausible mechanism for their formation was suggested based on ESI-MS and MSⁿ experiments. The results indicate that in both systems the commercial samples showed the best performances both for reaction rate and selectivity toward 1,3-dihydroxyacetone, glyceraldehyde, formic acid and CO₂.

Full text of DOI:10.1016/j.cattod.2010.01.022

Catalysis Today 151 (2010) 21–28
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Partial photocatalytic oxidation of glycerol in TiO2 water suspensions
b
a
c
V. Augugliaro a, , H.A. Hamed El Nazer , V. Loddo , A. Mele , G. Palmisano a,d, L. Palmisano a,
*
S. Yurdakal a,e,f
a
`
‘‘Schiavello-Grillone’’ Photocatalysis Group, Dipartimento di Ingegneria Chimica dei Processi e dei Materiali, Universita di Palermo, Viale delle Scienze, 90128 Palermo, Italy
b Photochemistry Department, National Research Centre, Al-Behoos St. (El-Tahrir street), 12622 Dokki-Cairo, Egypt
c Dipartimento di Chimica, Materiali e Ingegneria Chimica G. Natta, Politecnico di Milano, Via L. Mancinelli 7, 20131 Milano, Italy
d
`
Dipartimento di Ingegneria Elettronica, Universita degli Studi di Roma ‘‘Tor Vergata’’, Via del Politecnico 1, 00133 Roma, Italy
e
f
¨
Kimya Bo¨lu¨mu¨, Fen Faku¨ltesi, Anadolu Universitesi, Yunus Emre Kampu¨su¨, 26470 Eskis¸ehir, Turkey
¨
Kimya Bo¨lu¨mu¨, Fen-Edebiyat Faku¨ltesi, Afyon Kocatepe Universitesi, Ahmet Necdet Sezer Kampu¨su¨, 03100 Afyon, Turkey
A R T I C L E I N F O
A B S T R A C T
Article history:
The photo-oxidation of glycerol in aqueous suspensions containing commercial and home-prepared TiO2
samples in the anatase, rutile or anatase–rutile polymorphic phases was investigated by using two
different batch photoreactors, an annular photoreactor and a cylindrical photoreactor. The glycerol
oxidation products detected in the aqueous phase were: 1,3-dihydroxyacetone, glyceraldehyde, formic
acid and carbon dioxide. Two unknown compounds were also detected by electrospray ionization mass
spectrometry as peaks at m/z 268 and m/z 176. A plausible mechanism for their formation was suggested
based on ESI-MS and MSn experiments. The results indicate that in both systems the commercial samples
showed the best performances both for reaction rate and selectivity toward 1,3-dihydroxyacetone,
glyceraldehyde, formic acid and CO2.
Available online 18 February 2010
Keywords:
Glycerol partial oxidation
Selective photocatalytic oxidation
Titanium dioxide
Reaction pathways
ß 2010 Elsevier B.V. All rights reserved.
1. Introduction
Among alcohols, glycerol (1,2,3-propanetriol) has a widespread
application; this molecule is used in toothpastes, pharmaceuticals,
With an increase in environmental consciousness throughout the
world, there is a challenge for chemists and chemical engineers to
develop new products, processes and services that achieve necessary
social, economical and environmental objectives. The Green
Chemistry (environmentally benign chemistry) and Green Engineer-
ing areas, developed in these last years, are the answers of the
scientific community to that challenge. As well as using and
producing better chemicals with less waste, the challenge also
involves reducing other associated environmental impacts including
reduction in the amount of energy used in chemical processes [1–3].
The selective oxidation of alcohols is an important preparative
method in organic chemistry. The resulting products, aldehydes
and ketones [4–6], are extensively used as precursors and
intermediates for the production of flavours, fragrances, and
biologically active compounds. The application of these prepara-
tions, however, causes severe environmental problems due to
waste from heavy metals or stoichiometric amounts of reagents
and the generation of undesirable co-products [7–11]. From the
standpoint of green and sustainable chemistry, there is still a need
to develop cleaner catalytic oxidation systems.
cosmetics, soaps and as sweetener in cakes and wetting agent in
tobacco. Glycerol is now available in very huge amounts [12]; it is a
co-product of oils and fats industries which have grown very much
in last years. Moreover, nowadays glycerol is also obtained from
biodiesel: one ton of biodiesel produces in fact about 100 kg of pure
glycerol. Only small amounts of glycerol are used to synthesise
other useful molecules as for instance polyethers and esters [12]. It
is therefore of great concern to try to make value-added products
from glycerol as an alternative to disposal by incineration.
Consequently the researches aimed at the obtainment of
functionalized glycerol in order to take advantage of its wide
availability are up to date.
Many products can be obtained from glycerol partial oxida-
tion: 1,3-dihydroxyacetone (DHA), glyceraldehyde (GAD), glyce-
ric acid, glycolic acid, hydroxypyruvic acid, mesoxalic acid, oxalic
acid, tartronic acid and all of them are considered chemicals for
highly specialized applications. 1,3-Dihydroxyacetone is formed
when the secondary hydroxyl group of glycerol is selectively
oxidised and it is an important chemical used in the cosmetics
industry as a tanning substance [13]. In addition to its main
utilization in cosmetics, 1,3-dihydroxyacetone has also other
applications, for instance as a monomer in polymeric biomaterials
[12]. The oxidation of the primary hydroxyl groups of glycerol
produces GAD, an intermediate of the carbohydrate metabolism.
* Corresponding author. Tel.: +39 091 23863720; fax: +39 091 7025020.
E-mail address: augugliaro@dicpm.unipa.it (V. Augugliaro).
0920-5861/$ – see front matter ß 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2010.01.022

22
V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
Glyceraldehyde is also a standard by which chiral molecules of the
- or -series are compared.
Glycerol catalytic aerobic oxidation has been studied [12–20]
drop into a 200 mL beaker containing 50 mL of water. During the
addition, that lasted 5 min, the solution was magnetically stirred
by a cylindrical bar at 600 rpm. After that the beaker was closed
and mixing was prolonged for 12 h at room temperature,
eventually obtaining a clear solution. This solution was transferred
into a round-bottom flask fitted with a graham condenser. The
flask was heated in boiling water for 0.5 h, obtaining a white
suspension at the end of each treatment. The suspensions were
then dried at 323 K by means of a rotovapor machine (model Buchi
Rotovapor M) working at 50 rpm, in order to obtain the final
powdered catalysts. Hereafter the home-prepared anatase catalyst
is referred to as HPA.
D
L
by using Pd/C, Pt/C and more recently Au/C catalysts under
different experimental conditions, and the distribution of the
products and the selectivity were controlled by changing the
temperature and the pH values of the reaction environment. For
instance the oxidation in water over conventional precious metal
based catalysts such as Au/C and Pt/C yielded glyceric acid [19,20].
It has been described also the oxidation of both primary hydroxyl
groups of glycerol in the presence of Pt/CeO2 catalysts that
produced tartronic acid with a selectivity of 40% [21].
Electro-catalytic oxidation of glycerol mainly to DHA with
tetramethyl piperidine-1-oxyl (TEMPO) as the oxidant agent and
Ag/AgCl as the catalyst has been reported by Ciriminna et al. [13].
Only for long reaction times significant amount of hydroxyl-
pyruvic acid was found.
Photocatalytic oxidation of some aromatic alcohols in water has
been recently performed with high selectivity to the corresponding
aldehydes, using home-prepared anatase, brookite or rutile TiO2
[22–29].
As far as glycerol is concerned, Lichtin et al. [30] first reported
its oxidative decomposition in liquid water by a TiO2-catalyzed
photoprocess. As to concern the photocatalytic partial oxidation,
up to date there is only one published work by Maurino et al. [31]
reporting the photocatalytic selective oxidation of glycerol in
water to DHA by using Degussa P25 and Merck TiO2 catalysts and
their fluorinated derivatives. By using fluorinated Degussa P25 and
very low concentration of glycerol (ca. 1 mM), almost 100%
selectivity to DHA was obtained for conversions lower than 30%
and the intermediates were similar to the oxidation products in
thermal catalysis.
In this work photocatalytic oxidation of glycerol was performed
by using an annular and a cylindrical batch type photocatalytic
reactor in the presence of commercial and home-made TiO2
samples. The intermediates were determined and two of them
(DHA and GAD) resulted the same of those obtained by thermal
catalysis, while others were not found up to date both in thermal
catalytic and photocatalytic studies.
2.1.2. Home preparation of rutile TiO2
The preparation method has been reported elsewhere [28]. The
precursor solution was obtained by adding 20 mL of TiCl4 to
1000 mL water contained in a volumetric flask; the addition was
carried out very slowly in order to avoid warming of the sol as the
TiCl4 hydrolysis is a highly exothermic reaction. At the end the
resulting sol was mixed for 2 min by a magnetic stirrer and then
the flask was sealed and maintained at room temperature (ca.
298 K) for six days. The solid powder precipitated at the end of the
whole treatment was separated by centrifugation (20 min at
5000 rpm) and dried at room temperature. The final home-
prepared rutile catalyst is hereafter indicated as HPR.
2.2. Photocatalyst characterization
X-ray diffractometry (XRD) patterns of the powders were
recorded by a Philips diffractometer (operating at a voltage of
40 kV and a current of 30 mA) using the CuK
scan rate of 1.288/min. The crystalline sizes of the samples were
determined by using the Scherrer equation [33,34].
Environmental scanning electron microscopy (ESEM) images
were obtained using an ESEM microscope (Philips, XL30) operating
at 25 kV. A thin layer of gold was evaporated on catalysts samples,
previously sprayed on the stab and dried at room temperature. The
sizes of secondary particles were determined by analyzing the
images.
a radiation and a 2u
Brunauer–Emmett–Teller (BET) specific surface areas (SSA)
were measured by the single-point BET method using a Micro-
meritics Flow Sorb 2300 apparatus. Before the measurement, the
commercial samples were dried 1 h at 373 K, 2 h at 423 K and
degassed 0.5 h at 423 K; on the contrary the home-prepared
catalysts were only degassed at room temperature for 3 h. The
tolerance of BET measurements was 4% and the SSA values for
commercial specimens were almost the same of those furnished by
producers.
2. Experimental
2.1. TiO2 photocatalysts
Two commercial TiO2 catalysts and two home-prepared TiO2
powders were used for the photoreactivity runs. The commercial
photocatalysts were: pure rutile TiO2 (Sigma–Aldrich) and TiO2
P25 (Degussa). TiO2 P25 is a standard material in the field of
photocatalytic reactions, it contains anatase and rutile phases in a
ratio of about 4:1, the content of amorphous phase being very low
2.3. Photoreactivity set up and procedure
(about 1.0%) [32]. In Table
1 some physico-chemical and
The photocatalytic oxidation of glycerol in water was carried
out by testing the TiO2 samples (see Table 1) in the contempora-
neous presence of light, dissolved oxygen and glycerol in the
concentration range of 10–180 mM. Two different batch photo-
reactors, all made of Pyrex glass, were used: the first one was an
annular photoreactor (APR, internal diameter, 72 mm; annulus
morphological properties of these powders and two other justify
home-prepared TiO2 powders are reported.
2.1.1. Home preparation of anatase TiO2
The preparation method has been reported elsewhere [23]. The
precursor solution was obtained by adding 5 mL of TiCl4 drop by
Table 1
Crystalline phases (A, anatase; R, rutile), BET specific surface area (SSA), secondary particles and crystallites size of the TiO2 samples.
Photocatalyst
Crystalline phases
SSA [m2/g]
Secondary particle size [nm]
Crystallite size [nm]
Sigma–Aldrich
Degussa P25
HPR
R
2.5
50
240
80
52
30
6.8
5
A(80%), R(20%)
R
A
118
235
719
28
HPA

V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
23
had an average value of 1.2 mW/cm2. It was measured at 360 nm
by using a radiometer UVX Digital.
Before switching on the lamp, oxygen was bubbled into the
suspension for 30 min at room temperature to reach the
thermodynamic equilibrium.
Samples of reacting suspension were withdrawn at fixed time
intervals during the runs; they were immediately filtered through
a 0.45
mm hydrophilic membrane (HA, Millipore) before being
analysed. The pH of the starting glycerol solution in the presence of
the photocatalysts was adjusted to ca. 7. During the run pH values
decrease due to production of formic acid.
2.4. Analytical techniques
The quantitative determination and identification of glycerol
and its oxidation products were performed by means of a Beckman
Coulter HPLC (System Gold 126 Solvent Module and 168 Diode
Array Detector), equipped with an Alltech IOA 1000 column. The
analyses were performed at 192 nm. Standards of glycerol, GAD,
DHA and formic acid (FA) were purchased from Sigma–Aldrich
with a purity of >99%. Retention times and UV spectra of the
compounds were compared with those of standards. The eluent
consisted of 0.005 M of H2SO4. Flow rate and analysis temperature
were 0.4 mL/min and 25 8C, respectively.
Total organic carbon analyses were carried out by using a
5000 A Shimadzu TOC analyser. Five measurements were per-
formed for each sample the lowest and the highest values were
rejected and the average value of the others was considered.
Samples of initial concentrations equal to 10, 50, 100 and 180 mM
were diluted 10, 50, 100 and 180 times, respectively.
Electrospray ionization mass spectrometry (ESI-MS) experi-
ments were carried out on a Bruker Esquire 3000+ instrument with
quadrupole ion trap detector (QITD). Stock solutions containing
about 1 mg/mL of crude product were diluted 1:100 (v/v) in a
mixture H2O/MeOH (50:50 (v/v) both HPLC grade). The solutions
thus prepared were analyzed by direct infusion in the ESI source at
20
m
L minꢀ1. Typical settings for MS acquisitions: needle voltage
4.5 kV, skimmer potential 40 V, end cap potential 250 V, N2
temperature 280 8C, gas flow 11 L/min. Mass range was scanned
from m/z 50 to m/z 500 at a rate of 13,000 (m/z)/s. MS2 and MS3
experiments were carried out by isolation of precursor ion(s) in the
quadrupole ion trap and collision with He gas.
Fig. 1. Photoreactor schemes: (a) annular photoreactor and (b) cylindrical
photoreactor.
3. Results and discussion
gap, 20 mm) with immersed lamp (Fig. 1(a)) containing 0.5 L of
aqueous suspension. It was provided with ports in its upper section
for the inlet and outlet of oxygen and for sampling. A magnetic
stirrer guaranteed a satisfactory suspension of the photocatalyst
and the homogeneity of the reacting mixture. A 125 W medium
pressure Hg lamp (Helios Italquartz, Italy) was axially positioned
within the photoreactor; this non-coherent radiation source emits
energy at the wavelengths of 313, 334, and mostly 366 nm. The
lamp was cooled by water circulating through a Pyrex thimble; the
temperature of the suspension was about 300 K. The radiation
energy impinging on the suspension had an average value of
10 mW/cm2. It was measured at 360 nm by using a radiometer
UVX Digital.
The used commercial photocatalysts showed to have lower
specific surface areas (SSA) and higher crystallite sizes than the
home-prepared ones (see Table 1). The secondary particles size of
HPR sample showed to be the highest one, probably because the
preparation method of this sample (with an aging time of six days)
allowed the crystallites growth. XRD patterns of the used photo-
The second photoreactor was a cylindrical photoreactor (CPR,
internal diameter, 32 mm) (Fig. 1(b)), containing 220 mL of
aqueous suspension. It was irradiated from the outside by means
of six 15 W Philips fluorescent lamps whose main emission peak in
the near-UV region is at 365 nm. In the course of the photocatalytic
runs the temperature of the suspension was 318 K. A reflux
condenser avoids loss of volatile organic molecules. With the six
lamps turned on, the radiation energy impinging on the suspension
Fig. 2. XRD patterns of commercial and home-prepared TiO2 samples. P25, Degussa
P25; SA, Sigma–Aldrich Rutile; HPA, home-prepared anatase; HPR, home-prepared
rutile.

24
V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
Table 2
Results of photocatalytic oxidation of glycerol obtained by using APR reactor. Used catalyst, catalyst amount, glycerol initial concentration, irradiance measured at the
external wall of reactor, reaction time required for 35% conversion (t35%) and selectivities (S) towards the known products.
Catalyst
Catalyst
Glycerol
[mM]
Leaving irradiance
[mW/cm2]
t35% [h]
SGAD [%]
SDHA [%]
SFA [%]
SCO [%]
amount [g/L]
Homogeneous
Degussa P25
Degussa P25
Degussa P25
Degussa P25
Degussa P25
Degussa P25
Sigma–Aldrich
Sigma–Aldrich
Sigma–Aldrich
HPR
–
100
100
100
100
10
–
17a
24.5
21
4.5a
7
3.7a
4
6.5a
3
0.1
0.2
0.4
0.2
0.1
0.1
0.2
0.4
0.4
0.2
0.4
0.4
0.3
0.8
17.4
20.3
11.5
12.5
16
0.15
0.02
0.02
0.8
5.7
11
10
9
4
4.2
4.8
3.1
7
70
8
6.5
28
6.5
5
175
53
0.8
14.5
20
9
5.5
3
4
17.4
12.6
23.7
18.5
11.1
10.9
13.8
21.1
100
100
10
2.63
0.8
5.5
10.5
8
5
31
5.2
5
10
6.1
3.7
2.2
4
0.8
8.5
30
100
100
10
0.8
4
2.5
2.9
2.2
5
HPR
0.02
0.02
0.8
40
4
HPR
6
1.8
6.5
HPA
100
32
7.4
a
Data obtained for 5% conversion.
catalysts are provided in Fig. 2. The main peaks of anatase and rutile
phases can be easily recognized [23,28]. It can be however noticed
that the peaks of HP samples diffractograms are much broader than
the others. This is mainly due to the absence of a thermal treatment
in their preparation (see Section 2), so that HP samples are
nanostructured and less crystalline with respect to the commercial
ones.
The contemporary presence of irradiation, catalyst and
dissolved oxygen was needed for observing a significant oxidation
of glycerol. The photoreactivity runs carried out in the absence of
catalyst by using the APR photoreactor gave rise to a low glycerol
conversion, only 5% mol after 17 h illumination. The homogeneous
conversion in the CPR photoreactor was also quite low: 12.8 h of
illumination were needed for reaching 5% mol conversion.
Consequently the homogeneous process was not considered
throughout this work because it was negligible with respect to
the heterogeneous one.
concerned, the reaction rate decreased by increasing the photo-
catalyst amount from 0.2 to 0.4 g/l, remaining significantly higher
with respect to P25, but the selectivity increased. Selectivities
towards GAD were quite similar to those of P25, whereas
selectivities towards DHA and FA were higher for P25 and SA,
respectively. The best selective sample between HPR and HPA was
the latter one, although these selectivity values towards GAD and
DHA were lower than that obtained in the presence of the
commercial samples. The results obtained by using Degussa P25
showed a higher accumulation of products when the leaving
irradiance (i.e., the irradiance measured at the external wall of
reactor) was lower than ca. 1% of the incident one, differently from
the experiments carried out with values equal to ca. 8%. This could
be due to an unsatisfactory irradiation of a fraction of the
photocatalyst (present in excess when the transmitted light was
very low). For Degussa P25 an amount of 0.1 g/L appeared to be the
best compromise between reaction rate and selectivity.
The glycerol oxidation products detected in the liquid phase
were: 1,3-dihydroxyacetone, glyceraldehyde, formic acid and
carbon dioxide. Two unknown compounds were also detected
by ESI-MS as peaks at m/z 176 (hereafter indicated as ‘‘Product A’’)
and m/z 268 (hereafter indicated as ‘‘Product B’’).
The global selectivity to GAD, DHA, FA and CO2 had an average
value of ca. 35% mol with respect to the converted glycerol. HPLC
analyses indicated the presence of other compounds (named
Product A and Product B, with m/z 176 and m/z 268, respectively, as
showed in ESI-MS results reported in the following).
The process performance was followed by measuring the values
of glycerol, DHA, GAD and FA concentrations and calculating the
substrate conversion and the product selectivities, i.e., the ratio
between the produced moles of DHA, GAD or FA and the moles of
converted glycerol. CO2 concentration during the reaction was
determined by subtracting the total organic carbon (TOC)
concentration, experimentally determined at a certain time, from
the initial concentration of organic carbon (entirely deriving from
glycerol). As formic acid and carbon dioxide only contain one C
atom, FA and CO2 selectivities were calculated dividing by three
their concentrations in order to have a correct comparison.
At the end of each run the carbon contained in the remaining
glycerol and in the produced GAD, DHA, FA and CO2 was calculated.
As expected, the balance was not satisfactory because in the
absence of commercial standards it was not possible to determine
the amounts of Product A and Product B.
It is important to report that, even for long lasting runs, the
catalyst powder did not show any modification and activity loss
throughout the experiments.
Table
3 reports the results of glycerol photo-oxidation
performed by using CPR reactor. Under the used experimental
conditions, both reaction rate and selectivity increased for all the
samples, due to the different reactor geometries and irradiation
fields. It can be noticed that this system seemed globally more
efficient than APR reactor. Indeed, the times required for a
conversion of 35% are lower and the selectivities towards the
products are higher than those obtained by using the APR reactor.
The best performances were obtained by using 0.091 g/L of
Degussa P25 with an incident photon flow equal to 0.6 mW/cm2.
The global selectivity to GAD, DHA, FA and CO2 had an average
value of ca. 45% mol with respect to the reacted glycerol. HPLC
analyses, moreover, again indicated the presence of Product A and
Product B previously mentioned (with m/z 176 and m/z 268,
respectively, as showed in ESI-MS results in the following).
Fig. 3 shows the data of a representative run carried out in the
presence of Degussa P25 by using the APR reactor. The
3.1. Photocatalytic reactivity in APR and CPR reactors
Table 2 reports the results of glycerol photocatalytic oxidation
performed by using the APR reactor. It can be noticed that glycerol
reacted in homogeneous conditions under irradiation even though
with a very low reaction rate. The commercial samples showed the
best performances both for reaction rate and products selectivity.
The experiments carried out in order to study the influence of
photocatalyst amount both on reaction rate and selectivity
indicated that they do not appreciably change for Degussa P25
by increasing the amount from 0.1 to 0.2 g/L, whereas reaction rate
enormously decreased (see t35% column in Table 2) and selectivity
increased in the presence of 0.4 g/L. As far as Sigma–Aldrich (SA) is

V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
25
Table 3
Results of photocatalytic oxidation of glycerol obtained by using CPR reactor. Used catalyst, catalyst amount, glycerol initial concentration, reaction time required for 35%
conversion (t35%) and selectivities (S) towards the known products.
Catalyst
Catalyst amount [g/L]
Glycerol [mM]
t35% [h]
SGAD [%]
SDHA [%]
SFA [%]
SCO [%]
2
Homogeneous
Degussa P25
Degussa P25
Degussa P25
Degussa P25
Degussa P25
Degussa P25b
Degussa P25
Degussa P25
Sigma–Aldrich
HPR
–
100
100
100
100
100
100
100
185
55
12.8a
12
13a
13
8
7.2a
7.5
6.2
7
7a
8
0.023
0.046
0.091
0.182
0.364
0.091
0.091
0.091
0.091
0.091
0.091
20.3
29.1
26.6
31.2
37.7
11.7
14.3
20.6
7.1
12
6.3
8.8
6.8
4
10
10
8
13
7
13
7
7
14.5
12
13
11
12
9.5
6.5
6.9
8
8
6.3
7.5
4.5
4.6
5.4
14
6.5
8
5.2
7.7
11
100
100
100
5.5
6.9
29.4
18.3
HPA
18
a
Data obtained for 5% conversion.
Data obtained by using an incident photon flux of 0.6 mW/cm2.
b
It is worth noting that Degussa P25 showed to be more selective
towards GAD, DHA and FA formation, in contrast with other
photoreactions involving aromatic alcohols yielding higher selec-
tivities in the presence of less crystalline home-prepared TiO2
samples [22–29]. In the glycerol structure, indeed, differently from
aromatic alcohols, three hydroxyl groups are present and
consequently adsorption and desorption of reagents and products,
respectively, occur in different extents and with different strengths
[35,36].
Moreover it is very likely that Product A with m/z 176 derives
from two molecules of glyceric acid adsorbed onto the surface. This
hypothesis is in accord also with a preliminary investigation by
Amadelli [37] which found that glyceric acid is more strongly
adsorbed onto TiO2 surface than both glycerol and GAD. These
findings indicate that the selectivity depends not only on
photocatalyst samples, but also on substrate specificity and
oxidation products.
The results showed in Fig. 3 indicate that the products
concentration increased up to 15–20 h of reaction, then started to
decrease demonstrating a subsequent over-oxidation, whereas the
data obtained by using CPR reactor (Fig. 4) show a continuous
increase of the products concentration that reached a plateau for
times higher than 10 h.
Fig. 3. Concentration of 1,3-dihydroxyacetone, glyceraldehyde, formic acid and
carbon dioxide and conversion for a run carried out with the APR reactor. Glycerol
initial concentration, 100 mM. Photocatalyst: Degussa P25 (0.1 g/L). FA and CO2
concentrations were divided by three for comparison aim.
concentrations of GAD, DHA, CO2 (divided by three for the sake of
comparison) and FA (divided by three for the sake of comparison)
are represented along with the conversion values in the course of
irradiation. Fig. 4 shows analogous results obtained with the CPR
reactor.
In both cases the evolution of the oxidation products occurred
simultaneously, indicating that parallel reactions are responsible
for their production.
From the analyses of the results, it may be hypothesized that
the catalyst surface is not specialized to perform only one reaction
but, on the contrary, that it is a highly heterogeneous one with the
presence of different active sites that generate different products.
3.2. ESI-MS results
ESI-MS analyses were performed on samples withdrawn from
both systems and showed the same results.
The full scan spectrum of the reaction mixture showed the
presence of several peaks, mostly assigned by comparison with
analytical standards. Along with glycerol, the main component of
the blend, DHA and GAD were identified as peaks at m/z 145.
Sequential enrichment of the solution with DHA and GAD allowed
us to unambiguously assign the peak at m/z 145 as [DHA +
MeOH + Na]+ and [GAD + MeOH + Na]+ adducts.
The mass spectrum of the reaction mixture showed also two
unknown products at m/z 177 and m/z 269. The values of 176 and
268 correspond to the molecular weight of the unknown
compounds A and B. The ESI ionization source converts these
neutral molecules into positively charged ions detected at m/z 177
and m/z 269, respectively, by attachment of H+ during the
electronebulization process. This is the reason why the peaks
related to unknown A and B are reported as m/z 177 and m/z 269
rather than 176 and 268. The structure of the corresponding ions
Fig. 4. Concentration of 1,3-dihydroxyacetone, glyceraldehyde, formic acid and
carbon dioxide and conversion for a run carried out with the CPR reactor. Glycerol
initial concentration, 100 mM. Photocatalyst: Degussa P25 (0.1 g/L). FA and CO2
concentrations were divided by three for comparison aim.

26
V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
Table 4
of glycerol itself. This working hypothesis is also substantiated by
Summary of MS2 experiments.
MS3 experiments carried out on the fragment at m/z 223. The
results are displayed in Fig. 5. Isolation and fragmentation of m/z
223 leads, as the only collision induced fragment, to the ion at m/z
131 via 92 Da neutral loss.
Precursor ion m/z
Fragment ion m/z and relative intensity
269
223 (47)
195 (4)
177 (100)
131 (3)
Combining these two experimental findings, we can confidently
consider the carboxy functional group and the entire molecule of
glycerol as building blocks of the unknown product whose
protonated form gives rise to the peak at m/z 269. Additionally,
the neutral loss of 28 Da observed on passing from m/z 223 to m/z
195 in the MS2 experiment suggests the presence of a keto
functional group in the precursor ion. At this stage a reasonable
structure for m/z 269 and a consistent fragmentation pathway can
be postulated. The scheme is sketched in Fig. 6. In this figure the
large fragment is reported as protonated adduct of a compound of
molecular weight 268 Da. It is therefore reported as m/z 269. As we
are not able to locate the protonation site (e.g., H+ may be attached
to one of the two carbonyl groups or to any OH group) H+ is
generically indicated at the top of the molecular formula
corresponding to the putative structure of the fragment at m/z 269.
The second unknown product detected in the mixture as the
peak at m/z 177 is structurally related to the product discussed
above. Indeed, both unknown products share the fragment at m/z
177
149 (17)
131 (100)
was investigated by MS2 and MS3 experiments. In both cases the
collision potential was calibrated in such a way to retain 10% of the
isolated precursor ion intensity. The results of the MS2 experi-
ments are summarized in Table 4.
These fragmentations deserve further comments. The fragment
ion at m/z 223 comes from the precursor by neutral loss of 46 Da,
consistent with formic acid release. This fact is consistent with the
presence of a carboxy functional group in the structure of the
precursor. The base peak at m/z 177 can be seen as deriving from
the parent ion by neutral loss of the elements of glycerol (92 Da),
suggesting that the structure of the precursor ion could be thought
as the addition product of glycerol to a possible oxidation product
Fig. 5. MS3 experiment corresponding to 269 ! 223 ! fragmentation.
Fig. 6. Sketch of putative structures and fragmentation pathways from MS2 experiments on m/z 269 and m/z 177.

V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
27
Scheme 1.
Scheme 2.
131 and, as a matter of fact, the second product is also a fragment of
4. Conclusions
the first one. MS3 experiment on m/z 131 gave, as only fragment, m/
z 113, consistent with the presence of a primary OH functionality in
the precursor, as sketched in Fig. 6.
The partial photocatalytic oxidation of glycerol was performed
by using two different systems in the presence of commercial and
home-prepared TiO2 samples in the anatase, rutile or anatase–
rutile polymorphic phases. Contrarily to the results present in the
literature for the photo-oxidation of some aromatic alcohols to the
corresponding aldehydes, generally the most crystalline TiO2
commercial samples showed the best performances both for
reaction rates and products selectivities. Probably, the presence of
three OH groups in glycerol and the different experimental
conditions used determined this different behaviour. Moreover
the high heterogeneity of surface and the likely presence of
different active sites resulted in the formation of different products
with relatively low selectivities. Products with molecular weights
higher than that of glycerol were found as peaks at m/z 176 and m/z
268 of ESI-MS spectra.
According to HPLC and ESI-MS analyses, Schemes 1 and 2 for
products formation can be presented:
Scheme 1 shows the hypothesised reaction pathways in which
GAD, DHA, FA and CO2 are contemporaneously produced, whereas
Scheme 2 indicates a likely pathway for the formation of Product A
(m/z 176 compound) from glyceric acid molecules strongly
adsorbed on the photocatalytic surface. This compound can
subsequently react with a glycerol molecule giving rise to Product
B (m/z 268 compound).
The absence of a second keto-group in the molecule m/z 176
could be due to its desorption before another subsequent surface
oxidation. In fact the interaction of a CO group with the surface
hydroxyls is weaker than that of OH alcoholic group.

28
V. Augugliaro et al. / Catalysis Today 151 (2010) 21–28
[22] V. Loddo, S. Yurdakal, G. Palmisano, G.E. Imoberdorf, H.A. Irazoqui, O.M. Alfano, V.
References
Augugliaro, H. Berber, L. Palmisano, Selective photocatalytic oxidation of 4-
methoxybenzyl alcohol to p-anisaldehyde in organic-free water in a continuous
annular fixed bed reactor, Ind. J. Chem. React. Eng. 5 (2007) A57.
[1] M. Kidwai, R. Mohan, Green chemistry: an innovative technology, Found. Chem.
7 (2005) 269–287.
[23] G. Palmisano, S. Yurdakal, V. Augugliaro, V. Loddo, L. Palmisano, Photocatalytic
selective oxidation of 4-methoxybenzyl alcohol to aldehyde in aqueous suspen-
sion of home-prepared TiO2 catalyst, Adv. Synth. Catal. 349 (2007) 964–970.
[24] S. Yurdakal, G. Palmisano, V. Loddo, V. Augugliaro, L. Palmisano, Nanostructured
rutile TiO2 for selective photocatalytic oxidation of aromatic alcohols to alde-
hydes in water, J. Am. Chem. Soc. 130 (2008) 1568–1569.
[25] V. Augugliaro, T. Caronna, V. Loddo, G. Marci, G. Palmisano, L. Palmisano, S.
Yurdakal, Oxidation of aromatic alcohols in commercial and home-prepared
irradiated rutile TiO2 aqueous suspensions: a selectivity study, Chem. Eur. J.
14 (2008) 4640–4646.
[2] J.C. Warner, A.S. Cannon, K.M. Dye, Green chemistry, Environ. Impact Asses. Rev.
24 (2004) 775–799.
[3] M.M. Kirchhoff, Promoting sustainability through green chemistry, Res. Conserv.
Recycl. 44 (2005) 237–243.
[4] C.Y. Lorber, I. Pauls, J.A. Osborn, Catalyzed oxidation of alcohols by cis-dioxomo-
lybdenum(VI) complexes via oxygen atom transfer from sulfoxide, Bull. Soc.
Chim. Fr. 133 (1996) 755–758.
[5] K.S. Coleman, L.J.L. Bedel, J.A. Osborn, Catalytic oxidation of alcohols to aldehydes
or ketones using osmium–oxo complexes with sulfoxides or N-methylmorpho-
line-N-oxide as the co-oxidant: a comparative study, C. R. Acad. Sci. Ser. IIc: Chem.
3 (2000) 765–769.
´
[26] V. Augugliaro, H. Kisch, V. Loddo, M.J. Lo´pez-Mun˜oz, C. Ma´rquez-Alvarez, G.
Palmisano, L. Palmisano, F. Parrino, S. Yurdakal, Photocatalytic oxidation of
aromatic alcohols to aldehydes in aqueous suspension of home-prepared titani-
um dioxide. 1. Selectivity enhancement by aliphatic alcohols, Appl. Catal. A 349
(2008) 182–188.
[6] Y. Kurusu, An oxidation system for green chemistry: the oxidation of alcohols and
hydrocarbons in the presence of metal–polymer complexes, J. Inorg. Organomet.
Polym. Mater. 10 (2000) 127–144.
[7] G. ten Brink, I.W.C.E. Arends, R.A. Sheldon, Green, catalytic oxidation of alcohols in
water, Science 287 (2000) 1636–1639.
[8] R.A. Sheldon, I.W.C.E. Arends, G. ten Brink, A. Dijksman, Green, catalytic oxidations
of alcohols, Acc. Chem. Res. 35 (2002) 774–781.
[9] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.F. Carley, A.A. Herzing, M.
Watanabe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Solvent-free oxidation of
primary alcohols to aldehydes using Au–Pd/TiO2 catalyst, Science 311 (2006)
362–365.
[10] C. Li, L. Chen, Organic chemistry in water, Chem. Soc. Rev. 35 (2006) 68–82.
[11] K. Ohkubo, K. Suga, S. Fukuzumi, Solvent-free selective photocatalytic oxidation
of benzyl alcohol to benzaldehyde by molecular oxygen using 9-phenyl-10-
methylacridinium, Chem. Commun. 19 (2006) 2018–2020.
[12] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Improved utilisation of
renewable resources: new important derivatives of glycerol, Green Chem. 10
(2008) 13–30.
[13] R. Ciriminna, G. Palmisano, C. Della Pina, M. Rossi, M. Pagliaro, One-pot electro-
catalytic oxidation of glycerol to DHA, Tetrahedron Lett. 47 (2006) 6993–6995.
[14] C.L. Bianchi, P. Canton, N. Dimitratos, F. Porta, L. Prati, Selective oxidation of
glycerol with oxygen using mono and bimetallic catalysts based on Au, Pd and Pt
metals, Catal. Today 102–103 (2005) 203–212.
[15] S. Demirel-Gu¨len, M. Lucas, P. Claus, Liquid phase oxidation of glycerol over
carbon supported gold catalysts, Catal. Today 102–103 (2005) 166–172.
[16] P. Gallezot, Selective oxidation with air on metal catalysts, Catal. Today 37 (1997)
405–418.
[17] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, G.J. Hutchings, Selective oxidation
of glycerol to glyceric acid using a gold catalyst in aqueous sodium hydroxide,
Chem. Commun. (2002) 696–697.
[18] S. Demirel, K. Lehnert, M. Lucas, P. Claus, Use of renewables for the production of
chemicals: glycerol oxidation over carbon supported gold catalysts, Appl. Catal. B:
Environ. 70 (2007) 637–643.
[19] R. Garcia, M. Besson, P. Gallezot, Chemoselective catalytic oxidation of glycerol
with air on platinum metals, Appl. Catal. A: Gen. 127 (1995) 165–176.
[20] N. Dimitratos, J.A. Lopez-Sanchez, D. Lennon, F. Porta, L. Prati, Effect of particle size
on monometallic and bimetallic (Au,Pd)/C on the liquid phase oxidation of
glycerol, Catal. Lett. 108 (2006) 147–153.
[27] M. Addamo, V. Augugliaro, M. Bellardita, A. Di Paola, V. Loddo, G. Palmisano, L.
Palmisano, S. Yurdakal, Environmentally friendly photocatalytic oxidation of
aromatic alcohol to aldehyde in aqueous suspension of brookite TiO2, Catal. Lett.
126 (2008) 58–62.
[28] S. Yurdakal, G. Palmisano, V. Loddo, O. Alago¨z, V. Augugliaro, L. Palmisano,
Selective photocatalytic oxidation of 4-substituted aromatic alcohols in water
with rutile TiO2 prepared at room temperature, Green Chem. 11 (2009) 510–
516.
´
[29] V. Augugliaro, V. Loddo, M. Lo´pez-Mun˜oz, C. Ma´rquez-Alvarez, G. Palmisano, L.
Palmisano, S. Yurdakal, Home-prepared anatase, rutile, and brookite TiO2 for
selective photocatalytic oxidation of 4-methoxybenzyl alcohol in water: reactiv-
ity and ATR–FTIR study, Photochem. Photobiol. Sci. 8 (2009) 663–669.
[30] N.N. Lichtin, J. Dong, K.M. Vijayakumar, Photopromoted TiO2-catalyzed oxidative
decomposition of organic pollutants in water and in the vapor phase, Water
Pollut. Res. J. Can. 27 (1992) 203–210.
[31] V. Maurino, A. Bedini, M. Minella, F. Rubertelli, E. Pelizzetti, C. Minero, Glycerol
transformation through photocatalysis: a possible route to value-added chemi-
cals, J. Adv. Ox. Technol. 11 (2008) 184–192.
[32] T. Ohno, K. Sarukawa, K. Tokieda, M. Matsumura, Morphology of a TiO2 photo-
catalyst (Degussa, P-25) consisting of anatase and rutile crystalline phases, J.
Catal. 203 (2001) 82–86.
[33] P. Scherrer, Estimation of the size and internal structure of colloidal particles by
means of ro¨entgen, Nachr. Ges. Wiss. Go¨ttingen, Math.–Phys. Kl. 2 (1918) 96–100.
[34] A.W. Burton, K. Ong, T. Rea, I.Y. Chan, On the estimation of average crystallite size
of zeolites from the Scherrer equation: a critical evaluation of its application to
zeolites with one-dimensional pore systems, Micropor. Mesopor. Mater. 117
(2009) 75–90.
´
[35] V. Augugliaro, H. Kisch, V. Loddo, M.J. Lo´pez-Mun˜oz, C. Ma´rquez-Alvarez, G.
Palmisano, L. Palmisano, F. Parrino, S. Yurdakal, Photocatalytic oxidation of
aromatic alcohols to aldehydes in aqueous suspension of home prepared titanium
dioxide. 2. Intrinsic and surface features of catalysts, Appl. Catal. A: Gen. 349
(2008) 189–197.
[36] A. Da Pozzo, M. Mascia, S. Palmas, A.M. Polcaro, J. Rodriguez Ruiz, A. Vacca,
Innovative processes for oxidation of glycerol aqueous solutions, Electrochem 09,
Manchester, UK, September 2009.
[21] P. Fordham, M. Besson, P. Gallezot, Selective catalytic oxidation of glyceric acid to
tartronic and hydroxypyruvic acids, Appl. Catal. A: Gen. 133 (1995) L179–L184.
[37] R. Amadelli, Private communication.