Selective oxidation of alcohols in gas phase using light-activated titanium dioxide

Article abstract of DOI:10.1016/S0021-9517(02)93771-1

Selective oxidation of various primary and secondary alcohols (e.g., 1-, 2-, and 3-pentanol, cyclohexanol, benzyl alcohol, 1- and 2-phenyl ethanol) was studied in a gas-phase photochemical reactor using immobilized TiO₂ catalyst. An annular photoreactor was used at 463 K with an average contact time of 32 sec. Catalyst deactivation studies were performed by FTIR spectroscopic analysis of benzaldehyde, benzoic acid, acetophenone, and the spent catalyst to identify the potential intermediates formed on the TiO₂ surface during the photocatalytic oxidation of alcohols. The presence of a benzene ring in the alcohol generally increased the conversion, and the higher the chain length, the greater the conversion. 1-Phenyl ethanol yielded styrene almost selectively instead of its primary oxidation product, acetophenone, while 2-phenyl ethanol yielded benzaldehyde as the main product instead of the expected primary oxidation product, phenyl acetaldehyde. In the absence of oxygen, the main products were toluene and ethyl benzene, and no appreciable aldehyde formation was observed. When air was used instead of nitrogen (O₂/alcohol = 11), the main products were aldehydes, and there was hardly any hydrocarbon formation.

Full text of DOI:10.1016/S0021-9517(02)93771-1

Journal of Catalysis 211, 434–444 (2002)
doi:10.1006/jcat.2002.3771
Selective Oxidation of Alcohols in Gas Phase Using
Light-Activated Titanium Dioxide
Unnikrishnan R. Pillai and Endalkachew Sahle–Demessie1
National Risk Management Research Laboratory, Sustainable Technology Division, MS-443, U.S. Environmental
Protection Agency, 26 W, Martin Luther King Drive, Cincinnati, Ohio 45268
Received March 29, 2002; revised July 30, 2002; accepted August 6, 2002
manganese complexes, which produce a lot of heavy metal
waste (1–3). In addition, these reactions are often carried
out in environmentally unfriendly organic solvents. Hence
replacing them with heterogeneous catalytic oxidation us-
ing clean and atom-efficient oxidants such as molecular O2
and H2O2 is a definite need, as well as an important goal of
the “green chemistry” concept.
Selective oxidation of various primary and secondary alcohols
was studied in a gas-phase photochemical reactor using immobi-
lized TiO2 catalyst. An annular photoreactor was used at 463 K
with an average contact time of 32 s. The system was found to
be specifically suited for the selective oxidation of primary and
secondary aliphatic alcohols to their corresponding carbonyl com-
pounds. Benzylic alcohols gave higher conversions, however, with
more secondary reaction products. The reaction mechanism for
various products formed is explained. The effects of different re-
action parameters, such as O2/alcohol ratio, water vapor, UV light,
and contact time, were studied. The presence of oxygen was found
to be critical for the photooxidation. Water vapor in the feed was
also found to be helpful in the reaction, although it was not as
critical as in hydrocarbon oxidation, where it was necessary for
hydroxylating the catalyst surface and sustaining its activity. In
alcohol oxidation, surface hydroxylation could be partially pro-
vided by the hydroxyl groups of the alcohol itself. Catalyst deac-
tivation was also observed and is attributed to the surface accu-
mulation of reaction products. However, the catalyst regained its
original activity after regeneration by calcination in air for 3 h at
c
Despite the increasing demand for more-efficient cata-
lytic processes, few efficient catalytic oxidations have been
reported for the oxidation of alcohols to carbonyl com-
pounds. Most of these reports use the Mukaiyama route,
where O2 is used in the presence of at least a stoichiomet-
ric amount of a reactive aldehyde, which would form the
peracid as the actual oxidizing agent (5–7). However, there
have also been a few interesting reports which employ aero-
bic oxidation of alcohols that use copper (8–11), palladium
(12, 13), and ruthenium compounds (14–19). Some of these
methods are limited to benzylic alcohols and require two
equivalents of the catalyst per equivalent of the alcohol
(8–10). Marko et al. used a copper complex catalyst to ox-
idize a variety of alcohols, but the method works best in
toluene and in the presence of a base and additives such as
di(t-butyl azodihydrazine) (11). One major drawback with
the majority of the systems reported above is the use of
environmentally harmful organic solvents such as toluene
(12, 14–16), ethylene carbonate (13), trifluoro toluene (17),
dichloro methane (18), or chlorobenzene (19), which makes
the recovery of products and catalyst difficult. Aerobic ox-
idation of alcohols using Pd and Pt/C has also occurred but
it is limited to water-soluble alcohols (20). Sheldon and co-
workers recently reported aerobic oxidation of a variety
of alcohols in aqueous phase using a water-soluble palla-
dium(II) bathophenanthroline complex catalyst with high
conversion and selectivity (21). Even though the method
seems to be attractive, it involves long reaction times (5–
15 h) under pressure (3 MPa) as well as expensive and com-
plex catalyst preparation. In addition, most of these meth-
ods are carried out at very low concentration levels, which
may involve challenging product and catalyst separation
steps. Another study reports the oxidation of primary al-
cohols and α,β-unsaturated alcohols using heterogeneous
723 K.
ꢀ 2002 Elsevier Science (USA)
Key Words: photocatalytic oxidation; immobilized TiO2; annular
reactor; UV light; alcohol oxidation; carbonyl compounds; catalyst
deactivation.
INTRODUCTION
Selective catalytic oxidation of alcohols to carbonyls is
one of the most important chemical transformations in in-
dustrial chemistry. Carbonyl compounds such as ketones
and aldehydes are the precursors for many drugs, vitamins,
and fragrances and are also important intermediates for
many complex syntheses (1, 2). Most of these reactions,
however, use toxic, corrosive, and expensive oxidants, strin-
gent conditions such as high pressure or temperature, and
strong mineral acids (3, 4). For example, alcohol oxidation
is traditionally carried out in liquid phase by stoichiomet-
ric oxidants such as toxic and expensive chromium(VI) and
1 To whom correspondence should be addressed. Fax: 513-569-7677.
E-mail: Sahle-Demessie.Endalkachew@epa.gov.
434
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

SELECTIVE OXIDATION OF ALCOHOLS IN GAS PHASE
435
FeZSM-5 catalysts and H2O2 with high yield and selectivity; lysts for the oxidation of lower alcohols, such as methanol
however, it involves a high reaction time (20 h) and the use and ethanol (44–46). Chen et al. have studied the photo-
of organic solvents such as methanol (22). A very recent catalytic oxidation of isopropanol using amorphous man-
study by Jensen et al. reports a method for enantioselec- ganese oxide as the photocatalyst (47). This paper presents
tive aerobic oxidation of alcohols using a (-)-spartein/Pd(II) a systematic and rigorous study of selective photooxidation
complex catalyst (23). The method is good for benzylic alco- of higher alcohols on the TiO2 surface.
hols and its application to other substrates is being investi-
To achieve the desired oxygenates, air/O2 and the alco-
gated. This method also requires the presence of an organic hol are reacted at atmospheric pressure and at tempera-
solvent and a base. A review of the recent developments in tures in the range 423–463 K in a photochemical reactor
catalytic alcohol oxidation using O2 and H2O2 as primary that uses ultraviolet light and immobilized TiO2 on silica
oxidants has appeared recently (24). In industrial chem- cloth. This work also incorporates the concept of green
istry, heterogeneous catalyst systems have some advantages chemistry and engineering by allowing the production of
over homogeneous systems, such as easy recyclability and oxygenates in a selective manner and producing fewer by-
separability. However, most of the heterogeneous catalyst products and pollutants than the conventional techniques
systems lack good conversion and selectivity. Therefore, an that use stoichiometric oxidants. TiO2 is used in the form
important challenge is to develop active and selective het- of a deposited film, since the film photocatalyst system has
erogeneous catalyst systems for this type of reaction.
the following practical benefits not attainable in the pow-
The objective of this study was to conduct selective der system: (i) the catalyst filtration step after the reaction
photooxidation of alcohols on a TiO2 surface. TiO2 is bi- is not necessary, and (ii) measurement of the photocatalytic
ologically and chemically inert and stable with respect to effect can be repeated. Therefore, it is of high technologi-
photocorrosion and chemical corrosion and is relatively in- cal importance that a TiO2 film with high photoactivity be
expensive. Titanium dioxide particles absorb light in the prepared.
wavelength region of 315–380 nm and electron–hole pairs
EXPERIMENTAL
are generated at the particle/solution interfaces (25). The
photogenerated hole is believed to have an oxidation po-
tential of ca. 3.0 V and has, therefore, considerable ox-
idizing capability. In order to intercept the energetically
favorable and rapid process of electron–hole recombina-
tion, the species to be oxidized is believed to adsorb on
the surface of the particles (26). Therefore, photochemi-
cal reactions at particle/solution interfaces are controlled
by both relative redox energies and adsorption characteris-
tics. The anatase form of TiO2 has been the most extensively
employed in photocatalytic reactions because of its high ac-
tivity and chemical stability (27–31). The feasibility of this
technology on a commercial scale has also been demon-
strated by the implementation of numerous small-scale ap-
plications, including treatment of air and water streams.
Photocatalytic reactions are applicable to a wide range of
valuable industrial processes, including organic synthesis,
photodestruction of toxic compounds, and purification of
drinking water (32, 33). Over the past two decades, there
has been considerable work aimed at utilizing semiconduc-
tors as photocatalysts (34–43). The photocatalytic oxidation
of many organic molecules, including saturated hydrocar-
bons, by optically excited semiconductor oxides is thermo-
dynamically allowed in the presence of oxygen at room
temperature. Selectivities different from those obtained by
other oxidation means have been reported (34, 36), show-
ing the potential of the method for syntheses where the
expected product is obtained with an acceptable quantum
Reagents
All the alcohol substrates as well as titanium isopropox-
ide were obtained from Aldrich Chemical Company and
used as received without any further purification.
Catalyst Preparation by Dip-Coating Method
TiO2-coated pads were prepared by the dip-coating
method, where preshrunk silica quartz fiber mats (Ther-
mal Material Systems, Brentwood, CA) of dimension 7.6 ×
35 cm2 were dipped in a saturated solution of neat
titanium tetraisopropoxide followed by overnight dry-
ing at room temperature (48). During this drying step,
the isopropoxide was hydrolyzed by reaction with mois-
ture in the air to form amorphous TiO2 on the sur-
face. The dry coated fabrics were then calcined in air at
773 K for 3 h, during which time the amorphous titania
was converted into the catalytically more active anatase
phase. The primary particle size of the TiO2 film was
determined to be about 140 nm (48). These pads were
made in the form of hollow cylindrical shapes by wrap-
ping them around a suitable stainless-steel wire mesh.
Reactor Setup
The geometry of the photoreactor has a major effect on
yield. However, the use of this system for the selective ox- thisprocess. A30-cm-longstainless-steel(304 Grade)annu-
idation of alcohols has been rather limited. There are few lar reactor of 9.7-cm i.d. with an annular volume of 2500 cm3
reports that use TiO2 and metallized TiO2 as photocata- was employed. The reactor was fitted with a magnetic stirrer

436
PILLAI AND SAHLE–DEMESSIE
at the bottom and a cylindrical water-cooled glass jacket (Omega FMX 8400). Two mass flow controllers (Kobold
containing a 1000-W UV lamp (Jaylight Co., JSZ4378, peak MFC-5108) were used to establish the desired flows of air
radiation at 360 nm) at the middle. The magnetic stirrer was and supplementary oxygen and/or nitrogen. Water vapor
intended to improve the mixing inside the reactor and to was introduced into the flowing air by bubbling it through a
ensure turbulent conditions and uniform contact of the re- waterreservoirkeptatacontrolledtemperature. Thewater-
actants with the catalyst-coated surface. The intensity of laden air was then passed through a shell and tube heat
the UV light on the catalyst-coated surface was measured exchanger kept at 373 K using a high-pressure micrometer-
approximately (20 mW/cm2) using a photometer (Interna- ing plunger pump (Omega PHP-202) followed by mixing
tional Light, Model IL 1400). Three TiO2 cylindrical pads, with the vaporized alcohol in the static mixer, maintained at
prepared as mentioned above, with a total exposed surface 423 K. Excess water was drained from the tubing periodi-
area of 765 cm2, were fixed on the internal surface of the cally to avoid entrapment. The preheated mixture was in-
reactor. Temperatures at the inlet, middle, and bottom of troduced into the bottom of the annular reactor. The flow
the reactor were measured by three iron–constantan ther- system was allowed to reach steady state for 60 min with air
mocouples. The schematic of the reactor setup is shown in flowing and the lamp was turned on before introducing the
Fig. 1.
alcohol feed. An average total gas flow rate of 2 L min−1
was used with the bottom stirrer at 600 rpm, resulting in a
minimal mass transfer resistance, as verified by calculations
similar to those of Peral and Ollis (49). The exit stream from
the reactor was passed through a liquid nitrogen-cooled
trap to condense the oxidized products before being vented
to a fume hood.
The condensed liquid products were collected peri-
odically and analyzed by a Hewlett–Packard 6890 gas
chromatograph using a HP-5 5% phenyl methyl siloxane
Reaction Procedure
Photocatalytic oxidations of alcohols were performed in
gas phase using the annular reactor described above by
flowing a preheated, humidified air–oxygen mixture and the
alcohol vapor. Alcohol was introduced into the reactor sys-
tem via a metering pump at a particular rate wherein it was
heated and vaporized prior to being fed into a static mixer
N Purge
sample
2
Hygrometer
loop
Out
Coolant
In
On line
GC/MS-
T
T
T
P
Humidifier
Flow Totalizer
Flow
meter
Condensers
Air heater
Reactor
P
T
Mixer
MFC
Air
Vent
Mass flow
Controller
Collection
Flasks
Feedstock
N
2
or O
T
2
Heated
Tube
Evaporator
Metering
Pump
Drain
FIG. 1. Schematic diagram of the reactor assembly and experimental setup for the gas-phase photocatalytic oxidation of alcohols.

SELECTIVE OXIDATION OF ALCOHOLS IN GAS PHASE
437
capillary column (30 m × 320 µm × 0.25 µm) and a quadru-
RESULTS
Effect of the Nature of Alcohol
ple mass filter-equipped HP 5973 mass selective detector
under temperature-programmed heating from 313–473 K
at 10◦/min. Samples were analyzed with an injection volume
of 1 µL. Quantification of the oxygenated products was ob-
tainedusingamultipointcalibrationcurveforeachproduct.
The presence of any organic acids in the product mixture
was determined by HPLC (Finniganmat—LCQ) analysis
of the methanol extracts (neutral pH) from the spent cat-
alyst, where the catalyst was scratched off the silica fabric
after the reaction and mixed well with methanol (25 mL)
followed by filtration. The filtrate was concentrated under
vacuum and analyzed.
Oxidation results of different alcohols obtained after a
reaction period of 2 h are shown in Table 1. This amount
of time was chosen to make sure that activity and product
selectivity determinations were made at steady-state con-
ditions. All the alcohols except the two phenyl ethanols
formed the corresponding carbonyl compound selectively.
Generally, the conversions per pass were low for primary
alcohols, with a slightly higher value for secondary alcohols.
The presence of a benzene ring in the alcohol generally
increased the conversion, and the higher the chain length,
the greater the conversion. 1-Phenyl ethanol showed the
maximum conversion, 97%, among all the alcohols. Oxida-
tion of 3-methyl-3-butene-1-ol yielded 3-methyl-2-butenal
and not the expected 3-methyl-3-butene-1-al. This could
be due to a double bond migration to the secondary carbon
position during the oxidation process. 1-Phenyl ethanol and
2-phenyl ethanol, however, formed mainly the secondary
reaction products instead of their primary oxidation
carbonyl product. For example, 1-phenyl ethanol gave
styrene almost selectively instead of its primary oxidation
product, acetophenone. Similarly, 2-phenyl ethanol yielded
benzaldehyde as the main product instead of the expected
primary oxidation product, phenyl acetaldehyde. Turnover
frequencies (TOF) ranging from 7 to 32 × 10−3 s−1 were
estimated for the alcohols in Table 1 using a surface
coverage of 330 micromoles per gram of TiO2 (45).
The alcohol substrates studied include primary, sec-
ondary, and cyclic aliphatic and benzylic alcohols. The
effects on conversion and product selectivity of different
variables, such as O2/alcohol molar ratio, contact time,
water vapor, UV light, and time-on-stream, were in-
vestigated. Conversion and selectivity are defined as
follows:
Moles of reactant consumed
Conversion (%) =
× 100; [1]
Initial moles of reactant
Selectivity of product P (%)
Percentage formation of product P
=
× 100.
[2]
Percentage of total conversion
TGA Analysis
TABLE 1
Effect of the Nature of the Alcohol on Conversion and Selectivitya
Catalyst deactivation studies were carried out using ther-
mogravimetric analysis of the fresh TiO2 and the catalyst
after different reaction periods, using a Perkin–Elmer ther-
mogravimetric analyzer (Model TGA-7). Approximately
10 mg of the sample was heated in an atmosphere of air
in the temperature range 323–1023 K at a heating rate of
10◦/min.
Conversionb
(%)
Selectivity
(%)
Entry Alcohol
Product
1
2
18
20
>95
>95
3
4
24
26
>95
>95
5
6
7
37
14
21
>95
IR Spectroscopy
>95
Catalyst deactivation studies were also conducted by
FTIR spectroscopic analysis of benzaldehyde, benzoic acid,
acetophenone, and the spent catalyst to identify the po-
tential intermediates formed on the TiO2 surface during
the photocatalytic oxidation of alcohols using a Perkin–
Elmer FTIR spectrometer (Model Spectrum 2000). A dis-
posable IR card (3M, Type 61, polyethylene) adsorbed with
a methanolic solution of benzaldehyde, benzoic acid, ace-
tophenone, and the methanol extracts from the spent cata-
lysts was used for the analysis and further comparison.
The spectral region from 1100 to 2000 cm−1 was analyzed
for identification of adsorbed aromatic compounds such as
aldehydes and carboxylic acids (50).
>95
8
9
35
97
>95
7
(83% styrene)
26
10
53
(benzaldehyde = 48,
acetophenone = 10)
a Alcohol, 1.43 mmol min−1; O2/alcohol, 22; temperature, 463 K.
b After a 2-h reaction period.

438
PILLAI AND SAHLE–DEMESSIE
TABLE 2
Effect of O2/Alcohol Molar Ratio on 2-Phenyl Ethanol Photocatalytic Oxidationa
Selectivity (%)
O2/substrate Conv.
Phenyl
Ethyl
ratio
(%)
Benzaldehyde acetaldehyde Acetophenone Styrene Ester Toluene benzene Propionaldehyde
0 (only N2)
13
20
36
53
49
51
53
—
36
44
48
49
51
43
13
37
29
26
18
18
19
—
4
7
10
9
8
—
2
—
2
3
2
—
4
10
6
9
11
16
40
15
6
7
8
22
2
—
—
—
1
26
—
—
1
4
1
1
11
22
34
45
56
6
14
5
—
2
—
a Alcohol, 1.43 mmol min−1; temperature, 463 K.
Since all the alcohols except the phenyl ethanols formed good conversion and aldehyde formation. A large excess
only the corresponding carbonyl product selectively, 2- of oxygen (O2/alcohol ≥ 22) does not have any significant
phenyl ethanol was selected for investigating the effects effect on the formation of products. No overoxidation prod-
of various reaction parameters, as described below.
ucts, such as acid formation, were detected in any case. This
was confirmed by HPLC analysis of the methanol extracts
from the spent catalyst.
Effect of Oxygen-to-Alcohol Molar Ratio
The effect of the oxygen-to-alcohol molar ratio on con-
version and product selectivity for the gas-phase photocata-
lytic oxidation of 2-phenyl ethanol is shown in Table 2.
Conversion increased from 13 to 20% when a very small
quantity of oxygen (O2/alcohol = 1) was added to the nitro-
gen carrier gas. There was significant improvement in the
conversion, from 13 to 36%, when nitrogen was replaced
by air as the carrier gas (O2/alcohol = 11). As the oxygen
content increased further, conversion also increased, reach-
ing a plateau at an oxygen-to-alcohol molar ratio of 22.
Table 2 also shows the effect of the oxygen-to-alcohol mo-
lar ratio on product selectivity. In the absence of oxygen, the
main products were toluene and ethyl benzene, and no ap-
preciable aldehyde formation was observed. When air was
used instead of nitrogen (O2/alcohol = 11), the main prod-
ucts were aldehydes, and there was hardly any hydrocar-
bon formation. However, there was no significant change
in the product distribution with a further increase in the
concentration of oxygen in the feed. These results show
that the presence of oxygen is very critical for obtaining a
Effect of Gas Flow Rate (Contact Time)
Effect of contact time of the reactant was studied by vary-
ing the gas flow at a constant oxygen-to-reactant molar ra-
tio of 22 (Table 3). Conversion increased from 17 to 60%
as the total gas flow decreased from 4 to 0.8 L min−1 (con-
tact time increased from 16 to 80 s). Benzaldehyde selec-
tivity increased from 30 to 69%, whereas selectivity for
phenyl acetaldehyde decreased from 40 to 20% in the same
flow-range variation. Acetophenone was not formed sig-
nificantly under these conditions and hence its selectivity
variation was not important.
Effect of UV Light
The significant role of UV light on the reaction is shown
in Fig. 2. When the reactor was heated to 423 K by an ex-
ternal heating source, only 8% conversion was obtained.
Irradiation of the catalyst with the UV light increased the
conversion dramatically, to 48%. However, the selectivity
of phenyl acetaldehyde decreased from 80 to 22% when
TABLE 3
Effect of Total Gas Flow (Contact Time) on 2-Phenyl Ethanol Photocatalytic Oxidationa
Selectivity (%)
Total gas flow
Contact
time (s)
Conversion
(%)
Phenyl
acetaldehyde
(L min−1
)
Benzaldehyde
Acetophenone
4
2
1.2
0.8
16
32
53
80
17
53
57
60
30
48
62
69
40
26
23
22
—
10
12
8
a Alcohol, 1.43 mmol min−1; temperature, 463 K; O2/alcohol molar ratio, 22.

SELECTIVE OXIDATION OF ALCOHOLS IN GAS PHASE
439
ucts, such as ethyl benzene and esters, were also formed.
Humidity is known to play a key role in hydrocarbon ox-
idations through the formation of hydroxyl groups on the
catalyst surface (51). However, it does not seem to have any
dramatic effect on alcohol oxidation.
100
80
60
40
20
0
Benzaldehyde
Phenyl acetaldehyde
Acetophenone
Other Products
Conversion
Catalyst Deactivation: Effect of Time-on-Stream
Effect of time-on-stream on alcohol conversion is shown
in Table 4. Generally, conversion decreases with an increase
in time-on-stream for all alcohols studied. This is attributed
to the deactivation of the catalyst, which is very common
in this type of reaction (52–54). However, the magnitude
of the decrease was related to the conversion value and the
nature of the alcohol or the products formed. The decrease
was less pronounced in the case of alcohols that show low
conversions. The higher the conversion, the greater the ex-
tent of deactivation, except in the case of 1-phenyl ethanol,
which did not show any significant deactivation during the
period studied. For example, there was about a 10–30%
reduction in conversion for C5 alcohols, whereas C6 al-
cohols such as cyclohexanol, benzyl alcohol, and 2-phenyl
ethanol showed around a 65% reduction in conversion as
the reaction period increased from 1 to 4 h. The selectiv-
ity for the corresponding carbonyl compound was almost
100% for all the aliphatic alcohols irrespective of the ex-
tent of conversion. However, the product distribution var-
ied with reaction time for 1-phenyl and 2-phenyl ethanol
oxidations (Figs. 4a and 4b). It can be seen that the pro-
duct composition consisted of a significant amount of ace-
tophenone in both cases during the initial period (1 h).
As the reaction period increased, acetophenone selectiv-
ity decreased drastically and formation of other products
ensued. The catalyst after the reaction was extracted with
methanol and analyzed by HPLC for the possible formation
of overoxidation products, such as acids. The analysis
UV-Present
UV-Absent
FIG. 2. Effect of UV light on the activity and selectivity of 2-phenyl
ethanol oxidation (temperature, 423 K; O2/alcohol molar ratio, 22; alcohol,
1.43 mmol min−1).
UV light was used at the same temperature. In the pres-
ence of UV light, benzaldehyde was formed in preference
to phenyl acetaldehyde, in addition to small amounts of
other secondary and tertiary reaction products.
Effect of Water Vapor
The presence of humidity in the feed is also shown to af-
fect the reaction (Fig. 3). When the carrier gas was enriched
with water vapor (relative humidity of 90%), conversion re-
mained more or less the same but there was an increased
yield of aldehydes. In the absence of water vapor in the feed,
appreciable amounts of other secondary and tertiary prod-
TABLE 4
Effect of Time-on-Stream on Conversion and Selectivity
of Various Alcoholsa
Conversion (%) at different reaction times
Alcohol
1-Pentanol
2-Pentanol
3-Pentanol
Cyclopentanol
Hexyl alcohol
Cyclohexanol
3-Methyl-3-butene-1-ol
Benzyl alcohol
1-Phenyl ethanol
2-Phenyl ethanol
1 h
2 h
4 h
26
19
34
48
45
38
25
52
90
64
20
18
24
37
26
21
14
35
97
53
17
17
21
24
21
14
04
18
94
20
FIG. 3. Effectofwatervaporontheactivityandselectivityof2-phenyl
ethanol oxidation (temperature, 423 K; O2/alcohol molar ratio, 22; alcohol,
1.43 mmol min−1).
a Alcohol, 1.43 mmol min−1; temperature, 463 K; O2/alcohol molar
ratio, 22.

440
PILLAI AND SAHLE–DEMESSIE
100
a
b
99
98
97
c
d
e
96
373
473
573
673
773
873
973
Temperature (K)
FIG. 5. Thermograms of the fresh and used catalysts. (a) Fresh cata-
lyst; (b) regenerated catalyst; (c) catalyst after a 1-h reaction period;
(d) catalyst after a 2-h reaction period; and (e) catalyst after a 4-h reaction
period.
An increase in the reaction time to 2 and 4 h increased the
weight loss to 3 and 3.7%, respectively. The onset of weight
loss was around 650 K, followed by a slow and steady de-
crease up to around 773 K. The catalyst after 4 h of reaction,
however, showed a two-step weight loss process, with the
initial weight loss occurring at around 473 K and the sec-
ond one at the same temperature as that of the other two
catalysts. The figure also shows that calcining in air at 773 K
for 3 h could regenerate the catalyst, almost completely, to
its original form.
FIG. 4. (a) Effect of reaction run time on product selectivity for 1-
IR Spectroscopic Analysis
phenyl ethanol photooxidation. (b) Effect of reaction run time on prod-
uct selectivity for 2-phenyl ethanol photooxidation (temperature, 423 K;
O2/alcohol molar ratio, 22; alcohol, 1.43 mmol min−1).
IR spectra from the spent catalyst sample is shown in
Fig. 6 along with the spectra for benzaldehyde, benzoic acid,
and acetophenone for identification of the surface species
present on the catalyst surface which is believed to be re-
sponsible for the deactivation. A comparison of the spectra
suggests that the spent catalyst surface contains aromatic
and aldehydic species. The IR spectrum of the spent catalyst
also indicates the presence of some polymeric carboxylate
compound. No indication of any acid formation could be
observed. This is in agreement with both the HPLC analysis
and the TGA experiments, as discussed above.
showed the presence of the initial reactant and products
such as benzaldehyde and phenyl acetaldehyde as well as
some carboxylate-type compound on the surface. No de-
tectable acids could be identified. However, the methanol
extraction did not remove all the carbon compounds from
the catalyst, as evidenced by the dark color of the catalyst
even after methanol washing. The catalyst surface after the
extended reaction period (4 h) was slightly sticky, with a
yellowish-brown appearance.
TABLE 5
Thermogravimetric Analysis
Thermogravimetric Analysis Data of the Fresh
and Used Catalyst
Thermogravimetric analysis (TGA) of the fresh catalyst
and the catalysts after different reaction periods showed a
broad thermogram with a slow rate of weight loss in the
temperature range 373–973 K (Fig. 5). The values of per-
centage weight loss and temperature of the onset of weight
loss are given in Table 5. Weight loss is a measure of the
amount of product adsorbed on the catalyst surface. The
catalyst after 1 h of reaction showed a gross weight loss
of 2.6%, in comparison to 0.8% loss for the fresh catalyst.
Reaction time
of the catalyst (h)
Onset of weight Weight loss
loss (K)
(%)
0 (fresh)
—
652
649
648
—
0.8
2.6
3.0
3.7
1.0
1
2
4
Regenerated

SELECTIVE OXIDATION OF ALCOHOLS IN GAS PHASE
441
FIG. 6. FTIR spectra of adsorbed methanolic solutions of spent catalyst sample, acetophenone, benzoic acid, and benzaldehyde.
DISCUSSION
the alcohol (56). In addition, the alcohols may undergo de-
hydration on the catalyst surface during a photocatalytic
oxidation reaction (57).
The initial photoxidation step here may be the interac-
Effect of the Nature of Alcohol on Conversion
and Selectivity
Activity for photocatalytic oxidation of alcohols is found tion of a surface hole with the hydroxyl group of the alcohol
to be very much dependent on the nature of the alcohol forming a metal–oxo species with the removal of a proton
used. During photocatalytic oxidation, the initial reaction (Scheme 1). This proton removal step becomes easier with
of the alcohol takes place on the surface of TiO2 where the an increase in carbon chain branching as well as with an
primary hole reaches the surface and interacts with the sur- increase in carbon chain length, because of the increased
facehydroxylgroups, followedbyanelectrontransfertothe availability of adjacent removable protons. The higher the
·
·
≡
hole to form species such as OH and TiO (55). These number of adjacent hydrogen atoms present, the easier the
species react via a mediated pathway. At high alcohol con- removal and the greater the conversion. Temperature also
centration, as is the case here, there could also be a direct plays a role here; therefore, this effect may not be pro-
interaction of the surface hole with the hydroxyl group of nounced at higher temperature. The nature of the alkyl
R
R
R
CH2
CH2
O
R-CH=O
R-CH3
CH2
2-
2-
2-
Ti4+O
CH2
O
O Ti4+O
CH2
O
H
2-
2-
2-
Ti4+O
O Ti4+O
R-CH2-CH=O
2-
2-
2-
Ti4+O
O Ti4+O
R-CH2-CH
3
SCHEME 1. Photocatalytic oxidation of alcohols to aldehyde/ketone over TiO2 (ꢀ = surface hole).

442
PILLAI AND SAHLE–DEMESSIE
group also influences the above step, as is evident from Effect of UV Light
Table 1. Cyclopentanol (entry 5) showed a higher conver-
UV light is required for the activation of the catalyst sur-
sion than its open-chain counterparts, viz. 2- and 3-pentanol
(entries 2 and 3). This could be due to the higher strain
in the 5-carbon cyclic structure. However, cyclohexanol
(entry 7) gave a lower conversion than n-hexanol (entry 4)
due to the increased stability of the six-carbon cyclic struc-
ture (58). It was also observed that the presence of a ben-
zene ring enhanced the conversion (entries 8–10). This can
be attributed to the electron-deficient nature of the ben-
zene ring, which results in a reduced electron density at
the oxygen–hydrogen bond, thereby making the proton
abstraction relatively easy. Formation of styrene from 1-
phenyl ethanol (entry 9) may be due to the photocatalytic-
induced dehydration of the alcohol (57).
Two types of adsorption sites are present on TiO2 dur-
ing photocatalytic oxidation (45): a weakly adsorbed site,
where the carbonyl product formed readily gets desorbed,
and a strongly adsorbed site, where the primary product
undergoes further reactions to form other secondary and
tertiary oxidation products, as may be case for 2-phenyl
ethanol (Scheme 1).
face as well as the molecular oxygen (Fig. 2). Heating the
catalyst to the same temperature by an external source may
not produce the same type of active sites on the surface as is
produced by UV irradiation. In other words, the photocata-
lytic sites could be completely different from the thermal
reaction sites. It may be assumed that the UV irradiation
of the catalyst surface results in photocatalytically active
sites, which are stronger oxidants than the sites produced
by the thermal effect, which causes only mild oxidation.
This mild oxidation affords mainly the primary oxidation
product, whereas strong oxidation affords more secondary
products. This explains why phenyl acetaldehyde is selec-
tively formed in the absence of UV light whereas benzalde-
hyde is the main product in the presence of UV light. Both
the UV and thermal effect may be operating in parallel
during the photocatalytic oxidation (45c), even though the
thermal effect is not the predominant one determining the
overall activity and product selectivity. An increase in light
intensity accelerates the reaction; however, the increase in
photoactivity is not linear with the light intensity (59, 60).
Effect of Water Vapor
Effect of Oxygen-to-Substrate Ratio
Results show that humidity in the feed affects the alcohol
oxidation (Fig. 3). However, the role of humidity in photo-
catalytic oxidation of alcohols is not very critical. Water
vapor is known to contribute to the formation of surface
hydroxyl groups by coordination with surface holes. The
presence of surface hydroxyl groups is necessary for the ox-
idation process. In the absence of water vapor in the feed,
the degree of surface hydroxylation is less. This reduces the
formation of oxidation products (aldehydes) and enhances
dehydration products such as alkanes and alkenes. Expo-
sure of the catalyst to UV light under dry conditions leads
to irreversible consumption of the surface hydroxyl groups,
which in turn results in catalyst deactivation (61). In alcohol
oxidation, this surface hydroxylation can be partly achieved
by the alcohol itself.
The presence of oxygen is necessary for photocatalytic
oxidation to take place (Table 2). The effect of oxygen
concentration, however, diminishes at high oxygen levels
(O2/alcohol ≥ 22). Molecular oxygen is activated on the
catalyst surface in the presence of UV light to form elec-
trophilic species such as O2− and O−, which also helps in ac-
tivating the reactant molecule and eventually in producing
different oxidation products. When no oxygen is available,
the holes formed on the surface on UV irradiation abstract
the oxygen from the alcohol, producing hydrocarbons such
as toluene and ethyl benzene, as illustrated in Scheme 1.
Anaerobic conditions can result in photocatalytic-assisted
hydrolysis reaction (57a). It is also obvious from the scheme
that a large excess of oxygen has no additional benefits.
Boarini et al. have made a similar observation in the pho-
tocatalytic oxidation of cyclohexane (43).
Catalyst Deactivation: Time-on-Stream, TGA,
and IR Studies
Effect of Gas Flow Rate (Contact Time)
The decrease in conversion with reaction pass time is at-
The effect of residence time studied using the photocata- tributed to the deactivation of the catalyst surface, as men-
lytic oxidation of 2-phenyl ethanol shows that an increase tioned earlier. The product selectivity for the two phenyl
in the residence time of the alcohol in the reactor facilitates ethanols also varies with time-on-stream (Fig. 4). Ace-
the formation of secondary products such as benzaldehyde tophenone is observed as the main product during the ini-
at the cost of the primary oxidation product, phenyl ac- tial reaction period, indicating that it is one of the primary
etaldehyde (Table 3). In other words, a high residence time products of oxidation formed from the weakly adsorbed
of the molecule in the reactor means the intermediate prod- sites. With an increase in reaction pass time, more and more
ucts can adsorb and react further down the reactor to form strongly adsorbed species are formed, leading to the for-
more strongly bound products that cause catalyst deactiva- mation of secondary oxidation products, such as aldehydes
tion (45).
(45).

SELECTIVE OXIDATION OF ALCOHOLS IN GAS PHASE
443
TGA analysis of the fresh TiO2 and the TiO2 after dif- electrical energy needed to change the feed substrate con-
ferent reaction periods indicates that catalyst deactivation centration by an order of magnitude, is estimated to be 16
occurs in the first hour of the reaction itself. The initial rate to 18. However, the estimated electrical energy per mass
of the reaction is higher than at steady state for almost all of desired product formed for the photoxidation method is
the alcohols tested. This initial deactivation may be due within the order of magnitude of that of the conventional
to the accumulation of the product on the catalyst surface process.
(45b). Most of the weight loss occurs at around 650 K, indi-
cating the presence of a carbonaceous deposit on its surface.
However, the broad thermogram indicates that it is not just
CONCLUSIONS
simple carbon but some carboxylate compound. The con-
centrations of both the reactant and the carboxylate-type
material increase with an increase in the reaction period.
The yellowish-brown appearance with a slightly sticky na-
ture of the spent catalyst surface indicates the presence of
some polymeric material formed by the polymerization of
the aldehyde products. The longer the reaction period, the
greater its formation. Similar oligomerization was detected
during the gas-phase photocatalytic oxidation of acetalde-
hyde on the TiO2 surface (45d). The appearance of an ad-
ditional peak at 473 K in the thermogram may be due to
this reason. The fact that a 3-h calcination of the catalyst in
air at 773 K regenerates the catalyst almost completely also
suggests that it is mostly carbonaceous material that deacti-
vates the catalyst. HPLC analysis of the methanol extracts
from the catalyst surface after the reaction also confirmed
the presence of some carboxylate-type compounds on the
surface. This is further confirmed by the IR analysis of the
spent catalyst, which also shows the presence of some alde-
hydic and polymeric species.
In summary, UV light-assisted photocatalytic oxidation
is found to be suitable for selectively oxidizing primary
and secondary aliphatic alcohols to their corresponding car-
bonyl products using a 2.5-L annular photoreactor contain-
ing immobilized TiO2 catalyst. Aromatic alcohols, however,
form mainly secondary reaction products. Oxygen avail-
ability is found to be critical for the photooxidation. The
presence of water vapor in the feed is advantageous for
the reaction; however, it is not as significant as in the case
of hydrocarbon oxidation, where its presence is especially
important in sustaining the catalyst life (59). One disadvan-
tage of the system is the catalyst deactivation, which is at-
tributed to the surface accumulation of reaction products.
The catalyst activity, however, can be regained by calcin-
ing the catalyst in air at 723 K for approximately 3 h. The
reaction mechanism is described to explain the behavior
of the catalyst under different conditions. Conversion and
product distribution obtained are explained on the basis of
different oxidation mechanisms. The oxidation of alcohols
using light-activated TiO2 catalyst and molecular oxygen
could be a green alternative, but further investigation into
process development is required.
Catalyst deactivation during photocatalytic reactions
have been observed by many workers (52–54, 62, 63). Many
hypotheses have been used to explain the catalyst deacti-
vation during photooxidation. For example, Falconer and
co-workers identified the accumulation of acetaldehyde on
the catalyst surface for the deactivation of the catalyst
ACKNOWLEDGMENTS
The authors are grateful to the referees for their critical assessment and
during the photocatalytic oxidation of ethanol (45). Peral valuable suggestions for improving the clarity and quality of the paper.
The technical assistance and support in conducting the reaction studies by
Mr. Tom Deinlein is gratefully acknowledged. U.R. Pillai is a postgraduate
research participant at the National Risk Management Research Labora-
tory administered by the Oak Ridge Institute for Science and Education
and Ollis proposed that strongly adsorbed 1-butanoic acid
causes catalyst deactivation during the photocatalytic ox-
idation of 1-butanol (49). Cunningham and Hodnett sug-
gested that the deactivation is caused by the adsorption of
through an interagency agreement between the U.S. Department of En-
CO2 formed during the reaction on the active sites (62). ergy and the U.S. Environmental Protection Agency.
Blake and Griffin observed carboxylate-type species accu-
mulating on the catalyst surface during butanol oxidation
(64). Strongly adsorbed carbonate species are reported to
REFERENCES
cause catalyst deactivation during the photooxidation of
trichloroethylene (65). Ollis and co-workers have also re-
ported that carboxylate and carboxylic acid accumulation
are responsible for deactivation during the photooxida-
tion of toluene and toluene mixed with trichloroethylene,
trichloropropene, or perchloroethylene (66, 67).
The relative photonic efficiencies for the initial oxida-
tion of the organic substrates in the gas-phase reactor using
a TiO2 film are estimated to be in the range 0.8–1.2. The
estimated quantum yields ranged from 0.005 to 0.02 (68).
The “electrical energy per order” (EE/O), defined as the
1. Sheldon, R. A., and Kochi, J. K., “Metal-Catalyzed Oxidation of Or-
ganic Compounds.” Academic Press, New York, 1981.
2. Hudlicky, M., “Oxidations in Organic Chemistry.” Am. Chem. Soc.,
Washington, DC, 1990.
3. Larock, R. C., “ComprehensiveOrganicTransformations.”VCH, New
York, 1989 .
4. Cainelli, G., and Cardillo, G., “Chromium Oxidations in Organic
Chemistry.” Springer, Berlin, 1984.
5. Murahashi, S. I., Naota, T., and Hirari, J., J. Org. Chem. 58, 7328 (1993).
6. Inokuchi, T., Nakagawa, K., and Torii, S., Tetrahedron Letters 36, 3223
(1995).
7. Iwahama, T., Sakaguchi, S., Nishiyama, Y., and Ishii, Y., Tetrahedron
Lett. 36, 6923 (1995).

444
PILLAI AND SAHLE–DEMESSIE
8. Capdevielle, P., Sparfel, D., Baranne-Lafont, J., Cuong, N. K., and
41. Worsley, D., Andrew, M., Keith, S., and Hutchings, M. G., J. Chem.
Soc. Chem. Commun. 1119 (1995).
42. Mao, Y., and Bakac, A., Inorg. Chem. 35, 3925 (1996).
43. Boarini, P., Carassiti, V., Maldotti, A., and Amadelli, A., Langmuir 14,
2080 (1998).
Maumy, M., J. Chem. Res. 10, 1993 (1993).
9. Munakata, M., Nishibayashi, S., and Sakamoto, H., J. Chem. Soc.
Chem. Commun. 219 (1980).
10. Senmelhack, M. F., Schmid, C. R., Cortes, D. A., and Chon, C. S.,
J. Am. Chem. Soc. 106, 3374 (1984).
11. Marko, I. E., Giles, P. R., Tsukazaki, M., Brown, S. M., and Urch, C. J.,
Science 274, 2044 (1996).
12. Nishimura, T., Onoue, T., Ohe, K., and Uemura, S., J. Org. Chem. 64,
6750 (1999).
44. Chen, J., Ollis, D. F., Rulkens, W. H., and Bruning, H., Water Res. 33,
661 (1999).
45. (a) Muggli, D. S., McCue, J. T., and Falconer, J. L., J. Catal. 173, 483
(1998); (b) Muggli, D. S., and Falconer, J. L., J. Catal. 175, 213 (1998);
(c) Falconer, J. L., and Magrini-Bair, K. A., J. Catal. 179, 171 (1998);
(d) Muggli, D. S., Lowery, K. H., and Falconer, J. L., J. Catal. 180, 111
(1998).
13. Blackburn, T. F., and Schwartz, J., J. Chem. Soc. Chem. Commun. 157
(1977).
14. Bilgrien, C., Davis, S., and Drago, R. S., J. Am. Chem. Soc. 109, 3786
(1987).
46. Hussein, F. H., Pattenden, G., Rudham, R., and Russell, J. J., Tetrahe-
dron Lett. 25, 3363 (1984).
15. Tang, R., Diamond, S. E., Neary, N., and Mares, F., J. Chem. Soc. Chem.
Commun. 562 (1978).
47. Chen, J., Lin, J. C., Purohit, V., Cutlip, M. B., and Suib, S. L., Catal.
Today 33, 205 (1997).
16. Marko, I. E., Giles, P. R., Tsukazaki, M., Chelle-Regnaut, I., Urch, C. J.,
and Brown, S. M., J. Am. Chem. Soc. 119, 12661 (1997).
48. Sahle-Demessie, E., Gonzalez, M., Wang, Z. M., and Biswas, P., Ind.
Eng. Chem. Res. 38, 3276 (1999).
17. Hanyu, A., Takeyawa, E., Sakaguchi, S., and Ishii, Y., Tetrahedron Lett. 49. Peral, J., and Ollis, D. F., J. Catal. 136, 554 (1992).
39, 5557 (1998).
50. Socrates, G., “Infrared Characteristic Group Frequencies.” Wiley,
New York, 1980.
18. Lenz, R., and Ley, S. V., J. Chem. Soc. Perkin Trans. 1, 3291 (1997).
19. Dijkman, A., Arends, I. W. C. E., and Sheldon, R. A., Chem. Commun. 51. Lin, M. M., Appl. Catal. A 207, 1 (2001).
1591 (1999).
52. Sauer, M. L., and Ollis, D. F., J. Catal. 163, 215 (1996).
53. Colon, G., Hidalgo, M. C., and Navio, J. A., J. Photochem. Photobiol.
A 138, 79 (2001).
54. Mendez-Roman, R., and Cardone-Martinez, N., Catal. Today 40, 353
(1998).
55. Minero, C., Mariella, G., Maurino, V., Vione, D., and Pelizzetti, E.,
Langmuir 16, 8964 (2000).
56. Brezova, V., Vodny, S., Vesely, M., Ceppan, M., and Lapcile, L.,
J. Photochem. Photobiol. A 56, 125 (1991)
57. (a) Calza, P., Minero, C., and Pelizzetti, E., Environ. Sci. Technol.
31, 2198 (1997); (b) Djeghri, N., and Teichner, S. J., J. Catal. 62, 99
(1980).
58. Smith, M. B., and March, J., in “March’s Advanced Organic Chem-
istry: Reactions, Mechanisms and Structure,” 5th ed., p. 180. Wiley-
Interscience, New York, 2001.
20. Mallat, T., and Baiker, A., Catal. Today 19, 247 (1994).
21. Brink, G., Arends, I. W. C. E., and Sheldon, R. A., Science 287, 1636
(2000).
22. Srinivas, N., Radha Rani, V., Radha Kishan, M., Kulkarni, S. J., and
Raghavan, K. V., J. Mol. Catal. A 172, 187 (2001).
23. Jensen, D. R., Pugsley, J. S., and Signam, M. S., J. Am. Chem. Soc. 123,
7475 (2001).
24. Sheldon, R. A., Arends, I.W. C. E., and Dijkman, A., Catal. Today 57,
157 (2000).
25. Gerisher, H., and Willig, F., Top. Curr. Chem. 61, 50 (1976).
26. Arbour, C., Sharma, D. K., and Langford, C. H., J. Chem. Soc.
Chem.Commun. 917 (1987).
27. Matthew, R. W., J. Catal. 111, 264 (1988).
28. Tunesi, S., and Anderson, M. A., J. Phys. Chem. 95, 3399 (1991).
29. Sabate, J., Anderson, M. A., Kikkawa, H., Xu, Q., Cervera-March, S.,
and Hill, C. G., J. Catal. 134, 36 (1992).
59. Peterson, M. W., Jurner, J. A., and Nozik, A. J., J. Phys. Chem. 95, 221
(1991).
30. Wold, A.,Chem. Mater. 5, 280 (1993).
31. Vinodgopal, K., Hotchandani, S., and Kamat, P. V., J. Phys. Chem. 97,
9040 (1993).
32. Al-Ekabi, H., and Serpone, N., J. Phys. Chem. 92, 5726 (1988), and
references therein.
33. Matthews, R. W., J. Phys. Chem. 91, 3328 (1987).
34. Fox, M. A., Acc. Chem. Res. 16, 314 (1983).
35. Fox, M. A., New J. Chem. 11, 129 (1987).
36. Ollis, D. F., Hsiao, C. Y., Budiman, L., and Lee, C. L., J. Catal. 88, 89
(1984).
60. Mills, A., Davies, R. H., and Worsely, D., Chem. Soc. Rev. 22, 417
(1993).
61. Martra, G., Coluccia, S., Marchese, L., Augugliaro, V., and Loddo, V.,
Catal. Today 53, 353 (1998).
62. Cunningham, J., and Hodnett, B. K., J. Chem. Soc. Faraday Trans. 77,
2777 (1981).
63. Peral, J., and Ollis, D. F., in “Proc. Ist Int. Conf. TiO2 Photocata-
lytic Purification and Treatment of Water and Air” (D. F. Ollis
and H. Al-Ekabi, Eds.), p. 741. Elsevier, Amsterdam/New York,
1993.
37. Forments, M., Juillet, F., Meriaudeau, P., and Teichner, S. J.,
CHEMTECH 680 (1971).
38. Fox, M. A., and Chen, C. C., J. Am. Chem. Soc. 103, 6757 (1981).
39. Pincock, J. A., Alexandra, L. P., and Fox, M. A., Tetrahedron 41, 4107
(1985).
64. Blake, N. R., and Griffin, G. L., J. Phys. Chem. 92, 5697 (1988).
65. Larson, S. A., and Falconer, J. L., Appl. Catal. B 4, 325 (1994).
66. Sauer, M. L., Hale, M. A., and Ollis, D. F., Photochem. Photobiol. A
88, 169 (1995).
67. Luo, Y., and Ollis, D. F., J. Catal. 163, 1 (1996).
68. Serpone, N., J. Photochem. Photobiol. A 104, 1 (1997).
40. Fox, M. A., Photocatal. Environ. 237, 445 (1988).