Photocatalytic selective oxidation of hydrocarbons in the aqueous phase

Article abstract of DOI:10.1006/jcat.1999.2395

The sustainable transformation of an inert alkane into its corresponding oxygenates has been the subject of intense chemical research. These oxygenates typically produced from processes using stringent conditions and materials offer disadvantages that inc

Full text of DOI:10.1006/jcat.1999.2395

Journal of Catalysis 183, 159–162 (1999)
Article ID jcat.1999.2395, available online at http://www.idealibrary.com on
RESEARCH NOTE
Photocatalytic Selective Oxidation of Hydrocarbons
in the Aqueous Phase
Michael A. Gonzalez,1 S. Garry Howell, and Subhas K. Sikdar
National Risk Management Research Laboratory, Sustainable Technology Division, Clean Processes and Products Branch,
U.S. Environmental Protection Agency, 26 West Martin Luther King Drive, Mail Stop 443, Cincinnati, Ohio 45268
Received October 5, 1998; revised December 23, 1998; accepted December 26, 1998
ert, adding to the environmental friendliness of this process.
On irradiation with ultraviolet light below a wavelength of
390 nm, the inert anatase TiO2 generates electron holes
which are believed to have an oxidation potential of 3.2 V
(1). The high-degree of oxidation potential of this material
has been demonstrated by the photocatalytic activation of
halogenated hydrocarbons, aromatics, and phenols for the
complete destruction, i.e., mineralization, of these environ-
mental pollutants (2–5).
The sustainable transformation of an inert alkane into its corre-
sponding oxygenates has been the subject of intense chemical re-
search. These oxygenates typically produced from processes using
stringent conditions and materials offer disadvantages that include
decreased selectivities to the partial oxygenates and by-product for-
mation. As environmental concerns and regulations become more
rigorous, the need for alternative catalytic oxidation processes that
use mild or ambient conditions is increased. In this Note, the authors
have investigated the use of photocatalysis as a low-temperature
and “green” alternative for the direct and selective oxidation of
alkanes with molecular oxygen.
Previous attempts to use photocatalysis for the produc-
tion of oxygenates from saturated hydrocarbons have been
reported (6–10); however, many of these studies were per-
formed as a proof-of concept and the experimental con-
ditions were or may not have been fully optimized. Our
research efforts have been directed toward demonstrating
the feasibility and gaining an understanding of using photo-
catalysis as a synthesis tool. In doing so, the high oxidation
potential of TiO2 must be controlled to prevent overoxi-
dation of the primary oxygenates. This aqueous-based ox-
idation process is performed as a screening tool as well as
providing experimental information on the conditions nec-
essaryfor the activation ofhydrocarbons. Thisexperimental
information was then used in the design, construction, and
implementation of a gas-phase photocatalytic reactor sys-
tem (11, 12). It is anticipated the continuous-flow gas-phase
reactor, which eliminates a solvent stream, offers a more
controlled and characterized substrate : water : oxidant feed
stream, and the need for separation steps will increase the
feasibility of using a light-induced process for the oxidation
of alkanes.
Key Words: green chemistry; photocatalysis; hydrocarbon oxida-
tion; selective oxidation; TiO2; aqueous.
INTRODUCTION
The selective oxidation of alkanes and aromatics to their
corresponding alcohols, ketones, aldehydes, and carboxylic
acids offers an attractive synthesis of higher-value products
from relatively inexpensive feedstocks. However, many of
the current pathways to produce these products are envi-
ronmentally hazardous, offer low single-pass conversions,
and achieve product selectivities that are lower than de-
sired.
Recent research has seen an increased effort in the in-
vestigation of the use of alternative oxidation technologies
(AOTs) as catalytic processes for the oxidation of hydrocar-
bon substrates. One such alternative oxidation technology
that can possibly avoid most of the above-mentioned pit-
falls is photocatalysis. This light-induced catalytic reaction
possesses unique reaction conditions that include ambient
reaction temperatures and pressures, as well as the use of
an inexpensive semiconductive material and nontoxic sol-
vents. This material, TiO2, is biologically and chemically in-
In this Note, the authors have investigated the use of
photocatalysis as a low-temperature “green” alternative for
the direct and selective oxidation of saturated hydrocar-
bons with molecular oxygen in an aqueous environment.
Research has been directed toward the use of a heteroge-
neous liquid-phase reactor for the partial oxidation of cy-
clohexane, toluene, methylcyclohexane, ethylbenzene, and
cumene to their corresponding oxygenates.
1 To whom correspondence should be addressed. Fax: (513) 569–7677.
E-mail: gonzalez.michael@epamail.epa.gov.
159
0021-9517/99

160
GONZALEZ, HOWELL, AND SIKDAR
METHODS
sis was performed on a Perkin–Elmer Spectrum 2000 FT-IR
spectrometer equipped with a tungsten halogen source and
a DTGS detector. The sample was prepared as a KBr pellet.
Photoreactor. Oxidation experiments were performed
in an Ace Glass Micro Photochemical reactor assembly.
This apparatus consists of a 20-ml quartz batch annular
reactor. The TiO2 slurry was stirred mechanically using a
magnetic stirrer and additional agitation was provided by
bubbling air through the reactor. Air also served as an ad-
ditional oxidant in this process. Illumination was provided
by 5.5-W quartz Pen-Ray, low-pressure, cold cathode mer-
cury lamp. Maximum emission for this lamp occurs at a
wavelength of 365 nm. The lamp was placed in an 11-mm-
i.d. jacketed immersion well. Water was allowed to flow
through the jacket to remove any heat produced by the
lamp. The entire reactor assembly was wrapped with alu-
minum foil to prevent light from entering the apparatus.
RESULTS AND DISCUSSION
Prior to performing the oxidation reactions, a number
of blank experiments were performed. These experiments
would demonstrate if the desired oxidation reactions were
proceeding via a photocatalyzed pathway. Conditions for
the blank runs constituted reactions performed in the ab-
sence of catalyst and “dark” reactions, in the absence of
UV light. Each blank experiment exhibited no measurable
amount of oxidation products for each substrate, thereby
demonstrating the oxidation reaction was proceeding via a
photocatalyzed mechanism.
The substrates were chosen for their unique properties.
Toluene, ethylbenzene, and cumene, each beingan aromatic
and possessing benzyllic carbons, are expected to be quite
reactive toward a free radical oxidation. Cumene, with a ter-
tiary benzyllic carbon, offering a more stable free radical,
should be the most responsive toward oxidation. Cyclohex-
ane and methylcyclohexane, both saturated hydrocarbons,
were chosen to demonstrate the feasibility of using this ox-
idation reaction for the activation of alkanes. Methylcy-
clohexane, with its pendant methyl group, allowed us to
observe the regioselectivity of this free radical oxidation
reaction. In most cases, the overoxidation of the primary
oxygenates to aldehydes and ketones can also be observed.
Oxidation experiments performed with toluene (See
Table 1) exhibited a direct selectivity to benzaldehyde
(90.9% ), the result of overoxidation of benzyl alcohol
Reagents. Finely divided titanium dioxide (TiO2) P25
was used as received from Degussa. This catalyst is mostly
anatase phase and has a BET surface area and average par-
ticle size of 50 10 m2 g 1 and 30 nm, respectively. Toluene,
cyclohexane, methylcyclohexane, cumene, ethylbenzene,
and o-dichlorobenzene were purchased from Aldrich. All
chemicals were used as received, without any further pu-
rification.
Oxidation procedure. A specified quantity of TiO2 was
added to 13 ml of deionized water, and the mixture was
stirred for 10 min while being purged with oxygen. The
substrate (2.0 ml) was introduced into the reaction mixture
by pipett and the lamp assembly added. If additional oxi-
dant was needed, air was bubbled into the reactor during
the specified reaction time.
Analysis. All samples were centrifuged for 10 min to (9.1% selectivity). Detection for the formation of benzoic
facilitate separation of the TiO2 particles from the solution acid by derivatization of the aqueous and solid fractions
and then filtered. The filtrate was then collected for GC/MS with BF3/MeOH was attempted. Results from this test de-
analysis. The supernate was rinsed with ethanol to remove termined no benzoic acid had been produced. IR analysis
any additional products and then centrifuged and filtered. of the spent catalyst also confirmed no benzoic acid had
On separation the filtrate fraction was collected for GC/MS been generated. An overall conversion of 11.61% , based
analysis.
The reaction products were analyzed using a Hewlett–
on products, was obtained for this oxidation reaction.
Experiments for the oxidation of cumene (Table 1) were
Packard 6890 gas chromatograph outfitted with a low- not as successful. It was anticipated cumene would be more
bleed HP-5MS (30 m 0.25 mm 0.25 m) column and a reactive toward oxygenation under the current conditions.
split/splitless injector. The carrier gas used was helium. A However, no oxidation products were observed. Attempts
Hewlett–Packard 5973 mass selective detector equipped to increase the oxidizing power of the reaction, by adding
with a quadrupole mass filter was used as the detector. Each of 30% H2O2 (3.0 ml), adding larger quantities of catalyst
sample had 0.05 ml of o-dichlorobenzene added as an in- (twofold), bubbling air (10 ml min 1) through the reac-
ternal standard. Samples were then analyzed in duplicate tion slurry, and increasing reaction time (8 h), also demon-
with an injection volume of 1.0 l. Quantification of the strated lack of oxidized products. Navio and co-workers
oxygenated products was obtained using a multipoint cali- (13) and Nizova and Shul’pin (14) reported that the oxida-
bration curve for each product.
tion of cumene proceeds via formation of a cumylhydroper-
Analysis for CO and CO2 was performed on a Hewlett– oxide intermediate. This intermediate then forms acetophe-
Packard 5890 gas chromatograph equipped with a J.W. Sci- none and methanol via a C–C bond cleavage mechanism.
entific GS-GasPro capillary column (30 m 0.32 mm) and a The methanol is subsequently overoxidized to form CO2.
thermalconductivitydetector. Infrared spectroscopyanaly- The decreased activity in our process is possibly due to an

PHOTOCATALYTIC SELECTIVE OXIDATION OF HYDROCARBONS
161
TABLE 1
periments with methylcyclohexane yielded products with
oxidation at a number of positions about the ring. Of these
products the most prevalent were the four isomers for
methylcyclohexanol and the three isomers for methylcyclo-
hexanone. Higher selectivities were exhibited for the prod-
ucts with oxidation at the 2- and 3-positions. Also observed
was the direct oxidation of the primary methyl position to
produce 1-methanolcyclohexane (5.77% selectivity). Al-
though oxidation at this position is difficult, as indicated by
the low selectivity, this process at least allows for this oxida-
tion to occur. Oxidation at the tertiary carbon accounts for
15.52% of the oxygenates, while oxidation at the secondary
carbons accounts for 78.7% of the oxidized products. An
overall conversion of 8.18% was obtained for methylcyclo-
hexane.
The oxidation of cyclohexane also results in the pro-
duction of the corresponding oxygenates. Attempts to oxi-
dize cyclohexane in the absence of hydrogen peroxide did
not afford any oxygenates. On addition of hydrogen per-
oxide (3.0 ml), cyclohexanol and cyclohexanone were ob-
tained. It is speculated that the addition of hydrogen per-
oxide provides the reaction with a greater concentration of
OH species necessary for activation of the inert saturated
hydrocarbon. Oxidation products obtained were cyclohex-
anol (30.08% selectivity) and cyclohexanone (44.03% se-
lectivity), along with small quantities of 2-cyclohexeneone
(5.1% selectivity) and trace quantities of cyclohexylhy-
droperoxide. The coupling reaction of two cyclohexane rad-
ical species to produce 1,10-bicyclohexane (20.79% selectiv-
ity) also occurred. It is anticipated that on increasing the
oxidant concentration in the reaction slurry, the undesired
termination reaction can be inhibited and replaced with
oxygenation of the radical species. Blank runs performed
for the oxidation of cyclohexane with hydrogen peroxide
and UV light in the absence of TiO2 did produce trace
quantities of cyclohexanol and cyclohexanone. However,
the amounts produced were considerably less than those
generated in the presence of TiO2, demonstrating the de-
composition of hydrogen peroxide into OH radicals in the
presence of light is not solely responsible for the oxygena-
tion of cyclohexane.
Results for Liquid-Phase Photocatalyzed Oxidation
of Hydrocarbonsa
Selectivityc Conversiond
Substrate
Toluene
Products
(% )
(% )
Benzyl alcohol
Benzaldehyde
9.1
90.9
11.61
Methylcyclohexane 1-Methylcyclohexanol
2-Methylcyclohexanol
15.52
7.63
3-Methylcyclohexanol
4-Methylcyclohexanol
2-Methylcyclohexanone
3-Methylcyclohexanone
4-Methylcyclohexanone
1-Methanolcyclohexane
15.89
4.94
15.18
24.20
10.86
5.77
8.18
4.24
Cyclohexaneb
Cyclohexanol
30.08
44.03
5.10
Cyclohexanone
2-Cyclohexeneone
1,10-Bicyclohexane
20.79
Ethylbenzene
Cumene
Acetophenone
None
100
None
6.71
0.00
a Reaction conditions, 1.0 g TiO2 (Degaussa P25), 13 ml deionized H2O,
2 ml substrate, 5-W UV–vis Pen Ray lamp, reaction time 2 h.
b 3.0 ml 30% H2O2 added.
c Selectivity is calculated as moles of specified product/total moles of
oxygenates expressed as a percentage.
d Conversion is calculated as total moles of oxygenates/moles of sub-
strate (initial) expressed as a percentage.
insufficient concentration of cumylhydroperoxide. The in-
ability to form a large quantity of the alkylhydroperoxide
may be due to the low lamp power, an insufficient quantity
of oxidant, or solvent incompatibility. Further demonstrat-
ing the lack of reactivity of cumene is the inability to pro-
duce phenol or acetone. This result is attributed to the lack
of oxidation potential necessary for the cleavage of the ben-
zyllic and tertiary carbon bond. Additional investigation is
needed to determine the exact cause and how to improve
on the current results.
To further investigate the inability to activate cumene,
oxidation experiments were performed on ethylbenzene.
This molecule, which possesses a secondary benzyllic car-
bon, should afford acetophenone as the sole product.
Table 1 exhibits the results obtained for this oxidation af-
ter 2 h of reaction. The sole reaction product, acetophenone
(100% selectivity), was obtained with an overall conversion
of 6.71% . No additional products were obtained, including
CO or CO2. It is interesting that only one product was ob-
tained, in addition to the lack of cleavage products. It is
apparent the oxidation of ethylbenzene is proceeding in a
manner identical to that for toluene, benzyllic carbon attack
and subsequent oxidation.
Increasing the oxygen concentration in the reaction
slurry, by bubbling air or molecular oxygen, did not result
in an increase in conversion. Since the reactor is not a
closed system, bubbling gas increases the loss or escape of
the volatile substrates and oxygenates from the solution.
The design of the current reactor apparatus did not allow
for the system to be pressurized. This will be corrected in
the future to allow additional oxidant to be supplied to the
reaction slurry.
Although a batch reactor was used, a high degree of se-
lectivity toward the partial oxygenates was achieved. Fur-
ther oxidation of the alcohols to ketones and aldehydes
was obtained in the case of ethylbenzene, cyclohexane, and
Results obtained for the oxygenation of the saturated
alkanes were also encouraging (see Table 1). Oxidation ex-

162
GONZALEZ, HOWELL, AND SIKDAR
methylcyclohexane. However, subsequent overoxidation nology as a potential candidate for replacement of current
and ring opening did not occur, exhibiting the selective na- oxidation processes.
ture of this oxidation process. As with all batch processes
decreasing reaction time results in higher selectivities with
an overall decrease in conversion. Increasing reaction times
results in an opposite effect for selectivity and conversion.
Therefore, a medium must be obtained; in this case it was
2 h. Experiments performed with increased reaction times
(4 and 6 h) did lead to increased conversions, but not much
ACKNOWLEDGMENTS
This research was supported in part by an appointment to the Post-
graduate Research Program at the U.S. Environmental Protection Agency
(USEPA), Office of Research and Development, at the National Risk
Management Research Laboratory administered by the Oak Ridge Insti-
tute for Science and Education through an interagencyagreement between
improved over those represented in Table 1. Selectivities the U.S. Department of Energy and the USEPA.
toward the alcohols did decrease along with subsequent in-
crease in aldehyde and ketone formation, but degradation
of the oxygenates did not result. This may be attributed to
the low lamp power (5 W) used and the limited quantities
of oxygen within the reactor.
REFERENCES
1. Mills, A., and Le Hunte, S., J. Photochem. Photobiol. A 108, 1
(1998).
2. Sykora, J., Pado, M., Tatarko, M., and Izakovic, M., J. Photochem.
Photobiol. A 110, 167 (1997).
3. Pozzo, R. L., Baltanas, M. A., and Cassano, A. E., Catal. Today 39, 219
(1997).
CONCLUSIONS
The direct activation and selective oxidation of satu-
rated hydrocarbons to their corresponding oxygenates via a
“green” process offer an attractive pathway for the produc-
tion of commercial oxygenates. A catalytic oxidation reac-
tion that proceeds in water with molecular oxygen at room
temperatures offers many of the desired components for an
oxidation process. The results for the photocatalyzed oxi-
dation of toluene, ethylbenzene, cyclohexane, and methyl-
cyclohexane demonstrate the feasibility of using this al-
ternative oxidation technology as a synthesis tool. Further
research must be performed to increase the efficiency and
kinetics of the described reactions. Although this study is in
the early stages, the current results demonstrate the ability
to activate and selectively produce oxygenated products,
without extensive overoxidation and degradation of the
4. Lichtin, N. N., and Sadeghi, M., J. Photochem. Photobiol. A 113, 81
(1998).
5. Rideh, L., Wehrer, A., Ronze, D., and Zoulalian, A., Ind. Eng. Chem.
Res. 36, 4712 (1997).
6. Formenti, M., Juillet, F., Meriaudeau, P., and Teichner, S. J., Chem.
Tech., 680 (1971).
7. Fujihira, M., Satoh, Y., and Osa, T., Nature 293, 206 (1981).
8. Izumi, I., Dunn, W. W., Wilbourn, K. O., Fan, F. F., and Bard, A. J.,
J. Phys. Chem. 84, 3207 (1980).
9. Giannetti, C., Le Greneur, S., and Watts, S., Tetrahedron Lett. 24, 46,
5071 (1983).
10. Formenti, M., Juillet, F., and Teichner, S. J., Bull. Fr. Chem. Soc. 9/10,
1315 (1976).
11. Sahle-Demessie, E., and Gonzalez, M. A., manuscript in preparation.
12. Sahle-Demessie, E., Gonzalez, M. A., Wang, Z., and Biswas, P., sub-
mitted for publication.
13. Wittenberg, R., Pradera, M. A., and Navio, J. A., Langmuir 13, 2373
(1997).
desired oxygenates, leading to consideration of this tech- 14. Nizova, G. V., and Shul’pin, G. B., Russ. Chem. Bull. 45, 1353 (1996).