Photooxidation of 1-alkenes in zeolites: A study of the factors that influence product selectivity and formation

Article abstract of DOI:10.1021/ja984477a

Photochemical oxidation of hydrocarbons with molecular oxygen is potentially an environmentally benign method for the selective oxidation of hydrocarbons. In this study, in situ FT-IR spectroscopy and ex situ NMR spectroscopy were used to investigate the factors that influence product formation and selectivity in the room-temperature photooxidation of 1- alkenes in zeolites. Upon irradiation with broadband visible light, propylene, l-butene, and 1-pentene loaded in BaY were photooxidized with molecular oxygen. As discussed in the literature, initial excitation of alkene and molecular complexes results in the selective formation of unsaturated aldehydes and ketones, proposed to occur through a hydroperoxide intermediate. In addition, epoxide and alcohol products are formed when the hydroperoxide intermediate reacts with an unreacted parent alkene molecule. Here it is shown that saturated aldehydes and ketones are formed as well through both a thermal ring-opening reaction of the epoxide in BaY and a second photochemical oxidation route involving a dioxetane intermediate. The yield of saturated aldehydes and ketones increased with decreasing wavelength, increasing temperature, and at a given temperature and wavelength, increasing chain length. Photooxidation of propylene in BaX, BaZSM-5, and BaBeta zeolites was also investigated. Photooxidation in BaX is very similar to thai of BaY. In zeolites, BaZSM-5 and BaBeta, propylene polymerized upon adsorption. The polymer, polypropylene, also undergoes photooxidation with molecular oxygen to form an oxygenated polymer product. The results of this study show that product formation and selectivity in the photooxidation of 1-alkenes in zeolites depends on several factors. These factors include thermal reactions of the reactant and photoproduct molecules in the zeolite at ambient temperatures. Several reactions of l-alkenes in cation-exchanged zeolites contribute to the loss of selectivity; they include: epoxide ring opening, double-bond migration, and alkene polymerization. Some of these reactions are proposed to occur at Bronsted acid sites that are present in various amounts in cation-exchanged zeolites.

Full text of DOI:10.1021/ja984477a

J. Am. Chem. Soc. 1999, 121, 5063-5072
5063
Photooxidation of 1-Alkenes in Zeolites: A Study of the Factors that
Influence Product Selectivity and Formation
Y. Xiang, S. C. Larsen,* and V. H. Grassian*
Contribution from the Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
ReceiVed December 29, 1998
Abstract: Photochemical oxidation of hydrocarbons with molecular oxygen is potentially an environmentally
benign method for the selective oxidation of hydrocarbons. In this study, in situ FT-IR spectroscopy and ex
situ NMR spectroscopy were used to investigate the factors that influence product formation and selectivity in
the room-temperature photooxidation of 1-alkenes in zeolites. Upon irradiation with broadband visible light,
propylene, 1-butene, and 1-pentene loaded in BaY were photooxidized with molecular oxygen. As discussed
in the literature, initial excitation of alkene and molecular complexes results in the selective formation of
unsaturated aldehydes and ketones, proposed to occur through a hydroperoxide intermediate. In addition, epoxide
and alcohol products are formed when the hydroperoxide intermediate reacts with an unreacted parent alkene
molecule. Here it is shown that saturated aldehydes and ketones are formed as well through both a thermal
ring-opening reaction of the epoxide in BaY and a second photochemical oxidation route involving a dioxetane
intermediate. The yield of saturated aldehydes and ketones increased with decreasing wavelength, increasing
temperature, and at a given temperature and wavelength, increasing chain length. Photooxidation of propylene
in BaX, BaZSM-5, and BaBeta zeolites was also investigated. Photooxidation in BaX is very similar to that
of BaY. In zeolites, BaZSM-5 and BaBeta, propylene polymerized upon adsorption. The polymer, polypropylene,
also undergoes photooxidation with molecular oxygen to form an oxygenated polymer product. The results of
this study show that product formation and selectivity in the photooxidation of 1-alkenes in zeolites depends
on several factors. These factors include thermal reactions of the reactant and photoproduct molecules in the
zeolite at ambient temperatures. Several reactions of 1-alkenes in cation-exchanged zeolites contribute to the
loss of selectivity; they include: epoxide ring opening, double-bond migration, and alkene polymerization.
Some of these reactions are proposed to occur at Brønsted acid sites that are present in various amounts in
cation-exchanged zeolites.
Introduction
demonstrated that upon irradiation with visible light, small
hydrocarbons such as propylene and 2-butenes loaded in zeolites
BaY and NaY, respectively, can be selectively oxidized to allyl
hydroperoxide and 3-hydroperoxy-1-butene by molecular oxy-
gen at low temperatures near -100 °C.5,7 Upon warming to 0
°C, allyl hydroperoxide and 3-hydroperoxy-1-butene convert to
acrolein and methyl vinyl ketone, respectively; water is also a
product in the reaction. The mechanism for propylene photo-
oxidation is outline in Scheme 1A. Photolysis of the hydrocarbon‚
O2 complex with visible light results in a charge-transfer
complex in the zeolite cage.7 The hypothesis is that the
electrostatic field of the cation-exchanged zeolite (on the order
of several V nm-1) stabilizes the excited charge-transfer state
which otherwise can only be accessed by high-energy ultraviolet
photons. The allyl hydroperoxide is produced from the charge
transfer state which subsequently goes on to form acrolein. The
use of visible rather than ultraviolet irradiation allows access
to this low-energy pathway, eliminates many secondary pho-
toprocesses, and leads to product selectivity. The geometric
constraints of the zeolite framework are also thought to be
important in achieving high reaction selectivity. The implications
of this methodology for environmentally benign synthesis of
chemicals are 2-fold. The increased selectivity of the oxidation
reaction will decrease the production of unwanted side products.
One of the most important applications of homogeneous
catalysis is the oxidation of hydrocarbons to produce industrially
important chemicals, such as terephthalic acid and adipic acid,
which are used in the production of polymers.1,2 Oxidation by
O2 with homogeneous catalysts exhibits very low selectivity
because side reactions are difficult to control and intermediates
are often more easily oxidized than the parent under thermal
conditions.1,2 The problem with autoxidation is that it is slow
and nonselective.3 One challenge is to find new, more efficient
catalysts for the conversion of hydrocarbons into oxygenates.
A second challenge is to develop processes that have a minimal
environmental impact.
Recently, Frei and co-workers have demonstrated that
hydrocarbons in zeolite Y can be selectively oxidized by visible
light with relatively high quantum yields (∼0.3).4-11 It has been
* To whom correspondence should be addressed.
(1) Gates, B. C. Catalytic Chemistry; Wiley: New York, 1992.
(2) Parshall, G. W.; Ittel, S. D. Homogeneous Catalysis, 2nd ed.;
Wiley: New York, 1992.
(3) Roby, A.; Kingsley, J. P. CHEMTECH 1996, 39.
(4) Blatter, F.; Frei, H. J. Am. Chem. Soc. 1993, 115, 7501-7502.
(5) Blatter, F.; Frei, H. J. Am. Chem. Soc. 1994, 116, 1812-1820.
(6) Blatter, F.; Moreau, F.; Frei, H. J. Phys. Chem. 1994, 98, 13403-
13407.
(7) Blatter, F.; Sun, H.; Frei, H. Catal. Lett. 1995, 35, 1-12.
(8) Frei, H.; Blatter, F.; Sun, H. CHEMTECH 1996, 24-30.
(9) Sun, H.; Blatter, F.; Frei, H. J. Am. Chem. Soc. 1994, 116, 7951-
7952.
(10) Sun, H.; Blatter, F.; Frei, H. Catal. Lett. 1997, 44, 247-253.
(11) Blatter, F.; Sun, H.; Vasenkov; Frei, H. Catal. Today 1998, 41, 297-
309.
10.1021/ja984477a CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/15/1999

5064 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Xiang et al.
Scheme 1. Proposed Mechanism for the Formation of
Unsaturated Carbonyl Compounds
need to be done as close to ambient temperature as possible,
rather than low temperatures, if they are to be of industrial use.
The objective of this study is to further develop this
methodology for the environmentally benign synthesis of
industrially important chemicals as an alternative to conventional
liquid-phase oxidation and to determine its limitations.12 Here
we report on the photooxidation of 1-alkenes in zeolites Y, X,
ZSM-5, and Beta at room temperature. The Ba2+-exchanged
form of these zeolites was used to achieve high electrostatic
field strengths on the order of 4-8 V nm-1. These zeolites were
chosen because they differ in chemical composition and
structure, i.e., Si/Al ratio and framework topography. If it can
be shown that the photooxidation of hydrocarbons in zeolites
is a general phenomenon, then the shape and size selective
properties of these zeolites may potentially be used to impact
further control the selectivity of these reactions. For example,
ZSM-5 is an important shape-selective catalyst in many reac-
tions, such as the disproportionation reaction of toluene.13
p-Xylene is the dominant product because the transport of the
other isomers, o- and m-xylene, is restricted due to the pore
size of ZSM-5. Thus, stereochemical aspects of these reactions
may also be influenced by the zeolite and may be used in the
design of environmentally benign syntheses of industrially
important molecules.
In this study, in situ FT-IR spectroscopy and ex situ NMR
spectroscopy were used to analyze product formation and
product yields. This combined approach allows for a more
detailed analysis of the product distribution. As discussed here,
several factors influence product selectivity and formation in
these reactions. These factors include excitation wavelength,
temperature, hydrocarbon chain length, and the parent zeolite.
Experimental Section
In situ FT-IR spectra were recorded with a Mattson RS-10000
infrared spectrometer equipped with a narrowband MCT detector. Each
spectrum was taken by averaging 500 scans at an instrument resolution
of 4 cm-1. The infrared sample cell used in this study has been described
previously.14 Briefly, 50-70 mg zeolite is sprayed from a water slurry
onto a 3 × 2 cm2 photoetched tungsten grid held at 40 °C. The tungsten
grid coated with the zeolite is mounted onto nickel jaws that are attached
to a copper feedthrough. The sample can be heated and cooled from
150 to 1200 K. The temperature of the sample is measured by a
thermocouple wire spot-welded to the center of the grid. The entire
assembly is mounted inside a 2.75-in. stainless steel cube with BaF2
windows. The IR cell is then evacuated by a turbomolecular pump to
a pressure of 1 × 10-7 Torr. Zeolites are heated under vacuum to 300
°C for BaY and BaX, 350 °C for BaZSM-5, and 450 °C for Beta and
kept at the temperature overnight to remove adsorbed water.
Alkenes were loaded into the zeolite by adsorption under an
equilibrium gas pressure of 15-16 Torr at room temperature, lower
temperatures for some experiments, for 45 min. The excess hydrocarbon
was then pumped out for 5 min, and oxygen was added to the IR cell
at a pressure of approximately 200 Torr. A 500-W mercury lamp (Oriel
Corp.) with a water filter was used as the light source for photolysis.
Broadband long pass filters were placed in front of the lamp (Oriel
Corp., filter 59492, %T ) 0 at 495 nm; filter 59472, %T ) 0 at 400
nm; filter 59460, %T ) 0 at 348 nm; and filter 59423, %T ) 0 at 285
nm). The broadband light was then reflected off an aluminum coated
mirror and turned by a 1-in. quartz prism onto the zeolite sample. The
quartz prism is mounted inside the FT-IR sample compartment so that
the dry air purge was not broken during irradiation. For room-
(12) Myli, K. B.; Larsen, S. C.; Grassian, V. H. Catal. Lett. 1997, 48,
199-202.
In addition, a zeolite catalyst and gas-phase reactants and
products will eliminate the use of expensive and environmentally
toxic solvents. From a practical point of view, these reactions
(13) Venuto, P. B. Microporous Mater. 1994, 2, 297-411.
(14) Miller, T. M.; Grassian, V. H. J. Am. Chem. Soc. 1995, 117, 10969-
10975.

Photooxidation of 1-Alkenes in Zeolites
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5065
temperature experiments, even with the water filter in place, the zeolite
sample warmed to between 35 and 45 °C after several hours of
irradiation.
Ex situ proton NMR spectra were recorded at room temperature on
a Bruker NMR AC-300 spectrometer. Samples were prepared for ex
situ analysis as follows. After photooxidation of the alkene in the IR
cell at conversions on the order of 45%, the zeolite containing unreacted
parent and product molecules was removed from the tungsten grid and
suspended in CDCl3. The suspension was placed in a sonic bath for 1
h, and then the zeolite powder was filtered from the suspension. The
resulting solution was placed in a 5-mm NMR tube with TMS
(tetramethylsilane) as an internal standard before acquisition of the 1H
NMR spectrum. Several control experiments were done to ensure that
the extracted product distribution was representative of the reaction
product distribution. Except for very volatile products and reactants,
e.g., formaldehyde and propylene, the reported product distributions
from ex situ NMR experiments are representative of the chemistry
occurring in the zeolite. NMR experiments done after longer and shorter
reaction times did not show an appreciable change in the product
distribution.
Figure 1. Difference infrared spectra before and after photolysis of
propylene and oxygen in zeolite BaY near room temperature with (a)
λ > 495 nm for 210 min, (b) λ > 400 nm for 20 min, (c) λ > 348 nm
for 8 min, and (d) λ > 285 nm for 5 min. Propylene losses are indicated
by asterisks. FT-IR spectra of (e) 1:1 acetaldehyde and formaldehyde
in zeolite BaY.
The Ba2+-forms of zeolites Y, X, ZSM-5, and Beta were used so
that high electrostatic field strengths could be achieved. The shift in
the CO stretching frequency upon adsorption in the zeolite was used
to estimate the electrostatic field strength.15-17 For BaY, BaX, BaZSM-
5, and BaBeta the field strength is estimated to be 6.3, 4.8, 7.4, and
7.8 V nm-1. Zeolites BaY, BaX, BaZSM-5, and BaBeta were prepared
from NaY (Aldrich), NaX (Acros Organics), NaZSM-5 (Zeolyst), and
was placed at the output of the lamp so that only wavelengths
above 495 nm were incident on the sample. Figure 1a shows
the difference spectrum, i.e., the spectrum recorded before
irradiation subtracted from the spectrum recorded following
irradiation for 210 min. The sample warmed to 45 °C during
irradiation. The difference spectrum shows negative features due
to the loss of propylene; these bands are marked by asterisks.
The most intense positive features are observed at 1668 and
1366 cm-1 and are due to the formation of photoproduct
molecules. These two absorption bands match the bands
observed for an authentic spectrum of acrolein in zeolite BaY
(see spectrum shown in Figure 1e).
+
NH4 Beta (Zeolyst) by standard ion-exchange procedures at 90 °C
with aqueous 0.5 M BaCl2 (Aldrich) solutions. The Si/Al ratios for
BaY, BaX, BaZSM-5, and BaBeta were determined by ICP-AES
(inductively coupled plasma atomic emission spectroscopy) to be 2.4,
1.4, 19, and 16.5, respectively. The Ba/Al ratios for BaY, BaX, BaZSM-
5, and BaBeta were determined to be 0.39, 0.33, 0.26, and 0.49,
respectively.
To qualitatively test for Brønsted acidity in these cation-exchanged
zeolites, a colorimetric method was used. Approximately 50 mg of the
zeolite sample was pretreated at 300 °C under vacuum for ap-
proximately 12 h. One milliliter of a dilute solution of retinol (Aldrich)
or retinyl acetate (Aldrich) in dry hexane was injected onto the activated
zeolite. A color change to blue indicated the presence of Brønsted acid
sites.18
Propylene (Matheson, 99.6% purity), 1-butene (Matheson, 99.9%
purity), and O2 (Air Products, 99.6% purity) were used without further
purification. 1-Pentene (99% purity) was purchased from Aldrich and
subjected to several freeze-pump-thaw cycles prior to use. Standards
of acrolein (97% purity), allyl alcohol (99% purity), 3-butene-2-ol (99%
purity), butyraldehyde (99% purity), crotonaldehyde (99+% purity),
crotyl alcohol (97% purity), epoxybutane ((99% purity), epoxypropane
(99+% purity), ethyl vinyl ketone (97% purity), methyl vinyl ketone
(99% purity), trans 2-pentenal (99+% purity), cis 2-pentene (99%
purity), trans-2-pentenal (95% purity), and 3-penten-2-one and propi-
onaldehyde (97% purity) were purchased from Aldrich and subjected
to several freeze-pump-thaw cycles prior to use. Deuterated chloro-
form (Aldrich, CDCl3, 99.8 atom % D) was used for the NMR
experiments.
The effect of wavelength on the photooxidation of propylene
in BaY was examined. The same zeolite sample was used in
these experiments, and the irradiation times were chosen so that
the conversion at each wavelength was approximately the same.
Before changing the wavelength of light for irradiation, the
sample was heated under vacuum to 350 °C to remove adsorbed
propylene and adsorbed photoproducts. The difference spectra
upon irradiation at λ >400 nm for 20 min, λ >348 nm for 8
min, and λ >285 nm for 5 min, are shown in Figure 1b-d.
The conversion of propylene at each wavelength was estimated
from the integrated absorbance of the 1467 cm-1 propylene band
to be between 10 and 26% for each of the different wavelengths.
It is seen from the FT-IR spectra that acrolein is not the only
product observed in the IR spectra following the photooxidation
of propylene in BaY. A band indicative of saturated aldehydes
and ketones is observed near 1702 cm-1 as can be seen in the
photoproduct spectra (Figure 1a-d) and in the spectrum of a
1:1 mixture of acetaldehyde and formaldehyde loaded in BaY
(Figure 1f). A peak at 1502 cm-1 due to formaldehyde, not
obvious in the difference spectra, was observed in other FT-IR
experiments. Other saturated aldehydes, such as propionalde-
hyde, may also contribute to the band at 1702 cm-1. In addition,
the shoulder near 1645 cm-1, labeled δ H2O is due to the
bending mode of water, another product of the reaction.
In two separate experiments, photoproducts and unreacted
propylene in BaY were extracted in CDCl3 after irradiation for
several hours with broadband visible light at λ > 400 nm for
18 h and λ > 285 nm for 3.5 h. Room temperature proton NMR
spectra recorded for each sample are shown in Figure 2. Peak
assignments and integrated peak areas are presented in Table
Results
Photooxidation of Propylene in BaY: The Effect of
Wavelength. Propylene and molecular oxygen were added to
the zeolite as described in the Experimental Section. After an
initial spectrum of propylene and O2 was recorded, the sample
was irradiated with broadband light. A 495-nm broadband filter
(15) Bordiga, S.; Garrone, E.; Lamberti, C.; Zecchina, A.; Arean, C.;
Kazansky, V.; Kustov, L. J. Phys. Chem. 1994, 90, 3367-3372.
(16) Bordiga, S.; Lamberti, C.; Geobaldo, F.; Zecchina, A. Langmuir
1995, 11, 527-533.
(17) Angell, C. L.; Schaffer, P. C. J. Phys. Chem. 1966, 70, 1413-1418.
(18) Rao, V. J.; Perlstein, D. L.; Robbins, R. J.; Lakshminarasimhan, P.
H.; Kao, H.-M.; Grey, C. P.; Ramamurthy, V. Chem. Commun. 1998, 269-
270.

5066 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Xiang et al.
1
Table 2. Assignment of H NMR Spectrum for Photooxidation of
Propylene and Oxygen on BaY (λ > 285 nm)
peak
positiona
(ppm)
relative
integrated
area
multiplicity
unresolved
assignment
9.80
0.3
acetaldehyde (1H)
propionaldehyde (1H)
acrolein (1H)
9.59
6.40
3.00
2.76
2.46
doublet (6.8Hz)
multiplet
unresolved
triplet (4.5 Hz)
unresolved
1.0
4.0
0.5
0.5
0.9
acrolein (3H)
propylene oxide (1H)
propylene oxide (1H)
propylene oxide (1H)
propionaldehyde (2H)
acetaldehyde (3H)
acetone (3H)
propylene oxide (3H)
impurity
propionaldehyde (3H)
impurity
2.21
2.17
1.32
1.25
1.12
0.90
doublet (2.7 Hz)
singlet
doublet
singlet
triplet (7.3 Hz)
singlet
0.7
0.1
1.1
1.5
0.9
0.8
Figure 2. Proton NMR spectra of the products extracted in CDCl3
after photolysis of propylene and oxygen in BaY near room temperature
with (a) λ > 400 nm, 18 h and (b) λ > 285 nm, 3.5 h. Spectra are
referenced to an internal standard of TMS.
a Chemical shifts are measured relative to an internal standard of
TMS.
1
Table 1. Assignment of H NMR Spectrum for Photooxidation of
Propylene and Oxygen on BaY (λ > 400 nm)
weak, even after irradiation with the longest wavelength
broadband filters (λ > 495 nm and λ > 400 nm) used. However,
the amount of saturated aldehydes and ketones (as indicated by
the 1702 cm-1 band in the FT-IR spectrum) relative to acrolein
increases with excitation at the shorter wavelengths. The same
conclusions can be drawn from the ex situ NMR data. Only
very small amounts of saturated aldehydes and ketones are
observed after excitation at λ > 400 nm. However, the ratio of
acrolein to saturated aldehydes and ketones is approximately
1.0:0.3 after excitation at shorter wavelength, λ > 285 nm.
The mechanisms for the formation of acrolein and propylene
oxide, taken from the low temperature work of Frei and co-
workers, are shown in Schemes 1A and 2A.5-7 In both cases,
the allyl hydroperoxide is formed. At room temperature, inside
of the zeolite cage, the allyl hydroperoxide either fragments to
acrolein and water or reacts with an unreacted propylene
molecule to yield propylene oxide and allyl alcohol. The allyl
alcohol has been proven to be difficult to detect in ex situ NMR
experiments and may be due to the volatility of this compound.
The formation of saturated aldehydes, in particular the
nonselective, shorter chain ones at short wavelengths has been
attributed to a photooxidation mechanism of the alkene to form
a dioxetane intermediate that decomposes into shorter chain
aldehydes. This mechanism for propylene photooxidation is
shown in Scheme 3A. The other saturated aldehyde, propional-
dehyde, identified unambiguously in the ex situ NMR experi-
ments has not been previously detected in propylene photoox-
idation in zeolites, and its formation cannot be explained by
the dioxetane mechanism. A mechanism for propionaldehdye
formation which also accounts for the formation of acetone will
be discussed (vide infra).
peak
positiona
(ppm)
relative
integrated
area
multiplicity
unresolved
assignment
9.80
0.2
acetaldehyde (1H)
propionaldehyde (1H)
acrolein (1H)
9.59
6.40
3.00
2.76
2.46
doublet (6.8Hz)
multiplet
unresolved
unresolved
unresolved
1.0
4.0
0.3
0.3
0.5
acrolein (3H)
propylene oxide (1H)
propylene oxide (1H)
propylene oxide (1H)
propionaldehyde (2H)
acetaldehyde (3H)
acetone (3H)
propylene oxide (3H)
impurity
propionaldehyde (3H)
impurity
2.21
2.17
1.32
1.25
1.12
0.90
unresolved
unresolved
unresolved
broad peak
unresolved
singlet
0.3
trace
b
b
b
b
a Chemical shifts are measured relative to an internal standard of
TMS. b Integration unreliable due to broad background peak.
1. The products identified in Figure 2a after excitation with λ
> 400 nm are acrolein, propylene oxide, acetaldehyde, propi-
onaldehyde, and acetone. Through integration of the NMR
spectrum (Table 1), the ratio of acrolein:propylene oxide:
(acetaldehyde + propionaldehyde + acetone) is determined to
be approximately 1.0:0.3:0.2.19 The products identified in Figure
2b for excitation with λ > 285 nm are acrolein, propylene oxide,
acetaldehyde, propionaldehyde, and acetone. Using the inte-
grated areas in Table 2 the ratio of acrolein:propylene oxide:
(acetaldehyde + propionaldehyde + acetone) was determined
to be 1.0:0.5:0.3.
There are two conclusions that can be drawn from the
wavelength dependence of propylene photooxidation in BaY.
First, there is an increase in the rate of photoproduct formation
using the shorter wavelength broadband filters. For example,
the rate of formation of acrolein and propylene oxide is about
10 times higher at λ >400 nm than at λ >495 nm. Second,
there is a decrease in product selectivity at shorter wavelengths
as evidenced by the formation of nonselective saturated alde-
hydes, e.g., acetaldehyde, in the NMR spectrum. The appearance
of a band at 1702 cm-1 in the infrared spectrum is indicative
of saturated aldehydes and ketones. This band is apparent, albeit
Photooxidation of 1-Butene in BaY: The Effect of Tem-
perature. Photooxidation of 1-butene in BaY was investigated
at two different temperatures and two different wavelengths
using in situ FT-IR spectroscopy. The FT-IR spectra are shown
in Figure 3. An experimental procedure similar to the one used
for propylene was followed for 1-butene. The difference
spectrum in Figure 3a was made by spectral subtraction of the
spectrum recorded prior to irradiation and the spectrum recorded
after warming the sample to 25 °C. The negative absorption
bands are due to 1-butene depletion. The two most intense
features at 1629 and 1465 cm-1 are marked by asterisks. After
photooxidation at -20 °C for 5 h with λ > 400 nm followed
by warming to 24 °C, the intensity pattern of the bands in the
difference spectrum (Figure 2a), particularly bands near 1615
(19) Integrations shown in Table 1 for propylene oxide and the saturated
aldehydes and ketone (propionaldehyde, acetaldehyde, and acetone) are
highly uncertain because only trace quantities of these products are formed.

Photooxidation of 1-Alkenes in Zeolites
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5067
Scheme 2. Proposed Mechanism for the Formation of
Epoxides, Alcohols, Saturated Aldehydes, and Saturated
Ketones
Figure 3. Difference infrared spectra before and after photolysis of
1-butene and oxygen in zeolite BaY with (a) λ > 400 nm at -20 °C
for 5 h followed by warm-up at 24 °C without oxygen for 18 h and (b)
λ > 400 nm near room temperature for 1.5 h, (c) λ > 495 nm at 45 °C
for 14 h. Asterisks indicate the loss of 1-butene. FT-IR spectrum of
(d) methyl vinyl ketone in zeolite BaY.
Scheme 3. Proposed Mechanism for the Formation of
Saturated Aldehydes via a Dioxetane Intermediate
and 1584 cm-1, match the characteristic bands observed for
methyl vinyl ketone in BaY (Figure 3d). The shift of the
strongest band from 1660 cm-1 for methyl vinyl ketone standard
to 1653 cm-1 in the product spectrum is due to the coadsorption
of water, another product in the reaction. Heating briefly to 175
°C removes the coadsorbed water and shifts the bands to within
exact agreement with the authentic spectrum of methyl vinyl
ketone. The weak shoulder near 1702 cm-1 is assigned to the
formation of a small amount (∼5%) of saturated aldehydes and
ketones.
The difference spectrum shown in Figure 3 (b) was obtained
by spectral subtraction of the spectrum recorded prior to
irradiation and the spectrum recorded after irradiation with the
sample initially at room temperature with λ > 400 nm. The
sample warmed to ∼45 °C during the course of the experiment.
Note that the intensity pattern in the difference spectra shown
Figure 3b does not match that of methyl vinyl ketone in BaY
(Figure 3d). There are several additional bands in the spectrum,
and broadening of some of the bands in the methyl vinyl ketone
spectrum is observed, suggesting that other photooxidation
products with similar infrared spectra are formed. The identity
of these products can be more clearly determined by ex situ
NMR spectroscopy as discussed below. The FT-IR difference
spectrum obtained after irradiation with λ > 495 nm near room
temperature is shown in Figure 3c and is qualitatively very
similar to the spectrum obtained in Figure 3b after irradiation
of 1-butene with λ > 400 nm light. These data for the visible

5068 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Xiang et al.
1
Table 3. Assignment of H NMR Spectrum for Photooxidation of
on temperature rather than wavelength and, therefore, must be
due in part to a thermal reaction. Using in situ FT-IR and ex
situ NMR spectroscopy, we have observed that epoxides can
react in BaY to form saturated aldehydes and ketones at room
temperature. For example, when propylene oxide is adsorbed
on BaY and then extracted into CDCl3 for proton NMR
experiments, propylene oxide, acetone, and propionaldehyde are
observed in the NMR spectrum. In a parallel FT-IR experiment,
growth of the band at 1702 cm-1 was observed when propylene
oxide was adsorbed on BaY. Similarly, butyraldehyde and
2-butanone are observed when epoxybutane is adsorbed on BaY
and then extracted and analyzed with NMR spectroscopy. Thus,
ring opening and fragmentation of the epoxide product in the
zeolite at ambient temperatures results in the formation of
saturated aldehydes and ketones. These reactions for propylene
oxide and epoxybutane are shown in Scheme 2 A and B as
continued thermal chemistry of the epoxide. Similar epoxide
ring-opening reactions have been previously observed in zeo-
lites; it was determined that aliphatic epoxides rearrange to form
saturated aldehydes and ketones in zeolites that only contain a
few weakly acidic sites.20
As discussed in the case of propylene photooxidation, the
formation of nonselective, shorter chain oxidation products
increases when shorter wavelengths λ < 400 nm are used due
to a photooxidation mechanism involving a dioxetane intermedi-
ate (Scheme 3). The NMR data show that this route also occurs
for 1-butene and 1-pentene (vide infra) to some extent at λ >
400 nm as evidenced by the formation of propionaldehyde and
butyraldehyde (Scheme 3 B and C). This mechanism involving
a dioxetane intermediate is most likely the cause of the 1700
cm-1 shoulder observed in the low-temperature experiment as
well.
1-Butene and Oxygen on BaY (λ > 400 nm)
peak
positiona
(ppm)
relative
integrated
area
multiplicity
singlet
assignment
9.80
0.7
acetaldehyde (1H)
propionaldehyde (1H)
butyraldehyde (1H)
crotonaldehyde (1H)
crotonaldehyde (1H)
methyl vinyl ketone (2H)
methyl vinyl ketone (1H)
alcohol
9.77
9.50
6.89
6.27
5.92
4.50
singlet
doublet
0.9
1.0
1.2
6.5
2.5
multiplet (unresolved)
multiplet
doublet (9.8 Hz)
broad peak (narrow
peak at 4.52)
4.09
2.89
2.74
2.44
unresolved
alcohol product
multiplet (unresolved)
multiplet (unresolved)
multiplet
0.2
0.9
6.7
epoxybutane (1H)
epoxybutane (1H)
epoxybutane (1H)
propionaldehyde (2H)
butyraldehyde (2H)
2-butanone (2H)
2.29
2.21
2.14
2.04
1.69
singlet
5.5
0.6
2.1
2.1
2.7
methyl vinyl ketone (3H)
acetaldehyde (3H)
2-butanone (3H)
crotonaldehyde (3H)
epoxybutane (2H)
butyraldehyde (2H)
impurity
propionaldehyde (3H)
butyraldehyde (3H)
epoxybutane (3H)
2-butanone (3H)
doublet (unresolved)
singlet
doublet (6.6 Hz)
multiplet (unresolved)
1.28
1.03
multiplet (unresolved)
multiplet (unresolved)
7.1
4.6
a Chemical shifts are measured relative to an internal standard of
TMS.
photooxidation of 1-butene suggest that in these experiments
temperature, rather than wavelength is the more important factor
in product selectivity.
Detailed product distributions were determined by ex situ
NMR spectroscopy following irradiation of 1-butene and oxygen
in BaY with λ >400 nm light at 45 °C for 18 h. Products
adsorbed in BaY were extracted in CDCl3, and proton NMR
spectra were recorded. Peak assignments and integrated peak
areas are presented in Table 3. The reactant, 1-butene, and the
following photoproducts, methyl vinyl ketone, crotonaldehyde,
propionaldehyde, acetaldehyde, butyraldehyde, epoxybutane, and
2-butanone, are all observed in the proton NMR spectrum. There
is also evidence of an alcohol product in the 4 ppm region of
the proton NMR spectrum which is broadened, probably due
to hydrogen bonding of the alcohol product. The relative ratios
can be estimated from the integrated peak areas in Table 3. The
relative ratio of unsaturated carbonyls (methyl vinyl ketone and
crotonaldehyde):saturated aldehydes and ketones (propionalde-
hyde + acetaldehyde + butyraldehyde + 2-butananone) is
approximately 3.5:2.3. The relative ratio of methyl vinyl ketone
to crotonaldehyde is 2.5:1.0. Similar product ratios were
obtained following irradiation of 1-butene and oxygen in BaY
with λ >495 nm light at 45 °C for 18 h. Again, this suggests
that wavelength (λ > 400 vs λ > 495 nm) is not the factor that
controls the selectivity and formation of saturated aldehydes
and ketones in 1-butene photooxidation with visible light in
BaY. As discussed below, thermal reactions of some of the
products play a role in the formation of saturated aldehydes
and ketones.
In the case of room-temperature photooxidation of 1-butene,
the NMR data also shows the formation of acetaldehyde.
Acetaldehyde cannot be explained by any of the mechanisms
outlined in any of the schemes presented thus far. One plausible
mechanism for acetaldehyde production involves double bond
migration of 1-butene to yield 2-butene via a thermal reaction,
followed by photooxidation of 2-butene via a dioxetane mech-
anism. Additional evidence for double-bond migration is found
in the product distribution for 1-pentene photooxidation.
Photooxidation of 1-Pentene in BaY: The Effect of Alkene
Chain Length. The difference FT-IR spectra for the photooxi-
dation of 1-pentene in BaY at ambient temperatures are shown
in Figure 4. The more intense negative absorption bands of the
parent molecule are at 1625 and 1467 cm--1 and are indicated
by asterisks. The major product bands at 1700, 1688, 1669,
1650, 1613, and 1595 cm-1 in the FT-IR spectrum do not match
the spectrum of any single standard, including those of ethyl
vinyl ketone (not shown) and 2-pentenal (shown - Figure 4c)
very well. The 1H NMR results suggest that this is because there
are a number of other products being formed during the
photooxidation of 1-pentene in BaY at room temperature with
visible light. The band near 1700 cm-1 is assigned to a mixture
of saturated aldehydes and ketones that can be identified by ex
situ NMR experiments as described below.
The ex situ NMR data recorded after irradiation of 1-pentene
and oxygen in BaY with λ >400 nm light at 45 °C for 12 h
and extracted in CDCl3 are presented in Table 4. Ethyl vinyl
ketone, 2-pentenal, 3-penten-2-one, valeraldehyde, propional-
dehyde, butyraldehyde, and acetaldehyde products are all
Origin of the Band Near 1700 cm-1: Epoxide Ring-
Opening Reactions and Dioxetane Photooxidation Channels.
The origin of the band near 1700 cm-1 in the infrared spectra
is important since this band is indicative of the formation of
additional products namely saturated aldehydes and ketones. At
visible wavelengths, λ > 400 nm, the formation of saturated
aldehydes and ketones appears to be more strongly dependent
(20) Hoelderich, W. F.; Goetz, N. In Proceedings of the 9th International
Zeolite Conference I and II; von Ballmoss, R., Higgins, J. B., Treacy, M.
M. J., Eds.; Butterworth-Heinemann, Boston, MA, 1993; pp 309-317.

Photooxidation of 1-Alkenes in Zeolites
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5069
Figure 4. Difference infrared spectra before and after photolysis of
1-pentene and oxygen in BaY with (a) λ > 495 nm near room
temperature for 12 h, (b) λ > 400 nm near room temperature for 12 h.
Asterisks indicate the loss of 1-pentene. FT-IR spectrum of (c) trans-
2-pentenal in BaY.
Figure 5. (Left panel). Infrared spectrum of propylene adsorbed in
zeolites BaY, BaX, BaZSM-5, and BaBeta. (Right panel). Difference
infared spectra before and after irradiation of propylene and oxygen in
zeolites BaY, BaX, BaZSM-5, and BaBeta with λ > 400 nm near room
temperature.
1
Table 4. Assignment of H NMR Spectrum for Photooxidation of
1-Pentene and Oxygen on BaY (λ > 400 nm)
peak
positiona
(ppm)
relative
integrated
area
multiplicity
singlet
assignment
(0.5), 2-pentenal (1.0), 3-pentene-2-one (0.5)): saturated alde-
hydes (valeraldehyde, butyraldehyde, propionaldehyde, acetal-
dehyde) can be estimated to be 2:0.8. The product mixture
observed after 1-pentene photooxidation is more complex than
that observed after photooxidation of propylene or 1-butene.
The reasons for this difference will be discussed further in the
Discussion.
9.80
0.2
acetaldehyde (1H)
propionaldehyde (1H)
valeraldehyde (1H)
butyraldehyde (1H)
2-pentenal (1H)
9.77
singlet
0.6
9.52
6.87
doublet (8.1 Hz)
multiplet
1.0
1.4
2-pentenal (1H)
3-pentene-2-one (1H)
ethyl vinyl ketone (2H)
3-pentene-2-one (1H)
2-pentenal (1H)
ethyl vinyl ketone (1H)
alcohol product
Photooxidation of Propylene in Zeolites X, Y, ZSM-5, and
Beta: The Effect of the Parent Zeolite. In addition to
investigating the effect of excitation wavelength, temperature,
and alkene chain length on product selectivity, the role of the
parent zeolite in photooxidation of alkenes was also investigated.
The left panel of Figure 5 shows the infrared spectra of
propylene adsorbed in BaY, BaX, BaZSM-5, and BaBeta. The
infrared spectra of propylene in these four zeolites are very
different. When propylene is adsorbed in BaY and BaX, there
is an intense band in the infrared spectrum at 1631 cm-1 due to
the CdC stretching mode of propylene.21 When propylene is
adsorbed at room temperature in BaZSM-5 and BaBeta, the
intensity of the 1631 cm-1 band decreases substantially relative
to other absorption bands in the spectrum. In addition, absorption
bands between 1300 and 1500 cm-1 are broadened in the IR
spectra of propylene adsorbed on BaZSM-5 and BaBeta
compared to the IR spectra of propylene adsorbed in BaX and
BaY. The most intense peak in the IR spectra of propylene
adsorbed in BaZSM5 and BaBeta is at 1467 cm-1. These
changes in the propylene IR spectrum are consistent with the
polymerization of propylene upon adsorption in BaZSM-5 and
BaBeta. Polymerization of propylene at room temperature has
previously been observed in the acid zeolites, HY, HZSM-5,
and HMordenite using FT-IR spectroscopy.21,22 Polymerization
of propylene in HY, HZSM-5, and H Mordentite was detected
by a substantial decrease in the intensity of the CdC peak at
1630 cm-1. and the subsequent growth of a band due to -CH2-
6.24
multiplet
2.1
5.82
4.10
2.62
2.40
doublet (9.5 Hz)
doublet (4.6 Hz)
quartet (7.32 Hz)
multiplet (unresolved)
0.5
0.3
0.7
5.3
ethyl vinyl ketone (2H)
2-pentenal (2H)
valeraldehyde (2H)
propionaldehyde (2H)
butyraldehyde (2H)
3-pentene-2-one (2H)
acetaldehyde (3H)
acetone (3H)
2.24
2.22
2.17
singlet
doublet (2.7 Hz)
singlet
0.5
0.2
0.1
pentanone (3H)
2.14
1.92
1.62
singlet
doublet (6.8 Hz)
multiplet (unresolved)
0.5
0.7
1.9
2-3-pentene-2-one (1H)
3-pentene-2-one (3H)
valeraldehyde (2H)
butyraldehyde (2H)
valeraldehyde (2H)
impurity
1.37
1.28
1.13
multiplet (7.1 Hz)
triplet (unresolved)
multiplet
1.8
1.6
4.2
ethyl vinyl ketone (3H)
2-pentenal (3H)
0.93
multiplet
6.1
valeraldehyde (2H)
propionaldehyde (3H)
butyraldehyde (3H)
a Chemical shifts are measured relative to an internal standard of
TMS.
observed as indicated in Table 4. A weak peak at 2.17 ppm has
been assigned to a saturated ketone, most likely acetone or
2-pentanone. All other peaks expected for 2-pentanone overlap
with other reaction products, valeraldehyde and 3-penten-2-one,
and thus it is difficult to definitively assign the peak at 2.17
ppm. Using the integrated peak areas in Table 4, the relative
ratio of unsaturated carbonyl compounds (ethyl vinyl ketone
at 1468 cm-1 21,22
The formation of propylene oligomers on
.
BaZSM-5 and BaBeta is most likely due to the presence of
residual Brønsted acid sites on these zeolites. For BaBeta, these
(21) Kiricsi, I.; Forster, H.; Tasi, G. J. Mol. Catal. 1991, 65, L29-L34.
(22) Gosh, A. K.; Kydd, R. A. J. Catal. 1986, 100, 185-195.

5070 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Xiang et al.
sites result from the incomplete exchange of the ammonium
cation (NH4+) with Ba2+ as indicated by the ICP data (see
Experimental Section). Upon heating, NH4 is converted to
Brønsted acid sites and ammonia, which desorbs at high
temperature. The source of Brønsted acid sites is not as clear
for BaZSM-5 which was prepared from NaZSM-5. In a recent
study, Schiff bases were used to detect low levels of Brønsted
acidity in NaY and NaX zeolites.18 In the next section, the results
of the colorimetric test for Brønsted acid sites in zeolites will
be discussed.
°C. In the case of alkenes, the hydroperoxide fragments upon
warming to room temperature to form the corresponding
aldehyde and/or ketone products and water. The mechanism for
the photooxidation of propylene to yield acrolein is shown in
Scheme 1A. This mechanism is based on the mechanism
proposed by Frei and co-workers.5-7 A proposed reaction
mechanism for the photooxidation of 1-butene and 1-pentene
is presented in Scheme 1B and C. The mechanisms for the
photooxidation of 1-butene and 1-pentene in Scheme 1 B and
C are similar except for the resonance stabilization of the allyl
radical formed after hydrogen abstraction from the hydrocarbon‚
O2 charge-transfer state. The two resonance forms of the allyl
radical intermediate lead to the formation of two different
products.
For 1-butene, the two possible products resulting from this
scheme are methyl vinyl ketone and crotonaldehyde. In the FT-
IR spectrum after low-temperature irradiation of 1-butene in
BaY followed by warming to room temperature (shown in
Figure 2a), the spectrum is resolved, and there appears to be
one major product, methyl vinyl ketone. Methyl vinyl ketone
is the preferred product since it is formed from the more stable
secondary radical cation in Scheme 1 B. Broadening of the FT-
IR spectrum acquired after room-temperature irradiation of
1-butene with visible light suggests that a mixture of products
is present. The FT-IR spectra of methyl vinyl ketone and
crotonaldehyde are similar and therefore difficult to resolve.
However, the formation of crotonaldehyde is observed in the
proton NMR spectrum after room-temperature irradiation of
1-butene (λ > 400 nm, 18 h).
For 1-pentene photooxidation, the two possible products
resulting from Scheme 1 C are ethyl vinyl ketone and 2-pentenal.
Both products are clearly observed in the proton NMR spectra
of the extracted products. In addition, 3-pentene-2-one and
saturated aldehydes such as valeraldehyde, butyraldehyde,
propionaldehyde, and acetaldehyde are identified in the proton
NMR spectrum.
The most likely mechanism for the formation of 3-pentene-
2-one is through the isomerization of 1-pentene to 2-pentene in
the zeolite at room temperature, followed by photooxidation of
2-pentene to ethyl vinyl ketone and 3-pentene-2-one as depicted
in Scheme 4. The isomerization of alkenes in zeolites has been
previously documented.23,24 There is evidence of 1-butene to
2-butene isomerization and 1-pentene to 2-pentene isomerization
in some of the FT-IR and NMR experiments (not shown). The
formation of small quantities of acetaldehyde from 2-butene
photooxidation and acetaldehyde and propionaldehyde from
2-pentene oxidation also suggests that double-bond migration
occurs in BaY, followed by dioxetane chemistry.
+
Photooxidation following propylene adsorption in BaY, BaX,
BaZSM-5, and BaBeta was then examined. The difference
spectra following irradiation with broadband visible light (λ >
400 nm) at 318 K are shown in the right panel of Figure 5.
Irradiation of propylene in BaY and BaX yielded primarily
acrolein (1668 and 1366 cm-1) at this wavelength with small
amounts of saturated aldehydes and ketones (1702 cm-1). New
photoproduct bands are observed from adsorbed propylene in
BaZSM-5 and BaBeta upon irradiation with visible light. These
bands, however, are much broader and blue-shifted by ∼12 cm-1
from the authentic acrolein infrared spectrum in the correspond-
ing zeolite. We attribute these bands to the products from
photooxidation of propylene oligomers and the remaining
propylene. A broad band at 1683 cm-1 is observed in the
spectrum that can be attributed to a carbonyl stretch of the
oxidized species. In a previous FT-IR study, thermal reactions
of propylene oligomers on acidic zeolites yielded IR peaks at
1590 cm-1 (HY) and 1510 cm-1 (HZSM-5 and HMordenite),
attributed to coke and aromatics, respectively.22 Intensity in the
1500 cm-1 region of FT-IR spectrum of BaBeta suggests that
aromatic products could be formed and potentially be oxidized
during irradiation.
Colorimetric Test for Brønsted Acid Sites in Na and Ba-
Exchanged Zeolites X, Y, ZSM-5, and Beta. To clearly
determine that Brønsted acid sites are present in the zeolites
used in this study, the colorimetric test described by Rama-
murthy and co-workers was utilized.18 The colorimetric test is
based on differences in the electronic absorption properties of
neutral and cationic forms of probe molecules, such as retinol
and retinol acetate. Both retinol and retinol acetate form a blue
retinyl cation in an acid solution.18 The zeolites were actiVated
by heating to 300 °C under vacuum for approximately 12 h.
Samples of BaZSM-5 and BaBeta both turned dark blue when
a dilute solution of retinol was added to the activated zeolite,
indicating the presence of Brønsted acid sites. Activated BaY
zeolite turned light blue when retinol in hexane was added to
the sample and activated samples of BaX, NaX, and NaY
exhibited no color change when a dilute solution of retinol was
added to the zeolite. These results suggest that, for the zeolites
used in this study, the number and strength of Brønsted acid
sites trend in the following way: BaZSM-5, BaBeta > BaY >
BaX, NaY, NaX. This is qualitatively consistent with the
observed oligomerization of propylene which is proposed to
occur in BaZSM-5 and BaBeta, but not in BaY and BaX.
Schemes 1 and 4 account for all of the unsaturated aldehyde
and ketone photoproducts formed from the photooxidation of
1-alkenes. However, epoxides are observed in the ex situ NMR
experiments after photooxidation of 1-alkenes. The epoxides
are proposed to be formed from the thermal reaction of the
hydroperoxide intermediate with parent hydrocarbon as shown
in Scheme 2.4-7,9 In the case of propylene, the other product,
allyl alcohol, should also be formed. It may be that the alcohol
product is not observed because of its high volatility. The
samples for NMR analysis were exposed to air during the
extraction process, and some of the more volatile products were
most likely lost. In general, the alcohols appear to be difficult
to observe in proton NMR experiments due to hydrogen bonding
Discussion
Reaction Mechanism of 1-Alkene Photooxidation and
Product Selectivity in Zeolite Y. Frei and co-workers previ-
ously suggested that the photooxidation of hydrocarbons in
cation-exchanged zeolite Y proceeds through an intermolecular
hydrocarbon‚O2 charge-transfer state.5-7 The second step in the
proposed mechanism involves tranfer of a proton to O2- to form
an allyl/HO2• radical pair that reacts further to yield the selective
formation of a hydroperoxide at low temperatures near -100
(23) Kondo, J. N.; Liqun, S.; Wakabayashi, F.; Domen, K. Catal. Lett.
1997, 47, 129-133.
(24) Alvarez-Idaboy, J. R.; Eriksson, L. A.; Lunell, S. J. Phys. Chem
1993, 97, 12742-12744.

Photooxidation of 1-Alkenes in Zeolites
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5071
Scheme 4. Proposed Mechanism for the Formation of
3-Pentene-2-one
oxygen charge-transfer complex, the selectivity to the unsatur-
ated aldehyde or ketone is maximized, but the rate of product
formation is substantially decreased.
In addition, as the temperature is increased at a given
wavelength (Figure 3a and b), the amount of unsaturated
aldehyde and ketone being formed increased. In fact, all of the
observed chemistry once the initial hydroperoxide is formed is
thermal chemistry including the rearrangement of the hydrop-
eroxide to the corresponding unsaturated ketone and aldehyde
and epoxide formation from the reaction of the hydroperoxide
and an unreacted parent molecule. Thermal chemistry also
contributes to the formation of nonselective, short-chain satu-
rated aldehydes and ketones when the photooxidation experi-
ments are conducted at room temperature. As shown here, a
thermal reaction involving ring opening of the epoxide yields
both saturated aldehydes and saturated ketones on BaY. In
addition, a thermal reaction involving double-bond migration
followed by photooxidation of the 2-alkene gives rise to other
unsaturated carbonyls (e.g., 3-pentene-2-one) and saturated
carbonyls (e.g., acetaldehyde). However, it is important to point
out that although there are other products in the reaction, there
is a rather high selectivity for the formation of unsaturated
carbonyl compounds. It is seen for all of these compounds that
greater than 60% of the reacted parent alkene goes on to form
an unsaturated carbonyl compound and that the selectivity
increases as the carbon chain length decreases.
The Effect of Prereaction in Zeolites ZSM-5 and Beta on
the Photooxidation Product. Our efforts to exploit the shape-
selective properties of zeolites for the photooxidation of
hydrocarbons in zeolites have been complicated by prereaction
of alkenes to form polymers. The polymerization of alkenes
has been extensively studied on the acid form of zeolites, such
as HZSM-5.21,22,25,26 Prereaction in ZSM-5 and Beta involves
polymerization of the parent alkene which most likely occurs
on acid sites.
effects which cause spectral broadening. For 1-butene photo-
oxidation, evidence of the alcohol and the epoxide products has
been observed in the ex situ NMR experiments. For 1-pentene,
the alcohol product is observed in the NMR spectrum; however,
the epoxide is not. Instead a substantial amount of saturated
aldehydes are detected. As discussed here, epoxides can react
on BaY to form saturated aldehydes and ketones. Propylene
oxide and epoxybutane both react on BaY to form the saturated
aldehydes, propionaldehyde, and butyraldehyde and the saturated
ketones, acetone and 2-butanone, respectively. This suggests
that the presence of saturated aldehydes and ketones after
photooxidation of 1-alkenes with visible light at room temper-
ature is due in part to epoxide ring-opening reactions. The
rearrangement of aliphatic epoxides in zeolites has been
observed previously.20
Effect of Wavelength and Temperature on Product
Selectivity. The wavelength dependence of the photooxidation
of propylene in BaY was examined. It was found that the rate
of photoproduct formation increased as the wavelength of light
decreased. In addition, the results suggested that the formation
of nonselective, short-chain saturated aldehyde and ketone
products (acetaldehyde, formaldehyde, and propionaldehyde)
increased when short wavelength light (λ > 285 nm) was used
for photolysis of propylene and oxygen in BaY. The wavelength
dependence of the formation of saturated aldehydes is most
likely due to the increase in dioxetane chemistry (Scheme 2)
when the shorter wavelengths of light were used for photolysis.
Similar effects were observed in corresponding experiments of
1-butene and 1-pentene photooxidation in BaY. Frei and co-
workers previously observed that when 2,3-dimethyl-2-butene
(DMB) and oxygen were photooxidized in BaY at 223 K,
hydroperoxide formation increased and the formation of acetone
(by dioxetane chemistry) increased as the wavelength of
photolysis decreased from 633 to 458 nm.5 By using the longest
possible wavelength of light that will form the hydrocarbon-
The source of Brønsted acid sites in BaZSM-5 is due to the
following reaction that occurs during pretreatment or activation
of a divalent cation-exchanged zeolite at elevated temperature:
M
2+ + H O + Si-O-Al f M(OH)+ + Si-O(H+)-Al
(1)
2
where Si-O-Al represents a part of the zeolite framework.27
It has been shown that Brønsted acid sites can be present in
low concentrations in alkali-metal zeolites. For example,
Jayathurthe et al. have established the presence of low levels
of Brønsted acidity in NaY and NaX zeolites using the color
change of a base indicator.18 The cationic forms of retinyl
acetate, retinol, and retinyl Schiff bases are colored and easily
identified by both visual inspection and spectrophotometrically.
Using these bases they were able to determine whether Brønsted
acids were present in low concentrations and the relative
differences in acid strength for the Brønsted acid sites in the
different zeolites. Importantly, these studies showed that the
presence of small quantities of acid sites can alter the reactivity
of various alkenic substrates. In another study using solid-state
NMR spectroscopy, the presence of Brønsted acid sites, as well
as other types of sites, were identified in the alkaline earth zeolite
CaY.27 The presence and concentration of sites was found to
depend on whether the zeolite was activated in air or under
(25) Forster, H.; Kiricsi, I.; Seebode, J. Stud. Surf. Sci. Catal. 1988, 37,
435-442.
(26) Kirisci, I.; Forster, H. J. Chem. Soc., Faraday Trans. 1 1988, 84,
491-499.
(27) Kao, H.-M.; Grey, C. P.; Pitchumani, K.; Lakshminarasimhan, P.
H.; Ramamurthy, V. J. Phys. Chem. A 1998, 102, 5627-5638.

5072 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Xiang et al.
vacuum. These studies indicate that the reactivity and properties
of cation-exchanged zeolites can change from source to source
and from different pretreatment conditions. The zeolite samples
used in this study contained various quantities of Brønsted acid
sites. Using the colorimetric method with retinol as a probe
molecule, the activated forms BaZSM-5 and BaBeta were both
found to contain Brønsted acid sites. BaY contained much fewer
Brønsted acid sites than BaZSM-5 and BaBeta, and no Brønsted
acid sites were detected colorimetrically in BaX. These results
are qualitatively consistent with the observed oligomerization
of propylene which occurs on BaZSM-5 and BaBeta, but not
on BaY or BaX.
Photooxidation of the propylene oligomers and unreacted
propylene in BaZSM-5 and BaBeta results in broad, unresolved
FT-IR spectra, particularly in the 1550-1700 cm-1 range. The
observed broadening in the FT-IR spectra suggests a mixture
of oxygenated products. Loss of propylene was also observed
as indicated by the peaks with asterisks.
Hydrocarbon Photooxidation Reactions in Cation-Ex-
changed Zeolites: A Comparison to Other Studies. Recent
studies using zeolites as hosts for photooxidation reactions of
hydrocarbons show that selectivity is significantly enhanced,
compared to solution studies. Ramamurthy and co-workers have
shown remarkable selectivity in the formation of hydroperoxides
from alkenic systems through a singlet oxygen mechanism using
dye-exchanged zeolites in hexane solution.28,29 However, in
terms of the challenges for industrial oxidation processes
outlined in the Introduction, the use of hexane solvent is nonideal
for environmentally benign synthesis. The approach developed
by Frei and co-workers using gas-phase reactants and a host
zeolite is very promising in terms of minimal environmental
impact. However, most of the reactions have been done at low
temperatures near -100 °C, not very practical from an industrial
standpoint. As shown here, when reactions are run near -20
°C, they remain fairly selective; however, thermal reactions near
ambient temperatures can lead to loss of selectivity.30
zeolites is the wavelength dependence. Higher energy UV
photons open up reaction pathways, e.g., dioxetane chemistry,
that do not occur with lower energy visible photons. However,
as shown here, selectivity can also be lost when visible photons
are used. Although these reactions are initiated with light,
thermal reactions at ambient temperatures can result in a loss
of product selectivity. At room temperature, mixtures of
aldehydes, ketones, epoxides, and alcohols were observed by
FT-IR and NMR spectroscopy after visible excitation of
hydrocarbon‚O2 complexes with visible light. Studies of the
photooxidation of propylene in other zeolites, BaX, BaZSM-5,
and BaBeta, were also investigated to determine if the meth-
odology was of general use and if shape selectivity through the
use of zeolites ZSM-5 and Beta on the reactant and product
molecules could be imposed. Prereaction on residual acid sites
in the zeolite leads to polymerization of the alkene. These results
illustrate that the selectivity of the overall oxidation process is
controlled by a number of factors including thermal chemistry
at ambient temperatures. Several room temperature reactions
of reactants and products in Ba-exchanged Y, X, ZSM-5, and
Beta give rise to the loss of selectivity. These include: (i)
epoxide ring opening, (ii) alkene double bond migration, and
(iii) alkene polymerization. Therefore, if 100% selectivity for
the formation of a single oxygenated hydrocarbon product is to
be realized through the photooxidation reaction of a parent
hydrocarbon molecule in cation-exchanged zeolites, a careful
choice of chemical systems, such as those that are nonalkenic,
is necessary. In addition, the choice of zeolite host may also be
important in these systems. Further studies are currently
underway in developing and extending this methodology in
environmentally benign synthesis of partial oxidation products.31
Acknowledgment. Although the research described in this
article has been funded wholly or in part by the Environmental
Protection Agency through Grant No. R825304-01-0 to S.C.L.
and V.H.G., it has not been subjected to the Agency’s required
peer and policy review and therefore does not necessarily reflect
the views of the Agency and no official endorsement should
be inferred. The authors would like to thank Professor John
Wiencek for use of an ICP-AES instrument and Dr. Alexander
Panov for the colorimetric measurements of acid sites in the
zeolites used in this study. Professor Nancy I. Totah is
acknowledged for helpful discussions.
Conclusions
The photooxidation of a homologous series of 1-alkenes in
zeolites was investigated by using in situ FT-IR spectroscopy
and ex situ NMR spectroscopy. Propylene, 1-butene, and
1-pentene were photooxidized in BaY with excess oxygen. One
important factor in the photooxidation of hydrocarbons in
Supporting Information Available: 1H NMR spectra
showing the photoproducts extracted in CDCl3 after the pho-
tooxidation of 1-butene and 1-pentene in BaY (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(28) Robbins, R. J.; Ramamurthy, V. Chem. Commun. 1997, 1071-1072.
(29) Li, X.; Ramamurthy, V. J. Am. Chem. Soc. 1996, 118, 10666-
10667.
(30) Blatter et al. have also observed a loss of selectivity at room
temperature during photooxidation of propylene in BaY at room temperature
(Figure 2 of ref 7). They observed an FT-IR band at ∼1700 cm-1 that they
attributed to formaldehyde and acetaldehyde. In addition, in ref 5 during
the photooxidation of 2-butene, Blatter and Frei mention the appearance of
a band at 1720 cm-1. The band is attributed to acetaldehyde, and the growth
of the band was not reproducible.
JA984477A
(31) Panov, A.; Larsen, S. C. and Grassian, V. H., unpublished results.