Selective Solid State Photooxidant

Article abstract of DOI:10.1021/ja0368406

Irradiation of biphenyl encapsulated in the cavities of a NaZSM-5 zeolite framework has been reported to result in the formation of an extremely long-lived radical cation. Here, we show that such zeolite encapsulated radical cations can act as irreversible one-electron oxidants for simple alkenes and dienes, in a solid-state analogue to solution-phase cosensitization. Compared to the well-known semiconductor photooxidizers, such as titanium dioxide, the NaZSM-5 zeolite-based solid photooxidants exhibit enhanced selectivity based on oxidation potential, molecular size and shape, and Lewis base character.

Full text of DOI:10.1021/ja0368406

Published on Web 11/12/2003
Selective Solid State Photooxidant
Tracy L. Morkin,† Nicholas J. Turro,*,† Mark H. Kleinman,† Cheyenne S. Brindle,†
Wolfgang H. Kramer,‡ and Ian R. Gould*,‡
Contribution from the Department of Chemistry, Columbia UniVersity, 3000 Broadway,
New York, New York 10027 and Department of Chemistry and Biochemistry,
Arizona State UniVersity, Tempe, Arizona 85287-1604
Received June 23, 2003; E-mail: njt3@columbia.edu
Abstract: Irradiation of biphenyl encapsulated in the cavities of a NaZSM-5 zeolite framework has been
reported to result in the formation of an extremely long-lived radical cation. Here, we show that such zeolite
encapsulated radical cations can act as irreversible one-electron oxidants for simple alkenes and dienes,
in a solid-state analogue to solution-phase cosensitization. Compared to the well-known semiconductor
photooxidizers, such as titanium dioxide, the NaZSM-5 zeolite-based solid photooxidants exhibit enhanced
selectivity based on oxidation potential, molecular size and shape, and Lewis base character.
Introduction
is illustrated in Scheme 1. Excitation of an electron acceptor
(A) results in electron transfer from the cosensitizer (C) to form
Photoinduced one-electron-transfer reactions represent the
primary step in a wide variety of important natural and tech-
nological processes. Historically, photosynthesis, artificial solar
energy conversion, and various imaging processes have been
among the more obvious examples of these processes.1 More
recently, one-electron oxidation processes involving semicon-
ductor substrates have found application in the remediation of
organic waste materials.2 Excitation of the semiconductor,
commonly TiO2, results in the formation of a valence hole and
conduction band electron that can initiate redox reactions leading
to destruction of organic materials. In many cases, complete
oxidation to carbon dioxide can be achieved.2 Oxidation of the
organic material can be initiated either by the generation of
oxidizing species such as hydroxyl radicals, or by direct one-
electron oxidation of the absorbed organic compound D to form
its radical cation.2 The radical cation reaction products can be
further oxidized by this, or other mechanisms. In common with
all photoinduced electron-transfer processes,1,3 the utility of these
semiconductor-based oxidation processes are controlled to a
large extent by the efficiency of trapping of the initially formed
excited state (in this case, a valence band hole), and energy
wasting return electron transfer (in this case, from the conduction
band to the valence band) to reform the starting species.2b
In homogeneous solution, these problems have been circum-
vented to some extent by the use of cosensitizers.3,4 The concept
a geminate pair consisting of the acceptor radical anion and
the cosensitizer radical cation (A•-/C•+). The cosensitizer is
chosen so that separation in the geminate pair is much faster
than return electron transfer (i.e., ksep is much larger than k-et
,
Scheme 1). In this way, the efficiency of formation of separated
radical ions (A•- + C•+) is very high.3 The separated C•+ can
then oxidize the donor to form D•+ in a second step, kox. The
overall efficiency for formation of D•+ is high because formation
of separated C•+ occurs with high efficiency, and the follow-
up oxidation process of D is irreversible. Furthermore, the
concentration of the D can be low because it does not have to
react with a short-lived excited state, A*, but with the relatively
long-lived C•+.3,5 An ideal cosensitizer C has a high oxidation
potential, both to ensure a small k-et as a consequence of the
Marcus inverted region effect, and also so that the C•+ can
oxidize as wide a range of D molecules as possible,3 and
importantly, a long lifetime. The most widely used cosensitizer
in solution has been biphenyl.3,4 With 9,10-dicyanoanthracene
as the electron acceptor, biphenyl has one of the highest yields
of separated radical ions in polar solution yet measured.3
Biphenyl also has a relatively high oxidation potential, ca. 2 V
vs SCE.3 Because the biphenyl radical cation is also relatively
long-lived in solution, even endothermic oxidations of donor
molecules in solution have been observed.5 An obvious question
is whether the solution phase cosensitization concept can be
extended to a solid-state photooxidation system. Consequently,
our attention was drawn to reports of long-lived photogenerated
biphenyl radical cations in zeolites.6 We wondered whether these
† Department of Chemistry, Columbia University.
‡ Department of Chemistry and Biochemistry, Arizona State University.
(1) See, for example: (a) Electron Transfer in Chemistry, Vols. 1-5; Balzani,
V., Ed.; Wiley-VCH: New York, 2001. (b) AdVances in Electron-Transfer
Chemistry; Mariano, P. S., Ed.; Jai Press: Greenwich, CT, 1991-1994,
Vols. 1-4. (c) Photoinduced Electron Transfer; Fox, M. A., Chanon, M.,
Eds.; Elsevier: Amsterdam, 1988. (d) Kavarnos, G. J.; Turro, N. J. Chem.
ReV. 1986, 86, 401.
(2) (a) Photocatalytic Purification and Treatment of Water and Air; Ollis, D.
F.; Al-Ekabi, H., Eds.: Elsevier: New York, 1993. (b) Thurnauer, M. C.;
Rajh, T.; Dimitrijevic, N. M. In Electron Transfer in Chemistry; Balzani,
V., Ed.; Wiley-VCH: New York, 2001; Vol. 5, p 695.
(3) (a) Gould, I. R.; Ege, D.; Moser, J. E.; Farid, S. J. Am. Chem. Soc. 1990,
112, 4290. (b) Gould, I. R.; Farid, S. Acc. Chem. Res. 1996, 29, 522.
(4) Schaap, A. P.; Siddiqui, S.; Prasqad, G.; Palomino, E.; Lopez, L. J.
Photochem. 1984, 25, 167. (b) Arnold, D. R.; Snow, M. S.; Can. J. Chem.
1988, 66, 3012. (c) Majima, T.; Pac. C.; Nakasone, A.; Sakurai, H. J. Am.
Chem. Soc. 1981, 103, 4499.
(5) Mattes, S. L.; Farid, S. In Organic Photochemistry; Padwa, A., Ed.; Marcel
Dekker: New York, 1983; Vol. 6, p 233.
9
10.1021/ja0368406 CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 14917-14924
14917

A R T I C L E S
Morkin et al.
Scheme 1. Mechanism of Solution-Phase Cosensitization
species could be used to oxidize electron donor compounds in
an analogous manner to solution phase cosensitization.7
Previous work on related systems suggested that this should
be possible. Photoinduced reactions in zeolites, including
bimolecular processes, are well-known,8 and many studies of
bimolecular electron transfer reactions involving excited state
donors and acceptors have been reported.9 Studies of radiation
induced and thermal processes have demonstrated intermolecular
hole transfer and electron transfer within zeolite cages,10
including an example demonstrating size selectivity,10c and
another involving electron transfer to biphenyl radical cation
from a coadsorbed amine.10f In this and other work on organic
radical ions in zeolites,11 however, no systematic study of
secondary oxidation processes corresponding to solid-state
cosensitization has been reported.
Figure 1. Apparatus for EPR study of radical cations in zeolites. S,
stopcock; V, vacuum joint. The biphenyl and zeolite are heated, then
irradiated in bulb B and transferred to EPR tube E for analysis. The donor
is subsequently introduced from bulb D and any reaction products are
analyzed again in E.
complete oxidation of organics may be desirable in some
applications, control over the oxidizing power of a solid-state
photooxidant may allow the development of selective partial
oxidizing agents.12 The use of different organic radical cations
would allow much finer control over oxidizing power than that
available via manipulation of a semiconductor band levels.2b
Furthermore, by taking advantage of the known size and shape
selectivity of zeolites,8,13 we hoped to gain further control and
selectivity on the oxidation process beyond that possible in
solution or on the surface of a semiconductor particle.2
The present work sets out to address a variety of issues.
Although our primary interest is in extending the cosensitization
concept from solution to the solid state, the ultimate goal is to
explore the utility of zeolite-encapsulated organic radical cations
as solid-state photoactivatable oxidizing substrates. Although
(6) (a) Gener, I.; Buntinx, G.; Moissette, A.; Bremard, C. J. Phys. Chem. B
2002, 106, 10 322. (b) Gener, I.; Moissette, A.; Vezin, H.; Patarin, J.;
Bremard, C. Stud. Surf. Sci. Catal. 2001, 38, 1819. (c) Bremard, C.; Buntinx,
G.; De Waele, V.; Didierjean, C.; Gener, I.; Poizat, O. J. Mol. Struct. 1999,
480-481, 69. (d) Gener, I.; Buntinx, G.; Bremard, C. Angew. Int. Ed. Engl.
1999, 38, 1819-1822. (e) Gener, I.; Buntinx, G.; Bremard, C. Micropor.
Mesopor. Mat. 2000, 41, 253-268. (f) Erickson, R.; Lund, A.; Lindgren,
M. Chem. Phys. 1995, 193, 89. (g) Mao, Y.; Zhang, G.; Thomas, J. K.
Langmuir 1993, 9, 1299-1305.
In this paper, we demonstrate all of these. Biphenyl radical
cation has been photochemically generated in sodium cation
exchanged ZSM-5 zeolites, and selective irreversible oxidation
of a series of simple alkenes and dienes as electron donors
demonstrated. Fine control of the oxidation based on size, shape
and the Lewis basicity of the alkene is demonstrated. The
oxidation process is much more selective than the corresponding
process in homogeneous solution.
(7) For other approaches to photooxidative processes in zeolites, see: Vasenkov,
S.; Frei, H. Mol. Supramol. Photochem. 2000, 5, 295.
(8) See, for example: (a) Ramamurthy, V.; Robbins, R. J.; Thomas, J. K.;
Lakshminarasimhan, P. H. In Organized Molecular Assemblies in the Solid
State; Wiley: New York, 1999, p. 62. (b) Scaiano, J. C.; Garcia, H. Acc.
Chem. Res. 1999, 32, 783. (c) Turro, N. J. Acc. Chem. Res. 2000, 33, 637.
(d) Ramamurthy, V. J. Photochem. Photobiol., C: 2000, 1, 145. (e) Thomas,
J. K. International Journal of Photoenergy 2002, 4, 27. (f) Tung, C.-H.;
Wu, L.-Z.; Zhang, L.-P.; Chen, B. Acc. Chem. Res. 2003, 36, 39. (g)
Hashimoto, S. J. Photochem. Photobiol. C 2003, 4, 19.
Experimental Approach and Selection of Electron
Donors
(9) See, for example: (a) Yoon, K. B. Chem. ReV. 1993, 93, 321. (b) Thomas,
J. K. Chem. ReV. 1993, 93, 301. (b) Liu, X.; Mao, Y.; Ruetten, S. A.;
Thomas, J. K. Solar Energy Materials and Solar Cells 1995, 38, 199. (c)
Dutta, P. K.; Ledney, M. Prog Inorg. Chem. 1997, 44, 209. (d) Zhang, G.;
Liu, X.; Thomas, J. K. Radiat. Phys. Chem. 1998, 51, 135. (e) Yoon, K.
B. Mol. Supramol. Photochem. 2000, 5, 143. (f) Garcia, H.; Roth, H. D.
Chem. ReV. 2002, 102, 3947. (g) Vaidyalingam, A. S.; Coutant, M. A.;
Dutta, P. K. In Electron Transfer in Chemistry; Balzani, V., Ed.; Wiley-
VCH: New York, 2001; Vol. 4, p 412. (h) O’Neill, M. A.; Cozens, F. L.;
Schepp, N. P. J. Phys. Chem. B 2001, 105, 12 746.
(10) (a) Crockett, R.; Roduner, E. Stud. Surf. Sci. Catal. 1994, 84, 527. (b)
Barnabas, M. V.; Trifunac, A. D. Chem. Phys. Lett. 1992, 193, 298. (c)
Yoon, K. B.; Park, Y. S.; Kochi, J. K. J. Am. Chem. Soc. 1996, 118, 12 710.
(d) Werst, D. W.; Han, P.; Trifunac, A. D. Chem. Phys. Lett. 1997, 269,
333. (e) Vitale, M.; Castagnola, N. B.; Ortins, N. J.; Brooke, J. A.;
Vaidyalingam, A.; Dutta, P. K. J. Phys. Chem. B 1999, 103, 2408. (f) Zhang,
G.; Thomas, J. K. J. Phys. Chem. B 2003, ASAP.
To confirm oxidation of the donors in NaZSM-5, we need to
first observe formation of the BP•+ in the zeolite, demonstrate
reaction of the donor molecules with the BP•+, and confirm
that the product of this reaction is the radical cation of the donor,
D•+. EPR spectroscopy was chosen as the preferred method for
detection of the radical cations in the solid state. An EPR cell
was designed (Figure 1) to allow the study of photochemical
(12) (a) Zaera, F. J. Phys. Chem. B 2002, 106, 4043. (b) Rafelt, J. S.; Clark, J.
H. Catalysis Today 2000, 57, 33. (c) Heller, A. Acc. Chem. Res. 1995, 28,
503. (d) Pichat, P. Catal. Today 1994, 19, 313. (e) Zhan, B.-Z.; White, M.
A.; Sham, T.-K.; Pincock, J. A.; Doucet, R. J.; Ramana, K. V.; Robertson,
K. N.; Cameron, T. S. J. Am. Chem. Soc. 2003, 125, 2195.
(13) (a) Dwyer, F. G. Stud. Surf. Sci. Catal. 1991, 67, 179. (b) Corbin, D. R.;
Herron, N. J. Mol. Catalysis 1994, 86, 343. (c) De Vos, D. E.; Jacobs, P.
A. Catal. Sci. Series 2002, 3, 261.
(11) (a) Werst, D. W.; Han, P.; Trifunac, A. D. Radiat. Phys. Chem. 1998, 51,
255. (b) Park, Y. S.; Lee, K.; Lee, C.; Yoon, K. B. Langmuir, 2000, 16,
4470. (c) Folgado, J.-V.; Garcia, H.; Marti, V.; Espla, M. Tetrahedron,
1997, 53, 4947. (d) Erikson, R.; Benetis, N. P.; Lund, A.; Lindgren, M. J.
Phys. Chem. A 1997, 101, 2390. (e) Hashimoto, S. J. Photochem. Photobiol.
C 2003, 4, 19-49.
9
14918 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003

Selective Solid State Photooxidant
A R T I C L E S
Table 1. Oxidation Potentials, Reaction Free Energies, Rate
Constants, and Reactivity in Zeolites for Oxidation of Electron
Donors with Biphenyl Radical Cation in Solution and in Na-ZSM-5
Zeolites
Scheme 2. Structures and Abbreviations for Cosensitizer Biphenyl
(BP) and Electron Donors: 2-Methylpentane (2-MP),
Tetramethylethylene (TME), 2,3-Dimethyl-2-pentene (2,3-DMP),
2,4-Hexadiene (HD), 2,4-Dimethyl-2,4-hexadiene (DMHD),
cis,cis-1,3-Cyclooctadiene (COD) and 2,3-Dihydrofuran (2,3-DHF)
dimensionsb
(Å)
E
∆G c
(eV)
kox
(M-1 s-1
)
quenching in
Na-ZSM-5d
D
ox
et
donora
2-MP
COD
HD
(V vs SCE)
4.4 × 6.4
6.1 × 6.4
4.4 × 8.4
5.4 × 5.8
5.9 × 5.9
3.8 × 4.0
5.1 × 8.7
1.9816a
1.9016b
1.8216a
1.5616a
1.35e
0.02 4.0 × 108
-0.06 5.5 × 109
-0.14 9.5 × 109
-0.40 7.5 × 109
-0.61 5.8 × 109
-0.70 1.0 × 1010
-0.63 1.5 × 1010
no
no
yes
yes
no
TME
2,3-DMP
2,3-DHF
DMHD
1.26f
no
yes
1.3316a
a See Scheme 2 for structures. b Cylindrical dimensions of the molecule,
estimated as described in the text and the Experimental Section. The smaller
dimension represents the diameter of the cylinder, the second, its length.
The corresponding dimensions for biphenyl are 5.4 × 10.4 Å. c Calculated
C
using eq 1, with a value of 1.96 V vs SCE3a for Eox
.
d Indicates whether
quenching of the biphenyl radical cation EPR signal is observed upon
addition to BP•+@Na-ZSM-5. e Estimated from IP16c according to correla-
tion in ref 16e. f Estimated from IP16d according to correlation in ref 16e.
such as biphenyl. The decay is presumably due to oxidation of
low-oxidation potential impurities, or perhaps attack by adventi-
tious nucleophiles such as water.5
The oxidation process in solution can be characterized by
the quantum efficiency of formation of the separated C•+
(Scheme 1), and the bimolecular rate constant for the secondary
oxidation step, kox. Under the experimental conditions, the
quantum yield for formation of separated C•+ has a value of
ca. 0.7 and obviously does not vary with the alkene donor, D.3a
Summarized in Table 1 are the bimolecular rate constants
measured in acetonitrile solvent for the various donors, kox,
together with their oxidation potentials, EoxD. Bimolecular
electron transfer between the C•+ and the D should be close to
diffusion controlled for reactions that are more exothermic than
the reaction reorganization energy.14 The thermodynamics of
the oxidation reactions, ∆Get, are given by the difference
between the oxidation potential of the electron donor (EoxD) and
that of the biphenyl cosensitizer (EoxC), eq 1. Reactions that
are more exothermic than ca. 0.5 eV (i.e., -∆Get > 0.5 eV)
are usually close to diffusion controlled, although exothermici-
ties close to 1 eV may be required to actually reach this limit.14,15
However, thermodynamically meaningful oxidation
formation of the BP•+ first, and then the effects of subsequent
addition of a potential electron donor.
In a typical experiment, biphenyl is loaded into the zeolite
under vacuum (see the Experimental Section for details) and
BP@NaZSM-5 is irradiated. The irradiated zeolite mixture is
transferred into the EPR tube part of the apparatus and the
spectrum is recorded. The apparatus is removed from the EPR
spectrometer, and the donor is allowed to mix with the
BP@NaZSM-5. The tube is then returned to the EPR spec-
trometer to look for any possible reactions.
For secondary oxidation in NaZSM-5 zeolites, electron donors
that would be sufficiently volatile to be loaded into the zeolites
using the experimental apparatus were required. To investigate
the possible factors controlling oxidation, ready manipulation
of the donor structure was required. A series of small unsaturated
electron donors with a maximum of eight carbons, that include
a range of oxidation potentials, shapes, and atomic composition
were chosen, i.e., 2-methyl-2-pentene (2-MP), tetramethyleth-
ylene (TME), 2,3-dimethyl-2-pentene (2,3-DMP), 2,4-hexadiene
(HD), 2,5-dimethyl-2,4-hexadiene (DMHD), cis,cis-1,3-cyclooc-
tadiene (COD), and 2,3-dihydrofuran (DHF). Their structures
are summarized in Scheme 2. We have also measured the
solution-phase rate constants for one-electron oxidation of these
donors with biphenyl radical cation using laser flash photolysis,
to compare the results with those we obtained in the zeolite
environment.
C
∆Get ) EoxD - Eox
(1)
potentials are difficult to obtain (the errors in the potentials given
in Table 1 are not known), because electrochemical oxidation
is usually not reversible for simple alkenes such as those in
Scheme 2 due to fast follow-up reactions on the electrochemical
time scale. Furthermore, the solvent reorganization is not well-
defined for diffusive bimolecular reactions in solution, and
indeed may vary within a series of reactions depending upon
the exothermicity.15 Nevertheless, the expected behavior is
roughly observed for the present data. The oxidation reaction
of the trisubstituted alkene 2-MP is slightly endothermic, its
-∆Get is significantly smaller than the solution reorganization
energy, and its rate constant significantly smaller than diffusion
Results
Solution Phase Oxidation. Pulsed laser excitation of 9,10-
dicyanoanthracene in acetonitrile in the presence of biphenyl
resulted in the observation of absorptions due to separated
DCA•- and BP•+, as described previously.3 In the presence of
oxygen, the DCA•- undergoes rapid pseudo-first-order decay
due to electron transfer to oxygen. The decay of the BP•+ is
unaffected by added oxygen, and decays by mainly pseudo-
first-order processes under the experimental conditions with a
lifetime of ca. 1.5 µs. Such behavior is quite common for radical
cations of compounds with relatively high oxidation potential,
(14) (a) Marcus, R, A. J. Chem. Phys. 1957, 26, 872. (b) Marcus, R. A. Annu.
ReV. Phys. Chem. 1964, 15, 155. (c) Rehm, D.; Weller, A. Isr. J. Chem.
1970, 8, 259.
(15) Gould, I. R.; Young, R. H.; Mueller, L. J.; Farid, S. J. Am. Chem. Soc.
1994, 116, 8176.
9
J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003 14919

A R T I C L E S
Morkin et al.
Figure 3. EPR spectrum observed upon addition of 2,4-dimethyl-2,4-
hexadiene to biphenyl radical cation in Na-ZSM-5 zeolite.
confirmed complete quenching of the BP•+ in each case. With
DMHD, the BP•+ signal was replaced by a new EPR spectrum
similar to that previously attributed to the DMHD•+ on
H-mordenite, Figure 3.17 This spectrum is stable for at least 24
h. However, control experiments showed that the same spectrum
is also observed upon loading the DMHD into NaZSM-5 in
the absence of BP•+ and without irradiation. With HD, no new
paramagnetic signals are observed after addition of the diene.
The HD obviously reacts with the BP•+, but in this case, any
HD•+ that is formed must have a lifetime of less than 60 s,
otherwise we would have expected to observe it under the
experimental conditions. It is possible that significant broadening
of the signal could occur in the presence of the added donor
such that the radical cation might not be detected. However,
the disappearance of the characteristic green color is further
evidence to support the finding that the radical cation has indeed
been quenched.
Corresponding addition of the alkene 2-MP to the BP•+@
NaZSM-5 did not quench the color, or reduce the intensity of
the BP•+ EPR spectrum, indicating that this electron donor does
not react with the BP•+. Interestingly, no significant decrease
in the BP•+ EPR intensity is observed even 24 h after addition
of this donor, i.e., any reaction is at least 3 orders of magnitude
slower than the corresponding reactions for HD and DMHD.
Oxidation of 2-MP by BP•+ is essentially thermoneutral and
slow in solution (Table 1), and so the lack of reactivity is perhaps
not surprising in this case.
Figure 2. EPR spectra showing formation of biphenyl radical cation upon
UV irradiation in Na-ZSM-5 zeolite. (a) Experimentally measured spectrum;
(b) simulated spectrum.
controlled. The other reactions are all exothermic and their rate
constants exhibit a small increase with increasing reaction
exothermicity, with the most exothermic reactions approaching
the diffusion controlled limit.
Solid-State Oxidation. When biphenyl (BP) is sublimed into
activated NaZSM-5 under vacuum, the zeolite crystals remain
colorless. Irradiation of the BP occluded NaZSM-5 using a
Rayonet Reactor (254 nm) at 298 K caused the zeolite to turn
bright green. This color change has been described previously,
and a diffuse-reflectance UV-vis spectrum of the irradiated
zeolite matches that previously observed, and attributed to the
radical cation of biphenyl.6d EPR analysis of the sample gave
rise to the spectrum given in Figure 2a. The coupling constants
and line widths are consistent with literature values for BP•+
reported previously.6d,e A simulation of the EPR spectrum of
BP•+ is given in Figure 2b. The close match between the ob-
served and calculated spectra, and the similarity to previous
observations confirm generation of the BP•+ in the NaZSM-5
zeolite channels.
In contrast to the behavior in acetonitrile solvent (above),
and as observed previously, the EPR signal due to the BP•+ is
quite long-lived, and is stable on a time scale of many hours,
even days, in the absence of any added potential reactants. This
is a remarkable considering the high oxidizing power and
potential reactivity of this species. As discussed above, its
lifetime in acetonitrile solvent is shorter by more than 10 orders
of magnitude! Presumably, the zeolite cage protects the
BP•+@NaZSM-5 from unwanted reaction with adventitious
nucleophiles and reductants. If this is true, then the zeolite cage
also has the potential to prevent the desired redox reactions of
the BP•+ with the donors of Scheme 2. However, subsequent
addition of the electron donors demonstrated this not to be the
case (Table 1), although the reactions were found to be highly
selective.
Addition of the alkene 2,3-DMP, diene COD, and enol ether
2,3-DHF also resulted in no detectable quenching of the BP•+
EPR signal. These observations are considerably more surpris-
ing, because in each case reaction with BP•+ in solution is
exothermic, and their kox are all within an order of magnitude
of the diffusion controlled limit. In these cases, factors other
than reaction exothermicity must be important in determining
reactivity.
When tetramethylethylene is added to the BP•+@NaZSM-5,
quenching of both the color and the BP•+ EPR signal is
observed. In this case, appearance of a new EPR spectrum is
observed, that is consistent with that reported previously for
TME•+.18 Within 10 minutes of the addition of TME to BP•+,
the absorptions attributed to BP•+ have disappeared and the
(16) (a) Schepp, N. P.; Johnston, L. J. J. Am. Chem. Soc. 1996, 118, 2872-
2881. (b) Lew, C. S. Q.; Brisson, J. R.; Johnston, L. J. J. Org. Chem. 1997,
62, 4047. (c) Masclet, P.; Grosjean, D.; Mouvier, G.; J. Electron Spectrosc.
Relat. Phenom. 1973, 2, 225. (d) Bernhard, K.; Geimer, J.; Canle-Lopez,
M.; Reynisson, J.; Beckert, D.; Gleiter, R.; Steenken, S. Chem. Eur. J.
2001, 7, 4640-4650. (e) Miller, L. L.; Nordblom, G. D.; Mayeda, E. A. J.
Org. Chem. 1972, 37, 916.
Addition of the two linear dienes, 2,4-hexadiene (HD) and
2,5-dimethyl-2,4-hexadiene (DMHD) to the BP•+@NaZSM-5
resulted in rapid fading of the characteristic green color of the
BP•+@NaZSM-5 within a few seconds. Presumably, irreversible
electron transfer occurred from the dienes to the BP•+ within
the zeolite channels. Measurement of the EPR spectrum
(17) Roduner, E.; Wu, L. M.; Crockett, R.; Rhodes, C. J. Catal. Lett. 1992, 14,
373.
9
14920 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003

Selective Solid State Photooxidant
A R T I C L E S
Scheme 3. General Mechanism for Photooxidation of Donors by
Biphenyl Included and Irradiated NaZSM-5 Zeolites
electron. It has been proposed that the electron can be trapped
in the form of a Na4 species.20 The second step involves
3+
subsequent transfer of a single electron from the electron donor,
D to the BP•+ (via more than one possible mechanism, see
further below). Whether this process actually occurs depends
on a number of factors.
Apart from the higher oxidation potential donor 2-MP, the
efficiencies of the reactions between BP•+ and the donors in
acetonitrile solution vary over a fairly limited range (Table 1).
On the other hand, the efficiencies of the reactions between
BP•+ and the donors in NaZSM-5, as given by the observation
of quenching of the BP•+ EPR signal, vary dramatically (Table
1), and clearly depend on other factors in addition to oxidation
potential. A quantitative comparison of the solution phase
second-order rate constants kox with the EPR quenching
observations in the solid state is clearly not possible, but it is
obvious that the variation in reaction efficiency is much larger
in the zeolite than in solution, to the extent that either reaction
occurs rapidly (seconds to minutes), or not at all over a time
period of days.
The alkene TME and the two dienes HD and DMHD all
quench the BP•+ signal, but the alkene 2-MP does not. As
discussed above, the lack of reaction with the latter can be
explained as a consequence of the fact that electron transfer is
slightly endothermic in this case, and is exothermic for the
others, and thus the lack of reaction is not surprising.
It is interesting that the difference in reactivity between the
exothermic and endothermic reactions may be greater in the
zeolite, compared to solution. Although the present data do not
allow these to be distinguished, possible reasons for this include
differences in diffusion rates and/or reorganization energies in
the two media. Nevertheless, we conclude that when the ther-
modynamics of electron transfer are favorable, electron transfer
may occur efficiently. The interesting cases are those reactions
that do not occur, even when thermodynamically favorable.
These reactions are presumably controlled by one or more of
the other criteria mentioned above, including the topology,
geometry and atomic composition of the zeolite.
The known size and shape selectivity of zeolites prompted
us to estimate the molecular dimensions of the donors. These
were defined in terms of the minimum cylindrical dimensions,
i.e., the dimensions of the smallest cylinder that will accom-
modate the most reasonable conformation of the molecule. The
molecules were first subject to a molecular dynamics annealing
process to obtain a rough lowest energy conformation, which
was followed by a molecular mechanics energy minimization.21
The “length” of the cylinder was then assigned to the length of
the longest axis of the resulting structure, determined by visual
inspection, that was consistent with providing the appropriate
minimum orthogonal width. It is assumed that the molecule will
orient with the long axis aligned with the zeolite channel, which
Figure 4. EPR spectra showing reduction of biphenyl radical cation (BP•+
)
by tetramethylethylene (TME) in Na-ZSM-5 zeolite, and formation of the
TME•+. (a) Spectrum of BP•+ formed by irradiation of BP in Na-ZSM-5,
(b) spectrum observed 2 min after addition of TME; (c) spectrum observed
6 min after addition of TME.
spectrum of the TME•+ persists for several hours. Control
experiments show that TME does not spontaneously ionize on
NaZSM-5 at room temperature under our conditions, indicating
that reaction with BP•+ in the zeolite is necessary for the
oxidation of TME to occur. The spectra recorded during the
reaction are shown in Figure 4.
Discussion
The reactivity of supramolecular systems can be discussed
in terms of four main criteria.19 Topology describes the shape
of the three-dimensional space in which the components of the
supramolecular assembly are found. For example, the MFI
family of zeolites to which NaZSM-5 belongs defines a reaction
space of parallel channels intersecting with sinusoidal channels.13
Geometry describes the actual dimensions of that topology in
terms of size and angles. Structure encompasses the chemical
features of the species involved, including thermodynamics and
binding affinity of the guest molecules to various sites within
the zeolite environment. Finally, chemical dynamics that
describe the motion of the molecules within the restricted
reaction space must also be considered. With these four criteria
in mind, we can address the overall reactivity of “guest@host”
assemblies, or in our case, BP•+@ZSM-5.
A general mechanism for the intrazeolite oxidation is given
in Scheme 3. Irradiation of the biphenyl leads to ejection of an
(18) (a) Corio, P. L.; Shih, S. J. Phys. Chem. 1971, 75, 3475. (b) Chemerisov,
S. D.; Werst, D. W.; Trifunac, A. D. Chem. Phys. Lett. 1998, 291, 262. (c)
Roduner, E.; Crockett, R.; Wu, L. M. J. Chem. Soc. Far. Trans. 1993, 89,
2101.
(19) Turro, N. J.; Barton, J. K.; Tomalia, D. A. Acc. Chem. Res. 1991, 24, 332.
(20) Liu, X.; Zhang, G.; Thomas, J. K. J. Phys. Chem. 1995, 99, 10 024.
(21) Calculations were performed using Chem3D-Pro, Cambridgesoft Corp.,
Cambridge, MA. Details are given in the Experimental Section.
9
J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003 14921

A R T I C L E S
Morkin et al.
assumed to be too large to fit into the NaZSM-5 channels, and
is consistent with the cylindrical width of 6.1Å, Table 1.24
Indeed, no reaction can be detected, despite the fact that the
reaction occurs in solution with a similar rate constant to TME
(Table 1). Presumably, the zeolite can accommodate much
smaller quantities of this diene compared to the smaller donors,
although we have no way of determining the extent to which
the donors are actually absorbed. The fact that no reaction is
also observed for the 2,3-dimethyl-2-pentene (2,3-DMP), how-
ever, was initially surprising. The thermodynamics of this
oxidation are very similar to those for TME, and the additional
methyl group compared to TME was not expected to have a
large influence. However, the cylindrical width of the 2,3-DMP
is significantly larger than that for TME (5.9 vs 5.4Å), which
evidently is sufficient to restrict access to the BP•+ to such an
extent that no reaction can be observed!
More than one mechanism has been proposed for bimolecular
electron-transfer reactions in zeolites,25 from those involving
direct contact between the reactants, to electron transfer “at a
distance”, with the zeolite frame mediating electron transfer
between nondirectly interacting reactants.9h For example, it has
been proposed that electron migration could occur over the
electron deficient sites in the zeolite, such as cations,25d cation
clusters, and tri-coordinated aluminum centers.9h Alternatively,
hole migration (redistribution of positive charge) could occur
via electron rich sites, such as oxygen sites in the zeolite
framework.6a,d The current data do not provide any direct
evidence as to the exact mechanism of electron transfer;
however, some of the observations are interesting in this regard.
Quenching of the BP•+ occurs within seconds with the dienes,
but takes several minutes with TME. If direct contact between
the BP•+ and the donor is required, then mutual diffusion of
one or both partners is required. In this case, diffusion of the
dienes is presumably significantly faster than that of TME. The
cylindrical widths, i.e., 4.4 and 5.1 Å for the dienes and 5.4 Å
for TME support this suggestion. Alternatively, the dienes and
the TME may quench by different mechanisms. For example,
spontaneous ionization of the DMHD certainly can occur, and
transfer of this ionized electron to the BP•+ (perhaps via the
zeolite framework) could account for the observations. The TME
does not spontaneously ionize, and it is possible that diffusion
to direct contact between the TME and the BP•+ is required,
which would be consistent with the considerably slower reaction
in this case. The estimated widths for TME and BP are, in fact,
identical (5.4Å); however, their relative mobilities may be
controlled more by electronic effects with the framework.
Computational evidence from Bremard et al. suggests that
the preferred binding site for BP is close to the intersections
between the straight and sinusoidal channels of NaZSM-5.6d
For electron transfer to occur in contact, either the donor must
migrate to this site, or the BP•+ must migrate from this site.
EPR simulations of the hyperfine anisotropy in the EPR spectra
of TME•+ and DMHD•+ in HZSM-5 suggest restricted motion
for these radical cations,18c perhaps as a consequence of the
terminal methyl groups, and a location near the ellipsoid
Figure 5. Crystal structure of ZSM-5, showing both the straight and
sinusoidal channels. (b) Top view of ZSM-5 showing the straight channel
diameter where substrates can enter the internal surface of the zeolite. (c)
Molecular dynamics and mechanics minimized conformations for (left)
tetramethylethylene (TME) and (right) 2,3-dimethyl-2-pentene (2,3-DMP),
compared to the dimensions of the straight channel of ZSM-5. The estimated
molecular cylindrical dimensions are included in brackets (see text).
means that the width of the cylinder defines the size of the
molecule that the channel must accommodate. The dimensions
were determined as equivalent internuclear length and width
distances, plus 1Å to account for the size of the exterior
hydrogen atoms.22 The results are shown in Figure 5, for the
examples of 2,3-DMP and TME, and summarized in Table 1
for all of the donors. Several points should be noted. First, the
dimensions are of the most reasonable (low energy) conformers,
and not necessarily the “smallest” conformers. More compact
conformations can be attained, however, only at an energy cost.
Second, the longest axis is not necessarily along one of the
obvious dimensions, e.g., the carbon-carbon double bond. For
example, the length of the cylinder for 2-MP is longer than that
for 2,3-DMP, corresponding to a quite different axis, because
in this way, a smaller width dimension can be obtained. Third,
the dimensions determined in this way are clearly somewhat
arbitrary; however, we believe that they represent a relative
ranking that is adequate for the present purposes.
The kinetic width of the channels in NaZSM-5 are estimated
to be 5.5 Å.23 The cyclic COD was not expected to react with
the BP•+, based on the fact the eight-membered ring was
(24) Kokotailo, G. T.; Lawton, S. L.; Olson, D. H.; Meier, W. M. Nature 1978,
272, 437.
(25) See, for example: (a) Turberville, W.; Robins, D. S.; Dutta, P. K. J. Phys.
Chem. 1992, 96, 5024. (b) Liu, X.; Iu, K.-K.; Thomas, J. K. J. Phys. Chem.
1989, 93, 4120. (c) Iu, K.-K.; Thomas, J. K. Langmuir, 1990, 6, 471. (d)
Wernette, D. P.; Ichimura, A. S.; Urbin, S. A.; Dye, J. L. Chem. Mat. 2003,
15, 1441-1448.
(22) (a) 1Å was chosen as a rough compromise between the hydrogen covalent
(0.4) and van der Waals radii (1.1-1.3Å).22b (b) Cotton, F. A.; Wilkinson,
G AdVanced Inorganic Chemistry; Interscience: New York, 3rd ed.; 1972;
pp. 116-119.
(23) Olson, D. H.; Kokotallo, G. T.; Lawton, S. L.; Meier, W. M. J. Phys. Chem.
1981, 85, 2238.
9
14922 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003

Selective Solid State Photooxidant
A R T I C L E S
Summary
intersections of the straight and sinusoidal channels of the
zeolite. If the TME, indeed experiences restricted motion within
the zeolite framework, then diffusive motion may well be slow,
as observed.
The generation of BP•+ and analogous radical cations in
NaZSM-5 zeolites represents an example of a new class of
supramolecular solid-state irreversible photooxidizers. Evidence
is provided for one-electron oxidation of simple organic electron
donors in this manner. The oxidation reactions exhibit high
selectivity, and can be controlled by the thermodynamics of the
reactions (via the oxidation potential of the donor or a substituted
biphenyl), and also the molecular size, shape, and Lewis basicity
of the donors. In those cases where oxidation is observed,
complete quenching of the BP•+ EPR signal indicates that the
reactions are quantitative and thus efficient in a chemical sense.
An important unanswered question, however, relates to the
photochemical efficiency, which is controlled by the as yet
unknown quantum yield of formation of the BP•+. Studies to
further address this and other factors that control the selectivity,
reactivity, and product distributions in these systems are ongoing
in our laboratories.
Because of its relatively small width (3.8 Å), a size effect
seems unlikely to be the reason for the lack of reactivity for
the enol ether DHF. Note that this donor has the second lowest
oxidation potential of all of the donors included in the present
study. It is possible that the preferred location of the oxygen-
containing DHF in the zeolite is very different from that of the
hydrocarbons discussed thus far. Two possible consequences
of the basic oxygen can be envisioned. First, the lone pair
electrons on the oxygen could interact in a partially bonding
manner with the cations in the zeolite framework. One of our
groups has already shown that the lone pairs on the carbonyl
oxygen of a phenyl ketone derivative can coordinate with the
metal cations located on the external surface of the zeolite,26
and recently an interaction between the nonbonding electrons
on THF and the sodium ions in ZSM-35 was reported.27 The
coordination of the ketone was found to inhibit its diffusion
into the channels of the ZSM-5 zeolite. It is possible that it is
the restricted diffusive or orientational molecular motion of the
DHF in the zeolite framework that inhibits the reaction in this
case, either as a result of a corresponding coordination process,
or simply because of its larger dipole moment compared to the
other donors. Alternatively, interaction of the nonbonding
electrons on the enol ether with Lewis acid sites in the
framework, or with extraframework Na+ ions, will also increase
the oxidation potential of the enol ether. The effect on oxidation
potential of coordination to Na+ ions has been demonstrated in
electrochemical studies.28 Whether as a result of restricted
motion, or less favorable thermodynamics, it is clear that the
oxidative ability of BP•+@NaZSM-5 can also be controlled via
the Lewis basicity of the electron donor.
The product of electron transfer is D•+, and these radical
cations are observed for two of the three donors that react with
the BP•+. Although the mechanism of formation of the DMHD•+
is ambiguous, there is no doubt that the TME•+ is formed by
electron transfer to BP•+. One question is why HD•+ is not
observed? As discussed above, EPR spectra for both DMHD•+
and TME•+ have been observed previously on H-ZSM-5, but
EPR detection of HD•+ has not been reported. This presumably
reflects the fact that the lifetime of this particular radical cation
is too short under the experimental conditions. The factors that
control the reactivity and lifetimes of the radical cations of
simple alkenes on zeolites have not yet been completely
described.9f,11a,29 It has been suggested that a “tight fit” of the
substrate within the zeolite pores or channels is essential for
generating a long-lived radical cation.29 On the basis of the width
estimates, HD should certainly have the “loosest” fit of any of
the simple alkenes and dienes, which may result in a more
mobile radical cation that can undergo, for example, deproto-
nation or even dimerization reactions.30
Experimental Section
Sodium-exchanged ZSM-5 (Si/Al ) 20) was a Chemie Uetikon
product obtained as a generous gift from Dr. V. Ramamurthy, Depart-
ment of Chemistry, Tulane University, New Orleans, LA. Biphenyl
was obtained from the Aldrich Chemical Co. and sublimed prior to
use. Tetramethylethylene, 3-methyl-2-pentene, 2,3-dimethyl-2-pen-
tene, 2,3-dihydrofuran, cis,cis-1,3-cyclooctadiene, 2,5-dimethyl-2,4-
hexadiene, and 2,4-hexadiene were obtained from Aldrich and used
as received. 9,10-Dicyanoanthracene was available from previous
studies.3
The EPR spectra of the radical cations in the zeolites were recorded
using a Bruker EMX-EPR spectrometer operating at X-band (9.5 GHz).
The radical cations in solution were detected using a transient absorption
apparatus that has been described previously.3
Typical sample preparation for study in the zeolites involved the
calcination of NaZSM-5 in a furnace at 500°C for 12 h prior to use. A
9-mg portion of biphenyl and a 300-mg portion of hot zeolite were
placed in a custom-made EPR tube equipped with a small bulb
attachment containing the alkene (Figure 1). An equimolar amount of
alkene was placed in the small bulb on the side of the tube. All
stopcocks were closed, and the sample was placed on a vacuum line.
The portion of the tube containing the zeolite and biphenyl was
evacuated to 5 × 10-5 Torr. After the evacuation was complete, the
zeolite-biphenyl mixture was heated at 70 °C to sublime the biphenyl
into the channels of the zeolite, typically for a period of 12 h. The
contents of the small bulb then underwent 3 freeze-pump-thaw cycles
to deaerate the alkene. EPR spectra were measured before and after
irradiation was carried out in a Rayonet reactor equipped with 16 254-
nm lamps. Each sample was irradiated for 10 min. The extent of
inclusion of the biphenyl or the added donor into the zeolite is not
known.
Experiments in solution were performed in acetonitrile solvent
(Omnisolv Spectro grade, used as received). Typically, 9,10-dicyanoan-
thracene (ca. 10-4 M) was excited at 355 nm using a nanosecond laser
in the presence of 0.15 M biphenyl. The absorption of the biphenyl
radical cation was monitored at 670 nm. The pseudo-first-order decay
of the biphenyl radical cation was measured as a function of added
donor to yield the bimolecular rate constant for reaction, kox.
(26) Turro, N. J.; Lei, X.; Niu, S.; Liu, Z. Org. Lett. 2000, 2, 3991.
(27) Lin, D.-C.; Zhou, W.-Z.; Guo, J.; He, H.-Y.; Long, Y.-C. J. Phys. Chem.
B. 2003, 107, 3798.
(28) Bard, A. J.; Faulkner, L. R. Chapter 8 In Electrochemical Methods;
Fundamentals and Applications, 2nd ed.; Wiley and Sons: New York,
2001.
(29) Ramamurthy, V.; Caspar, V. J.; Corbin, D. R. J. Am. Chem. Soc. 1991,
113, 594.
(30) (a) Crockett, R.; Roduner, E. J. Chem. Soc., Perkin Trans. 2 1993, 1503.
(b) Werst, D. W.; Piocos, E. A.; Tartakovsky, E. E.; Trifunac, A. D. Chem.
Phys. Lett. 1994, 229, 421. (c) Leu, T. M.; Roduner, E. Stud. Surf. Sci.
Catal. 2001, 135, 2014. (d) Ramamurthy, V.; Lakshminarasimhan, P.; Grey,
C. P.; Johnston, L. J. Chem. Commun. 1998, 2411. (e) Garcia, H.; Marti,
V.; Casades, I.; Fornes, V.; Roth, H. D. Phys. Chem. Chem. Phys. 2001, 3,
2955.
9
J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003 14923

A R T I C L E S
Morkin et al.
The cylindrical widths were estimated by first performing a molec-
ular dynamics annealing calculation on the molecules using the pro-
gram Chem3D-Pro, from CambridgeSoft. The cooling rate was 1 kcal/
atom/ps, and the final temperature was 300 K. An additional molec-
ular mechanics optimization was performed to a minimum RMS
gradient of 0.1. The relevant axes were then determined by visual
inspection. Equivalent atomic center-to-center distances were mea-
sured graphically, and 1.0Å was added to this distance to take into
account the dimensions of the exterior hydrogen atoms. Repeated
calculations of this type gave dimensions that were reproducible to ca.
0.15Å.
Acknowledgment. The authors at Columbia University thank
the National Science Foundation for its generous support of this
research (Grant No. NSF CHE01-10655). We also thank Lloyd
Abrams for helpful comments and suggestions regarding the
calculation of molecular dimensions and Kirsten Ostberg for
the graphic in Figure 1. The authors at Arizona State University
also wish to thank the National Science Foundation for financial
support Grant No. NSF CHE-0213445.
JA0368406
9
14924 J. AM. CHEM. SOC. VOL. 125, NO. 48, 2003