Long-lived charge-separation by retarding reverse flow of charge-balancing cation and zeolite-encapsulated Ru(bpy)32+ as photosensitized electron pump from zeolite framework to externally placed viologen

Article abstract of DOI:10.1021/ja020446a

K⁺-exchanged, Ru(bpy)₃²⁺-encapsulating zeolite-Y [K⁺-Ru(bpy)₃²⁺Y] and N-[3-(dicyclohexyl-methyl)oxypropyl-N'-methyl-4,4′-bipyridinium [DCH-MV²⁺] were prepared, and visible light-

Full text of DOI:10.1021/ja020446a

Published on Web 05/23/2002
Long-Lived Charge-Separation by Retarding Reverse Flow of
2+
Charge-Balancing Cation and Zeolite-Encapsulated Ru(bpy)3
as Photosensitized Electron Pump from Zeolite Framework to
Externally Placed Viologen
Yong Soo Park, Eun Jung Lee, Yu Sung Chun, Young Deok Yoon, and
Kyung Byung Yoon*
Contribution from the Center for Microcrystal Assembly and Department of Chemistry,
Sogang UniVersity, Seoul 121-742, Korea
Received March 20, 2002
Abstract: K+-exchanged, Ru(bpy)32+-encapsulating zeolite-Y [K+-Ru(bpy)32+Y] and N-[3-(dicyclohexyl-
methyl)oxypropyl-N′-methyl-4,4′-bipyridinium [DCH-MV2+] were prepared, and visible light-induced electron
transfer from the zeolite-encapsulated Ru(II) complex to the size-excluded viologen was studied in aceto-
nitrile. Addition of a series of crown ethers (CEs) into the heterogeneous solution leads to over a 10-fold
increase in the yield of DCH-MV•+, where the yield linearly increases as the formation constant of CE with
K+ [Kf(K+)CE] increases. The following two sequential events are attributed to be responsible for the above
novel phenomenon. First, K+ ions are liberated from the zeolite to solution during interfacial electron transfer
from the photoexcited Ru(II) complexes to DCH-MV2+. Second, the liberated K+ ions form strong host-
guest complexes with the added CE molecules, which leads to retardation of the reverse flow of the cations,
hence the charge-balancing electrons, from the solution to the zeolite. Surprisingly, the yield of DCH-MV•+
reaches more than ∼50 times the amount of Ru(bpy)32+ situated in the outermost supercages, despite the
absence of electron relay in the zeolite. This is attributed to photosensitized electron pumping from the
2+
zeolite framework to viologen by the outermost Ru(bpy)3 ions. In support of the above conclusion,
Ru(bpy)33+ does not accumulate in the zeolite host while DCH-MV•+ accumulates in the supernatant solution.
Consistent with the above, the independently prepared hexafluorophosphate salt of Ru(bpy)33+ is reduced
to Ru(bpy)3 in acetonitrile upon contact with Ru(bpy)32+-free M+Y (M+ ) Li+, Na+, K+, Rb+, and Cs+),
2+
where the yield increases as the donor strength of the framework oxygen increases. Although small, thermal
electron transfer also takes place from the zeolite framework to DCH-MV2+, where the yield increases
upon increasing the donor strength of the framework, concentration of DCH-MV2+, temperature, and Kf(K+)CE
(when K+Y is the zeolite host). The photoyield is always higher than the thermal yield by 4-30 times,
2+
confirming that the zeolite-encapsulated Ru(bpy)3 serves as the photosensitized electron pump.
Introduction
As a possible means to demonstrate the strategies aimed to
achieve long-lived CSS, zeolites have often been employed as
The spatial separation of the intimate ion pair (D+,A-)
generated by photoinduced electron transfer (PET) from a donor
(D) to an acceptor (A) is commonly referred as charge separation
(CS). The ability to induce and maintain the charge-separated
state (CSS) is essential for the vitality of the natural and artificial
light energy harvesting systems.1 One way to achieve long-lived
CSS is to retard the energy wasting thermal back electron
transfer (TBET) from the intimate ion pair, namely the electron
transfer (ET) from A- to D+ (eq 1).1-18
(3) Kalyanasundaram, K. Photochemistry in Microheterogeneous Systems;
Plenum: New York, 1987.
(4) Yoon, K. B. In Solid State and Surface Photochemistry; Ramamurthy, V.,
Schanze, K. S., Eds.; Marcel Dekker: New York, 2000; Vol. 5, Chapter 4.
(5) Yoon, K. B. Chem. ReV. 1993, 93, 321.
(6) Kim, Y. I.; Mallouk, T. E. J. Phys. Chem. 1992, 96, 2879.
(7) Krueger, J. S.; Mayer, J. E.; Mallouk, T. E. J. Am. Chem. Soc. 1988, 110,
8232.
(8) Yonemoto, E. H.; Kim, Y. I.; Schmehl, R. H.; Wallin, J. O.; Shoulders, B.
A.; Richardson, B. R.; Haw, J. F.; Mallouk, T. E. J. Am. Chem. Soc. 1994,
116, 10557.
(9) Dutta, P. K.; Incavo, J. A. J. Phys. Chem. 1987, 91, 4443.
(10) Kalyanasundaram, K. Coord. Chem. ReV. 1982, 46, 159.
(11) Borja, M.; Dutta, P. K. Nature 1993, 362, 43.
PET
D + A {
} [D+, A-] {
intimate
\
CS} D+ + A-
CSS
(1)
(12) Bossmann, S. H.; Turro, C.; Schnabel, C.; Pokhrel, M. R.; Payawan, L.
M., Jr.; Baumeister, B.; Wo¨rner, M. J. Phys. Chem. B. 2001, 105, 5374.
(13) Sykora, M.; Kincaid, J. R. Nature 1997, 387, 162.
(14) Bhuiyan, A. A.; Kincaid, J. R. Inorg. Chem. 2001, 40, 4464.
(15) Fukuzumi, S.; Urano, T.; Suenobu, T. Chem. Commun. 1996, 213.
(16) Persaud, L.; Bard, A. J.; Campion, A.; Fox, M. A.; Mallouk, T. E.; Webber,
S. E.; White, J. M. J. Am. Chem. Soc. 1987, 109, 7309.
(17) Kim, Y. I.; Keller, S. W.; Krueger, J. S.; Yonemoto, E. H.; Saupe, G. B.;
Mallouk, T. E. J. Phys. Chem. 1997, 101, 2491.
TBET
ion pair
(1) Fox, M. A., Chanon, M., Eds. Photoinduced Electron Transfer; Elsevier:
New York, 1988.
(2) Ramamurthy, V., Ed. Photochemistry in Organized and Constrained Media;
VCH: New York, 1991.
(18) Hurst, J. K.; Thompson, D. H. P.; Connolly, J. S. J. Am. Chem. Soc. 1987,
109, 507.
9
10.1021/ja020446a CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 7123-7135
7123

A R T I C L E S
Park et al.
prototypical organizing media for a variety of D and A pairs
for systematic PET reactions.2-18 Prominent among these are
the interfacial PET reactions from the photosensitized donors
situated on the external surfaces of zeolites to the acceptors in
zeolites6-8 or from the photosensitized donors encapsulated in
zeolites to the acceptors dissolved in solution.11-15
donor [Ru(4-mmb)32+] in the adjacent supercage. Overall,
Ru(bpy)2bpz2+ acts as a photosensitized electron pump from
Ru(4-mmb)3 to PVS. This eventually leads to slower TBET
from PVS•- to Ru(4-mmb)33+, which is a weaker acceptor than
Ru(bpy)2bpz3+. Similarly, it has been demonstrated that the
photoyield of MV•+ obtained from the Y incorporating MV2+
and two Ru(II) complexes with different E° values, Ru(bpy)2bpz2+
and Ru(bpy)2(H2O)2+ (E° ) 0.63 V in H2O vs NHE), in the
2+
Mallouk and co-workers demonstrated that PET readily takes
2+
place from Ru(bpy)3 (bpy ) 2,2′-bipyridine) placed on the
external surfaces of zeolites (Y, L, and mordenite) to MV2+
ions incorporated within the pores.6 The PET takes place much
more readily when N,N′-ethylene-2,2′-bipyridinium (2DQ2+) is
tethered, as the electron relay, to one of the bpy ligands of the
Ru(II) complex and when the relay is intercalated into the
channels of L incorporating viologen acceptors.7 The rate of
PET decreases upon increasing the length of the spacer between
the Ru(II) complex and the electron relay.8 The above results
demonstrate that the spatial organization of the photosensitized
donor and the corresponding acceptor at the external and internal
surfaces of zeolites, respectively, leads to much longer-lived
CSSs [Ru(bpy)33+,V•+] as compared to those in which the D
and A pairs are placed in the neighboring cages of Y9 or in
homogeneous solutions.10 It was also demonstrated that intro-
duction of promethazine cation (PMZ+), a size-excluded
secondary electron donor, in solution (CH3CN) ultimately leads
to CS between the oxidized form of PMZ+ (PMZ2+) in solution
and MV•+ in zeolite.8 In this case, the Ru(II) complex on the
external surface of the zeolite behaves as a photosensitized
electron pump from PMZ+ (E° ) 1.18 V vs NHE in CH3CN)
to MV2+ (E° ) -0.21 V vs NHE in CH3CN), since the ET
does not take place otherwise.
adjacent cages is 4 times higher than that from the Y incorporat-
2+ 14
ing MV2+ and only one Ru(II) photosensitizer, Ru(bpy)3
.
Fukuzumi et al.15 demonstrated long-lived CSS between Fe3+
in zeolite-Y and the anion radical of 7,7,8,8-tetracyanoquino-
dimethane (TCNQ•-) in solution, arising from ET from Fe2+
ions exchanged in zeolite-Y to TCNQ in acetonitrile by
employing simultaneously exchanged N-methylacridinium (AC+)
as the photosensitized electron pump.
Mallouk and co-workers have also made efforts to exploit
the advantages of the spatial separation of D-A systems on
and in zeolites for H2 generation from water.16,17 For this
purpose, Zn(II)-porphyrin complexes tethering intercalatable
electron relays were employed as the externally placed photo-
sensitized electron pumps for eventual ET from ethylenedi-
aminetetraacetic acid (EDTA), a sacrificial electron donor in
solution, to the internally incorporated MV2+ and Pt ag-
gregates.16 It was revealed that incorporation of TiO2 or Nb2O5
semiconductor as the electron mediator between the photo-
sensitized electron pump and MV2+ leads to a marked increase
in the quantum yield.17
Hurst and co-workers18 employed vesicles instead of zeolites
as the organizing media as a means to provide insights into the
long-lived CS. For instance, they demonstrated that the triplet
excited state of the negatively charged zinc(II) porphyrin
complex [5,10,15,20-tetrakis(4-sulfonatophenyl)porphinato]-
zinc(II) (3*ZnTPPS4-) undergoes ET quenching by the long-
alkyl chain viologen ions (CnMV2+), despite that they are
incorporated into the negatively charged dihexadecyl phosphate
(DHP) vesicles (CnMV2+-DHP). However, the lifetime of CSS
(ZnTPPS3-, CnMV+) increases significantly when the viologen
radical cation is placed within the negatively charged vesicle,
as a result of facile dissociation of the ion pair into free ions
due to charge repulsion between the negatively charged
ZnTPPS3- and CnMV+-incorporating DHP vesicles. The utiliza-
tion of electrostatic repulsion between DHP vesicles and
ZnTPPS3- for long-lived CS in the above system is likened to
The PET from zeolite-encapsulated Ru(II) complexes to
viologen acceptors placed in the supernatant solution has also
received great attention. For instance, Dutta11 and co-workers
employed Ru(bpy)32+ as the zeolite-encapsulated photosensitized
donor [E° ) -0.87 V in the triplet metal-to-ligand charge-
transferred (3MLCT) excited state, in H2O vs NHE] and propyl
viologen sulfonate (PVS, E° ) -0.41 V in H2O vs NHE) as an
electron acceptor dissolved in water. The photoyield of the CSS
between Ru(bpy)3 and PVS•- increases dramatically when
3+
N,N′-tetramethylene-2,2′-bipyridinium, (4DQ2+, E° ) -0.65 V
in H2O, vs NHE) is incorporated as the electron relay into every
empty supercage of Na+Y. Thus, 4DQ2+ promotes electron
2+
conduction from every photoexcited Ru(bpy)32+ [*Ru(bpy)3
in Na+Y to PVS in solution.
]
that between the negatively charged zeolite framework (Zn-
)
Bossmann and co-workers demonstrated that TiO2 nanowires
incorporated within the interior of Y also serve as the electron
and PVS•- in the systems of Dutta, Kincaid, and their co-
workers11,13 and that between Zn- and TCNQ•- in the system
of Fukuzumi and co-workers.15
2+
relay between the zeolite-encapsulated Ru(bpy)3 and a size-
excluded Co(III) complex placed in solution.12
None of the above reports, however, paid attention to the
phenomenon of simultaneous movement of the charge-balancing
cations (often alkali metal ions) during the interfacial PET and
TBET, except the one by Dutta and co-workers, which discussed
the possible role of the simultaneously migrating charge-
It was demonstrated that the placement of an electron richer
secondary donor adjacent to a photosensitized primary donor
gives rise to further increase in the photoyield of CSS. For
instance, Kincaid13 and co-workers demonstrated that further
increase (∼4-fold) in the photoyield of CSS results when a
stronger donor [Ru(4-mmb)32+, 4-mmb ) 4-methyl-2,2′-bi-
pyridine, E° ) 1.18 V vs NHE] is elegantly placed into the
supercages adjacent to those occupied by a photosensitized
donor [Ru(bpy)2bpz2+, bpz ) bipyrazine, E° ) 1.50 V in H2O
vs NHE]. The key idea of this work is to immediately reduce
2+
balancing cation during intercage ET from *Ru(bpy)3 in a
supercage of Y to MV2+ in the neighboring supercages.19
Accordingly, no attempts have been made to explore the possible
routes to achieve long-lived CSS by blocking the reverse flow
of charge-balancing cations during TBET. We now report that
a charge-balancing cation is indeed liberated from the interior
the oxidized form of the photosensitized donor, Ru(bpy)2bpz3+
,
back to the reduced state, Ru(bpy)2bpz2+, by the stronger
(19) Dutta, P. K.; Turbeville, W. J. Phys. Chem. 1992, 96, 9410.
9
7124 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
Table 1. Composition of M+Y Zeolites Used in This Study and the
sidearms of the viologen, the access of the viologen into the
interior of the zeolite-Y is supposed to be denied. Indeed,
analysis of the supernatant solution of the stirred (48 h)
acetonitrile solution of DCH-MV2+(PF6-)2 (15 mL) suspended
with K+-exchanged zeolite-Y (K+Y, 50 mg, 196 µmol of K+)
revealed that the exchanged amount varies only from 0.34 to
Yields of DCH-MV•+ and Ru(bpy)3 Produced from
2+
-
DCH-MV2+(PF6
)
and Ru(bpy)33+(PF6-)3, Respectively, upon
2
Contact with the Zeolite Powdersa
+
b
2+ c
M Y
composition
−δO
DCH-MV
Ru(bpy)33+ c
Li+Y
Li37Na16Al53Si139O384
0.238
0.265
0.276
0
1
2
4
6
46
175
251
501
2620
Na+Y Na53Al53Si139O384
0.38% (from 0.66 to 0.74 µmol) of the total K+ ions in K+Y,
K+Y
K49Na4Al53Si139O384
Rb+Y Rb35K13Na2H3Al53Si139O384 0.284
-
despite the increase in the total amount of DCH-MV2+(PF6
)
2
Cs+Y
Cs37K14Na2Al53Si139O384
0.304
in the solution from 0.6 µmol (equivalent to 1.2 µmol of K+)
to 3.1 µmol (equivalent to 6.2 µmol of K+). Thus, the ion
exchange of K+ in K+Y with DCH-MV2+ is saturated at around
a 10 mg of each calcined zeolite powder was stirred for 24 h in 5 mL of
5.4 mM acetonitrile solution of DCH-MV2+(PF6-)2 or 1.3 mM acetonitrile
-
solution of Ru(bpy)33+(PF6
)
3
under anaerobic conditions. b Sanderson’s
-
0.4%. This contrasts with the fact that all of the PF6 salt of
partial negative charge density on the framework oxygen. c Yields are
expressed as the number of ions per 100 unit cells of each zeolite.
methyl viologen, N,N′-dimethyl-4,4-bipyridinium [MV2+(PF6-)2],
is exchanged into K+Y under the identical conditions (50 mg
of K+Y, 0.6 and 3.1 µmol of MV2+(PF6-)2, respectively, in 15
mL of CH3CN). Furthermore, in the case of MV2+(PF6-)2, the
exchange level easily reaches to 20% upon stirring K+Y (50
mg) with 20 mL of 11 mM solution of the salt. Therefore, it is
concluded that the access of DCH-MV2+ into the supercage of
K+Y is denied by the supercage window and the exchange of
the bulky ion is limited only onto the external surfaces of the
zeolite particles.
of zeolite to the solution during interfacial ET from a zeolite-
encapsulated photosensitized donor to an acceptor placed in
solution, and over a 10-fold increase in the photoyield of CSS
can be readily achieved through complexation of the liberated
cation with a multiple chelating agent such as crown ether (CE).
During the course of demonstrating the above strategy, we also
discovered a novel phenomenon where dehydrated Zn- acts as
3+
the electron donor to a viologen acceptor and Ru(bpy)3 in
Effect of CE on the Photoyield of DCH-MV•+. We first
hypothesized that charge-balancing cations (K+) should
leave the zeolite host to solution during interfacial ET from
2+
CH3CN and, as a result, the zeolite-encapsulated Ru(bpy)3
acts as a photosensitized electron pump from Zn- to the viologen
in solution.
*Ru(bpy)3 to DCH-MV2+ as a means to compensate for the
2+
Results and Discussion
loss of a negative charge in the zeolite host. As a corollary,
3+
Ru(bpy)3 was chosen as the prototypical photosensitized
TBET from DCH-MV•+ to Ru(bpy)3 should accompany the
2+
electron donor encapsulated in zeolite-Y. Encapsulation of the
Ru(II)-complex in K+-exchanged zeolite-Y (K+Y) was carried
out according to well-established procedures.20-24 The unit cell
reverse migration of the liberated K+ ions in solution back to
the zeolite host. We further hypothesized that introduction of
various CEs into the acetonitrile solution will form complexes
with the liberated K+ ions, which will eventually lead to
retardation of TBET, since CEs are known to form strong
complexes with alkali metal ions through chelation.26 The CEs
employed for this purpose are cis-dicyclohexano-18-crown-6
(DCH-18C6), 18-crown-6 (18C6), and 15-crown-5 (15C5), and
their formation constants for complexation with K+ [Kf(CE-
K+)] are 106.6, 105.7, and 103.0, respectively.27
composition of K+Y used for incorporation of Ru(bpy)3 is
2+
2+
described in Table 1. The loading of Ru(bpy)3 was 54 µmol
per gram of the dried zeolite, which corresponds to one
2+
Ru(bpy)3 per every 10 supercages.
As a means to embody a true interfacial ET from the zeolite-
2+
encapsulated Ru(bpy)3 complexes to an electron acceptor
placed in solution, we prepared the hexafluorophosphate salt
of N-[3-(dicyclohexylmethyl)oxypropyl]-N′-methyl-4,4-bipyri-
dinium [DCH-MV2+(PF6-)2] as a bulky, size-excluded electron
acceptor (see the Experimental Section for the synthetic
procedure).
The estimated kinetic diameters of the CEs in the uncom-
plexed states are 10.3 Å (DCH-18C6 and 18C6) and 9.3 Å
(15C5), respectively.25 Therefore, they are all size excluded by
the zeolite-Y aperture (7.4 Å). In fact, 18C6 and 15C5 have
often been employed as templates for the synthesis of high-
silica and hexagonal zeolite-Y.28 When they are employed for
this purpose, the CEs complexed with K+ or Na+ always remain
trapped within the supercages of the resulting zeolites, due to
their sizes being larger than those of the apertures. Accordingly,
calcination of the resulting zeolites under flowing oxygen has
The kinetic diameter (minimum cross sectional width) of the
acceptor ion was estimated to be 9.1 Å,25 which is larger than
the size of the aperture of zeolite-Y supercage (7.4 Å). Thus,
due to the bulky dicyclohexylmethyl unit tethered to one of the
(20) (a) DeWilde, W.; Peeters, G.; Lunsford, J. H. J. Phys. Chem. 1980, 84,
2306. (b) Quayle, W. H.; Lunsford, J. H. Inorg. Chem. 1982, 21, 97.
(21) Incavo, J. A.; Dutta, P. K. J. Phys. Chem. 1990, 94, 3075.
(22) Dutta, P. K.; Turbeville, W. J. Phys. Chem. 1992, 96, 9410.
(23) Turbeville, W.; Robins, D. S.; Dutta, P. K. J. Phys. Chem. 1992, 96, 5024
(24) Laine, P.; Lanz, M.; Calzaferri, G. Inorg. Chem. 1996, 35, 3514.
(25) The kinetic diameter was estimated on the basis of the three-dimensional
structure of the compound produced by the commercial program Chem-
Draw.
(26) Gokel, G. W. Crown Ethers and Cryptands; The Royal Society of
Chemistry: Cambridge, 1991.
(27) Izatt, R. M.; Bradshaw, J. S.; Nielsen, S. A.; Lamb, J. D.; Christensen, J.
J. Chem. ReV. 1985, 85, 271.
(28) Delprato, F.; Delmotte, L.; Guth, J. L.; Huve, L. Zeolites 1990, 10, 546.
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7125

A R T I C L E S
Park et al.
of light repeatedly, the pellets were rather securely placed on a
supporting glass plate placed within the cell, with a dihedral
angle of 60° from the flat bottom of the rectangular cell.
Addition of less strongly complexing CEs such as 18C6 and
15C5 leads to a progressive decrease in the yield, as shown in
Figure 2A. Thus, after 24 h of irradiation, the resulting yields
are 15.8 and 7.4%, respectively. Figure 3A shows that there
exists a linear relationship between the yield (taken after 24 h)
of DCH-MV•+ and [Kf(CE-K+)].
On the basis of the above results, we initially proposed
Scheme 1 to account for the CE-dependent increase in the yield
of CSS, employing DCH-18C6 as the prototypical CE. Thus,
while the interfacial ET takes place from *Ru(bpy)32+ to DCH-
MV2+, K+ ions also leave the zeolite framework in order to
maintain the overall charge balance in the zeolite system. The
liberated K+ ions then form strong host-guest complexes with
DCH-18C6 [Kf(CE-K+) ) 106.6]. Therefore, in order for the
electron residing in DCH-MV•+ to return to the zeolite-
encapsulated Ru(bpy)33+, the K+ ion should also return to the
zeolite framework again to maintain the charge balance.
However, since the CE complex of K+ cannot pass through the
zeolite-Y aperture (vide supra), the K+ ions should first be
decomplexed from the CE to return to the zeolite host to allow
the interfacial TBET from DCH-MV•+ to Ru(bpy)33+. Thus, as
a result of the strong chelating power of the CE, the complexed
K+ ions can remain in solution for a longer period of time than
in the absence of the CE, which causes retardation of the
interfacial TBET from DCH-MV•+ to Ru(bpy)33+. The linear
relationship established between the yield and Kf(CE-K+) in
Figure 3 further verifies that the yield is governed by the
complexing power of CE.
Figure 1. Progressive increase of DCH-MV•+ in the supernatant solution
of DCH-MV2+(PF6-)2 upon irradiating a pellet of K+-Ru(bpy)32+Y at 450
( 20 nm in the presence of DCH-18C6.
been a routine practice for removal of the encapsulated CEs
from the resulting zeolites. This fact further verifies that the
CEs cannot pass the supercage windows of zeolite-Y also in
the complexed state.
To test the hypotheses, a thin, round, and self-supporting
pellet (d ) 7 mm) made from 10 mg of the dry, K+-exchanged,
Ru(bpy)32+-incorporating zeolite-Y [K+-Ru(bpy)32+Y] [540
nmol of Ru(bpy)32+] was placed in an airtight cell containing
the acetonitrile solution of DCH-MV2+(PF6-)2 [MW ) 698.5,
19 mg, 27 µmol, 50 equiv with respect to the amount of
Ru(bpy)32+] and DCH-18C6 (MW ) 372.5, 6 mg, 16 µmol,
30 equiv). Upon irradiation of the pellet at the wavelengths of
430-470 nm using a 450-nm band-pass filter, the solution
turned blue and the color intensified with time as irradiation
continues. The UV-vis spectral measurement of the blue
solution showed the characteristic spectrum of the viologen
radical cation, DCH-MV•+, whose intensity progressively
increased as shown in Figure 1.
Alternatively, it can be imagined that DCH-MV2+ ions are
first ion-exchanged onto the surface, and ET takes place mostly
from *Ru(bpy)3 to the surface-exchanged DCH-MV2+ ions.
2+
Subsequently, the DCH-MV•+ ions generated on the surface
are then exchanged with the unreduced ones (DCH-MV2+) in
solution. One might then propose that the presence of CEs with
higher affinities for K+ leads to an increase in the net
concentration of the surface-exchanged DCH-MV2+ during ion
exchange of K+ ions on the zeolite surface with DCH-MV2+
in solution, which in turn leads to an increase in the yields of
DCH-MV2+. However, we found that the amounts of DCH-
MV2+ ion exchanged onto the surface of K+Y are nearly the
same, regardless of the type of CE; thus, they are 0.44% (15C5),
0.46% (18C6), and 0.46% (DCH-18C6) of the total exchange-
able cations in K+. Therefore, the above alternative interpretation
for the CE-dependent variation of yield of DCH-MV•+ can be
excluded.
The profile of the yield with respect to time is shown in Figure
2A. The yield of DCH-MV•+ with DCH-18C6 as CE after 24-h
irradiation is 124 nmol, which corresponds to 23.0% with respect
2+
to the amount of Ru(bpy)3 present in the 10-mg pellet (540
nmol). The corresponding yield in the absence of the CE is 12
nmol (2.2%).29 Thus, the addition of DCH-18C6 gives rise to
as much as a 10.3-fold increase in the yield. The molar
extinction coefficients (ꢀ) of DCH-MV•+ independently meas-
ured in acetonitrile for the quantitative analysis of the radical
cation are 46 500 and 12 690 M-1 cm-1 for 397- and 605-nm
bands, respectively, which are quite comparable with those of
MV•+ in acetonitrile.30
The irradiation angle also sensitively affects the overall yield
for a given reaction period. Thus, even much higher yields
(>50%) can be achieved within 24 h by carefully positioning
the thin zeolite pellets vertically so that they can absorb incident
light normal to the surface. However, due to the experimental
difficulties to position the pellet normal to the incident beam
One of the requirements to establish Scheme 1 is to provide
evidence that K+ ions are indeed liberated from the zeolite host
to solution with the amount equivalent to that of DCH-MV•+.
To prove this, the reaction scale was increased by 10 times,
and for this particular experiment, K+-Ru(bpy)32+Y was used
in the powder form (see the Experimental Section for detail).
The quantitative analysis of the supernatant solution at two
different yields of DCH-MV•+, namely 564 and 993 nmol,
revealed the presence of 568 and 1182 nmol of K+ ion,
respectively. Since the amount of K+ ion liberated from K+Y
to solution through ion exchange of K+ by DCH-MV2+ is
negligible (vide supra), the above close matching between the
(29) The measurement of the quantum yield was not attempted since the reaction
system is heterogeneous.
(30) Watanabe, T.; Honda, K. J. Phys. Chem. 1982, 86, 2617.
9
7126 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
Figure 2. CE-dependent (as indicated) progressive increase in the total (A), the corresponding thermal (B), and the genuine photoinduced (C) yields of
DCH-MV•+ with time.
Figure 3. The linear relationship between the formation constant of CE with K+ [Kf(CE-K+)] and the total (A), the corresponding thermal (B), and photoinduced
(C) yields of DCH-MV•+ after 24 h.
Scheme 1. Liberation of K+ from the Zeolite Host to Solution at
DCH-MV2+ and they return to the zeolite framework during
2+
the Time of Interfacial ET from *Ru(bpy)3 to DCH-MV2+ and the
3+
TBET from DCH-MV•+ to Ru(bpy)3
.
Subsequent Complexation of the Liberated K+ by DCH-18C6, as a
Result of Photoexcitation (430-470 Nm) of K+-Ru(bpy)32+Ya
Thermal ET from K+-Ru(bpy)32+Y and M+Y to DCH-
MV2+. It is most likely that PET from *Ru(bpy)3 to DCH-
2+
MV2+ is limited to those Ru(II) complexes placed at the
outermost supercages of zeolite-Y crystals. From the fact that
the Ru complexes are homogeneously distributed within the
interiors of zeolite crystals when the loading levels are less than
or equal to two Ru complexes per three supercages,24 it is
2+
reasonable to expect that about 1% of the total Ru(bpy)3
complexes in the K+-Ru(bpy)32+Y with the loading level of 1
Ru(bpy)3 per 10 supercages is situated at the outermost
2+
supercages of the zeolite particles. In this respect, considering
that Zn- is insulating, the observed yield (g23.0%) of DCH-
MV•+ in the presence of DCH-18C6 (after 24 h) is surprisingly
high. This unexpected result suggests that there must be some
sacrificial electron donors within the zeolite host, if not in
2+
3+
solution, that keep regenerating Ru(bpy)3 from Ru(bpy)3
.
The control experiment with the zeolite-free solution containing
only DCH-MV2+ and DCH-18C6 showed no indication of
DCH-MV•+ formation, even after irradiation for 24 h, even by
careful spectroscopic analyses. This verifies that there are no
impurities in the solvent, DCH-MV2+(PF6-)2, and DCH-18C6
that serve as the electron donors to DCH-MV2+. This therefore
leads to the conclusion that sacrificial electron donors must
reside within the zeolite host.
a E˚ Values Are Reported vs NHE in Acetonitrile
yield of DCH-MV•+ and the amount of K+ in solution estab-
lishes that K+ ions are indeed liberated from the zeolite frame-
work to solution during the interfacial ET from *Ru(bpy)32+ to
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7127

A R T I C L E S
Park et al.
Interestingly, the control experiment in the dark revealed an
interesting fact that about 13% (16 nmol) of the total yield
originates from simultaneous thermal electron transfer (TET).
This means that the genuine photoinduced yield of DCH-MV•+
after subtraction of the thermal yield (16 nmol) corresponds to
20% (108 nmol). This indicates that the yields shown in Figure
2A, actually represent the combined (total) yields originating
from both TET and PET. The corresponding CE-dependent
profiles of the thermal and the genuine photoinduced yields are
shown in panels B and C of Figure 2, respectively. It was also
revealed that the linear relationship also holds between Kf(CE-
K+) and the corresponding yield obtained either from TET or
PET, as demonstrated in panels of B and C of Figure 3,
respectively.
acts as the electron donor from the well-established fact that
Zn- can serve as the electron donor, and the donor strength of
Zn- increases with increasing the size or the electropositivity
of the charge-balancing cation.33-48 For instance, the results from
the XPS studies34-37 of Zn-, the FT-IR studies of various probe
molecules incorporated into zeolites,38-42 the charge-transfer
(CT) interactions between Zn- and iodine43 or MV2+ 44, and
the UV-vis spectral shift of the arene donor-acceptor com-
plexes encapsulated within zeolite hosts45 have all served as
solid experimental bases in establishing the fact that Zn- serves
as the electron donor and the donor strength of Zn- increases
with increasing the electropositivity of the charge-balancing
cation.
There are numerous other examples that demonstrate the
donor property of Zn-. For instance, it has long been known
that trinitrobenezene49,50 and tetracyanoethylene50,51 immediately
form the corresponding anion radicals upon contact with Na+Y,
even at room temperature. It has been demonstrated that ET
takes place from the framework to photosensitized electron
acceptors such as pyrene,52-54 tetracyanobenezene,55 1,4-di-
cyanobenzene,55 pyromellitic dianhydride,55 and methyl viologen
(MV2+).56 In close relation to the above, zeolite frameworks
have been known to eject electrons upon exposure to high-
energy radiations such as γ-ray,57 X-ray,58 electron beams,59,60
Zeolite Framework as Electron Donor. Knowing that
2+
Ru(bpy)3 does not thermally reduce DCH-MV2+ to DCH-
2+/3+
MV•+ (E° for Ru(bpy)3
is 1.53 V31 and that for DCH-
MV2+/•+ is -0.25 V,32 vs NHE in CH3CN), the discovery of
the TET process was highly intriguing. This, coupled with the
aforementioned fact that the acetonitrile solutions of CEs contain
no species that can reduce DCH-MV2+, prompted us to examine
K+Y itself as a possible electron donor for DCH-MV2+. For a
more systematic study, we prepared a series of alkali metal-
exchanged zeolites-Y (M+Y, M+ ) Li+, Na+, K+, Rb+, and
Cs+) with the compositions listed in Table 1 (column 2), and
they were calcined at 550 °C for 15 h under flowing oxygen to
remove any organic contaminants. Immediately after calcination,
they were evacuated at 350 °C for 10 h before they were finally
transferred into a glovebox where an aliquot (10 mg) of each
zeolite was introduced into a vial containing 5 mL of acetonitrile
and DCH-MV2+(PF6-)2 (19 mg, 27 µmol).
(33) (a) Mortier, W. J.; Schoonheydt, R. A. Prog. Solid St. Chem. 1985, 16, 1.
(b) Mortier, W. J. J. Catal. 1978, 55, 138. (c) Heidler, R.; Janssens, G. O.
A.; Mortier, W. J.; Schoonheydt, R. A. J. Phys. Chem. 1996, 100, 19728.
(d) Van Genechten, K. A.; Mortier, W. J. Zeolites 1988, 8, 273.
(34) (a) Barr, T. L.; Lishka, M. A. J. Am. Chem. Soc. 1986, 108, 3178. (b)
Barr, T. L.; Chen, L. M.; Mohsenian, M.; Lishka, M. A. J. Am. Chem.
Soc. 1988, 110, 7962. (c) Barr, T. L. Zeolites 1990, 10, 760.
(35) Okamoto, Y.; Ogawa, M.; Maezawa, A.; Imanaka, T. J. Catal. 1988, 112,
427.
(36) (a) Huang, M.; Adnot, A.; Kaliaguine, S. J. Am. Chem. Soc. 1992, 114,
10005. (b) Huang, M.; Adnot, A.; Kaliaguine, S. J. Catal. 1992, 137, 322.
(37) Kaushik, V. K.; Bhat, S. G. T.; Corbin, D. R. Zeolites 1993, 13, 671.
(38) (a) Barthomeuf, D. J. Phys. Chem. 1984, 88, 42. (b) Barthomeuf, D. Stud.
Surf. Sci. Catal. 1991, 65, 157. (c) Murphy, D.; Massiani, P.; Franck, R.;
Barthomeuf, D. J. Phys. Chem. 1996, 100, 6731.
Interestingly, although the presence of DCH-MV•+ was not
visually apparent, the UV-vis analysis of the supernatant
solutions from the acetonitrile suspensions of each zeolite
powder stirred for 24 h in the dark revealed the presence of
DCH-MV•+ in each solution. Furthermore, as listed in Table 1
(column 4), the produced amount of DCH-MV•+ progressively
increased with increasing the size of M+, where the amounts
correspond to a few DCH-MV•+ radical ions, i.e., a few
electrons per 100 unit cells of M+Y.
Knowing that the formation of DCH-MV•+ cannot be
attributed to the presence of electron-rich organic contaminants
inadvertently present within the zeolites, as they were calcined
(vide supra) prior to treatment with DCH-MV2+ under flowing
oxygen, one might then alternatively suspect the transition metal
impurities such as Fe(II) as the electron donors. Such a
possibility is also very low, since the transition elements are
likely to exist in their highest oxidation states after calcination.
Furthermore, if there were still some transition elements capable
of reducing DCH-MV2+, the produced amount of DCH-MV•+
from each M+Y should be nearly the same, since the zeolites
are prepared from the same source (batch) of zeolite-Y.
Having concluded that there are no other candidates that could
be assignable as electron donors, we now propose that Zn- itself
(39) Huang, M.; Kaliaguine, S. J. Chem. Soc., Faraday Trans. 1992, 88, 751.
(40) (a) Barthomeuf, D.; Ha, B.-H. J. Chem. Soc., Faraday Trans. 1973, 69,
2158. (b) de Mallmann, A.; Barthomeuf, D. Zeolites 1988, 8, 292. (c) de
Mallmann, A.; Barthomeuf, D. J. Phys. Chem. 1989, 93, 5636. (d) Dzwigaj,
S.; de Mallmann, A.; Barthomeuf, D. J. Chem. Soc., Faraday Trans. 1990,
86, 431.
(41) Uytterhoeven, L.; Dompas, D.; Mortier, W. J. J. Chem. Soc., Faraday Trans.
1992, 88, 2753.
(42) (a) Mirodatos, C.; Pichat, P.; Barthomeuf, D. J. Phys. Chem. 1976, 80,
1335. (b) Mirodatos, C.; Abou Kais, A.; Vedrine, J. C.; Pichat, P.;
Barthomeuf, D. J. Phys. Chem. 1976, 80, 2366.
(43) Choi, S. Y.; Park, Y. S.; Hong, S. B.; Yoon, K. B. J. Am. Chem. Soc.
1996, 118, 9377.
(44) Park, Y. S.; Um, S. Y.; Yoon, K. B. J. Am. Chem. Soc. 1999, 121, 3193.
(45) Hashimoto, S. Tetrahedron 2000, 56, 6957.
(46) Yashima, T.; Sato, K.; Hayasaka, T.; Hara, N. J. Catal. 1972, 26, 303.
(47) Ono, Y. Stud. Surf. Sci. Catal. 1980, 5, 19.
(48) (a) Hattori, H. Chem. ReV. 1995, 95, 537. (b) Tanabe, K.; Misono, M.;
Ono, Y.; Hattori, H. Stud. Surf. Sci. Catal. 1989, 51, 1.
(49) Turkevich, J.; Ono, Y. AdV. Catal. 1969, 20, 135.
(50) Flockhart, B. D.; Mcloughlin, L.; Pink, R. C. J. Catal. 1972, 25, 305.
(51) Khulbe, K. C.; Mann, R. S.; Manoogian, A. Zeolites 1983, 3, 360.
(52) (a) Liu, X.; Iu, K.-K.; Thomas, J. K. J. Phys. Chem. 1994, 98, 7877. (b)
Liu, X.; Iu, K.-K.; Thomas, J. K. Chem. Phys. Lett. 1993, 204, 163.
(53) Thomas, J. K. Chem. ReV. 1993, 93, 301.
(54) Iu, K.-K.; Thomas, J. K. Colloids Surf. 1992, 63, 39.
(55) Hashimoto, S. J. Chem. Soc., Faraday Trans. 1997, 93, 4401.
(56) McManus, H. J. D.; Finel, C.; Kevan. L. Radiat. Phys. Chem. 1995, 45,
761.
(57) (a) Liu, X.; Thomas, J. K. Chem. Phys. Lett. 1992, 192, 555. (b) Liu, X.;
Thomas, J. K. Langmuir 1992, 8, 1750. (c) Liu, X.; Iu, K.-K.; Thomas, J.
K. J. Phys. Chem. 1994, 98, 13720. (d) Iu, K.-K.; Liu, X.; Thomas, J. K.
J. Phys. Chem. 1993, 97, 8165. (e) Liu, X.; Zhang, G.; Thomas, J. K. J.
Phys. Chem. 1995, 99, 10024.
(58) Alvaro, M.; Garc´ıa, H.; Garc´ıa, S.; Marquez, F.; Scaiano, J. C. J. Phys.
Chem. B. 1997, 101, 3043.
(31) Ghosh, P. K.; Brunschwig, B. S.; Chou, M.; Creutz, C.; Sutin, N. J. Am.
Chem. Soc. 1984, 106, 4772.
(32) The E° values for Ru(bpy)2+/3+ and DCH-MV+/2+ were measured in
acetonitrile using Ag/Ag+ electrode as the reference. LiClO4 (0.1 M) was
used as the electrolyte. As for the filling solution of the reference electrode,
AgNO3 (0.01 M) was dissolved into the above LiClO4 electrolytic solution.
To each measured E° value, 0.5082 V is added to convert the values with
respect to NHE. See Mann, C. K.; Barnes, K. K. Electrochemical Reactions
in Nonaqueous Systems; Marcel Dekker Inc.: New York, 1970; p 26.
(59) (a) Liu, X.; Zhang, G.; Thomas, J. K. J. Phys. Chem. B 1997, 101, 2182.
(b) Zhang, G.; Liu, X.; Thomas, J. K. Radiat. Phys. Chem. 1998, 51, 135.
(c) Liu, X.; Mao, Y.; Ruetten, S. A.; Thomas, J. K. Solar Energy Mater.
Solar Cells 1995, 38, 199.
9
7128 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
Table 2. CE-Dependent Change of Yield of DCH-MV•+ from K+Y
and K+-Ru(bpy)32+Y, Respectivelya
[K -Ru(bpy)32+]Y
+
+
K Y
CE
none
15C5
18C6
thermalb
thermal (T)b
photoinduced (P)b
P/T
0.0
0.6
1.3
2.3
0.0
0.6
1.3
2.1
1.6
4.7
10.2
14.6
7.8
7.8
7.0
DCH-18C6
a A 10 mg pellet of each zeolite was placed in an airtight fluorescence
cell containing 4 mL of acetonitrile solution of DCH-MV2+(PF6-)2 (50 equiv
with respect to a unit cell of each zeolite) in the absence or in the presence
of each CE (30 equiv), and the cell was placed in the dark (thermal) or
under visible light (photoinduced) for 24 h at 30 °C. b Yields are expressed
as the number of ions per 100 unit cells of each zeolite.
Figure 4. The linear relationship between the Sanderson’s (average) partial
negative charge on the framework oxygen (-δO) of M+Y (as indicated)
-
and the yield of DCH-MV•+ produced from DCH-MV2+(PF6
) upon
2
stirring with each zeolite powder in acetonitrile in the dark at ambient
temperature in a glovebox charged with high-purity argon after 24 h.
and high-energy UV (such as 185, 193, and 248 nm).56,61 The
electrons generated from the framework by the high-energy
radiations have been shown to reduce clusters of alkali metal
ions (such as 4 Na+),57 MV2+ 56,57 and O2.59 Thus, the donor
,
property of Zn- has been unambiguously established.
In the meantime, the Sanderson’s electronegativity equaliza-
tion principle has served as a good theoretical basis to correlate
between the experimentally observed donor strength of the
framework and the calculated partial charge of the framework
oxygen.33 Although the yields are very small, we also found
that there exists a good linear relationship between the Sand-
erson’s partial negative charge density on the framework
oxygen62 of each M+Y and the yield of DCH-MV•+, as shown
in Figure 4. This relationship further supports our proposal that
the zeolite framework indeed acts as the electron donor for
DCH-MV2+, and as a result, the degree of TET increases as
the size of M+ increases.
Figure 5. Plot of thermal yield of DCH-MV•+ from the mixture of DCH-
MV2+(PF6-)2 and a pellet of K+-Ru(bpy)32+Y in acetonitrile with respect
to time at different temperatures (A) and the linear relationship between
the thermal yield taken after 24 h and different temperatures (B).
Although the yields in the absence of CEs are in fact very
small, they become significant in the presence of CEs, and the
yield increases as Kf(CE-K+) increases. For instance, as shown
in Table 2 (column 2), the thermal yield of DCH-MV•+ increases
in the presence of CE in the order none < 15C5 < 18C6 <
DCH-18C6. The yield of DCH-MV•+ from K+Y in the absence
of any CE being smaller in Table 2 than that in Table 1 seems
to arise due to the difference in the form of the zeolite; i.e.,
while the zeolites were used in the powder form in Table 1,
they were used in pellets in Table 2. Table 2 also shows that
there is essentially no difference in the thermal yield, regardless
2+
of the presence (column 3) or absence (column 2) of Ru(bpy)3
2+
in the zeolite. This further confirms that Ru(bpy)3 does not
play any role during TET from Zn- to DCH-MV2+. This result
also verifies that there are no remaining organic impurities in
K+-Ru(bpy)32+Y capable of reducing DCH-MV2+
.
Effect of Temperature on Thermal Yield. We found that
the ambient temperature sensitively affects the thermal yield.
Thus, as displayed in Figure 5A, the thermal yield of DCH-
MV•+ from K+-Ru(bpy)32+Y in the presence of DCH-18C6
increases with increasing the temperature of the reaction mixture
over the entire reaction period of 24 h, and there is a linear
relationship between the temperature and the yield of DCH-
MV•+ taken after 24 h, as shown in Figure 5B. We also found
that the yield obtained at 40 °C after 24 h slowly reduces to a
value similar to the one obtained at 30 °C over a period of
several days upon decreasing the temperature to 30 °C, which
rises back to the initial value upon increasing the temperature
(60) Park, Y. S.; Um, S. Y.; Yoon. K. B. Chem. Phys. Lett. 1996, 252, 379.
(61) (a) Liu, X.; Iu, K.-K.; Thomas, J. K. Chem. Phys. Lett. 1994, 224, 31. (b)
Liu, X.; Thomas, J. K. J. Chem. Soc., Faraday Trans. 1995, 91, 759.
(62) The Sanderson’s (average) partial charge of the framework oxygen (δO)
was calculated using the equation δO ) (SZ - SO)/(2.08SO1/2), where SZ
and SO represent the Sanderson’s intermediate electronegativity of each
M+-exchanged zeolite and the Sanderson’s electronengativity of an oxygen
p
q
r
atom. SZ was calculated according to the equation SZ ) (SM SSi SAl
SO )
1/(p+q+r+t), where SM, SSi, and SAl represent the Sanderson’s electro-
t
negativities of the alkali metal cation, silicon, and aluminum, respectively,
and p, q, r, and t represent the number of the corresponding element,
respectively, in a unit cell. The following are the Sanderson’s electro-
negativity for each element: Si, 2.14; Al, 1.71; O, 3.65; Li, 0.89; Na, 0.56;
K, 0.45; Rb, 0.31; Cs, 0.22. The values were taken from: Huheey, J. E.;
Keiter, E. A.; Keiter, R. L. Inorganic Chemistry, 4th ed.; Harper Collins
College Publications: New York, 1993; pp 187ff.
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7129

A R T I C L E S
Park et al.
Figure 7. The time-dependent variation of the ratio of photo-to-thermal
yields.
Figure 6. Plot of thermal yield of DCH-MV•+ from the mixture of DCH-
MV2+(PF6-)2 and a pellet of K+-Ru(bpy)32+Y in acetonitrile with respect
to the concentration of DCH-MV2+(PF6-)2 in the presence (top) and absence
(bottom) of DCH-18C6 at 30 °C.
its absence, i.e., while the yield of DCH-MV•+ increases to 0.09
in the presence of DCH-18C6, it increases only to 0.01 puc in
the absence of the CE upon varying the concentration of DCH-
MV2+ from 0 to 75 puc in the supernatant solution. This result,
coupled with the larger increase in the overall yield, again
emphasizes that the TET reaction is also effectively driven by
complexation between K+ and CEs.
back to 40 °C. This indicates that the TET is an equilibrium
reaction, as expressed in eq 2.
-
Z
n- + DCH-MV2+ h Z(n-1) + DCH-MV•+
(2)
As noticed, the yield of DCH-MV•+ does not linearly increase
in both cases, and we confirmed the reproducibility of the
nonlinear plot. Although the reason for the observed nonlinearity
is yet to be elucidated, the above result further confirms that
TET from Zn- to DCH-MV2+ is an equilibrium reaction, as
depicted in eq 2. This result also explains the unexplained
The above result also indicates that the equilibrium for TET
from the framework to DCH-MV2+ shifts to the right upon
increasing the temperature. In other words, the degree of
temperature-induced increase in the rate of TET from Zn- to
DCH-MV2+ is higher than that from DCH-MV•+ to Z(n-1)-
.
Knowing that the temperature sensitively affects the yield of
DCH-MV•+, we therefore carried out all the reactions at 30 °C,
unless specified otherwise.
phenomenon we used to observe while working with MV2+
-
exchanged zeolites; the intensity of blue color developed on
the zeolite at elevated temperatures is higher with increasing
the degree of ion exchange of MV2+ in the zeolite. This result
also explains why we did not observe MV•+ in our previous
report regarding the charge-transfer interaction between MV2+
and the zeolite framework,44 since the degree of MV2+ exchange
in the zeolites for the previous work was only one out of eight
supercages, which corresponds to 1 puc in the X-axis of Figure
6, where the amount of MV•+ is negligible, in particular, at
room temperature.
The above result now sheds lights on the question we used
to raise while working with MV2+-exchanged Na+Y (∼1 MV2+
per supercage):4,5,63 why does it develop a light blue color during
evacuation at elevated temperatures, such as 150-200 °C, only
to have the color slowly bleach with time upon cooling to room-
temperature, despite the glass vessel containing the dried blue
zeolite being stored in a glovebox charged with high-purity
argon while the vacuum inside the glass vessel was maintained?
Effect of Concentration of DCH-MV2+ on Thermal Yield.
We also found that the thermal yield of DCH-MV•+ from K+Y
increases with increasing the concentration of DCH-MV2+ in
the supernatant solution, as shown in Figure 6. Although the
presence of DCH-MV•+ in all of the supernatant solutions is
always apparent from the spectroscopic analysis, it becomes
visually detectable only when the yield reaches 0.08/unit cell
(puc) or higher. Accordingly, visual observation of the pale blue
coloration of the supernatant solution becomes possible only
when the concentration of DCH-MV2+ is very high and in the
presence of DCH-18C6. This indicates that one is apt to miss
the presence (formation) of MV•+ in the acetonitrile solutions
of MV2+ suspended with dry zeolites unless the concentration
of MV2+ is extremely high or CEs are present in the solution.
Interestingly, the degree of concentration-dependent increase
in the yield is also larger in the presence of DCH-18C6 than in
Ru(bpy)32+ as Photoinduced Electron Pump from Zn- to
DCH-MV2+. Although it was demonstrated earlier that an
equivalent amount of K+ ion is liberated from the zeolite host
to solution while DCH-MV•+ is generated in solution (vide
supra), it is still necessary to prove that the stoichiometric
3+
amount of Ru(bpy)3 is simultaneously generated within the
zeolite host during the PET from K+-Ru(bpy)32+Y to DCH-
MV2+ to establish Scheme 1. However, despite our repeated
trials, the characteristic ESR spectrum of Ru(bpy)33+, which is
supposed to give resonances at g ) 2.67 and g ) 1.24,20b
||
was not detected, even from the K+-Ru(bpy)32+Y, which gave
over 15% true photoyield of DCH-MV•+, while the supernatant
solution showed a strong resonance at g ) 2.00 due to DCH-
MV•+ accumulated in the solution. Nevertheless, it is clear that
the yield of DCH-MV•+ is always higher upon irradiation than
not, as shown in Figure 7. Thus, the ratio of the photo-to-thermal
yields of DCH-MV•+ initially ranges from 8 to 31 and then
equilibrates to ∼7 as the reaction time elapses (see column 5
(63) (a) Yoon, K. B.; Kochi, J. K. J. Am. Chem. Soc. 1989, 111, 1128. (b)
Yoon, K. B.; Kochi, J. K. J. Phys. Chem. 1991, 95, 3780. (c) Yoon, K. B.;
Huh, T. J.; Corbin, D. R.; Kochi, J. K. J. Phys. Chem. 1993, 97, 6492.
9
7130 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
2+
Scheme 2. A Diagram Showing the Role of Ru(bpy)3 as the
Photoinduced Electron Pump for the Net ET Reaction from Zn- to
DCH-MV2+ and the Simultaneous Liberation of K+ from the Zeolite
Host to Solution and the Subsequent Complexation of the Cation
by CE
Figure 8. The spectrum (solid line) of Ru(bpy)33+(PF6-)3 in acetonitrile
(0.6 mM, 5 mL) and the appearance and growth of the absorption of
Ru(bpy)32+ at 450 nm with time upon introduction of a pellet of K+Y (10
mg) into the solution, after 3 (dashed line) and 15 h (dotted line),
respectively. The inset shows an isosbestic point at 570 nm.
MV2+. Monitoring of the supernatant solution containing a pellet
of K+Y by UV-vis spectrometry revealed that the rate of
production of Ru(bpy)32+ is slow but steady, as shown in Figure
8.
To achieve faster equilibrium, we added the acetonitrile
solution of Ru(bpy)33+(PF6-)3 (5 mL, 2 mM) into each powder
of M+Y (10 mg), and the heterogeneous mixtures were stirred
for 24 h. As listed in Table 1 (last column), analysis of the
supernatant solution revealed that the yield of Ru(bpy)32+, hence
the number of electron liberated from the M+Y, gradually
increases from ∼0.5 to ∼5 puc of each zeolite as the size of
M+ increases from Li+ to Rb+, i.e., as the donor strength of
Zn- increases. When compared to the number of electrons
liberated from each M+Y to DCH-MV2+, the numbers are at
least 175 times larger. This can be rationalized from the fact
3+
in Table 2). Therefore, the discrepancy that Ru(bpy)3 does
not accumulate in the zeolite host while the yield of DCH-MV•+
keeps increasing in the supernatant solution, coupled with the
one that the overall yield of DCH-MV•+ is much higher than
the amount of Ru(bpy)32+ present in the outermost supercages
of Y, urged us to examine the reactivity of Ru(bpy)33+ toward
Zn-
.
For the above examination, we synthesized anhydrous
Ru(bpy)33+(PF6-)3 by modification of the procedure reported
3+
by DeSimone and Drago.64 It has been known that Ru(bpy)3
is stable only in strongly acidic solutions such as 6 M aqueous
H2SO4 solution, but it is instantaneously reduced to Ru(bpy)3
3+
that the acceptor strength of Ru(bpy)3 is much higher than
2+
that of DCH-MV2+. Interestingly, Cs+Y was exceptional and
the amount of electrons liberated from the zeolite looks to be
unusually high. A more detailed study is necessary to resolve
the anomalous behavior of Cs+Y.
by water in neutral aqueous media.64 It has generally been
believed that the powerful oxidant is also unstable in organic
solvents such as acetonitrile.64 However, we found that
Ru(bpy)33+(PF6-)3 is in fact unlimitedly stable in acetonitrile
at room temperature provided that the solvent is rigorously
purified before use and the storage bottle is well capped in a
moisture- and organic-vapor-free environment. For purification
of the solvent, treatment with KMnO4 is essential to remove
ESR investigation of the Ru(bpy)33+-treated M+Y zeolites
was also performed with the hope to detect the hole centers in
3+
the zeolites. However, all the zeolites, including the Ru(bpy)3
-
treated Cs+Y, were ESR silent, even at 77 K. Consistent with
our result, no reliable reports about the detection of hole centers
in Zn- have been made. The detection of the hole centers by
ESR may not be possible if the electrons liberated from Zn-
for reducing DCH-MV2+ and Ru(bpy)33+ come from the valence
bands of insulating inorganic solids, from the view of band
theory. Despite the loss of some electrons from Zn-, analysis
of the M+Y zeolites with X-ray powder diffractometry showed
that the structures (including Cs+Y) remain intact.
3+
the organic impurities that cause reduction of Ru(bpy)3 to
Ru(bpy)32+. Figure 8 shows the UV-vis spectrum of Ru(bpy)3
3+
in acetonitrile. The absorption maximums appear at 416 (ꢀ ) 2
230) and 674 nm (ꢀ ) 410 M-1 cm-1), closely matching with
those of Ru(bpy)33+ in acidic aqueous solutions, i.e., 418 (ꢀ )
∼3000)65 and 675 (ꢀ ) 440 M-1cm-1).31 It is also worth noting
that the molar extinction coefficient of the 416-nm band is about
6 times lower than that of the 450-nm band of Ru(bpy)32+, which
is 13 800 M-1cm-1 in acetonitrile.66
On the basis of the above results we now propose that
2+
Ru(bpy)3 actually behaves as a photoinduced electron pump
3+
We found that the characteristic green color of Ru(bpy)3
or photocatalyst for ET from Zn- to DCH-MV2+ according to
slowly disappears from the solution, while the characteristic
orange color of Ru(bpy)32+ appears in the supernatant solution
when a pellet of M+Y is introduced into the acetonitrile solution
of Ru(bpy)33+(PF6-)3. Again, the production of Ru(bpy)3
cannot be attributed to some organic contaminants inadvertently
present in the zeolite hosts, since the M+Y zeolites were calcined
prior to introduction into the Ru(II) solution. Accordingly, we
propose that Zn- itself also acts as the electron donor to
Ru(bpy)33+ as it does even to a weaker electron acceptor, DCH-
2+
the scheme depicted in Scheme 2. First, Ru(bpy)3 reaches
3MLCT excited state *Ru(bpy)32+ upon excitation. Among the
2+
*Ru(bpy)3 complexes encapsulated within the zeolite hosts,
2+
those that reside in the outermost supercages of zeolite-Y then
transfer electrons (one electron each) to DCH-MV2+ ions,
perhaps through the contact at the zeolite apertures. The
3+
2+
generated Ru(bpy)3 are then reduced back to Ru(bpy)3 by
Z
n-. When an electron is discharged from *Ru(bpy)32+ to DCH-
MV2+, a charge-balancing cation K+ is simultaneously liberated
from the zeolite host to the external solution to maintain charge
balance in the zeolite system. This is indispensable, since
Ru(bpy)33+ can balance three negative centers in the zeolite host.
(64) DeSimone, R. E.; Drago, R. S. J. Am. Chem. Soc. 1970, 92, 2343.
(65) Braddock, J. N.; Meyer, T. J. J. Am. Chem. Soc. 1973, 95, 3158.
(66) Young, R. C.; Meyer, T. J.; Whitten, D. G. J. Am. Chem. Soc. 1976, 98,
286.
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7131

A R T I C L E S
Park et al.
center in the framework initially generated at the periphery of
the zeolite crystals into the interior. The temperature-induced
increase in the rate demonstrated in Figure 5 may then arise
due to faster propagation of the hole center in the framework.
Knowing that Zn- can readily reduce Ru(bpy)3 from this
3+
work, now is the time to discuss the possibility whether Zn-
can also reduce the oxidized forms of the photosensitizers
employed by Mallouk,6-8,16,17 Dutta,11 Kincaid,13 and their co-
workers. However, such a pathway is less likely in their systems,
since the reactions were carried out in aqueous solutions and
hydrated zeolite frameworks usually do not act as electron
donors.44 Instead, since their PET reactions are carried out in
aqueous solutions, the possible role of water molecules as the
sacrificial electron donors should be raised from the fact that
3+
Ru(bpy)3 is instantaneously reduced by water in nonacidic
aqueous media, as mentioned earlier.67
Indeed, Lunsford,20b Dutta68 and their co-workers prepared
Ru(bpy)33+-encapsulating Na+Y and demonstrated that the
Ru(III) complex is slowly reduced to Ru(bpy)32+ upon contact
with water. Therefore, in the systems of Mallouk,6-8,16,17 Dutta,11
Kincaid,13 and their co-workers, water is likely to play the role
of the secondary donors such as anhydrous Zn- does in this
paper, unless the reactions are carried out in highly acidic media,
Figure 9. Increase in the photoinduced yield of DCH-MV•+ from the
mixture of DCH-MV2+(PF6-)2 and a pellet of M+-Ru(bpy)32+Y at 30 °C
upon changing the countercation (M+) from Na+ to K+.
Upon reduction of Ru(bpy)33+ to Ru(bpy)32+, a positive charge
is transferred from Ru(bpy)33+ to the framework. Accordingly,
Zn- becomes Z(n-1)-. Independent from ET from Zn- to
Ru(bpy)33+, the liberated K+ ion undergoes complexation with
one of the CEs residing outside the zeolite hosts through
chelation.
or in the presence of good secondary electron donors such as
2+ 13
PMZ+,8 Ru(4-mmb)3
,
Ru(bpy)2(H2O)2+ 14 or EDTA.16,17
,
In the Fukuzumi and co-workers’ system,15 however, the
simultaneous ET from Zn- to *AC+ should also be considered
as a competitive route as well as the proposed ET from Fe2+ to
*AC+, since the reaction was carried out in acetonitrile. As a
result, the CSS between Z(n-1)- and TCNQ•- is likely to coexist
together with the one between Fe3+ and TCNQ•-. Alternatively,
the final CSS between Fe3+ and TCNQ•- may also proceed via
the one between Z(n-1)- and TCNQ•-.
The above scheme coincides well with the absence of
Ru(bpy)33+ within the zeolite host, despite the continual increase
of the true photoinduced yield of DCH-MV•+ in solution, with
the facts that the yield of DCH-MV•+ is always higher upon
irradiation than not and that the photoinduced yield is much
higher than the amount of Ru(bpy)32+ residing in the outermost
supercages. On the basis of Scheme 2, DCH-MV•+ has to now
undergo TBET to the oxidized form of Zn- [Z(n-1)-] but not to
Ru(bpy)33+. Since Z(n-1)- is a much weaker electron acceptor
than Ru(bpy)33+, as judged by the facile ET from Zn- to
Ru(bpy)33+, the corresponding thermodynamic driving force for
the ET from DCH-MV•+ to Z(n-1)- is much less than that from
DCH-MV•+ to Ru(bpy)3+. This situation is likened to the
schemes of Mallouk and co-workers,8 where the surface-
exchanged Ru(II) complex acts as the photosensitized electron
pump from the ultimate electron donor, PMZ+, to MV2+, and
of Sykora and Kincaid,13 where Ru(bpy)2bpz2+ acts as the
The result shown in Figure 9 also indicates that the residual
moisture present in the Ru(bpy)32+Y, due to not high enough
dehydration temperature (200 °C), does not play the role of
secondary electron donor in our case, since the amount of
residual moisture is likely to be higher with Na+ as the
countercation than K+, in compliance with the well-known
phenomenon that the electrostatic field strength in the vicinity
of the cation is higher with decreasing the size.69 Otherwise,
the yield of DCH-MV•+ should be higher with Na+ as the
countercation than K+.
2+
photosensitized electron pump and Ru(mmb)3 as the neigh-
3+
The fact that Ru(bpy)3 can in fact be incorporated safely
boring ultimate electron donor to the oxidized form of the
electron pump. Alternatively, in the respect that Ru(bpy)32+ acts
as the photoinduced electron relay, the situation is also closely
related to the system of Fukuzumi and co-workers, in which
AC+ acts as the photoinduced electron relay from Fe2+ ions in
zeolite-Y to TCNQ in solution.15
within Na+Y (vide supra) as described above seems to contradict
with our proposal that Zn- reduces Ru(bpy)33+. However, a
3+
careful analysis of the procedure to generate Ru(bpy)3 in
3+
Na+Y resolves the discrepancy. Thus, generation of Ru(bpy)3
in Na+Y has been carried out by treating the dried Na+-
The fact that Zn- is the ultimate electron donor is further
supported by the experimental result that the photoinduced yield
from K+-Ru(bpy)32+Y in the absence of any CE is also higher
than that from Na+-Ru(bpy)32+Y, as shown in Figure 9, despite
3+
(67) The fact that water is readily oxidized by Ru(bpy)3 at neutral pHs is
conceivable from the fact that the E° value of water (1.23 V for 2 H2O a
O2 + 4H+ + 4e- vs NHE; Bard, A. J.; Faulkner, L. R. Electrochemical
Methods: Fundmentals and Applications; John Wiley & Sons: New York,
1980; p 700.) is lower than that of Ru(bpy)33+, which can be obtained
only in highly acidic aqueous solutions (1.26 V in 6 M aqueous H2SO4
solution vs NHE).
2+
that the amount of Ru(bpy)3 loading in each zeolite is the
(68) (a) Ledney, M.; Dutta, P. K. J. Am. Chem. Soc. 1995, 117, 7687. (b) Das,
S. K.; Dutta, P. K. Langmuir 1998, 14, 5121.
same (1 per 10 supercages). If the framework serves only as an
inert compartmentalizing host, then there would be no such
difference in the yield, since the number of Ru(bpy)32+ located
in the outermost supercages is the same in both cases.
The reason for both PET and TET reactions being slow in
our system may be attributed to slow propagation of the hole
(69) (a) Rabo, J. A.; Angell, C. L.; Kasai, P. H.; Schomaker, V. Discuss. Faraday
Soc. 1966, 41, 328-349. (b) Breck, D. W. Zeolite Molecular SieVes;
Wiley: New York, 1974. (c) Dempsey, E. J. Phys. Chem. 1969, 73, 3660.
(d) Dempsey, E. Molecular SieVes; Society of Chemical Industry, London,
1968; p 293. (e) Ward, J. W. In Zeolite Chemistry and Catalysis; Rabo, J.
A., Ed.; ACS Monograph 171; American Chemical Society: Washington,
DC, 1976; Chapter 3.
9
7132 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
K+ with the CE. Therefore, for such a reversible reaction like
this case, complexation of K+ by the CE renders the overall
equilibrium of the ET reaction shift to the forward ET.
In summary, the present work reports the following novel
experimental results. The zeolite framework (Zn-) of alkali-
Scheme 3. A Simplified Version of Scheme 2 Showing Only the
Net ET Reaction from Zn- to DCH-MV2+ and the Liberation of K+
from the Zeolite Host to Solution and the Subsequent
Complexation of the Cation by CE
metal-exchanged zeolite-Y (M+Y) acts as an electron donor to
DCH-MV2+ and Ru(bpy)3
.
The zeolite-encapsulated
3+
2+
Ru(bpy)3 acts as a photosensitized electron pump from Zn-
to DCH-MV2+ in solution. The interfacial ET from the zeolite-
encapsulated *Ru(bpy)32+ or Zn- to DCH-MV2+ in solution and
the TBET from DCH-MV•+ to the oxidized form of Zn-
[Z(n-1)-] accompany the simultaneous migration of the charge-
balancing cation, M+. Complexation of the M+ liberated from
the zeolite host to solution by CEs leads to long-lived CSS. A
PET reaction can be driven forward by complexation of the
simultaneously migrating cation with complexing agents.
Ru(bpy)32+Y with chlorine gas, which has been known to
oxidize Zn- as well. For instance, it has been known that
treatment of dry zeolites X and mordenite with chlorine results
- 70
in the formation of Cl2
.
Now, considering the fact that the
E° of Cl2 (1.36 V, in water vs NHE)71 is higher than that of
Ru(bpy)3 ,
2+ 67 it is expected that chlorine will oxidize the zeolite
framework more readily than Ru(bpy)33+ does. Therefore, it is
highly reasonable to expect that all the valence electrons or the
Experimental Section
Materials. Na+Y (LZY-52, Si/Al ratio ) 2.6, Lot No. 968087061020-
S) was purchased from Union Carbide. Hexaammineruthenium(III)
chloride was purchased from Strem. Dicyclohexylmethanol, sodium
hydride, allyl bromide, borane-tetrahydrofuran complex (1 M), tri-
phenyl phosphite, methyl iodide, DCH-18C6, 18C6, 15C5, and 4,4′-
bipyridine were purchased from Aldrich. 18C6 was recrystallized from
hot water. Methyl iodide was distilled before use. Tetrahydrofuran
(THF) was distilled over sodium-benzophenone prior to use. Acetontirile
was stirred for 24 h in the presence of KMnO4 until the purple color
of the oxidant persists in the solution. The KMnO4-treated acetonitrile
was first simple distilled and then redistilled over P2O5 under argon.
The distilled acetonitrile was transferred to a Schlenk flask using a
Teflon cannular under argon atmosphere. The Schlenk flask was then
stored in a glovebox charged with high-purity argon. Diethyl ether was
stirred over the aqueous solution of KMnO4 (diethyl ether:water ) 9:1)
for 15 h. The collected diethyl ether was washed with water and then
treated with (concentrated) sulfuric acid for 15 h. The collected and
water-washed diethyl ether was dried over anhydrous CaCl2 and distilled
over the eutectic mixture of sodium and potassium under argon. The
distilled diethyl ether was also transferred into a Schlenk storage bottle
under argon and kept in the glovebox. Na+Y was washed once with 1
M NaCl solution to ensure incorporation of Na+ as the charge-balancing
cation. Li+Y and K+Y were prepared from Na+Y by treating it with 1
M LiCl and KCl solutions, respectively, five times. Rb+Y and Cs+Y
were prepared from K+Y by treating it with 0.5 M RbCl and CsCl
solutions, respectively, three times. The resulting zeolites after ion
exchanges were washed with distilled deionized water until the silver
ion test for chloride was negative. The washed, ion-exchanged zeolites
were calcined at 550 °C for 10 h under flowing oxygen to remove
organic impurities prior to use.
3+
basic sites that could possibly be oxidized by Ru(bpy)3 are
already removed by chlorine during formation of the Ru(III)
3+
complex, enabling the zeolite framework to host Ru(bpy)3
.
From the conclusion that ET also takes place directly from
Zn- to DCH-MV2+ (vide supra), the pathway described in
Scheme 3 also exists. Therefore, coupled with the result shown
in Figure 7 and Table 2, it is now clear that there exist two
pathways in yielding DCH-MV•+, photoinduced (Scheme 2) and
thermal (Scheme 3). In the case of the former, Ru(bpy)32+ acts
as the electron pump. It is also worth noting that Scheme 3
also represents the net ET reaction described in Scheme 2.
2+
For the electrons to be transferred from *Ru(bpy)3 to the
external DCH-MV2+, many possible mechanisms can be
imagined. One way is to transfer an electron at the aperture of
zeolite-Y through direct contact between the electron donor in
the outermost supercage and the acceptor in solution. An
2+
alternative pathway is to inject an electron from *Ru(bpy)3
to the conduction band of Zn- in the absence of water, and
subsequently, the electron in the conduction band of Zn- is then
transferred to DCH-MV2+ at the terminal of the framework. If
this is the case, then the role of the zeolite framework is likened
to that of 4-DQ2+, MV2+ tethered to the Ru(II) complexes, and
the semiconductors such TiO2 or Nb2O5 employed by Dutta,11
Mallouk,7,8 Bossmann,12 and co-workers. However, for the ET
2+
from *Ru(bpy)3 to the conduction band of Zn- to take place
readily, the energy level of the conduction band of Zn- should
be lower than that of *Ru(bpy)32+. Therefore, without knowing
the precise energy level of the conduction band of Zn-, further
discussion is yet premature.
Incorporation of Ru(bpy)3 into the K+-exchanged zeolite-Y was
2+
carried out according to the well-established procedure.20-24 To remove
Ru(bpy)32+ complexes that might have formed on the external surface
of the zeolite-Y crystals, the Ru(bpy)32+-incorporating zeolites (10 g)
were washed with aqueous KCl solution (0.1 M, 100 mL) for three
times and subsequently with distilled deionized water until the silver
ion test for chloride is negative. The analysis of the amount of
Ru(bpy)32+ in K+-Ru(bpy)32+Y was initiated by dissolving the zeolite
framework with 10% aqueous HF solution according to the reported
Thermodynamic Consideration of the CE-Dependent
Yield. The effect of CE on the increase in the yield of DCH-
MV•+ can also be analyzed from the thermodynamic point of
view. With DCH-18C6 as the prototypical complexing agent,
the estimated standard free energy change (-∆G°) for com-
plexation of K+ is 8.9 kcal mol-1, which corresponds to 62%
of the free energy change (14.3 kcal mol-1) for ET from
2+
procedure.24 The analysis revealed that the loading of Ru(bpy)3 in
2+
*Ru(bpy)3 to DCH-MV2+ if the forward ET occurs directly
K+-Ru(bpy)32+Y is 1 Ru complex per 10 supercages. K+-Ru(bpy)32+Y
(10 mg each) was pressed into round, thin pellets (d ) 7 mm) using a
press at the pressure of 600 kgf cm-2. The pellets of K+-Ru(bpy)32+Y
were dehydrated at 200 °C for 10 h under vacuum and then stored in
a glovebox charged with high purity argon.
from *Ru(bpy)32+ to DCH-MV2+, as shown in Scheme 1. This
means that the overall thermodynamic driving force for the
reaction increases by 62% via simultaneous complexation of
(70) Coope, J. A. R.; Gardner, C. L.; McDowell, C. A.; Pelman, A. I. Mol.
Phys. 1971, 21, 1043.
(71) See the reference in ref 67.
Preparation of DCH-MV2+. As a means to prepare the viologen
with a bulky sidearm that can prevent the access of the viologen in
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7133

A R T I C L E S
Park et al.
Scheme 4. A Synthetic Procedure to Prepare
3-(Dicyclohexylmethyl)oxy-1-iodopropane (1c)
the temperature reached to about 130 °C, as a result of the consumption
of most of the low boiling methyl iodide in the reaction mixture. It
requires about 48 h to reach the temperature at which the reaction
mixture turns dark and fuming becomes vigorous. The tarry reaction
mixture was then cooled to room temperature and distilled water (50
mL) was added to the flask to remove polar byproducts. The residue
was extracted with diethyl ether three times (50 mL each). The diethyl
ether solution was dried using anhydrous MgSO4 and the solvent was
removed by evaporation. The crude product (1c) was isolated by column
chromatography (ethyl acetate:n-hexane ) 1:36): C16H29IO 364.31;
yield (6.13 g, 89%); 1H NMR (CDCl3, ppm) 3.51 (t, 2H), 3.31 (t, 2H),
2.66 (t, 1H), 2.06 (quintet, 2H), 1.80-1.14 (m, 22H).
Preparation of 1-Methyl-4-(4′-pyridyl) Pyridinium Iodide
(MPP+I-). Methyl iodide (1.7 mL, 28 mmol, 1.25 equiv) was added
to a dichloromethane solution of 4,4′-bipyridine (3.5 g, 22 mmol) and
the reaction mixture was refluxed overnight. The produced yellow
precipitate was isolated by filtration and subsequently purified by
recrystallization from methanol and diethyl ether: C11H11IN2 298.13;
1
yield (6.23 g, 95%); H NMR (DMSO, ppm) 9.27 (d, 2H), 8.98 (dd,
2H), 8.74 (d, 2H), 8.17 (t, 2H), 4.52 (s, 3H).
Preparation of DCH-MV2+(PF6-)2. Into an acetonitrile solution (50
mL) of MPP+I- (1.1 g, 3.7 mmol) was added 1c (1.48 g, 4.05 mmol,
1.1 equiv) and the solution was refluxed for 48 h. After cooling to
room temperature, the red precipitate [DCH-MV2+(I-)2] was isolated
by filtration and dissolved in distilled deionized water. The aqueous
solution (50 mL) of NH4PF6 (5 g) was added dropwise to the clean,
pale yellow solution of the iodide salt of DCH-MV2+. The precipitated,
white hexafluorophosphate salt of DCH-MV2+ was isolated by filtration
and recrystallized twice from acetonitrile and diethyl ether: C27H40F12-
N2OP2 698.56; yield (1.68 g, 65%); 1H NMR (CD3CN, ppm,) 8.99 (d,
2H), 8.89 (d, 2H), 8.40 (m, 4H), 4.79 (t, 2H), 4.45 (s, 3H), 3.68 (t,
2H), 2.71 (t, 1H), 2.33 (m, 2H), 1.80-1.14 (m, 22H)
Preparation of Ru(bpy)33+(PF6-)3. Ru(bpy)33+(PF6-)3 was synthe-
sized using the procedure reported by DeSimone and Drago64 by
employing. PbO2 as the oxidizing agent in 6 M H2SO4 aqueous solution.
The precipitated Ru(bpy)33+(PF6-)3 in the acidic solution was filtered
over a glass frit and the filtered green powder (microcrystals) was
immediately transferred into a glovebox charged with high purity argon.
The surface of the green powder turned reddish during the course of
transfer to the dry chamber. The reddish green powder (5 g) was
dissolved into the rigorously purified acetonitrile (30 mL). The color
of the solution was initially green but slowly turned orange with time.
Upon addition of rigorously purified diethyl ether (150 mL), the solution
slowly turned orange while green powder precipitates at the bottom of
the flask. After standing still for 4 h in the glovebox, the supernatant
solution was removed by decanting. The above procedure was repeated
four additional times until the supernatant solution consists of diethyl
ether and the acetonitrile was free of the orange hue due to the presence
of Ru(bpy)32+(PF6-)2. The collected dark green powder (0.5 g) was
kept in a glass vial. The UV-vis spectrum of the green powder
solution into the interior of the zeolite, 3-(dicyclohexylmethyl)oxy-1-
iodopropane (Scheme 4, 1c) was prepared with the intent to prepare
DCH-MV2+
.
Preparation of Dicyclohexylmethyloxyallyl Ether (Scheme 4, 1a).
The THF solution (90 mL) of dicyclohexylmethanol (7.46 g, 38.02
mmol) was added to a suspension of dry NaH (1.09 g, 45.60 mmol) in
THF (30 mL) with vigorous stirring at room temperature. After 1-h
reflux, the THF solution (30 mL) of allyl bromide (4.84 g, 40.01 mmol)
was introduced into the above mixture, and the reflux was allowed to
proceed for an additional 15 h with vigorous stirring. After cooling to
room temperature, the precipitated NaBr was filtered off from the
mixture and THF in the clear solution was removed by evaporation in
vacuo. Into the flask containing the residue was introduced diethyl ether
(200 mL) to dissolve the residue. The diethyl ether solution was washed
successively with brine (30 mL) and water (30 mL) to remove the
residual NaBr dissolved in the THF solution. The NaBr-free solution
was dried using anhydrous MgSO4 and concentrated in vacuo. The crude
product was isolated by column chromatography (ethyl acetate:n-hexane
) 1:36): C16H28O 236.39; yield (6.37 g, 71%); 1H NMR (CDCl3, ppm)
5.98 (m, 1H), 5.25 (d, 1H), 5.08 (d, 1H), 4.03 (d, 2H), 2.78 (t, 1H),
1.22-1.85 (m, 22H).
Preparation of 3-(Dicyclohexylmethyl)oxypropan-1-ol (1b). A
solution of BH3-THF (22 mL, 1 M, 22 mmol) was added dropwise to
a THF solution (10 mL) of 1a (6.37 g, 26.97 mmol). The clear, colorless
reaction mixture was stirred at room temperature for 30 min to complete
the reaction. Excess hydride was destroyed by the careful addition of
5 mL of water. After the cease of hydrogen evolution, 10 mL of aqueous
sodium hydroxide (3 M) was added to the reaction mixture and
hydrogen peroxide (10 mL, 30% aqueous solution) was added dropwise
to the stirred reaction mixture at a rate such that the temperature did
not exceed approximately 40 °C. The reaction mixture was subsequently
heated at 50 °C for 1 h to ensure complete oxidation. The reaction
mixture was cooled to room temperature and most of the solvent (THF
and water) was stripped off in vacuo. Distilled water (30 mL) was
introduced into the flask containing the residue and 1b was extracted
from the aqueous slurry using diethyl ether three times (50 mL each).
The diethyl ether solution was dried with anhydrous MgSO4, and
evaporation of the ethereal solution yielded 1b. The crude product was
purified by column chromatography (ethyl acetate:n-hexane ) 1:9):
C16H30O2 254.41; yield (5.60 g, 82%); 1H NMR (CDCl3, ppm) 3.82 (t,
2H), 3.71 (t, 2H), 2.68 (t, 1H), 1.14-1.95 (m, 25H).
dissolved in fresh acetonitrile did not show the characteristic 450-nm
3+
band of Ru(bpy)32+ but only the 416- and 674-nm bands of Ru(bpy)3
.
,
The corresponding extinction coefficients were 2230 and 410 M-1 cm-1
respectively.
PET from Zeolite-Encapsulated Ru(bpy)3 to DCH-MV2+ in
Solution. In a glovebox charged with high purity argon, a piece of
thin zeolite pellet (10 mg, d ) 7 mm) was placed on a piece of glass
plate (∼6 × 18 mm2), which was placed in an airtight fused silica
fluorescence cell. The dihedral angle between the supporting glass plate
and the flat bottom of the rectangular cell was about 60°. An aliquot
of an acetonitrile solution of each crown ether [2 mL, 8.1 mM, 30
equiv with respect to Ru(bpy)32+] was introduced into the cell, and an
additional 2 mL of acetonitrile solution of DCH-MV2+(PF6-)2 (13.5
mM, 50 equiv) was subsequently added. The airtight fluorescence cell
was then brought out of the glovebox and exposed to visible light
generated from a 200-W Hg lamp at the wavelengths between 430 and
2+
Preparation of 1c. Triphenyl phosphite (7.09 g, 22.85 mmol), methyl
iodide (4.05 g, 28.53 mmol), and 1b (4.83 g, 18.99 mmol) were charged
into a two-necked round-bottomed flask fitted with a Hg thermometer
and a reflux condenser connected to a drying tube charged with calcium
chloride. The neat reaction mixture was heated under gentle reflux until
9
7134 J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002

Long-Lived Charge-Separation
A R T I C L E S
470 nm using a 450-nm band-pass filter (Corion, P70-450-F), in a
homemade temperature-controlled irradiation chamber. The temperature
inside the chamber was controlled to 30 ( 1 °C. The solution was
gently stirred during irradiation with the help of a small magnetic stirring
bar placed under the tilted glass plate. The stirring did not cause
disintegration of the pellet into powder. The solution turned blue upon
irradiation, due to formation of DCH-MV•+, and the blue color
intensified with time. After each 1-h period of irradiation, the cell was
removed from the light source and the spectrum of DCH-MV•+ was
measured on a UV-vis spectrophotometer. The fluorescence cell was
placed in the cell holder in such way that the plane of the glass support
in the cell lies parallel to the direction of the incident beam, to minimize
the interference caused by the glass plate during the measurement of
the yield. The sample pellets did not interfere the passage of the
analyzing beam.
TET from Zeolite Framework to DCH-MV2+ in Solution. For
this experiment all of the Ru(bpy)32+-free M+Y zeolites with different
cations (either in the form of pellet or powder) were calcined separately
from one another at 550 °C for 10 h under flowing oxygen. Immediately
after calcination they were each transferred to a glass tube equipped
with a greaseless stopcock and further evacuated in each tube at 350
°C for 10 h before they were finally transferred into a glovebox, where
a pellet or an aliquot of each zeolite was introduced into a vial
containing an acetonitrile solution of DCH-MV2+(PF6-)2.
mM solution of DCH-MV2+(PF6-)2 in the glovebox under a dimmed
red light. The tightly capped vial was wrapped with aluminum foil to
prevent light from going into the vial. The acetonitrile suspension of
the vial was magnetically stirred for 20 h, and the zeolite powders were
allowed to sediment during the period of 4 h. Subsequently, 4 mL of
the clear supernatant solution was transferred to a fused silica UV-
vis cell equipped with an airtight cap. The cell was removed from the
glovebox for the spectral measurement.
Reaction of Ru(bpy)33+(PF6-)3 with Zeolite. Into a vial containing
-
5 mL of 1.3 mM acetonitrile solution of Ru(bpy)33+(PF6
) was
3
introduced 10 mg of dried M+Y (see Table 1) in the form of powder
in the glovebox. The tightly capped vial was wrapped with aluminum
foil to prevent light from going into the vial. The rest of the procedure
is the same with that employed to investigate the effect of countercation
on the thermal yield. In the case Cs+Y, which consumed all of the
Ru(bpy)33+ complex present in the solution, an additional 5-mL aliquot
of the Ru(III) solution was added into the vial containing the zeolite.
For the in situ measurement of the spectral change from Ru(III) to
Ru(II) in the presence of zeolite, a pellet of 10 mg of K+Y was placed
inside the fluorescence cell containing 4 mL of 0.65 mM acetonitrile
solution of Ru(bpy)33+(PF6-)3, and the cell was placed in the light-
protected compartment of a UV-vis spectrophotometer. The spectral
change of the supernatant solution was monitored by taking the
spectrum at a regular interval (1 h).
Quantitative Analysis of K+ Liberated from Zeolite to Solution.
For quantitative analysis of K+, the scale of the reaction was increased
by 10 and K+-Ru(bpy)32+Y was used in powder form. The blue
supernatant solution from the irradiated mixture was collected by
filtration in the glovebox. The solution was then taken out of the
glovebox and the solvent was removed by evaporation. Into the residual
solid was added 20 mL of 1 M aqueous NH4Cl solution to convert the
hexafluorophosphate (PF6-) salt of K+ and DCH-MV2+ into the
corresponding water-soluble chloride (Cl-) salts for the subsequent ICP-
AES analysis of potassium.
Instrumentation. The X-ray diffraction patterns for the identification
of the zeolite crystals were obtained from a Rigaku diffractometer (D/
MAX-1C) with the monochromatic beam of Cu KR. The UV-vis
spectra of the samples were recorded on a Shimadzu UV-3101PC. The
diffuse reflectance UV-vis spectra of solid samples were obtained using
an integrating sphere. The solution 1H and 13C NMR spectra were
recorded on a Varian Jemini 300 NMR spectrometer. Irradiation of
the samples was carried out using a 200-W super high-pressure Hg
lamp placed in an Oriel 66011 housing. The quantitative analysis of
K+ was carried out on a Shimadzu ICPS-1000 IV.
A General Procedure. An airtight, fused silica fluorescence cell
containing a piece of thin zeolite pellet (10 mg, d ) 7 mm) of K+-
Ru(bpy)32+Y, 2 mL of 8.1 mM acetonitrile solution of each CE [30
equiv with respect to the amount of Ru(bpy)32+], and 2 mL of 13.5
mM acetonitrile solution of DCH-MV2+(PF6-)2 (50 equiv) was prepared
similarly in the glovebox charged with high-purity argon. When CE is
not required, 2 mL of pure acetonitrile was introduced into the cell
instead of the CE solution. The cell was then placed in a water bath
placed in a dark chamber. Both the water bath and the supernatant
solution inside the cell were magnetically stirred with the aid of
appropriate magnetic bars. The desired temperature (20, 30, and 40
°C) of the bath was controlled using an immersion heater/cooler. The
spectral change of the supernatant solution was measured at ambient
temperature in the dark by removing the cell from the bath, and the
cell was placed back to the controlled temperature bath immediately
after spectral measurement. The spectral measurement was performed
every hour until the overall elapsed time reached 24 h. The surface of
the cell was briefly cleaned using a tissue wetted with ethanol prior to
each spectral measurement.
To obtain the effect of DCH-MV2+ concentration on the thermal
yield, a weighed amount of DCH-MV2+(PF6-)2 was introduced into
the cell containing a pellet of K+Y and 4 mL of 4.05 mM acetonitrile
solution of DCH-18C6. The amounts were 1.2 mg (3 equiv), 2.4 mg
(6 equiv), 3.6 mg (9 equiv), 4.7 mg (12 equiv), 9.4 mg (25 equiv),
18.8 mg (50 equiv), and 28.3 mg (75 equiv with respect to 10
supercages of K+Y).
Acknowledgment. We thank the Ministry of Science and
Technology (MOST), Korea, for supporting this work through
the Creative Research Initiatives (CRI) program. We are also
very grateful to the referees for the valuable comments and
corrections that greatly helped improve the quality of this paper.
To obtain the effect of countercation on the thermal yield, 10 mg of
M+Y (see Table 1) was introduced into a vial containing 5 mL of 5.4
JA020446A
9
J. AM. CHEM. SOC. VOL. 124, NO. 24, 2002 7135