Single-step electron transfer on the nanometer scale: Ultra-fast charge shift in strongly coupled zinc porphyrin-gold porphyrin dyads

Article abstract of DOI:10.1002/chem.200701335

The synthesis, electrochemical properties, and photoinduced electron transfer processes of a series of three novel zinc(II)-gold(III) bisporphyrin dyads (ZnP-S-AuP⁺) are described. The systems studied consist of two trisaryl porphyrins connected directly in the meso position via an alkyne unit to tert-(phenylenethynylene) or penta(phenylenethynylene) spacers. In these dyads, the estimated center to center interporphyrin separation distance varies from 32 to 45 A. The absorption, emission, and electrochemical data indicate that there are strong electronic interactions between the linked elements, thanks to the direct attachment of the spacer on the porphyrin ring through the alkyne unit. At room temperature in toluene, light excitation of the zinc porphyrin results in almost quantitative formation of the charge shifted state ⁺ZnP-S-AuP, whose lifetime is in the order of hundreds of picoseconds. In this solvent, the charge-separated state decays to the ground state through the intermediate population of the zinc porphyrin triplet excited state. Excitation of the gold porphyrin leads instead to rapid energy transfer to the triplet ZnP. In dichloromethane the charge shift reactions are even faster, with time constants down to 2 ps, and may be induced also by excitation of the gold porphyrin. In this latter solvent, the longest charge-shifted lifetime (τ = 2.3 ns) was obtained with the penta(phenylenethynylene) spacer. The charge shift reactions are discussed in terms of bridge-mediated super-exchange mechanisms as electron or hole transfer. These new bis-porphyrin arrays, with strong electronic coupling, represent interesting molecular systems in which extremely fast and efficient long-range photoinduced charge shift occurs over a long distance. The rate constants are two to three orders of magnitude larger than for corresponding ZnP-AuP⁺ dyads linked via mesophenyl groups to oligo-phenyleneethynylene spacers. This study demonstrates the critical impact of the attachment position of the spacer on the porphyrin on the electron transfer rate, and this strategy can represent a useful approach to develop molecular photonic devices for long-range charge separations.

Full text of DOI:10.1002/chem.200701335

FULL PAPER
DOI: 10.1002/chem.200701335
Single-Step Electron Transfer on the Nanometer Scale: Ultra-Fast Charge
Shift in Strongly Coupled Zinc Porphyrin–Gold Porphyrin Dyads
JØrôme Fortage,[a] Julien Boixel,[a] Errol Blart,[a] Leif Hammarstrçm,*[b]
Hans Christian Becker,*[b] and Fabrice Odobel*[a]
+
ꢀ ꢀ
C
C
Abstract: The synthesis, electrochemi-
cal properties, and photoinduced elec-
tron transfer processes of a series of
charge shifted state
ZnP S AuP ,
charge shift reactions are discussed in
terms of bridge-mediated super-ex-
change mechanisms as electron or hole
transfer. These new bis-porphyrin
arrays, with strong electronic coupling,
represent interesting molecular systems
in which extremely fast and efficient
long-range photoinduced charge shift
occurs over a long distance. The rate
constants are two to three orders of
whose lifetime is in the order of hun-
dreds of picoseconds. In this solvent,
the charge-separated state decays to
the ground state through the inter-
mediate population of the zinc porphy-
rin triplet excited state. Excitation of
the gold porphyrin leads instead to
rapid energy transfer to the triplet ZnP.
In dichloromethane the charge shift re-
actions are even faster, with time con-
stants down to 2 ps, and may be in-
duced also by excitation of the gold
porphyrin. In this latter solvent, the
longest charge-shifted lifetime (t=
2.3 ns) was obtained with the penta-
three novel zinc(II)–goldACHTRE(UGN III) bispor-
+
ꢀ ꢀ
phyrin dyads (ZnP S AuP ) are de-
scribed. The systems studied consist of
two trisaryl porphyrins connected di-
rectly in the meso position via an
alkyne unit to tert-(phenylenethyny-
lene) or penta(phenylenethynylene)
spacers. In these dyads, the estimated
center to center interporphyrin separa-
tion distance varies from 32 to 45 Š.
The absorption, emission, and electro-
chemical data indicate that there are
strong electronic interactions between
the linked elements, thanks to the
direct attachment of the spacer on the
porphyrin ring through the alkyne unit.
At room temperature in toluene, light
excitation of the zinc porphyrin results
in almost quantitative formation of the
magnitude larger than for correspond-
+
ꢀ
ing ZnP AuP dyads linked via meso-
phenyl groups to oligo-phenyleneethy-
nylene spacers. This study demon-
strates the critical impact of the attach-
ment position of the spacer on the por-
phyrin on the electron transfer rate,
and this strategy can represent a useful
approach to develop molecular photon-
ic devices for long-range charge separa-
tions.
(phenylenethynylene)
spacer.
The
Keywords: electron transfer · fem-
tosecond spectroscopy · gold · mo-
lecular wires · porphyrinoids · zinc
Introduction
[a] Dr. J. Fortage, J. Boixel, Dr. E. Blart, Dr. F. Odobel
UniversitØ de Nantes, Nantes Atlantique UniversitØs
CNRS
The development of molecular systems for long-range
charge separation has attracted considerable attention over
the past decades. One driving force for the growing interest
in this area is the realization that such molecular devices
can be used for a variety of applications, ranging from solar
energy conversion (artificial photosynthesis and photovolta-
ic) to molecular electronics.
Generally, the devices designed for light-induced charge
separation are composed of donors and acceptors covalently
linked via a molecular spacer.[1–7] The development of ap-
propriate spacers to promote electronic coupling is an im-
portant issue from the viewpoint of keeping a high rate of
electron transfer while increasing the distance between the
donor and the acceptor. Extensive theoretical analyses have
been devoted to this problem,[1,2,8] but surprisingly there are
few molecular devices for photoinduced charge separation
FacultØ des Sciences et des Techniques
Laboratoire de Synth›se Organique (LSO), UMR CNRS 6513
2, rue de la Houssini›re - BP 92208–44322 Nantes Cedex 3 (France)
Fax : (+33)2-51-12-54-29
E-mail: Fabrice.Odobel@univ-nantes.fr
[b] Prof. L. Hammarstrçm, Dr. H. C. Becker
Department of Photochemistry and Molecular Science
The Šngstrçm Laboratories, Uppsala University
Regementsvägen 1, 752 37 Uppsala (Sweden)
Fax : (+46)18-471 6844
E-mail: Leif.Hammarstrom@fotomol.uu.se
hcb@fotomol.uu.se
Supporting information for this article (synthesis of compounds,
1H NMR spectra of the dyads D1–D3 and of the compounds 25, 27,
29, 31, 32 and 33, spectroelectrochemical data, and additional transi-
ent absorption data) is available on the WWW under http://www.che-
mistry.org or from the author.
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3467

that rely on the utilization of long conjugated spacers to
connect the active units. Relevant examples are the para-
(phenylenevinylene) spacers used to connect a tetracene
sensitizer to a pyromellitimide acceptor described by Wasie-
lewski and co-workers.[5] More recently the use of a similar
spacer was reported by Guldi and co-workers to prepare
fullerene-based dyads.[9]
In the present work, we extend a concept that we have
previously reported, which involves anchoring the spacer di-
rectly on the porphyrin core to promote, in a single step, a
long-range photoinduced electron transfer between porphy-
rin units.[10,11] In a previous investigation, we showed that
the bis-ethynylenequaterthiophene spacer is suitable for
promoting fast energy transfer from a zinc porphyrin to a
free base porphyrin.[12] We are now interested in testing
whether this concept is appropriate for photoinduced elec-
tron transfer with oligo(phenyleneethynylene) spacers. To
this end, we replaced the free base porphyrin with a gold
porphyrin, because the couple ZnP/AuP+ is well suited for
photoinduced charge separation. Pioneering work by Sauv-
age and co-workers[13] with elegant rotaxane type architec-
tures, and more recently the groups of Albinsson and co-
workers,[6,14] and Fukuzumi and co-workers,[15] have shown
that efficient charge separation generally occurs in porphy-
would inevitably cause a decrease of the HOMO–LUMO
gap that would therefore create an accessible p–p* state lo-
calized on the bridge. As a consequence, energy transfer
from the photoexcited porphyrin to the bridge, which is in
close proximity to it, could compete with electron transfer
between the porphyrins. To circumvent this potential para-
site reaction, phenyleneethynylene spacers appear to be
better suited than oligothiophenes if one envisions using
very long spacers. para-(Phenyleneethynylene)s are linear,
relatively rigid p-conjugated molecules, and they exhibit rel-
atively high-lying electronic excited states, even when they
contain a large number of repeating units.[17]
In this work, para-(phenyleneethynylene)s were substitut-
ed by electron-withdrawing groups such as esters (D1), or
by electron-donating groups such as alkoxy chains (D2, D3)
(Scheme 1). These substituents were used not only to in-
crease the solubility but also to tune the electronic proper-
ties of the bridge.
This paper focuses on the synthesis, and the electrochemi-
cal and photoinduced electron transfer of novel dyads of
particular relevance to the above discussion. We show that
the properties of the dyads D1–D3 are not simple additions
of those of their molecular components, since electronic in-
teractions are clearly manifested in the ground and excited
states. We show that attachment of a para-(phenyleneethy-
nylene) spacer directly on the meso position of the porphy-
rin via a triple bond is an effective strategy to promote very
fast light-induced charge separation in a single step between
the porphyrins, over very long distances up to 45 Š. Indeed,
light excitation of the zinc porphyrin in dyads D1–D3 in di-
chloromethane leads to an almost quantitative electron
transfer reaction from ZnP to AuP+ on the picosecond time
scale.
rin arrays composed of zinc and gold porphyrins. Besides
+
ꢀ ꢀ
this, the system ZnP S AuP offers the possibility to study
the influence of the electronic properties of the spacer on
the charge separation by photoexcitation of either the zinc
or gold porphyrin, since both are known to bring about a
charge separation. In each case, the role of the spacer can
differ because the rate of the electron transfer, by a through
bond mechanism, is highly dependent on the energy match-
ing and the degree of overlap of the frontier molecular orbi-
tals of the photoexcited chromophore and the spacer.[2,5,8,16]
Furthermore, the utilization of long conjugated spacers
Scheme 1. Structures of the dyads at the focus of the present study.
3468
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
Results and Discussion
quantitative yield.[19] The iodoporphyrin 3 can be metallated
by zinc in a 94% yield using classical conditions. Iodo gold
porphyrin 6 was obtained in a modest yield (32%) using
Synthesis of the dyads: The preparation of dyads D1–D3
was accomplished according to a stepwise modular approach
in which the appropriate spacer is first linked to a zinc por-
phyrin and then to a gold porphyrin via a Sonogashira cross-
coupling reaction (Scheme 2). The key building blocks for
the synthesis of dyads D1–D3 are the zinc iodo porphyrin 4,
the differently protected bis-ethynyl spacer 18, 23, and 20
and the gold halogeno porphyrin 7, 8 (Scheme 2).
[AuACTHREU(NG tht)2]SbF6 in refluxing chloroform in presence of luti-
dine (tht=tetrahydrothiophene).[20] The low yield of this re-
action prompted us to test the classical conditions (HAuCl4
in AcOH + AcONa)[21] but they did not give the expected
porphyrin 6, but a mixture of different decomposition by-
products. Alternatively, the porphyrin 2 could be brominat-
ed with N-bromosuccinimide (NBS) and then submitted to
gold metalation using HAuCl4
in refluxing acetic acid with
sodium acetate. During the re-
action, the bromo group was
partly substituted by a chloro
group, leading to an equimolar
mixture of two porphyrins 7
and 8 in overall 75% yield.
These two porphyrins were not
separated, but used as a mix-
ture in the next step since
their reactivities proved to be
high with respect to the next
Sonogashira cross-coupling re-
action.
Scheme 2. Retrosynthetic strategy for the stepwise synthesis of dyads D1–D3.
Preparation of the spacers:
para-(Phenyleneethynylene)
spacers 19 and 21 were pre-
Preparation of the porphyrin building blocks: Starting with
the known diaryl porphyrin 1, the third aryl group was intro-
duced according to Sengeꢁs methodology by nucleophilic
substitution of one meso porphyrin proton by an excess of
phenyl lithium (Scheme 3).[18] This procedure presents the
advantage of being simpler than our previous methods[10,12]
and offers a high yield for the preparation of trisaryl por-
phyrin 2. The resulting porphyrin 2 is then iodinated using
bis(trifluoroacetoxy)iodobenzene at room temperature in a
pared according to the synthetic route reported in
Scheme 4. 1-Ethynyl-4-triisopropylsilylethylenyl benzene 11
was easily obtained through two successive Sonogashira
cross-coupling reactions starting from commercially avail-
able para-iodobromobenzene 9.[22] 1-ethynyl-4-trimethylsily-
lethylenyl benzene 14 was synthesized from para-bromoben-
zaldehyde 12 starting with a Sonogashira cross-coupling re-
action with trimethylsilyl acetylene followed by a Corey–
Fuchs reaction with overall 71% yield.[23] Dibromoterephtal-
ic acid was esterified with octa-
nol catalyzed by PTSA (para-
toluenesulfonic acid) in 87%
yield. The last step of the syn-
thesis consists of a double stat-
istical Sonogashira cross-cou-
pling reaction between the bis-
halogenophenyl units 16 or 17
and one equivalent of each
acetylenic derivative 11 and 14.
The separation of the three
products, formed during this
reaction, was easily achieved
by column chromatography
and afforded the desired
spacers 18 and 20 in about
50% isolated yield.
Finally, the preparation of
the long spacer 23 is described
Scheme 3. Synthesis of the porphyrin building blocks. Reagents and conditions: i) 6 equiv PhLi, THF, ꢀ788C,
then H2O and DDQ (92%); ii) I2, (CF3CO2)2IPh, CHCl3 (99%); iii) Zn(OAc)2, CH2Cl2, MeOH (94%);
iv) [Au
(tht)2]SbF6, 2,6-lutidine, CHCl3 (38%); v) NBS, CH2Cl2, ꢀ208C to RT. (90%); vi) HAuCl4·3H2O,
NaOAc, CH3CO2H, 1208C (75%). DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone.
AHCTREUNG
ACHTREUNG
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3469

L Hammarstrçm, H. C. Becker, F. Odobel et al.
Scheme 4. Synthesis of the para-tris(phenyleneethynylene) spacers 19 and 21. Reagents and conditions : i) trimethylsilylacetylene, PPh3, CuI, [Pd
CHCl3, THF, Et3N, iPr2NH, 08C, 1 h (98%); ii) triisopropylsilylacetylene, PPh3, CuI, [Pd2A(dba)3]-CHCl3, THF, Et3N, iPr2NH, 808C, 1 h (72%); iii) K2CO3,
MeOH, CH2Cl2, RT, 1 h (90%); iv) trimethylsilylacetylene, PPh3, CuI, [Pd2A(dba)3]-CHCl3, THF, Et3N, iPr2NH, 908C, 15 h (73%); v) CBr4, PPh3, Zn,
2ACHTREUNG(dba)3]-
CTHREUNG
CTHREUNG
1
CH2Cl2, RT, 15 h (100%); vi) LDA, THF, ꢀ808C, = h then aqueous NH4Cl, RT, 24 h (98%); vii) n-OctOH, APTS-H2O, toluene, 1108C, 20 h (87%);
2
viii) PPh3, CuI, [Pd2ACHTREUNG(dba)3]-CHCl3, THF, Et3N, iPr2NH, 808C, 24 h (41% for 18 and 55% for 21); ix) K2CO3, CH2Cl2, MeOH, RT (100%).
in Scheme 5. The iodo dimer partner 22 was obtained by the
coupling of 14 with 17 and could be separated from the sym-
metrical trimer which inevitably formed during this reaction.
The elongation of the trimer 21 was then realized by anoth-
er Sonogashira cross-coupling with 22 in 70% yield after
column chromatography.
The trimethylsilyl protecting group of the spacers 18, 20,
and 23 was cleaved with potassium carbonate in a quantita-
tive yield (Scheme 6 and 7).
to the spacer, because we observed that the cleavage of the
triisopropylsilyl (TIPS) group in presence of the gold por-
phyrin leads to the formation of by-products, which was not
the case with the zinc porphyrin. Most probably, the nucleo-
philic fluoride anion could react with the electron-deficient
gold porphyrin.[24] The zinc iodo porphyrin 4 was therefore
reacted with the appropriate spacer using different catalytic
systems (phosphine ligands such as tBu3P, (oTol)3P, AsPh3),
but the best conditions found were essentially those initially
described by Lindsey and co-workers (Scheme 6).[25]
Assembly of the building blocks: The final steps of the syn-
thesis involve connecting the porphyrins to the spacer in a
stepwise manner. The zinc iodoporphyrin was first coupled
The deprotection of the triisopropylsilyl group was carried
out with tetrabutylammonium fluoride, and the resulting
acetylenic derivatives were subjected to the final Sonoga-
Scheme 5. Synthesis of the spacers 29. Reagents and conditions: i) PPh3, CuI, [Pd
2ACHTERU(GN dba)3]-CHCl3, THF, Et3N, iPr2NH, 808C, 1 h (25%); ii) PPh3, CuI,
[Pd2A(dba)3]-CHCl3, THF, Et3N, iPr2NH, 808C, 12 h (70%); iii) K2CO3, CH2Cl2, MeOH, RT (100%).
CHTREUNG
3470
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
relies on a Sonogashira cross-
coupling reaction of the halo-
genogold porphyrin with the
appropriate spacer S1–S3 using
the proper catalytic conditions
discussed above (Scheme 7).
The porphyrin center-to-
center distance in the dyads
was estimated by molecular
mechanics calculations to be
32 Š for D1 and D2 and 45 Š
for D3 (see Supporting Infor-
mation for synthetic proce-
dures).
Electronic absorption spectra:
The UV/Vis absorption spectra
of the dyads D1–D3 along
Scheme 6. Synthesis of the dyads D1–D3. Reagents and conditions: i) AsPh3, [ PdCHTRE(UNG dba)3]-CHCl3, toluene, Et3N,
608C (19–36%); ii) Bu4NF, THF, RT (99%); iii) equimolar 7:8, CuI, [PdACHTREU(NG dppf)Cl2], Et3N, DMF, 40–608C
(50%)
2A
with those of the zinc and gold
tetraaryl porphyrins are shown
in Figure 1 and 2. Table 1 sum-
shira cross-coupling reaction with the mixture of halogeno-
gold porphyrins 7 and 8. Classical Sonogashira conditions
and Lindsey conditions were not suitable for this coupling
since they did not afford the expected coupled product.[25]
Such phenomena were previously observed with a gold por-
phyrin substrate by Mårtensson and co-workers.[6b] We
marizes the absorption data of the porphyrin systems de-
scribed in this work.
Qualitatively, the spectra of these series of dyads differ
from those of the reference tetraaryl porphyrins ZnTdBPP
found that a [PdACHTREUNG(dppf)Cl2] catalyst (dppf=diphenylphosphi-
noferrocene) with copper iodide in a mixture of triethyla-
mine and dimethylformamide proved to be particularly effi-
cient for this coupling, since the two porphyrins 7 and 8
were completely consumed after 5 h at 508C and the expect-
ed dyads D1–D3 were formed with 50% yields. During the
course of this work, Mårtensson and co-workers reported
ꢀ
that the classical catalyst [Pd2CAHTRE(UNG dba)3] CH3Cl (dba=trans,-
trans-dibenzylideneacetone) with AsPh3 is active for this
type of coupling provided that the polarity of the medium is
sufficiently high to ease the deprotonation of the terminal
alkyne proton.[26] However, these new conditions were not
tested, since they appeared in the literature after the com-
pletion of the synthesis of D1–D3. It is worth noting that
the ester groups of spacer 19 tend to hydrolyze during the
desylilation and the Sonogashira reactions explaining, thus,
the lower yields obtained with the spacer S1.
Figure 1. UV/Vis absorption spectra of D1–D3 (black, green, red, respec-
tively), along with those of the tetra(3,5-diterbutylphenyl) zinc(II)
The spacer–AuP+ systems 31–33 were also prepared as
references (Scheme 7). The synthesis of these molecules
(ZnTdBPP; blue) and tetra(3,5-diterbutylphenyl) porphyrin gold
ACHTREUNG(III)
(AuTdBPP+; magenta) recorded in dichloromethane.
Scheme 7. Synthesis of the reference porphyrin systems 31–33. Reagents and conditions: i) equimolar mixture of 7 and 8, CuI, [Pd
40–608C (40%).
ACHTRE(UNG dppf)Cl2], Et3N, DMF,
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3471

L Hammarstrçm, H. C. Becker, F. Odobel et al.
spacer and the nature of its substituents (alkoxy or ester)
have little effect on the electronic transitions of the porphy-
rin, since the spectra of all three dyads are quite similar in
the Soret and Q-bands regions (Figure 1). This indicates
that the average conjugation length of the phenyleneethyny-
lene in the ground state is not longer than three ethynylphe-
nylene units. It is important to note that the spacer unit of
all the dyads D1–D3 show p–p* absorptions that are blue-
shifted relative to the Soret band of the zinc and gold por-
phyrins (Table 1). This characteristic eliminates the risk of
energy transfer from the porphyrin to the spacer.
Electrochemistry: The half-wave redox potentials were de-
termined by square-wave voltammetry (frequency 15 Hz) in
dichloromethane and they are collected in Table 2.
Figure 2. UV/Vis absorption spectrum of D3 (red) along with those of its
components ZnP-S3 (29; black), S3-AuP+ (33; blue), and S3 (24; magen-
ta) recorded in dichloromethane.
Table 2. Redox potentials of the compounds prepared for this study. The
electrochemistry was recorded in dichloromethane with 0.15m of
Bu4NPF6 as supporting electrolyte and all potentials are reported versus
the saturated calomel electrode (SCE). When limiting values are given
the process occurred outside the limit of the electroactivity of the electro-
lyte in our conditions.
Table 1. UV/Vis absorption data of the compounds recorded in dichloro-
methane. zinc(II)
ZnTdBPP=tetra(3,5-diterbutylphenyl)
and
AuTdBPP+ =tetra(3,5-diterbutylphenyl)porphyrin gold
AHCTRE(UNG III).
Compound lmax [nm]/e [mꢀ1 cmꢀ1
]
S2: 20
256 (3.63”104); 275 (2.14”104); 323 (6.98”104); 384
(7.24”104)
ZnP+/
ZnP
[V]
ZnP2+
ZnP+
[V]
/
S+/S
[V]
AuP+/
AuP /
S/Sꢀ
[V]
C
ZnP-S2: 27 323 (8.01”104); 380 (6.04”104); 446 (5.04”105); 570
(2.34”104); 620 (4.96”104)
AuPꢀ
[V]
C
AuP
[V]
AuP-S2: 32 320 (4.59”104); 384 (5”104); 424 (1.91”105); 540
(1.63”104); 587 (2.22”104)
S1: 18
>1.6
>1.5
>1.5
>1.5
1.30
1.32
1.34
>1.4
1.26
1.27
ꢀ1.39
D2
317 (2.52”104); 447 (1.86”105); 545 (1.12”104); 586
(1.69”104); 619 (2.18”104)
ZnP-S1: 25
AuP-S1: 31
D1
0.77
0.76
0.76
0.74
0.73
0.74
0.68
1.20
1.16
1.18
1.18
1.14
1.14
1.03
<ꢀ1.3
<ꢀ1.4
ꢀ1.28
ꢀ0.56
ꢀ0.58
ꢀ1.06
ꢀ1.05
S1: 18
268 (3.61”104); 326 (4.21”104); 373 (5”104)
ZnP-S1: 25 262 (1.2”104); 323 (1.45”104); 367 (1.31”104); 447
(1.04”105); 567 (4.8”103); 620 (1.09”104)
S2: 20
<ꢀ1.6
<ꢀ1.3
<ꢀ1.4
ZnP-S2: 27
AuP-S2: 32
D2
AuP-S1: 31 328 (3.25”104); 384 (4.16”104); 427 (8.9”104); 440
(9.84”104); 543 (1.14”104); 585 (1.56”104)
ꢀ0.56
ꢀ0.58
ꢀ1.08
ꢀ1.05
D1
315 (2.77”104); 382 (2.93”104); 446 (2.29”105); 543
(1.28”104); 582 (2.07”104); 620 (2.28”04)
S3: 23
<ꢀ1.6
<ꢀ1.5
<ꢀ1.5
<ꢀ1.5
ZnP-S3: 29
AuP-S3: 33
D3
AuP ref
ZnP ref
S3: 23
314 (2.39”104); 330 (2.58”104); 398 (4.33”104)
1.28
1.30
ꢀ0.57
ꢀ0.56
ꢀ0.59
ꢀ1.06
ꢀ1.05
ꢀ1.15
ZnP-S3: 29 325 (6.05”104); 400 (8.35”104); 450 (2.97”105); 534
(4.9”103); 570 (1.36”104); 620 (2.93”104)
AuP-S3: 33 330 (5.2”104); 413 (1.41”105); 424 (1.54”105); 489
(2.65”104); 542 (1.33”104); 590 (1.91”104)
D3
280 (7.39”104); 294 (8.65”104); 324 (3.6”104); 427
(1.54”105); 447 (2.41”105); 545 (1.29”104); 586
(2.03”104); 618 (2.38”104)
The zinc porphyrin unit of all new compounds exhibits
two oxidation processes that are located around 0.75 V and
1.16 V, respectively. They correspond to the formation of
the radical cation and dication on the p-porphyrin ring.[29]
The gold porphyrin unit displays a first metal-based reduc-
tion at around ꢀ0.6 V followed by a porphyrin-based reduc-
tion around ꢀ1.1 V.[30] The spacers S2 and S3 bearing the
alkoxy substituents are relatively electron-rich and display
similar potentials for oxidation around 1.3 V (Table 2),
while their reductions occurred outside the solvent electro-
activity window. Conversely, the more electron-deficient
spacer S1 could not be oxidized below 1.6 V, while its poten-
tial for reduction could be determined as ꢀ1.39 V. Com-
pared to the tetraaryl porphyrins the attachment of the
spacer induces an increase of both zinc porphyrin potentials
and the second potential of the gold porphyrin. The poten-
tials for oxidation of the spacers in dyads D2 and D3 are
also anodically shifted by 40 mV compared to the isolated
bridges 20 and 23. These results are consistent with a stabili-
ZnTdBPP
423 (3.13”105); 550 (1.32”104); 590 (4.3”103)
AuTdBPP+ 416 (2.57”105); 524 (1.26”104); 564 (shoulder)
and AuTdBPP+ by a red-shift of the maximum absorption
positions of the Soret and Q-transitions and a broadening of
the Soret band in particular (Figure 1). The direct attach-
ment of the bridge to the meso-position extends the porphy-
rin p-conjugation, leading to a lower HOMO–LUMO gap
and bathochromic absorption shifts. Interaction with the
spacer in the ZnP–SX and AuP–SX mono-porphyrinic com-
pounds broadens the Soret band, presumably by splitting
the degeneracy of the x- and y-transitions as observed for
other meso-ethynyl substituted porphyrins.[7,11,12,27,28] The ab-
sorption spectra of the dyads above 400 nm are in good
ꢀ
ꢀ
agreement with a sum of the ZnP SX and AuP SX compo-
nents, indicating that the spectrum is not strongly affected
by inter-porphyrin interactions. Also, the length of the
3472
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
zation of the HOMO and LUMO orbitals of the porphyrin
and of the spacers, which is caused by the extended p-conju-
gation between the spacer and the porphyrins. On the other
hand, the negligible shift of the first reduction potential of
the gold porphyrin is in agreement with a metal-based pro-
cess that is naturally less affected by the appended bridge.[32]
(DG8ht), and finally of the subsequent back electron transfer
to reform the ground state reactants (DG8bet) As these are
charge shift reactions with one uncharged unit in both the
reactant and product states, the simplified Rehm-Weller
equation (Equations (1)–(3)) could be used (the columbic
interaction term equals zero).[31]
1
DGꢁ ¼ eðE0ðZnPþ=ZnPÞꢀE0ðAuPþ=AuPÞꢀE00ð ZnPÞÞ
Steady-state fluorescence spectroscopy: The steady-state
fluorescence spectra of the dyads D1–D3 and the parent
compounds ZnP–Spacer 25, 27, and 29 were recorded with
identical absorbance at the maximum absorption Q-band of
the zinc porphyrin.
ET
ð1Þ
3
DG0 ¼ eðE0ðZnPþ=ZnPÞꢀE0ðAuPþ=AuPÞꢀE00ð AuPþÞÞ
HT
ð2Þ
Both the fluorescence and phosphorenscence spectra are
red shifted in the compounds with the spacer attached, com-
pared to zinc and gold tetraaryl porphyrins, in analogy to
the absorption spectra. Again, the number of ethynylpheny-
lene units and the nature of the substituents on the spacer
play a minor role on the energy of the excited state of the
porphyrin units. The zinc porphyrin phosphorescence maxi-
mum could not be clearly determined and the triplet energy
was estimated to lie 0.46 eV below the corresponding sin-
glet, as for zinc tetraaryl porphyrin. It is noteworthy that the
zinc porphyrin fluorescence in all the dyads D1–D3, is
almost completely quenched by the appended gold porphy-
rin independently of the interporphyrin separation distance
(Table 3). In the current dyads, the fluorescence quenching
of ZnP originates from intramolecular electron transfer with
the gold porphyrin (see below).
DGꢁ
¼ eðE0ðAuPþ=AuPÞꢀE0ðZnPþ=ZnPÞÞ
ð3Þ
BET
Also, the driving force should be independent of solvent po-
larity for an ideal system in a dielectric continuum, with
compensatory charge redistributions on the two porphyrin
units. In practice, there may be small differences due to dif-
ferences in solvation of the ZnP+/0 and AuP+/0 species.[32]
Time-resolved fluorescence: The fluorescence decays of
ZnP references and ZnP/AuP+ dyads were recorded using a
streak camera/spectrograph combination (Hamamatsu,
<5 ps fwhm) and 400 nm, 100 fs excitation pulses (see Fig-
ure S20 in the Supporting Information). Excitation with
620 nm light yielded the same ZnP lifetimes, but because of
the small Stokes shift scattered light contaminated the
decays. The fluorescence lifetimes are summarized in
Table 5. It should be noted that the measured lifetimes for
Table 3. Emission characteristics of the porphyrins.
[e]
lem
(1ZnP*)
E
N
R
00A
[eV][d]
[nm][a]
[eV][b]
[nm][c]
Table 4. Free energy for electron transfer, hole transfer, and back elec-
tron transfer in the dyads D1–D3, calculated from Equations (1)–(3).
ZnP-S1: 25 635, 689
AuP-S1: 31
1.97
n.d.
768
(1.51)[f]
ACHTREUNG
1.61
D1
634, 691
1.97
1.97
0.01
0.02
0.04
Dyad
DG8ET
[eV]
DG8HT
[eV]
DG0
[eV]
BET
ZnP-S2: 27 636, 687
AuP-S2: 32
D2
ZnP-S3: 29 634, 687
AuP-S3: 33
D3
n.d.
768
ACHTREUNG
D1
D2
D3
ꢀ0.63
ꢀ0.65
ꢀ0.67
ꢀ0.28
ꢀ0.29
ꢀ0.31
ꢀ1.34
ꢀ1.32
ꢀ1.30
634, 687
1.97
1.97
n.d.
769
(1.51)[f]
ACHTREUNG
1.61
633, 686
1.97
AuP ref.
ZnP ref.
1.82[g]
1.64[g]
600, 650
2.10
[a] Recorded in dichloromethane at room temperature. [b] Determined
from the average of the 0–0 band maxima for absorption and fluores-
cence. [c] Recorded in ethanol (ZnP) and 2-methyl THF (AuP+) at 77 K.
[d] Determined from the phosphorescence band maximum recorded at
77 K. [e] Fem: ZnP fluorescence yield in the dyad relative to that of the
parent ZnP-SX, in dichloromethane at room temperature. [f] Estimated
(see text). [g] From Ref. [13a]
dyads D1 and D2 are comparable to the instrument
FWHM, and may possibly be shorter (see transient absorp-
ꢀ
tion results below). Compared to the ZnP SX references,
the ZnP emission lifetime in the dyads is strongly reduced
in both toluene and dichloromethane. This is consistent with
the low relative emission yields in Table 3. A minor compo-
nent of longer-lived emission is also present, which we tenta-
tively attribute to small quantities of ZnP lacking the AuP
moiety remaining after purification, or having formed
during irradiation. By combining the fluorescence yield and
lifetime data we can conclude that the level of fluorescent
impurity is <2% in all dyads.
Calculation of the reaction free energy for photoinduced
charge separation in the dyads: The redox potentials along
with the emission data allow for the determination of the
free energy changes of the photoinduced electron transfer
from the zinc porphyrin singlet excited state to the gold por-
phyrin (DG8et), of the photoinduced hole transfer from the
gold porphyrin triplet excited-state to the zinc porphyrin
Femtosecond transient absorption spectroscopy: Transient
spectra after Q-band excitation of the reference compounds
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3473

L Hammarstrçm, H. C. Becker, F. Odobel et al.
as well as the AuP+ unit, both in toluene and dichlorome-
thane. We analyzed the transient absorption data using a
global fitting strategy. The wavelengths were chosen to rep-
resent significant spectral features in the bleaching and ex-
cited state absorption regions, and to minimize contributions
from the solvent. Light-induced electron transfer from the
ꢀ
ꢀ
ZnP S1 and S1 AuP in toluene and dichloromethane are
shown in Figure 3. The spectra for the other references are
virtually identical to those shown, and the dynamics are
3
1ZnP unit and hole transfer from the AuP+ unit were ob-
served, both leading to the charge-shifted (CSh) product
+
ꢀ
ꢀ
C
C
ZnP SX AuP , as well as subsequent back electron trans-
fer. Kinetic data for these processes are summarized in
Table 6.
Table 5. Fluorescence lifetimes of dyads D1–D3 and the corresponding
ZnP-SX references.
Dyad
Solvent
tf [ps] (tf ZnP-SX)
D1
D1
D2
D2
D3
D3
CH2Cl2
toluene
CH2Cl2
toluene
CH2Cl2
toluene
7 (1800)
3 (1800)
6 (1800)
7 (1800)
40 (1600)
390 (1600)
+
C
To estimate the expected transient spectrum of the ZnP
ꢀ
ꢀ
C
SX AuP state, spectroelectrochemistry difference spectra
+
ꢀ
ꢀ
for oxidation of ZnP SX and reduction of SX AuP were
recorded (see Figure S21 in the Supporting Information).
These show mainly bleaching of the ZnP and AuP+ absorp-
Figure 3. Transient absorption spectra of ZnP-S1 (top) and AuP-S1
(bottom) in toluene (left) and dichloromethane (right). The pump wave-
length was 620 nm for ZnP-S1 and 580 nm for AuP-S1.
tions. Some new, broad absorption bands appear, due
+
C
mainly to the ZnP species around 480 and 690 nm, but
also a narrow difference absorption feature around 460 nm
1
very similar. The initially populated AuP+ state of gold por-
due to the formation of AuP . These additional absorption
C
phyrins is known to convert to the lowest AuP + state on a
features are very similar to the excited state features, which
is typical for this type of porphyrin.[13] However, inspection
of the transient absorption decays reveal the formation of
intermediates in the dyads. The spectral changes in the
Soret- and Q-band regions show that the intermediate in
both ET and HT pathways involves the porphyrin moiety
that was not excited (ZnP for HT, AuP+ for ET pathways).
Moreover, the ET intermediate has a transient absorption
around 460 nm that is somewhat blue-shifted compared to
any of the locally excited states observed, consistent with
3
sub-picosecond time scale.[13,33] We observe small-amplitude
dynamics on the time scale of a few picoseconds for both
1
3
the ZnP and AuP states, tentatively attributed to vibration-
al and solvent relaxation, but apart from this, the transient
spectra decay cleanly. As is typical for Zn-porphyrins, the
main decay pathway of the ZnP singlet excited state is inter-
system crossing to the long-lived (@10 ns) ZnP triplet with
an observed singlet lifetime of around 2 ns. The Q-band
bleaches are distinctly different for the two chromophores,
1
with bleach minima at approximately 565 and 620 nm for
the CSh state spectrum. Finally, comparison with ZnP emis-
+
ꢀ
ꢀ
ZnP S1 and at 545 and 585 nm for S1 AuP . Also the
Soret bleach occurs at somewhat shorter wavelengths for
the AuP chromophore. These signatures are useful for inter-
preting the transient spectra of the dyads below. In line with
the thermodynamics for charge and energy transfer, there
are no indications that the bridge is involved in any electron
or energy transfer reactions in these reference compounds.
sion lifetimes allows us to kinetically identify the ET pro-
cess. Transient absorption traces for D1 ZnP and AuP exci-
tation are shown in Figure 4, and the associated transient
absorption spectra in Figure 5. Additional data is shown in
the Supporting Information.
ZnP excitation—high polarity: Electron transfer can be ach-
ieved with high selectivity by exciting the ZnP moiety at
620 nm. Careful analysis of the transients obtained for dyads
D1 and D2 reveals the formation of an intermediate, attrib-
utable to the CSh state, with a time constant of 2 ps for both
dyads, similar to that for the fluorescence decay (see also
Figure S20 in the Supporting Information). The somewhat
slower kinetics deduced from the fluorescence experiments
3
Interestingly, the AuP state is significantly more long-lived
(7–10 ns) in toluene than in dichloromethane (ꢂ2 ns). Also,
the magnitude of transient absorption above 600 nm is much
larger compared to the other features in dichloromethane
than in toluene, which was seen also in the dyads below.
In the dyads D1–D3, we observe multi-exponential transi-
ent absorption kinetics following excitation of the ZnP unit
3474
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
Figure 5. Transient absorption spectra for D1. Top spectra are for excita-
tion of the ZnP unit (620 nm pump), lower spectra are for excitation of
predominantly the AuP+ unit (580 nm pump), left spectra are in toluene,
right spectra in dichloromethane. Black: ꢀ10 ps, light blue: 0.6 ps, vio-
let:5 ps, green:50 ps, red: 500 ps. The 0.6 ps spectrum is chirp-corrected.
See Supporting Information for corresponding spectra for D2 and D3.
Figure 4. Transient absorption decays for D1. Top traces are for excita-
tion of the ZnP unit (620 nm pump), lower traces are for excitation of
predominantly the AuP+ unit (580 nm pump), left traces are in toluene,
right traces in dichloromethane. Violet: 422 nm blue: 440 nm, dark green
(toluene only): 465 nm, light green: 490 nm, red: 690 nm. Solid lines are
best global bi- or triexponential fits to multiple wavelengths. See Sup-
porting Information for corresponding traces for D2 and D3.
quent back electron transfer is slower, with a time constant
of 2.1 ns.
can be attributed to the limited temporal resolution of the
streak camera. The Soret band bleach recovers partially on
this time scale, as expected from the spectral differences of
ZnP excitation—low polarity: The results obtained with se-
lective ZnP excitation in toluene are initially very similar to
those in dichloromethane. Thus, an intermediate attributable
to the CSh state is formed with approximately the same
time constant as the fluorescence decay, namely 8 ps for D1,
and 7 ps for D2. This is only slightly slower than in the more
polar solvent. For the subsequent back electron transfer the
results are instead qualitatively quite different. The CSh
state disappears with time constants of 240 ps (D1) and
130 ps (D2), but the product has the clear spectral character-
istics of the 3ZnP state and shows no decay on the time
scale of experiment (6 ns). The rise and decay behavior of
the signal at 440 nm (Figure 4, top left panel) illustrates
1
+
C
the ZnP and ZnP species. On the blue side, however,
there is no recovery because also the AuP+ transition be-
comes bleached. Also at the AuP+ Q-band positions a fur-
ther bleaching appears. Concomitantly, the absorption fea-
ture around 470 nm blue-shifts, and a weak but significant
absorption increase is seen at 690 nm, all consistent with for-
mation of the charge-shifted state. Subsequently, the transi-
ent features of the CSh state decay with time constants of
90 ps (D1) and 80 ps (D2), consistent with back electron
transfer to reform the ground state reactants. In addition we
observe a residual 2–3 ns transient, with a spectrum resem-
bling that of the ZnP triplet, although the lifetime is shorter
than expected. The photostability is lower in dichlorome-
thane than in toluene (see below), and the presence of the
nanosecond component can be due to degradation. We note
that the degree of fluorescent ZnP impurity in the freshly
prepared solutions was very low (see above) and cannot ac-
count for these signals. Finally, also in the longer dyad D3
we observe very rapid ET with a time constant of 90 ps, in
this case over a 45 Š center-to-center distance. The subse-
1
clearly the dynamics of the reaction sequence ZnP ! CSh
3
! ZnP, showing that the triplet is indeed formed via the
1
CSh state. By comparison of the ZnP–3ZnP absorption dif-
ꢀ
ference in the dyad and the reference ZnP SX, we conclude
that the ZnP formation is approximately quantitative. Thus
3
the rate of back electron transfer to form the ground state is
much slower than the observed reaction leading to the trip-
let state, and much slower than the back electron transfer in
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3475

L Hammarstrçm, H. C. Becker, F. Odobel et al.
dichloromethane. Formation of a long-lived 3ZnP state is ob-
viously direct evidence that this state is lower in energy than
the CSh state, at least in toluene. In the dyad D3 the ET re-
action is somewhat slower, with a time constant of 370 ps,
similar to the fluorescence decay time. As for D1 and D2
this is about four times slower than the corresponding pro-
cess in dichloromethane. The lifetime of the CSh state in D3
is also comparatively long, around 12 ns. In contrast to the
shorter dyads, back electron transfer seems to reform both
or 130 ps (D2). As in dichloromethane, some dynamics due
to ZnP excitation were also observed, but no other inter-
3
mediates were observed. The magnitude of the ZnP signal
formed is consistent with quantitative, direct triplet energy
3
transfer from AuP+ to ZnP.
In contrast, no energy transfer is observed in the long
dyad D3 in toluene. Only small amplitude dynamics with a
roughly 400 ps time constant was observed, attributable to
1
ET from the fraction of ZnP created. The main transient
3
3
the ground state and the ZnP state in ratio of approximate-
appears to be due to the AuP+ state decaying with a life-
3
+
ꢀ
ly 2:1 (judging from the ZnP signal amplitude).
time of several nanoseconds as in the SX AuP reference.
AuP+ excitation—high polarity: Although not as selective
as ZnP excitation, pumping at 590 nm predominantly targets
AuP+. This is expected to result in HT from the excited
Discussion of the of charge shift processes: A simplified
state energy diagram is given in Figure 6. Formation of the
CSh state in dichloromethane is an exergonic process both
1
AuP+ to ZnP. Since the singlet excited AuP+ undergoes in-
tersystem crossing in a few hundred femtoseconds,[33] the
HT rates cannot be determined by our fluorescence meas-
urements. Fortunately, the transient absorption spectra of
3
D1 show the decay of AuP + and formation of an intermedi-
ate species with time constants of 2 ps in both cases, fol-
lowed by its disappearance with time a constant of 170 ps.
The identification of the intermediate as the CSh based on
spectral changes is less obvious than with ZnP excitation,
partly because the less selective excitation gives some ZnP
bleach features already from the beginning. Nevertheless,
the blue shift of the 460 nm absorption is seen. As we ob-
Figure 6. Energies of the excited and charge-shifted states in toluene (tol)
and dichloromethane (DCM).
3
via ET and HT mechanisms, and the ZnP state clearly lies
above the CSh state. From the discussion around the data in
Table 4 one would expect the energetic situation to be very
similar in toluene, and it has even been reported that the
+
ꢀ
serve almost total reformation of the ground state much
CSh states of ZnP AuP dyads are stabilized in non-polar
+
faster than in the SX AuP reference, the dynamics cannot
solvents.[32] In contrast, we see back electron transfer in tolu-
ene that quantitatively forms the ZnP state. Thus the CSh
ꢀ
3
be explained solely by the fraction of ZnP-excited dyads.
Also the faster kinetics compared to the reference shows
that the reaction of the AuP+-excited dyad involves the
ZnP moiety. Finally, this intermediate is observed only in
the dyads, it is spectrally different from the ZnP triplet, and
the spectra at 20–50 ps are identical to those obtained with
ZnP excitation. Thus we assign it to the CSh state formed
state must lie at least 0.2 V higher in toluene than in di-
chloromethane. The reason could possibly be ion pairing of
the AuP+ in toluene so that the effective charge actually in-
creases upon reduction, which would decrease the reduction
potential. As the 3ZnP state is estimated to lie only
3
100 meV below the AuP+ state, the lack of HT from the
3
by HT from AuP +. The decay traces on the >100 ps time
latter in toluene may be because the process has a very
weak driving force or is even endergonic.
scale are very similar to those after 620 nm excitation, show-
ing that back electron transfer is rather independent on
whether the CSh state was formed from a singlet or triplet
excited state (see below). Also, the residual 2–3 ns transient,
attributed to partial decomposition, appears just as upon
620 nm excitation above.
The very rapid ET and HT processes over several nano-
meters are remarkable and they are clearly mediated by the
bridge. In the super-exchange mechanism virtual states in-
volving the bridge increase the electronic coupling.[2,5,8,18]
This can involve either the HOMO or LUMO of the bridge,
In contrast, for D2 and D3 we see only small transient
amplitude changes with a 2 ps (D2) or 23 ps (D3) time con-
stant, at least partially due to the fraction of ZnP excitation,
Table 6. Time constants from transient absorption experiments. Note
that the main product of back electron transfer in toluene is the 3ZnP
state, while it is the ground state in dichloromethane.
3
but no clear evidence for a CSh state formed from AuP+.
Instead the spectrum is very similar to the initial 3AuP+
state and decays with a time constant of about 2 ns, very
Dyad
Solvent
ZnP excitation
AuP+ excitation
tET
tBET
tHT
tBET
3
+
[ps]
[ps]
[ps]
[ps]
170
ꢀ
close to the life time of the SX AuP reference. We see
substantial degradation of these dyads with electron-rich
bridges during measurements with 590 nm excitation in di-
chloromethane. Thus, we cannot determine the HT time
constants in these cases.
D1
D1
D2
D2
D3
D3
CH2Cl2
toluene
CH2Cl2
toluene
CH2Cl2
toluene
2
8
2
7
80
240
90
130
2
(170[a]
)
)
[b]
ACHTREUNG
[b]
85
370
2100
ꢂ12000
[b]
[b]
[b]
[b]
AuP+ excitation—low polarity: In toluene the 3AuP+ is con-
[a] Time constant for excitation energy transfer from 3AuP+ to 3ZnP.
[b] Not observed (see text).
3
verted to the ZnP state with a time constant of 170 ps (D1)
3476
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
thus defining HT and ET, respectively. Figure 7 shows a rela-
tive orbital energy diagram involving the bridges. This is
based on the pairing theorem, stating that the HOMO–
The estimated ZnP–bridge LUMO gap in the present series
varies only from 0.34 to 0.39 eV between the electron-rich
and electron-deficient bridges (Figure 8). This can explain
why we observed no difference for ET in D1 and D2. Also
the AuP+–bridge HOMO gaps are similar, and is in fact
largest in D1 where we did observe a very rapid HT. Thus,
in the absence of sample decomposition and parallel excita-
tion of the ZnP unit, we would have expected to observe a
similarly rapid HT process in D2.
We estimate the reorganization energy two ways. First, it
is given by Equation (5),[35] where we take the inner reor-
ganization energy li to be 0.2 eV.[13a–d,6a,b]
ꢀ
ꢁꢀ
ꢁ
e2
1
1
1
1
1
l ¼ li þ lo ¼ li þ
þ
ꢀ
ꢀ
n2 es
4pe0 2RD 2RA 2RDA
ð5Þ
Figure 7. Relative orbital energy diagram of the donor, bridge, and ac-
ceptor calculated from redox data in dichloromethane and corresponding
E00 values.
For the dyads D1–D3, using RD =RA =4.8 Š, RCC =33 Š for
D1 and D2, and RCC =45 Š for D3, in dichloromethane we
estimate the total reorganization energy to be about 1.2 eV.
For the charge shift via the ET mechanism in D3, DG0 =
ꢀ0.67 eV. From these values we predict an activation energy
of 58 meV, which is close to the observed value of 28 meV
above. Starting instead from the experimental activation
energy and driving force, Equation (4) gives a value of l=
1.0 eV Using the value of 28 meV in Equation (4) we esti-
mate the electronic coupling to be about 150 cmꢀ1 for ET in
D1 and D2, and about 20 cmꢀ1 in D3.
LUMO energy difference of an organic molecule is equal to
the first excitation energy.[34] Thus, for each porphyrin and
bridge unit the relative energy for either the HOMO or
LUMO is given by the potential for either the first oxidation
or first reduction. The other frontier orbital energy is then
given by adding or subtracting the singlet excitation energy.
ꢀ
This is an approximate procedure, used before for ZnP
AuP+ dyads.[14b] The electrochemical potentials give equilib-
rium free energy differences for relaxed states, while virtual
bridge states are never populated and the relevant energy
should be for the geometry of the neutral bridge. Neverthe-
less, the diagram should be useful to discuss the differences
between the dyads and the different processes. First, the dia-
gram shows that charge shift via electron or hole hopping
onto the bridge is unlikely, as the bridge states lie approxi-
mately 0.3–0.4 eV higher in energy than the electron and
hole donor states. Indeed, we measured the activation
energy of ET in D3 in dichloromethane, where the energet-
ics are most in favor of hopping. From the fluorescence life-
time at temperatures between 7 and 348C we calculate the
activation energy to be very small, only 0.028 eV, which is
clearly too low for a hopping mechanism. Since the shorter
links disfavor hopping we can conclusively say that ET
occurs via superexchange in D1–D3, and that this probably
also holds for HT in D1.
It may be surprising that the rates of ET and HT for D1
in dichloromethane are very similar, as the driving force for
HT is much smaller, DG0 =ꢀ0.28 eV. The close similarity in
+
rate is however typical for ZnP AuP dyads.[13] As the reor-
ꢀ
ganization energy is not expected to be significantly differ-
ent for ET and HT, this implies that a stronger electronic
coupling for the HT process compensates for the smaller
driving force. However, the origin of stronger coupling is
not easily rationalized by inspection of frontier orbitals, (see
Ref. [14b]). With l=1.0–1.2 eV the predicted activation
energy for HT is 0.13–0.18 eV, and Equation (4) implies a
coupling that is four to ten times that for ET. While this is
not impossible, we note that already the coupling strength
estimated for ET puts this reaction on the limit of an adia-
batic reaction. For a solvent-controlled adiabatic electron
transfer reaction the pre-exponential factor of Equation (4)
should be replaced by (16pkbT/l)1/2/tL, which for lꢂ1.0 eV
at room temperature becomes ꢂ1/tL.[8] Thus, the reaction
rate is then controlled by the longitudinal relaxation time of
the solvent rather than by the electronic coupling. We note
that the time constants for ET and HT in D1 are very close
to the solvent relaxation times obtained by dynamic Stokes
shift measurement: <t>=0.56 ps for dichloromethane and
<t>=2.72 ps for toluene.[36] Finally, the rapid ET and HT
reactions in D1 and D2 occur on a time scale comparable to
relaxation observed in the reference compounds. To con-
clude, Equation (4) may not be perfectly valid in these
cases, which may explain why the difference in driving force
for ET and HT does not result in significantly different rates.
For a single-step reaction such as these the rate constant
(kET) should follow a Marcus equation for non-adiabatic
electron transfer,[8,35] as given in Equation (4), where l is the
reorganization energy and VDA the electronic coupling. The
coupling is mediated by the bridge, and is inversely propor-
tional to the energy difference between the donor and
bridge states.[2,5,8,16]
ꢀ
ꢁ
2
4p2
ꢀðDGꢁ þ lÞ
4lkBT
2
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
kET
¼
jVDAj exp
ð4Þ
h
4plkBT
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3477

L Hammarstrçm, H. C. Becker, F. Odobel et al.
With this in mind, the somewhat slower ET in toluene
than in dichloromethane could be due to slower solvent dy-
namics, and/or due to a higher free energy of activation. The
reorganization energy should be very small in toluene as n2
ꢂ es[36] ([Eq. (2)]) so that ET occurs in the Marcus inverted
region, but at the same time the driving force is smaller
than in dichloromethane. We can not estimate these param-
eters for toluene with meaningful precision, but as the acti-
vation energy is very small in dichloromethane it is likely to
be higher in toluene.
Back electron transfer (BET) is much slower than solvent
controlled and should follow the non-adiabatic Equation (4).
It is slower in toluene than in dichloromethane, and as the
main product is 3ZnP, BET to the ground state is even
slower. This is most readily explained by the low reorganiza-
tion energy in toluene that puts BET to the ground state far
into the inverted region. In dichloromethane instead BET to
the ground state should be near activationless, based on the
parameter estimates above. BET to form a triplet excited
chromophore is a phenomenon observed also in photosyn-
thetic reaction centers, and it has been investigated and suc-
cessfully exploited in a series of papers by Wasielewski and
coworkers.[37] This opens up the possibility of controlling the
product formation, and also the lifetime, of the product,
using the polarity of the medium in which the photoinduced
reaction takes place. One application could be to use the
long-lived ZnP triplet as a reservoir for photon energy if the
energy levels can be tuned so that the triplet-charge-shift
gap is small. Unfortunately, the solubilities of the dyads pre-
clude a more comprehensive investigation of the solvent de-
pendence in the present dyads.
While ET would result in a singlet CSh state, HT from
3AuP+ should form the corresponding triplet state. While
the subsequent BET in D1 is somewhat slower upon AuP+
excitation (170 ps, versus 80 ps after ZnP excitation) the dif-
ference is not dramatic and given the complexity of data
analysis perhaps not significant. Thus, it seems that the pres-
ence of the heavy gold ion induces sufficient singlet–triplet
mixing in the CSh state to make spin effects negligible, as
has been observed before.
Figure 8. Distance dependence for ET and BET in the present dyads (cir-
cles) and those of reference [14b] (squares). Solid lines are best linear
fits, with slopes (-b) of ꢀ0.31 (ET) and ꢀ0.39 (BET). Filled symbols: ET
1
from ZnP to AuP+; open symbols: BET from the charge shifted state to
form the ground state. Dashed lines are a visual aid.
ꢀ
ET and BET as a function of distance in a series of ZnP
AuP+ dyads, linked by oligo-ethynylphenylene bridges via a
meso-phenyl group of the porphyrins.[14b] For comparison,
we also plot our data for the present series of dyads in di-
chloromethane. The important structural difference between
the series is that D1–D3 are linked with an ethynyl group di-
rectly on the porphyrin core. The effect on the observed
rate is dramatic, with rates three orders of magnitude larger
for both ET and BET in D1–D3. As Albinsson and co-work-
ers estimated activation energies very similar to ours, the
effect is due to a much stronger electronic coupling in our
dyads. Rotation of the cylindrically symmetric p-system of
polyethynyls should not alter the electronic coupling. Thus,
the main mechanism for reducing the coupling in oligo-ethy-
nylphenylenes is probably rotation of the phenyls. If we
compare data for dyads of the two series in Figure 8 that
have the same number of phenyl groups between the por-
pyrin cores (that is, D1 and D2 with the points at 26.5 Š in
the lower series, and D3 with the points at 40.3 Š) we still
find that the rates are about two orders of magnitude higher
in D1–D3. A possible rationale for this is the hindered rota-
tion of meso-phenyl groups in the weakly coupled system,
which for steric reasons prefer to be oriented orthogonally
to the porphyrin plane.[6,14,16] This is probably the main de-
coupling factor as rotation of the other phenyls should be
unhindered. It is obvious from the perturbed ground state
properties that the direct ethynyl link couples the bridge
very strongly to the porphyrin, and the excited and radical
states are most likely partially delocalized onto the
bridge.[11,12,27a] As discussed above, however, the ground
state perturbation does not seem to extend beyond three
ethynylphenylene units. In line with this observation, the
rate constants for ET and BET decrease about as steeply
with distance between D1/D2 and D3 as in the more weakly
coupled series of Reference [14b]. This may suggest that
while ET in the shorter dyads is on the adiabatic limit, ET
in D3 and BET in all dyads follow the exponential decrease
of rate with distance as expected for weakly coupled systems
It is further interesting to compare the rate constants for
ET and back electron transfer (BET) in this strongly cou-
+
ꢀ
pled system with data from other ZnP AuP dyads. The
rate constant for ET in D1–D3 is 20–40 times larger than for
BET, which is a favorable property for efficient and yet
long-lived photoinduced charge separation, and is a long-
standing target of artificial photosynthesis. These values
+
ꢀ
seems to be typical for a range of published ZnP AuP
dyads with very different bridges and geometries, which all
give a kET/kBET ratio of 10–100.[13,14] The strong coupling ob-
tained by direct coupling of the ethynyl group to the por-
phyrin meso-position obviously gives a similar effect for the
excited state ET and HT reaction as for the ground state
BET.
Albinsson and co-workers have made extensive, quantita-
tive investigations of bridge-mediated electron transfer in
bis-porphyrins.[6,14] In Figure 8 we plot the rate constants for
3478
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480

Electron Transfer in Porphyrin Dyads
FULL PAPER
[2] V. Balzani, F. Scandola, Supramolecular Photochemistry, Ellis Hor-
wood, Chischester, 1991.
(non-adiabatic reaction). With only two data points, and ET
rates similar to those for excited state relaxation, we refrain
from reporting b values for our systems.
[3] a) M. R. Wasielewski, Chem. Rev. 1992, 92, 435–461; b) D. Gust,
T. A. Moore, A. L. Moore, Acc. Chem. Res. 1993, 26, 198–205; c) D.
Gust, T. A. Moore, A. L. Moore, S. J. Lee, E. Bittersmann, D. K.
Luttrull, A. A. Rehms, J. M. DeGraziano, X. C. Ma et al., Science
1990, 248, 199–201; d) J. H. Alstrum-Acevedo, M. K. Brennaman,
T. J. Meyer, Inorg. Chem. 2005, 44, 6802–6824; e) E. Baranoff, J.-P.
Collin, L. Flamigni, J.-P. Sauvage, Chem. Soc. Rev. 2004, 33, 147–
155.
[4] a) D. M. Guldi, H. Imahori, K. Tamaki, Y. Kashiwagi, H. Yamada,
Y. Sakata, S. Fukuzumi, J. Phys. Chem. A 2004, 108, 541–548; b) H.
Imahori, Y. Sekiguchi, Y. Kashiwagi, T. Sato, Y. Araki, O. Ito, H.
Yamada, S. Fukuzumi, Chem. Eur. J. 2004, 10, 3184–3196; c) H. Im-
ahori, Y. Sekiguchi, Y. Kashiwagi, T. Sato, Y. Araki, O. Ito, H.
Yamada, S. Fukuzumi, Chem. Eur. J. 2004, 10, 3184 ꢀ3196; d) H.
Imahori, K. Tamaki, D. M. Guldi, C. Luo, M. Fujitsuka, O. Ito, Y.
Sakata, S. Fukuzumi, J. Am. Chem. Soc. 2001, 123, 2607–2617.
[5] W. B. Davis, M. A. Ratner, M. R. Wasielewski, J. Am. Chem. Soc.
2001, 123, 7877–7886; W. B. Davis, W. A. Svec, M. A. Ratner, M. R.
Wasielewski, Nature 1998, 396, 60–63; R. T. Hayes, M. R. Wasielew-
ski, D. Gosztola, J. Am. Chem. Soc. 2000, 122, 5563–5567.
[6] a) K. Pettersson, J. Wiberg, T. Ljungdahl, J. Mårtensson, B. Albins-
son, J. Phys. Chem. B 2006, 110, 319–326; b) K. Kilsa, J. Kajanus,
A. N. Macpherson, J. Mårtensson, B. Albinsson, J. Am. Chem. Soc.
2001, 123, 3069–3080; c) ; K. Kilsa, J. Kajanus, S. Larsson, A. N.
Macpherson, J. Mårtensson, B. Albinsson, Chem. Eur. J. 2001, 7,
2122–2133; d) M. U. Winters, K. Pettersson, J. Mårtensson, B. Al-
binsson, Chem. Eur. J. 2005, 11, 562–573.
[7] M. U. Winters, E. Dahlstedt, H. E. Blades, C. J. Wilson, M. J. Framp-
ton, H. L. Anderson, B. Albinsson, J. Am. Chem. Soc. 2007, 129,
4291–4297.
[8] a) Molecular Electronics (Eds.: J. Jortner, M. Ratner), Blackwell,
London, 1997; b) M. Bixon, J. Jortner, Adv. Chem. Phys. 1999, 106,
35–202; c) P. F. Barbara, T. J. Meyer, M. A. Ratner, J. Phys. Chem.
1996, 100, 13148–13168.
[9] a) F. Giacalone, J. L. Segura, N. Martin, J. Ramey, D. M. Guldi,
Chem. Eur. J. 2005, 11, 4819–4834; b) S. A. Vail, P. J. Krawczuk,
D. M. Guldi, A. Palkar, L. Echegoyen, J. P. C. Tome, M. A. Fazio,
D. I. Schuster, Chem. Eur. J. 2005, 11, 3375–3388; c) F. Giacalone,
J. L. Segura, N. Martin, D. M. Guldi, J. Am. Chem. Soc. 2004, 126,
5340–5341; d) G. De La Torre, F. Giacalone, J. L. Segura, N. Martin,
D. M. Guldi, Chem. Eur. J. 2005, 11, 1267–1280.
[10] a) F. Odobel, F. Suzenet, E. Blart, J.-P. Quintard, Org. Lett. 2000, 2,
131–133; b) E. Blart, F. Suzenet, J.-P. Quintard, F. Odobel, J. Por-
phyrins Phthalocyanines 2003, 7, 207–213.
[11] C. Monnereau, J. Gomez, E. Blart, F. Odobel, S. Wallin, A. Fallberg,
L. Hammarstrçm, Inorg. Chem. 2005, 44, 4806–4817.
[12] F. Odobel, S. Suresh, E. Blart, Y. Nicolas, J.-P. Quintard, P. Janvier,
J.-Y. Le Questel, B. Illien, D. Rondeau, P. Richomme, T. Haupl, S.
Wallin, L. Hammarstrçm, Chem. Eur. J. 2002, 8, 3027–3046.
[13] a) A. M. Brun, A. Harriman, V. Heitz, J. P. Sauvage, J. Am. Chem.
Soc. 1991, 113, 8657–8663; b) A. M. Brun, S. J. Atherton, A. Harri-
man, V. Heitz, J. P. Sauvage, J. Am. Chem. Soc. 1992, 114, 4632–
4639; c) J. C. Chambron, A. Harriman, V. Heitz, J. P. Sauvage, J.
Am. Chem. Soc. 1993, 115, 6109–6114; d) A. Harriman, V. Heitz,
J. C. Chambron, J. P. Sauvage, Coord. Chem. Rev. 1994, 132, 229–
234; e) A. Harriman, V. Heitz, J. P. Sauvage, J. Phys. Chem. 1993,
97, 5940–5946; f) J. C. Chambron, S. Chardon-Noblat, A. Harriman,
V. Heitz, J. P. Sauvage, Pure Appl. Chem. 1993, 65, 2343–2392;
g) J. C. Chambron, A. Harriman, V. Heitz, J. P. Sauvage, J. Am.
Chem. Soc. 1993, 115, 7419–7425; h) J.-P. Collin, A. Harriman, V.
Heitz, F. Odobel, J.-P. Sauvage, J. Am. Chem. Soc. 1994, 116, 5679–
5690; i) L. Flamigni, I. M. Dixon, J.-P. Collin, J.-P. Sauvage, Chem.
Commun. 2000, 2479; j) I. M. Dixon, J.-P. Collin, J.-P. Sauvage, L.
Flamigni, Inorg. Chem. 2001, 40, 5507–5517; k) M. Andersson, M.
Linke, J.-C. Chambron, J. Davidsson, V. Heitz, J.-P. Sauvage, L.
Hammarstrçm, J. Am. Chem. Soc. 2000, 122, 3526–3527; l) M. An-
It has recently been pointed out that OPE linkers are not
as rigid as they may appear.[38] However, even in view of
these results, thermal fluctuations at room temperature
should change the end-to-end distance of the bridge by less
than 10%. Since the donor–acceptor coupling at these dis-
tances is entirely through-bond, rotation of the phenylene
and pophyrin planes are likely to give a much larger effect.
In principle the bending fluctuations will also affect the
energy of the bridge, but as the D–B and B–A energy gap is
much larger than kT we believe that the effect is negligible.
Conclusions
We have reported the synthesis of three new porphyrin
dyads consisting of a zinc porphyrin linked to a gold porphy-
rin via a long and rigid conjugated spacer anchored directly
on the meso position through an ethynylene group. These
new systems exhibit strong porphyrin-bridge electronic cou-
pling as evidenced by absorption and emission spectra and
electrochemical potentials. This leads to very rapid electron
and hole transfer processes upon light excitation of one of
the porphyrin units. The rate constants for these processes
are two to three orders of magnitude higher than for very
+
ꢀ
similar ZnP AuP dyads linked via meso-phenyl groups.
These results give evidence for highly efficient electron
transfer mediation by the oligo(ethynylphenylene) spacers
when they are directly connected to the porphyrin ring via a
triple bond, since very rapid electron transfer can be ach-
ieved over distances as long as 45 Š. We[11] and others[27a]
have previously reported ET from ethenylphenylene-bridg-
ed Zn-porphyrin-acceptor systems, but although the bridge
was much shorter the rates were slower than for D1 and D2
of the present study. Our results also underscore the utility
of connecting a conjugated spacer directly on the porphyrin
core to develop porphyrin arrays for long-distance photoin-
duced charge separation (see Ref [7]). This result opens up
new design strategies for vectorial electron transfer over
very long distances with porphyrin systems, with a view to
applications for solar energy transformations and molecular
electronics.
Acknowledgements
We gratefully acknowledge financial support for this project from The
French Research Ministry through the programs “ACI Jeune Chercheur”
and the ANR blanc “PhotoCumElec”, The Swedish Foundation for Stra-
tegic Research, The K&A Wallenberg Foundation, The Swedish Energy
Agency and The European Commission (COST D35).
[1] V. Balzani, P. Piotrowiak, M. A. J. Rodgers, J. Mattay, D. Astruc,
H. B. Gray, J. Winkler, S. Fukuzumi, T. E. Mallouk, Y. Haas, A. P.
de Silva, I. Gould, Electron Transfer in Chemistry, Wiley-VCH,
Weinheim, 2001.
Chem. Eur. J. 2008, 14, 3467 – 3480
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3479

L Hammarstrçm, H. C. Becker, F. Odobel et al.
dersson, M. Linke, J.-C. Chambron, J. Davidsson, V. Heitz, L. Ham-
marstrçm, J.-P. Sauvage, J. Am. Chem. Soc. 2002, 124, 4347–4362.
[14] a) J. Andreasson, G. Kodis, T. Ljungdahl, A. L. Moore, T. A. Moore,
D. Gust, J. Mårtensson, B. Albinsson, J. Phys. Chem. A 2003, 107,
8825–8833; b) J. Wiberg, L. Guo, K. Pettersson, D. Nilsson, T.
Ljungdahl, J. Mårtensson, B. Albinsson, J. Am. Chem. Soc. 2007,
129, 155–163.
[15] a) M. J. Crossley, P. J. Sintic, J. A. Hutchison, K. P. Ghiggino, Org.
Biomol. Chem. 2005, 3, 852–865; b) K. Ohkubo, P. J. Sintic, N. V.
Tkachenko, H. Lemmetyinen, W. E. Z. Ou, J. Shao, K. M. Kadish,
M. J. Crossley, S. Fukuzumi, Chem. Phys. 2006, 326, 3–14.
[16] M. P. Eng, B. Albinsson, Angew. Chem. 2006, 118, 5754–5757;
Angew. Chem. Int. Ed. 2006, 45, 5626–5629.
[17] a) P. F. H. Schwab, M. D. Levin, J. Michl, Chem. Rev. 1999, 99,
1863–1934; b) U. H. F. Bunz, Chem. Rev. 2000, 100, 1605–1644;
c) R. E. Martin, F. Diederich, Angew. Chem. Int. Ed. 1999, 38, 1351–
1377; d) J. M. Tour, J. S. Schumm, D. L. Pearson, Angew. Chem.
1994, 106, 1445–1448; Angew. Chem. Int. Ed. Engl. 1994, 33, 1360–
1363; e) J. M. Tour, Acc. Chem. Res. 2000, 33, 791–804; f) Y. Zhao,
Y. Shirai, A. D. Slepkov, L. Cheng, L. B. Alemany, T. Sasaki, F. A.
Hegmann, J. M. Tour, Chem. Eur. J. 2005, 11, 3643–3658.
[18] a) M. O. Senge, Acc. Chem. Res. 2005, 38, 733–743; b) M. O. Senge,
W. W. Kalisch, I. Bischoff, Chem. Eur. J. 2000, 6, 2721–2738; c) X.
Feng, M. O. Senge, J. Chem. Soc. Perkin Trans. 1 2001, 1030–1038.
[19] R. W. Boyle, C. K. Johnson, D. Dolphin, J. Chem. Soc. Chem.
Commun. 1995, 527–528.
[20] J. C. Chambron, V. Heitz, J. P. Sauvage, New J. Chem. 1997, 21, 237–
240.
[21] a) E. B. Fleischer, A. Laszlo, Inorg. Nucl. Chem. Lett. 1969, 5, 373–
376; b) J. W. Buchler, Synthesis and properties of metalloporphyrins,
Vol. 1 , Academic Press, New York, 1978.
[22] A. Godt, J. Org. Chem. 1997, 62, 7471–7474.
[23] J.-F. Nierengarten, S. Zhang, A. Gegout, M. Urbani, N. Armaroli, G.
Marconi, Y. Rio, J. Org. Chem. 2005, 70, 7550–7557.
[24] H. Segawa, R. Azumi, T. Shimidzu, J. Am. Chem. Soc. 1992, 114,
7564–7565.
[25] R. W. Wagner, T. E. Johnson, F. Li, J. S. Lindsey, J. Org. Chem. 1995,
60, 5266–5273.
ien, J. Am. Chem. Soc. 1996, 118, 11854–11864; c) R. Shediac,
M. H. B. Gray, H. T. Uyeda, R. C. Johnson, J. T. Hupp, P. J. Angiolil-
lo, M. J. Therien, J. Am. Chem. Soc. 2000, 122, 7017–7033.
[28] a) H. L. Anderson, Chem. Commun. 1999, 2323–2330; b) H. L. An-
derson, Inorg. Chem. 1994, 33, 972.
[29] a) R. H. Felton, Primary redox reactions of metalloporphyrins,
Vol. 5 , Academic Press, New York, 1978; b) A. Wolberg, J. Mana-
ssen, J. Am. Chem. Soc. 1970, 92, 2982–2991; c) J. Fajer, D. C. Borg,
A. Forman, D. Dolphin, R. H. Felton, J. Am. Chem. Soc. 1970, 92,
3451–3459.
[30] Z. Ou, K. M. Kadish, W. E. J. Shao, P. J. Sintic, K. Ohkubo, S. Fuku-
zumi, M. J. Crossley, Inorg. Chem. 2004, 43, 2078–2086.
[31] D. Rehm, A. Weller, Isr. J. Chem. 1970, 8, 259–271; A. Weller, Z.
Phys. Chem. 1982, 133, 93–98.
[32] S. Fukuzumi, K. Ohkubo, W. O. E, Z. , J. Shao, K. M. Kadish, J. A.
Hutchison, K. P. Ghiggino, P. J. Sintic, M. J. Crossley, J. Am. Chem.
Soc. 2003, 125, 14984–14985.
[33] a) J. Andreasson, G. Kodis, S. Lin, A. L. Moore, T. A. Moore, D.
Gust, J. Mårtensson, B. Albinsson, Photochem. Photobiol. 2002, 76,
47–50; b) M. P. Eng, T. Ljungdahl, J. Andreasson, J. Mrtensson, B.
Albinsson, J. Phys. Chem. A 2005, 109, 1776–1784.
[34] W. B. Davis, M. A. Ratner, M. R. Wasielewski, Chem. Phys. 2002,
281, 333–346.
[35] R. A. Marcus, N. Sutin, Biochim. Biophys. Acta Rev. Bioenerg. 1985,
811, 265–322.
[36] a) L. Reynolds, J. A. Gardecki, S. J. V. Frankland, M. L. Horng, M.
Maroncelli, J. Phys. Chem. 1996, 100, 10337–10354; b) M. L. Horng,
J. A. Gardecki, A. Papazyan, M. Maroncelli, J. Phys. Chem. 1995,
99, 17311–17337.
[37] a) K. Hasharoni, H. Levanon, S. R. Greenfield, D. J. Gosztola, W. A.
Svec, M. R. Wasielewski, J. Am. Chem. Soc. 1995, 117, 8055–8056;
b) M. R. Wasielewski, J. Org. Chem. 2006, 71, 5051–5066, and refer-
ences therein.
[38] a) A. Godt, M. Schulte, H. Zimmermann, G. Jeschke, Angew. Chem.
2006, 118, 7722–7726; Angew. Chem. Int. Ed. 2006, 45, 7560–7564;
b) B. D. Allen, A. C. Benniston, A. Harriman, L. J. Mallon, C. Paria-
ni, Phys. Chem. Chem. Phys. 2006, 8, 4112–4118.
[26] T. Ljungdahl, K. Pettersson, B. Albinsson, J. Mårtensson, J. Org.
Chem. 2006, 71, 1677–1687.
[27] a) N. P. Redmore, I. V. Rubtsov, M. J. Therien, J. Am. Chem. Soc.
2003, 125, 8769–8778; b) S. M. LeCours, S. G. DiMagno, M. J. Ther-
Received: August 27, 2007
Published online: February 12, 2008
3480
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3467 – 3480