Long-range electron transfer in porphyrin-containing [2]-rotaxanes: Tuning the rate by metal cation coordination

Article abstract of DOI:10.1021/ja0119907

A series of [2]-rotaxanes has been synthesized in which two Zn(II)-porphyrins (ZnP) electron donors were attached as stoppers on the rod. A macrocycle attached to a Au(III)-porphyrin (AuP⁺) acceptor was threaded on the rod. By selective excitat

Full text of DOI:10.1021/ja0119907

Published on Web 03/27/2002
Long-Range Electron Transfer in Porphyrin-Containing
[2]-Rotaxanes: Tuning the Rate by Metal Cation Coordination
Mikael Andersson,† Myriam Linke,‡ Jean-Claude Chambron,‡ Jan Davidsson,†
Vale´rie Heitz,‡ Leif Hammarstro¨m,*,† and Jean-Pierre Sauvage*,‡
Contribution from the Department of Physical Chemistry, Uppsala UniVersity, Box 532,
S-751 21 Uppsala, Sweden, and Laboratoire de Chimie Organo-Mine´rale,
UMR 7513 au C.N.R.S., Institut Le Bel, UniVersite´ Louis Pasteur, 4, Rue Blaise Pascal,
F-67070 Strasbourg, France
Received August 17, 2001. Revised Manuscript Received November 8, 2001
Abstract: A series of [2]-rotaxanes has been synthesized in which two Zn(II)-porphyrins (ZnP) electron
donors were attached as stoppers on the rod. A macrocycle attached to a Au(III)-porphyrin (AuP+) acceptor
was threaded on the rod. By selective excitation of either porphyrin, we could induce an electron transfer
from the ZnP to the AuP+ unit that generated the same ZnP•+-AuP• charge-transfer state irrespective of
which porphyrin was excited. Although the reactants were linked only by mechanical or coordination bonds,
electron-transfer rate constants up to 1.2 ×1010 s-1 were obtained over a 15-17 Å edge-to-edge distance
between the porphyrins. The resulting charge-transfer state had a relatively long lifetime of 10-40 ns and
was formed in high yield (>80%) in most cases. By a simple variation of the link between the reactants,
viz. a coordination of the phenanthroline units on the rotaxane rod and ring by either Ag+ or Cu+, we could
enhance the electron-transfer rate from the ZnP to the excited 3AuP+. We interpret our data in terms of an
enhanced superexchange mechanism with Ag+ and a change to a stepwise hopping mechanism with Cu+,
2+
involving the oxidized Cu(phen)2 unit as a real intermediate. When the ZnP unit was excited instead,
electron transfer from the excited 1ZnP to AuP+ was not affected, or even slowed, by Ag+ or Cu+. We
discuss this asymmetry in terms of the different orbitals involved in mediating the reaction in an electron-
and a hole-transfer mechanism. Our results show the possibility to tune the rates of electron transfer between
noncovalently linked reactants by a convenient modification of the link. The different effect of Ag+ and Cu+
on the rate with ZnP and AuP+ excitation shows an additional possibility to control the electron-transfer
reactions by selective excitation. We also found that coordination of the Cu+ introduced an energy-transfer
reaction from 1ZnP to Cu(phen)2+ (k ) 5.1 × 109 s-1) that proceeded in competition with electron transfer
3
to AuP+ and was followed by a quantitative energy transfer to give the ZnP state (k ) 1.5 × 109 s-1).
1. Introduction
Small structural changes in the protein could affect the electronic
coupling along the different paths, thereby controlling the rate
of the electron-transfer reactions. As an example, evidence has
been found that the second quinone acceptor, QB, of purple
bacterial reaction centers moves ca. 4.5 Å upon the first
reduction,4 which is probably important for subsequent reaction
steps. Also, in cytochrome bc1 large motions of the [2Fe2S]
domain have been shown necessary for the electron-transfer
In the photosynthetic reaction center of purple bacteria,
photoinduced charge separation proceeds via a series of electron-
transfer steps between porphyrin-like components and quinones.1
The reactant molecules are held together by the protein without
being covalently linked. In general, the electronic coupling
between electron donor and acceptor molecules is mediated by
the intervening medium to different extents, depending on its
electronic structure. Extensive studies of synthetic model
systems with a variable reactant distance have demonstrated the
efficiency of through-bond mediation,2 in particular through
covalent bonds. A protein medium has a more complicated
structure. Multiple electron-transfer pathways may exist, com-
prising covalent bonds, hydrogen bonds, space jumps, etc.3
(2) (a) Oevering, H.; Paddon-Row, M. N.; Heppener, M.; Oliver, A. M.;
Cotsaris, E.; Verhoeven, J. W.; Hush, N. S. J. Am. Chem. Soc. 1987, 109,
3258. (b) Johnson, M. D.; Miller, J. R.; Green, N. S.; Closs, G. L. J. Phys.
Chem. 1989, 93, 1173. (c) Helms, A.; Heiler, D.; McLendon, G. J. Am.
Chem. Soc. 1992, 114, 6227. (d) Finckh, P.; Heitele, H.; Volk, M.; Michel-
Beyerle, M. E. J. Phys. Chem. 1988, 92, 6584. (e) Ribou, A. C.; Launay,
J. P.; Takahashi, K.; Nihira, T.; Tarutani, S.; Spangler, C. W. Inorg. Chem.
1994, 33, 1325. (f) Wasielewski, M. R.; Johnsson, D. G.; Svec, W. A.;
Kersey, K. M.; Cragg, D. F.; Minsek, D. W. In Photochemical Energy
ConVersion; Norris, J. R., Meisel, D., Eds.; Elsevier: New York, 1989.
(3) For different opinions, see: (a) Beratan, D. N.; Betts, J. N.; Onuchic, J. N.
Science 1991, 252, 1285. (b) Beratan, D. N.; Onuchic, J. N.; Winkler, J.
R.; Gray, H. B. Science 1992, 258, 1740. (c) Gray, H. B.; Winkler, J. R.
Annu. ReV. Biochem. 1996, 65, 537. (d) Moser, C. C.; Keske, J. M.;
Warncke, K.; Farid, R. S.; Dutton, L. Nature 1992, 355, 796. (e) Page, C.
C.; Moser, C. C.; Xiaoxi, C.; Dutton, L. Nature 1999, 402, 47.
* To whom correspondence should be addressed. L.H.: Leifh@fki.uu.se.
J.-P.S.: sauvage@chimie.u-strasbg.fr.
† Uppsala University.
‡ Universite´ Louis Pasteur.
(1) (a) Anoxygenic Photosynthetic Bacteria; Blankenship, R. E., Madigan, M.
T., Bauer, C. E., Eds.; Kluwer Academic Publishers: Dordrecht, 1995. (b)
Diesenhofer, J.; Michel, H. Angew. Chem., Int. Ed. Engl. 1989, 28, 829.
(c) Huber, R. Angew. Chem., Int. Ed. Engl. 1989, 28, 848.
(4) Stowell, M. H. B.; McPhillips, T. M.; Rees, D. C.; Soltis, S. M.; Abresch,
E.; Feher, G. Science 1997, 276, 812.
9
10.1021/ja0119907 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 4347-4362
4347

A R T I C L E S
Andersson et al.
function of the protein.5 However, for electron transfer over
longer distances, additional redox cofactors or redox active
amino acid residues such as tyrosine often act as real intermedi-
ates in a charge-hopping mechanism,3e to reach sufficient rates.
There is currently a great interest in the mediation mechanism
in different systems6-8 and the distinction between super-
exchange9 and a charge-hopping mechanism. In a super-
exchange, the mediating states of the link only induce a small
perturbation, while in a hopping mechanism, the intermediates
are transiently populated which can lead to a very efficient
electron transfer over long distances.6-8
In contrast to the case in proteins, relatively few synthetic
systems exhibiting long-range electron transfer have employed
links other than covalent links to assemble the donor and
acceptor. Hydrogen bonds have attracted attention for this
purpose10 due to their possible importance in proteins, and also
coordination bonds have been employed.11 Recently, donor/
acceptor interactions and pure mechanical links have been
used.13 Also, a few of these systems have used molecular
mobility to enhance charge separation and suppress back
reactions.
Figure 1. Principle of transition metal-templated synthesis of a [2]-rotaxane.
A thick line represents a dpp chelate, a black dot represents a metal cation,
a hatched diamond represents a Au(III) porphyrin, and an empty diamond
represents a Zn(II) porphyrin. The transition metal controls the threading
of Au(III) porphyrin-pendant macrocycle A onto chelate B, to form
prerotaxane C. Construction of the porphyrin stoppers at the X functions
leads to the metal complex [2]-rotaxane D. Removal of the template cation
forms the free rotaxane E.
In the present paper, we present a series of [2]-rotaxanes
(Figure 4) that exhibits photoinduced electron transfer between
Zn(II)-porphyrin (ZnP) and Au(III)-porphyrin (AuP+) units.
Rotaxanes are described as a ring threaded on a rod with
stoppers that prevent dissociation. In the present case, the ZnP
electron donors also act as stoppers on the rod, and the AuP+
acceptor is attached to the ring. Thus, in Zn2//Au+ the redox
active porphyrins are held together only by mechanical bonds,
and no covalent electron-transfer pathway exists. By binding
different metal cations to the phenanthrolines, forming Zn2/
Ag+/Au+, Zn2/Cu+/Au+, and Zn2/Li+/Au+, the rod and ring
become connected by coordination bonds. There is no readily
identifiable single pathway for electron transfer between the ZnP
and AuP+ units, and all possible pathways include link segments
with different electronic structure and a jump through space
Figure 2. Precursors to the synthesis of [2]-rotaxanes.
and/or coordination bonds, somewhat resembling the case in
proteins. Some mobility of the rotaxane units is also possible,
which might affect the electron-transfer rates.
The aim of this study was to investigate the possibility of
enhancing the electron-transfer rates by coordination of the metal
cations and to compare the effects of space jumps versus metal-
ligand bonds in the electron-transfer pathway. Coordination of
the metal cations will increase the electronic coupling between
the phenanthrolines considerably, change the energy of their
molecular orbitals, and also introduce new, metal-based orbitals.
Thus, if the electron-transfer pathway involves the metal-
phenanthroline complex moiety, one would expect a significant
increase in reaction rate.
(5) Darrouzet, E.; Valkova-Valchanova, M.; Moser, C. C.; Dutton, L.; Daldal,
F. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 4567.
(6) Bixon, M.; Jortner, J. AdV. Chem. Phys. 1999, 106, 35.
(7) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R. Nature
1998, 396, 60.
(8) (a) Lewis, F. D.; Liu, X.; Liu, J.; Miller, S. E.; Hayes, R. T.; Wasielewski,
M. R. Nature 2000, 406, 51. (b) Schuster, G. B. Acc. Chem. Res. 2000,
33, 253. (c) Bixon, M.; Giese, B.; Wessely, S.; Langenbacher, T.; Michel-
Beyerle, M. E.; Jortner, J. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11713.
(d) Meggers, E.; Michel-Beyerle, M. E., Giese, B. J. Am. Chem. Soc. 1998,
120, 12950. (e) Lewis, F. D.; Wu, T.; Zhang, Y.; Letsinger, R. L.;
Greenfield, S. R.; Wasielewski, M. R. Science 1997, 277, 673.
(9) McConnell, H. M. J. Chem. Phys. 1961, 35, 508.
(10) (a) Turro, C.; Chang, C. K.; Leroi, G. E.; Cukier, R. I.; Nocera, D. G. J.
Am. Chem. Soc. 1992, 114, 4013. (b) Berman, A.; Izraeli, E. S.; Levanon,
H.; Wang, B.; Sessler, J. L. J. Am. Chem. Soc. 1995, 117, 8252. (c) Kirby,
J. P.; Roberts, J. A.; Nocera, D. G. J. Am. Chem. Soc. 1997, 119, 9230. (d)
Yang, J.; Seneviratne, D.; Arbatin, G.; Andersson, A. M.; Curtis, J. C. J.
Am. Chem. Soc. 1997, 119, 5329. (e) Harriman, A.; Kubo, Y.; Sessler, J.
L. J. Am. Chem. Soc. 1992, 114, 388. (f) Cukier, R. I.; Nocera, D. G. Annu.
ReV. Phys. Chem. 1998, 49, 337. Ghaddar, T. H.; Castner, E. W.; Isied, S.
J. Am. Chem. Soc. 2000, 122, 1233. (g) Osuka, A.; Yoneshima, R.; Shiratori,
H.; Okada, T.; Taniguchi, S.; Mataga, N. Chem. Commun. 1998, 1567.
(11) (a) Harriman, A.; Sauvage, J.-P. Chem. Soc. ReV. 1996, 41. (b) Harriman,
A.; Ziessel, R. Chem. Commun. 1996, 1707. (c) Dixon, I. M.; Collin, J. P.;
Sauvage, J.-P.; Barigelletti, F.; Flamigni, L. Angew. Chem., Int. Ed. 2000,
39, 1292. (d) Imahori, H.; Yoshizawa, E.; Yamada, K.; Hagiwara, K.;
Okada, T.; Sakata, Y. J. Chem. Soc., Chem. Commun. 1995, 11, 1133.
(12) (a) Linke, M.; Chambron, J.-C.; Heitz, V.; Sauvage, J.-P. J. Am. Chem.
Soc. 1997, 119, 11329. (b) Ashton, P. R.; Balzani, V.; Kocian, O.; Prodi,
L.; Spencer, N.; Stoddart, J. F. J. Am. Chem. Soc. 1998, 120, 11190. (c)
Hu, Y.-Z.; Bossmann, S. H.; van Loyen, D.; Schwartz, O.; Du¨rr, H. Chem.-
Eur. J. 1999, 5, 1267. (d) Hu, Y.-Z.; Tsukiji, S.; Shinkai, S.; Oishi, S.;
Du¨rr, H.; Hamachi, I. Chem. Lett. 2000, 442.
Both the ZnP and the AuP+ units can be selectively excited,
so that electron transfer can be initiated from different excited
states. In both cases the product is the ZnP•+-AuP• charge-
transfer state. When ZnP was excited, we did not observe any
effect of the Ag+ or Cu+ cations on the electron-transfer
reaction. Interestingly, however, when AuP+ was excited Ag+
increased the electron-transfer rate, which we have reported in
a previous communication.13 We interpreted this as a super-
+
exchange mediation involving the Ag(phen)2 unit, that is
enhanced as compared to the case in Zn2//Au+. In the present
paper, we have expanded the series of rotaxanes and found that
with Cu+, the rate was further increased, and the reaction
2+
presumably involved the oxidized Cu(phen)2 complex as a
(13) Andersson, M.; Linke, M.; Chambron, J.-C.; Davidsson, J.; Heitz, V.;
Sauvage, J.-P.; Hammarstro¨m, L. J. Am. Chem. Soc. 2000, 122, 3526.
real intermediate. Thus, we propose that the role of the metal-
9
4348 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
Figure 3. Synthesis of (H2)2/Cu+/Au+.
phenanthroline complex changes from that of an efficient
superexchange mediator in the case of Ag+ to an intermediate
site in a rapid charge-hopping mechanism in the case of Cu+.
We also propose that the different mediation effects with ZnP
and AuP+ excitation can be rationalized by the involvement of
different mediating orbitals, that is, an “electron-transfer”2b,6,14
pathway with ZnP excitation but “hole-transfer”2b,6,14 with AuP+
excitation. A complete account of our photochemical results is
given, including the subsequent recombination and ZnP triplet
reactions as well as data for three new rotaxanes. Furthermore,
the synthesis and structural characterization of the series of
rotaxanes are described.
Figure 4. Illustration of the demetalation/remetalation reactions carried
out on the Cu(I)-complexed [2]-rotaxane Zn2/Cu+/Au+ to afford free
[2]-rotaxane Zn2//Au+ and metallo-[2]-rotaxanes Zn2/Ag+/Au+ and Zn2/
Li+/Au+. In the free [2]-rotaxane Zn2//Au+, the macrocycle is more deeply
buried inside the cavity formed by the bis-porphyrin dumbbell as compared
to the complexed [2]-rotaxanes.
and a difunctionalized thread B, thanks to the gathering
properties of a transition metal (black dot). Construction of the
porphyrin blocking groups is the kinetically templated key step
leading to the metal-complexed rotaxane structure D. The
desired rotaxane E is obtained after removal of the metal
template from D.
2. Results and Discussion
2.1. Design and Synthesis. Principle of Rotaxane Con-
struction. Rotaxanes made of a pendant gold-porphyrin macro-
cycle threaded inside a bis-zinc porphyrin stoppered dumbbell
are built using a template approach to assemble the two parts
of the system, the dumbbell and the macrocycle. This has
already led to the synthesis of various catenanes,15 rotaxanes,16
and knots.17 The principle of template construction is depicted
in Figure 1. The precursor C to the rotaxane E called prerotaxane
is obtained in one step from a gold porphyrin macrocycle A
Synthesis of [2]-Rotaxanes. The precursors to the synthesis
of [2]-rotaxanes are represented in Figure 2. The macrocycle
Au+ bearing a 2,9-diphenyl-1,10-phenanthroline chelate (dpp)
and a pendant gold(III) porphyrin was synthesized and used as
a building block for the construction of porphyrin-containing
[2]-catenanes.18 Mixing Au+ with equimolar amounts of Cu-
(CH3CN)4PF619 followed by addition of 2,9-di(p-formylphenyl)-
1,10-phenanthroline20 1 gave quantitatively the prerotaxane Cu+/
Au+. This copper(I) complex was submitted to porphyrin
synthesis21 at the aldehyde function by reaction with 3,5-di-
tert-butylbenzaldehyde 2 (8 equiv),22 (3,3′-dihexyl-4,4′dimethyl-
2,2′-dipyrryl)methane (10 equiv) 3, and a few drops of
(14) Newton, M.Chem. ReV. 1991, 91, 767.
(15) (a) Dietrich-Buchecker, C.; Sauvage, J.-P. Chem. ReV. 1987, 87, 795. (b)
Dietrich-Buchecker, C.; Sauvage, J.-P. Tetrahedron 1990, 46, 503. (c)
Sauvage, J.-P. Acc. Chem. Res. 1990, 23, 327.
(16) (a) Chambron, J.-C.; Heitz, V.; Sauvage, J.-P. J. Chem. Soc., Chem.
Commun. 1992, 1131. (b) Chambron, J.-C.; Heitz, V.; Sauvage, J.-P. J.
Am. Chem. Soc. 1993, 115, 12378.
(18) (a) Amabilino, D. B.; Sauvage, J.-P. Chem. Commun. 1996, 2441-2442.
(b) Amabilino, D. B.; Sauvage, J.-P. New. J. Chem. 1998, 22, 395.
(19) Kubas, G. J.; Monzyk, B.; Crumbliss, A. L. Inorg. Synth. 1990, 28, 90.
(20) Linke, M.; Chambron, J.-C.; Heitz, V.; Sauvage, J.-P.; Encinas, S.;
Barigelletti, F.; Flamigni, L. J. Am. Chem. Soc. 2000, 122, 11834.
(21) (a) Lindsey, J. S.; Hsu, H. C.; Schreiman, I. C. Tetrahedron Lett. 1986,
27, 4969. (b) Lindsey, J. S.; Schreiman, I. C.; Hsu, H. C.; Kearney, P. C.;
Marguerettaz, A. M. J. Org. Chem. 1987, 52, 827. (c) Lindsey, J. S.;
Wagner, R. W. J. Org. Chem. 1989, 54, 828.
(17) (a) Dietrich-Buchecker, C. O.; Sauvage, J.-P. Angew. Chem., Int. Ed. Engl.
1989, 28, 189. (b) Dietrich-Buchecker, C. O.; Sauvage, J.-P.; Kintzinger,
J.-P.; Malte`se, P.; Pascard, C.; Guilhem, J. New J. Chem. 1992, 16, 931.
(c) Dietrich-Buchecker, C. O.; Nierengarten, J.-F.; Sauvage, J.-P.; Armaroli,
N.; Balzani, V.; De Cola, L. J. Am. Chem. Soc. 1993, 115, 11237. (d)
Carina, R.; Dietrich-Buchecker, C.; Sauvage, J.-P. J. Am. Chem. Soc. 1996,
118, 9110. (e) Dietrich-Buchecker, C. O.; Sauvage, J.-P.; De Cian, A.;
Fischer, J. J. Chem. Soc., Chem. Commun. 1994, 2231.
(22) Chardon-Noblat, S.; Sauvage, J.-P. Tetrahedron 1991, 47, 5123.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4349

A R T I C L E S
Andersson et al.
Table 1. Selected 1H NMR Data for the Rotaxanes and Reference Compoundsa
dumbbell protons
macrocycle protons
compound
Cu+/Au+
meso
5′,6′
4′,7′
3′,8′
o′
m′
Me1
py1
py2
5,6
4,7
3,8
o
m
pp3
op3
9.69b
10.16
10.12
10.09
10.11
10.03
8.46
8.86
8.86
8.77
8.89
7.98
8.94
9.21
9.27
9.38
9.21
8.57
8.11
8.52
8.54
8.58
8.48
8.90
7.67
8.07
8.07
8.09
7.97
9.48
7.12
7.46
7.47
7.54
7.44
7.72
9.44
9.55
9.56
9.54
9.56
9.42
9.29
9.45
9.45
9.44
9.45
9.24
7.84
7.54
9.55
7.33
7.39
4.85
8.43
8.32
8.26
8.14
8.23
6.72
7.91
8.06
8.04
7.52
8.12
7.53
7.22
7.60
7.58
7.76
7.58
8.63
6.12
6.33
6.35
6.45
6.36
7.00
7.33
7.21
7.20
7.16
7.19
7.86
7.59
7.63
7.67
7.62
7.63
7.70
(H2)2/Cu+/Au+
Zn2/Cu+/Au+
Zn2/Ag+/Au+
Zn2/Li+/Au+
Zn2/Au+
1.78
1.76
1.70
1.75
2.03
a Chemical shifts in ppm downfield from TMS. Solvent: CD2Cl2. b CHO proton.
Figure 5. Numbering of protons for the two components of a [2]-rotaxane, the bis-porphyrin dumbbell, and the gold(III)-porphyrin-pendant macrocycle.
1
trifluoroacetic acid in CH2Cl2 (Figure 3). The mixture was stirred
at room temperature for 16 h. After addition of tetrachloro-p-
benzoquinone to oxidize the intermediate porphyrinogen groups
formed, the mixture was refluxed for 1.5 h. The crude product
was subjected to chromatographic purifications on silicagel,
which resulted in 18% isolated yield of the free base copper(I)
[2]-rotaxane (H2)2/Cu+/Au+. Complexation of the porphyrin
stoppers was achieved by reacting (H2)2/Cu+/Au+ for 1 h with
2 equiv of Zn(OAc)2‚2H2O in a refluxing mixture of CHCl3
and CH3OH. The Zn2/Cu+/Au+ [2]-rotaxane was obtained in
82% yield.
To remove the copper template, a solution of Zn2/Cu+/Au+
was reacted with 50 equiv of KCN in a ternary mixture of
CHCl3/H2O/CH3CN (1/1/4).23 Demetalation did not go further
after refluxing the solution for 30 min; therefore, the metal-
free [2]-rotaxane was separated by column chromatography, and
the remaining Zn2/Cu+/Au+ was subjected to the same reaction
again. This procedure was repeated five times to lead to the
free rotaxane Zn2//Au+ in 92% yield.
Coordination Chemistry Performed on the dpp Chelates
of the [2]-Rotaxane Zn2//Au+. The free coordination site
formed by the dpp chelate can be occupied by other metals
accommodating a tetrahedral geometry. Thus, treating the free
Zn2//Au+ rotaxane with lithium or silver salt resulted quanti-
tatively in the formation of two new complexed rotaxanes, Zn2/
Li+/Au+ and Zn2/Ag+/Au+ (Figure 4). These complexes have
a conformation similar to the Zn2/Cu+/Au+ complex. They have
been characterized by mass spectroscopy, H NMR spectros-
copy, and in the case of Zn2/Li+/Au+, also by 7Li NMR
spectroscopy.
2.2. Conformation of the Different [2]-Rotaxanes Studied
by NMR Spectroscopy. Relevant protons, which can be
considered as probes for the geometry of the different [2]-ro-
taxanes, are reported in Table 1. The proton labeling of the
1
components of these rotaxanes is shown in Figure 5. The H
NMR spectrum of prerotaxane Cu+/Au+ was discussed in a
previous paper and is in agreement with an extended conforma-
tion of this molecule.18b Introduction of the porphyrin stoppers
led to a downfield shift of the protons localized on the thread
of (H2)2/Cu+/Au+ due to the deshielding effects exerted by the
porphyrin rings. Zinc coordination on both porphyrins results
in almost no change in the chemical shift of Zn2/Cu+/Au+ as
compared to that of (H2)2/Cu+/Au+. The postulated conforma-
tion of Zn2/Cu+/Au+, an extended one, in which the two zinc
porphyrins and the gold porphyrin are as far as possible from
each other is supported by several dipolar correlations observed
on 2-D ROESY spectra. For example, protons o located on the
macrocycle correlate both with protons o′ and with Me(1) of
the dumbbell (Figure 4).
Exchanging either Li+ or Ag+ for Cu+ led to similar 1H NMR
spectra. In the case of Zn2/Ag+/Au+, a downfield shift of 0.5
ppm for protons m and 3,8 of the macrocycle is observed. This
may result from the larger size of the Ag+ cation, which would
move the dpp chelate away from the shielding field of the other
chelate. In the case of the lithium rotaxane Zn2/Ag+/Au+, the
chemical shift of the lithium is 2.79 ppm, a value close to the
chemical shift of lithium in the lithium (I) [2]-catenane (δ )
(23) (a) Albrecht-Gary, A.-M.; Saad, Z.; Dietrich-Buchecker, C. O.; Sauvage,
J.-P. J. Am. Chem. Soc. 1985, 107, 3205. (b) Dietrich-Buchecker, C. O.;
Sauvage, J.-P.; Kern, J.-M. J. Am. Chem. Soc. 1984, 106, 3043.
9
4350 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
Table 2. Electrochemical Potentialsa and Excited State Energies
reference compoundb
or componentc
E
E
1E
(eV)
3E
(eV)
ox
red
(V vs SCE)
(V vs SCE)
AuP+
ZnP
/Cu/
1.6d
-0.52 (-0.55)
-1.55e
2.18
2.13
1.75
1.72
1.85e
2.49e
0.70 (0.71)
(1.00)
-1.50f
+
Ag(dpp)2
∼1.5d
-0.7g
>3h
a See Figures S3 and S4. b The reference compounds are Zn, Au+, and
Ag(dpp)2+, and /Cu/ denotes the Cu(dpp)2+ component of the [2]-rotaxane
Zn2/Cu+/Au+. c The numbers in brackets are determined for the corre-
sponding components in the [2]-rotaxane Zn2/Cu+/Au+. d Reference 43c.
e Reference 36. f Value for related unit, refs 28, 39. g Reference 43a.
h Reference 46.
fluorescent 1ZnP singlet excited state and continuing with those
3
initiated from the nonemitting AuP+ triplet state.
2.3.1. Properties of the Individual Porphyrin Components.
The absorption spectra of the rotaxanes were equal to a sum of
the spectra of the model porphyrins Zn and Au+ (Supporting
Information, Figure S1), indicating that the electronic coupling
between the porphyrins is very weak. The spectra show that
the ZnP units can be excited at 575 nm with a >90% selectivity,
while a 70% selectivity of the AuP+ excitation is achieved at
528 nm. The excited-state energies and redox potentials of the
ZnP and AuP+ units are given in Table 2. The lifetime of the
excited singlet state of Zn was 2.0 ns, as determined from the
fluorescence lifetime. In similar Zn(II)-porphyrins the 1ZnP state
is converted to the triplet 3ZnP state with a yield of 0.8-0.9.25-27
3
From transient absorption measurements of the ZnP state of
Figure 6. Structures of the model compounds Zn and Zn2/Cu+/.
Zn, a lifetime of 7.5 µs was obtained. The lowest excited singlet
state of the AuP+ units undergoes a rapid intersystem crossing
within a few picoseconds,25,28,29 due to strong spin-orbit
coupling. Thus, the electron-transfer reactions induced by AuP+
excitation started from the 3AuP+ state. The lifetime of this state
in Au+ was 2.0 ns, as determined from the decay of its transient
absorption.
2.2 ppm).24 This result was a good proof of the presence of
lithium coordinated by the two dpp chelate in Zn2/Li+/Au+.
Removal of the copper template resulted in some important
modifications of the Zn2//Au+ spectra, explained by a less
constrained molecule in which the macrocycle has moved
toward the cavity formed by the bis-porphyrin thread. Thus,
the phenanthroline protons (5,6), (4,7), and (3,8) now lie in the
shielding field of the two zinc porphyrins and are shielded by
-2.7, -1.94, and -0.51 ppm, respectively. Protons o and m
that do not experience the phenanthroline field anymore are
deshielded by +1.05 and +0.65 ppm, respectively. This new
conformation is supported by dipolar coupling observed between
proton m and both Me(1) and o′. Other important chemical shifts
resulting from demetalation are the upfield shift of protons 5′,6′
(-0.88 ppm) and 4′,7′ (-0.80 ppm)and the deshielding of proton
pp3 (0.66 ppm). A folding of the macrocycle placing the gold
porphyrin macrocycle above the dpp fragment of the dumbbell
can be proposed to account for these observations.
2.3. Photophysics and Photochemistry. The photoinduced
reactions in the rotaxanes Zn2/Cu+/Au+, Zn2/Ag+/Au+, and
Zn2//Au+ of Figure 4 were studied with time-resolved emission
and transient absorption spectroscopy. (The results from Zn2/
Li+/Au+ are not included here because of the instability of this
Li+ complex in DMF, the solvent used for the photophysical
experiments.) Model compounds were macrocycle Au+ with
attached gold(III) porphyrin (Figure 2), Zn(II) porphyrin Zn,
and Cu(I)-complexed [2]-rotaxane Zn2/Cu+/ lacking the ap-
pended Au(III) porphyrin (Figure 6). The reactions in the
rotaxanes are discussed starting with those initiated from the
Zn(II)-tetraphenyl and -octaethyl porphyrins usually have
triplet lifetimes of 0.1-1 ms25-28 in deoxygenated solutions,
which is much longer than the value observed here. However,
in a recent publication,27a similarly short triplet lifetimes were
measured for mono- and di-meso-phenyl-substituted ZnOEP.
It was suggested that steric interaction between the phenyl and
alkyl substituents induced a conformational change to a non-
planar porphyrin in the triplet state, leading to a more rapid
relaxation to the ground state. In our transient absorption spectra
of the triplet state of Zn, we noted a shift of the transient
absorption maximum from 454 to 462 nm, with isosbestic points
at 460 and 510 nm, consistent with a conversion between two
triplet states. From the transient absorption kinetics, we found
that this conversion occurred with a time constant of 100 ns.
(25) Kalyanasundaran, K. Photochemistry of Polypyridine and Porphyrin
Complexes; Academic Press: London, 1992.
(26) (a) Kalyanasundaran, K.; Neumann-Spallert, M. J. Phys. Chem. 1982, 86,
5163. (b) Flamigni, L.; Armaroli, N.; Barigelletti, F.; Balzani, V.; Collin,
J.-P.; Dalbavie, J.-O.; Heitz, V.; Sauvage, J.-P. J. Phys. Chem. B 1997,
101, 5936.
(27) (a) Knyukshto, V.; Zenkevich, E.; Sagun, E.; Shulga, A.; Bachilo, S. Chem.
Phys. Lett. 1998, 297, 97. (b) Darwent, J. R.; Douglas, P.; Harriman, A.;
Porter, G.; Richoux, M.-C. Coord. Chem. ReV. 1982, 44, 83. (c) Tran-Thi,
T. H.; Desforge, C.; Thiec, C.; Gaspard, S. J. Phys. Chem. 1989, 93, 1226.
(d) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic Press:
New York, 1978; Vol. 3.
(28) Brun, A. M.; Harriman, A.; Heitz, V.; Sauvage, J.-P. J. Am. Chem. Soc.
1991, 113, 8657.
(29) Antipas, A.; Dolphin, D.; Gouterman, M.; Johnson, E. C. J. Am. Chem.
Soc. 1978, 100, 7705.
(24) Dietrich-Buchecker, C. O.; Sauvage, J.-P.; Kern, J.-M. J. Am. Chem. Soc.
1989, 111, 7791.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4351

A R T I C L E S
Andersson et al.
Table 3. Summary of Forward Electron-Transfer Rate Constants, Calculated as kET ) 1/τ - 1/τ0, Where τ and τ0 Are the Excited State
Lifetimes in the Rotaxane and Model Compounds
k
ET (s-1) (rel. amplitudes)
+
ZnP excitation
AuP excitation
Zn2//Au+
1.2 × 1010 (35%)
1.2 × 1010 (35%)
1.2 × 1010 a
1.3 × 109 (65%)
1.1 × 109 (65%)
a
1.1 × 1010 (40%)
2.0 × 1010 (40%)
3.9 × 1010 (75%)
5 × 108 (60%)
2.0 × 109 (60%)
1.6 × 109 (25%)
Zn2/Ag+/Au+
Zn2/Cu+/Au+
a See text.
Similar data have recently been presented for Zn(II)-di-meso-
diaryloctaalkyl porphyrins.30 Thus, it seems that a conversion
3ZnPinitial f 3ZnPdistorted occurs with a lifetime of 100 ns in these
porphyrins, leading to a substantial reduction in the intrinsic
triplet lifetime. In donor/acceptor assemblies involving this type
of Zn(II)-porphyrins, the unusually short lifetimes should thus
not be misinterpreted as indicating any interaction between the
3ZnP and other intramolecular components. In the present work,
we monitored the electron-transfer reactions on the nanosecond
time scale at 460 nm (the isosbestic point), to avoid any
interference from the triplet dynamics.
2.3.2. Reactions in Zn2//Au+. Excitation of the ZnP Unit.
Selective excitation at 575 nm initially created the lowest excited
1
singlet state of a ZnP unit. The ZnP fluorescence intensity in
Zn2//Au+ was quenched by 88.5% as compared with the model
Zn. Also the lifetime of the fluorescence was reduced. A double
exponential fit gave lifetimes of 80 ps (35%) and 570 ps (65%),
which should be compared with 2.0 ns for Zn. The reduction
in lifetime was due to electron transfer between the ZnP and
AuP+ units, forming a charge-transfer state:
Figure 7. Transient absorption spectra after excitation of the ZnP unit in
Zn2//Au+ with a 150 fs, 575 nm laser pulse.
still has some of the typical 1ZnP characteristics, consistent with
the fluorescence lifetime data. In contrast, the spectrum after
4700 ps is consistent with that of the charge-transfer state
comprising the oxidized ZnP•+ and the reduced AuP•
radicals.28,31b,32 For this state, the differential absorption is
1
Zn1Zn//Au+ f ZnZn•+//Au•
∆G° ) -0.91 eV (1)
1
smaller than that for ZnP over the whole spectrum. Further-
more, the peak around 460 nm is slightly blue shifted, the
1ZnP shoulder at 500 nm is absent, and both the stimulated
emission at 650 nm and the bleach at 550 nm have disappeared.
Instead, a weak absorption from ZnP•+ can be seen around 670
nm.
as shown in the transient absorption experiments below. Rate
constants for the forward electron transfer (eq 1) were calculated
from the lifetime data and are given in Table 3. The value of
∆G° was calculated using the data in Table 2. The observation
of biexponential fluorescence decay traces can be explained by
the presence of different conformations of the rotaxanes, due
to the flexibility of the macrocycle Au+ with attached gold
porphyrin, in which the 1ZnP-AuP+ electronic coupling is
different (see discussion below). Because we were following
The kinetics of the transient absorption decay were studied
with various probe light wavelengths in the range 450-650 nm.
Figure 8 shows a typical transient trace for Zn2//Au+ measured
at 460 nm. The excitation pulse induced a strong absorption
from the 1ZnP state, that decreased during the first nanosecond
as it was converted to the charge-transfer state (eq 1). This is
consistent with the ca. 50% smaller extinction coefficient for
1
the ZnP fluorescence, it is clear that it reflected the excited-
state decay and not subsequent reactions.
Transient absorption spectra were collected at various time
delays after excitation of Zn2//Au+ with a 150 fs laser pulse at
575 nm (Figure 7). The spectra show the initial formation of
the excited 1ZnP state and the subsequent evolution to the
charge-transfer state (eq 1), in agreement with the fluorescence
results. Thus, the spectra after 10-100 ps display features of
1
the ZnP•+-AuP• state as compared to that of ZnP.28,32 The
charge-transfer state was more long-lived and had decayed only
partially after 5 ns, which was the limit of the femtosecond setup.
The best fit of the decay curves (Figure 8) was obtained using
a triple exponential function.33 The faster two components
corresponded to electron transfer from the 1ZnP state, forming
the charge-transfer state (eq 1). The lifetimes obtained, 80 ps
(45%) and 570 ps (55%), were in good agreement with the
fluorescence decay data. A much slower component, with a
1
the excited singlet ZnP,25,26,28,31 with a peak at 465 nm and a
broad shoulder around 500 nm. Bleaching of the Q(1,0) band
can be seen centered around 550 nm, while the Q(0,0) bleaching
around 575 nm was covered by stray light from the excitation
laser pulses. In the 600-750 nm region, there is a broad
absorption, with a bleach feature around 650 nm, due to
(32) (a) Neta, P. J. Phys. Chem. 1981, 85, 3678. (b) Abou-Gamra, Z.; Harriman,
A.; J. Chem. Soc., Faraday Trans. 2 1986, 82, 2337. (c) Fajer, J.; Borg, D.
C.; Forman, A.; Dolphin, D.; Felton, R. H. J. Am. Chem. Soc. 1970, 92,
3451. (d) Shimidzu, T.; Segawa, H.; Iyoda, T.; Honda, K. J. Chem. Soc.,
Faraday Trans. 2 1987, 83, 2191. (e) Carnieri, N.; Harriman, H. Inorg.
Chim. Acta 1982, 62, 103. (f) Mosseri, S.; Mialocq, J. C.; Perly, B.;
Hambright, P. J. Phys. Chem. 1991, 95, 2196. (g) Hayes, R. T.;
Wasielewski, M. R.; Gosztola, D. J. Am. Chem. Soc. 2000, 122, 5563. (h)
Osuka, A.; Noya, G.; Taniguchi, S.; Okada, T.; Nishimura, Y.; Yamazaki,
I.; Mataga, N. Chem.-Eur. J. 2000, 6, 33.
1
stimulated ZnP emission. The transient spectrum after 500 ps
(30) Andre´asson, J.; Zetterqvist, H.; Kajanus, J.; Mårtensson, J.; Albinsson, B.
J. Phys. Chem. A 2000, 104, 9307.
(31) (a) Rodriguez, J.; Kirmaier, C.; Holten, D. J. Am. Chem. Soc. 1989, 111,
6500. (b) Asahi, T.; Ohkohchi, M.; Matsusaka, R.; Mataga, N.; Zhang, R.
P.; Osuka, A.; Maruyama, K. J. Am. Chem. Soc. 1993, 115, 5665.
9
4352 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
Figure 10. Transient absorption spectra after excitation of the AuP+ unit
in Zn2//Au+ with a 150 fs, 528 nm laser pulse.
Figure 8. Transient absorption decay at 460 nm for Zn2//Au+ and Zn2/
Ag+/Au+ after ZnP excitation at 575 nm. The inset shows the decays at
early times.
forward reaction (eq 1). This suggests that the different rotaxane
conformations do not interconvert faster than on the time scale
of nanoseconds and that they give different rates also for the
back reaction (eq 2).
The triplet quantum yield in porphyrins similar to Zn is 0.8-
0.9.25-27 In Zn2//Au+ 88.5% of the fluorescence is quenched
1
by electron transfer, leaving 11.5% of the ZnP state to decay
via the intrinsic pathways. Thus, about 10% of triplet would be
formed in competition with electron transfer. The differential
extinction coefficient at 460 nm for the triplet is about twice
that for the charge-transfer state and would therefore give about
20% of the total signal. Instead, the signal level after the
recombination reaction was less than half of that before the
recombination (compare initial and final amplitudes in Figure
9) and decayed with a lifetime similar to that for the triplet
state of the model Zn. Thus, it seems that most of the 3ZnP in
Zn2//Au+ decayed much faster on a time scale similar to that
of the recombination reaction. Electron transfer to the AuP+
Figure 9. Transient absorption decay at 460 nm for Zn2//Au+ on the nano-
to microsecond time scale after ZnP excitation at 575 nm with a 5 ns laser
pulse.
3
unit is the only conceivable quenching process for the ZnP
state. Presumably also the 3ZnP state undergoes electron transfer
to the AuP+ unit, but on the same time scale as the subsequent
recombination reaction, so that this reaction cannot be resolved.
The fraction of Zn2//Au+ reacting via the 3ZnP state is too small
to account for the full amplitude of either the 10 ns or the 35
ns components in Figure 9. Both of these can be attributed to
the recombination reaction, although the 3ZnP decay could
contribute to the nanosecond decay signal amplitude.
lifetime of 10 ns, represented back electron transfer from the
charge-transfer state, regenerating the ground state:
ZnZn+•//Au• f Zn2//Au+
∆G° ) -1.22 eV (2)
Using nanosecond flash photolysis, we could follow the
complete decay of the transient absorption (Figure 9). A sum
of two exponents was needed to fit the data on the nanosecond
time scale, giving lifetimes of 10 ns (45%) and 35 ns (55%)
for the charge-transfer state. Although the 10 ns component is
close to the time resolution limit of the nanosecond setup, the
lifetime value is in good agreement with that obtained in the
femtosecond experiments. The relative amplitudes of the
components show a weight similar to that observed in the
The reactions in Zn2//Au+ are summarized in the scheme of
Figure 15. From the fluorescence lifetimes of Zn2//Au+ and
the model Zn, a quantum yield of 85% for charge transfer was
calculated. This is also in agreement with the relative magnitudes
of the transient absorption changes at 460 nm (Figure 8) where
the initial 1ZnP state has about twice the differential extinction
coefficient of the ZnP•+-AuP• state.28,32
Excitation of the AuP+ Unit. Excitation of the AuP+ unit
generates the 3AuP+ state within a few picoseconds.25,28,29 The
3
triplet AuP+ state displays no emission at room temperature,
(33) The following procedure was used to analyze the transient decays on the
picosecond time scale; the 570 ps lifetime obtained in the fluorescence
lifetime measurements for Zn2//Au+ was used as a fixed parameter in the
triple exponential fit, while the other lifetimes and all amplitudes were
unrestricted. Thus, we obtained lifetimes of 80 ps (45%) and 570 ps (55%)
for the forward electron transfer (eq 1), and a much slower decay lifetime
of 10 ns corresponding to back electron transfer (eq 2). Note that if instead
only the 10 ns component was fixed, the 570 ps lifetime was recovered
((10%), showing the robustness of the fit. The same lifetimes were
observed for all probe wavelengths (450-650 nm). Only the relative
amplitudes varied in the fit results; either the faster two components, or
the slowest one, decreased in amplitude at wavelengths where the
corresponding reaction gave small absorption changes.
and the electron-transfer reactions were followed with transient
absorption spectroscopy only. Figure 10 shows transient absorp-
tion spectra after excitation of Zn2//Au+ with 528 nm light. At
10-100 ps, the spectra show characteristic features of the 3AuP+
state28 with peaks around 460 and 620 nm. After 4700 ps
instead, the spectrum has features of the charge-separated
state,28,32 where the shoulder at 500 nm is absent, and the 620
nm signal is much smaller. This shows that electron transfer
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4353

A R T I C L E S
Andersson et al.
2.3.3. Reactions in Zn2/Ag+/Au+. Excitation of the ZnP
Unit. The results for Zn2/Ag+/Au+ obtained with excitation of
the ZnP units at 575 nm were quantitatively very similar to
1
those for Zn2//Au+. The steady-state fluorescence from the -
ZnP state in Zn2/Ag+/Au+ was quenched by 74% as compared
to that of the model Zn. The fluorescence lifetime measurements
displayed the same biexponential behavior as for Zn2//Au+, with
a fast component of 80 ps (31%) and a slower component of
700 ps (66%).34 Also the transient absorption spectra (not
shown) match those for Zn2//Au+ in Figure 7.
As for Zn2//Au+, the transient absorption kinetics for Zn2/
Ag+/Au+ were best fit by a triple exponential function (Figure
8). The fastest two components corresponded to electron transfer
1
from the ZnP state:
Zn1Zn/Ag+/Au+ f ZnZn+•/Ag+/Au•
∆G° ) -0.91 eV (1′)
Figure 11. Transient absorption decay at 460 nm for Zn2//Au+ and Zn2/
Ag+/Au+ after AuP+ excitation at 528 nm. The inset shows the decay at
early times.
and the lifetimes of 90 ps (40%) and 650 ps (60%) obtained
were in good agreement with the emission data. A lifetime of
10 ns was obtained for the third, slower component which was
attributed to the recombination reaction:
from the ZnP to the AuP+ unit occurred also with AuP+
excitation:
Zn2//3Au+ f ZnZn+•//Au•
∆G° ) -0.53 eV
(3)
ZnZn+•/Ag+/Au• f Zn2/Ag+/Au+
∆G° ) -1.22 eV
(2′)
The transient absorption decay at 460 nm for Zn2//Au+ is shown
in Figure 11 (together with that of Zn2/Ag+/Au+, see below).
As for excitation of the ZnP unit above, the best fit to the decay
curves was obtained with a triple exponential function. For Zn2//
Au+, this gave lifetimes of 90 ps (40%) and 1000 ps (60%),
respectively, followed by a slow decay with a lifetime of 10
ns. The transient spectra in Figure 10 display the absorption
The recombination was followed to completion using the
nanosecond setup, and a biexponential fit to the decays at 460
nm yielded lifetimes of 10 ns (50%) and 40 ns (50%). Thus,
the rate constants for both forward and back electron transfer
were almost identical in Zn2//Au+ and Zn2/Ag+/Au+. Also the
yield of charge separation was the same in the two rotaxanes
as determined from the 460 nm transient absorption level at
the end of the traces in Figure 8. Consequently, we conclude
that the presence of the Ag+ cation has no significant effect on
the electron-transfer reactions initiated from the 1ZnP state. This
is in contrast to the effect for the reactions initiated from the
3AuP+ triplet, as discussed below. Note that electron transfer
3
signature of the AuP+ triplet (the shoulder at 500 nm and the
peak at 620 nm) also after 500 ps, showing that both of the
two faster decay components reflected the forward electron-
transfer process (eq 3). The observation of biexponential electron
transfer is consistent with the findings for excitation of the ZnP
unit above. The different conformations present are likely to
give different deactivation rates also for the AuP+ triplet.
3
from the 1ZnP to the Ag(phen)2 unit is also energetically
+
1
The 10 ns lifetime of the slowest component is identical to
that for the recombination of the charge-transfer state (eq 2)
observed after excitation of the ZnP unit. Using nanosecond
flash photolysis, the recombination was followed to completion.
The transient absorption decay was fitted with a sum of two
exponents, giving lifetimes of 10 ns (40%) and 40 ns (60%).
The observation of nearly identical recombination lifetimes with
excitation of either the ZnP or the AuP+ unit strongly suggests
that the same charge-transfer state is formed, independent of
excitation wavelength. Thus, the ZnP•+-AuP• state shows no
spin memory of the excited state from which it was formed.
Note that although energy transfer from the 3AuP+ triplet to
form the 3ZnP triplet is energetically possible (Table 2), it cannot
explain the transient absorption curves at, for example, 460 nm.
The extinction coefficient of the 3ZnP state is somewhat larger
possible (∆G° ) -0.7 eV, Table 2). However, since the ZnP
lifetime in Zn2/Ag+/Au+ is even slightly longer than that in
Zn2//Au+, this reaction seems to be insignificant.
As for Zn2//Au+ above, the absorption signal remaining after
complete decay of the charge-transfer state in the nanosecond
experiments was too small to account for the expected fraction
3
of ZnP generated. Thus, it seems that also for Zn2/Ag+/Au+
the 3ZnP state reacted by electron transfer to the AuP+ unit on
the same time scale as the subsequent recombination.
Excitation of the AuP+ Unit. Excitation of the AuP+ unit
at 528 nm resulted in transient absorption spectra for Zn2/Ag+/
Au+ 13 that were qualitatively identical to those for Zn2//Au+
given in Figure 10. The spectra show characteristic features of
the 3AuP+ state evolving to the charge-transfer state according
to
3
than that of AuP+,28 so that the transient absorption would
Zn2/Ag+/3Au+ f ZnZn+•/Ag+/Au•
∆G° ) -0.53 eV (3′)
display a rise rather than a decay. Instead, the data are consistent
with electron transfer (eq 3), as the sum of the ZnP•+ and AuP•
differential extinction coefficients at 460 nm are approximately
one-half of that for the initial 3AuP+ state.28,32 From the
absorption levels in Figure 11, we estimate that the charge-
separation yield is about 50%. From the lifetime data instead,
we estimate a yield of 65%.
The transient absorption decay at 460 nm is shown in Figure
11. A triple exponential fit gave lifetimes of 50 ps (40%) and
(34) A small fraction (3%) with a lifetime of 2.0 ns was also found,
corresponding to the lifetime of the singlet excited state in the model Zn.
9
4354 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
400 ps (60%), respectively, for the forward electron transfer
and 10 ns for the subsequent charge recombination. This should
be compared with the forward reaction in Zn2//Au+ that gave
lifetimes of 90 and 1000 ps (see above). Thus, we note in the
case of AuP+ excitation a clear increase of the electron-transfer
rate when the Ag+ is coordinated to the phenanthroline. This
important observation is further supported by the yield of charge
transfer. For Zn2/Ag+/Au+, the transient absorption after 3 ns
had decayed to 45% of the initial value, suggesting a near unity
charge-separation yield. In contrast, we see that the yield for
Zn2//Au+ was only about one-half of that.
Alternative reactions cannot explain the higher reaction rates
in Zn2/Ag+/Au+. Excitation, oxidation, and reduction of the
1
Ag+(phen)2 unit, as well as energy transfer to form the ZnP
singlet, can all be ruled out on thermodynamic grounds (Table
2). Energy transfer forming the 3ZnP triplet is exergonic but is
not consistent with the transient absorption traces at 460 nm,
as discussed for Zn2//Au+ above. Finally, the enhanced rate
cannot be explained by a simple Ag+-induced structural change
bringing the ZnP and AuP+ units closer together, since then a
faster electron transfer would have been observed also when
the ZnP unit was excited. Instead, we propose that coordination
of Ag+ increased the rate of direct electron transfer between
the ZnP and AuP+ units in a superexchange mechanism.9 In
contrast, there was no rate enhancement in Zn2/Ag+/Au+ when
the ZnP unit was excited. The Ag+-induced rate enhancement
for electron transfer between the noncovalently linked ZnP and
AuP+ units is one of the main results of the photochemical
studies and will be further discussed below.
Figure 12. Transient absorption changes at 460 nm for Zn2/Cu+/ after
ZnP excitation at 575 nm.
the two-step energy-transfer mechanism. The initial transient
absorption at 460 nm from 1ZnP decreased rapidly as expected
(Figure 12), since the excited *Cu(phen)2+ absorption is small
at this wavelength.25,35 The subsequent absorption rise up to
90% of the initial level showed that formation of 3ZnP occurred,
because this is the only state with an extinction coefficient as
high as 1ZnP. A fit according to eq 4 gave a 180 ps lifetime for
the 1ZnP state, in excellent agreement with the fluorescence data,
and a 650 ps lifetime for the following energy transfer from
*Cu(phen)2+ to 3ZnP. Given the intrinsic lifetimes of 1ZnP and
*Cu(phen)2+, we expect a 90% yield of ZnP by the two-step
3
mechanism. Furthermore, since the main competing process is
direct intersystem crossing from 1ZnP to 3ZnP we expect a near
unity triplet yield. Thus, from the absorption levels in Figure
12, we estimate that the differential extinction coefficient of
the 3ZnP state at 460 nm is 90-95% of that for the 1ZnP state.
The lifetime of the 3ZnP state was 6 µs, as determined in flash
photolysis experiments (not shown), close to the value for the
model Zn.
The charge-separated state decayed with lifetimes of 10 ns
(35%) and 40 ns (65%), as determined from the nanosecond
flash photolysis traces (not shown). This is very close to the
results obtained for Zn2//Au+, showing that the recombination
reaction was not significantly effected by Ag+, as was also
concluded from the results from ZnP excitation above.
2.3.4. Reactions in Zn2/Cu+/. Zn2/Cu+/ was studied as a
model for the reactions between the ZnP and Cu(phen)2+ units
in Zn2/Cu+/Au+. Excitation of the ZnP units at 575 nm showed
that the 1ZnP fluorescence intensity of Zn2/Cu+/ was quenched
by 90% as compared to that of the model Zn. The fluorescence
decay of Zn2/Cu+/ was single exponential with a lifetime of
180 ps, which corresponds to a 91% quenching as compared to
that of the model Zn.
2.3.5. Reactions in Zn2/Cu+/Au+. Excitation of the ZnP
Unit. Upon excitation at 575 nm, the 1ZnP fluorescence intensity
in Zn2/Cu+/Au+ was quenched by 93% as compared to that
for the model Zn. A double exponential fit to the fluorescence
decay curves gave lifetimes of 60 ps (70%) and 300 ps (30%).37
The fast (60 ps) component was faster than in both Zn2/Cu+/
+
and Zn2//Au+, suggesting that energy transfer to Cu(phen)2
,
1
In Zn2/Cu+/, electron transfer from the singlet ZnP to the
3
followed by generation of the ZnP state (eq 4′′), occurred in
competition with electron transfer to the AuP+ unit (eq 1′′):
+
Cu (phen)2 unit is conceivable, but somewhat endergonic
(Table 2, Figure 16). This suggests that the emission quenching
is instead due to energy transfer to the Cu(phen)2+ unit, which
has an excited state below the 1ZnP state (Table 2). The excited
*Cu(phen)2+, which is probably a singlet and triplet in equi-
librium,35 has an expected lifetime of 175 ns25,35,36 and is higher
in energy than the 3ZnP state (Table 2). Thus, a two-step energy-
transfer process is energetically possible (Figure 16):
Zn1Zn/Cu+/Au+ f ZnZn+•/Cu+/Au•
∆G° ) -0.91 eV (1′′)
Zn1Zn/Cu+/Au+ f Zn2/*Cu+/Au+ f 3ZnZn/Cu+/Au+
(4′′)
The transient absorption data support this suggestion. This is
demonstrated most clearly in the data at 460 nm (Figure 13),
where an initial decay of the 1ZnP absorption was followed by
1ZnZn/Cu+/ f Zn2/*Cu+/ f 3ZnZn/Cu+/
(4)
3
a partial recovery as the ZnP state was populated, just as for
The corresponding process was reported by Flamigni et al.36 in
a Zn(II)-porphyrin-stoppered Cu(I) [3]-rotaxane with structural
similarities to Zn2/Cu+/. Our transient absorption data support
Zn2/Cu+/. Because direct electron transfer to the AuP+ unit
(37) The observation of a 300 ps lifetime for the slow conformation in the
1
fluorescence lifetime measurement, that is, a slower ZnP decay than that
(35) Everly, M. R.; McMillin, D. R. J. Phys. Chem. 1991, 95, 9071.
(36) Flamigni, L.; Armaroli, N.; Barigelletti, F.; Chambron, J.-C.; Sauvage, J.-
P.; Solladie, N. New J. Chem. 1999, 23, 1151.
in the Zn2/Cu+/, probably reflects the increased uncertainty as the lifetimes
of the two conformations become similar and more difficult to deconvolute.
The true value is probably ca. 150 ps.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4355

A R T I C L E S
Andersson et al.
Figure 14. Transient absorption decay at 460 nm for Zn2/Cu+/Au+ after
Figure 13. Transient absorption changes at 460 nm for Zn2/Cu+/Au+ after
ZnP excitation at 575 nm. The inset shows the decay at early times.
AuP+ excitation at 528 nm. The inset shows the decay at early times.
and electron transfer to the AuP+ unit (∆G° ) -0.1and -0.25
eV, respectively). In Figure 16, we present the reaction pathways
in Zn2/Cu+/Au+. The reaction scheme proposed is consistent
with the transient absorption kinetics over the whole range
investigated (450-650 nm) and with the transient spectra (see
Figure S2). The most notable difference in the transient spectra
as compared to that of the other rotaxanes is that the initial
decrease in absorption at 450-520 nm during the first few
hundred picoseconds was followed by a recovery that gave a
larger signal again after 1000-5000 ps. Similarly, the Q-band
bleach at 550 nm grew stronger again after 5000 ps, all
competes with the two-step energy transfer, both the 3ZnP and
the charge-transfer states were formed on this time scale. Thus,
the transient absorption did not recover to the same level in
Zn2/Cu+/Au+ as it did in Zn2/Cu+/ (compare the signals after
ca. 2 ns in Figures 12 and 13). On a slower, nanosecond time
scale, the transient absorption decreased again when the charge-
separated state recombined to the ground state, as in Zn2//Au+.
A fit to the kinetic data38 gave a lifetime of 55 ps for the
1
initial decay of the ZnP state and a 650 ps lifetime for the
3
subsequent rise as the ZnP was populated. The 55 ps lifetime
1
3
of ZnP was in good agreement with the fluorescence results
consistent with formation of ZnP.
The decay of the charge-separated and 3ZnP states was
followed by nanosecond flash photolysis. Similar to the case
in Zn2//Au+ and Zn2/Ag+/Au+ above, the back electron transfer
occurred with lifetimes of 10 ns (40%) and 40 ns (60%):
and also with the lifetime expected from a combination of the
two competing processes, since a 90 ps electron transfer to the
AuP+ unit (as for Zn2/Ag+/Au+) and a 180 ps energy transfer
to Cu(phen)2+ (as in Zn2/Cu+/) give an expected 1ZnP lifetime
in Zn2/Cu+/Au+ of 60 ps. Thus, the rate of electron transfer
from 1ZnP to AuP+ was apparently unaffected by the presence
of Cu+. The only effect of Cu+ was consequently to open an
additional energy-transfer reaction in parallel to electron transfer.
ZnZn+•/Cu+/Au• f Zn2/Cu+/Au+
∆G° ) -1.22 eV
(2′′)
Thus, the kinetics are unaltered by the insertion of Cu+ also
for the recombination reaction. From the absorption levels at
the end of Figure 13, it is clear that the 3ZnP state did not decay
significantly within 5 ns. Because its differential extinction
coefficient is twice that of the charge-transfer state, the triplet
yield estimate above indicates that at least one-half of the signal
1
The lifetime data suggest that about two-thirds of the ZnP
singlet states reacted by electron transfer to the AuP+ unit, and
+
one-third underwent energy transfer to the Cu(phen)2 unit,
followed by generation of 3ZnP with τ ) 650 ps (eq 4).
However, the differential absorption at 460 nm from the charge-
3
transfer and ZnP states after 1500 ps was ∼0.65 of the initial
3
after 5 ns (at the end of the trace) would be due to the ZnP
value (Figure 13), whereas it was ∼0.40 for the charge-transfer
state in Zn2//Au+ (Figure 8). Because the 3ZnP state has twice
the extinction coefficient of the charge-transfer state, this
indicates that the total fraction of 1ZnP undergoing energy
transfer was instead about one-half. The additional 3ZnP
generation probably came from the rotaxane conformation that
was slow in electron transfer,37 as seen in the fluorescence decay,
although it could not be resolved from the other processes in
the transient absorption traces.
state. However, when the decay of this signal was followed on
the nanosecond time scale, the recombination lifetimes of 10
and 40 ns obtained represent 80% of the total nanosecond signal
amplitude. Only 20% of the signal decayed with a longer
lifetime (ca. 3 µs). This indicates that a large fraction of the
3ZnP states reacted on the time scale of the recombination
reaction, and the remaining part reacted much slower. This is
again very similar to the case in the other rotaxanes, except
3
that more ZnP was generated in Zn2/Cu+/Au+.
The identical rise times of 650 ps obtained in Zn2/Cu+/ and
Excitation of the AuP+ Unit. The transient absorption
spectra after excitation of Zn2/Cu+/Au+ with 528 nm light were
qualitatively the same as for the other rotaxanes, consistent with
+
Zn2/Cu+/Au+ show that the deactivation of *Cu (phen)2
occurred mainly by energy transfer to 3ZnP, and not by energy
3
a AuP+ state being converted to the ZnP•+-AuP• state (not
(38) The kinetic traces were fitted to a stepwise reaction A f B f C, where A
is the 1ZnP state, B represents a mixture of the charge-separated state and
the excited *Cu(phen)2+, and C represents a mixture of the charge-separated
state and 3ZnP. A slower exponential decay of the charge-transfer state,
fixed to 10 ns according to the flash photolysis results, was also included
in the fit. The result was a lifetime of 55 ps for the initial decay (A f B)
and 650 ps for the subsequent rise (B f C).
shown). Figure 14 shows the transient absorption decay at 460
nm, which was fitted to a triple exponential function as for the
other rotaxanes. The faster two components, representing decay
3
of the AuP+ state, gave lifetimes of 25 ps (75%) and 470 ps
9
4356 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
(25%). The third, slower component of 12 ns represented the
decay of the charge-transfer state. By nanosecond flash pho-
tolysis, this recombination could be followed to completion,
giving charge-transfer lifetimes of 12 ns (35%) and 40 ns (65%),
close to the values for the other rotaxanes.
The 3AuP+ decay is clearly much faster in Zn2/Cu+/Au+ than
in the other rotaxanes. The transient absorption traces show that
energy transfer from 3AuP+ cannot explain the higher rates, as
discussed for Zn2//Au+. Instead, the higher rates can be
attributed to a Cu+-induced enhancement of the rate of electron
transfer from ZnP to 3AuP+. This is supported by the resulting
transient spectrum that is identical to that for the charge-transfer
state in the other rotaxanes. Two different mechanisms for the
Cu+-enhanced electron transfer are possible. Either a bridge-
assisted, direct electron transfer from the ZnP to the 3AuP+ unit
occurred as in Zn2/Ag+/Au+:
Figure 15. Internal relaxation and electron-transfer processes proposed in
Zn2//Au+: (a) k ) 5.0 × 108 s-1; (b) k ) 1.7 × 105 s-1; (c) k > 3 × 1011
s-1; (d) k ) 2.0 × 108 s-1; (e) k ) 1.2 × 1010 s-1 and 1.3 × 109 s-1; (f)
see text; (g) k ) 1.1 × 1010 s-1 and 5.0 × 108 s-1; (h) k ) 1.0 × 108 s-1
and 2.9 × 107 s-1. The corresponding values for the electron-transfer
processes (e-h) Zn2/Ag+/Au+ are (e) k ) 1.2 × 1010 s-1 and 1.1 × 109
s-1; (f) see text; (g) k ) 2.0 × 1010 s-1 and 2.0 × 109 s-1; (h) k ) 1.0 ×
Zn2/Cu+/3Au+ f ZnZn+•/Cu+/Au•
∆G° ) -0.53 eV (3′′)
108 s-1 and 2.5 × 107 s-1
.
or the reaction proceeded in a sequential manner via an initial
formation of the Zn2/Cu2+/Au• state, it seems likely that the
electron-transfer mediated by the Cu+ unit proceeds via the two-
step reaction mechanism (eqs 5 and 6).
electron transfer from the Cu(phen)2 unit to the 3AuP+,
+
followed by a secondary reaction between the ZnP and Cu-
2+
(phen)2 units:
2.4. General Discussion. We have shown that electron
transfer occurs with a lifetime of 100 ps or faster in all the
rotaxanes, both from the 1ZnP state to the AuP+ unit and from
Zn2/Cu+/ 3Au+ f Zn2/Cu2+/Au•
Zn2/Cu2+ /Au• f Zn Zn+•/Cu+/Au•
∆G° ) -0.20 eV
(5)
3
the ground-state ZnP to the excited AuP+. The same ZnP•+-
AuP• charge-transfer state is formed in both cases. This state
has a relatively long lifetime, 10-40 ns, giving a ratio of 100
for the forward/backward electron transfer. The charge-transfer
state lifetime is much longer than in previous rotaxanes where
the ZnP and AuP+ units were both linked to the rod.28,39Despite
the weaker coupling in the present case, the charge-separation
yield is high, and for Zn2/Ag+/Au+ the yield is >80%
irrespective of excitation wavelength. The rate constants for the
different processes are given in Figures 15 and 16.
∆G° ) -0.30 eV (6)
The fact that the electron-transfer rate is enhanced by Cu+ shows
that Cu(phen)2+ orbitals are involved in mediating the reaction.
However, the energy of the Cu(phen)22+-AuP• state is between
the energies of the initial 3AuP+ and the final ZnP•+-AuP• state.
Consequently, it would rather be formed as a real intermediate
according to eqs 5 and 6 than being a virtual state in a
superexchange mechanism.6,9,14 Because the distance between
the Cu(phen)2+ and 3AuP+ units is relatively small, it is likely
that the sequential electron transfer may compete with the direct
electron-transfer reaction from the ZnP unit. The transient
absorption data show no sign of the Cu(phen)22+-AuP• inter-
mediate, although this would give a much smaller signal than
the final ZnP•+-AuP• state, particularly at 460 nm.25,32,35 Thus,
The observation of biexponential electron-transfer kinetics
for both the forward and the backward electron transfer may
seem puzzling at first. However, the rotaxanes are large and
not entirely rigid, and the reactants are not covalently linked to
each other. Consequently, one can expect some structural
flexibility. The biexponential kinetics may thus be attributed to
the presence of different conformations of the rotaxane, which
exhibit different donor/acceptor coupling.40 If the conformational
motions occur on the same time scale as electron transfer, or
slower, the reaction kinetics will not be single exponential. The
data for Zn2/Cu+/, where only processes between the ZnP and
3
if the observed decay times of AuP+ of 25 and 470 ps reflect
the formation of the intermediate according to eq 5, in parallel
with the direct electron transfer from the ZnP unit (eq 1′′), the
subsequent charge transfer to the ZnP unit (eq 6) must be faster
+
+
than that. This is not unlikely, since the ZnP and Cu(phen)2
Cu(phen)2 units occur, show single exponential fluorescence
units are covalently linked via phenyl groups, so that the
electronic coupling would be relatively strong.
decays. Thus, we conclude that the conformation differences
(39) (a) Chambron, J.-C.; Harriman, A.; Heitz, V.; Sauvage, J.-P. J. Am. Chem.
Soc. 1993, 115, 6109. (b) Brun, A. M.; Atherton, S. J.; Harriman, A.; Heitz,
V.; Sauvage, J.-P. J. Am. Chem. Soc. 1992, 114, 4632. (c) Chambron, J.-
C.; Harriman, A.; Heitz, V.; Sauvage, J.-P. J. Am. Chem. Soc. 1993, 115,
7419.
Note that the relative amplitude of the fastest (25 ps) reaction
component is 75%, while it was 40% for Zn2//Au+ and Zn2/
Ag+/Au+. The components reflect reactions for different ro-
taxane conformations, and if the same ZnP and 3AuP+ reactants
are involved as in the other rotaxanes, the relative amplitudes
are not expected to vary. However, the marked difference for
Zn2/Cu+/Au+ is consistent with a change to a reaction between
the Cu(phen)2+ and AuP+ units instead, for which the effect of
the different conformations could be quite different. In conclu-
sion, although we have no direct spectral evidence for the
(40) Note that the two different 1ZnP lifetimes we obtain cannot reflect the
lifetimes of one of the ZnP units on the same rotaxane reacting faster than
the other, since the components do not have a 1:1 amplitude relation, and
3
we see the same biexponential behavior for the AuP+ lifetime. Also, the
biexponential kinetics cannot be explained by excitation of both ZnP and
AuP+ units in the same experiment. This is shown by the contrasting results
obtained for ZnP- and AuP+-excitation, and the fact that also the 1ZnP
fluorescence was biexponential. Finally, the biexponential fluorescence
cannot be due to delayed fluorescence (repopulation) since there are no
states available close enough in energy to 1ZnP.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4357

A R T I C L E S
Andersson et al.
is no entirely covalent electron-transfer pathway between the
porphyrins. From simple molecular models we estimate a
porphyrin edge-to-edge distance of ca. 15-17 Å. Nevertheless,
we observed electron-transfer rate constants up to 1.2 × 1010
s-1 for a direct reaction between the ZnP and AuP+ units. This
is much faster than for activationless electron transfer in proteins,
where a rate constant at 15 Å is typically <1 × 108 s-1.3 In
fact, it is comparable to the values obtained in donor/acceptor
molecules covalently linked by norbornyl units,2a which were
found to be more efficient mediators (â ) 0.86 Å-1 2a
) than other
saturated bonds. In addition, the electron transfer to the 3AuP+
in our rotaxanes is probably not free energy optimized (∆G° )
-0.45 eV), so that an activationless rate constant would be even
larger than the value observed here. These comparisons suggest
that, although the metal-phenanthroline unit is involved, the
electron-transfer pathway may take some shortcuts. A plausible
pathway could, for example, involve a direct space jump from
the phenanthroline on the rod to the AuP+ unit, which could
come close due to breathing motions of the macrocycle.
However, the mediating effect of Ag+- and Cu+-coordination
upon AuP+ excitation shows that the ZnP and AuP+ units did
not come close enough for a direct, through-solvent electron
transfer, because then the same effect would have been observed
with ZnP excitation. Instead, the electron-transfer pathway with
AuP+ excitation involves the metal-phenanthroline complex,
which enhances the electronic coupling.
Figure 16. Internal-relaxation, energy-transfer (EnT), and electron-transfer
(ET) processes proposed in Zn2/Cu+/Au+. For processes a-d the rate
constants are the same as in Zn2//Au+: (e) EnT, k ) 5.1 × 109 s-1; (f)
EnT, k ) 1.5 × 109 s-1; (g) ET, k ) 1.25 × 1010 s-1; (h) ET, see text; (i)
sequential ET, k ) 3.9 × 1010 s-1 and 1.6 × 109 s-1; (j) ET, not resolved;
(k) ET, k ) 1.0 × 108 s-1 and 2.5 × 107 s-1
.
mainly involve motion of the nonrigid Au+-macrocycle. A
rotation of the macrocycle, positioning the AuP+ between the
ZnP units, has been observed in a related rotaxane,20 but this
seems to be excluded in the present case by our NMR data.
Furthermore, when Ag+ or Cu+ is coordinated so that rotation
is prevented, we still observe biexponential kinetics. Instead,
twisting and breathing motions of the polyether chain, changing
the relative position of the ZnP and AuP+ units, are plausible
origins for the different conformations. Because the recombina-
tion reaction was also biexponential, it seems that the conforma-
tions do not interconvert faster than on the nanosecond time
scale. Note that the rate enhancement due to coordination of
Ag+ or Cu+ is only observed with AuP+ excitation and not with
ZnP excitation. This shows that it is the through-bond electronic
coupling - not through-space - that is important for the
reaction. Thus, the different conformations must differ in the
through-bond coupling and not just in the through-space ZnP-
AuP+ distance.
In Table 3, we compare the forward electron-transfer rate
constants in the different rotaxanes, and it is clear that
coordination of Ag+ and Cu+ increases the reaction rate for
AuP+ excitation (eqs 3, 3′, 3′′), but not for ZnP excitation (eqs
1, 1′, 1′′). The different effect of Ag+ on the electron-transfer
rate, depending on which porphyrin is excited, is very interest-
ing. In both cases the electron is transferred from the ZnP to
the AuP+ unit, and the same charge-separated state is generated.
1
The difference is that the initial excited state is either ZnP or
3AuP+. We propose that the results can be explained by
considering that different orbitals may be involved in mediating
the electronic coupling in the two cases. In Figure 17, we show
a schematic picture of HOMO and LUMO orbitals for the ZnP
and AuP+, and the MO:s of the link are represented by only
one HOMO/LUMO couple for simplicity. In a superexchange
reaction, the donor/acceptor coupling is mediated via a small
perturbation by the link.6,9,14 The mediating contribution from
each link unit is inversely proportional to the energy difference
(∆En) between the mediating link state and the donor state.
Furthermore, the donor/acceptor coupling (HAB) is proportional
to the sum of all pairwise transfer integrals (V) between all the
For electron-transfer reactions in general, the rate constant
(k) decreases exponentially with the distance (d) between the
reactants: k
exp(-âd), when corrected for the distance
dependence of other factors. In a superexchange mechanism,6,9,14
the parameter â that characterizes the distance dependence of k
varies with the nature of the link. It decreases with a decrease
in the energy difference between the virtual redox states of the
link and the donor/acceptor states. Consequently, saturated
hydrocarbon links give relatively large values, â ) 0.8-1.0
Å-1 2a,b while unsaturated ones give â ) 0.2-0.6 Å-1 2c-f
,
,
6,14,42
states of the link (ln) and the donor and acceptor:
allowing for rapid long-range electron transfer. Noncovalent
bonds and space jumps, however, are believed to be less efficient
in mediating electron transfer.3,6 When the energy of the
mediating states of the link approaches that of the donor and
acceptor states, a change to a hopping mechanism may occur,
in which the link is actually a reduced or oxidized intermedi-
ate.6,7 In that case, the distance dependence is much weaker
than that for a superexchange. As an example, charge hopping
between guanine bases has recently been shown8 to explain
reports of efficient charge transport over several tens of
angstro¨ms in DNA.8,41
HAB
)
V(Al )V(Bl )/∆E
n
(7)
∑
n
n
As seen in Figure 17, the reaction from 1ZnP is the transfer of
an electron from the ZnP LUMO to the AuP+ LUMO. Because
of the proximity in energy, the LUMO:s of the bridge are likely
to give the largest contribution to the coupling. This is referred
to as the “electron-transfer mechanism”2b,6,14 since the virtual
(41) (a) Kelley, S. O.; Barton, J. K. Science 1999, 283, 375. (b) Nunez, M. E.;
Hall, D. B.; Barton, J. K. Chem. Biol. 1999, 6, 85. (c) Murphy, C. J.; Arkin,
M. R.; Jenkins, Y.; Ghatlia, N. D.; Bossmann, S. H.; Turro, N. J.; Barton,
J. K. Science 1993, 262, 1025.
(42) May, V.; Ku¨hn, O. Charge and Energy Transfer Dynamics in Molecular
Systems; Wiley-VCH: Berlin, 2000.
In contrast to many synthetic systems, the rotaxanes of this
study have a link comprising different types of bonds, and there
9
4358 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
generation of this intermediate (see above), the increased
+
reaction rates show that Cu(phen)2 is involved. We therefore
suggest that the electron transfer in this case proceeds by a
hopping mechanism. A recent study of electron transfer between
ZnP and AuP+ units covalently linked by conjugated hydro-
carbon bridges reports a change from superexchange mediation
to a hopping mechanism when the energy of the bridge state
1
was gradually decreased.44 Only electron transfer from ZnP
was studied, however, so it is not clear if the mediation behavior
with AuP+ excitation would be different.
Finally, for the recombination reaction of the charge-transfer
state, the case is different from both the forward reactions, since
an electron should be transferred from the AuP• LUMO to the
ZnP•+ HOMO, and all states of the link lie at a very different
energy than at least one of these two porphyrin MO:s involved.
We observed no effect of Ag+ or Cu+ in the experiments,
indicating that either the metal-phenanthroline unit was not
involved in this reaction or that the changes induced were too
small to significantly alter the parameters of eq 7.
Figure 17. Schematic MO diagram illustrating the effect of bridge energy
on through bond electron transfer. (A) Electron transfer from the ground-
state ZnP to the 3AuP+ triplet (“hole transfer”). (B) Electron transfer from
1
the singlet ZnP to the ground-state AuP+ (electron transfer).
To conclude, we propose that the electron transfer following
excitation of the AuP+ unit changes from a superexchange in
Zn2/Ag+/Au+, which is enhanced as compared to the case in
Zn2//Au+, to a hopping mechanism for Zn2/Cu+/Au+, involving
the oxidized Cu(phen)22+ unit as a real intermediate. The effect
is a gradual increase in reaction rate throughout the series. In
contrast, when the ZnP unit was excited we observed no effect
of the coordinated metal cations or even a slight rate decrease.
We propose that this can be explained by the different mediating
orbitals of the link involved, HOMO for AuP+ excitation and
LUMO for ZnP excitation. This behavior is in contrast to
previously reported systems with repeated, identical units of
hydrocarbon σ bonds between the reactants, in which a
symmetric HOMO/LUMO superexchange mediation was
observed.2b A mediation asymmetry has been reported for a
covalently linked system, in which a synthetic modification of
the bridge increased the HOMO-mediated “hole transfer”, while
no significant effect on the LUMO-mediated “electron transfer”
was observed.45
intermediate state involved has an additional electron on the
link. In the reaction starting with 3AuP+, however, an electron
is transferred from the ZnP HOMO to the AuP+ HOMO. For
the same energy proximity reason as before, the bridge HOMO:s
will probably give the largest mediating contribution. This is
the “hole-transfer mechanism”2b,6,14 in which the virtual state
involves the oxidized link.
The coordination of metal cations changes the energy of the
phenanthroline MO:s and introduces new metal-based orbitals.
Also, the coupling between the phenanthroline units will be
greatly enhanced when they are coordinated to the same metal
ion. In the electron-transfer mechanism, starting from 1ZnP, the
+
LUMO of the Cu(phen)2 unit is phenanthroline-based, while
Ag(phen)2+ has low lying unoccupied orbitals on both the metal
and the ligand.43 From the reduction potentials in Table 2, it is
clear that the LUMO is stabilized to the same level as that for
the 1ZnP donor state (for Cu+), or even below (for Ag+). If the
electron-transfer pathway involved the metal-phenanthroline
unit, a significant increase in the reaction rate would be expected.
Instead, we observed no change, or even a slight decrease with
Ag+. This suggests that the electron-transfer pathway from the
1ZnP state does not involve these units. For the hole-transfer
mechanism instead, starting from 3AuP+, also the phenanthroline
HOMO is stabilized by the metal cation, so that the energy
difference to the donor state increases. Still, for Zn2/Ag+/Au+
3. Conclusions
By selective excitation of either porphyrin, we could follow
the electron transfer from the ZnP to the AuP+ unit in this series
of rotaxanes. A rapid, photoinduced electron transfer between
the noncovalently linked porphyrins was observed, over an ca.
15-17 Å edge-to-edge distance, with rate constants up to 4 ×
1010 s-1. The resulting ZnP•+-AuP• charge-transfer state had a
relatively long lifetime, 10-40 ns, and was formed in high yield.
For Zn2/Ag+/Au+ it was >80%, irrespective of excitation
wavelength. In Zn2/Cu+/Au+, electron transfer to AuP+ was
competing with a two-step energy-transfer process: (i) from
1ZnP to Cu(phen)2+ with k ) 5.1 × 109 s-1, and then (ii) from
3
the electron transfer starting with the AuP+ state was faster
than for Zn2//Au+, suggesting that the Ag(phen)2+ unit mediated
the reaction. This apparently contradicting behavior can be
explained by the strongly increased coupling between the two
phenanthrolines as they become coordinated to the same metal,
leading to an increase of the V(Aln)V(Bln) factors in the
numerator of eq 7 which compensates for the increase in ∆En.
In Zn2/Cu+/Au+, the Cu(phen)2+ HOMO is Cu-based, and the
redox potentials (Table 2) show that it lies between the donor
and acceptor states. This state would therefore not be involved
in a superexchange mechanism, but could instead be a real
intermediate. Thus, even if there is no direct evidence for the
+
3
*Cu(phen)2 to give ZnP with k ) 1.5 × 109 s-1
.
By a simple variation of the linking structure, viz. the
coordination of either Ag+ or Cu+ to the phenanthroline units,
we could enhance the electron-transfer rate from the ZnP to
3
the excited AuP+. We interpret our data as a change from a
superexchange mechanism for Zn2/Ag+/Au+, which is enhanced
(43) (a) Dietrich-Buchecker, C.; Sauvage, J.-P.; Kern, J.-M. J. Am. Chem. Soc.
1989, 111, 7791. (b) Billon, M.; Divisia-Blohorn, B.; Kern, J.-M.; Sauvage,
J.-P. J. Mater. Chem. 1997, 7, 1169. (c) Schmittel, M.; Ganz, A.; Fenske,
D.; Herderich, M. J. Chem. Soc., Dalton Trans. 2000, 353.
(44) Kilså, K.; Kajanus, J.; Macpherson, A. N.; Mårtensson, J.; Albinsson, B.
J. Am. Chem. Soc. 2001, 123, 3069.
(45) Wasielewski, M. R.; Niemczyk, M. P.; Johnson, D. G.; Svec, W. A.; Minsek,
D. W. Tetrahedron 1989, 45, 4785.
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4359

A R T I C L E S
Andersson et al.
as compared to the case in Zn2//Au+, to a hopping mechanism
(background and scan) were averaged for each curve. The response
time of the whole system, as measured by the rise of an instantaneous
absorption increase, was ca. 5 ns (10%-90% rise), and the time jitter
between different shots was <1 ns.
2+
for Zn2/Cu+/Au+, involving the oxidized Cu(phen)2 unit as
a real intermediate. In contrast, electron transfer from the excited
1ZnP to AuP+ was not affected, or even slowed, by Ag+ or
Cu+. This asymmetry may possibly be explained by the different
orbitals involved in mediating the reaction in an electron- and
a hole-transfer mechanism.6,14 A consideration of the orbital
energies involved suggests that also the electron transfer pathway
is different dependent on which porphyrin is excited. Our results
demonstrate the possibility to tune the rates of electron transfer
between the noncovalently linked reactants by a convenient
modification of the linking structure. The different effect of Ag+
and Cu+ on the rate with ZnP and AuP+ excitation shows an
additional possibility to control the electron-transfer reactions
by selective excitation.
Electrochemical Measurements. Cyclic voltammetry measurements
were performed with an EG&G Princeton Applied Research poten-
tiostat, model 273A. The working electrode was a Pt disk, the counter
electrode was a Pt wire, and a SCE was the reference electrode. The
electrolyte was n-Bu4NPF6 in dichloromethane.
1
Synthesis. H NMR spectra were obtained on either a Bruker WP
200 SY (200 MHz) or an AM 400 (400 MHz) spectrometer. Chemical
shifts in ppm are referenced downfield from tetramethylsilane. Labels
of the protons of the [2]-rotaxanes and some of their precursors are
provided in Figure 5. Mass spectral data were obtained on either a
ZAB-HF (FAB) or a Bruker Protein TOF (MALDI-TOF) spectrometer.
Melting points were determined in open capillary tubes on a Bu¨chi
SMP-20 apparatus and are uncorrected. Elemental analyses were
performed by the “Service de Microanalyse de l’Institut de Chimie de
Strasbourg”. UV-visible absorption spectra were recorded with a
Kontron-Uvikon 860 spectrophotometer.
Schlenk techniques were used for the reactions performed under an
atmosphere of argon. Macrocycles M30,15b Au+,18 2,9-bis-(p-formylphe-
nyl)-1,10-phenanthroline 1,20 3,5-di-tert-butylbenzaldehyde 2,22 and Cu-
(CH3CN)4PF619 were prepared according to literature. A procedure for
the preparation of (3,3′-dihexyl-4,4′-dimethyl-2,2′-dipyrryl)methane 3
is also given in the Supporting Information.
4. Experimental Section
Photophysics and Photochemistry. All photophysical measurements
were made in spectroscopic grade DMF at 298 K, unless otherwise
stated. The samples used for triplet lifetime measurement were
deoxygenated by repeated freeze-pump-thaw cycles or by purging
with nitrogen gas. The triplet emission spectra were measured in a
1
EtOH/ MeOH 4:1 mixture at 77 K. The ZnP state energy of Zn was
calculated from the midpoint of the absorption and emission peaks (the
Q(0,0)-band), while the energy of the nonemitting singlet excited state
of Au+ was calculated from the red end of the lowest absorption peak.
The triplet state energies for Zn and Au+ were calculated from the
blue edge of the phosphorescence peak at 77 K. The redox potentials
are reported as the average between the anodic and cathodic peak
A. Synthesis of (3,3′-Dihexyl-4,4′-dimethyl-2,2′-dipyrryl)methane
3. This is described in the Supporting Information.
B. Synthesis of [2]-Rotaxanes. Prerotaxane [/Cu+/Au+](PF6-)2.
A degassed solution of Cu(CH3CN)4‚BF4 (0.0342 g; 0.11 mmol) in
CH3CN (10 mL) was added via cannula to a solution of macrocycle
Au+ (0.200 g; 0.11 mmol) in dry CH2Cl2 (20 mL). After stirring for
20 min, a solution of 2,9-bis-(p-formylphenyl)-1,10-phenanthroline
(0.042 g; 0.11 mmol) in dry CH2Cl2 was added via cannula to the
solution of the complex formed. After stirring for 2 h, the mixture was
treated with a saturated aqueous solution of KPF6. The organic layer
potentials. The driving force for the electron-transfer reaction was
00
calculated from the Weller equation:46 ∆G° ) zFE0
E
D+/D-
-
D*
zFE0
- ∆w. The work term (∆w) is zero for the charge shift
A/A-
reactions between the ZnP and AuP+ units.
The fluorescence lifetimes were measured with a time correlated
single photon counting setup, with an instrumental response function
of 80 ps. Excitation pulses at 575 nm (150 fs duration, 200 kHz, <10
nJ/pulse) were delivered an OPA, pumped by a regenerative amplifier/
Ti-sapphire oscillator system. Steady-state emission measurements were
made on a Spex Fluorolog spectrometer.
Femtosecond transient absorption measurements were performed
using a setup that has been described in detail elsewhere.47 The pump
light is the 150 fs, 1 kHz tunable output of an optical parametric
amplifier (TOPAS). The intensity of the pump pulses was kept below
3 µJ to avoid multiple excitation of the same rotaxane. As probe light,
a white light continuum created from a 3 mm sapphire window was
used. In our experiments, we used an optical delay line that allows a
maximum delay of 5 ns relative to the pump light.
In the nanosecond flash photolysis, the following instrumentation
was used: a Q-switched Nd:YAG laser (Quantel Brilliant B) pumped
an OPO (Opotek) that delivered ca. 5 ns (fwhm) pulses at 528 or 575
nm (ca. 10 nm bandwidth). The pulses were focused in the sample in
a standard quartz 1 × 1 cm2 cell, via a cylindrical lens, giving ca. 15
mJ/pulse excitation light on a 8 × 3 mm2 spot (w × h). In the flash
photolysis spectrometer (Applied Photophysics LKS60), the analyzing
light from a pulsed, 150 W xenon lamp was focused in the sample
though 2 × 2 mm pinholes, being overlapped by the laser beam in a
cross-beam fashion. After the sample, the analyzing light passed a single
holographic grating monochromator with two 1-2 mm slits and was
focused onto a side-on P928-type PMT. The PMT output was fed into
the 50 Ω input of a HP Infinium digital oscilloscope (2 Gs/s) and
analyzed using the commercial processor software. At least 10 shots
was extracted and evaporated to dryness to give pure [/Cu+/Au+](PF6-
)
2
(0.264 g; 0.11 mmol; 100%) as a red solid. 1H NMR (400 MHz, CD2-
3
Cl2): δ 9.69 (s, 2H, CHO); 9.44 (d, 2H, Hpy1, J ) 5.3 Hz); 9.32 (s,
4H, Hpy3+py4); 9.29 (d, 2H, Hpy2, 3J ) 5.3 Hz); 8.94 (d, 2H, H4′,7′, 3J )
3
8.4 Hz); 8.46 (s, 2H, H5′,6′); 8.43 (d, 2H, H4,7, J ) 8.4 Hz); 8.11 (d,
3
4
2H, H3′,8′, J ) 8.4 Hz); 8.06 (d, 4H, Hop1, J ) 1.7 Hz); 7.92 (t, 1H,
4
3
H
pp2, J ) 1.7 Hz); 7.91 (d, 2H, H3,8, J ) 8.4 Hz); 7.90 (d, 2H, Hop2,
4J ) 2.0 Hz); 7.84 (s, 2H, H5,6); 7.84 (t, 2H, Hpp1, J ) 1.7 Hz); 7.67
(d, 4H, Ho′, 3J ) 8.1 Hz); 7.59 (d, 2H, Hop3, 4J ) 2.0 Hz); 7.22 (d, 4H,
Ho, 3J ) 8.9 Hz); 7.33 (t, 1H, Hpp3 , 4J ) 2.0 Hz); 7.12 (d, 4H, Hm′, 3J
) 8.1 Hz); 6.12 (d, 4H, Hm, 3J ) 8.9 Hz); 4.46 (m, 4H, Hδ); 4.07 (m,
4H, Hγ); 3.84 (m, 8H, HR+â); 1.53 (s, 18H, HtBu2); 1.49 (s, 36H, HtBu1).
4
+
-
FAB-MS: m/z ) 2279.9 [M - PF6 ] , calcd 2279.87, 25%; 2134.5
[M - 2PF6 + e ] , calcd 2134.91, 20%; 1746.0 [Au+ + Cu+ + e-]+,
-
+
-
calcd 1746.43, 95%; 1682.8 [Au+], calcd 1682.94, 100%; 1069.6 [M
- 2PF6 ]2+/2, calcd 1067.45, 84%; 873.3 [Au+ + Cu+]2+/2, calcd 873.0,
-
95%.
[2]-Rotaxane [(H2)2/Cu+/Au+](PF6-)2. Prerotaxane [/Cu+/Au+]-
(PF6-)2 (0.264 g; 0.11 mmol), 3,5-di-tert-butylbenzaldehyde (0.190 g;
0.87 mmol), and 3 (0.345 g; 1.1 mmol) were dissolved in dry CH2Cl2
(100 mL) under argon. Trifluoroacetic acid (four drops) was then added,
and the mixture was stirred at room temperature for 16 h. Subsequently,
chloranil (tetrachloro-p-benzoquinone) (0.800 g; 3.27 mmol) was added,
and the reaction mixture was refluxed for 1.5 h. After neutralization
by NaHCO3, the organic layer was washed three times with water,
treated with a saturated aqueous solution of KPF6, and then evaporated
to dryness. The crude product was purified by repeated column
chromatography on silicagel (eluent: CH2Cl2/0.5% MeOH) and on
(46) Rehm, O.; Weller, A. Ber. Bunsen-Ges. Phys. Chem. 1969, 73, 834.
(47) Andersson, M.; Davidsson, J.; Hammarstro¨m, L.; Korppi-Tommola, J.;
Peltola, T. J. Phys. Chem. B 1999, 103, 3258.
alumina (eluent: CH2Cl2) affording pure [(H2)2/Cu+/Au+](PF6
)
2
-
1
(0.081 g; 0.02 mmol; 18% yield) as a red solid. H NMR (400 MHz,
9
4360 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002

Porphyrin-Containing [2]-Rotaxanes
A R T I C L E S
CD2Cl2): δ 10.16 (s, 4H, Hmeso); 9.55 (d, 2H, Hpy1, 3J ) 5.1 Hz); 9.45
(d, 2H, Hpy2, J ) 5.5 Hz); 9.39 (d, 2H, Hpy3, J ) 5.4 Hz); 9.35 (d,
Hpp2, 3J ) 1.8 Hz); 7.98 (s, 2H, H5′,6′); 7.92 (t, 2H, Hpp1, 3J ) 1.8 Hz);
3
3
3
3
7.86 (t, 1H, Hpp3, J ) 1.8 Hz); 7.84 (d, 4H, Hop′, J ) 1.8 Hz); 7.80
3
3
(t, 2H, Hpp′, 3J ) 1.8 Hz); 7.72 (d, 4H, Hm′, 3J ) 8.1 Hz); 7.70 (t, 2H,
2H, Hpy4, J ) 5.4 Hz); 9.21 (d, 2H, H4′,7′, J ) 8.5 Hz); 8.86 (s, 2H,
5′,6′); 8.52 (d, 2H, H3′,8′, 3J ) 8.4 Hz); 8.32 (d, 2H, H4,7, 3J ) 8.1 Hz);
H
op3, 3J ) 1.8 Hz); 7.53 (d, 2H, H3,8, 3J ) 8.4 Hz); 7.00 (d, 4H, Hm, 3J
H
4
4
8.14 (d, 4H, Hop1, J ) 1.8 Hz); 8.10 (d, 2H, Hop2, J ) 1.8 Hz); 8.07
) 9.0 Hz); 6.72 (d, 2H, H4,7, 3J ) 8.4 Hz); 4.85 (s, 2H, H5,6); 4.44 (m,
4H, Hδ); 4.10 (m, 4H, HR); 4.01 (m, 4H, Hγ); ∼3.9 (m, 8H, HR′2); 3.85
(m, 4H, Hâ); 3.75 (m, 8H, HR′1); 2.38 (s, 12H, HMe2); ∼2.1 (m, 8H,
Hâ′2); 2.03 (s, 12H, HMe1); ∼2.0 (m, 8H, Hâ′1); ∼1.7 (m, 16H, Hγ′1+γ′2);
1.52 (s, 18H, HtBu2); 1.47 (s, 36H, HtBu1); 1.46 (s, 36H, HtBu′); ∼1.3
(d, 4H, Ho′, 3J ) 8.4 Hz); 8.06 (d, 2H, H3,8, 3J ) 8.4 Hz); 7.98 (t, 3H,
4
4
H
pp1+pp2, J ) 1.6 Hz); 7.88 (d, 4H, Hop′, J ) 1.8 Hz); 7.84 (t, 2H,
H
pp′, 4J ) 1.8 Hz); 7.63 (d, 2H, Hop3, 4J ) 2.2 Hz); 7.60 (d, 4H, Ho, 3J
3
) 8.4 Hz); 7.54 (s, 2H, H5,6); 7.46 (d, 4H, Hm′, J ) 8.1 Hz); 7.21 (t,
1H, Hpp3); 6.33 (d, 4H, Hm, J ) 8.4 Hz); 4.44 (m, 4H, Hδ); 4.24 (m,
3
(m, 32H, Hδ′1+δ′2+ꢀ′1+ꢀ′2); 0.9 (m, 24H, HMe hex). FAB-MS: m/z ) 3920.4,
+
[M - PF6]+, calcd 3920.9, 70%; 2238.3, [M - 2PF6 - Au - Cu+
4H, Hγ); 4.08 (m, 16H, HR+â+R′2); ∼3.9 (m, 8H, HR′1); 2.45 (s, 12H,
-
+ H+]+, calcd 2238.98, 22%; 1960.4, [M - PF6 - e ]2+/2, calcd
-
H
Me2); ∼2.2 (m, 8H, Hâ′2); 2.08 (m, 8H, Hâ′1); 1.78 (s, 12H, HMe1);
-
1960.5, 16%; 1682.4, [Au]+, calcd 1682.94, 100%. UV-visible (CH2-
Cl2), λ, nm (ꢀ, M-1 cm-1): 414 (870 000); 538 (39 300); 574 (16 500).
∼1.7 (m, 16H, Hγ′1+γ′2); 1.53 (s, 36H, HtBu1); 1.52 (s, 18H, HtBu2); 1.47
(s, 36H, HtBu′); ∼1.3 (m, 32H, Hδ′1+δ′2+ꢀ′1+ꢀ′2); 0.8 (m, 24H, HMe hex);
-2.57 (s, 2H, NH); -2.67 (s, 2H, NH). FAB-MS: m/z ) 4002.1 [M
[2]-Rotaxane [Zn2/Ag+/Au+](PF6-)2. To a solution of [Zn2//Au+]-
(PF6-)2 (0.004 g; 9.8 × 10-4 mmol) in dry CH2Cl2 was added under
+
- +
-
-
- PF6 ] , calcd 4002.7, 26%; 3856.8 [M - 2PF6 + e ] , calcd 3857.7,
44%; 2173.4 [M - 2PF6 - Au+]+, calcd 2174.7, 24%; 1928.4 [M -
-
argon, via syringe, a solution of AgBF4 (0.75 mL of a 1.5 × 10-3
M
2PF6 ]2+/2, calcd 1928.9, 100%; 1682.8 [Au+], calcd 1682.9, 68%.
-
solution, 1.2 × 10-6 mmol) in CH3CN. After stirring the solution for
1 h, the solvents were evaporated. The residue was taken-up in CH2-
Cl2, washed with water, and treated by an aqueous solution of KPF6.
The crude product was purified by column chromatography on alumina
(eluent: CH2Cl2) giving pure [Zn2/Ag+/Au+](PF6-)2 (0.004 g, 9.7 ×
10-4 mmol, 99%) as a red solid. 1H NMR (400 MHz, CD2Cl2): δ 10.09
(s, 4H, Hmeso); 9.54 (d, 2H, Hpy1, 3J ) 5.2 Hz); 9.44 (d, 2H, Hpy2, 3J )
UV-visible (CH2Cl2), λ, nm (ꢀ, M-1 cm-1): 416 (762 000); 511
(43 600); 543 (18 300); 574 (15 800); 626 (2300).
[2]-Rotaxane [Zn2/Cu+/Au+](PF6-)2. A solution of [(H2)2/Cu+/Au+]-
(PF6-)2 (0.065 g; 0.016 mmol) and Zn(OAc)2‚2H2O (0.032 mmol) in
a mixture of CHCl3 (20 mL) and MeOH (10 mL) was refluxed under
argon for 1 h. After evaporation of the solvent, the residue was taken
up in CH2Cl2, washed twice with water, and evaporated. The crude
product was purified by column chromatography on alumina (eluent:
CH2Cl2) to leave pure [Zn2/Cu+/Au+](PF6-)2 (0.055 g; 0.013 mmol;
3
5.2 Hz); 9.38 (d, 2H, H4′,7′, J ) 8.2 Hz); 9.37 (s, 4H, Hpy3+py4); 8.77
(s, 2H, H5′,6′); 8.58 (d, 2H, H3′,8′, 3J ) 7.9 Hz); 8.14 (d, 2H, H4,7, 3J )
4
4
7.9 Hz); 8.13 (d, 4H, Hop1, J ) 1.8 Hz); 8.10 (d, 2H, Hop2, J ) 1.8
1
82%) as a red solid. H NMR (400 MHz, CD2Cl2): δ 10.12 (s, 4H,
3
Hz); 8.09 (d, 4H, Ho′, J ) 7.9 Hz); 7.98 (m, 3H, Hpp1+pp2); 7.87 (d,
H
meso); 9.56 (d, 2H, Hpy1,3J ) 5.5 Hz); 9.45 (d, 2H, Hpy2,3J ) 5.2 Hz);
4H, Hop′, 4J ) 1.9 Hz); 7.84 (t, 2H, Hpp′, 4J ) 1.8 Hz); 7.76 (d, 4H, Ho,
9.37 (d, 2H, Hpy3,3J ) 5.2 Hz); 9.35 (d, 2H, Hpy4,3J ) 5.1 Hz); 9.27 (d,
2H, H4′,7′, 3J ) 8.6 Hz); 8.86 (s, 2H, H5′,6′); 8.54 (d, 2H, H3′,8′, 3J ) 8.2
3J ) 8.6 Hz); 7.62 (d, 2H, Hop3, J ) 2.1 Hz); 7.54 (d, 4H, Hm′, J )
4
3
8.2 Hz); 7.52 (d, 2H, H3,8, 3J ) 8.2 Hz); 7.33 (s, 2H, H5,6); 7.16 (t, 1H,
Hz); 8.26 (d, 2H, H4,7,3J ) 8.2 Hz); 8.16 (d, 4H, Hop1, J ) 2.1 Hz);
4
4
3
H
pp3, J ) 2.1 Hz); 6.45 (d, 4H, Hm, J ) 8.6 Hz); 4.36 (m, 4H, Hδ),
8.11 (d, 2H, Hop2, 4J ) 1.7 Hz); 8.07 (d, 4H, Ho′, 3J ) 8.2 Hz); 8.04 (d,
4.02 (m, 4H, Hγ); ∼3.9 (m, 12H, Hâ+R′2); 3.87 (m, 4H, HR); 3.70 (t,
3
4
2H, H3,8, J ) 8.2 Hz); 7.95 (t, 3H, Hpp1+pp2); 7.90 (d, 4H, Hop′, J )
3
8H, HR′1, J ) 7.8 Hz); 2.42 (s, 12H, HMe2); ∼2.2 (m, 8H, Hâ′2); 1.94
4
4
1.7 Hz); 7.81 (t, 2H, Hpp′, J ) 1.7 Hz); 7.67 (d, 2H, Hop3, J ) 2.4
(m, 8H, Hâ′1); 1.7 (s, 12H, HMe1); ∼1.7 (m, 16H, Hγ′1+γ′2); 1.54 (s,
36H, HtBu1); 1.50 (s, 18H, HtBu2); 1.49 (s, 36H, HtBu′); ∼1.4 (m, 32H,
Hδ′1+δ′2+ꢀ′1+ꢀ′2); 0.8 (t, 12H, HMe hex). FAB-MS: m/z ) 4173.1 [M -
Hz); 7.58 (d, 4H, Ho, 3J ) 8.6 Hz); 7.55 (s, 2H, H5,6); 7.47 (d, 4H, Hm′,
3J ) 9.0 Hz); 7.20 (t, 1H, Hpp3); 6.35 (d, 4H, Hm, J ) 8.6 Hz); 4.46
3
(m, 4H, Hδ); 4.11 (m, 4H, Hγ); 4.08 (m, 16H, HR+â+R′2); ∼3.97 (m,
8H, HR′1); 2.42 (s, 12H, HMe2); ∼2.15 (m, 8H, Hâ′2); 2.05 (m, 8H, Hâ′1);
1.76 (s, 12H, HMe1); ∼1.68 (m, 16H, Hγ′1+γ′2); 1.52 (s, 36H, HtBu1);
1.50 (s, 18H, HtBu2); 1.49 (s, 36H, HtBu′); ∼1.3 (m, 32H, Hδ′1+δ′2+ꢀ′1+ꢀ′2);
0.89 (t, 12H, HMe2 hex,3J ) 7.2 Hz); 0.89 (t, 12H, HMe1 hex,3J ) 7.2 Hz).
+
- +
-
-
PF6 ] , calcd 4173.7, 43%; 4027.4 [M - 2PF6 + e ] , calcd 4028.8,
46%; 2344.6 [M - 2PF6 - Au+ - Ag+ + H+]+, calcd 2345.8, 22%;
-
+
+
-
-
2236.7 [M - 2PF6 - Au ] , calcd 2237.9, 17%; 2013.5 [M - 2PF6
]
2+/2, calcd 2014.4, 41%; 1789.4 [Au+ + Ag+ + e-]+, calcd 1790.8,
35%; 1682.1 [Au]+, calcd 1682.9, 100%. UV-visible (CH2Cl2), λ, nm
+
-
FAB-MS: m/z ) 4128.8 [M - PF6 ] , calcd 4129.4, 51%; 3983.1 [M
(ꢀ, M-1 cm-1): 415 (884 000); 538 (42 200); 574 (17 000).
- 2PF6 + e-]+, calcd 3984.5, 54%; 2300.0 [M - 2PF6 - Au+]+,
-
-
[2]-Rotaxane [Zn2/Li+/Au+](PF6-)2. To a solution of [Zn2//Au+]-
(PF6-)2 (0.009 g; 2.2 × 10-3 mmol) in dry CH2Cl2 was added, under
argon, a solution of LiBF4 in MeOH (0.7 mL of a 6.3 × 10-3 M, 4.4
10-3 mmol) via syringe. The solution was stirred for 2 h. The solvents
were evaporated, leaving pure [Zn2/Li+/Au+](PF6-)2 (0.009 g; 2.2 ×
calcd 2301.5, 25%; 2238.7 [M - 2PF6 - Au+ - Cu+ + H+]+, calcd
-
2238.98, 13%; 1991.6 [M - 2PF6 ]2+/2, calcd 1992.2, 52%; 1745.2
-
[Au+ + Cu+ + e-]+, calcd 1682.9, 91%; 1684.4 [Au+], calcd 1682.9,
91%. UV-visible (CH2Cl2), λ, nm (ꢀ, M-1 cm-1): 415 (890 000); 538
(44 300); 574 (19 600).
1
10-3 mmol; 100%) as a red solid. H NMR (400 MHz, CD2Cl2): δ
[2]-Rotaxane [Zn2//Au+](PF6-)2. A solution of rotaxane [Zn2/Cu+/
Au+](PF6-)2 (0.030 g; 0.007 mmol) and KCN (22 mg; 0.35 mmol) in
a mixture of CHCl3/H2O/CH3CN, 1/1/4 (3 mL), was heated at 40 °C
for 30 min. Solvents were evaporated, and the residue was taken-up in
CH2Cl2. The organic layer was washed twice with water, treated with
a saturated aqueous solution of KPF6, and evaporated to dryness. The
crude product was purified by column chromatography on alumina
(eluent: hexane/CH2Cl2, 50/50) affording pure [Zn2//Au+](PF6-)2 and
some starting material, which was subjected to the same reaction
conditions again. This procedure was repeated five times. The overall
yield after a total of five reactions was 92%. [Zn2//Au+](PF6-)2 (0.026
10.11 (s, 4H, Hmeso); 9.56 (d, 2H, Hpy1, 3J ) 5.2 Hz); 9.45 (d, 2H, Hpy2
,
3
3J ) 5.2 Hz); 9.37 (s, 2H, Hpy3+py4); 9.21 (d, 2H, H4′,7′, J ) 8.6 Hz);
3
8.89 (s, 2H, H5′,6′); 8.48 (d, 2H, H3′,8′, J ) 8.5 Hz); 8.23 (d, 2H, H4,7
,
4
3
3J ) 8.2 Hz); 8.13 (d, 4H, Hop1, J ) 1.8 Hz); 8.12 (d, 2H, H3,8, J )
4
8.2 Hz); 8.10 (d, 2H, Hop2, J ) 1.8 Hz); 7.98 (m, 3H, Hpp1+pp2); 7.97
(d, 4H, Ho′, 3J ) 8.0 Hz); 7.88 (d, 4H, Hop′, 4J ) 1.8 Hz); 7.86 (m, 2H,
4
3
H
pp′); 7.63 (d, 2H, Hop3, J ) 2.1 Hz); 7.58 (d, 4H, Ho, J ) 8.8 Hz);
7.44 (d, 4H, Hm′, 3J ) 8.0 Hz); 7.39 (s, 2H, H5,6); 7.19 (t, 1H, Hpp3, 4J
) 2.4 Hz); 6.36 (d, 4H, Hm, 3J ) 8.6 Hz); 4.42 (m, 4H, Hδ); 4.06 (m,
4H, Hγ); 3.99 (m, 4H, Hâ); 3.92 (m, 12H, HR+R′2); 3.78 (m, 8H, HR′1);
2.41 (s, 12H, HMe2); ∼2.15 (m, 8H, Hâ′2); 2.04 (m, 8H, Hâ′1); 1.75 (s,
12H, HMe1); ∼1.7 (m, 16H, Hγ′1+γ′2); 1.53 (s, 18H, HtBu2); 1.52 (s, 36H,
1
mg; 0.0063 mmol), red solid. H NMR (400 MHz, CD2Cl2): δ 10.03
3
3
(s, 4H, Hmeso); 9.48 (d, 8H, Ho′, J ) 8.1 Hz); 9.42 (d, 2H, Hpy1, J )
5.1 Hz); 9.36 (d, 2H, Hpy3, J ) 5.1 Hz); 9.34 (d, 2H, Hpy4,3J ) 5.1
3
H
tBu1); 1.48 (s, 36H, HtBu′); ∼1.3 (m, 32H, Hδ′1+δ′2+ꢀ′1+ꢀ′2); 0.85 (t, 12H,
7
Hz); 9.24 (d, 2H, Hpy2, 3J ) 5.1 Hz); 8.90 (d, 2H, H3′,8′, 3J ) 8.4 Hz);
HMe hex). Li NMR (400 MHz, CDCl3/LiCl, D2O): δ 2.79 (s). UV-
visible (CH2Cl2), λ, nm (ꢀ, M-1 cm-1): 415 (878 000); 538 (37 400);
574 (16 700).
3
8.63 (d, 4H, Ho, J ) 8.7 Hz); 8.57 (d, 2H, H4′,7′, 3J ) 8.4 Hz); 8.08
(d, 2H, Hop2, 3J ) 1.8 Hz); 8.03 (d, 4H, Hop1, 3J ) 1.8 Hz); 7.99 (t, 1H,
9
J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002 4361

A R T I C L E S
Andersson et al.
Scheme 1
Zn(OAc)2‚2H2O, using the same experimental conditions as for
[2]-rotaxane [Zn2/Cu+/Au+](PF6-)2. 1H NMR (400 MHz, CD2Cl2): δ
10.18 (s, 4H, Hmeso); 9.07 (d, 2H, H4′,7′, J ) 8.5 Hz); 8.57 (s, 2H,
3
H
5′,6′); 8.42 (d, 2H, H3′,8′, 3J ) 8.3 Hz), 8.36 (d, 2H, H4,7, 3J ) 8.3 Hz);
8.10 (d, 2H, H3,8, 3J ) 8.3 Hz); 8.03 (d, 4H, Ho′, 3J ) 8.1 Hz); 7.89 (d,
4
3
4H, Hop′, J ) 1.8 Hz); 7.86 (t, 2H, Hpp′, J ) 1.7 Hz); 7.55 (d, 4H,
Ho,3J ) 8.8 Hz); 7.53 (s, 2H, H5,6); 7.44 (d, 4H, Hm′, J ) 8.1 Hz);
3
6.24 (d, 4H, Hm, 3J ) 8.6 Hz); 3.92 (s, 4H, Hꢀ); ∼3.75 (m, 32H,
HR′1+R′2+R+â+γ+δ); 2.47 (s, 12H, HMe2); ∼2.13 (m, 16H, Hâ′1+â′2); 1.81
(s, 12H, HMe1); ∼1.53 (m, 48H, Hγ′1+γ′2+δ′1+δ′2+ꢀ′1+ꢀ′2); 1.51 (s, 36H,
HtBu′); 0.89 (t, 24H, HMe hex, J ) 7.1 Hz).
Acknowledgment. We thank J. Korppi-Tommola for the loan
C. Synthesis of [2]-Rotaxane [Zn2/Cu+/](PF6-)2. Prerotaxane
[/Cu+/](PF6-)2. The synthesis of this molecule was achieved using the
same experimental conditions as for the prerotaxane [/Cu+/Au+](PF6-)2,
with the following quantities: solution of macrocycle M30 (0.100 g;
0.17 mmol) in CH2Cl2 (5 mL), solution of [Cu(CH3CN)4]PF6 (0.66 g;
0,18 mmol) in CH3CN (3 mL), and solution of 2,9-bis-(p-formylphenyl)-
1,10-phenanthroline (0.068 g; 0.17 mmol) in CH2Cl2 (5 mL). [/Cu+/]
(PF6-)2 was obtained quantitatively as a red-brown solid (Scheme 1).48
of an optical delay line, R. Hueber for the FAB mass spectra,
and J.-D. Sauer and M. Martigneaux for the high-field NMR
spectra. Support from the European TMR program (network
contract CT96-0031), the Wallenberg Foundation, and the
Swedish Research Council for Engineering Sciences is gratefully
acknowledged. M.L. thanks the French Ministry of Science for
a fellowship.
+
-
[2]-Rotaxane [Zn2/Cu /](PF6 )2. The synthesis of this molecule was
Supporting Information Available: Pyrrole and dipyr-
romethane synthesis description, cyclic voltammograms of Zn,
Au+, and Zn2/Cu+/Au+, ground-state absorption spectra of Zn
and Au+, and transient absorption spectra for Zn2/Cu+/Au+
after excitation of the AuP+ unit (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
achieved using the same experimental conditions as for the prerotaxane
[(H2)2/Cu+/Au+](PF6-)2, with the following quantities: [/Cu+/](PF6
)
2
-
(0.195 g, 0.17 mmol), 3 (0.538 g; 1.70 mmol), 3,5-di-tert-butyl-
benzaldehyde (0.296 g; 1.36 mmol), and CH2Cl2 (100 mL). After
workup, the crude intermediate [(H2)2/Cu+/](PF6-)2 was metalated with
(48) Chambron, J.-C.; Heitz, V.; Sauvage, J.-P. Bull. Soc. Chim. Fr. 1995, 132,
340.
JA0119907
9
4362 J. AM. CHEM. SOC. VOL. 124, NO. 16, 2002