Probing flexibility in porphyrin-based molecular wires using double electron electron resonance

Article abstract of DOI:10.1021/ja905796z

A series of butadiyne-linked zinc porphyrin oligomers, with one, two, three, and four porphyrin units and lengths of up to 75 A, have been spin-labeled at both ends with stable nitroxide TEMPO radicals. The pulsed EPR technique of double electron electron

Full text of DOI:10.1021/ja905796z

Published on Web 09/08/2009
Probing Flexibility in Porphyrin-Based Molecular Wires Using
Double Electron Electron Resonance
Janet E. Lovett,*,†,‡ Markus Hoffmann,§ Arjen Cnossen,§ Alexander T. J. Shutter,§
Hannah J. Hogben,§ John E. Warren,⊥ Sofia I. Pascu,|| Christopher W. M. Kay,#
Christiane R. Timmel,†,§ and Harry L. Anderson*,§
Centre for AdVanced Electron Spin Resonance, Inorganic Chemistry Laboratory, UniVersity of
Oxford, South Parks Road, Oxford OX1 3QR, United Kingdom, Sir William Dunn School of
Pathology, UniVersity of Oxford, South Parks Road, Oxford OX1 3RE, United Kingdom,
Deptartment of Chemistry, Chemistry Research Laboratory, UniVersity of Oxford, Mansfield
Road, Oxford, OX1 3TA, United Kingdom, Synchrotron Radiation Source, Daresbury
Laboratory, Warrington WA4 4AD, United Kingdom, Department of Chemistry, UniVersity of
Bath, Bath BA2 7AY, United Kingdom, and Institute of Structural and Molecular Biology and
London Centre for Nanotechnology, UniVersity College London, Gower Street,
London WC1E 6BT, United Kingdom
Received July 13, 2009; E-mail: janet.lovett@path.ox.ac.uk; harry.anderson@chem.ox.ac.uk
Abstract: A series of butadiyne-linked zinc porphyrin oligomers, with one, two, three, and four porphyrin
units and lengths of up to 75 Å, have been spin-labeled at both ends with stable nitroxide TEMPO radicals.
The pulsed EPR technique of double electron electron resonance (DEER) was used to probe the distribution
of intramolecular end-to-end distances, under a range of conditions. DEER measurements were carried
out at 50 K in two types of dilute solution glasses: deutero-toluene (with 10% deutero-pyridine) and deutero-
o-terphenyl (with 5% 4-benzyl pyridine). The complexes of the porphyrin oligomers with monodentate ligands
(pyridine or 4-benzyl pyridine) principally adopt linear conformations. Nonlinear conformations are less
populated in the lower glass-transition temperature solvent. When the oligomers bind star-shaped
multidentate ligands, they are forced to bend into nonlinear geometries, and the experimental end-to-end
distances for these complexes match those from molecular mechanics calculations. Our results show that
porphyrin-based molecular wires are shape-persistent, and yet that their shapes can deformed by binding
to multivalent ligands. Self-assembled ladder-shaped 2:2 complexes were also investigated to illustrate
the scope of DEER measurements for providing structural information on synthetic noncovalent nanostructures.
X-ray scattering5 techniques, it remains difficult to probe the
conformations of molecules in solution on the 2 to 10 nm length-
Introduction
Long π-conjugated oligomers, or molecular wires, are widely
scale. Fo¨rster resonance energy transfer (FRET) is widely used
to monitor the conformations of biopolymers on this length-
scale, after attachment of suitable donor and acceptor chro-
mophores.6 However, it would be difficult to use FRET to
measure the distance between two points on a molecular wire
because the π-system of the wire would participate in the energy
transfer process and differential labeling is typically required.
Double electron electron resonance (DEER) is a pulsed EPR
technique for measuring distances between paramagnetic centers,
using dipole-dipole coupling.7-11 If the distance, R, between
two localized organic radical centers is more than about 1.5
investigated because of their ability to mediate electron transfer,1
and because of their strong nonlinear optical behavior.2,3 To
understand the structure-property relationships governing these
molecules, we need information on their three-dimensional
shapes. It is often assumed that molecular wires are rigid units,
but as they become longer they inevitably become more flexible.
Despite recent advances in small-angle neutron scattering4 and
† Centre for Advanced Electron Spin Resonance, University of Oxford.
‡ Sir William Dunn School of Pathology, University of Oxford.
§ Department of Chemistry, University of Oxford.
⊥ Daresbury Laboratory.
(5) Easwaramoorthi, S.; Jang, S. Y.; Yoon, Z. S.; Lim, J. M.; Lee,
C.-W.; Mai, C.-L.; Liu, Y.-C.; Yeh, C.-Y.; Vura-Weis, J.; Wasielews-
ki, M. R.; Kim, D. J. Phys. Chem. A 2008, 112, 6563–6570.
(6) Hass, E. ChemPhysChem 2005, 6, 858–870.
|| University of Bath.
# University College London.
(1) Tao, N. J. Nat. Nanotechnol. 2006, 1, 173–181.
(2) Tykwinski, R. R.; Gubler, U.; Martin, R. E.; Diederich, F.; Bosshard,
C.; Gunter, P. J. Phys. Chem. B 1998, 102, 4451–4465.
(3) Pawlicki, M.; Collins, H. A.; Denning, R. G.; Anderson, H. L. Angew.
Chem., Int. Ed. 2009, 48, 3244–3266.
(7) Larsen, R. G.; Singel, D. J. J. Chem. Phys. 1993, 98, 5134–5146.
(8) Milov, A. D.; Maryasov, A. G.; Tsvetkov, Y. D. Appl. Magn. Reson.
1998, 15, 107–143.
(9) Tsvetkov, Y. D.; Milov, A. D.; Maryasov, A. G. Russ. Chem. ReV.
2008, 77, 487–520.
(4) Lei, S.; Ver Heyen, A.; De Feyter, S.; Surin, M.; Lazzaroni, R.;
Rosenfeldt, S.; Ballauff, M.; Lindner, P.; Mo¨ssinger, D.; Ho¨ger, S.
Chem.sEur. J. 2009, 15, 2518–2535.
(10) Schiemann, O.; Prisner, T. F. Q. ReV. Biophys. 2007, 40, 1–53.
(11) Jeschke, G.; Polyhach, Y. Phys. Chem. Chem. Phys. 2007, 9, 1895–
1910.
9
13852 J. AM. CHEM. SOC. 2009, 131, 13852–13859
10.1021/ja905796z CCC: $40.75  2009 American Chemical Society

Flexibility in Porphyrin-Based Molecular Wires
A R T I C L E S
nm, then the dipole-dipole coupling energy is inversely
proportional to R.3 The DEER technique measures the coupling
frequency by monitoring how a refocused echo, at one
microwave frequency, is affected by a 180° microwave pulse
at a second frequency. DEER has been widely used to measure
distances in the 1.5-7.5 nm range in peptides, proteins, protein
complexes, and polynucleotides,9–17 and it is a promising
technique for probing the conformations of molecular wires18-21
and other rodlike molecules.22-25
their electronic delocalization.33,37,38 Coordination to multiden-
tate ligands can also be used to bend porphyrin-based molecular
wires, as illustrated by the template-directed synthesis of
porphyrin nanorings,39,40 which leads to the question of whether
these molecules should be regarded as shape-persistent. Here
we report a series of DEER distance measurements on butadiyne-
linked porphyrin oligomers terminated with stable TEMPO
radicals, P1-P4 (Figure 1). The results show that they are
shape-persistent, and yet that their shapes can readily be
deformed by binding to ligands such as L1 and L2. Assembly
of ladder complexes by binding oligomers P2 and P3 with ligand
L3 has also been confirmed by DEER distance measurements.
The crystal structure of a bis-TEMPO porphyrin monomer P1′
(identical to P1 except with different solubilizing side chains)
has been solved and the crystallographic radical-radical distance
compares well with the DEER results.
The high polarizability and electronic delocalization of the
porphyrin macrocycle make it an ideal unit from which to
construct molecular wires,26-29 and conjugated porphyrin oli-
gomers exhibit a range of wirelike behavior including long-
range charge transport,30-33 high conductance,34 rapid exciton
migration,35 and large two-photon absorption cross-sections.36,37
Supramolecular self-assembly can be used to control the
torsional angles in these molecular wires and thus to control
Methods and Materials
Synthesis. Full experimental details for the synthesis and
characterization of compounds P1-P4, P1′, L1 and L3 are provided
in the Supporting Information. The synthesis of L2 has been
reported previously.39
(12) Jeschke, G.; Bender, A.; Paulsen, H.; Zimmermann, H.; Godt, A. J.
Magn. Reson. 2004, 169, 1–12.
(13) Denysenkov, V. P.; Prisner, T. F.; Stubbe, J.; Bennati, M. Proc. Natl.
Acad. Sci. 2006, 103, 13386–13390.
UV-Vis-NIR Titrations. The stability constants and stoichi-
ometries of the complexes P2·L1, P4·L2, and (P2)2 ·(L3)2 were
measured by UV-vis-NIR titration in toluene at 298 K. See the
Supporting Information for titration spectra and binding isotherms.
X-ray Crystallography. Crystals of the monomer with 3,5-di-
tert-butylphenyl side chains, P1′, were grown by vapor diffusion
of methanol into a solution of the porphyrin in dioxane. X-ray
diffraction data were collected at the synchrotron at Daresbury UK.
See the Supporting Information for further details; CCDC 725782.
Molecular Dynamics (MD) Calculations. Molecular dynamics
simulations were performed in HyperChem with a temperature of
180 K without explicit solvent for a simulation period of 500 ps.
The X-ray crystal structure of porphyrin monomer P1′ was used
to refine the MM+ force field parameters. More details are given
in the Supporting Information.
DEER Experiments. The oligomers without templates were
dissolved in d8-toluene with 10% d5-pyridine (d8 -PhMe/Py) and
d14-o-terphenyl with 5% 4-benzylpyridine (d14-oTP/BnPy). The
monodentate ligands Py and BnPy were added to coordinate the
zinc centers to prevent aggregation. The complexes of L1, L2, and
L3 were measured in d8-PhMe.
The four-pulse DEER sequence was performed at X-band (∼9.3
GHz) with a Bruker ELEXSYS 680 EPR spectrometer. The tubes
with frozen sample were inserted into a 3 mm split ring resonator
(Bruker EN 4118X-MS3). The experiments were conducted at 50
K using a (π)/(2)-τ1-π-(τ1 + t)-πpump-(τ2 - t)-π-τ2-echo
pulse sequence with 32 ns pulses at the observer frequency and a
12 ns pump frequency pulse. The frequency of the πpump pulse was
set to the maximum of the nitroxide echo detected field sweep
spectrum and the observer frequency was offset by 65 MHz (see
the Supporting Information for exceptions). A ( phase cycle was
applied to the first observer pulse to remove receiver offsets. The
τ1 was set at 400 ns for deuterated solvents and was stepped eight
times in 56 ns increments to average out deuterium modulations
from hyperfine coupling to the solvent matrix. τ2 ranged between
4.5 and 16 µs. The time, t, was incremented in either 8 or 16 ns
steps and hence each time trace normally consisted of between 500
(14) Georgieva, E. R.; Ramlall, T. F.; Borbat, P. P.; Freed, J. H.; Eliezer,
D. J. Am. Chem. Soc. 2008, 130, 12856–12857.
(15) Alexander, N.; Bortolus, M.; Al-Mestarihi, A.; Mchaourab, H.; Meiler,
J. Structure 2008, 16, 181–195.
(16) Kay, C. W. M.; El Mkami, H.; Cammack, R.; Evans, R. W. J. Am.
Chem. Soc. 2007, 129, 4868–4869.
(17) Banham, J. E.; Baker, C. M.; Ceola, S.; Day, I. J.; Grant, G. H.;
Groenen, E. J. J.; Rodgers, C. T.; Jeschke, G.; Timmel, C. R. J. Magn.
Reson. 2008, 191, 202–218.
(18) Martin, R. E.; Pannier, M.; Diederich, F.; Gramlich, V.; Hubrich, M.;
Spiess, H. W. Angew. Chem., Int. Ed. 1998, 37, 2834–2837.
(19) Godt, A.; Schulte, M.; Zimmermann, H.; Jeschke, G. Angew. Chem.,
Int. Ed. 2006, 45, 7560–7564.
(20) Margraf, D.; Bode, B. E.; Marko, A.; Schiemann, O.; Prisner, T. F.
Mol. Phys. 2007, 105, 2153–2160.
(21) Bode, B. E.; Margraf, D.; Plackmeyer, J.; Du¨rner, G.; Prisner, T. F.;
Schiemann, O. J. Am. Chem. Soc. 2007, 129, 6736–6745.
(22) Pornsuwan, S.; Bird, G.; Schafmeister, C. E.; Saxena, S. J. Am. Chem.
Soc. 2006, 128, 3876–3877.
(23) Bird, G. H.; Pornsuwan, S.; Saxena, S.; Schafmeister, C. E. ACS Nano
2008, 2, 1857–1864.
(24) Lovett, J. E.; Bowen, A. M.; Timmel, C. R.; Jones, M. W.; Dilworth,
J. R.; Caprotti, D.; Bell, S. G.; Wong, L. L.; Harmer, J. Phys. Chem.
Chem. Phys. 2009, 11, 6840–6848.
(25) Raitsimring, A. M.; Gunanathan, C.; Potapov, A.; Efremenko, I.;
Martin, J. M. L.; Milstein, D.; Goldfarb, D. J. Am. Chem. Soc. 2007,
129, 14138–14139.
(26) Crossley, M. J.; Burn, P. L. J. Chem. Soc., Chem. Commun. 1991,
1569–1571.
(27) Lin, V. S.-Y.; DiMagno, S. G.; Therien, M. J. Science 1994, 264,
1105–1111.
(28) Anderson, H. L. Chem. Commun. 1999, 2323–2330.
(29) Tsuda, A.; Osuka, A. Science 2001, 293, 79–82.
(30) Kang, B. K.; Aratani, N.; Lim, J. K.; Kim, D.; Osuka, A.; Yoo, K.-H.
Chem. Phys. Lett. 2005, 412, 303–306.
(31) Susumu, K.; Frail, P. R.; Angiolillo, P. J.; Therien, M. J. J. Am. Chem.
Soc. 2006, 128, 8380–8381.
(32) Winters, M. U.; Dahlstedt, E.; Blades, H. E.; Wilson, C. J.; Frampton,
M. J.; Anderson, H. L.; Albinsson, B. J. Am. Chem. Soc. 2007, 129,
4291–4297.
(33) Grozema, F. C.; Houarner-Rassin, C.; Prins, P.; Siebbeles, L. D. A.;
Anderson, H. L. J. Am. Chem. Soc. 2007, 129, 13370–13371.
(34) Sedghi, G.; Sawada, K.; Esdaile, L. J.; Hoffmann, M.; Anderson, H. L.;
Bethell, D.; Haiss, W.; Higgins, S. J.; Nichols, R. J. J. Am. Chem.
Soc. 2008, 130, 8582–8583.
(38) Screen, T. E. O.; Thorne, J. R. G.; Denning, R. G.; Bucknall, D. G.;
Anderson, H. L. J. Am. Chem. Soc. 2002, 124, 9712–9713.
(39) Hoffmann, M.; Ka¨rnbratt, J.; Chang, M.-H.; Herz, L. M.; Albinsson,
B.; Anderson, H. L. Angew. Chem., Int. Ed. 2008, 47, 4993–4996.
(40) Hoffmann, M.; Wilson, C. J.; Odell, B.; Anderson, H. L. Angew.
Chem., Int. Ed. 2007, 46, 3122–3125.
(35) Chang, M.-H.; Hoffmann, M.; Anderson, H. L.; Herz, L. M. J. Am.
Chem. Soc. 2008, 130, 10171–10178.
(36) Drobizhev, M.; Stepanenko, Y.; Dzenis, Y.; Karotki, A.; Rebane, A.;
Taylor, P. N.; Anderson, H. L. J. Phys. Chem. B. 2005, 109, 7223–
7236.
(37) Drobizhev, M.; Stepanenko, Y.; Rebane, A.; Wilson, C. J.; Screen,
T. E. O.; Anderson, H. L. J. Am. Chem. Soc. 2006, 128, 12432–12433.
(41) Taylor, P. N.; Huuskonen, J.; Rumbles, G.; Aplin, R. T.; Williams,
E.; Anderson, H. L. Chem. Commun. 1998, 909–910.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13853

A R T I C L E S
Lovett et al.
Figure 1. Structures of the TEMPO-terminated porphyrin oligomers P1-P4 and ligands L1-L3 [Ar ) 3,5-bis(octyloxy)phenyl]. The lengths marked on
P1-P4 are the distances between the centers of the N-O bonds in the fully extended molecules (RXtal), estimated from the crystal structures of P1′ and a
butadiyne-linked porphyrin dimer (Zn-Zn distance: 13.53 Å).41
and 1030 data points. The result was averaged by multiple scanning
until an acceptable signal-to-noise ratio was acquired, which took
between 3 and 24 h.
degree of freedom value between the value belonging to the best
fit and the value plus one. The fits were noticeably poor when the
parameters corresponding to the extremes of the uncertainties were
used.
The data were background corrected using a three-dimensional
background function in DeerAnalysis2006.42 Some DEER time
traces needed to be truncated before background correction since
pulse overlap was evident (manifested as an artificial upturn in the
time trace intensity). The Tikhonov regularization results from
DeerAnalysis2006 are shown in the Supporting Information,
however the distance distributions from this approach can be poorly
determined. A program in Matlab was therefore developed that
assumes a predetermined model for the distance distribution, unlike
the Tikhonov regularization procedure. The models employed are
presented in the relevant Results and Discussion sections. To correct
for noise levels in the data, the time traces were normalized to unit
intensity at time ) 0 by fitting a low-order polynomial to the early
part of each curve. Analyzing the data without this correction can
lead to results with erroneously short distances. The new program
utilizes the pcf2deer function in DeerAnalysis2006 which transforms
the predetermined distribution model into a DEER time trace for
fitting to the experimental data.42 Best fits to the DEER time traces
were given by those corresponding to the smallest ꢀ2. The ꢀ2 is the
sum over all points of the squared difference between the
experimental and simulated data divided by the error for each point.
The error associated with each point in the experiment was the
overall linear best fit to the variance between each point and the
mean of the 6 surrounding points ((3). The uncertainty in the fitted
parameters were estimated from collating all fits that had a ꢀ2 per
Results and Discussion
Synthesis and Crystallography. The TEMPO-terminated
porphyrin oligomers P1-P3 were synthesized by Sonogashira
coupling of 4-iodobenzoic acid TEMPO ester to alkyne-
terminated porphyrin oligomers as shown in Scheme 1. In the
case of the tetramer, this route generated byproducts which were
difficult to remove, so P4 was prepared by an alternative route
involving Glaser coupling of a mono-TEMPO-terminated por-
phyrin dimer (see the Supporting Information). Oligomers
P1-P4 have flexible side chains [Ar ) 3,5-bis(octyloxy)phenyl]
for high solubility. A TEMPO-terminated porphyrin monomer
with t-butyl side chains (Ar ) 3,5-di-tert-butylphenyl), P1′, was
also synthesized for crystallographic studies. This compound
crystallized with two axial dioxane ligands coordinated to the
zinc center (Zn-O distances 2.46 and 2.49 Å). The structure is
shown in Figure 2. The 6-coordinate zinc atom is in the plane
of the porphyrin and the 25-atom porphyrin core is planar
(deviation from mean plane: ( 0.125 Å). The meso-aryl
substituents are twisted at an angle of 63° to the mean plane of
the porphyrin, whereas the conjugated 4-ethynylbenzoate sub-
stituents are more nearly coplanar with the porphyrin (torsional
angles: 20 and 31°). The TEMPO groups have regular chair
geometries, with both oxygen substituents in equatorial posi-
(42) Jeschke, G.; Chechik, V.; Ionita, P.; Godt, A.; Zimmermann, H.;
Banham, J.; Timmel, C. R.; Hilger, D.; Jung, H. Appl. Magn. Reson.
2006, 30, 473–498.
9
13854 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

Flexibility in Porphyrin-Based Molecular Wires
A R T I C L E S
Scheme 1. Synthesis of TEMPO-Labeled Wiresa
Figure 3. Distance profiles from molecular dynamics simulations (circles)
and the WLC fits (see eqs 1 and 2) for each of the wire molecules. p(R) is
the probability density that the distance between the centers of the N-O
bonds is R.
a Reagents and conditions: (a) Pd2(dba)3, CuI, PPh3, NEt3, PhMe,
40° C.
Table 1. Most-Probable End-to-End Distance (R0) Results from
Fitting the MD Simulations and DEER Experiments with the WLC
Model Compared to the Lengths Estimated from Crystal
Structures, R
(see Figure 1)
Xtal
RXtal
(nm)
R0
(MD) (nm)
R0
(DEER,d8-PhMe/Py) (nm)
R0
(DEER,d14-oTP/BnPy) (nm)
wire
P1
P2
P3
P4
3.44
4.79
6.14
7.50
3.40
4.68
5.96
7.20
3.36
4.64
5.88
7.34
3.36
4.60
5.78
7.18
the loss of discretization. This is obtained by converting the
infinite sum based on eq 1 (which uses the first two terms) to
an integral, solving and using the first two terms of the solution
to give eq 2.
Figure 2. Structure of TEMPO-terminated monomer P1′ in the crystal
(50% probability ellipsoids, omitting hydrogens and coordinated dioxanes
for clarity). The intramolecular radical-radical distance, defined as the
distance between the centers of the N-O bonds, is 34.36 ( 0.01 Å.
-1
κ
1
p(R) )
e-(4κ(1 - r))
- 2 +
[
(
)
8(πκ(1 - r))3/2
κ(1 - r)
-1
9
e-9(4κ(1 - r))
- 2
tions. The intramolecular radical-radical distance, defined as
the distance between the centers of the N-O bonds, is 34.36 (
0.01 Å.
]
(
)
κ(1 - r)
(2)
Molecular Dynamics Simulations. The spin-to-spin distances,
R, in each conformation of each oligomer was measured as the
distance between the centers of the N-O bonds. The distance
distributions given by binning molecular dynamic (MD) tra-
jectories are shown in Figure 3. These are very well fitted by
assuming a semiflexible polymer based on the Kratky-Porod
worm-like chain (WLC) model using equations derived by
Wilhelm and Frey, which assume a continuously bendable
system.43 The two fitting parameters are the contour length, L,
and the persistence length, Lp. The model is based on the
statistics for a self-avoiding chain: when L , Lp, the chain tends
toward a rigid rod, whereas when L . Lp, correlations are lost
and the polymer becomes better described as a Gaussian chain.
If κ is Lp/L and r is R/L, when κ (1-r) > 0.2 the probability
density p(R) of the spin-to-spin distance R can be approximated
by the two terms in eq 142,43
The WLC model fits to the molecular dynamics results are
shown as continuous curves in Figure 3 and the most probable
distances from these fits, R0 (MD) are presented in Table 1.
DEER Results for Wires Bound to Monodentate Ligands. The
oligomers P1 to P4 were measured with DEER in d8-PhMe/Py
and d14-oTP/BnPy solvents. All these experiments were carried
out at 50 K, but the temperature determining the Boltzmann
distribution of conformations is probably the glass-transition
temperature of the solvent (120 K for d8-PhMe/Py and 250 K
for d14-oTP/BnPy).44,45 The DEER time traces are shown in
Figure 4a (d14-oTP/BnPy) and Figure 4c (d8-PhMe/Py). The
shapes of the Pake patterns obtained by Fourier transforming
the DEER time traces (see the Supporting Information) do not
indicate any exchange coupling; the data appear to result entirely
from through-space dipole-dipole coupling.46-48 The results
were fitted with the WLC model described above giving the
2
2
2κ
4π
p(R) )
(π2e-κπ (1-r) - 4π2e-4κπ (1-r)
)
(1)
(44) Murthy, S. S. N.; Gangasharen Nayak, S. K. J. Chem. Soc., Faraday
Trans. 1993, 89, 509–514.
whereas toward the rigid chain limit when κ(1 - r) e 0.2 (i.e.,
r f 1), a more complicated expression is required because of
(45) The glass-transition temperatures of oTP and toluene are 247.7 and
119.4 K, respectively.44 DSC experiments showed that the glass-
transition temperature of 5% 4-benzyl pyridine in d14-oTP is the same
as that of pure oTP. We assume that the presence of 10% pyridine in
toluene has a negligible effect on the glass-transition temperature.
(43) Wilhelm, J.; Frey, E. Phys. ReV. Lett. 1996, 77, 2581–2584.
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13855

A R T I C L E S
Lovett et al.
Figure 4. DEER results for the wires P1-P4: (a) experimental and fitted time traces in d14-oTP/BnPy; (b) distance results in d14-oTP/BnPy; (c) experimental
and fitted time traces in d8-PhMe/Py; (d) distance results in d8-PhMe/Py including the Gaussian line shape contributions for P2-P4.
distance profiles plotted in Figure 4b and Figure 4d. The data
for solutions in d14-oTP/BnPy fit extremely well to this model,
resulting in the modal lengths R0 listed in Table 1, which agree
with the MD simulations. The oligomers dissolved in d8-PhMe/
Py exhibit a component with a fast decay which corresponds
to a broad distribution of short distances. We attribute this effect
to some disordered aggregation (see the Supporting Informa-
tion). To fit the WLC model to the data, we included a variable
Gaussian line shape to take account of this faster frequency
component (the Gaussian parameters were held constant to
estimate the uncertainty in the WLC parameters). The modal
distance results are summarized in Table 1 and plotted in Figure
5 (for more information including uncertainties, see the Sup-
porting Information). Comparing the DEER data from P1-P4
dissolved in d14-oTP/BnPy and d8-PhMe/Py, it can be seen that
Figure 5. R0 for the single-strand wires plotted against the number of
the spin-labeled wires are more rigid when frozen in the lower
glass-transition temperature d8-PhMe/Py.49 For example, con-
trasting the P1 time traces in panels a and c in Figure 4 for the
two solvent systems shows that the d8-PhMe/Py sample has less
evident damping of the modulations than the wires dissolved
in d14-oTP/BnPy, which is indicative of a more rigid system.
porphyrin units, n, from the DEER results in d8-PhMe/Py (squares) and
d14-oTP/BnPy (circles) with fitting uncertainties as error bars (only visible
for P4). The fully extended radical-radical distances estimated from
crystallographic data, RXtal (see Figure 1) are shown as the blue line, for
comparison.
As expected therefore, the end-to-end lengths in the lower
temperature d8-PhMe/Py glass are slightly longer than in d14-
oTP/BnPy for the longer porphyrin oligomers. The disordered
aggregation in the lower temperature glass may also favor the
more extended conformation. The radical-radical distances
estimated for the fully extended wires from crystallographic data,
RXtal (blue line in Figure 5) are longer than the R0 values from
the MD calculation as expected for these semiflexible systems.
The MD R0 values in turn are slightly longer than the
(46) Weber, A.; Schiemann, O.; Bode, B.; Prisner, T. F. J. Magn. Reson.
2002, 157, 277–285.
(47) Bode, B. E.; Plackmeyer, J.; Prisner, T. F.; Schiemann, O. J. Phys.
Chem. A 2008, 112, 5064–5073.
(48) Bode, B. E.; Plackmeyer, J.; Bolte, M.; Prisner, T. F.; Schiemann, O.
J. Organomet. Chem. 2009, 694, 1172–1179.
(49) Similar temperature effects were observed on end-labeled oligo(p-
phenyleneethynylene)s (Prof. G. Jeschke, Physical Chemistry, ETH
Zu¨rich, personal communication).
9
13856 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

Flexibility in Porphyrin-Based Molecular Wires
A R T I C L E S
Figure 6. Structures of the Bent Wires, P2·L1 and P4·L2, and the (P2)2 ·(L3)2 ladder. The most probable distances, R0, from the MD simulations are
shown.
experimental DEER distances for P1, P2, and P3, though closer
to those of the oligomers dissolved in d8-PhMe/Py. The
uncertainty in the persistence length fitting parameter, Lp, from
the DEER measurements was large, particularly with P3 and
P4 where the experimental signal-to-noise ratio was worse and
fewer modulations were observed. Within a particular data set
the values of Lp and L were correlated, with the longest L
corresponding to the smallest Lp and vice versa. However, the
R0 (most probable) end-to-end length is a much more stable
parameter and so our discussion above is focused on R0 values.
L and Lp fitting results are given in the Supporting Information.
Although there is a significant uncertainty in the values of Lp,
it is clear that for DEER in both solvents, and also for the MD
results, that Lp increases as the chain length increases from
P1-P3. The trend is less evident with P4 which probably
reflects the lower quality of the data for this longest oligomer.
This increase in persistence length with increasing chain length
indicates that the TEMPO linked end-group is more flexible
than the conjugated porphyrin wire backbone. We conclude that
a longer porphyrin oligomer of this type would have a
persistence length at least as long as that of P3, which in d14-
oTP/BnPy is about 19 nm; in frozen d8-PhMe/Py the persistence
length appears to be substantially longer.
stretching, and bending potential as fitted parameters. Jeschke
and co-workers also compared their method to the WLC model
used in this study and showed that results from the two methods
converge as the chain length increased.19 The second study by
Prisner and co-workers exploited the information given by
altering the positions of the two microwave frequencies with
respect to the nitroxide EPR absorption shape when the nitroxide
spin labels have some level of fixed orientation.20 We have
shown that within the bounds of our DEER and MD results,
the simple WLC model with two fitting parameters provides
adequate information on the flexibility of the wires at two
different temperatures.
DEER Results for Wires Bound to Star-Shaped Templates. It
has previously been shown that porphyrin oligomers can bend
to bind radial template ligands.35,39,40 Here we show that DEER
confirms the bent stuctures of these complexes. We have
investigated two complexes: the porphyrin dimer P2 on the two
legged template L1, and the porphyrin tetramer P4 on the six-
legged template L2 (Figure 6).
The WLC model, as applied to the linear wires, is not appropriate
for the template-bound molecules because they are not free to move
in a random way. Therefore, a Gaussian distribution was assumed
for these systems because the distribution should be mainly due to
flexibility of the spin label moiety. The Gaussian fitting was
implemented in a similar way to above with the mean and full-
width at half-maximum (fwhm) of a Gaussian function as fitting
parameters.50
The DEER data, corresponding distance distributions and
calculated MD distributions are shown in Figure 7. At low
concentrations (2.7 µM for P4·L2 and 7 µM for P2·L1) in a
regime where the UV-vis titration showed clean formation of
Two previous DEER studies with bis-spin-labeled semiflex-
ible conjugated oligomers (poly-p-phenyleneethynylenes) mod-
eled the polymer as one or more jointed stiff units, which can
bend with respect to each other and are capped at each end by
a stiff spin label which can rotate within a defined cone.19,20
The earlier study showed that the spin label contribution to the
flexibility could be successfully removed from the distance di-
stribution using spin label bending, spin label length, backbone
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13857

A R T I C L E S
Lovett et al.
Figure 7. Results for the wires bound to templates in d8-toluene: (a) P2·L1 experimental and fitted time trace; (b) MD simulation (circles) and Gaussian
distance distribution result from fitting the time trace for P2·L1; (c) P4·L2 experimental and fitted time trace; (d) MD simulation (circles) and Gaussian
distance distribution result from fitting the time trace for P4·L2.
Table 2. MD and Fitted DEER Most-Probable (R0) Distance
1:1 complexes (see the Supporting Information), we saw only
Results for the (P2)2 ·(L3)2 Complex
the distance between nitroxides indicative of the template-bound,
ladder component
R0 (MD) (nm)
R0 (DEER, d8-PhMe/Py) (nm)
rather than unbound, oligomer. We also measured the P4·L2
at a higher concentration (50 µM) and saw more than the single
distance (see the Supporting Information). Therefore, DEER
may be a good way of quantifying the degree of template
binding since in favorable cases a ratio of template-bound to
linear oligomer could be assessed.
For the 1:1 complexes, fitting a Gaussian line shape to the
MD simulations for P2·L1 result gave a mean at 4.30 nm and
a fwhm of 0.46 nm. The experimental data were best fitted with
a Gaussian centered around 4.15 nm (+0.1/-0.14 nm) and a
fwhm of 0.68 nm (+0.47/-0.26 nm). Similarly, the MD
simulation for P4·L2 was fitted with a Gaussian of mean )
2.79 nm, fwhm ) 0.94 nm, whereas the experimental data
revealed a slightly shorter best distance with mean 2.49 nm
(+0.81/-0.35 nm) and fwhm 0.87 nm (+0.86/-0.43 nm). The
distance distribution is broader than for the linear wires,
reflective of conformational freedom of the TEMPO group
around the ester linkage, which will affect the distribution more
in the bent structure than in the linear one.
width, w
length, l
diagonal, d
2.69
4.70
5.08
2.69
4.62
4.76
of the “frame” and these are bound by other “rung” molecules
(which may be of variable length) via the metal in the porphyrin.
Figure 6 shows an example of a ladder system made up from
the P2 wire with dipyridyl rungs L3, (P2)2 ·(L3)2. These ladder
structures have increased hole mobility and two-photon absorp-
tion compared to the single-strand wires.33,37
For these systems, there will be three distances corresponding
to the inter strand width (w, roughly equivalent to the length of
the rung joining the two oligomers), which is the shortest distance
in this case, wire length (l) and diagonal length (d), see Figure 6
and Table 2. The distributions of the longer distances (l and d)
from MD calculations fitted well to the WLC model, whereas the
shorter distances (w) were fitted better by a Gaussian model. The
DEER data were modeled using two WLC distributions and one
Gaussian distribution; the shapes of these three distributions were
fixed from the MD results and their modal distances were adjusted
to fit the time traces.52,53 The fit is shown in Figure 8 and the results
are given in Table 2 and the Supporting Information. The ꢀ2 per
degree of freedom error test confirms that the three contributing
distances are necessary for a good fit. If the ladder had rectangular
(D2h) symmetry, as drawn in Figure 6, then the inter strand distance
(w ) 2.69 nm) and oligomer length (l ) 4.62 nm) would lead to
DEER Results for the (P2)2 ·(L3)2 Ladder Complex. The
porphyrin wire structures discussed in this paper may be used
as building blocks for larger arrays.51 An example of such a
structure is analogous to a ladder where the wire forms the side
(50) The full-width at half-maximum (fwhm) corresponds to 2(2 ln 2)1/2σ,
where σ is the standard deviation of a Gaussian lineshape.
(51) Taylor, P. N.; Anderson, H. L. J. Am. Chem. Soc. 1999, 121, 11538–
11545.
9
13858 J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009

Flexibility in Porphyrin-Based Molecular Wires
A R T I C L E S
application of this equation to the DEER data gives a twist angle
of 44°. The wide distance distributions for the width and diagonal
length demonstrate that a range of conformations were measured
by DEER. Preliminary DEER data were also recorded on a three-
rung ladder complex, (P3)2 ·(L3)3, giving an experimental distance
distribution which is consistent with the result from MD calcula-
tions, although the diagonal distance and length could not be
resolved, see the Supporting Information.
Conclusions
We have synthesized porphyrin-based molecular wires,
P1-P4, with terminal nitroxide spin-labels in order to assess
their conformational flexibility using DEER spectroscopy. The
results from these EPR measurements, in two different frozen
solvent glasses (with glass-transition temperatures of 120 and
250 K) show that these molecular wires are remarkably rigid.
In all cases, the ratio of the experimental modal end-to-end
distance R0 to the corresponding distance in the fully extended
conformation, calculated from crystallographic data, RXtal is in
the range R0/RXtal ) 0.94-0.98. A slight increase in rigidity in
the lower temperature glass is evident from the less damped
DEER time traces, narrower distance distributions, and longer
modal lengths recorded in toluene. The distance distributions
from these DEER experiments fit well to a wormlike chain
model, giving a well-defined modal length R0. The persistence
lengths Lp are poorly determined by the data but the results
indicate that for longer oligomer chains of this type, Lp is greater
than 19 nm at 250 K. Molecular dynamics calculations (using
the MM+ force field at 180 K) gave distance distributions which
match well with the DEER results. Previously the conformations
of two other types of conjugated oligomers have been investi-
gated by DEER;18,19 comparison with these previous studies
indicates that butadiyne-linked porphyrin oligomers have similar
flexibility to poly(p-phenyleneethynelene)s.
Figure 8. Results for the (P2)2 ·(L3)2 ladder in d8-toluene (30 µM): (a)
DEER time trace and fit; (b) MD simulation (circles) and shifted MD
components to give best fit to time trace with the resulting R0 values.
A special feature of metalloporphyrin-based molecular wires
is that their conformations can be controlled by noncovalent
self-assembly with multidentate amine ligands. Three supramo-
lecular self-assembled structures, P2·L1, P4·L2, and (P2)2 ·
(L3)2 were investigated by DEER. The experimental spin-to-
spin distances in these complexes match the predictions from
molecular dynamics calculations, demonstrating that DEER is
a valuable technique for gaining structural information on
nanoscopic self-assembled supramolecular structures. DEER
should become a useful addition to the tool-kit for supramo-
lecular chemistry.
Figure 9. Two orthogonal views of a low-energy conformation of
(P2)2 ·(L3)2 from the MD simulation, showing a twisted geometry consistent
with the DEER data (meso-aryl substituents and hydrogen atoms omitted
for clarity).
Acknowledgment. We thank the EPSRC for funding, the STFC
for granting access to the SRS Daresbury, and the EPSRC Mass
Spectrometry Service (Swansea) for mass spectra. S.I.P. acknowl-
edges a URF from the Royal Society. J.E.L. was funded jointly by
the EPSRC and University College, Oxford. We thank Dr. Jeffrey
Harmer for maintaining the EPR spectrometer and Professor Gunnar
Jeschke for useful discussions.
a diagonal of d ) 5.35 nm, which is significantly longer than the
measured diagonal of 4.76 nm. Thus the results indicate that the
structure is twisted in solution, as illustrated in Figure 9. If we
assume that the ladder has a regular D2 geometry, then the twist
angle θ can be calculated from the equation cos θ ) (l2 - w2)/d2;
Supporting Information Available: Synthetic procedures,
characterization data, and expanded DEER and simulation
results (PDF and CIF). This material is available free of charge
via the Internet at http://pubs.acs.org.
(52) Recently, Jeschke and co-workers have presented a method for
avoiding systematic errors in DEER measurements on systems with
multiple radical correlations.53 Here, we show that the DEER results
are consistent with a ladder structure formation. We hope to extend
our investigation to incorporate their new methodology in future work.
(53) Jeschke, G.; Sajid, M.; Schulte, M.; Godt, A. Phys. Chem. Chem. Phys.
2009, 11, 6580–6591.
JA905796Z
9
J. AM. CHEM. SOC. VOL. 131, NO. 38, 2009 13859