Redox-Active Star Molecules Incorporating the 4-Benzoylpyridinium Cation: Implications for the Charge Transfer Efficiency along Branches vs Across the Perimeter in Dendrimers

Article abstract of DOI:10.1021/ja0390247

We report the redox properties of four star systems incorporating the 4-benzoyl-N-alkylpyridinium cation; the redox potential varies along the branches but remains constant at fixed radii. Bulk electrolysis shows that at a semi-infinite time scale all red

Full text of DOI:10.1021/ja0390247

Published on Web 03/16/2004
Redox-Active Star Molecules Incorporating the 4-Benzoylpyridinium Cation:
Implications for the Charge Transfer Efficiency along Branches vs Across the
Perimeter in Dendrimers
Nicholas Leventis,*,† Jinhua Yang, Eve F. Fabrizio, Abdel-Monem M. Rawashdeh,
‡
§
‡
‡
,‡
Woon Su Oh, and Chariklia Sotiriou-Leventis*
NASA Glenn Research Center, 21000 Brookpark Road M.S. 49-1, CleVeland, Ohio 44135, Department of Chemistry,
UniVersity of Missouri-Rolla, Rolla, Missouri 65409, and Ohio Aerospace Institute, 22800 Cedar Point Road,
CleVeland, Ohio 44142
Received October 13, 2003; E-mail: Nicholas.Leventis@nasa.gov; cslevent@umr.edu
Dendrimers are self-repeating globular branched star molecules,
whose fractal structure continues to fascinate, challenge, and
1
inspire. Functional dendrimers may incorporate redox centers, and
potential applications include antennae molecules for light harvest-
ing, sensors, mediators, and artificial biomolecules. Dendrimers with
2
a redox core show no significant inhibition but also shielding (by
3
the branches) and orientation effects (in asymmetric dendrimers)
4
upon the rate of heterogeneous electron transfer. Dendrimers with
redox centers in the branches are also well-known. Using thin-
layer-cell electrochemistry, all 21 tetrathiafulvalenes embedded in
the branches of a third-generation dendrimer are electrochemically
5
accessible leading to a 42+ cation. Using pulse radiolysis, Fox
reported fast electron transfer (e-transfer) between peripheral
6
biphenyls (donors) and a core acceptor (a Ru(II)-complex). On
the other hand, Crooks has reported incomplete electrolysis of amido
7
Figure 1. CV at 0.1 V s-1 of 1 (1.10 mM) in Ar-degassed DMF/0.1 M
TBAP containing 1.44 mM of dMeFc as internal standard, using a Au disk
electrode (1.6 mm in diameter). (Inset A): Same solution Randles-Sevcik
amine dendrimers with a viologen functionalized perimeter, while
Amatore and Abru n˜ a using ultrafast voltammetry have reported fast
e-transfer, despite the 2-nm separation from one another, among
the 64 Ru-complexes in the perimeter of a fourth-generation
-
1
plots for 1 and dMeFc in the range 0.02-10 V s . For 1: slope ) 0.89 (
1/2
-1/2
2
0
s
.01 µA s (mV) ; R ) 0.999. For dMeFc: slope ) 0.629 ( 0.003 µA
8
1/2
-1/2
2
dendrimer adsorbed on a microelectrode; the fast charge propaga-
(mV) ; R ) 1.0. (Inset B): DPV’s of 1/dMeFc (s; 1.02/1.59 mM)
tion was attributed to the branch flexibility.
and of the free base of 1-FB/dMeFc (‚‚‚; 0.98/1.78 mM).
On the basis of those reports, e-transfer across the perimeter of
dendrimers should depend on their rigidity, but it is unclear whether
it would be more or less efficient than e-transfer along the branches.
As these questions have important implications for molecular
design, they were investigated with star systems 1-4, serving as
models of first- and second-generation redox dendrimers. Cations
and infer how easily charge randomizes across the perimeter of a
1
relatively small, rigid redox star-system. Thus, n was assessed both
at a semi-infinite time scale (by bulk electrolysis) and at the cyclic
-
1
voltammetric (CV) time scale of 0.02-10 V s . Next, the
voltammetry of 2-4 was used to assess whether within the same
time scale redox centers within the branches are as accessible as
redox centers across the perimeter.
Figure 1 shows the redox processes of 1 versus decamethylfer-
rocene (dMeFc: internal standard).10 One-electron-like waves
1
1
indicate that redox units behave independently of one another.
Bulk electrolysis of 1 at -0.85 V vs Ag/AgCl affords n ) 3.01 (
.03. The linearity of the Randles-Sevcik plot in inset A shows
that the number of redox centers reduced remains unchanged in
1
0
-
1
1
the time scale of 0.02-10 V s . Within this time-scale, n was
determined first by comparing the ratio of the slopes of the
Randles-Sevcik plots of 1 and dMeFc with the limiting current
ratio obtained in the same solution with an ultramicroelectrode (∼10
µm diam).12 Thus, n
) 2.08 ( 0.01. Alternatively, the diffusion
1
-
6
2
-1
coefficients D
1
2
) 9.48 + 0.05 10 cm s and DdMeFc ) 2.51 (
9
-5
-1
1
-4 incorporate 4-benzoyl-N-alkylpyridinium (BP), whose redox
0.03 10 cm s were determined by an HPLC method (see
potential (a) varies along the branches and (b) remains constant at
Supporting Information) and their values were introduced directly
fixed radius. Our strategy was to measure the number of electrons,
in the slope ratio of the Randles-Sevcik plots. Thus, n ) 2.02 (
1
n
1
, exchanged between 1 and the electrode at different time-scales,
0.07. The two values of n are comparable. Meanwhile, differential
1
pulse voltammetry (DPV) allows baseline separation of the two
†
NASA Glenn Research Center.
reduction waves of 1 (solid line, Figure 1, inset B), and by fitting
‡
University of Missouri-Rolla.
§
12
Ohio Aerospace Institute.
the nonlinear expression for (δCurrent)max
,
it is calculated from
4094
9
J. AM. CHEM. SOC. 2004, 126, 4094-4095
10.1021/ja0390247 CCC: $27.50 © 2004 American Chemical Society

C O MMU N I C A T I O N S
b) and that n
4
-internal ) 1.46 ( 0.06 (from wave c). Thus, the
number of accessible redox centers within the branches is about
equal to the number of accessible redox centers across the perimeter.
Given the large separation between the external and internal BP
groups (∼1.6 nm from CdO to CdO), through-bond e-transfer is
expected to be slow. Tunneling through the branches must be also
8
ruled out. Electron hopping from the perimeter to the center is
thermodynamically unfavorable as reduced external redox centers
(
-NO
2
and pyridinium) do not have the power to reduce the internal
- spacers along the
carbonyl. On the other hand, the -(CH
2 6
)
branches may coil up, bringing the internal redox centers closer to
the perimeter. This hypothesis is supported by the UV spectrum of
the core, which is independent of the concentration but is affected
by the substituents along the perimeter. Thus, the λmax values of 1
and 3 are at 334 and 324 nm, respectively, corresponding to a
-
1
stabilization by 2.1 kcal mol , which is consistent with π-π
1
7
interactions between the external BP groups and the core. In
summary, the rigidity of 1 provides a complementary view of the
fact that fast e-transfer along the perimeter of core-branch systems
requires flexible branches.8 From a practical viewpoint, redox
equivalents emerging from the core of a rigid light-harvesting
system would be localized at the tips of the branches they emerge
from, creating issues of efficient bimolecular e-transfer to redox
quenchers in their immediate environment. Flexible branches may
not only facilitate e-transfer along the perimeter but may also fold,
rendering internal redox centers more accessible.
Figure 2. (A) CV of 2/dMeFc (‚‚‚; 1.10/1.44 mM) and 3/dMeFc (s; 0.95/
1.56 mM) under the same conditions as in Figure 1. (Inset): DPVs of
2/dMeFc (0.96/2.70 mM) and of 3/dMeFc (0.98/1.84 mM). (B) DPV of
4
/dMeFc (0.87/1.02 mM). (Inset): Corresponding CV (0.1 V s-1).
Acknowledgment. We thank the NASA Glenn Research Center
Director’s Discretionary Fund (DDF) for financial support.
the first wave of 1 that n ) 1.81 ( 0.01. Assuming that the
1
Supporting Information Available: Experimental Section. This
material is available free of charge via the Internet at http://pubs.acs.org
diffusion coefficient of the first-wave reduction product is 0.7 ×
3-n )+
D
1
, i.e., that the ratio of the diffusion coefficients of 1 and 1(
1
is equal to the ratio of the diffusion coefficients of BP and its 1-e
References
reduced form,13 it is calculated from the second wave of 1 (Figure
(1) Grayson, S. M.; Fr e´ chet, J. M. Chem. ReV. 2001, 101, 3819-3867.
(
2) (a) Cameron, C. S.; Gorman, C. B. AdV. Funct. Mater. 2002, 12, 17-20.
1
, inset B) that n
equal number of electrons in the first and second reductions. The
voltammetric values of n
1
) 1.84 ( 0.06. Thus, by DPV 1 uptakes an
(
b) Toba, R.; Quintela, J. M.; Peinador, C.; Rom a´ n, E.; Kaifer, A. E. Chem.
15
Commun. 2001, 857-858.
(∼2) probably reflect the average
(3) Ceroni, P.; Vicinelli, V.; Maestri, M.; Balzani, V.; M u¨ ller, W. M.; M u¨ ller,
1
U.; Hahn, U.; Oswald, F.; V o¨ gtle, F. New J. Chem. 2001, 25, 989-993.
orientation of 1 approaching the electrode and signify that e-hopping
across the perimeter does not occur to a significant extend within
this time frame. In this regard, 1 is a rigid system, and the distance
between the pyridinium nitrogens is fixed at ∼2 nm (by modeling).
Since the rate of through-space e-transfer decreases exponentially
(
4) Ong, W.; Kaifer, A. E. J. Am. Chem. Soc. 2002, 124, 9358-9359.
(5) Christensen, C. A.; Goldenberg, l. M.; Bryce, M. R.; Becher, J. Chem.
Commun. 1998, 509-510.
(6) Ghaddov, T. H.; Wishart, J. E.; Kirby, J. P.; Whitesell, J. K.; Fox, M. A.
J. Am. Chem. Soc. 2001, 123, 12832-12836.
(7) Baker, W. S.; Lemon, B. I., III; Crooks, R. M. J. Phys. Chem. B 2001,
105, 8885-8894.
-
1
with distance (attenuation factor ) 10 nm ), the e-exchange rate
(
8) (a) Amatore, C.; Bouret, Y.; Maisonhaute, E.; Abru n˜ a, H. D.; Goldsmith,
J. I. C. R. Chimie 2003, 6, 99-115. (b) Amatore, C.; Bouret, Y.;
Maisonhaute, E.; Goldsmith, J. I.; Abru n˜ a, H. D. Chem. Eur. J. 2001, 7,
9
between the BP centers in 1 would be ∼10 times slower than the
rate in contact either in solution or via branch flexibility.8
2206-2226.
Figure 2 summarizes the redox chemistry of 2-4. The overlap-
ping waves of Figure 2A imply that both external and internal redox
centers of 2 and 3 are accessible.16 Bulk electrolysis of 3 at -0.95
(9) (a) Leventis, N.; Zhang, G.; Rawashdeh, A.-M. M.; Sotiriou-Leventis, C.
Electrochim. Acta 2003, 48, 2799-2806. (b) Leventis, N.; Rawashdeh,
A.-M. M.; Zhang, G.; Elder, I. A.; Sotiriou-Leventis, C. J. Org. Chem.
2002, 67, 7501-7510.
(
10) Leventis, N.; Oh, W.-S.; Gao, X.; Rawashdeh, A.-M. M. Anal. Chem.
V vs Ag/AgCl affords n ) 5.9 ( 0.1. To determine how many
3
2003, 75, 4996-5005.
redox centers within a branch are accessible within a voltammetric
time scale, it was necessary to resolve two redox waves, one as-
signed unambiguously to an external and one to an internal redox
center. In that regard, the -NO group of 4 is reduced between the
2
two BP-based reductions (wave b, Figure 2B), and while the first
(
11) The peak-current potential separations (∆Ep-p) are 68 ( 5, 78 ( 5, and
70 ( 5 mV for the first- and second-reduction waves of BP and for the
dMeFc wave, respectively.
(
12) Bard, A. J.; Faulkner, L. R. Electrochemical Methods, Fundamentals and
Applications, 2nd ed; John Wiley and Sons: New York, 2001.
13) Experimentally, the second redox wave of 4-benzoyl-N-methylpyridinium
(
1
4
cation is 0.85× the first one. Looking up that ratio in Table 3 of ref
reduction waves of the external and internal pyridinium groups
14a, it is deduced that the diffusion coefficient of the 1-e reduced form is
∼0.7× the diffusion coefficient of the parent species.
overlap (wave a, Figure 2B), upon further reduction the -NO
2
(14) (a) Leventis, N.; Gao, X. J. Phys. Chem. B 1999, 103, 5832-5840. (b)
group turns from a good electron acceptor (σp-NO2 ) 0.78) to an
Leventis, N.; Gao, X. J. Electroanal. Chem. 2001, 500, 78-94.
extremely strong electron donor (σp-NO2•- ) -0.97),9 pushing
(15) A similar analysis for the reduction of the carbonyl groups in the free
base of 1 (1-FB) (dotted line in Figure 1, inset B) yields n1-FB ) 1.66 (
negative the reduction wave of the “external” carbonyl group of 4
-5
2
-1
0.02. (D1-FB ) 1.20 ( 0.02 10 cm s by HPLC.)
(wave d). Thus, the carbonyl-based reductions of the internal (wave
(16) For the two waves of 2, ∆EP-P ) 95 ( 3 mV and 114 ( 5 mV,
respectively. Thus, their composite nature had to be confirmed by DPV:
Figure 2A, inset.
c) and external (wave d) redox centers are resolved. Using D
4
)
-
6
2
-1
9
.24 ( 0.01 10 cm s (by HPLC) as an approximation of the
(17) For comparison, the π-π interaction energy in methylene blue, as reflected
-1
14a
in the absorption spectrum, is ∼4 kcal mol . For example, see:
true D
i
’s of the intermediate reduced forms, it is calculated from
Rabinowitch, E.; Epstein, L. F. J. Am. Chem. Soc. 1941, 63, 69-78.
4
the DPV of Figure 2B that n -external ) 1.38 ( 0.06 (from wave
JA0390247
J. AM. CHEM. SOC.
9
VOL. 126, NO. 13, 2004 4095