Communications
triarylammonium radical.[22] Similar results were obtained by
and could be detected by averaging over the dipolar
interactions. The observed reappearance of the resonance
signals that disappeared after light exposure supports this
claim, although an in-depth comparison between the two
tensors of dipolar coupling and electronic paramagnetism did
not allow for a definitive conclusion to assess the predominant
phenomena (see Figure S3 in the Supporting Information).
Importantly, chlorinated solvents appeared to be necessary to
initiate the oxidation of the triarylamine. For instance, both
the green color of the solutions and the disappearance of
irradiating solutions that were strictly deoxygenated by
freeze–thaw cycles. To obtain quantitative data, EPR spec-
troscopy was performed on a nonirradiated solution, the
1H NMR spectrum of which showed a full set of resonance
signals (Figure 2a). As expected, no trace of radicals was
detected in the dark (t = 0; point A). However, after irradiat-
1
signals in the H NMR spectrum also occurred when deu-
terated dichloromethane and 1,2-tetrachloroethane were
used as solvents, but not when deuterated methanol, toluene,
acetonitrile, or dimethylsulfoxide were used. These data are
consistent with reported studies in which the excited state of
the triarylamine can reversibly transfer one free electron to
chloroform and liberate chloride counterions.[24–26]
We then carried out a structure–property study by
synthesizing derivatives 2–8 and investigating the sensitivity
of their corresponding solutions in chloroform to light (see
chart in Figure 1a). Here, the criterion of the response to light
excitation was the disappearance (yes) or the persistance (no)
of the NMR signals of the aromatic protons when the samples
were submitted to irradiation for one hour. The results reveal
Figure 2. a) Quantitative EPR data, as a function of time, showing the
evolution of the ratio of the triarylammonium radical 1C+ over neutral
1: without visible light excitation at RT (point A); upon visible light
excitation (A–B); in the absence of light at RT (B–C); and after
subsequent heating (608C) in the dark (from C; [1init.]=10 mm).
Dotted line: y-axis value of 6ꢂ10ꢀ3. b) Time autocorrelation function
of the scattered electric field vector for an irradiated solution of 1
([1init.]=7.5 mm in chloroform at T=208C) and for a scattering angle
q=908. Solid line: exponential fit.
ꢀ
that both the free N H bond of the amide moiety (H-bonding
interactions) and the long alkyl (or benzyl) R1 chains must be
present, thus indicating the requirement for additional van
der Waals (or supplementary p-stacking) stabilizations of the
hypothesized self-assembled structures.
ing the sample with a white light (20 W) for one hour (A to
B), the formation of a radical was clearly observed in a
proportion of up to 11% compared to the total number of
molecules (see the Supporting Information). The observed
paramagnetic hyperfine pattern was consistent with that
expected for a radical localized on the nitrogen atom of a
triarylammonium derivative (see Figure S2 in the Supporting
Information).[23] After one hour of irradiation, the amount of
radicals was recorded as a function of time in the dark
(Figure 2a; point B). A very smooth decay was observed over
a period of 16 hours; the decay asymptotically reached a
plateau at a concentration of 6 radicals per 1000 triarylamine
molecules (points B to C). This concentration then remained
constant for at least one week as verified by an EPR control
experiment, thus reinforcing the idea of a radical-stabilization
mechanism in the solution. Finally, the subsequent heating of
the solution (point C) led to fast destruction of the remaining
stabilized radicals over a period of 2 hours. This quantification
of the radical confirmed its very low concentration at
equilibrium and its unusual stability. As a consequence of
We further investigated the solutions by dynamic light
scattering (DLS) experiments (Figure 2b).[27,28] Prior to
irradiation with visible light and consistent with previous
observations, no evidence for the presence of assemblies or
large objects could be deduced from DLS measurements on a
solution of 1 in chloroform. However, after one hour of
irradiation with visible light, the normalized time autocorre-
lation function of the concentration fluctuations, g(1)(q,t),
could be characterized by a simple exponential relaxation
(Figure 2b). The angular dependance revealed a diffusive
relaxation with a characteristic time inversely proportional to
q2, where q is the scattered wave vector. In the case of a
diffusive process, g(1)(q,t), is given by Equation (1):
À
Á
hdcðq; 0Þdcðq; tÞi
gð1Þðq; tÞ ¼
¼ exp ꢀDq2t
ꢀ
ꢁ
ð1Þ
2
dcðq; 0Þ
where dc(q,t) and dc(q,0) represent fluctuations of the
concentration at time t and zero, respectively. The formation
of self-assemblies in solution was thus demonstrated by the
presence of this diffusive relaxation in the correlation
function. In highly diluted solutions, the fit of g(1)(q,t) allows
to determine the diffusive coefficient, D = 2 ꢁ 10ꢀ12 m2 sꢀ1,
and, by using Equation (2), the average hydrodynamic radius
can be determined, where hw is the solvent viscosity.
1
the stability, the H NMR spectra observed at equilibrium
after light exposure (spectrum B in Figure 1b) might indicate
that the total disappearance of the aromatic signals is not due
to the presence of 6 paramagnetic species per 1000 molecules,
but more likely caused by a self-association/delocalization
process by charge transfer of these radicals when self-
assembled with their neutral precursors. To test this assump-
tion, we performed a high-resolution magic angle spinning
(HRMAS) 1H NMR experiment in solution; the possibly
highly p-stacked aromatic parts of the molecules would
behave as a very anisotropic domain in such an experiment
kBT
q!0 6phwD
ð2Þ
RH ¼ lim
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 6974 –6978