Triplet state dynamics within cyclodextrin solid complexes

Time-resolved diffuse-reflectance laser-flash photolysis techniques were employed to determine the triplet state dynamics of 9,10-anthraquinone, benzil, 1,4-benzoquinone, 2,5-dimethyl-1,4-benzoquinone, 9,10-phenanthrenequinone, and α -terthienyl within β -cyclodextrin solid complexes. The α - and γ -cyclodextrin solid complexes of benzil and α -terthienyl were studied likewise. The average triplet lifetime for α -terthienyl increases as the size of the host cavity decreases, while that for benzil remains practically unchanged. In addition, the solid-state photoreduction of 1,2-dimethoxy-4-nitrobenzene triplets by 1-phenylethylamine within β -cyclodextrin solid complexes was investigated. Observed lifetimes tend to diminish as the 1-phenylethylamine-to-1,2-dimethoxy-4-nitrobenzene molar ratio increases.


Introduction
Cyclodextrins (CDs), cyclic oligosaccharides of six to eight α-D-glucose units (namely α, β, and γ-CD, respectively), have been widely used as hosts over the past decades, 1 and continue to be the subjects of current interest both in basic and applied research. 2][6][7][8][9][10][11][12][13] The work reported here focuses on the application of time-resolved diffuse-reflectance laser-flash photolysis techniques to characterize the triplet state dynamics of several aromatic compounds within cyclodextrin solid complexes.The aromatic guest substrates employed in this study are summarized in Chart 1; all of these substrates are known to form excited triplet states in essentially quantitative yield. 14In addition, the kinetics associated with the solid-state photoreduction of 1,2-dimethoxy-4-nitrobenzene triplets by 1-phenylethylamine within β-CD complexes is presented.

Results and Discussion
Triplet-triplet absorption spectra Laser excitation (at 355 nm) of any of the cyclodextrin solid complexes for the guest species shown in Chart 1, leads to the instantaneous formation (i.e., within the time response of the nanosecond laser-flash photolysis system used) of the corresponding guest excited triplet state.This in inferred by comparison of the resulting transient absorption spectra with the spectral data reported in the literature for excited triplet states in solution and in other solid systems.For instance, in the case of 1,4-benzoquinone (BQ) and 2,5-dimethyl-1,4-benzoquinone (DMBQ)/β-CD solid complexes, transient spectra are characterized by an absorption band centered at ca. 400-410 nm (Figures 1A-B).This band agrees very well with the triplet-triplet (T-T) absorption spectrum observed in aqueous solutions, albeit slightly blue-shifted in the case of DMBQ. 15egardless of the type of CD employed, laser excitation of benzil (BZ)/CD solid complexes leads to a transient absorption band centered at ca. 480 nm (Figure 1C).This band agrees excellently with the T-T absorption spectra observed in cyclohexane and acetonitrile solutions, 16 as well in silicalite 17 and β-CD 10 solid matrices.Likewise, laser excitation of α-terthienyl (αT)/CD solid complexes leads to a transient absorption band centered at ca. 470 nm (Figure 1D), independently of the type of CD employed.These spectra are very similar to the T-T spectra observed for αT in organic solvents such as methanol, 18 as well as in β-CD aqueous solutions. 19he transient absorption spectra shown in Figures 1A-D seem to retain their shape through time, i.e., only one transient species is being detected.On the contrary, in the case of 9,10-anthraquinone (AQ) and 9,10-phenanthrenequinone (PQ)/β-CD solid complexes, the spectral changes seem to indicate the presence of (at least) two absorbing species (Figures 1E-F The ease with which triplet states of quinones abstract hydrogen from hydrogen donating reactants is very well documented.14a The transient spectra shown in Figures 1E-F are consistent with the presence of excited triplet states and corresponding semiquinone radicals.For instance, the transient absorption spectra of PQ/β-CD solid complexes show peaks centered at ca. 390, 460 and 700 nm.The peaks at 460 and 700 nm are attributed to the excited triplet state, 20 while the peak at 390 nm is assigned to the radical (semiquinone) species. 21In the case of AQ/β-CD solid complexes, analysis of the transient spectra is rather difficult as a result of the significant overlap between the T-T and radical absorptions in the 390-500 nm region. 22However, the coexistence of triplets and radicals is inferred from the fact that kinetic traces (see next section) vary with monitoring wavelength (the shorter the monitoring wavelength, the longer the lifetime).Furthermore, it should be pointed out here that photoexcited anthraquinone-substituted β-CDs have been shown to react, in the absence of a hydrogen donating solvent, by rapid intramolecular hydrogen abstraction from the cyclodextrin itself. 23s already mentioned, in the case of BQ and DMBQ/β-CD solid complexes the corresponding transient absorption spectra retain their shape through time.Thus, if any hydrogen abstraction is indeed taking place the corresponding semiquinone radicals (λ max ≈ 440 nm) 24 must be decaying rapidly (i.e.within the time response of the nanosecond laser-flash photolysis system used) by back hydrogen abstraction.Likewise, in the case of BZ/CD solid complexes no evidence for benzil ketyl radicals (λ max ≈ 360 nm) 25 formation was obtained.

Triplet state dynamics
Transient decay traces corresponding to air-equilibrated CD solid complexes were collected on different time scales. 26Typical stretched representations (i.e., reflectance change vs. log(time)) of such traces are shown in Figure 2. Representations of these types are most convenient for systems involving transient kinetics covering several orders of magnitude in time scale, as usually observed in solid systems.; multi-component (dispersed) kinetics are generally attributed to diversity in environments/inclusion sites and of conformational structures.
Several models and methods of analysis for transient kinetics in (heterogeneous) solid systems have been reported.In recent years, lifetime distribution analysis and the Albery model have been applied to analyze transient kinetics in CD solid complexes.The Albery model accounts for dispersed kinetics assuming a Gaussian distribution of the logarithm of the rate constant about some mean value, and has two adjustable parameters: a mean rate constant ( k) and the width of the distribution (γ).Alternatively, lifetime distribution analysis (carried out by using the exponential series method, ESM) consists of a fitting function containing up to 100 exponential terms (with fixed, logarithmically spaced lifetimes) and variable pre-exponential factors.
The results of the ESM analysis of the transient decay traces for the CD solid complexes used in this study are illustrated in Figure 3.The lifetime range for the ESM analysis was set between 10 ns and 250 µs (limits of the laser-flash photolysis system employed); however, in no case components of significant amplitude were obtained below 1 µs.From these ESM distributions, amplitude average lifetimes (<τ> a ) were calculated according to eq. ( 1), 27 where a i (N) represents the normalized amplitude (i.e.Σa i (N) = 1) of the ith component, τ i is its lifetime, and M is the number of components of the fit.The resulting <τ> a values are summarized in Table 1.With the exception of the ESM analysis for PQ/β-CD at 700 nm, lifetime distributions are overall characterized by one main decay mode, which accounts for at least 80% of the decay kinetics.As illustrated in Figure 3 and summarized in Table 1, the amplitude average lifetime for αT clearly diminishes as the size of the host cavity increases, while that for BZ seems to be independent of the host size, albeit a small increase (if at all significant) is observed when going from αto β-CD.In the case of AQ and PQ, it is clear that the amplitude average lifetimes diminish as the monitoring wavelength increases from λ max of the radical (ca.390 nm) to λ max of the triplet (> 450 nm).
Transient decay traces were also fitted to a logarithm-normal Gaussian lifetime distribution (using the Albery model) and, based on the relatively simple pattern of the ESM distributions, to either a single or double-exponential function.As shown in Table 1, the resulting mean rate constants ( k) and lifetimes (τ ) are overall in good agreement with the corresponding amplitude average lifetimes.However, it is necessary to point out that for those samples characterized by (at least) two decay modes (e.g., PQ/β-CD @ 700 nm), fittings to the Albery model render residual plots (not shown) lacking a flat and random pattern, particularly in the initial part of the decay traces.
As indicated in Table 1, the CD solid complexes being studied are far from being stoichiometric host-guest complexes; most CD cavities are believed to be empty following loss (during drying) of diethyl ether from CD-solvent complexes that co-precipitate under the experimental conditions employed (see Experimental Section).The increase in host-to-guest molar ratio observed for αT with decreasing host size reflects the limited space available for inclusion; this trend in host-to-guest molar ratio (and hence, in space limitations) clearly parallels the increase observed in lifetimes with decreasing host size.In the case of BZ, the host-to-guest molar ratio increases in the order α-< γ-<β-CD, in agreement with the fact that benzene has been shown to have maximum host-guest contact with the cavity of β-CD. 28iplet state quenching ESM analyses were also carried out to characterize the transient kinetics associated with the solid-state photoreduction of 1,2-dimethoxy-4-nitrobenzene (DMNB) triplets by 1-phenylethylamine (PhEA) within β-CD solid complexes.
Laser excitation of DMNB/β-CD solid samples leads to the formation of DMNB excited triplet states, which are characterized by an absorption band centered at ca. 470 nm (Figure 4A).Upon co-inclusion of PhEA, laser excitation yields a transient absorption with a maximum at ca. 480 nm, which undergoes a small hypsochromic shift at longer timescales (Figure 4B).These spectral changes are attributed to the formation of DMNB-H radicals (i.e., conjugate acid of DMNB radical anions), upon electron and proton transfers from PhEA to DMNB triplets within the β-CD complexes. 13nlike the results presented in the previous section, ESM analysis corresponding to DMNB/β-CD and PhEA/DMNB/β-CD solid complexes leads to more complex lifetime distributions (Figure 4, Inset).
Solid complexes of different PhEA-to-DMNB molar ratios were prepared by changing the amount of PhEA (relative to that of DMNB and of β-CD) initially dissolved in water (see Experimental Section).As illustrated in Figure 4, lifetime distributions for PhEA/DMNB/β-CD solid samples are much broader than those corresponding to DMNB/β-CD complexes.Due to the significant overlap between the triplet and radical absorptions, it is very difficult to ascribe physical meaning to each of the (main) components of the distribution.However, it is of interest to note that as the PhEA-to-DMNB molar ratio increases, the contribution of the shorter-lived components (tentatively ascribed to DMNB triplets being quenched) tends to increase as well.Thus, amplitude average lifetimes are found to diminish with increasing PhEA concentration (Table 2).While these experimental observations are in full agreement with those previously reported for PhEA/DMNB/β-CD solid samples, 8,13 the data given in Table 2 represent the first set of lifetimes obtained, as a function of varying quencher concentration, for triplet quenching within cyclodextrin solid complexes by means of time-resolved diffuse-reflectance laser-flash photolysis techniques.resolved laser-flash photolysis system employed in this study are reported elsewhere. 10All measurements were carried out at (21 ± 1)°C.Lifetime distribution analyses were carried out by using the exponential series method software (modified version for transient absorption) from Photon Technology International.The fitting function consisted of 100 exponentials terms (with fixed, logarithmically spaced lifetimes) and variable pre-exponential factors.Lifetimes ranged from 10 ns (shortest lifetime allowed) to 250 µs (limit of our laser flash photolysis system).Alternatively, fittings to the Albery model (using a convenient fitting function as described in ref. 30), and to a single or double exponential function were carried out by using the general curve fitting procedure of Kaleidagraph 3.0.5 software from Abelbeck Software.

Table 1 .
Amplitude average lifetimes (<τ> a ), results from the Albery model, and exponential (least-squares fitting) analysis results for air-equilibrated CD solid complexes a Host-to-guest molar ratio.b Monitoring wavelength.c Percent contribution to total decay given in brackets.d Lifetime(s) > 250 µs (longest lifetime accessible with the laser-flash photolysis system employed).