Article Cite This: J. Phys. Chem. B 2017, 121, 11180−11188
pubs.acs.org/JPCB
Photophysics of Zinc Porphyrin Aggregates in Dilute Water−Ethanol Solutions Amy L. Stevens, Neeraj K. Joshi, Matthew F. Paige,* and Ronald P. Steer* Department of Chemistry University of Saskatchewan Saskatoon, Saskatchewan, Canada S7N 5C9 S Supporting Information *
ABSTRACT: Dimeric and multimeric aggregates of a model metalloporphyrin, zinc tetraphenylporphyrin (ZnTPP), have been produced in a controlled manner by incrementally increasing the water content of dilute aqueous ethanol solutions. Steady state absorption, fluorescence emission, and fluorescence excitation spectra have been measured to identify the aggregates present as a function of solvent composition. The dynamics of the excited states of the aggregates produced initially by excitation in the Soret region have been measured by ultrafast fluorescence upconversion techniques. Only the monomer produces measurable emission from S2 with a picosecond lifetime; all Soret-excited aggregates, including the dimer, decay radiationlessly on a femtosecond time scale. The S1 state is the only significant product of the radiationless decay of the S2 state of the excited monomer, and the aggregates also produce substantial quantum yields of S1 fluorescence when initially excited in the Soret region. The resulting fluorescent aggregates all decay on a subnanosecond time scale, likely by a mechanism that involves dissociation of the excited monomer from the excitonic multimer. The ZnTPP dimers excited at their ground state geometries in the Soret region exhibit a dynamic behavior that is quite different from those produced following noncoherent triplet−triplet annihilation under the same conditions. The important implications of these observations in determining the aggregation conditions promoting efficient photon upconversion by excitonic annihilation in a variety of media are thoroughly discussed.
■
INTRODUCTION The spectroscopic, photophysical, and dynamic properties of organic and organometallic materials employed as the absorbing and/or emitting components of electroluminescent and photovoltaic devices are determined in large measure by their state of molecular aggregation in the solid state.1−10 For example, it is now well-established that electronically excited aggregates of aromatic molecules with extended π systems behave in a markedly different fashion from their monomeric constituents and that different aggregate geometries (e.g., J vs H) have different spectroscopic and dynamic signatures.11,12 The initial explanations of these effects were the result of examinations of the photochemistry and photophysics of twocomponent aggregates, excimers and exciplexes,13,14 often formed in fluid media by complexation during the lifetime of the excited monomer. Finding and understanding the exact aggregation conditions needed to promote the most efficient operation of electroluminescent and photovoltaic devices in the solid state has, however, proved to be rather challenging.4,5,7,10 Recently attention in this area has focused on noncoherent triplet exciton annihilation phenomena in which short-range interactions between two triplet species results in electronic energy “pooling”, often forming a radiative excited singlet product that exhibits delayed “upconverted” fluorescence.15 The very nature of the triplet−triplet annihilation (TTA) process requires that the two interacting but uncorrelated species lie within van der Waals distances of one another at the © 2017 American Chemical Society
instant of energy transfer. The post energy transfer pair therefore must also be equally close to one another, at least initially, and could well assume the properties of excimers (if the interaction involves two of the same species) or of exciplexes (if two different triplets are involved). Similar considerations apply to the correlated triplet pair formed initially during singlet fission in organic aggregates.16,17 Whether the aggregated excited dimer resulting from TTA behaves in a fashion similar to that of the excited monomer will depend, among other factors, on the lifetime of the electronically excited product species. In solution, a singlet excited product with a lifetime in the nanosecond or longer range will exist long enough, post energy transfer, for diffusional separation of the species to occur so that the properties of the excited product such as its fluorescence quantum yield, averaged over its lifetime, can be similar to that of the excited monomer in the same medium. On the other hand, if the electronically excited product of the annihilation interaction is very short-lived, as it will be if a more highly excited triplet (Tn, n > 1) or singlet (Sn, n > 1) results from TTA, then its measured properties will be much more susceptible to the effects of aggregation. More subtle effects due to aggregation Received: October 5, 2017 Revised: November 16, 2017 Published: November 21, 2017 11180
DOI: 10.1021/acs.jpcb.7b09868 J. Phys. Chem. B 2017, 121, 11180−11188
Article
The Journal of Physical Chemistry B
6000i UV−vis−NIR spectrophotometer operating in dual beam mode. For concentration-dependent absorption experiments, quartz cuvettes with path lengths from 10 to 1 mm were used to keep the absorbance of the solutions in the Soret band below 3 and thus avoid instrumental saturation effects. Detectorcorrected fluorescence emission and source-corrected fluorescence excitation spectra were obtained using a PTI QuantaMaster spectrofluorometer. The spectra of dilute solutions were obtained with 10 mm × 2 mm rectangular cuvettes and with triangular cuvettes and front-face illumination for more concentrated samples, thus circumventing artifacts from reabsorption and fluorescence collection efficiency effects. Control experiments were performed using ZnTPPS at the same molar concentrations. ZnTPPS is fully soluble in all the water−ethanol solutions employed and showed no evidence of the aggregation phenomena observed with ZnTPP. Time-resolved-fluorescence spectra were collected on two different time scales. Nanosecond lifetimes were measured using a Ti:sapphire laser (Mira, Coherent) that provided modelocked pulses in the 700−1000 nm range. The 76 MHz pulse train was sampled using a pulse picker to provide excitation pulses at an acceptable repetition rate and was frequency doubled using a second harmonic generator. An excitation wavelength of 400 nm was chosen to excite the ZnTPP samples to the blue of the main Soret band in an absorption feature assigned to a vibronic band of the monomer but also coincident with absorption features of the aggregates (vide infra). The instrument response function (IRF) was measured at the excitation wavelength using a Ludox scatterer, yielding an IRF with a width of 100 ps (full width at half-maximum (fwhm)). Time-correlated single photon counting (TCSPC) was used to collect fluorescence at the magic angle relative to the excitation beam polarization, and decay curves were accumulated with a minimum of 10 000 counts in the peak channel. Picosecond-time-scale fluorescence emission measurements were conducted using fluorescence upconversion spectroscopy as described previously in detail.18 Briefly, a Vitesse Duo pumped RegA 9000 (Coherent) produced a pulsed 800 nm beam with an average power of ca. 50 mW. A beamsplitter sent 20% of this light to a 0.5 mm thick BBO nonlinear crystal where it was frequency doubled, separated from the fundamental via a dielectric mirror, and focused into the sample contained in a 2 mm path length quartz cuvette. The average power of the 400 nm excitation beam before the focusing lens was 1.5 mW, and the sample was continuously stirred by a magnetically controlled stir bar. The fluorescence was collected and focused onto another 0.5 mm thick BBO crystal along with the remaining 80% of the 800 nm light. The resulting sum-frequency signal was detected using a monochromator and photomultiplier tube (PMT) combination. To measure the lifetime of the S2 electronic state accessed by exciting in ZnTPP’s Soret band, the fluorescence was detected at 436 nm at a repetition rate of 10 kHz, while for the rise of the S1 fluorescence, the emission wavelength was set to 620 nm at a repetition rate of 100 kHz. To ensure that there was no laser-induced sample damage, the peak signal amplitudes were monitored while measuring each sample and remained consistent over the length of the experiment. By adjusting the PMT voltage, the peak count rate was limited to ≤5% of the repetition rate of the laser in order to eliminate photoncounting artifacts in the detection electronics.
geometry and anisotropic interaction potential can then be expected to have significant influence. Experimental evidence of such subtleties is particularly strong when weakly fluorescent upper singlet states (Sn, n > 1) are the product of TTA, as in photoexcited d0 and d10 metalloporphyrins.18−20 Excitation in the Q bands of these absorbers is followed by rapid intersystem crossing to a long-lived triplet that can undergo efficient TTA yielding the second excited singlet state (S2) of the porphyrin as the major initial energypooled product.20,21 Although these S2 states are very shortlived in the directly photoexcited monomer (of the order of 1 ps due to rapid intramolecular internal conversion), they are also radiatively coupled to the ground state by a fully allowed electric dipole transition. Consequently, S2−S0 fluorescence is readily observable, despite rapid competing radiationless decay, and can be used to monitor the dynamics of the system. Thus, it is relatively easy to compare the spectroscopic behavior of the one-photon directly excited S2 state of the monomer with the same state when it is produced by TTA in the same medium, but inevitably in close proximity to a second ground state molecule. As a result, the quantum yield of S2−S0 emission can be significantly smaller when the S2 state is produced by TTA as compared with the same state obtained on one-photon direct excitation of the monomer under otherwise identical conditions.21,22 The inferred conclusion is that the excited dimer (excimer) resulting from homomolecular TTA in these metalloporphyrins provides an additional radiationless path for exciton decay not available to the excited monomer. Such net quantum yield differences, together with observations of the TTA behavior of metalloporphyrin aggregates in Langmuir−Blodgett films23 and in polymer films,19,20 raise the question of how the ground state geometry and extent of aggregation affect the dynamic properties and hence upconversion efficiencies of these systems. To assist in answering this question, we have investigated the aggregation of zinc tetraphenylporphyrin (ZnTPP) in water−ethanol mixtures of variable composition, using the water-soluble tetrasulfonated derivative (ZnTPPS) as a control. The onset of aggregation of the ZnTPP as the water content of the solvent increases is clearly observed in both the spectra and dynamics. Careful control of the solvent composition at a given total solute concentration then allows the photophysical properties of the dimers and higher aggregates in the ground state equilibrium mixtures to be measured.
■
EXPERIMENTAL SECTION Materials. Zinc meso-tetraphenylporphyrin (ZnTPP) and zinc meso-tetra(4-sulfonatophenyl)porphine tetrasodium salt (ZnTPPS) were purchased from Sigma-Aldrich and Frontier Scientific respectively and were used without further purification. The solvents used were spectroscopic-grade ethanol and ultrapure water (18 MΩ·cm, Millipore). Techniques. All experiments were performed at room temperature in air. To avoid any time-dependent changes in aggregate formation, the samples were prepared at least a day before being tested. At all other times, the samples were stored in a refrigerator to prevent solvent evaporation and each was thoroughly vortexed, before being measured, to avoid aggregate precipitation. A detailed procedure for sample preparation is presented in the Supporting Information. Steady-state absorption, fluorescence emission, and fluorescence excitation measurements were performed on the samples. Absorption spectra were determined using a Cary 11181
DOI: 10.1021/acs.jpcb.7b09868 J. Phys. Chem. B 2017, 121, 11180−11188
Article
The Journal of Physical Chemistry B The TCSPC decay curves were analyzed by deconvolution of the observed decay from the IRF to obtain the fluorescence decay function represented as a sum of discrete exponentials: I (t ) =
⎛ t⎞ ⎟ ⎝ τi ⎠
∑ αi exp⎜− i
where I(t) is the fluorescence intensity at time t and αi is the amplitude of the ith lifetime such that ∑i αi = 1. The picosecond upconverted fluorescence data were fitted by the convolution of an exponential with a Gaussian function representing the IRF: ⎛ B t − t0 ⎞ ⎛ B 2 ⎞ 1 − erf⎜ τi − B ⎟ n ⎜ τ ⎟ ⎛ t − t0 ⎞ ⎝ 2 ⎠ i ⎟ I(t ) = ∑ Ci exp⎜ − ⎟ exp⎜ τi ⎠ 2 ⎝ ⎜⎜ 2 ⎟⎟ i ⎝ ⎠
()
where t0 is the time value at half the maximum of the initial rise of the decay curve, Ci is the amplitude of the ith lifetime, τi is a , where a is the fwhm of the ith decay time, and B = 2 2 ln(2)
the IRF. For each data set, the goodness of fit to the experimental data was evaluated by considering the reduced chi-squared values (χr2 ≤ 1.1 if acceptable) and analyzing the statistical randomness of the weighted residuals.
■
RESULTS AND DISCUSSION The favorable absorption, photochemical, photophysical, and biocompatibility properties of porphyrinoids have, for decades, prompted researchers to seek derivatives of them that are water-soluble, while at the same time retaining their desired properties and avoiding aggregation.24,25 Here we take the opposite approach by using a poorly water-soluble model metalloporphyrin, ZnTPP, and dissolving it in ethanol solutions with increasing water content in order to induce aggregate formation. The process is adapted from that used earlier to form colloidal aggregates of protoporphyrins and amphiphilic tetraphenylporphyrins dissolved initially in DMSO.26 With appropriate control of the solution composition, the photophysical properties of the ZnTPP dimer and some larger aggregates can then be examined and compared with those of the monomer. Figure S1A−C in the Supporting Information presents the near-UV−visible absorption spectra of ZnTPP as a function of its concentration in ethanol−water solutions as the mole fraction of water, f H2O, is increased from 0 to 0.5, i.e., from 0 to 24% v/v water in ethanol. In ethanol-only solutions and ZnTPP concentrations of ≤10 μM the spectra of the dilute porphyrin are assigned exclusively to the monomer and are identical to those previously published.18 However, as expected, as the mole fraction of water increases, the absorption spectra at the higher ZnTPP concentrations broaden and the Soret and Q band maxima shift slightly to the blue. A weak feature at ca. 450 nm, to the red of the monomer’s Soret band, begins to appear at the higher ZnTPP concentrations within the range up to 10 μM when the water content has reached f H2O = 0.5 (cf. Figure S1B(i) in the Supporting Information), and this feature grows to a clearly observable band in solutions of higher water content. This is accompanied by a slight decrease in the peak intensity of the monomer at 422 nm and a small increase in the intensity of the feature at 401 nm, assigned in the monomer to the overlap of several vibronic bands. Then, as shown in Figure 1A, when the water content of the solvent has increased to f H2O
Figure 1. (A) Absorption spectra as a function of ZnTPP concentration in a water−ethanol solution of f H2O = 0.75. (B) Absorbances at monomer absorption peak wavelengths as a function of ZnTPP concentration, extracted from experiments such as those of (A). Background equal to apparent absorbance at 1000 nm has been subtracted in these spectra.
= 0.75, the spectra at all ZnTPP concentrations greater than 1.0 μM are broadened and more complex, providing clear evidence of porphyrin aggregation. The solutions with higher water content also exhibit considerable light scattering at the higher solute concentrations. The clear conclusion is that ZnTPP aggregation is occurring in these solutions and that the concentration of the porphyrin at the onset of observable aggregation is sensitive to the solvent composition, decreasing as the water content of the solvent increases. These observations are completely consistent with the known modest solubility of ZnTPP in ethanol but poor solubility in water. At f H2O > 0.65 the variations in absorbance as a function of total ZnTPP concentration at the wavelengths of maximum absorbance of the Soret and Q bands of the monomer (near 402, 422, and 557 nm for f H2O = 0.75, Figure 1B) show evidence of sequential molecular aggregation. As the nominal ZnTPP concentration increases, note the initial decrease in all 11182
DOI: 10.1021/acs.jpcb.7b09868 J. Phys. Chem. B 2017, 121, 11180−11188
Article
The Journal of Physical Chemistry B monomer peak absorbances to a first minimum at ca. 1.0−1.5 μM, followed by local minima at ca. 3.5, 5, 7, and 9.5 μM superimposed on a generally increasing absorbance at all three wavelengths. We attribute the changes in the spectra in the 0− 1.5 μM total ZnTPP concentration range to the formation, initially, of the dimer, thereby decreasing the absorbance at the monomer wavelengths while increasing the absorbance at wavelengths to both the red and the blue as dimer forms. This is followed at higher total ZnTPP concentrations by the addition of more monomer to the existing aggregates, leading to the additional dips at the monomer absorbance maxima at solution conditions most highly favorable to monomer capture. The broad underlying features that trail off into the nearinfrared and dominate the spectra at the higher ZnTPP concentrations of Figure 1A are readily assignable to diffusive light scattering by large aggregates.26 All experiments designed to extract quantitative information were then carried out at solvent compositions and solute concentrations where such scattering was minimal. We note as well that if the total integrated area of all the absorption features in the Soret band region is taken as a function of total ZnTPP concentration (Figure S1B(iii) in the Supporting Information for the case of f H2O = 0.5), the correlation remains linear well into the region in which dimers have formed. This observation suggests that the integrated molar absorptivity of the Soret band of the dimer exhibits a value about twice that of the monomer. Linearity of the Beer’s law plot is thus maintained because the oscillator strength of the dimer transition compensates for the 1:2 stoichiometry of the dimer, as expected in theory when the absorbing porphyrin dimer is π-stacked at close range.27 Linearity is lost at higher concentrations (greater than ca. 6.5 μM in Figure S1B(iii)) where multimers are being formed, confirming that the multimers exhibit integrated molar absorptivities that are somewhat smaller than that of the monomer multiplied by the number of monomers in the aggregate, again consistent with theory and observation.26,27 These data then define the conditions under which one can expect to observe the ground state monomer−dimer equilibrium at each f H2O. Similar trends can be seen in the fluorescence emission spectra as shown in Figure 2 where 2.5 × 10−6 M solutions of ZnTPP in ethanol−water solvents of varying water content are excited at a fixed wavelength of 400 nm. The spectra in pure ethanol are identical to those previously published,18,28 and exhibit fluorescence from both the S2 and S1 excited states with quantum yields of ca. 1.8 × 10−3 and 2.6 × 10−2 respectively.18 The emission spectra in the mixed solvent, corrected for both detector sensitivity and variation in absorbance at the excitation wavelength of 400 nm, remain almost unchanged up to a water content of ca. 30% v/v water ( f H2O = 0.58). In particular, both the S2 and S1 fluorescence intensities remain almost constant. However, a marked change in the spectra occurs at f H2O > 0.6. The S2 emission intensity falls rapidly to near zero, whereas the S1 fluorescence is reduced considerably (but not to zero), broadens, and shifts to the blue by 10−12 nm. No new bands appear. The solvent composition at which this marked change occurs is coincident with the appearance of significant ZnTPP aggregation in the absorption spectra (cf. Figure 1A for a 2.5 × 10−6 M solution). Note, however, that the emission in the Qband region, though slightly blue-shifted, remains characteristic of the excited monomer (although in a more polar environ-
Figure 2. Corrected fluorescence emission spectra of ZnTPP at a total concentration of 2.5 × 10−6 M as a function of solvent composition (v/v % H2O) excited at a fixed wavelength of 400 nm. The spectra have been corrected to account for the changes in absorbance at 400 nm as the solvent composition varies. They have not been corrected for a small amount of fluorescence reabsorption on the blue edge of the Soret band.
ment) even in water−ethanol solutions in which the absorption spectra suggest that mainly aggregates are being excited. In order to confirm the identities of the absorbing species responsible for the emission in the Q-band region of the spectrum, fluorescence excitation spectra were taken by observing emission at the band maxima (near 655 nm in the solutions of low water content) and scanning through the 350− 500 nm region, which includes the strong Soret bands of the monomer and its aggregates. Figure 3 presents typical fluorescence excitation spectra for 1.5 × 10−6 M ZnTPP solutions of varying water content. In pure ethanol, the
Figure 3. Excitation spectra taken through the Soret region and observing fluorescence at the Q-band emission maximum (near 655 nm). The total concentration of ZnTPP is 1.5 × 10−6 M, and the solvent composition is varied from pure ethanol to 90% v/v water in ethanol. Measurements were made using front-face illumination of solutions of low ZnTPP total concentration in a triangular cell to avoid absorbance artifacts. 11183
DOI: 10.1021/acs.jpcb.7b09868 J. Phys. Chem. B 2017, 121, 11180−11188
Article
The Journal of Physical Chemistry B absorption spectrum of the unaggregated monomer is faithfully reproduced18 and this spectrum is retained at constant intensity at solvent compositions up to 30% v/v water. However, as the water content of the solution increases still further, the intensity of the monomeric signal first decreases, and then, near 40% v/v water ( f H2O = 0.68), the spectra begin to broaden and develop significant intensity to the red of the monomeric Soret bands. At 50% v/v these excitation spectra have lost most of the identifiable features of the monomer, and their broad residual intensities coincide well with the wavelengths assigned to dimeric aggregates29,30 (cf. the absorption spectra of Figure 1). At still larger water contents, the excitation spectra lose their dimeric character and appear to consist of composites of the spectra of a multitude of aggregates. In solutions with somewhat higher total ZnTPP concentrations (e.g., 2.5 × 10−6 M), these aggregation phenomena appear in solutions of somewhat lower water content. These steady state absorption, emission, and excitation spectra then enable excitation conditions to be chosen that probe the differences between the dynamic photophysical properties of the ZnTPP monomer and its aggregates that are formed in the ground state. Here we focus primarily on the monomer and dimer, both of which are present in measurable quantities in solutions in which the mole fraction of water is varied between 0.5 and 0.7 when the total concentration of ZnTPP is in the 0.5−2.5 μM range. We excite with femtosecond pulsed lasers on the blue edge of the Soret band at 400 nm where the monomer and the aggregates all absorb, but sufficiently removed from the S2 emission range to avoid both excitation and Raman scatter. The fluorescence decay of the S2 state of the monomer, observed at 436 nm, and the fluorescence rise of the S1 state, observed at 620 nm, are measured by fluorescence upconversion with ca. 100 fs time resolution. The S1 fluorescence decay, also observed at 620 nm, is obtained by time-correlated single photon counting with ca. 100 ps resolution. Typical time-resolved data are shown in Figures 4, 5, and 6, and the parameters derived from them using fits to mono-, bi-, and triexponential decay functions are collected in Table 1. Individual S1 decays, together with their distribution of weighted residuals and values of the corresponding reduced chi-squared function, are shown in Figure S2 in the Supporting Information. The S2 decay times are numerically equal to the S1 rise times within experimental error in all cases for which it was possible to measure the S2 decay times accurately, including those for solutions of large water content in which the fluorescence intensities are greatly reduced and a large fraction of the 400 nm excitation light is absorbed by aggregates. Nevertheless, fitting these data to single exponential functions in which τS2,decay = 2.2 ± 0.1 ps and τS1,rise = 2.1 ± 0.2 ps remains satisfactory even though the emission intensities are greatly reduced when the water content is >40% v/v (f H2O > 0.68). At still larger water contents where aggregation has consumed almost all of the ZnTPP monomer, the S2 emission intensity is reduced to near zero, but the S1 fluorescence intensity is still measurable, and exhibits a significantly shorter rise time. The conclusion is that excitation at 400 nm populates the Soret states of many aggregated ZnTPP species in these solutions, but only the monomers are giving rise to measurable S2 fluorescence. Because aggregates excited in their Soret bands at their ground state equilibrium geometry (geometries) do not fluoresce significantly from S2 but do yield emission from S1
Figure 4. Normalized S2 fluorescence decays of ZnTPP monomer in water−ethanol solutions of varying water content. The total ZnTPP concentration is 4.8 × 10−6 M, and the observation wavelength is 436 nm. The data are fit to a monoexponential functions with τS2 = 2.2 ± 0.1 ps for solutions with the lowest water content. The dotted line is the instrument response function.
Figure 5. Normalized S1 fluorescence rise of 4.8 × 10−6 M ZnTPP in water−ethanol solutions of varying water content. The emission is observed at 620 nm. The data are fit to monoexponential functions; values of the time constants are collected in Table 1.
monomers, we conclude that the excited aggregates, including dimers, undergo rapid (≪1 ps) internal conversion, and that a significant fraction of that relaxation populates the corresponding S1 state(s) of the aggregates from which the fluorescing monomers are derived. The overall S1−S0 fluorescence intensities, corrected for changes in the total absorptivity at 400 nm, decrease significantly as the mole fraction of water increases beyond the onset of aggregation, suggesting that excitation in the Soret bands of the aggregates does not produce a quantitative yield of S1 fluorescent product. Nevertheless, the transition from monoexponential to biexponential decay at solvent compositions near the onset of aggregation is obvious and values of the reduced chi-squared function applied to biexponential 11184
DOI: 10.1021/acs.jpcb.7b09868 J. Phys. Chem. B 2017, 121, 11180−11188
Article
The Journal of Physical Chemistry B
assembly of larger aggregates (cf. the corrected emission spectra, Figure 2). In order to extract more quantitative data, the absorption spectra, expressed in terms of molar absorptivities vs wavenumber, were modeled by a sum of Gaussians, using four Gaussians for the Q band and three for the Soret band for the monomer. Additional Gaussians were then added to account for the presence of significant quantities of aggregates when the water content of the solvent increased, whereas the Gaussians for the monomer were kept at the same relative sizes, but multiplied uniformly by a factor 0 < B < 1 to scale the monomer’s concentration. The analysis was restricted to solution conditions where the spectroscopic and lifetime evidence suggest that the only aggregate present in significant quantity is the dimer. Further details of the modeling procedure are provided in the Supporting Information. In pure ethanol and in ethanol−water solutions of low water content and ZnTPP concentrations up to 2.5 × 10−6 M, the integrated absorptivities of the Q and Soret band solutions are 2.54 × 107 and 4.21 × 108 M−1 cm−1 cm−1 respectively. The ratio is 16.6, in excellent agreement with the experimental ratio of the oscillator strengths for the two transitions in the monomer. Moreover, the ratio of the integrated molar absorptivity of the dimer to that of the monomer is within ca. 10% of the expected ratio of 2:1.27 The 2M ⇆ D equilibrium is expected to dominate at 35% v/v water and 2.5 × 10−6 M total ZnTPP concentration. Modeling of the data from this sample produced a B factor of 0.43 and a monomer concentration, CM, of 1.1 × 10−6 M. At this particular solution composition, then, a value of the dimerization equilibrium constant, KD, equals 5.8 × 105 M−1, using the stoichiometrically derived expression
Figure 6. S1 fluoresence decays of ZnTPP (total concentration 4.8 × 10−6 M) in water−ethanol solutions of varying water content. The decays are fit using monoexponential functions at low water content and by biexponential functions when the water content of the solvent promotes significant dimer formation. At f H2O > 0.7, the decays converge on an average lifetime of ca. 0.4 ns.
deconvolutions in these solutions indicate that the data can be adequately modeled with a single ground state monomer− dimer equilibrium in this range. At still higher water contents, multiple equilibria involving larger aggregates that also absorb at 400 nm produce a statistically unsatisfactory fit of the data using a biexponential decay function. The amplitude of the longest-lived component, assigned to excitation of the fluorescent monomer, decreases to zero and multiple exponential functions produce a better, but still statistically unsatisfactory, fit to the data. Nevertheless, in the larger water content solutions the trend to a near uniform average S1 fluorescence decay time of ca. 0.4 ns is obvious (cf. Figure 6 and Figure S2C) despite the fact that many different aggregate species are initially excited at 400 nm. This trend, together with the observation that the fluorescence spectrum in the S1 range changes very little as the composition of the solution changes, suggests that, after excitation of the aggregates to S2 and rapid (≪1 ps) internal conversion to the aggregate S1 state, the aggregate dissociates, forming the fluorescing monomer with quantum efficiencies that diminish to about 0.1 for the
KD =
CT − CM 2CM 2
(1)
where CT is the total concentration of the solute. Similar information can be derived from the S1 fluorescence decay data displayed in Table 1. The values of Fi provide a measure of the relative integrated emission intensities of the emitting species captured in the fluorescence decays. At a 40% v/v water content, the decay data are well-modeled by a biexponential function and the emission spectra suggest that the only significant emitting species is the excited monomer. We follow the well-established kinetic scheme of Man-Him Hui and Ware31,32 and assign the longer-lived emitter to the
Table 1. Photophysical Parameters for ZnTPP’s S2 Decay and S1 Rise and Decay from Single and Double Exponential Fits to Measured Temporal Fluorescence Profiles as a Function of the Water Content of Ethanol−Water Solutionsa S1 decay water (% v/v)
f H2O
0 10 20 30 40 50 60b 80b
0 0.27 0.45 0.58 0.68 0.76 0.82 0.93
a
S2 decay τ436 (ps) 2.2 ± 2.2 ± 2.2 ± 2.2 ± 2.1 ± 2.1 ±
