Article pubs.acs.org/JPCA
Rapid “Step Capture” of Holes in Chloroform during Pulse Radiolysis Andrew R. Cook,*,† Matthew J. Bird,† Sadayuki Asaoka,‡,§ and John R. Miller† †
Chemistry Department, Brookhaven National Laboratory, Upton, New York 11793-5000, United States Chemical Resources Laboratory, Tokyo Institute of Technology, Yokohama 226-8503, Japan
‡
S Supporting Information *
ABSTRACT: The fundamental process of hole capture in solution was investigated following pulse radiolysis with polyfluorene and 4-cyano-4″-pentyl-p-terphenyl scavengers. Contrary to expectation, a large fraction of holes were captured in experimental time-resolution limited ∼20 ps steps, by a process much faster than diffusion of the initially formed solvent molecular cation. At the highest concentrations, 1.92 mM for a 52 unit long polyfluorene and 800 mM for 4cyano-4″-pentyl-p-terphenyl, 66% and 99%, respectively, of the initially formed holes were captured by 20 ps, with radiation chemical yield G = 1.2 × 10−7 and 1.7 × 10−7 mol J−1. The data can be explained by capture of presolvated holes, analogous to presolvated electrons, possibly possessing extended wave functions, high mobilities, or excess kinetic energy for the first few picoseconds after their creation. Such a process is not generally known in solution; however, the observed step capture as a function of solute concentration is shown to be well explained by this model. In addition to understanding the capture process in solution, the very large step yields formed in 20 ps will provide the ability to resolve subsequent hole transfer on the polymers with >2 orders of magnitude better time resolution than expected.
■
INTRODUCTION Pulse radiolysis has a special ability to rapidly inject charges into solute molecules, sometimes free of the counter charge, to produce species difficult to prepare in other ways. Charge capture by solutes following ionization of the solvent is a fundamental reaction in radiation chemistry. Time resolution can be limited by the width of the ionizing pulse, such as a high energy electron, ion, gamma, or X-ray pulse. Electron accelerator facilities used for pulse radiolysis with picosecond and shorter electron pulses,1−10 such as the Laser Electron Accelerator Facility (LEAF) at Brookhaven National Laboratory,11 enable studies on fast time scales; however, the capture rate of the charges produced often determines the time resolution in experiments. At a typical diffusion controlled rate of 2 × 1010 M−1 s−1, a solute concentration of ∼5 M is required to take advantage of the 10 ps time resolution possible at LEAF. Few samples are this soluble; thus, it is critical to understand the rate and yield at which capture occurs. For example, conjugated polymers, which are of interest because of their use as molecular wires in applications such as organic photovoltaic devices12 and plastic electronics,13 are typically only soluble to a few mM; thus, diffusional capture of charges limits experimental time resolution. Rates of attachment for both holes and electrons by solutes have been measured in many solvents. Two key mechanisms for enhancing rates and yields of electron capture have recently been observed for conjugated polymers: time-dependent rates and capture of electrons prior to solvation. Increases of 1−2 orders of magnitude in capture rate of solvated electrons at short times due to the transient term in the Smoluchowski reaction rate equation14−16 © 2013 American Chemical Society
k(t) = k inf (1 + R eff /(πDt )1/2 )
(1)
k inf = 4πR eff DNA
(2)
have been observed, extending into the nanosecond regime,17−19 due to the large effective capture radius, Reff, of the polymers, well described by theoretical models that extend eq 1 by expressing Reff for small particles approaching ellipsoids or lines of spheres.20−27 Similar enhancement of hole capture rates by polymers may also be possible; however, they will be limited due to the lower solvent cation diffusion coefficients, D, compared to those for solvated electrons. In certain solvents such as cyclohexane and methyl-cyclohexane but not others like n-hexane or iso-octane, capture of holes more than an order of magnitude faster than expected by diffusion controlled rates were reported, attributed to a high mobility positive ion.28−33 While such mechanisms may increase capture yield of cations at short times, data shown below have little indication that their effect in the first 20 ps is substantial; more effective means are needed to capture a large number of holes rapidly to take full advantage of the time resolution possible at LEAF. Efficient capture of electrons before they are fully solvated, so-called “dry” electrons, is known in many solvents. This process occurs on time scales faster than solvation times for the electrons, typically 25 mM, decay of CPT+• is evident in Figure 2. This decay, which continues to longer times, is not present in the pF data and is due to geminate recombination of CPT+• and Cl− to reform CPT and make Cl atoms. Note that, unlike pF, this implies that Cl• cannot appreciably oxidize CPT; thus, the observed diffusional growth is predominantly due to oxidation by CDCl3+•. Like the pF samples, most of the CPT+• production occurs in a fast and nearly electron pulse-width limited process, referred to here as step capture. Step solute oxidation was modeled considering only species formed at short times following ionization of the solvent, chloroform, with methods analogous to those previously used
Figure 2. Step attachment of holes to 4-cyano-4″-pentyl-p-terphenyl in CDCl3 after pulse radiolysis followed by slower diffusional growth within the short lifetime of the solvent radical cation collected with the OFSS experiment, probed at 900 nm. The lack of additional absorbance at ∼20 ps after the highest concentration indicates that nearly all solvent cations that can oxidize solutes were captured.
for reduction pF samples.19,34,35 The main species likely relevant for oxidation of pF and CPT in the observed 20 ps rise steps is the initially formed solvent radical cation, CDCl3+•, possibly before it is solvated. Possible mechanisms for the large amount of rapid capture are explored in the discussion section. For pF only, oxidation by chlorine atoms, many of which may be present at short times, is energetically feasible but however would occur in competition with H atom abstraction from chloroform and the solubilizing alkyl side chains on pF.49 In the first 20 ps, it might be expected that only Cl• formed in contact with pF, within ∼6 Å, may oxidize it, accounting for less than 10% of the observed step. The contribution may be larger than this, as explored in the discussion section. Other oxidizing radicals and ions are formed in chloroform, but they are not energetically capable of oxidizing the solutes in this study. To determine the magnitude of solute oxidation that occurs during the first 20 ps, it is necessary to account for other species that absorb at 580 and 900 nm in the step absorbance rise in Figures 1 and 2, respectively. Consideration was given to potential production of solute radical anions, solute excited states, and the species that absorbs in neat chloroform, also seen in Figures 1 and 2. Reduction of solutes is not expected due to dissociative attachment of the ionized electron to solvent molecules50 or rapid recombination that produces many radicals such as Cl•.51,52 However, if an electron is ejected to a site in contact with a solute molecule, it may be possible that capture by pF or CPT is faster. To determine if solute radical anions are formed, pF solutions were saturated with oxygen (11.6 mM53), which will quench radical anions in ∼10 ns but not radical cations. These measurements did not show any discernible decrease in the observed signals, though will not be sensitive to those few that recombine faster than oxygen can quench them. These results suggest that if formed at all, radical anions are less than a few percent of the signal and will thus be neglected in the analysis. Solute excited states are typically formed in radiolysis experiments by absorption of Cerenkov light, direct excitation by fast electrons, and by ion recombination. Singlets and triplets of pF both have broad absorption bands with maxima near 760−780 nm54,55 but still have significant oscillator strength at 580 nm where CDCl3+• absorbs, while pF+• only has significant absorption at 580 nm. 7714
dx.doi.org/10.1021/jp405349u | J. Phys. Chem. A 2013, 117, 7712−7720
The Journal of Physical Chemistry A
Article
Table 1. Summary of Contributions to the Observed Step Rise in Absorption for Polyfluorenes and 4-Cyano-4″-pentyl-pterphenyl in Deuterated Chloroform solute
[S] (mM)a
f (S+)b
f (EX)c
f (CHCl3+•)d
pF52 pF52 pF52 pF28 CPT CPT CPT CPT CPT CPT CPT
0.48 0.96 1.92 1.79 12.5 25 50 100 200 400 800
0.52 0.74 0.85 0.76 0.44 0.68 0.83 0.92 0.97 1.00 1.00
0.06 0.09 0.10 0.07 0 0 0 0 0 0 0
0.42 0.17 0.05 0.17 0.56 0.32 0.17 0.08 0.03 0.005 0.0002
G (step, mol J−1)e 0.21 0.57 1.17 0.59 0.12 0.31 0.57 0.90 1.40 1.68 1.74
× × × × × × × × × × ×
10−7 10−7 10−7 10−7 10−7 10−7 10−7 10−7 10−7 10−7 10−7
a
Concentrations of 4-cyano-4″-pentyl-p-terphenyl (CPT) and polyfluorene (pF) polymer molecules for pFs with average lengths of 52 and 28 repeat units. bFraction of the observed absorbance step due to solute radical cations. cFraction of step due to solute excited states. dFraction of step due to solvent radical cations. eRadiation chemical yield of step produced solute radical cations.
was determined considering only absorbance of the solvent species at the same wavelength. As noted earlier, absorption of the solvent species was treated assuming it was the primary oxidizing species, CDCl3+•; thus, its contribution to the step rise decreases as the amount of capture increases. The primary difficulty with the analysis was lack of authentic knowledge of the initial concentration of CHCl3+•, that might have been provided if its spectrum and extinction coefficient were wellknown. A reasonable proxy was the highest concentration samples of CPT, where no additional production of CPT+• was found as the concentration was increased. This is further explored in the discussion section. Oxidation after the step in both samples was treated using a single exponential capture rate to produce adequate fits of the data, extrapolated back to t = 0, or halfway up the step rise, to give the step height without contribution from diffusive capture. These fitted rates are faster than expected for diffusive capture, due to contributions from the time-dependent term in eq 1; the possibility of faster rates in the first 10−20 ps where we cannot observe them are neglected. The observed rate may include components due to both CDCl3+• and Cl• in the case of pF. Fitting results shown in Table 1 give fractions of the step absorbance due to solute radical cations, solute excited states, and the solvent species. At the highest concentration of pF and CPT, the step rise in absorption is found to be overwhelmingly due to solute radical cation absorption, with 87% and 98% from pF+• and CPT+•, respectively. If the absorption in the neat solvent was not due to CDCl3+•, it presents the significant source of error. For example, if it is either not bleached by the presence of pF or completely bleached, the fraction of pF+• formed in the step at the highest [pF] would increase by 9% or decrease by 5%, respectively. Errors are larger for lower concentrations where the absorbance in the neat solvent is larger compared to the observed absorption step in the presence of solute. Radiation chemical yields of pF+• and CPT+• formed in the steps per Joule of energy absorbed (G) are also shown in Table 1. The maximum G determined with 800 mM CPT and 1.92 mM pF52 was 1.74 and 1.17 × 10−7 mol J−1, respectively. G values were determined by comparison to solvated electrons produced in identical cells containing a standard water solution shot with the same dose, containing 20% methanol and 1 M NaOH. It was assumed that this solution has the same G value of 4.1 × 10−7 mol J−1 at 20 ps as neat water.57 Solvated electrons in neat water have a known extinction coefficient of
Data collected in chloroform at 780 nm show production of pF excited states, seen in Figure S3 (Supporting Information), with a magnitude at 780 nm 30−35% the size of the signal at 580 nm. Previously, we determined that 1pF* and 3pF* extinction coefficients were 2.71 and 3.53 times smaller at 580 than at 780 nm in THF, respectively.34 On the basis of these relative extinction coefficients, excited states account for 7−10% of the observed step rise at 580 nm for pFs. As noted earlier, CPT is simpler with no excited state absorption at 900 nm. Difficulties in the quantitative analysis of step capture data stem from lack of knowledge of the yields and spectroscopic identity of chloroform radical cations. There is a broad absorption seen throughout the visible to the near-IR following radiolysis of neat chloroform, as seen in Figure S2 (Supporting Information). The identity of this absorption is important, as it contributes to the observed signals in Figures 1 and 2. It may be due to more than one species with overlapping spectra. Washio has suggested that the absorption at 550 nm is due to CHCl3+•, with a lifetime of ∼2 ns.56 At 580 nm in neat CDCl3, OFSS data is best represented with two decay components of nearly equal magnitude with lifetimes of 0.26 and 1.78 ns, suggestive of rapid geminate recombination of CDCl3+• with Cl− in this reasonably nonpolar liquid. Deuterated chloroform was used because it gave approximately double the lifetime of the visible band, though if this band is due to CDCl3+•, the reason for the longer lifetime is not clear. The lifetime and magnitude of the absorption in neat CDCl3 was reduced by addition of high ionization potential solutes, such as hexafluorobenzene, that are not expected to react with species other than CDCl3+•. The slowest ∼1 ns growth seen for CPT is consistent with a short CDCl3+• lifetime in competition with capture. In lower concentrations of pF, slower growths with a lifetime up to ∼3 ns are observed, likely due to additional oxidation by Cl•, in competition with H atom abstraction from chloroform and the alkyl groups on both pF and CPT.49 These observations are consistent with Washio’s assignment and thus, the visible and NIR solvent absorptions were treated as due to CHCl3+•; however, uncertainty in the identity of the species involved may be a source of error in the results. Contributions to the observed step increases in absorption due to both pF* and the solvent species were determined in an iterative fashion between 580 and 780 nm where they have differing extinction coefficients, similar to the technique used for anions in previous work.34 At 900 nm, CPT+• production 7715
dx.doi.org/10.1021/jp405349u | J. Phys. Chem. A 2013, 117, 7712−7720
The Journal of Physical Chemistry A
Article
22 500 M−1 cm−1 at 720 nm;58 corrections were made for the slightly blue-shifted spectrum and ∼9% stronger peak absorption observed in the standard solution compared to neat water. Corrections were also made for the differences in energy deposition due to densities of the standard solution and the chloroform solutions. Relative error of ±20% in this calculation is possible due to uncertainties in solute radical cation extinction coefficients, and the difference in scattering of the electron beam between the standard water and chloroform solutions. For both pF and CPT, the yield of oxidation in the step is surprisingly large. While yields were not determined in previous work on electron capture by pF in THF,34 the yield of hole capture in CDCl3 at the same 50 mM (repeat units) of pF determined here likely is quite similar. Such large unprecedented yields of hole capture in
