THE JOURNAL OF CHEMICAL PHYSICS 142, 084309 (2015)

Observation of structural relaxation during exciton self-trapping via excited-state resonant impulsive stimulated Raman spectroscopy J. G. Mance,a) J. J. Felver, and S. L. Dexheimerb) Department of Physics and Astronomy, Washington State University, Pullman, Washington 99164-2814, USA

(Received 3 October 2014; accepted 3 February 2015; published online 26 February 2015) We detect the change in vibrational frequency associated with the transition from a delocalized to a localized electronic state using femtosecond vibrational wavepacket techniques. The experiments are carried out in the mixed-valence linear chain material [Pt(en)2][Pt(en)2Cl2]·(ClO4)4 (en = ethylenediamine, C2H8N2), a quasi-one-dimensional system with strong electron-phonon coupling. Vibrational spectroscopy of the equilibrated self-trapped exciton is carried out using a multiple pulse excitation technique: an initial pump pulse creates a population of delocalized excitons that self-trap and equilibrate, and a time-delayed second pump pulse tuned to the red-shifted absorption band of the self-trapped exciton impulsively excites vibrational wavepacket oscillations at the characteristic vibrational frequencies of the equilibrated self-trapped exciton state by the resonant impulsive stimulated Raman mechanism, acting on the excited state. The measurements yield oscillations at a frequency of 160 cm−1 corresponding to a Raman-active mode of the equilibrated self-trapped exciton with Pt-Cl stretching character. The 160 cm−1 frequency is shifted from the previously observed wavepacket frequency of 185 cm−1 associated with the initially generated exciton and from the 312 cm−1 Raman-active symmetric stretching mode of the ground electronic state. We relate the frequency shifts to the changes in charge distribution and local structure that create the potential that stabilizes the self-trapped state. C 2015 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4908155]

I. INTRODUCTION

The localization of electronic excitations to form selftrapped polaronic states reflects the fundamental physics of coupling between electronic and vibrational dynamics. In the self-trapping process, a carrier in a delocalized free electronic state interacts with a deformable lattice to induce displacements of the atoms from their equilibrium positions in the absence of the carrier, creating the potential well that stabilizes the localized self-trapped state. Self-trapping is a highly nonlinear process because the potential well that traps the electronic excitation is itself dependent on the electronic wavefunction of the carrier that induces the local distortion of the lattice. The dynamics of the self-trapping process have attracted considerable interest, and recent theoretical work has focused on the development of models for the evolution of both the electronic and vibrational degrees of freedom during the localization process.1–3 Measurement of the vibrational frequencies provides a probe of the evolution of the potential that stabilizes the selftrapped state. In the work reported here, we investigate the vibrational frequency shifts associated with the formation of a self-trapped exciton (STE, also referred to as an excitonpolaron or neutral bipolaron) by comparing the vibrational frequency associated with the initially generated exciton to that of the final equilibrated STE that is fully dressed by the phonon a)Present address: Spectral Energies, LLC, Dayton, Ohio 45431, USA. b)Author to whom correspondence should be addressed. Electronic mail:

[email protected] 0021-9606/2015/142(8)/084309/6/$30.00

field. Previous studies have revealed vibrational wavepacket modulations as the initially excited system evolves to form the self-trapped state.4,5 However, since the wavepacket oscillations dephase as the STE forms, information about the vibrational frequencies is lost as the system evolves to its final equilibrated structure. In the work presented here, we use a multiple pulse method that applies the resonant impulsive stimulated Raman mechanism to the excited state6 to measure the vibrational frequency of the equilibrated STE. The measurements reveal a softening of the vibrational frequency that reflects the change in electronic distribution and local structure as the excitation undergoes the transition from a delocalized to a localized state. The experiments are carried out in a mixed-valence metal-halide linear chain complex, a class of quasi-one-dimensional materials that are of interest as model systems for correlated electron physics because of their structural tunability: variation of their chemical composition allows systematic control of the strengths of the electronphonon and electron-electron interactions that determine their electronic properties.4,7–9 The platinum-chloride material studied here has an exceptionally strong electron-phonon coupling, resulting in the formation of highly localized selftrapped electronic excitations with a large change in charge distribution from the ground electronic state. The expected changes in charge distribution and atomic displacements upon formation of the STE are shown schematically in Fig. 1. Fig. 1(a) shows a schematic representation of the chain axis structure of the ground electronic state of the [Pt(en)2][Pt(en)2Cl2]·(ClO4)4 complex (en = ethylenediamine, C2H8N2), referred to as PtCl(en), in which the one-dimensional

142, 084309-1

© 2015 AIP Publishing LLC

084309-2

Mance, Felver, and Dexheimer -

+(3+δ)

-

+(3+δ)

(a)

Cl - Pt

(b)

Cl - Pt

+(3-δ)

-

- Cl

-

Cl

+3

Pt

+(3+δ)

-

Pt

Cl - Pt -

Cl

J. Chem. Phys. 142, 084309 (2015)

+3

Pt

-

- Cl

-

Cl

+(3-δ)

Pt

+(3-δ)

Pt

FIG. 1. (a) Schematic structure of the PtCl(en) complex along the chain axis, showing two repeat units of the Peierls-distorted charge density wave ground state. The charge disproportionation parameter δ ∼ 0.91,12,13 so that the Pt ions have alternating valences of nearly +4 and +2. (b) Schematic structure of the equilibrated STE created by structural relaxation following excitation of the optical intervalence charge transfer transition.

geometry is defined by an extended linear chain of Pt and Cl ions. Ethylenediamine ligands fill the transverse bonding sites of the Pt ions, and in the three-dimensional crystal structure,10,11 parallel chains are separated by ClO4− counterions, with sufficiently weak interchain interactions for the material to maintain a quasi-one-dimensional character. The structure of the ground electronic state is a Peierlsdistorted commensurate charge density wave (CDW), with alternating charges on the Pt ions and a corresponding variation in the Pt–Cl bond lengths. The strong electronphonon coupling in PtCl(en) is reflected in its exceptionally large Peierls distortion, with a ratio of the short- to longPt–Cl bond lengths of 0.743,10 and its nearly complete charge disproportionation,12,13 so that Pt ions of nominal valence +3 yield a CDW with alternating valences of nearly +4 and +2. Formation of the STE is triggered by excitation of the optical intervalence charge transfer (IVCT) transition, a strong optical absorption polarized along the chain axis that effectively transfers charge between inequivalent platinum ions. Owing to the one-dimensional band structure formed by bonding along the platinum-halide chain, the final state of the transition is delocalized along the chain, and theoretical modeling indicates that the resulting electronic state has strong excitonic character.14 The system undergoes rapid structural relaxation to form the metastable STE, with the expected changes in charge distribution and atomic displacements shown schematically in Fig. 1(b). In the mixed-valence linear chain materials, the STE consists of a localized region in which the CDW is reduced and the Peierls distortion of the metal-halide bond lengths is accordingly relaxed. Modeling of the properties of the electronic excitations in the platinumhalide linear chain complexes using many-body techniques has shown that their localization length, or spatial extent, varies systematically with interaction strength as controlled by the bridging halide ion.7,15–18 The STE in the strongly coupled PtCl(en) complex approaches the small polaron limit, in which the excitation is confined to a single repeat unit of the chain. This limit corresponds to essentially complete charge transfer between two adjacent Pt ions, yielding nominal charge states of Pt+3 with relaxed bond lengths, as shown in Fig. 1(b). The STE in PtCl(en) is strongly trapped, as evidenced by its large Stokes shift, with the IVCT optical absorption peak energy at 2.5 eV and the luminescence peak energy at 1.2 eV.12 In our previous work, we time-resolved the coupled vibrational and electronic dynamics of the formation of the STE in a series of platinum-halide linear chain materials using femtosecond vibrational wavepacket techniques.4,5,19–22

In these experiments, a short optical pump pulse excites the IVCT band, generating excitations in the delocalized free exciton state. The formation of the STE is detected by monitoring the appearance of its characteristic red-shifted absorbance with a time-delayed probe pulse. The associated initial vibrational dynamics are detected by carrying out the pump-probe measurements in the vibrationally impulsive limit, in which the optical pulses are short compared to the periods of the characteristic vibrations of the material. In this case, the vibrational modes coupled to the electronic transition are impulsively excited, creating nonstationary vibrational wavepackets made up of coherent superpositions of vibrational states. The wavepackets oscillate at the characteristic vibrational frequencies of the material and are observable as a timedependent modulation of the pump-probe signal.23 In general, impulsive optical excitation resonant with an electronic transition will generate vibrational wavepackets on both the ground state and the excited state potential energy surfaces. Ground state vibrational wavepackets are generated by the resonant impulsive stimulated Raman process, yielding oscillations at the same frequencies observed in a cw resonance Raman measurement. Excited state wavepackets involve vibrational modes coupled to the electronic excitations, and their evolution reflects the vibrational dynamics in the excited electronic state, which in the present case involve structural changes associated with self-trapping. The quasi-one-dimensional structure of the linear chain complexes simplifies the observation of the vibrational dynamics associated with self-trapping because the exciton is coupled to a limited number of phonon branches. The structural relaxation results primarily from coupling to optical phonons involving motion of the halide ions, and the Raman-active X-Pt-X symmetric stretch mode, which is strongly coupled to the IVCT transition as a result of the dependence of the Pt-X bond lengths on Pt valence state, is expected to significantly contribute to the formation of the localized state. In our previous studies of exciton self-trapping in PtCl(en),4,5 we observed the formation of the STE following excitation of the IVCT transition with pulses 25 fs in duration centered at 600 nm. The characteristic red-shifted optical absorption of the STE, detected with probe pulses 35 fs in duration centered at 800 nm, appears with a formation time of 180 fs, on the order of a single vibrational period of the optical phonon response, consistent with barrierless self-trapping as theoretically predicted for one-dimensional systems.24 The optical response is strongly modulated by excited state wavepacket oscillations that damp during the formation of the STE absorbance, with an oscillation frequency of 185 cm−1 (period of 180 fs), strongly shifted from the 312 cm−1 frequency of the Raman-active symmetric stretching mode of the ground electronic state. These measurements capture the initial coupled electronic and vibrational dynamics of the self-trapping process, in which the 185 cm−1 wavepacket modulation reflects the atomic motions that carry the photoexcited system from its initial structure, which according to the Franck-Condon principle is the Peierls-distorted structure of the ground electronic state, toward the structure that stabilizes the STE. A small low frequency oscillatory component at ∼ 70 cm−1 also appears in the response, consistent with

084309-3

Mance, Felver, and Dexheimer

contributions to self-trapping from coupling to acoustic phonons in the small polaron limit.19,25–28 Excited-state wavepacket motion associated with the initial vibrational dynamics of exciton-self trapping has now been detected in a number of studies on related linear chain materials using transient absorption or time-resolved luminescence.4–6,19–22,29–34 An inherent limitation of these experiments comes from the dephasing of the wavepacket oscillations as the STE forms and relaxes, which results in the loss of information on the vibrational properties as the system evolves to its final equilibrated state. The multiplepulse excitation method used here accesses the vibrational properties of the STE in its equilibrated structure, probing the nature of the potential well that stabilizes the localized self-trapped state. In this method, excited-state resonant impulsive stimulated Raman spectroscopy,6 an initial pump pulse produces a population of excitons. After a delay to allow self-trapping and relaxation, a second pump pulse that is spectrally shifted to lie within the STE absorption band impulsively excites the equilibrated STEs, generating wavepacket oscillations at the frequency of the Raman-active mode of the STE via the resonant impulsive stimulated Raman mechanism acting on the excited state. This approach allows the measurement of excited-state vibrational frequencies in a time and frequency range that is not easily accessible using other techniques:35 conventional time-resolved Raman spectroscopy is limited in temporal and frequency resolution by the time-bandwidth constraints of the uncertainty principle, precluding the measurement of low-frequency vibrational modes on short time scales, and while femtosecond stimulated impulsive Raman spectroscopy provides an elegant means of accessing high frequency vibrational modes with high time resolution, the phase structure of the continuum probe at frequencies close to that of the laser fundamental can make detection of low frequency modes difficult. The multiple pulse time-domain approach employed here has now been successfully applied in both molecular and solid-state systems to detect low frequency excited state vibrational modes on short time scales.6,36–39

J. Chem. Phys. 142, 084309 (2015)

was monitored as a function of time delay between the second pump and the probe pulse. All beams were polarized parallel to the chain axis. To enhance the oscillatory response, the probe beam was wavelength resolved following transmission through the sample using a 10-nm bandwidth bandpass filter. The excited state response was detected by simultaneously chopping both pump beams at different chopping rates with a dual-frequency chopper wheel and monitoring the probe intensity at the sum of the modulation frequencies using lockin amplification. This method allows a clear measurement of the resonant impulsive stimulated Raman response of the equilibrated STE because the measurement is sensitive only to the double-modulated signal that requires the presence of both pump pulses (i.e., ∆(∆T/T)) and is not sensitive to the individual pump responses. Transient absorption measurements were also carried out using single-pump-probe pulse sequences for comparison with the double-modulated pumppump-probe response. Oriented single-crystal samples of [Pt(en)2][Pt(en)2Cl2] ·(ClO4)4 were mounted on 1-mm thick sapphire plates and were maintained at room temperature. The samples used in these studies were synthesized using established methods40 that yield materials that are largely free of defects and that are resistant to photo-induced damage. Component analysis of the measurements was carried out using both linear prediction/singular value decomposition (LPSVD) and nonlinear fitting to a sum of exponentially damped cosine waves and decaying exponentials (i.e., the same functional form that is inherent to the LPSVD signal processing technique). In general, successful multi-parameter nonlinear fitting requires careful choice of the initial values for the fitting parameters, and for the fitting results presented here, the number of oscillatory components and their starting parameters were determined from the LPSVD analysis. The methods for determination of the frequency components yielded consistent results, and the nonlinear fitting results are presented here, since this method also provides error limits. III. RESULTS

II. METHODS

Measurements were carried out using optical pulses derived from a 35-fs, 1 kHz Ti:sapphire regenerative amplifier. A portion of the 800-nm amplifier output was frequency-doubled in a 200 µm thick β-barium borate crystal to generate the first pump pulse that excites the IVCT transition to produce the population of excitons. This 400-nm pump pulse was partially compressed with a pair of fused silica prisms to a duration of ∼100 fs and was kept sufficiently long to avoid excitation of long-lived wavepacket oscillations from the 312 cm−1 ground state Raman-active vibrational mode. A portion of the 35-fs 800-nm output was passed through a beamsplitter to generate the second pump pulse and the probe pulse. The experimental geometry was the same as used in our previous excited-state resonant impulsive stimulated Raman studies:6 the three pulses (pump 1, pump 2, and probe) were focused into the sample with a fixed delay between the first and second pump pulses, and the transmitted probe intensity

To determine an appropriate time delay between the first and second pump pulses in the pump-pump-probe measurements, the relaxation of the STEs was first monitored in a single-pump-probe measurement in which excitons were generated by excitation of the IVCT using the same 400 nm, ∼100 fs pulse used as the first pump pulse in the pump-pumpprobe sequence and probed in the red-shifted STE absorption band using the same 800 nm, 35-fs pulse used as the probe. The result of this measurement is shown in Fig. 2, which presents the pump-induced change in transmittance ∆T/T detected on the red side of the probe spectrum at 820 nm. The response includes the same STE absorbance and excited-state wavepacket components observed in our previously reported studies of the initial self-trapping dynamics in PtCl(en),4,5 though the STE formation component and wavepacket oscillations are convolved with the longer duration pump pulse used here and result in a substantially lower excited-state wavepacket modulation amplitude. The relatively long pump

084309-4

Mance, Felver, and Dexheimer

FIG. 2. Response of PtCl(en) to the first pump pulse used to the set time delay between the first and second pump pulses in the excited-state resonant impulsive stimulated Raman pump-pump-probe measurement. The IVCT band is excited with a single 400 nm, 100 fs pump pulse, and the timeresolved differential transmittance is probed within the red-shifted absorption band of the STE. Though the response is distorted by the 100 fs pump pulse duration, it contains the same excited state components detected in previous measurements carried out at higher time resolution,4,5 as discussed in the text. The 800 fs delay time for the second pump pulse used in the pump-pump-probe measurements is chosen to allow decay of the vibrational wavepacket response and relaxation to the final self-trapped structure and is indicated by an arrow.

pulse duration also prevents the observation of the higher frequency wavepacket oscillations at the ground state Raman frequency of 312 cm−1 in this measurement. The key result for the present work from the measurement shown in Fig. 2 is that the excited state wavepacket response is seen to fully damp within 800 fs, and this is the chosen as the time delay between the first and second pump pulses in the double pulse excitation measurements so as to avoid contributions from residual vibrational coherence originating from the first pump pulse. Excited-state resonant impulsive stimulated Raman measurements to probe the vibrational properties of the equilibrated STE are presented in Fig. 3(a). In the pump-pumpprobe sequence, the first pump pulse (400 nm, 100 fs) excites the IVCT transition to generate a population of excitons, and after an 800 fs delay to allow relaxation to the equilibrated

J. Chem. Phys. 142, 084309 (2015)

self-trapped structure and decay of the wavepacket response generated by the initial pump pulse, the second pump pulse (800 nm, 35 fs) resonant with the red-shifted STE absorption spectrum impulsively excites the population of STEs. The resulting wavepacket motion is observed as a function of time delay between the second pump pulse and the probe pulse, also 35 fs in duration and centered at 800 nm. The signal is detected at a probe wavelength of 820 nm, on the side of the probe pulse spectrum, to enhance the amplitude of the resonant impulsive stimulated Raman wavepacket oscillations.41 A nonlinear fit to a sum of exponentially damped cosine waves and a decaying exponential background is shown along with the data in Fig. 3(a), and the resulting oscillatory components are displayed in Fig. 3(b). The analysis of the pump-pump-probe response reveals two frequency components, a long-lived component at the 312 cm−1 frequency of the strongly Raman-active ground state symmetric stretch chain axis mode and a new oscillatory component at 160 cm−1. We assign the new frequency component at 160 cm−1 to a Raman-active vibrational mode of the equilibrated STE, excited by resonant impulsive stimulated Raman excitation of the STE. We note that this component is detected only in the double-modulated measurement: it is neither observed in the single-pump-probe measurement using excitation at 400 nm, shown in Fig. 2, nor in a single-pump-probe measurement using excitation at 800 nm (data not shown). The 160 cm−1 frequency is also absent from the ground state Raman spectrum. The oscillatory component at the ground state symmetric stretch frequency of 312 cm−1 likely appears in the double-modulated signal because the first pump pulse depletes the population of the ground electronic state, thereby reducing the amplitude of the impulsive stimulated Raman ground state response generated by the second pump pulse. IV. DISCUSSION

The observation of softening of the 185 cm−1 wavepacket oscillation, corresponding to the initial vibrational motion after generation of the free exciton, to the 160 cm−1 frequency of

FIG. 3. (a) Vibrational response of the equilibrated STE detected via excited-state resonant impulsive Raman spectroscopy using a pump-pump-probe sequence (blue), with a fit to a sum of two exponentially damped cosine waves and a decaying exponential background (red). Signals near zero delay include features from nonlinear interaction of the overlapped pump and probe pulses and are not included in the fit. (b) Individual oscillatory components of the fit, including a new frequency at 160 ± 5 cm−1, assigned to the equilibrated STE, and 312 ± 1 cm−1, the frequency of the Raman-active symmetric stretch mode of the ground electronic state.

084309-5

Mance, Felver, and Dexheimer

the final equilibrated STE reveals a new component of the selftrapping dynamics, indicating that further structural evolution takes place during the self-trapping process beyond the initial vibrational wavepacket dynamics. Both the 185 cm−1 frequency from the initially excited state and the 160 cm−1 frequency of the equilibrated STE are significantly lower than the 312 cm−1 frequency of the Raman-active ground state symmetric stretch vibration that is coupled to the IVCT. All of these vibrational modes are expected to involve Pt–Cl stretching motions with frequencies that are strongly dependent on the local charge states of the Pt ions. Force constant analysis of the ground state vibrational properties of PtCl(en)42 has shown that the ground state symmetric stretch frequency is determined primarily from interactions within a Cl–Pt+4–Cl unit, since the Pt+4–Cl force constant is much greater than that for Pt+2–Cl, and has also shown that the dispersion of the ground state symmetric stretch phonon mode is very weak, with an increase in frequency of only ∼7 cm−1 from the zone center to the edge of the first Brillouin zone.4,42 In principle, dispersion in the phonon branch coupled to the self-trapping dynamics could result in vibrational frequency shifts associated with localization: in a simple semi-classical model, the initially delocalized electronic excitation would couple primarily to phonons at low q-vector, followed by coupling to phonons of higher q as the electronic wavefunction collapses to the final localized state. Such a mechanism would result in a shift from the zone center frequency (i.e., the value observed in a ground state Raman measurement) to the frequency at a q corresponding to the inverse localization length, which in the small polaron limit would reach the zone edge. However, this mechanism clearly does not explain the observed frequency shifts for selftrapping in PtCl(en), both because of the small magnitude of the dispersion of the symmetric stretch optical phonon branch and because of the sign of the dispersion, which shows a frequency that increases, rather than decreases, from the zone center to the zone edge.4 We note that both the 185 cm−1 and 160 cm−1 frequencies lie outside of the ground state phonon dispersion bands, and moreover, we note that because of the large charge redistribution associated with the formation of the STE, the associated vibrational modes would not be expected to correspond to a simple linear combination of the ground state phonon modes.16 Since the Pt–Cl elastic constants are strongly dependent on the Pt ion charge density, we expect that the shift in the excited state frequency from 185 cm−1 to 160 cm−1 reflects, in large part, the degree of delocalization of the excited state electronic wavefunction through the change in charge density on the Pt ions. The electronic wavefunction of the initial “free exciton” state created by excitation of the IVCT transition is expected to be relatively more delocalized and results in partial reduction of the amplitude of the CDW and a corresponding change in the Pt–Cl elastic constants from their ground state values. After the structure relaxes to the final STE state, the electronic wavefunction is more localized, with the Pt ion charges approaching Pt+3 states, as shown in Fig. 1(b), which presents a schematic structure for the STE in the small polaron limit of confinement to a single unit cell. The evolution of the charge density to form the more localized final STE state

J. Chem. Phys. 142, 084309 (2015)

results in an additional weakening of the force constants, reducing the local vibrational frequency. Our observation of a Raman frequency for the equilibrated STE that is significantly lower than that of the 312 cm−1 ground state symmetric stretch optical phonon is qualitatively consistent with the theoretical modeling of Bishop and co-workers, who calculated the vibrational properties of polaronic states in the platinum-halide linear chain materials using many-body models.7,16,17 These calculations predict the presence of localized Raman-active vibrational modes associated with the polaronic states that involve displacements of atoms that lie within approximately the same region as the envelope of the electronic wavefunction, corresponding to “shape modes” of the nonlinear electronic excitations. For PtCl(en), we find a frequency shift larger than that predicted by the calculations,17 suggesting that the change in charge distribution upon formation of the STE is larger than that estimated by the parameterized model used in the calculation, with a final state closer to the small polaron limit. Since these authors note that the ionic character of PtCl(en) results in anharmonic interactions, the observation of a large frequency shift associated with ionic charge redistribution suggests that anharmonic effects may play a significant role in the localization dynamics.43 The structural tunability of the PtX(en) linear chain complexes allows a comparison of the properties of the selftrapped states in materials with different coupling strengths. In our previous work, we investigated the vibrational frequency shift associated with relaxation to the final equilibrated STE state in PtBr(en) using a similar experimental approach.6 This material has a smaller electron-phonon coupling as evidenced by its ground state properties: a smaller Peierls distortion, with a ratio of the short- to long-Pt–halide bond lengths of 0.828, compared to 0.743 in PtCl(en),10 and a lower amplitude CDW, with a charge disproportionation parameter δ ∼ 0.64, compared to δ ∼ 0.91 in PtCl(en).12,13 As a result of the smaller coupling strength, the STE in PtBr(en) is more delocalized, with a theoretically predicted spatial extent of ∼5 Pt2Br2 units, compared to 1 and 2 units in PtCl(en).7,16,17 Our excited-state resonant impulsive stimulated Raman measurements on PtBr(en) revealed a Raman frequency for the equilibrated STE of 125 cm−1, again significantly shifted from the frequency of the Raman-active symmetric stretch mode in the ground electronic state, which is 171 cm−1 in this complex. The frequency shift between the ground state and the equilibrated STE seen in PtBr(en) is not as large as that reported here for PtCl(en), consistent with a smaller change in the charge distribution upon formation of the self-trapped state, as the charge transfer in PtBr(en) is spread over a larger number of Pt2X2 repeat units. The experiments on PtBr(en) also showed a small, but observable, frequency shift in the excited state, with a wavepacket oscillation frequency of 110 cm−1 associated with the initially generated “free exciton” state compared to 125 cm−1 for the final equilibrated STE. Interestingly, the excited state vibrational frequency shifts to a higher value upon equilibration in PtBr(en), compared to a shift to lower frequency in PtCl(en). This may result from the larger spatial extent of the STE in PtBr(en), which could result in more complex vibrational modes for the STE compared to

084309-6

Mance, Felver, and Dexheimer

those in PtCl(en), where the STE is highly localized, or from differences in the extent of contributions from anharmonic interactions to the potential that stabilizes the self-trapped state, which are expected to be less significant in PtBr(en) compared to PtCl(en). In any case, the observation of a shift in the excited state vibrational frequency between the initially generated state and the final self-trapped state indicates that the structure continues to evolve during the self-trapping process in both of the materials.

V. CONCLUSION

In conclusion, we have detected structural relaxation associated with the formation of a self-trapped exciton in the small polaron limit by observing the change in vibrational frequency upon relaxation to the final equilibrated selftrapped state, using a pump-pump-probe sequence to probe the resonant impulsive stimulated Raman response of the excited state. The observed vibrational frequency shift can be understood in terms of the evolution of the system from a delocalized to a localized electronic state and reflects the evolution of the potential that stabilizes the self-trapped state, created by the coupled vibrational and electronic dynamics of the self-trapping process.

ACKNOWLEDGMENTS

This work is supported by the National Science Foundation under Grant Nos. DMR-1106379 and DMR-0706407. We thank J. A. Brozik for synthesis of the PtCl(en) complex and A. Saxena for discussions of Ref. 17. We also thank S. A. Trugman for discussions of his work supported by the Center for Integrated Nanotechnologies at Los Alamos National Laboratory. 1L.-C.

Ku and S. A. Trugman, Phys. Rev. B 75, 014307 (2007). Fehske, G. Wellein, and A. R. Bishop, Phys. Rev. B 83, 075104 (2011). 3D. Golez, J. Bonca, L. Vidmar, and S. A. Trugman, Phys. Rev. Lett. 109, 236402 (2012). 4F. X. Morrissey, J. G. Mance, A. D. Van Pelt, and S. L. Dexheimer, J. Phys.: Condens. Matter 25, 144204 (2013). 5A. D. Van Pelt and S. L. Dexheimer, in Ultrafast Phenomena XII, edited by T. Elsaesser, S. Mukamel, M. M. Murnane, and N. F. Scherer, Springer Series in Chemical Sciences, (Springer-Verlag, New York, 2001), Vol. 66, p. 393. 6F. X. Morrissey and S. L. Dexheimer, J. Phys. Chem. B 116, 10582 (2012). 7J. T. Gammel, A. Saxena, I. Batistic, A. R. Bishop, and S. R. Phillpot, Phys. Rev. B 45, 6408 (1992). 8B. Scott, S. P. Love, G. S. Kanner, S. R. Johnson, M. P. Wilkerson, M. Berkey, B. I. Swanson, A. Saxena, X. Z. Huang, and A. R. Bishop, J. Mol. Struct. 356, 207 (1995). 9W. E. Buschmann, S. D. McGrane, and A. P. Shreve, J. Phys. Chem. A 107, 8198 (2003). 2H.

J. Chem. Phys. 142, 084309 (2015) 10S.

C. Huckett, B. Scott, S. P. Love, R. J. Donohoe, C. J. Burns, E. Garcia, T. Frankcom, and B. I. Swanson, Inorg. Chem. 32, 2137 (1993). 11B. I. Swanson, J. A. Brozik, S. P. Love, G. F. Strouse, A. P. Shreve, A. R. Bishop, W. Z. Wang, and M. I. Salkola, Phys. Rev. Lett. 82, 3288 (1999). 12M. Yamashita, N. Matsumoto, and H. Kida, Inorg. Chim. Acta 31, L381 (1978). 13Y. Wada, T. Mitani, M. Yamashita, and T. Koda, J. Phys. Soc. Jpn. 54, 3143 (1985). 14K. Iwano and K. Nasu, J. Phys. Soc. Jpn. 61, 1380 (1992). 15S. M. Weber-Milbrodt, J. T. Gammel, A. R. Bishop, and J. E. Y. Loh, Phys. Rev. B 45, 6435 (1992). 16I. Batistic and A. R. Bishop, Phys. Rev. B 45, 5282 (1992). 17I. Batistic, X. Z. Huang, A. R. Bishop, and A. Saxena, Phys. Rev. B 48, 6065 (1993). 18A. Saxena, Z. Shuai, J. T. Gammel, I. Batistic, M. Alouani, J. L. Bredas, and A. R. Bishop, Synth. Met. 71, 1659 (1995). 19F. X. Morrissey and S. L. Dexheimer, Phys. Rev. B 81, 094302 (2010). 20J. G. Mance, J. J. Felver, and S. L. Dexheimer, J. Phys. Chem. C 118, 11186 (2014). 21S. L. Dexheimer, A. D. Van Pelt, J. A. Brozik, and B. I. Swanson, Phys. Rev. Lett. 84, 4425 (2000). 22S. L. Dexheimer, A. D. Van Pelt, J. A. Brozik, and B. I. Swanson, J. Phys. Chem. A 104, 4308 (2000). 23L. Dhar, J. A. Rogers, and K. A. Nelson, Chem. Rev. 94, 157 (1994). 24D. Emin and T. Holstein, Phys. Rev. Lett. 36, 323 (1976). 25J. G. Mance, F. X. Morrissey, A. D. V. Pelt, and S. L. Dexheimer, in Ultrafast Phenomena XVII, edited by M. Chergui, D. M. Jonas, E. Riedle, R. W. Schoenlein, and A. J. Taylor (Oxford University Press, New York, 2011), p. 326. 26Y. Toyozawa, Prog. Theor. Phys. 26, 29 (1961). 27Y. Toyozawa, in Relaxation of Elementary Excitations, edited by R. Kubo and E. Hanamura (Springer-Verlag, New York, 1980), p. 3. 28D. W. Brown, K. Lindenberg, and B. J. West, J. Chem. Phys. 84, 1574 (1986). 29S. Tomimoto, H. Nansei, S. Saito, T. Suemoto, J. Takeda, and S. Kurita, Phys. Rev. Lett. 81, 417 (1998). 30S. Tomimoto, S. Saito, T. Suemoto, K. Sakata, J. Takeda, and S. Kurita, Phys. Rev. B 60, 7961 (1999). 31S. Tomimoto, S. Saito, T. Suemoto, J. Takeda, and S. Kurita, Phys. Rev. B 66, 155112 (2002). 32K. Yasukawa, Y. Takahashi, S. Kurita, and T. Suemoto, Solid State Commun. 140, 197 (2006). 33A. Sugita, T. Saito, H. Kano, M. Yamashita, and T. Kobayashi, Phys. Rev. Lett. 86, 2158 (2001). 34Y. Wada, N. Matsushita, and H. Haneda, J. Lumin. 108, 285 (2004). 35D. W. McCamant, P. Kukura, S. Yoon, and R. A. Mathies, Rev. Sci. Instrum. 75, 4971 (2004). 36F. X. Morrissey and S. L. Dexheimer, in Ultrafast Phenomena XV, edited by P. Corkum, D. Jonas, R. J. D. Miller, and A. M. Weiner, Springer Series in Chemical Sciences, (Springer-Verlag, New York, 2006), Vol. 88, p. 240. 37S. Takeuchi, S. Ruhman, T. Tsuneda, M. Chiba, T. Taketsugu, and T. Tahara, Science 322, 1073 (2008). 38M. Trigo, J. Chen, M. P. Jiang, W. L. Mao, S. C. Riggs, M. C. Shapiro, I. R. Fisher, and D. A. Reis, Phys. Rev. B 85, 081102 (2012). 39J. P. Kraack, A. Wand, T. Buckup, M. Motzkus, and S. Ruhman, Phys. Chem. Chem. Phys. 15, 14487 (2013). 40J. A. Brozik, B. L. Scott, and B. I. Swanson, Inorg. Chim. Acta 294, 275 (1999). 41W. T. Pollard, S. L. Dexheimer, Q. Wang, L. A. Peteanu, C. V. Shank, and R. A. Mathies, J. Phys. Chem. 96, 6147 (1992). 42S. P. Love, S. C. Huckett, L. A. Worl, T. M. Frankcom, S. A. Ekberg, and B. I. Swanson, Phys. Rev. B 47, 11107 (1993). 43A. Saxena, private communication (2014).

Journal of Chemical Physics is copyrighted by AIP Publishing LLC (AIP). Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. For more information, see http://publishing.aip.org/authors/rights-and-permissions.

Observation of structural relaxation during exciton self-trapping via excited-state resonant impulsive stimulated Raman spectroscopy.

We detect the change in vibrational frequency associated with the transition from a delocalized to a localized electronic state using femtosecond vibr...
286KB Sizes 2 Downloads 7 Views