Observation of breather excitons and soliton in a substituted polythiophene with a degenerate
ground state
Takayoshi Kobayashi,1,2,3,4Juan Du,1,2 Wei Feng,5Katsumi Yoshino,6Sergei Tretiak,7 Avadh Saxena,7 and Alan R. Bishop7
1Department of Applied Physics and Chemistry and Institute for Laser Science, University of Electro-Communications,
1-5-1 Chofugaoka, Chofu, Tokyo 182-8585, Japan
2JST, ICORP, Ultrashort Pulse Laser Project, , 4-1-8 Honcho, Kawaguchi, Saitama 332-0012, Japan 3Department of Electrophysics, National Chiao Tung University, 1001 Ta Hsueh Road, Hsin-Chu 3005, Taiwan
4Institute of Laser Engineering, Osaka University, 2-6 Yamada-Oka, Suita, Osaka 565-0871, Japan 5School of Materials Science and Engineering, Tianjin University, Tianjin 300072, People’s Republic of China
6Shimane Institute for Industrial Technology, 1 Hokuryo-cho, Matsue, Shimane 690-0816, Japan
7Theoretical Division and Center for Nonlinear Studies, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA
共Received 25 October 2009; revised manuscript received 22 December 2009; published 5 February 2010兲 We investigated the ultrafast dynamics of a unique polythiophene derivative that has a degenerate ground state due to an alternating benzenoid-quinoid resonance. We probed 128 different wavelengths at the same time by using a sub-5-fs pulse laser and a detection system composed of a polychromator and a multichannel lock-in amplifier. The method allowed us to study the electronic relaxation and vibrational dynamics in completely same conditions at the same time. Because the polythiophene derivative has degenerate ground state, solitons are expected to be generated after photoexcitation. The dynamics of a breather composed of a dynamic bound state of solitons generated immediately after photoexcitation was time-resolved to reveal coupling between the vibrational modes and the solitons. The CuC and CvC stretching modes were found to be modulated by the breather, whose lifetime was determined to be 30–50 fs. The results of quantum-chemical excited-state mo-lecular dynamics simulation are consistent with experimental results. Our modeling results allow to identify related vibrational normal modes strongly coupled to the electronic degrees of freedom. Moreover, analysis of calculated trajectories of excited state shows appearance of short-lived breather excitation decaying due to intramolecular vibrational energy equilibration on a time scale of hundreds of femtoseconds, which also agrees well with the experimental results.
DOI:10.1103/PhysRevB.81.075205 PACS number共s兲: 78.47.J⫺, 78.40.Me
I. INTRODUCTION
Unusual wave propagation, later termed a soliton, was first discovered in 1844.1 Solitons have been identified in many fields of nonlinear physics2–7 including water waves, sound waves, matter waves, and electromagnetic waves.8 Ac-cording to simulations performed using the Su-Schrieffer-Heeger 共SSH兲 Hamiltonian,9a photogenerated electron-hole 共e−h兲 pair evolves into a soliton-antisoliton pair 共SS¯兲 within 100 fs after photoexcitation because of barrier-free relaxation in a one-dimensional system. Matter-wave solitons have given rise to many interesting phenomena in the simplest conducting polymer, trans-polyacetylene,10 including anomalous conductivity and huge optical nonlinearity.11The formation times of solitons in polyacetylene have been de-termined to be ⬍150 fs.12 Even though the existence of a soliton in trans-polyacetylene is well known and has been extensively studied, there has been no systematic study of conjugated polymer systems besides trans-polyacetylene. This is probably because of the scarcity of polymers with a degenerate ground state.
It has also been theoretically predicted9,13–16that the soli-ton pair is spatially localized to form a dynamic bound state called a breather. The excess energy of the photogenerated 共e−h兲 pair over that of the soliton pair induces collective carbon-carbon 共CuC兲 oscillations, namely, the breather mode, due to electron-phonon coupling. Breathers predicted in Ref. 17 have been observed in trans-polyacetylene,18
which was found to have a period of 44 fs and an extremely short lifetime of⬃50 fs. However, there is not yet currently consensus among researchers as to whether breather is the primary photogenerated excitations and how they affect the ultrafast vibronic dynamics.18–22
In this paper, we report the study of the dynamics of soli-tons and breather in a quinoid-benzenoid polythiophene with a degenerate ground state which chemical structure is shown in Fig.1共a兲. It is well-known that polythiophene is one of the most promising materials for various device applications, so that a detailed understanding of the dynamics of electronic state and vibrational dynamics photoexcitations in them and their derivatives is highly desirable for practical reasons. The polymer system studied in the present paper appears osten-sibly not to have a degenerate ground state, however, it has degeneracy due to the tautomerization of the inner structure of the polymer, as described below. And its repeat unit is much larger than that of trans-polyacetylene. The repeat unit in the latter is composed of one single bond and one double bond, while the one in the former is more bulky since it has two thiophene rings with both cis and trans configurations. Since there has been no other spectroscopic study of solitons except for trans-polyacetylene before, it is of great interest to study its dynamics and compare any differences between it and trans-polyacetylene in terms of the repeat unit structure. To the best of our knowledge, this is the first observation of the existence of a breathe and solitons and their dynamics in a system other than trans-polyacetylene. To rationalize the
experimental data, we performed a quantum-chemical excited-state molecular dynamics simulation, and its results are consistent with experimental data, enabling breather ex-citations to be analyzed and related coupled vibrational nor-mal modes to be identified.
II. EXPERIMENTAL DESCRIPTION
In the present experiment, we utilized a nearly Fourier-transform limited visible-near-IR pulse generated from a noncollinear optical parametric amplifier 共NOPA兲 seeded by a white-light continuum, as shown in Fig. 2. The pump source of this NOPA system was a regenerative amplifier 共Spectra Physics, Spitfire兲 with the following operating pa-rameters: central wavelength, 800 nm; pulse duration, 50 fs; repetition rate, 5 kHz; average output power, 650 mW. We used a 1-mm-thick sapphire plate to generate the continuum
spectrum, and great care was taken to introduce only the formation of a single filament. The NOPA output pulse was compressed with a pair of chirp mirrors and then with a prism pair, resulting in a nearly Fourier transform共FT兲 lim-ited pulse duration of 6.3 fs. Both the pump and probe pulses covered the spectral range extending from 515 to 716 nm,23 and the energies of them are about 50 and 6 nJ, respectively. The pump-probe signal at 128 different wavelengths was de-tected by a combined system of a polychromator and a mul-tichannel lock-in amplifier. Thanks to the extreme stability of the light source and noise reduction by the lock-in-detector, the time resolution is better than 1fs, which was recognized by the difference between the time-resolved spectra at neigh-boring delay step obtained using the time step of 0.2 fs.
The sample studied in this study was a thin film of quinoid-benzenoid polythiophene, poly关3-hexylthiophene-2,5-diyl兴-关p-dimethylaminobenzylidenequinoidmethene兴 共P-HTDMABQ兲, which was dissolved in methanol and cast on a quartz substrate for stationary and time-resolved spectra measurements. All experiments were performed at room temperature 共293⫾1 K兲.
III. MOLECULE STRUCTURE
The structure of PHTDMABQ is depicted in Fig. 1共a兲, whose monomer is a derivative of thiophene. It appears os-tensibly not to have a degenerate ground state. There is no degenerate ground state in polythiophene because quinoid thiophene has a higher energy thanbenzenoid thiophene. Therefore, it has a benzenoid structured ground state. After photoexcitation, this benzenoid structure is converted into a quinoid structure, but this excitation with the latter structure cannot propagate freely along the polymer chain, unlike soli-tons in trans-polyacetylene. Thus, the absence of solisoli-tons in polythiophenes is due to the energy difference between ben-zenoid thiophene and quinoid thiophene. Bipolarons are gen-erated by photoexcitation. The two polarons in the bipo-larons cannot be separated from each other because they are nondegenerate. However, in the case of PHTDMABQ, there are two tautomers, benzenoid and quinoid. Both the ben-zenoid thiophene and quinoid thiophene are next to each other, to form a repeat unit and they can exchange their tau-tomeric structures without energy requirement. Therefore, the ground state can be either –共 quinoid-benzenoid兲–n or –共 benzenoid-quinoid兲–nas shown in Fig.1共a兲. Because of the tautomerization of the internal structure of PHTDMABQ, it can have a degenerate ground state and thus solitons can be generated in it.
IV. QUANTUM-CHEMICAL METHODOLOGY
To analyze the experimental data, we used the Austin Model 1 共AM1兲 Hamiltonian and an excited-state molecular dynamics 共ESMD兲 computational package, which is de-scribed in detail in Refs.24and25, to follow photoexcitation adiabatic dynamics on a picosecond time scale for all lations presented in this study. The ESMD approach calcu-lates the excited-state potential energy as Ee共q兲=Eg共q兲 +⍀共q兲 in the space of nuclear coordinates q that span the
S S C6H13 C6H13 N N S S C6H13 C6H13 N N (b) (a)
FIG. 1. 共Color online兲 Two quantum-mechanical resonant struc-tures of共a兲 PHTDMABQ and 共b兲 the excited-state molecular dy-namics simulations results.共b兲 shows the normalized Fourier spec-tra of the lowest dipolar allowed excited-state spec-transition energy⍀共t兲 and its respective oscillator strength f共t兲 trajectories 共top two plots兲, and the amplitudes of dimensionless displacements⌬ 共stick spec-trum, bottom panel兲 along normal modes calculated in the oligomer with 10 repeat units, as shown in the inset. The three molecular structures at the top schematically show vibrational normal modes with frequencies strongly coupled to the electronic system, which lead to the formation of the breather excitation. These correspond to the CvC vibration and CuC stretches as schematically shown in the middle panel.
entire共3N−6兲 dimensional space, where N is the total num-ber of atoms in the molecule. Here, ⍀共q兲 is the electronic transition frequency to the lowest 1Bu共band-gap兲 state of the photoexcited molecule. The program efficiently calculates analytical derivatives of Ee共q兲 with respect to each nuclear coordinate qi to evaluate forces and to subsequently step along the excited-state hypersurface using these gradients. All computations start from vertical excitation at the optimal ground-state molecular geometry. The total molecular energy
Ee共q兲 is conserved if no dissipative processes are included. Subsequent analysis of the photoexcited trajectories of the excitation energy ⍀共q,t兲 and oscillator strength f共q,t兲 in Fourier space allows us to identify periods of participating vibrational motions. Alternatively, the minimum of the excited-state potential energy surface can be calculated by including an artificial dissipative force in the equations of motion corresponding to the relaxed excited-state geometry. To understand the formation of photoexcited breathers, we calculated the dynamics of the band-gap excited state in the 10-unit thiophene oligomer shown in the inset of Fig. 1共b兲, where the alkyl side chain have been replaced by hydrohens, effectively reducing the molecular size for calculations. This molecule is sufficiently long共10 nm兲 compared to the char-acteristic exciton size of about 2 nm for the infinite chain limit to be valid.
V. RESULTS AND DISCUSSION A. Vibrational real-time spectra and analysis
As described in the Experimental section, the stability of our NOPA laser and 128 channel-lock-in amplifiers helps us in very detailed analysis of dynamics of both electronic re-laxation and real-time molecular vibration. In order to obtain a complete view of excited-state evolution triggered by the ultrashort pulse, it is of vital importance to measure the dy-namics at different probe wavelength to obtain most reliable
information by the experiment to be performed. High sensi-tivity can be achieved by using a lock-in amplifier based on a phase-sensitive detection scheme, but in traditional experi-ment it is performed by single wavelength measureexperi-ment. Spectroscopic measurement can be performed by using a lin-ear array detector or two-dimensional detector such as CCD camera. However, in this method is not based on the phase-sensitive detection scheme, the data accumulate associated background noise and the data acquisition time is limited by the overflow of noise accumulated. The method of the present paper is not suffered from this problem.
The method presented in this paper is based on the modu-lation of transmitted light intensity induced by molecular vi-bration. This modulation can be classified into three mecha-nisms. One is the transition intensity change in probing wavelength due to “vibronic coupling.” In this case the amount of the transmission intensity change is proportional to the coupling strength and vibrational amplitude. The sec-ond one is the amplification and deamplification of the probe light induced by the stimulated Raman process. The third one is due to the modulation of the refractive index associated with molecular vibration resulting in the spectral shift pro-portional to the change in the refractive index and hence to the change in the phase velocity.
The transition probability change due to the “vibronic coupling” mechanism above mentioned is classified again into two subgroups.26The first subgroup is Condon-type in which the change in the transmitted light intensity to be ob-served is proportional to the first derivative of the absorption 共or emission兲 spectrum depending on the relative configura-tion of the potential curves of the relevant initial and final sates of absorption共or emission兲. The first subgroup is again into potential-minimum-shift type and potential-curvature-change type. The probe wavelength dependence of ampli-tudes of the Fourier transform of the former and the latter are obtained by the first and second derivative of the relevant electronic transition spectra, respectively. The second sub
Periscope BS
Mirror
Lens BBO(SHG) Filter
FS Prism Lens BBO(NOPA) Sapphire Lens Stage Iris
Concave mirror Chirped mirror Chirped mirror
Periscope Iris Mirror Prism Prism Mirror Mirror Iris BS Sample Iri Mirror VND Lens Ti:sapphire laser 140J,800nm,5kHz PMT Iris Mirror Mirror Concave Mirror Mirror Mirror Concave Mirror Shortpass filter Concave mirror Mirror VND Mirror Mirror Mirror Iris Lens Mirror Mirror Mirror Beam splitter Stage Iri Mirror Concave mirror Mirror Iris Iris Iris Fiber Polychromator + Multi-channel lock-in Fiber Concave mirror
FIG. 2.共Color online兲 Block diagram of the apparatus of vibration real-time spectroscopy based on a pump-probe scheme. VND: variable neutral-density filter; FS: fused silica; BS: beam splitter.
group is non-Condon type, in which the vibration modulate the transition dipole moment and hence the transmitted light intensity change is proportional to the zero-th derivative of the absorption 共or emission兲 spectrum.
The method can provide the information of the change in intensity of absorption 共or emission兲 as a function of probe frequency ⌬A共兲 from which the electronic relaxation is studied. The signal shows the decay dynamics and spectral change associated with the change in the electronic state. On top of that signal the modulation␦⌬A共兲 of the ⌬A共兲 due to molecular vibration can be used to study the vibrational dynamics of the system completely in the same condition as that electronic dynamics study. This situation is difficult to be realized in experiment made by using conventional UV 共VIS兲 pump-UV 共VIS兲 probe experiment and time-resolved vibrational experiment. In this paper the decay dynamics of the electronic state and vibrational dynamics highly corre-lated through excitonic coupling 共vibronic coupling兲 were observed under the same condition at the same time.
B. Electronic relaxation and molecular vibration dynamics
The difference absorbance 共⌬A兲 signals in Fig. 3共a兲 ex-hibit oscillation due to molecular vibrations. As shown in the ⌬A traces, the lifetimes of the electronic states consist of three components: 62⫾2 fs, 750⫾20 fs, and ⬎3 ps.27The
Fourier power spectra in Fig. 4 have peaks at 1111⫾7,
1343⫾7, and 1465⫾7 cm−1共P1, P
2, and P3, respectively兲. Theoretical calculations allow assigning these peaks to CuC stretching modes with different bond orders. Figure
1共b兲 shows the calculated dimensionless displacements ⌬ along the vibrational coordinates of the optimal geometries between the ground and excited states. This immediately en-ables us to identify the P1− P3vibrational normal modes that are strongly coupled to the electronic excitation. These cor-respond to intra- and interthiophene ring CuC and CvC stretching motions 关see Fig.1共b兲, top structures兴. The high-est, medium, and lowest frequencies are considered to corre-spond to CvC double bonds, a mixture of double and single bonds, and single bonds, respectively. The calculated vibra-tional frequencies 共1345, 1533, and 1675 cm−1兲 are overes-timated by about 200 cm−1 compared to the experimental values, which is typical for semiempirical calculations.
C. Dynamics of breather and soliton
The requirement for the existence of solitons is a degen-erate ground-state structure, and as shown in Fig. 1共a兲, our quinoid-benzenoid polythiophene sample has a degenerate ground state because the two quantum-mechanical resonance structures have equivalent energies. In Fig.3共b兲, the transient absorption spectra exhibit negative absorbance changes in the photon energy range 1.91 to 2.38 eV, while for photon energies smaller than 1.91 eV, the absorbance change is posi-tive. This increase in the absorbance is attributed to the tail of the solitons not fully relaxing to the band-gap center, as is the case in trans-polyacetylene.18Therefore, the induced ab-sorption observed in poly-共substituted thiophene兲 is attrib-uted to solitons. The positive value indicates the increased contribution of induced absorption due to a soliton with a peak near the mid gap, which is estimated to be around 1.4 eV.
Spectrogram is an analysis method of the time-dependent spectrum where the instantaneous frequency changes with time. It is suited to study the dynamic process where the (b)
(a)
FIG. 3. 共Color online兲 共a兲 Two-dimensional real-time absor-bance change. One example of the time-dependent absorabsor-bance change at 630 nm was plotted as the top figure. 共b兲 The time-resolved pump-probe spectra probed at 10 center delay time points from 100 to 1700 fs with an integration time width of 200 fs.
FIG. 4. 共Color online兲 Fourier-transform power spectra. The Fourier-transform power spectrum at 615 nm is plotted as an example.
molecular geometrical relaxation or chemical reaction is ac-companied with change in its vibrational frequency due to molecular structural change during the processes. Figure 5
shows the contour map obtained by the spectrogram analysis,28 it has three prominent peaks corresponding to peaks P1, P2, and P3. In addition to these three modes, five more peaks appear as side bands of P1− P3at 270, 500, 640, 1960, and 2200 cm−1. These five modes are breather modes which are not visible in the two-dimensional Fourier power spectrum共Fig.4兲 or in the stationary resonance Raman
spec-trum because they have extremely broad widths due to their short lifetimes. In other words, the breather cannot appear in Raman spectra because it is a short-lived transient vibration mode, and hence it is not a normal mode or eigenmode. We also note that there are no detectable normal modes with substantial intensity in this spectral region coupled to the electronic excitation as evidenced by lack of significant dis-placements calculated in Fig. 1共b兲. The amplitudes of the high-frequency modes in the 2000– 2200 cm−1 range de-crease rapidly due to their short vibrational periods, which cannot be properly resolved by the finite pulse widths of both pump and probe pulses. However, the frequency is not af-fected and the breather excitation is clearly visible for the three sidebands with the lower frequencies.
The average frequency difference between the three main bands 共1111, 1343, and 1465 cm−1 and their corresponding side bands 共270 and 1960 cm−1 for 1111 cm−1, 500, and 2200 cm−1for 1343 cm−1, and 640 cm−1for 1465 cm−1兲 is calculated to be 843⫾12 cm−1, which is close to the experi-mentally observed value of about 760 cm−1 reported for polyacetylene,18 and theoretically expected values of 660– 1000 cm−1.14–17This frequency separation between the main bands and the satellite bands indicates that the breather modulates the CuC stretching modes with a period of ⬃40 fs generating the sidebands. The reason why the mode with frequency of 1465 cm−1has only one lower sideband is because of the reduced intensity of upper sideband expected to be located around 2308 cm−1 corresponding to 14.4 fs close to the convoluted duration共8.9 fs兲 of the pulse duration time of both pump and probe pulses. Because of the limited
time resolution, the upper band signal could not be observed. The lifetimes of the side bands were determined to be about 50, 32, and 40 fs for sidebands of P1, P2, and P3, respectively. They are corresponding to the life time of the breather mode as observed decay time of 50 fs in polyacetylene.18The theoretically predicted decay time is shorter than 100 fs is again consistent with our observation. The result means the time for the breather mode to disappear followed by the separation into isolated solitons, and even shorter than that in polyacetylene. These short lifetimes are also corresponding to the shortest lifetime component of ⬃62 fs in the ⌬A trace.27The ultrafast relaxation of⬃50 fs of this nonlinear excitation ensures an ultrafast nonlinear re-sponse that can be used in all-optical switches.
Consequently, these sideband peaks appear due to nonlin-ear electron-vibrational excited-state dynamics. To analyze these processes from theoretical calculations, we compute the power spectra of 750 fs photoexcited trajectories of the excitation energy ⍀共q,t兲 and oscillator strength f共q,t兲 shown in Fig.1共b兲.24,29Both plots show an additional broad peak centered around 50 fs, corresponding to a vibration in the range 550 to 800 cm−1, which does not correspond to any of the vibrational normal modes that exhibit substantial coupling to the electronic degrees of freedom 关compare the displacements peaks with the fast Fourier transform 共FFT兲 trajectories in Fig. 1共b兲兴. Based on our previous computa-tional studies, we assign this peak to a nonlinear breather excitation that occurs due to coupling of CuC vibrational motions. This vibrational excitation decays gradually over long time scales due to dissipation of vibrational energy to internal vibrational degrees of freedom that are weakly coupled to the electronic system; this is similar to previous findings by us.24,29Power spectra of longer excited-state tra-jectories共not shown兲 demonstrate diminishing breather peak. Our calculations estimate the breather lifetime to be about 0.2 ps without accounting for intermolecular dissipation channels共baths兲. Such fast decay is considered to be reason-ably in agreement with our experiment data of ⬃50 fs, in case we take into accounts of both intermolecular interac-tions and phonon energy leakage through the boundaries of conjugated segments共defects兲. These processes may reduce breather lifetime from 0.2 ps to ⬃50 fs.29As expected, the breather peak is more pronounced in the power spectrum of the oscillator strength共see Fig.5兲, since it is directly related
to the modulation of the respective transition dipole moment.29
In the previous studied for breather in polyacetylene,18we only measured the real-time traces at only four different probe wavelengths共620, 720, 740, and 750 nm兲. However, it was very difficult to study the probe wavelength dependence of the modes due to modulation because of the small number of probe wavelengths 共i.e., probe photon energies兲. In the present study, thanks to the multichannel detection system developed, we could observe the molecular vibration in-duced modulation␦⌬A of the absorbance change ⌬A at 128 different wavelengths simultaneously. In this way, the probe wavelength dependence can be used in detailed discussions on the photon energy dependence of modulation amplitudes of various modes. As shown in Fig. 6, the breather strongly contributes to the variations in the oscillator strength. Except FIG. 5. Contour maps of the two-dimensional Fourier power of
the vibrational components obtained by spectrogram calculation for real-time data covering 680–690 nm.
for the modes at 1960 and 2200 cm−1, the signal sizes of all the modes exhibit an almost monotonic increase when the photon energy is reduced from 1.82 to 1.73 eV. This is con-sistent with the theoretical discussion in Ref. 24. Similar feature is found for the positive absorbance change, as shown in Fig. 3共b兲. The magnitude of the positive absorbance change increases when the probe photon energy decreases. This provides the evidence of the larger contribution of the breather to the modulation 共␦⌬A兲 of the difference absor-bance change共⌬A兲 due to soliton in the lower energy range. We can interpret this feature in terms of the modulation of the transient spectrum of soliton by the vibrational modes as follows.
Molecular vibration associated with the breather and soli-ton is expected to modulate the transition energy and transi-tion probability, and the amplitudes are expected to be pro-portional to the zeroth, first, and second derivatives of the absorption and/or stimulated emission spectrum depending on the mechanism of induction of the wavepacket motions.30 Since the absorption spectrum of soliton is expected to be close to the mid gap, which is located at 1.4 eV in the case of polyacetylene, the spectral range of the present observation is higher energy tail of the soliton absorption as seen from the time-resolved spectrum as shown in Fig.3共b兲. Therefore, all of the zeroth, first, and second derivatives of the absorp-tion spectrum are expected to increase monotonically with decreasing probe photon energy.
Here, we go back to the discussion of the above men-tioned exceptional behavior of the probe wavelength depen-dence of the amplitudes of the two modes at 1960 and 2200 cm−1. It can be explained in the following way. The molecular vibration modulates the transition probability and transition energy because of the electronic distribution change nearly instantaneously following the motion of nuclei
during the molecular vibration. The frequencies of the modes at 1960 and 2200 cm−1are nearly equal to the overtones of 1111 cm−1, which may affect the dynamics of the breather. The electronic transition is modulated at periodically with period of ⬃30, 17, and 15 fs corresponding to the frequen-cies of 1111, 1960, and 2200 cm−1, respectively. Then, at integer multiple of about 32 fs, all of them contribute. In case the vibrations of the modes are in-phase or out-of-phase then the amplitudes of them may be affected by either construc-tive or destrucconstruc-tive interference.
We now discuss the effect of the differences in the sizes and structures of the repeat units. The repeat unit in the
trans-polyacetylene is composed of one single bond and one
double bond, while in PHTDMABQ, it is composed of two thiophene rings, which show antiphase tautomerism共see Fig.
1兲, which produces the degenerate ground state in the
poly-mer. Since PHTDMABQ has a cis configuration in the thiophene ring, it is interesting to compare it with cis-polyacetylene. The lifetime of breather is expected to be longer than that in trans-polyacetylene. However, the decay time of the breather seems to be even shorter than that in
trans-polyacetylene. That is because the lifetime is not
deter-mined by the separation of soliton pairs from the originally generated site in the polymer chain, but by the energy dissi-pation to internal vibrational freedom. Since quinoid-benzenoid polythiophene has many more internal vibrational modes than trans-polyacetylene, its breather lifetime is even shorter than that of trans-polyacetylene.
VI. CONCLUSIONS
In summary, using a polythiophene derivative with a de-generate ground state, we investigated the ultrafast dynamics that take place immediately after excitation. The simulation results of quantum-chemical excited-state molecular dynam-ics agree reasonably well with the experimental data by showing formation of short-lived breather excitation. The breather lifetime was experimentally determined from the electronic spectral dynamics to be ⬃62 fs, which agrees with the time constants determined by the time-dependent signal intensity that appears as side peaks associated with the breather. Even though extensive theoretical studies have been conducted, to the best of our knowledge, this is the first time that modulation of the CuC single and double stretch-ing modes by the breather have been experimentally ob-served other than in trans-polyacetylene.
ACKNOWLEDGMENTS
This work was partly supported by a grant from the Min-istry of Education共MOE兲 in Taiwan under the ATU Program at National Chiao Tung University. A part of this work was performed under the joint research project of the Institute of Laser Engineering, Osaka University under Contract No. B1-27.
FIG. 6. 共Color online兲 The probe photon energy dependence of the vibrational amplitude probed at a 110 fs gate delay time with a gate width of 120 fs 共half-width at half maximum兲 in the spectrogram.
1J. S. Russell, in Report of the Fourteenth Meeting of the British
Association for the Advancement of Science共Murray, London, 1844兲, pp. 311-390.
2N. Manton and P. Sutcliffe, Topological Solitons 共Cambridge
University Press, Cambridge, England, 2004兲.
3L. Khaykovich, F. Schreck, G. Ferrari, T. Bourdel, J. Cubizolles,
L. D. Carr, Y. Castin, and C. Salomon, Science 296, 1290 共2002兲.
4K. E. Strecker, G. B. Partridge, A. G. Truscott, and R. G. Hulet,
Nature共London兲 417, 150 共2002兲.
5M. Nakazawa, E. Yamada, and H. Kubota, Phys. Rev. Lett. 66,
2625共1991兲.
6W. P. Su, J. R. Schrieffer, and A. J. Heeger, Phys. Rev. Lett. 42,
1698共1979兲.
7A. M. Kosevich, B. A. Ivanov, and A. S. Kovalev, Phys. Rep. 194, 117共1990兲.
8G. I. Stegeman and M. Segev, Science 286, 1518共1999兲. 9W. P. Su and J. R. Schrieffer, Proc. Natl. Acad. Sci. U.S.A. 77,
5626共1980兲.
10H. Shirakawa, Rev. Mod. Phys. 73, 713共2001兲.
11H. Naarmann and N. Theophilou, Synth. Met. 22, 1共1987兲. 12C. V. Shank, R. Yen, R. L. Fork, J. Orenstein, and G. L. Baker,
Phys. Rev. Lett. 49, 1660共1982兲.
13D. K. Campbell, Nature共London兲 432, 455 共2004兲.
14M. Sasai and H. Fukutome, Prog. Theor. Phys. 79, 61共1988兲. 15S. R. Phillpot, A. R. Bishop, and B. Horovitz, Phys. Rev. B 40,
1839共1989兲.
16S. Block and H. W. Streitwolf, J. Phys.: Condens. Matter 8, 889
共1996兲.
17A. R. Bishop, D. K. Campbell, P. S. Lomdahl, B. Horovitz, and
S. R. Phillpot, Phys. Rev. Lett. 52, 671共1984兲.
18S. Adachi, V. M. Kobryanskii, and T. Kobayashi, Phys. Rev.
Lett. 89, 027401共2002兲.
19G. S. Kanner, Z. V. Vardeny, G. Lanzani, and L. X. Zheng,
Synth. Met. 116, 71共2001兲.
20K. M. Gaab and C. J. Bardeen, J. Phys. Chem. B 108, 4619
共2004兲.
21G. Lanzani, G. Cerullo, C. Brabec, and N. S. Sariciftci, Phys.
Rev. Lett. 90, 047402共2003兲.
22R. Lécuiller, J. Berréhar, C. Lapersonne-Meyer, and M. Schott,
Phys. Rev. Lett. 80, 4068共1998兲.
23A. Baltuška, T. Fuji, and T. Kobayashi, Opt. Lett. 27, 306
共2002兲.
24S. Tretiak, A. Saxena, R. L. Martin, and A. R. Bishop, Proc.
Natl. Acad. Sci. U.S.A. 100, 2185共2003兲.
25S. Tretiak, A. Saxena, R. L. Martin, and A. R. Bishop, Phys.
Rev. Lett. 89, 097402共2002兲.
26T. Kobayashi and Z. Wang, IEEE J. Quantum Electron. 44, 1232
共2008兲.
27J. Du, Z. Wang, W. Feng, K. Yoshino, and T. Kobayashi, Phys.
Rev. B 77, 195205共2008兲.
28M. J. J. Vrakking, D. M. Villeneuve, and A. Stolow, Phys. Rev.
A 54, R37共1996兲.
29S. Tretiak, A. Piryatinski, A. Saxena, R. L. Martin, and A. R.
Bishop, Phys. Rev. B 70, 233203共2004兲.
30T. Kobayashi, Z. Wang, and I. Iwakura, New J. Phys. 10,
