Open Access
Add to BibSonomyAbstract
A train of highly-stable, high-beam-quality ultrashort pulses is successfully produced by synthesizing phase-coherent rotational-Raman-sidebands in parahydrogen. The intensity-waveform of this ultrashort-pulse-train is directly evaluated in time domain based on a sum-frequency-generation autocorrelation-technique. It is shown that a 10.6-THz ultrahigh-repetition-train of short pulses is formed with an effective-duration of 20 fs and a high peak-power of 2 MW.
©2005 Optical Society of America
1. Introduction
It is possible to drive a Raman process adiabatically by controlling the frequencies of excitation laser-fields in terms of a small two-photon detuning to the Raman resonance [1]. This adiabatic Raman process, at the same time, collinearly generates broad, phase-coherent sidebands with a wide, equidistant frequency-spacing [1–6]. It has been shown recently that a train of monocycle pulse may be generated through the Fourier-synthesis of such Raman sidebands, where the vibrational sidebands in deuterium were utilized [6]. A light source produced in this way has attractive potentials such as an ultrahigh-pulse-repetition-rate and wide tunability regarding the carrier frequency. The establishment as an actual light source will open many attractive applications.
In this paper we demonstrate the generation of an ultrahigh-repetition-rate train of short pulses by synthesizing phase-coherent rotational sidebands in parahydrogen. This paper focus on the following. First, a highly-stable, high-beam-quality Raman-sidebands is produced through an adiabatic process by employing a newly-developed dual-wavelength, injection-locked, pulsed Ti:sapphire laser [7]. Second, based on a measurement of the intensity-waveform in the time domain, it is demonstrated that the Fourier-synthesis of such Raman sidebands can produce a stable, ultrahigh-repetition-rate train of short pulses.
There has been a number of related studies regarding this type of light source. Yoshikawa and Imasaka suggested the application of multicomponent rotational Raman spectra to femtosecond-time-scale pulses [8, 9]. Yavuz et al. demonstrated a rotational sideband generation in deuterium with near-unity conversion efficiency [10]. Yavuz et al. produced a comb of about 200 spectral components using two molecular species, synthesizing a quasiperiodic waveform with pulse envelopes about three cycles long [11]. Katsuragawa et al. produced more than 100 spectral components using three correlated Raman-coherences in a molecular system [12]. Wittmann et al. produced a sequence of compressed femtosecond pulses using an impulsive Raman approach [13]. Sokolov et al. showed near-single cycle-pulse structure using Raman sidebands more than an octave of optical bandwidth [14].
2. Experimental
The experimental setup is illustrated in Fig. 1. The Raman sidebands were generated by driving the v = 0, J =0 → v = 0, J = 2 rotational transition in gaseous parahydrogen. The parahydrogen was produced in the liquid phase at a temperature just above the triple point (13.8 K) by converting normal hydrogen using a catalyst (oxidized-iron-powder). The purity of the parahydrogen was estimated to be greater than 99.9%. After the conversion process, the parahydrogen was transferred in a 13-cm-long sample-cell made of copper that was installed in a liquid-nitrogen-cryostat. The density of the parahydrogen was set to 2.5 x 1020 cm-3, giving a density-interaction-length-product of 3.3 x 1021 cm-2. The temperature was adjusted to ∼100 K. This temperature gives the narrowest Raman-transition-width of 300 MHz (HWHM) for this density, in addition to having a sufficient population-probability of 97% in the ground state [15].
The coherent rotation was driven by the dual-wavelength, injection-locked, pulsed Ti:sapphire laser. The key performance of this laser system is to emit two-wavelength, Fourier-transform-limited, pulses from a single resonator. This enables the perfect overlaps of the dual nanosecond long-pulses in both time (6 ns at FWHM) and space automatically, leading to the highly-stable generation of Raman-sidebands.
The wavelengths of the two driving-fields, Ω0 and Ω-1, were 763.180 and 784.393 nm, respectively, the difference frequency of which was slightly detuned by 600 MHz from the exact Raman-resonance (10.631 THz; see the inset in Fig. 1) to satisfy the adiabatic conditions [1]. The wavelengths were monitored by a wavemeter (Burleigh, 1500 WA). The driving-pulses were loosely focused into the parahydrogen with f = 1000 mm lens. The focus size at the beam-waist was 120 μm in diameter; and the Rayleigh-length was 10 cm. The typical incident energy of the driving pulse was 1.5 mJ each, corresponding to a peak excitation-intensity of 0.5 GW/cm2.
The generated sidebands were collimated with a lens of f = 300 mm, and then guided to a chirp-variable-device to compensate for the group-velocity-dispersion (GVD) across the sidebands. The GVD of the chirp-variable-device is - 41 fs2 at each reflection with a working spectral-region of 700 - 900 nm. The chirp-variable-device is specifically designed such that the negative-chirp can be widely changed by setting the reflection number from between 1 and 45 times [17]. This process is achieved without any deviation of the output optical-axis by moving a small mirror, as shown in Fig. 1. After appropriate GVD compensations, the sidebands were introduced to the background-free, sum-frequency-generation (SFG) autocorrelator, in which a thin β-barium-borate crystal of 20-μm-thickness was used in order to measure the sum-frequency signal without distorting the spectral-phase of the sidebands. The sum-frequency was detected by a photo diode. The signal was integrated over the 6ns-driving-pulse and digitized using a Boxcar integrator and an AD-converter, before being processed by a computer. It was checked that the autocorrelator had a potential to measure a pulse of down to 10 fs. We evaluated the temporal intensity-profile both quantitatively and systematically based on this well-established technique.
Fig. 1. Schematic diagram of the experimental setup. LN2-Cryostat: liquid-nitrogen-cryostat, SFG-Autocorrelator: sum-frequency-generation-autocorrelator, BBO: β-barium-borate crystal, SWP: shortwave-pass filter.
3. Results
All of the sidebands were generated collinearly without being restricted by the phase-matching condition. Figures 2(a) and 2(b) show the beam pattern of the output beam and that of each sideband after dispersion with a prism, respectively, taken by a digital camera after the collimating lens. 11-sidebands including the two driving-fields, Ω0 and Ω-1, are clearly seen with nearly the same Gaussian profile. The sufficiently-high Raman-coherence driven through the adiabatic process allows for this collinear-broad-sideband-generation [1–6]. As can be clearly seen in Figs. 2(a) and 2(b), each sideband was generated with an extremely-high beam-quality. Furthermore this sideband generation was very stable. The shot-by-shot fluctuation for the principal sidebands was typically less than 5%. The high beam-quality and stability observed here was possible because of the perfect temporal and spatial overlaps of the two driving-pulses of the dual-wavelength, injection-locked laser. Figure 2(c) shows the corresponding spectrum at the peak of the driving-fields, as measured by an optical-multichannel analyzer (OMA; Oriel, MS257 & Andor, DU420-OE). The relative wavelength-sensitivity of the detection system was calibrated using a standard light source (ORIEL Instruments, Model 68931). In this sensitivity-scale 7-sidebands were confirmed in the spectral region of 706 - 831 nm, with the equidistant frequency-spacing of 10.6 THz. The pump intensity of 0.5 GW/cm2 in the present experiment was selected such that the substantial sidebands were covered in the working spectral-region (700 ∼ 900 nm) of the chirp-variable-device, as confirmed here.
Fig. 2. Transverse-mode patterns of a: the output beam after the collimating lens and b: the generated sidebands. c: The relative output energy of each generated sideband at the peak of the driving-fields. Ω0 and Ω-1 indicate the two driving-fields.
Figure 3 shows the autocorrelation traces of the sidebands measured by the SFG autocorrelator for four different reflection-numbers, a: 9 times, b: 35 times, c: 37 times, and d: 45 times. A unique autocorrelation profile was observed for each fixed reflection-number. This is another manifestation of the Raman process obeying good adiabaticity, i.e., the relative phases of the sidebands were reasonably fixed over both the nano-second pulse-envelope and the whole beam-cross-section. The autocorrelation trace gradually became sharper as the reflection number increased. At 37 reflections, the trace was the sharpest and proceeded to broaden as the reflection number was further increased. The 37-reflection trace corresponds to a negative-chirp of 1520 fs2 (= 41 fs2 x 37). Since the total positive-dispersion of the intervening optics is estimated to be 700 fs2, it is reduced that the sidebands are generated with a GVD of 820 fs2 under the present generation process. This dispersion can be attributed to the sum of the GVDs from the refractive index of parahydrogen, the Gouy phase of the sideband beams, and the sideband-generation-process itself.
Fig. 3. Autocorrelation traces at a: 9, b: 35, c: 37, d: 45 reflections on the chirp variable-device. The sharp noise peaks seen in the observed traces are artifacts.
Figure 4 is, again, the autocorrelation trace for 37 reflections. The green line is the best-fit trace (3 fs at FWHM) for the observed data reduced from numerical-trials on the temporal-intensity-waveform. The resulting temporal waveform was similar to a sech-function, which is shown with the blue curve. It can be clearly seen that the reduced temporal waveform constitutes a train of short pulses with a high peak-pedestal-contrast. The effective pulse-duration was estimated to be 2 fs. The pedestal due to the overlap of the tail of each pulse was 3.7% of the pulse-peak. Through this pulse compression process, the produced train of short pulses reached a high peak-power of ∼ 2 MW. The measured period of the pulse-train was 94 fs, which was shown to be equivalent to a rotational-frequency of 10.6 THz. It should be emphasized that the observed trace is not averaged. This is direct evidence that a highly stable, ultrahigh-repetition-rate train of short pulses is realized here [17].
Fig. 4. The autocorrelation trace for 37 reflections. The green and blue curves show the best-fit to the observed trace and the corresponding temporal waveform, respectively.
4. Discussions
The Raman coherence follows the nanosecond-envelope (6 ns FWHM) of the driving pulses adiabatically in the present system. As a result the produced sideband spectrum and the waveform of the pulse-train on the femto-scale also stationally follow this pulse-envelope. Figure 5(a) shows the observed temporal profile for each rotational sideband. As seen in Fig. 5(a), the Raman coherence reaches a maximum around the peak of the driving pulses (t = 0), producing the multisidebands (Ω-3 - Ω3), the spectrum of which is shown in Fig. 2(c). On the other hand, at the initial and final stages of the adiabatic process, i. e., in the outer regions over ± 4 ns, only the driving fields (Ω0, Ω-1) exist. Since the actual autocorrelation trace in Fig. 4 was measured by integrating over this nanosecond-pulse-envelope, the traces and the reduced temporal waveforms are understood as averaged traces and waveforms over the nanosecond-pulse, respectively. In order to know the details, we calculated the local temporal-waveform from Fig. 5(a), assuming the transform-limit. At t = 0, the reconstructed temporal-waveform had a width of 17 fs at FWHM, shown with the solid blue-curve in Fig. 5(b). This waveform was also nearly steady with a duration of 1 fs within a 3-ns region around t = 0 (the shadow area in Fig. 5(a)). The corresponding local-autocorrelation-trace is shown with the green-dotted-curve in Fig. 5(b), the width of which was 24 fs at FWHM. According to the actual procedure, we integrated these local traces over the nanosecond pulse-envelope. Due to this integration effect, the trace broadened by 17% (28 fs FWHM) and the pedestal increased by a factor of 3 as compared with the local trace at t = 0. This is then a similar trace to the observed trace in Fig. 4. In spite of this modification, it was also found that the integrated trace substantially reflects the local autocorrelation trace around t = 0. Based on a comparison of these calculations, it is reasonably deduced for the 3-ns region around t = 0 that the train of short pulses was actually produced while keeping a duration of less than 20 fs. This corresponds to the quasi-steady generation of an ultrahigh-repetition-rate train with more than 30,000 short pulses.
Fig. 5. (a) Observed temporal profiles of the generated rotational sidebands.; (b): Waveform (blue solid-curve) on a femto-scale at t = 0 calculated from the temporal profiles (a), and the corresponding autocorrelation trace (green dotted-curve). The green solid-curve shows the autocorrelation trace integrated over the nanosecond driving-pulse-envelope.
5. Conclusions
In conclusion, we have shown that highly-stable, high-beam-quality phase-coherent Raman-sidebands can be produced by applying a dual-wavelength, injection-locked Ti:sapphire laser. By directly measuring the intensity temporal-profile, it has also been demonstrated that an ultrahigh-repetition-rate train of short pulses can be produced through a Fourier-synthesis of the phase-coherent Raman sidebands. It has been shown that a stable, 10.6-THz ultrahigh-repetition-rate train was formed with an effective-duration of 20 fs and a high peak-power of 2 MW. These results are an important step towards establishing Raman-sideband-synthesis ultrashort-pulses as an actual light source. The key idea for the potential applications of the light source would be manipulations of light-matter interactions on the basis of a resonance between an ultrahigh-repetition-rate and characteristic frequencies of the interactions.
Acknowledgments
The authors acknowledge K. Hakuta, R. Lang, M. Suzuki, and K. Ueda for their useful discussions and encouragement. They also thank K. Azegami and H. Satoh for their help with the experiment and the data analysis. This work was supported in part by a Grant-in-Aid for Scientific Research (B) from the Ministry of Education, Science, Sports, and Culture, the Research Foundation for Opto-Science and Technology, the Matsuo Foundation, and 21st Century COE program on Coherent Optical Science.
References and links
1. S. E. Harris and A. V. Sokolov, “Broadband spectral generation with refractive index control,” Phys. Rev. A 55, R4019–4022 (1997). [CrossRef]
2. A. V. Sokolov, D. R. Walker, D. D. Yavuz, G. Y. Yin, and S. E. Harris, “Raman generation by phased and antiphased molecular states,” Phys. Rev. Lett. 85, 562–565 (2000). [CrossRef] [PubMed]
3. M. Katsuragawa, J. Q. Liang, J. Z. Li, M. Suzuki, and K. Hakuta, “Stimulated Raman Scattering in Solid Hydrogen based on Adiabatic Preparation of Anti-Phased State,” CLEO/QELS ’99, Technical Digest, QthE2, pp.195-196, Baltimore, USA, May 23-28 (1999).
4. J. Q. Liang, M. Katsuragawa, F. Le Kien, and K. Hakuta, “Sideband generation using strongly driven Raman coherence in solid hydrogen,” Phys. Rev. Lett. 85, 2474–2477 (2000). [CrossRef] [PubMed]
5. M. Katsuragawa, J. Q. Liang, F. Le Kien, and K. Hakuta, “Efficient frequency conversion of incoherent fluorescent light,” Phys. Rev. A. 65, 025801–025804 (2002). [CrossRef]
6. M. Y. Shverdin, D. R. Walker, D. D. Yavuz, G. Y. Yin, and S. E. Harris, “Generation of a single-cycle optical pulse,” Phys. Rev. Lett. 94, 033904–033907 (2005). [CrossRef] [PubMed]
7. M. Katsuragawa and T. Onose, “A dual-wavelength injection-locked pulsed laser,” Opt. Lett. , in print; Japan Patent 2004-56879.
8. S. Yoshikawa and T. Imasaka, “A new approach for the generation of ultrashort optical pulses,” Opt. Commun. 96, 94–98 (1993). [CrossRef]
9. H. Kawano, Y. Hirakawa, and T. Imasaka, “Generation of more than 40 rotational Raman lines by picosecond and femtosecond Ti:sapphire laser for Fourier synthesis,” Appl. Phys. B 65, 1–4 (1996). [CrossRef]
10. D. D. Yavuz, D. R. Walker, G. Y. Yin, and S. E. Harris, “Rotational Raman generation with near-unity conversion efficiency,” Opt. Lett. 27, 769–771 (2002). [CrossRef]
11. D. D. Yavuz, D. R. Walker, M. Y. Shverdin, G. Y. Yin, and S. E. Harris, “Quasiperiodic Raman technique for ultrashort pulse generation,” Phys. Rev. Lett. 91, 233602–223605 (2003). [CrossRef] [PubMed]
12. M. Katsuragawa, Y. Ono, Fam Le Kien, and K. Hakuta, “Comb generation of collinear Raman sidebands by three correlated coherent molecular oscillations,” CLEO/IQEC 2004, Technical Digest CMO6, San Francisco CA, USA, May. 23-27 (2004).
13. M. Wittmann, A. Nazarkin, and G. Korn, “fs-Pulse synthesis using phase modulation by impulsively excited molecular vibrations,” Phys. Rev. Lett. 84, 5508–5511 (2000). [CrossRef] [PubMed]
14. A. V. Sokolov, D. R. Walker, D. D. Yavuz, G. Y. Yin, and S. E. Harris, “Femtosecond light source for phase-controlled multiphoton ionization,” Phys. Rev. Lett. 87, 033402–033405 (2001). [CrossRef] [PubMed]
15. The study regarding the optimization of adiabatic Raman processes will be reported in the future.
16. I. Matsuda, K. Misawa, and R. Lang, “Femtosecond chirp-variable apparatus using a chirped mirror pair for quantum coherent control,” Opt. Comm. 239, 181–186 (2004). [CrossRef]
17. There were no difficulties in maintaining this stability over several hours in the present experiment.
