Optical Kerr switching technique for the production of ... - OSA Publishing
Optical Kerr switching technique for the production of a picosecond, multiwavelength CO2 laser pulse Catalin V. Filip, Ritesh Narang, Sergei Ya. Tochitsky, Christopher E. Clayton, and Chandrashekhar Joshi
A wavelength-independent method for optical gating, based on the optical Kerr effect, has been demonstrated. Using this method, we produced 100-ps, 10-kW, two-wavelength pulses 共10.3 and 10.6 m兲 with a signal-to-background ratio contrast of 105 by slicing a long CO2 pulse. The capability of gating consecutive pulses separated on a picosecond time scale with this method is also shown. © 2002 Optical Society of America OCIS codes: 190.3270, 140.3470, 320.4240.
High-power lasers operating on two wavelengths are of importance for many applications, ranging from nonlinear optics1–3 to plasma-based electron acceleration.4,5 For these applications a high-power laser operating simultaneously on a pair of lines with a controllable line ratio is needed. A typical approach in building an infrared highpower laser is to produce a low-power, short seed pulse with good signal-to-background ratio contrast 共usually of the order of 105–106兲, followed by the injection of this pulse into a series of amplifiers.6 For CO2 lasers, several methods can be used to gate a low-power one-wavelength pulse on a picosecond time scale.7–11 Among them, the semiconductor switching technique10 –12 is widely used because it can provide one-wavelength CO2 pulses as short as 130 fs,12 with a contrast of up to 106, by chopping a much longer, nanosecond pulse. Recently, with this method, picosecond 共40 –200-ps兲, two-wavelength 共10.3- and 10.6-m兲 pulses with peak powers of around 1 TW were obtained.13,14 However, the best contrast achieved so far for a two-wavelength 100-ps seed pulse is 103 共see Ref. 13兲. Furthermore, the contrast is expected to decrease dramatically when C. V. Filip 共[email protected]兲. R. Narang, S. Ya. Tochitsky, C. E. Clayton, and C. Joshi are with the Department of Electrical Engineering, University of California, Los Angeles, Los Angeles, California 90095. Received 7 August 2001; revised manuscript received 14 March 2002. 0003-6935兾02兾183743-05$15.00兾0 © 2002 Optical Society of America
the separation between the CO2 lines is increased. Gating multiline CO2 laser pulses with semiconductor switching is limited because the two or three surfaces used to control the transmission or reflection of the CO2 radiation have to be oriented at the Brewster’s angle for one or the other of the CO2 wavelengths. The contrast of the switched-out pulse thus depends on the separation between the lines. Further amplification of picosecond pulses in highpressure broadband CO2 amplifiers can be achieved only if the short seed pulse has a good enough contrast to be preferentially amplified compared with the long-duration injected leakage.6 The addition of more semiconductor switching surfaces significantly complicates the system and reduces the gating efficiency. In addition to the contrast limitation, the semiconductor switching technique cannot be used to gate a train of consecutive pulses if the pulse separation is shorter than the plasma recombination time for that particular semiconductor material, typically tens of nanoseconds. Semiconductor switches grown by molecular beam epitaxy were reported15 having a recombination time close to 1 ps, although at the expense of the switching efficiency and a significant decrease in the damage threshold. To overcome these two important limitations in producing short CO2 laser pulses, we developed a polarization gating technique based on the optical Kerr effect 共OKE兲. OKE has found a wide range of applications16 since it was first experimentally reported in 1964.17,18 A typical application of OKE for gating of optical pulses is the detection of ultrafast infrared signals on a pi20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3743
Fig. 1. Schematic of the experimental setup for Kerr modulation of a two-wavelength CO2 laser pulse. LPDS, low-pressure discharge sections; M, mirror; TEA, transversely excited atmosphere.
cosecond time scale. The infrared pulse acts as a pump pulse that gates a probe laser operating at wavelengths where fast detectors such as streak cameras are available.19 In this paper we present the reverse situation where a short 1-m pump pulse gates a CO2 laser pulse as a method of production of 10-m two-wavelength picosecond radiation. The OKE gating technique is based on the rotation of the polarization of a relatively long probe pulse during the time when the Kerr medium is birefringent. In our experiment the birefringence is excited in this nonlinear medium by a much shorter, 100-ps, 1.06-m pump pulse through the molecular orientation effect. The Kerr cell is placed between a polarizer and an analyzer. The function of the polarizer is to clean the polarization of the incoming CO2 pulse and to transmit only linearly polarized radiation. The analyzer is used to dump the long CO2 pulse and to transmit only the short gated pulse that has its polarization rotated by 兾2 rad in the Kerr medium. In the middle infrared region the best choice for the Kerr medium is CS2 that transmits both 10- and 1-m pulses, has a large nonlinear index of refraction n2 ⫽ 10⫺20 m2兾V2, and a fast ⬃2-ps relaxation time.20 In our experiment the pulse is generated by a laser operating on two CO2 lines, 10P共20兲 and 10R共16兲, with wavelengths equal to 10.6 and 10.3 m, respectively. The experimental setup is presented in Fig. 1. To obtain simultaneous oscillation on two CO2 lines with an equal line ratio, a grating was used to separate the optical cavities for the lines, as sug3744
APPLIED OPTICS 兾 Vol. 41, No. 18 兾 20 June 2002
Fig. 2. Temporal dynamics of the pulses involved in the gating process: 共a兲 the relative timing between the 10.6- and the 10.3-m probe pulses from the CO2 laser and the 1-m pump pulse 共the 10.3-m trace is inverted兲; 共b兲 switched-out CO2 pulse as recorded by a fast detector after the Kerr modulator.
gested by Sheffield et al.21 To obtain singlelongitudinal-mode generation on each wavelength, a hybrid technique is used.22 Thus the folded hybrid cavity 共Fig. 1兲 has two low-pressure discharge sections and a transversely excited atmosphere CO2 module capable of producing a 400-ns, 70-mJ, twowavelength pulse with good shot-to-shot reproducibility 共5–7%兲. The gain-switched CO2 pulse contains ⬃120 kW of peak power. Such a multiline laser resonator can be tuned on an arbitrary pair of lines with the line ratio controlled simply when the diffraction losses in the two cavities are varied.13 The pump pulse is produced by a Nd:YAG laser system that delivers pulses with energies up to 150 mJ in ⬵100 ps. The oscilloscope traces of the pump 共1m兲 and probe 共10.3- and 10.6-m兲 pulses are presented in Fig. 2共a兲. Relative delay between these pulses was adjusted to provide gating at the maximum of the CO2 laser pulse. In this experiment a 106:1 extinction ratio for the 10-m polarizer–analyzer system was necessary to provide a good signal-to-background ratio contrast for the switched-out pulse. Any coated optical element, such as a beam combiner, present in the CO2 beam path in between the polarizer and the analyzer significantly decreases the polarizer–analyzer contrast. That is why a noncollinear interaction geometry between the pump and the probe pulses in the
Kerr cell was used. A drawback of one employing a noncollinear geometry is that the CO2 pulse has to fit within the 1-m beam inside the CS2 cell. For this purpose we reduced the beam size of the 10-m laser and made the angle between the two beams as small as possible. The beam size of the 10-m radiation coming out of the CO2 laser was reduced four times to approximately 2w0 ⫽ 2 mm, with the help of a beam compressor, whereas the beam size of the 1-m laser was around 2w0 ⫽ 6 mm. The angle between the beams was 1.5°, and the 2-cm-long Kerr cell was translated from the intersection point ⫾1 cm along the CO2 path without significant change in the gated signal. Therefore it can be concluded that the interaction between the two laser pulses was present along the entire length of the cell. It is known that the Kerr effect is most efficient when the phase angle between the polarizations of the probe and pump beams is 兾4 rad. That is why we rotated and cleaned the polarization of the 1-m laser pulse using a combination of a half-wave plate and a high-energy polarizer cube with an extinction ratio of 104:1. Although the CO2 laser output is p polarized, the optics used to transport the beam from the oscillator to the Kerr cell induce significant depolarization of the infrared beam. To obtain a linearly polarized CO2 pulse, two Ge plates, one set at the Brewster’s angle for 10.6 m and the other for 10.3 m, were used as a polarizer. For the analyzer, six Ge plates and a low-power wire polarizer were used. Using this polarizer–analyzer combination, we achieved a maximum extinction ratio of 106. It was important to use uncoated NaCl windows for the Kerr cell to maintain this high extinction ratio when the cell was placed in between the polarizer and the analyzer. The total losses introduced in the CO2 beam, including absorption in the CS2 cell, were approximately 50% of the output of the oscillator. The transmission of the probe beam through the Kerr cell is given by19
冋
T共t兲 ⫽ A sin2 2.34 ⫻ 10 6
冉 冊
册
L n 2 I pump共t兲 ,
(1)
where A is the total attenuation of the 10-m beam that is due to the optical elements used in the path 共 A ⫽ 0.5兲, L is the interaction length 共L ⫽ 2 cm兲, is the wavelength of the probe beam 共 ⫽ 10 m兲, and Ipump共t兲 is the intensity of the pump pulse as a function of time in watts per square centimeter 共for a 100-mJ pulse the peak intensity is 3.5 GW兾cm2兲. For the above-mentioned parameters, the peak transmission of the 10-m beam becomes 12%, corresponding to a maximum rotation in polarization of the CO2 pulse of 兾6 rad. For these parameters, we measured the transmission of the Kerr switch using both a fast-response 共⬍1-ns兲 photoelectromagnetic detector 共Boston Electronics PEM-L-3兲 and an energy meter. In Figure 2共b兲 a trace of the switched-out CO2 pulse that we obtained using the fast detector is shown. The detection system is not fast enough to resolve picosec-
ond CO2 pulses and was mainly used to monitor the contrast of the gated pulse. We measured pulse energy using a Gentec energy meter with a sensitivity of 217 V兾J, coupled to a Gentec 100⫻ amplifier. The maximum energy recorded in the switched-out pulse was 1.1 J when pump pulses with intensities of 3.5 GW兾cm2 were utilized. The degradation of the CS2 that was due to dissociation begins to occur beyond pump intensities of 3.5 GW兾cm2. After approximately 5 h of operation at a 1-Hz repetition rate, the Kerr medium became yellowish and the CS2 was replaced to avoid any degradation of the performance of the system. Because self-focusing of the CO2 pulse inside the CS2 cell could affect the interaction geometry, we monitored the CO2 beam profile using a pyroelectric two-dimensional array detector 共Spiricon Pyrocam兲. No significant changes in the spatial profile of the 10-m beam were observed. The energy in the switched-out pulse of approximately 1 J corresponds to a 10% power switching efficiency for a pulse that is 100 ps long. This number is close to the theoretically calculated efficiency 关Eq. 共1兲兴 of approximately 12%. The slight difference could be attributed to the absorption of the pump pulse inside the Kerr cell, which is neglected in Eq. 共1兲. Because it is difficult to record mid-infrared pulses with picosecond time resolution, we investigated the switched-out pulse by replacing the probe CO2 pulse with a 658-nm single-longitudinal-mode laser diode pulse in a similar switching geometry. As a detector we used a streak camera 共Hadland Photonics Imacon 500兲 with ⬍10-ps resolution. In Fig. 3, three streaks of the gated visible pulse taken at pump intensity levels of 0.85, 1.6, and 2.6 GW兾cm2 are shown. The temporal shape of the switched-out visible pulse is shown in Fig. 3共a兲. The polarization of the visible beam is rotated by ⬃兾2 rad when the peak power of the pump beam reaches approximately 0.85 GW兾cm2. The pulse length 共FWHM兲 is ⬃100 ps. Slightly smaller values were recorded when the rotation in polarization was less than 兾2 rad. Because of the nonlinearity of the switching process, the width of the gated pulse is not necessarily the same as the width of the pump pulse. To investigate this issue, we modeled the shapes and the widths of the switched-out pulses when Gaussian, Lorentzian, and triangular pump pulses are employed. The theoretical calculations show that the gated pulses have widths comparable to the pump when the rotation of the probe beam polarization is close to 兾2 rad. Even though for smaller values of rotation the calculations show that the switched-out pulse is slightly shorter than the pump pulse, we do not expect significant shortening of the switched-out CO2 pulses when 10% switching efficiency is recorded. Therefore the duration of the gated 10-m pulses should be approximately equal to the duration of the pump pulse that is 100 ps 共FWHM兲. Also, in this experiment, with lasers operating on 10- and 1-m wavelengths, the dispersion of CS2 of approximately 0.9 ps兾cm does not produce any significant broadening of 20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3745
ther injected into an eight-atmosphere CO2 laser 共Lumonics TE-280兲. Injection mode locking25 was simultaneously achieved for both frequencies, resulting in a pulse train having a gain-switched temporal envelope. Because the injected short pulse has a different polarization than the leakage, the polarization-selective cavity of the amplifier was able to maintain the contrast of the amplified pulse. The energy in the strongest pulse was approximately 1 mJ, which represents an amplification factor of 103. In conclusion, a wavelength-independent technique for gating CO2 laser pulses on a picosecond time scale with parameters suitable for further amplification was demonstrated. This has produced two-wavelength, 10.3- and 10.6-m, CO2 laser pulses with a duration of ⬃100 ps and a signal-tobackground ratio contrast of 105. The energy contained in these pulses is 1 J, and this corresponds to a 10% power switching efficiency. The high contrast of these pulses has been preserved after amplification of a factor of 103 in an injection mode-locked amplifier. The capability of this method to gate a train of mid-infrared pulses separated on a picosecond time scale has been shown. We thank K. Marsh and P. Muggli for their help with this research. This research was supported by the U.S. Department of Energy under contract DEFG03-92ER40727. References
Fig. 3. Temporal profile of the switched-out 658-nm pulse recorded with a streak camera at three different pump intensities. These intensities correspond to a rotation of the polarization of the probe pulse of 共a兲 兾2 rad, 共b兲 rad, and 共c兲 3兾2 rad.
the gated pulse. The phenomena of dispersion in the Kerr medium and the limited response time of CS2 become important when shorter pump pulses 共1–10 ps兲 are employed. Ultimately, mid-infrared pulses as short as 2 ps can be produced with this technique. Figures 3共b兲 and 3共c兲 show the temporal profile of the gated visible pulse when pump pulse intensities of 1.6 and 2.6 GW兾cm2 were employed, respectively. For these levels of pump power, the OKE induced in the Kerr cell is strong enough to produce overrotation23 of the polarization of the probe pulse with rad 关Fig. 3共b兲兴 and 3兾2 rad 关Fig. 3共c兲兴. This shows that the switching mechanism presented here based on OKE can be used to gate a series of consecutive pulses separated in time by few tens of picoseconds and still maintain the same high-contrast and the same switching efficiency. The ability to construct an optical modulator based on OKE in CS2 may become important when a train of CO2 pulses is required.24 The gated 100-ps two-wavelength pulses were fur3746
APPLIED OPTICS 兾 Vol. 41, No. 18 兾 20 June 2002
1. L. F. Tiemeijer, “Effects of nonlinear gain on four-wave mixing and asymmetric gain saturation in a semiconductor laser amplifier,” Appl. Phys. Lett. 59, 499 –501 共1991兲. 2. J. Zhou, N. Park, J. W. Dawson, K. J. Vahala, M. A. Newkirk, and B. I. Miller, “Terahertz four-wave mixing spectroscopy for study of ultrafast dynamics in a semiconductor optical amplifier,” Appl. Phys. Lett. 63, 1179 –1181 共1993兲. 3. V. G. Dmitriev, G. G. Gurzadyan, and D. N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Vol. 64 of the Springer Series in Optical Sciences, 3rd ed. 共Springer, New York, 1999兲 p. 340. 4. C. E. Clayton, K. A. Marsh, A. Dyson, M. Everett, A. Lal, W. P. Leemans, R. Williams, and C. Joshi, “Ultrahigh-gradient acceleration of injected electrons by laser-excited relativistic electron plasma waves,” Phys. Rev. Lett. 70, 37– 40 共1993兲. 5. M. Everett, A. Lal, D. Gordon, C. E. Clayton, K. A. Marsh, and C. Joshi, “Trapped electron acceleration by a laser-driven relativistic plasma wave,” Nature 共London兲 368, 527–529 共1994兲. 6. P. B. Corkum, “Amplification of picosecond 10 m pulses in multiatmospheric CO2 lasers,” IEEE J. Quantum Electron. QE-21, 216 –232 共1985兲. 7. J. F. Figueira, W. H. Reichelt, G. T. Schappert, T. F. Stratton, and C. H. Fenstermacher, “Nanosecond pulse amplification in electron-beam-pumped CO2 amplifiers,” Appl. Phys. Lett. 22, 216 –218 共1973兲. 8. E. Yablonovitch and J. Goldhar, “Short CO2 laser pulse generation by optical free induction decay,” Appl. Phys. Lett. 25, 580 –582 共1974兲. 9. R. A. Kaindl, D. C. Smith, M. Joschko, M. P. Hasselbeck, M. Woerner, and T. Elsaesser, “Femtosecond infrared pulses tunable from 9 to 18 m at an 88-MHz repetition rate,” Opt. Lett. 23, 861– 863 共1998兲. 10. A. J. Alcock and P. B. Corkum, “Ultra-fast switching of infra-
11.
12. 13.
14.
15.
16. 17.
red radiation by laser-produced carriers in semiconductors,” Can. J. Phys. 57, 1280 –1290 共1979兲. I. V. Pogorelsky, J. Fisher, K. P. Kusche, M. Babzien, N. A. Kurnitt, I. J. Bigio, R. F. Harrison, and T. Shimada, “Subnanosecond multi-gigawatt CO2 laser,” IEEE J. Quantum Electron. QE-31, 556 –566 共1995兲. C. Rolland and P. B. Corkum, “Generation of 130-fsec midinfrared pulses,” J. Opt. Soc. Am. B 3, 1625–1629 共1986兲. S. Ya. Tochitsky, R. Narang, C. Filip, B. Blue, C. E. Clayton, K. A. Marsh, and C. Joshi, “Amplification of two-wavelength CO2 laser pulses to terawatt level,” in Proceedings LASERS’99, V. J. Corcoran and T. A. Corcoran, eds. 共STS, McLean, Va., 2000兲, pp. 265–272. S. Ya. Tochitsky, C. Filip, R. Narang, C. E. Clayton, K. A. Marsh, and C. Joshi, “Efficient shortening of self-chirped picosecond pulses in a high-power CO2 amplifier,” Opt. Lett. 26, 813– 815 共2001兲. A. Y. Elezzabi, J. Meyer, M. K. Y. Hughes, and S. R. Johnson, “Generation of 1-ps infrared pulses at 10.6 m by use of lowtemperature-grown GaAs as an optical semiconductor switch,” Opt. Lett. 19, 898 –900 共1994兲. K. Sala and M. C. Richardson, “Optical Kerr effect induced by ultrashort laser pulses,” Phys. Rev. A 12, 1036 –1047 共1975兲. G. Mayer and F. Gires, “Action d’une onde lumineuse intense sur l’indice de refraction des liquides,” C. R. Acad. Sci. 共Paris兲 258, 2039 共1964兲.
18. P. D. Maker, R. W. Terhune, and C. M. Savage, “Intensitydependent changes in the refractive index of liquids,” Phys. Rev. Lett. 12, 507–509 共1964兲. 19. T. C. Owen, L. W. Coleman, and T. J. Burgess, “Ultrafast optical Kerr effect in CS2 at 10.6 m,” Appl. Phys. Lett. 22, 272–273 共1973兲. 20. M. A. Duguay and J. W. Hansen, “An ultrafast light gate,” Appl. Phys. Lett. 15, 192–194 共1969兲. 21. R. L. Sheffield, S. Nazemi, and A. Javan, “An independently controllable multiline laser resonator and its use in multifrequency injection locking,” Appl. Phys. Lett. 29, 588 –590 共1976兲. 22. D. M. Tratt, A. K. Kar, and R. G. Harrison, “Spectral control of gain-switched lasers by injection-seeding: application to TEA CO2 systems,” Prog. Quantum Electron. 10, 229 –265 共1985兲. 23. T. Morioka and M. Saruwatari, “Ultrafast all-optical switching utilizing the optical Kerr effect in polarization-maintaining single-mode fibers,” IEEE J. Sel. Areas Commun. 6, 1186 – 1198 共1988兲. 24. T. Hirose, “Project of polarized positron beams for future linear colliders,” in Proceedings LASERS’98, V. J. Corcoran and T. A. Goldman, 共STS, Mclean, Va., 1999兲, pp. 836 – 844. 25. P. A. Belanger and J. Boivin, “Gigawatt peak-power pulse generation by injection of a single short pulse in a regenerative amplifier above threshold 共RAAT兲,” Can. J. Phys. 54, 720 –727 共1976兲.
20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3747
A wavelength-independent method for optical gating, based on the optical Kerr effect, has been demonstrated. Using this method, we produced 100-ps, 10-kW, two-wavelength pulses 共10.3 and 10.6 m兲 with a signal-to-background ratio contrast of 105 by slicing a long CO2 pulse. The capability of gating consecutive pulses separated on a picosecond time scale with this method is also shown. © 2002 Optical Society of America OCIS codes: 190.3270, 140.3470, 320.4240.
High-power lasers operating on two wavelengths are of importance for many applications, ranging from nonlinear optics1–3 to plasma-based electron acceleration.4,5 For these applications a high-power laser operating simultaneously on a pair of lines with a controllable line ratio is needed. A typical approach in building an infrared highpower laser is to produce a low-power, short seed pulse with good signal-to-background ratio contrast 共usually of the order of 105–106兲, followed by the injection of this pulse into a series of amplifiers.6 For CO2 lasers, several methods can be used to gate a low-power one-wavelength pulse on a picosecond time scale.7–11 Among them, the semiconductor switching technique10 –12 is widely used because it can provide one-wavelength CO2 pulses as short as 130 fs,12 with a contrast of up to 106, by chopping a much longer, nanosecond pulse. Recently, with this method, picosecond 共40 –200-ps兲, two-wavelength 共10.3- and 10.6-m兲 pulses with peak powers of around 1 TW were obtained.13,14 However, the best contrast achieved so far for a two-wavelength 100-ps seed pulse is 103 共see Ref. 13兲. Furthermore, the contrast is expected to decrease dramatically when C. V. Filip 共[email protected]兲. R. Narang, S. Ya. Tochitsky, C. E. Clayton, and C. Joshi are with the Department of Electrical Engineering, University of California, Los Angeles, Los Angeles, California 90095. Received 7 August 2001; revised manuscript received 14 March 2002. 0003-6935兾02兾183743-05$15.00兾0 © 2002 Optical Society of America
the separation between the CO2 lines is increased. Gating multiline CO2 laser pulses with semiconductor switching is limited because the two or three surfaces used to control the transmission or reflection of the CO2 radiation have to be oriented at the Brewster’s angle for one or the other of the CO2 wavelengths. The contrast of the switched-out pulse thus depends on the separation between the lines. Further amplification of picosecond pulses in highpressure broadband CO2 amplifiers can be achieved only if the short seed pulse has a good enough contrast to be preferentially amplified compared with the long-duration injected leakage.6 The addition of more semiconductor switching surfaces significantly complicates the system and reduces the gating efficiency. In addition to the contrast limitation, the semiconductor switching technique cannot be used to gate a train of consecutive pulses if the pulse separation is shorter than the plasma recombination time for that particular semiconductor material, typically tens of nanoseconds. Semiconductor switches grown by molecular beam epitaxy were reported15 having a recombination time close to 1 ps, although at the expense of the switching efficiency and a significant decrease in the damage threshold. To overcome these two important limitations in producing short CO2 laser pulses, we developed a polarization gating technique based on the optical Kerr effect 共OKE兲. OKE has found a wide range of applications16 since it was first experimentally reported in 1964.17,18 A typical application of OKE for gating of optical pulses is the detection of ultrafast infrared signals on a pi20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3743
Fig. 1. Schematic of the experimental setup for Kerr modulation of a two-wavelength CO2 laser pulse. LPDS, low-pressure discharge sections; M, mirror; TEA, transversely excited atmosphere.
cosecond time scale. The infrared pulse acts as a pump pulse that gates a probe laser operating at wavelengths where fast detectors such as streak cameras are available.19 In this paper we present the reverse situation where a short 1-m pump pulse gates a CO2 laser pulse as a method of production of 10-m two-wavelength picosecond radiation. The OKE gating technique is based on the rotation of the polarization of a relatively long probe pulse during the time when the Kerr medium is birefringent. In our experiment the birefringence is excited in this nonlinear medium by a much shorter, 100-ps, 1.06-m pump pulse through the molecular orientation effect. The Kerr cell is placed between a polarizer and an analyzer. The function of the polarizer is to clean the polarization of the incoming CO2 pulse and to transmit only linearly polarized radiation. The analyzer is used to dump the long CO2 pulse and to transmit only the short gated pulse that has its polarization rotated by 兾2 rad in the Kerr medium. In the middle infrared region the best choice for the Kerr medium is CS2 that transmits both 10- and 1-m pulses, has a large nonlinear index of refraction n2 ⫽ 10⫺20 m2兾V2, and a fast ⬃2-ps relaxation time.20 In our experiment the pulse is generated by a laser operating on two CO2 lines, 10P共20兲 and 10R共16兲, with wavelengths equal to 10.6 and 10.3 m, respectively. The experimental setup is presented in Fig. 1. To obtain simultaneous oscillation on two CO2 lines with an equal line ratio, a grating was used to separate the optical cavities for the lines, as sug3744
APPLIED OPTICS 兾 Vol. 41, No. 18 兾 20 June 2002
Fig. 2. Temporal dynamics of the pulses involved in the gating process: 共a兲 the relative timing between the 10.6- and the 10.3-m probe pulses from the CO2 laser and the 1-m pump pulse 共the 10.3-m trace is inverted兲; 共b兲 switched-out CO2 pulse as recorded by a fast detector after the Kerr modulator.
gested by Sheffield et al.21 To obtain singlelongitudinal-mode generation on each wavelength, a hybrid technique is used.22 Thus the folded hybrid cavity 共Fig. 1兲 has two low-pressure discharge sections and a transversely excited atmosphere CO2 module capable of producing a 400-ns, 70-mJ, twowavelength pulse with good shot-to-shot reproducibility 共5–7%兲. The gain-switched CO2 pulse contains ⬃120 kW of peak power. Such a multiline laser resonator can be tuned on an arbitrary pair of lines with the line ratio controlled simply when the diffraction losses in the two cavities are varied.13 The pump pulse is produced by a Nd:YAG laser system that delivers pulses with energies up to 150 mJ in ⬵100 ps. The oscilloscope traces of the pump 共1m兲 and probe 共10.3- and 10.6-m兲 pulses are presented in Fig. 2共a兲. Relative delay between these pulses was adjusted to provide gating at the maximum of the CO2 laser pulse. In this experiment a 106:1 extinction ratio for the 10-m polarizer–analyzer system was necessary to provide a good signal-to-background ratio contrast for the switched-out pulse. Any coated optical element, such as a beam combiner, present in the CO2 beam path in between the polarizer and the analyzer significantly decreases the polarizer–analyzer contrast. That is why a noncollinear interaction geometry between the pump and the probe pulses in the
Kerr cell was used. A drawback of one employing a noncollinear geometry is that the CO2 pulse has to fit within the 1-m beam inside the CS2 cell. For this purpose we reduced the beam size of the 10-m laser and made the angle between the two beams as small as possible. The beam size of the 10-m radiation coming out of the CO2 laser was reduced four times to approximately 2w0 ⫽ 2 mm, with the help of a beam compressor, whereas the beam size of the 1-m laser was around 2w0 ⫽ 6 mm. The angle between the beams was 1.5°, and the 2-cm-long Kerr cell was translated from the intersection point ⫾1 cm along the CO2 path without significant change in the gated signal. Therefore it can be concluded that the interaction between the two laser pulses was present along the entire length of the cell. It is known that the Kerr effect is most efficient when the phase angle between the polarizations of the probe and pump beams is 兾4 rad. That is why we rotated and cleaned the polarization of the 1-m laser pulse using a combination of a half-wave plate and a high-energy polarizer cube with an extinction ratio of 104:1. Although the CO2 laser output is p polarized, the optics used to transport the beam from the oscillator to the Kerr cell induce significant depolarization of the infrared beam. To obtain a linearly polarized CO2 pulse, two Ge plates, one set at the Brewster’s angle for 10.6 m and the other for 10.3 m, were used as a polarizer. For the analyzer, six Ge plates and a low-power wire polarizer were used. Using this polarizer–analyzer combination, we achieved a maximum extinction ratio of 106. It was important to use uncoated NaCl windows for the Kerr cell to maintain this high extinction ratio when the cell was placed in between the polarizer and the analyzer. The total losses introduced in the CO2 beam, including absorption in the CS2 cell, were approximately 50% of the output of the oscillator. The transmission of the probe beam through the Kerr cell is given by19
冋
T共t兲 ⫽ A sin2 2.34 ⫻ 10 6
冉 冊
册
L n 2 I pump共t兲 ,
(1)
where A is the total attenuation of the 10-m beam that is due to the optical elements used in the path 共 A ⫽ 0.5兲, L is the interaction length 共L ⫽ 2 cm兲, is the wavelength of the probe beam 共 ⫽ 10 m兲, and Ipump共t兲 is the intensity of the pump pulse as a function of time in watts per square centimeter 共for a 100-mJ pulse the peak intensity is 3.5 GW兾cm2兲. For the above-mentioned parameters, the peak transmission of the 10-m beam becomes 12%, corresponding to a maximum rotation in polarization of the CO2 pulse of 兾6 rad. For these parameters, we measured the transmission of the Kerr switch using both a fast-response 共⬍1-ns兲 photoelectromagnetic detector 共Boston Electronics PEM-L-3兲 and an energy meter. In Figure 2共b兲 a trace of the switched-out CO2 pulse that we obtained using the fast detector is shown. The detection system is not fast enough to resolve picosec-
ond CO2 pulses and was mainly used to monitor the contrast of the gated pulse. We measured pulse energy using a Gentec energy meter with a sensitivity of 217 V兾J, coupled to a Gentec 100⫻ amplifier. The maximum energy recorded in the switched-out pulse was 1.1 J when pump pulses with intensities of 3.5 GW兾cm2 were utilized. The degradation of the CS2 that was due to dissociation begins to occur beyond pump intensities of 3.5 GW兾cm2. After approximately 5 h of operation at a 1-Hz repetition rate, the Kerr medium became yellowish and the CS2 was replaced to avoid any degradation of the performance of the system. Because self-focusing of the CO2 pulse inside the CS2 cell could affect the interaction geometry, we monitored the CO2 beam profile using a pyroelectric two-dimensional array detector 共Spiricon Pyrocam兲. No significant changes in the spatial profile of the 10-m beam were observed. The energy in the switched-out pulse of approximately 1 J corresponds to a 10% power switching efficiency for a pulse that is 100 ps long. This number is close to the theoretically calculated efficiency 关Eq. 共1兲兴 of approximately 12%. The slight difference could be attributed to the absorption of the pump pulse inside the Kerr cell, which is neglected in Eq. 共1兲. Because it is difficult to record mid-infrared pulses with picosecond time resolution, we investigated the switched-out pulse by replacing the probe CO2 pulse with a 658-nm single-longitudinal-mode laser diode pulse in a similar switching geometry. As a detector we used a streak camera 共Hadland Photonics Imacon 500兲 with ⬍10-ps resolution. In Fig. 3, three streaks of the gated visible pulse taken at pump intensity levels of 0.85, 1.6, and 2.6 GW兾cm2 are shown. The temporal shape of the switched-out visible pulse is shown in Fig. 3共a兲. The polarization of the visible beam is rotated by ⬃兾2 rad when the peak power of the pump beam reaches approximately 0.85 GW兾cm2. The pulse length 共FWHM兲 is ⬃100 ps. Slightly smaller values were recorded when the rotation in polarization was less than 兾2 rad. Because of the nonlinearity of the switching process, the width of the gated pulse is not necessarily the same as the width of the pump pulse. To investigate this issue, we modeled the shapes and the widths of the switched-out pulses when Gaussian, Lorentzian, and triangular pump pulses are employed. The theoretical calculations show that the gated pulses have widths comparable to the pump when the rotation of the probe beam polarization is close to 兾2 rad. Even though for smaller values of rotation the calculations show that the switched-out pulse is slightly shorter than the pump pulse, we do not expect significant shortening of the switched-out CO2 pulses when 10% switching efficiency is recorded. Therefore the duration of the gated 10-m pulses should be approximately equal to the duration of the pump pulse that is 100 ps 共FWHM兲. Also, in this experiment, with lasers operating on 10- and 1-m wavelengths, the dispersion of CS2 of approximately 0.9 ps兾cm does not produce any significant broadening of 20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3745
ther injected into an eight-atmosphere CO2 laser 共Lumonics TE-280兲. Injection mode locking25 was simultaneously achieved for both frequencies, resulting in a pulse train having a gain-switched temporal envelope. Because the injected short pulse has a different polarization than the leakage, the polarization-selective cavity of the amplifier was able to maintain the contrast of the amplified pulse. The energy in the strongest pulse was approximately 1 mJ, which represents an amplification factor of 103. In conclusion, a wavelength-independent technique for gating CO2 laser pulses on a picosecond time scale with parameters suitable for further amplification was demonstrated. This has produced two-wavelength, 10.3- and 10.6-m, CO2 laser pulses with a duration of ⬃100 ps and a signal-tobackground ratio contrast of 105. The energy contained in these pulses is 1 J, and this corresponds to a 10% power switching efficiency. The high contrast of these pulses has been preserved after amplification of a factor of 103 in an injection mode-locked amplifier. The capability of this method to gate a train of mid-infrared pulses separated on a picosecond time scale has been shown. We thank K. Marsh and P. Muggli for their help with this research. This research was supported by the U.S. Department of Energy under contract DEFG03-92ER40727. References
Fig. 3. Temporal profile of the switched-out 658-nm pulse recorded with a streak camera at three different pump intensities. These intensities correspond to a rotation of the polarization of the probe pulse of 共a兲 兾2 rad, 共b兲 rad, and 共c兲 3兾2 rad.
the gated pulse. The phenomena of dispersion in the Kerr medium and the limited response time of CS2 become important when shorter pump pulses 共1–10 ps兲 are employed. Ultimately, mid-infrared pulses as short as 2 ps can be produced with this technique. Figures 3共b兲 and 3共c兲 show the temporal profile of the gated visible pulse when pump pulse intensities of 1.6 and 2.6 GW兾cm2 were employed, respectively. For these levels of pump power, the OKE induced in the Kerr cell is strong enough to produce overrotation23 of the polarization of the probe pulse with rad 关Fig. 3共b兲兴 and 3兾2 rad 关Fig. 3共c兲兴. This shows that the switching mechanism presented here based on OKE can be used to gate a series of consecutive pulses separated in time by few tens of picoseconds and still maintain the same high-contrast and the same switching efficiency. The ability to construct an optical modulator based on OKE in CS2 may become important when a train of CO2 pulses is required.24 The gated 100-ps two-wavelength pulses were fur3746
APPLIED OPTICS 兾 Vol. 41, No. 18 兾 20 June 2002
1. L. F. Tiemeijer, “Effects of nonlinear gain on four-wave mixing and asymmetric gain saturation in a semiconductor laser amplifier,” Appl. Phys. Lett. 59, 499 –501 共1991兲. 2. J. Zhou, N. Park, J. W. Dawson, K. J. Vahala, M. A. Newkirk, and B. I. Miller, “Terahertz four-wave mixing spectroscopy for study of ultrafast dynamics in a semiconductor optical amplifier,” Appl. Phys. Lett. 63, 1179 –1181 共1993兲. 3. V. G. Dmitriev, G. G. Gurzadyan, and D. N. Nikogosyan, Handbook of Nonlinear Optical Crystals, Vol. 64 of the Springer Series in Optical Sciences, 3rd ed. 共Springer, New York, 1999兲 p. 340. 4. C. E. Clayton, K. A. Marsh, A. Dyson, M. Everett, A. Lal, W. P. Leemans, R. Williams, and C. Joshi, “Ultrahigh-gradient acceleration of injected electrons by laser-excited relativistic electron plasma waves,” Phys. Rev. Lett. 70, 37– 40 共1993兲. 5. M. Everett, A. Lal, D. Gordon, C. E. Clayton, K. A. Marsh, and C. Joshi, “Trapped electron acceleration by a laser-driven relativistic plasma wave,” Nature 共London兲 368, 527–529 共1994兲. 6. P. B. Corkum, “Amplification of picosecond 10 m pulses in multiatmospheric CO2 lasers,” IEEE J. Quantum Electron. QE-21, 216 –232 共1985兲. 7. J. F. Figueira, W. H. Reichelt, G. T. Schappert, T. F. Stratton, and C. H. Fenstermacher, “Nanosecond pulse amplification in electron-beam-pumped CO2 amplifiers,” Appl. Phys. Lett. 22, 216 –218 共1973兲. 8. E. Yablonovitch and J. Goldhar, “Short CO2 laser pulse generation by optical free induction decay,” Appl. Phys. Lett. 25, 580 –582 共1974兲. 9. R. A. Kaindl, D. C. Smith, M. Joschko, M. P. Hasselbeck, M. Woerner, and T. Elsaesser, “Femtosecond infrared pulses tunable from 9 to 18 m at an 88-MHz repetition rate,” Opt. Lett. 23, 861– 863 共1998兲. 10. A. J. Alcock and P. B. Corkum, “Ultra-fast switching of infra-
11.
12. 13.
14.
15.
16. 17.
red radiation by laser-produced carriers in semiconductors,” Can. J. Phys. 57, 1280 –1290 共1979兲. I. V. Pogorelsky, J. Fisher, K. P. Kusche, M. Babzien, N. A. Kurnitt, I. J. Bigio, R. F. Harrison, and T. Shimada, “Subnanosecond multi-gigawatt CO2 laser,” IEEE J. Quantum Electron. QE-31, 556 –566 共1995兲. C. Rolland and P. B. Corkum, “Generation of 130-fsec midinfrared pulses,” J. Opt. Soc. Am. B 3, 1625–1629 共1986兲. S. Ya. Tochitsky, R. Narang, C. Filip, B. Blue, C. E. Clayton, K. A. Marsh, and C. Joshi, “Amplification of two-wavelength CO2 laser pulses to terawatt level,” in Proceedings LASERS’99, V. J. Corcoran and T. A. Corcoran, eds. 共STS, McLean, Va., 2000兲, pp. 265–272. S. Ya. Tochitsky, C. Filip, R. Narang, C. E. Clayton, K. A. Marsh, and C. Joshi, “Efficient shortening of self-chirped picosecond pulses in a high-power CO2 amplifier,” Opt. Lett. 26, 813– 815 共2001兲. A. Y. Elezzabi, J. Meyer, M. K. Y. Hughes, and S. R. Johnson, “Generation of 1-ps infrared pulses at 10.6 m by use of lowtemperature-grown GaAs as an optical semiconductor switch,” Opt. Lett. 19, 898 –900 共1994兲. K. Sala and M. C. Richardson, “Optical Kerr effect induced by ultrashort laser pulses,” Phys. Rev. A 12, 1036 –1047 共1975兲. G. Mayer and F. Gires, “Action d’une onde lumineuse intense sur l’indice de refraction des liquides,” C. R. Acad. Sci. 共Paris兲 258, 2039 共1964兲.
18. P. D. Maker, R. W. Terhune, and C. M. Savage, “Intensitydependent changes in the refractive index of liquids,” Phys. Rev. Lett. 12, 507–509 共1964兲. 19. T. C. Owen, L. W. Coleman, and T. J. Burgess, “Ultrafast optical Kerr effect in CS2 at 10.6 m,” Appl. Phys. Lett. 22, 272–273 共1973兲. 20. M. A. Duguay and J. W. Hansen, “An ultrafast light gate,” Appl. Phys. Lett. 15, 192–194 共1969兲. 21. R. L. Sheffield, S. Nazemi, and A. Javan, “An independently controllable multiline laser resonator and its use in multifrequency injection locking,” Appl. Phys. Lett. 29, 588 –590 共1976兲. 22. D. M. Tratt, A. K. Kar, and R. G. Harrison, “Spectral control of gain-switched lasers by injection-seeding: application to TEA CO2 systems,” Prog. Quantum Electron. 10, 229 –265 共1985兲. 23. T. Morioka and M. Saruwatari, “Ultrafast all-optical switching utilizing the optical Kerr effect in polarization-maintaining single-mode fibers,” IEEE J. Sel. Areas Commun. 6, 1186 – 1198 共1988兲. 24. T. Hirose, “Project of polarized positron beams for future linear colliders,” in Proceedings LASERS’98, V. J. Corcoran and T. A. Goldman, 共STS, Mclean, Va., 1999兲, pp. 836 – 844. 25. P. A. Belanger and J. Boivin, “Gigawatt peak-power pulse generation by injection of a single short pulse in a regenerative amplifier above threshold 共RAAT兲,” Can. J. Phys. 54, 720 –727 共1976兲.
20 June 2002 兾 Vol. 41, No. 18 兾 APPLIED OPTICS
3747
