202
OPTICS LETTERS / Vol. 24, No. 4 / February 15, 1999
Enhancement of terahertz-wave output from LiNbO3 optical parametric oscillators by cryogenic cooling Jun-ichi Shikata, Manabu Sato,* Tetsuo Taniuchi, and Hiromasa Ito Research Institute of Electrical Communication, Tohoku University, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan
Kodo Kawase Department of Applied Physics, Tohoku Gakuin University, Tagajo 985-8537, Japan Received October 19, 1998 In recent years widely tunable terahertz- (THz-) wave generation from LiNbO3 optical parametric oscillators (OPO’s) has been successfully demonstrated by use of the prism output-coupler method. However, there remains a problem of large absorption loss for generated terahertz waves inside the crystal, so we investigated the cryogenic characteristics of the OPO. We achieved 125-times-higher THz-wave output and 32% reduction of the generation threshold by cooling the crystal to 78 K. This scheme also provides direct loss measurement at THz frequency, and we found that the THz-wave enhancement mechanism is improvement of the gain as well as the reduction of the absorption coeff icient. 1999 Optical Society of America OCIS codes: 190.2620, 190.4410, 190.4970, 190.5890, 300.1030.
During the past several years terahertz- (THz-) wave generation and detection have attracted much attention from both fundamental and applied points of view. Most studies have utilized the ultrabroad-bandwidth characteristics of the mode-locked subpicosecond laser pulses with a sacrif ice in temporal coherence.1 – 3 In contrast, we recently demonstrated the method of coherent tunable THz-wave generation (frequency, 0.9– 2.1 THz; wavelength, 140 310 mm) from a LiNbO3 optical parametric (OPO) by introducing outputcoupling methods for the THz wave to improve substantially the generation efficiency.4 – 6 This scheme is based on parametric downconversion of laser light by use of the lowest A1 -symmetry polariton mode of LiNbO3 (stimulated polariton scattering),7 and all the interacting waves are polarized along the z axis of the crystal. We can easily achieve a widely tunable THz wave by changing slightly the angle between the near-infrared pump and idler (Stokes) wave vectors, based on the noncollinear phase-matching condition kp ki 1 kT , where the indices p, i, and T denote pump, idler, and THz wave, respectively. LiNbO3 is one of the most suitable materials for efficient generation of THz waves because of its large nonlinear coefficient and its transparent characteristics in the wide-wavelength range. However, a large absorption coefficient (approximately a few tens of inverse centimeters) at the THz frequency is a serious problem, because eff iciently generated THz waves suffer large absorption losses inside the crystal (in the case of a 1-mm propagation distance THz-wave power drops by several orders). On the other hand, the absorption coeff icient is known to be reduced by cooling of the crystal,8 which is expected to be a very effective approach to THz-wave generation. In this Letter we report the cryogenic characteristics of the OPO, which have been investigated in terms of the absorption coefficient as well as of parametric gain in the THz region. Stimulated polariton scattering occurs in both infrared and Raman-active materials, such as LiNbO3 , 0146-9592/99/040202-03$15.00/0
LiTaO3 , and GaP.9 Although a nontunable Ramanscattering process is dominant in the resonant region, a continuously tunable parametric process occurs in the lower-energy region. As thermal activation is thought to perturb the coherent interaction process, highly efficient generation of THz waves is expected at low temperature. In fact, a three-times-higher stimulated Raman gain (lowest A1 -symmetry mode) in LiNbO3 was observed when the crystal was cooled to 80 K, owing to the decrease in the bandwidth of the phonon mode.10 A similar effect is expected even in the nonresonant parametric region. According to the plane-wave approach, analytical expressions of exponential gains for idler and THz waves are given by11 aT gT gi cos f 2
µ
Ω∑
g0 1 1 16 cos f aT
∂2 ∏1/2
æ
21 , (1)
where f denotes the phase-matching angle between the pump and the THz wave vectors; g0 is the parametric gain in the low-loss limit and aT is an absorption coeff icient in the THz region. In cgs units, they are written as11 g0
µ
pvT vi Ip 2c3 nT ni np
∂1/2 µ
dE 0 1
X Sj v0 2 dQj 0 ∂ , 2 2 j v0j 2 vT
(2)
aT 2jIm kT j ∂1/2 µ X Sj v0j 2 2vT , (3) Im e` 1 c v0j 2 2 vT 2 2 ivT Gj j where nb sb p, i, T d is a refractive index, Ip is the pump intensity, and dE 0 and dQ 0 are nonlinear coeff icients related to pure parametric (second-order) 1999 Optical Society of America
February 15, 1999 / Vol. 24, No. 4 / OPTICS LETTERS
and Raman (third-order) scattering processes.12 Parameters v0j , Sj , and Gj denote eigenfrequency, oscillator strength, and the bandwidth of the jth A1 -symmetry phonon mode, respectively. Figure 1 shows the calculated parametric gain, gT , for two temperatures at various pump intensities, where the pump wavelength is fixed at 1.064 mm. The gain increases with pump intensity, because it is monotonically increasing with g0 [Eq. (1)], which is proportional to Ip [Eq. (2)]. In addition, at low temperature the gain is enhanced, and the peak position (near 1 THz at room temperature) shifts to the higher-energy side. The enhancement of the gain at low temperature can be understood in terms of the decrease in the bandwidth of the A1 -symmetry phonon mode, or the reduction of aT (see below), because g0 is not strongly dependent on temperature. Thus the reduction of thermal activation leads to improved transparency of THz waves as well as improved conversion eff iciency, which is significant in the higher-energy region. To verify the enhancement of the THz-wave output we investigated the cryogenic characteristics of the OPO by use of the experimental setup shown in Fig. 2.6 The pump source was a Q-switched Nd:YAG laser (wavelength, 1.064 mm; pulse width, 7 ns; repetition rate; 16.7 Hz; spot size, 1.2 mm). A 5-mm-thick LiNbO3 z plate was cut 75 mm long along the x axis. Two surfaces in the x plane were cut parallel, polished, and antiref lection coated for operation at 1.07 mm. The LiNbO3 crystal with a Si prism coupler6 was placed inside the compact cryostat, which had antiref lectioncoated BK7 windows for near-infrared lights (pump and idler) and a TPX (4-methyl penten-1) window for THz waves. Mirrors M1 and M2 were high-ref lection (HR) coated in the half-area and formed an external resonator for the idler wave. Since absorption or scattering loss inside LiNbO3 is small for idler waves, the idler power well ref lects the gain or conversion efficiency and was monitored with a pyroelectric detector through mirror M2. The THz-wave output was measured by a 4-K Si bolometer, and its wavelength was measured by a scanning Fabry–Perot etalon with metallic mesh.13 During the measurements from 297 to 78 K, no optical damage owing to cooling was observed in the crystal, and the angle-tuning characteristics were in good agreement with the theoretical curves.8,14 Figure 3 shows the input –output characteristics of the coherent THz-wave at various temperatures with a fixed incident angle of the pump beam (1.55±). The measured THz wavelength shifted slightly from 184 to 173 mm owing to changes of the refractive index of LiNbO3 . The output increased monotically as the temperature decreased, and we achieved 125-times-higher THz-wave output at a pump energy s11.4 mJypulsed and 32% reduction of the OPO threshold by cooling the crystal to 78 K. The maximum output of the THz wave was measured to be ,50 pJypulse (peak power, 7.2 mW), which we calibrated by using a bolometer sensitivity of 5.72 3 103 sVyWd. The conversion efficiency was found through measurement of the idler power to increase by eight times.
203
Fig. 1. Calculated parametric gain at 300 and 80 K with variation of pump intensity, Ip . The pump wavelength, lp , is fixed at 1.064 mm. The enhancement of the gain at lower temperature is due mainly to the reduction of the absorption coefficient in the THz region.
Fig. 2. Schematic diagram of the experimental apparatus. A Si prism is used as an output coupler for the THz wave, and the output is detected by a 4-K Si bolometer. Cooling the crystal to liquid-nitrogen temperature is possible with a Dewar. The phase-matching condition among the pump, the idler, and the THz waves is shown at the bottom. See text for definitions.
Fig. 3. Temperature dependence of the variation of the THz-wave output with input pump energy. The output is enhanced by 125 times and the threshold is reduced by 32% at 78 K compared with the values obtained at room temperature.
204
OPTICS LETTERS / Vol. 24, No. 4 / February 15, 1999
Fig. 4. Absorption coefficient in the 1 – 2-THz region at 296 and 78 K, measured by use of a widely tunable coherent THz-wave source. The solid (78 K) and the dashed (300 K) theoretical curves were calculated from the A1 -symmetry polariton wave vector.
We further investigated the absorption coefficient aT of LiNbO3 . The present THz OPO scheme provides direct loss measurement at THz frequency. We achieved this measurement by simply shifting the crystal along the y axis to change the propagation distance of the THz wave (see Fig. 2). Since the absorption coeff icient of Si in the THz region is relatively small s,0.6 cm21 d and has negligible frequency dependence, that of LiNbO3 , we determined aT by measuring the THz-wave output as a function of the propagation distance to the crystal edge in the x z plane. The THz-wave power should be proportional to expf2saT ysin ddyg, where d denotes the phasematching angle between the idler and the THz wave vectors. Figure 4 shows the measured aT of LiNbO3 in the 1–2-THz region at 296 and 78 K, where relatively precise measurements were accomplished, in contrast with measurements obtained with other methods, such as a Fourier transform infrared spectrometer with a weak source in the THz region. This result clearly shows a decrease of the absorption coefficient by nearly a factor of 3 and is in fairly good agreement with theoretical curves calculated from the complex polariton wave vector [Eq. (3)]. The reduction mechanism of the absorption coeff icient could be well understood in terms of the decrease of the bandwidth G of the lowest A1 -symmetric phonon mode, v0 . From Eq. (3), it is easily shown that well below v0 the absorption coeff icient aT is approximately proportional to the bandwidth G, and it is known to be reduced by nearly a factor of 3 by cooling of the crystal to liquid-nitrogen temperature.10 The measured 125-times enhancement of THz-wave power is interpreted as follows: The effect of the reduced loss of THz-wave propagation is expressed by the factor exph2fsaT ysin ddT297 K 2 saT ysin ddT 78 K gDyj, where Dy denotes the propagation distance to the crystal edge (x z plane) along the y axis. Using the measured aT (Fig. 4) and considering the conf iguration of our experimental setup (Fig. 2), we estimated this factor to be ,12. As mentioned above, the enhancement
of THz-wave output by improvement of conversion efficiency was measured to be eight times, so the total contribution to the power is concluded to be ,100. Thus the 125-enhancement is well explained by the enhanced gain as well as by the reduced absorption coeff icient. To study the enhancement mechanism further, it is important to measure the gain itself, which we are currently doing. In conclusion, we have investigated the cryogenic characteristics of THz-wave generation from a LiNbO3 OPO and demonstrated theoretically and experimentally that cooling was very effective for obtaining large enhancement of THz-wave output as well as for reduction of the threshold. Our THz-wave source is expected to play an important role in various applications, owing to advantages such as coherency, wide tunability, and compactness of its system. Further study is required for higher efficiency, narrower linewidth, and establishment of cw operation by use of domaininverted structures.15,16 The authors are greatly indebted to C. Takyu and T. Shoji for their excellent coating and polishing of the crystal and mirrors. This work was partly supported by a grant-in-aid from the Ministry of Education, Science and Culture of Japan and from the Research Foundation for Opto-Science and Technology. J. Shikata’s e-mail address is
[email protected]. *Present address, Graduate School of Engineering, Yamagata University, Yonezawa 992-8510, Japan. References 1. D. H. Auston, K. P. Cheung, and P. R. Smith, Appl. Phys. Lett. 45, 284 (1984). 2. D. M. Mittleman, R. H. Jacobsen, and M. C. Nuss, IEEE J. Sel. Topics Quantum Electron. 2, 679 (1996). 3. Q. Wu, T. D. Hewitt, and X.-C. Zhang, Appl. Phys. Lett. 69, 1026 (1996). 4. H. Ito, K. Kawase, and J. Shikata, Inst. Electron. Inf. Commun. Eng. E81-C, 264 (1998). 5. K. Kawase, M. Sato, K. Nakamura, T. Taniuchi, and H. Ito, Appl. Phys. Lett. 71, 753 (1997). 6. K. Kawase, M. Sato, T. Taniuchi, and H. Ito, Appl. Phys. Lett. 68, 2483 (1996). 7. M. A. Piestrup, R. N. Fleming, and R. H. Pantell, Appl. Phys. Lett. 26, 418 (1975). 8. D. R. Bosomworth, Appl. Phys. Lett. 9, 330 (1966). 9. Y. R. Shen, The Principles of Nonlinear Optics (Wiley, New York, 1984). 10. W. D. Johnston, Jr., and I. P. Kaminow, Phys. Rev. 168, 1045 (1968). 11. S. S. Sussman, ‘‘Tunable light scattering from transverse optical modes in lithium niobate,’’ Microwave Lab. Rep. 1851 (Stanford University, Stanford, Calif., 1970). 12. C. H. Henry and C. G. B. Garrett, Phys. Rev. 171, 1058 (1968). 13. K. Sakai, T. Fukui, Y. Tsunawaki, and H. Yoshinaga, Jpn. J. Appl. Phys. 8, 1046 (1969). 14. G. J. Edwards and M. Lawrence, Opt. Quantum Electron. 16, 373 (1984). 15. K. Kawase and H. Ito, Nonlinear Opt. 7, 225 (1994). 16. Y. J. Ding and J. B. Khurgin, Opt. Commun. 148, 105 (1998).
