Picosecond carrier lifetime and large optical ... - OSA Publishing

108

OPTICS LETTERS / Vol. 22, No. 2 / January 15, 1997

Picosecond carrier lifetime and large optical nonlinearities in InGaAsP grown by He-plasma-assisted molecular beam epitaxy L. Qian, S. D. Benjamin, and P. W. E. Smith Department of Electrical and Computer Engineering, University of Toronto, and Ontario Laser and Lightwave Research Centre, 10 King’s College Road, Toronto, Ontario M5S 3G4, Canada

B. J. Robinson and D. A. Thompson Centre for Electrophotonic Materials and Devices, McMaster University, Hamilton, Ontario, L8S 4L7, Canada Received September 19, 1996 We report the measurement of a fast carrier lifetime and large band-gap-resonant optical nonlinearities in an InGaAsP sample grown by He-plasma-assisted molecular beam epitaxy. Using a 2-mm-thick sample grown on an InP substrate, we observed a carrier lifetime of 15 ps and an index change as large as 0.077 induced by an intense 1-ps pulse at a wavelength of 1.57 mm. Good crystalline structure is maintained in the material during growth, and the absorption spectrum shows a sharp band edge. These properties indicate that materials produced by He-plasma-assisted growth have potential applications in compact ultrafast photonic devices.  1997 Optical Society of America

Materials that exhibit large optically induced index changes and ultrafast optical response have attracted much research interest in recent years because of their potential applications in ultrafast all-optical devices for optical switching and signal processing.1 Semiconductors, having large optical nonlinearities in the vicinity of their band edge, are desirable for these types of photonic device because they can easily be integrated with other devices. However, a subpicosecond carrier lifetime is required in order to achieve a terahertz switching or processing rate, which is compatible with the transmission rate of an optical f iber. Various methods have been employed to reduce the carrier lifetime of semiconductors from nanoseconds toward subpicoseconds by introduction of deep-level trapping –recombination centers in the band gap. Among these methods, impurity doping,2 proton bombardment,3 and ion implantation,4 although successful in reducing carrier lifetime, also cause a high level of damage to the materials that results in high absorption associated with poor crystalline quality or in a reduction of the magnitude of the optical nonlinearity. A popular technique that has emerged recently involves growing the material at a much lower substrate temperature than the normal growth temperature. Extensive studies and measurements performed on low-temperature-grown GaAs (see, for example, Refs. 5 –7) have shown that the large concentration of defects that result from clusters of excess arsenic formed during low-temperature growth and subsequent annealing is responsible for the ultrashort carrier lifetime. A range of carrier lifetimes from subpicoseconds to many tens of picoseconds has been observed under various annealing conditions. However, a high absorption is also associated with 0146-9592/97/020108-03$10.00/0

low-temperature-grown materials, mainly because of reabsorption of light by the carriers trapped in the defects.8,9 Another disadvantage of low-temperature growth is that, if more layers are to be grown on top of the low-temperature layer, these layers will also undergo annealing and possible associated changes in properties. In a previous paper10 we reported a unique growth technique— He-plasma-assisted gas source molecular beam epitaxy growth— that yields a material of high resistivity, a characteristic property of a material expected to have an ultrashort carrier lifetime. In this Letter we demonstrate that this growth technique results in a significant reduction of the carrier lifetime of a quaternary InGaAsP sample while the large band-gap-resonant nonlinearity of the material is maintained and no significant additional absorption is introduced. Our interest in the quaternary InGaAsP derives from the fact that the band-gap energy can be engineered to correspond to a photon wavelength of 1.55 mm, which is in one of the important f iber communication wavelength zones. The sample was grown by conventional gas source molecular beam epitaxy at 435 ±C, while the growth surface was simultaneously exposed to a He-plasma stream generated by electron cyclotron resonance. For comparison, two additional InGaAsP samples were grown. Motivated by a recent report,11 we grew one sample without He plasma at a lower substrate temperature estimated to be 200 ±C and doped it with Be at a concentration of 9 3 1017 cm23 . The second comparative sample was grown without the plasma at a temperature similar to that for the He-plasma sample. All three samples were 2 mm thick and were grown on InP substrates. The laser source used for both the carrier lifetime measurement and the index  1997 Optical Society of America

January 15, 1997 / Vol. 22, No. 2 / OPTICS LETTERS

Fig. 1. (a) Absorption spectra of InGaAsP samples at wavelengths spanning the band edge. The lowtemperature-grown sample is of poor crystalline quality, as evidenced by the lack of an identifiable band edge. ( b) Enlargement of (a) near the band edge. Based on their almost identical absorption curves, the crystalline quality of the He-plasma-grown sample is comparable with that of the standard growth sample. (Data for the standard and He-plasma samples have been modif ied to compensate for the Fabry – Perot effects.)

change measurement was an optical parametric oscillator that produces 1-ps transform-limited pulses tunable from 1.44 to 1.64 mm.12 The linear absorption spectra of the three samples [Fig. 1(a)] were examined with a Fouriertransform infrared spectrometer. The low-temperature growth produced a material without an identifiable band edge and with substantial absorption out to 5 mm. The apparent poor crystalline quality of this material probably is due to the growth temperature’s being too low. Although it may be possible for one to improve the material properties by growing crystals at slightly higher temperatures, this growth technique was not pursued because of the diff iculty in measuring and controlling the substrate temperature at such temperatures. On the other hand, the He-plasma-assisted growth produces InGaAsP of high crystalline quality with a sharp band edge aligned with that of standard

109

material, as shown in Fig. 1(b). Since the sharpness of the band edge is comparable with that of the standard grown material, we would expect the magnitude of the carrier-induced index changes to be maintained as well. This shows one of the major advantages of the He-plasma-assisted growth over most other techniques and has significant device implications: It allows one to design devices to be operated at a wavelength near the band tail of the material, where the absorption is small but a significant index change can still be obtained. The carrier lifetime of the He-plasma sample was obtained by a standard pump–probe measurement of the recovery of absorption after saturation by a 1-ps pump pulse at 1.55 mm, as shown in Fig. 2. The results indicate that the absorption recovers with a single exponential lifetime of 15 ps as a result of the decay of excited carrier population. (The small peak near 6 ps is due to back surface ref lection of the pump pulse.) This decay is dramatically shorter than that for the standard sample, which was measured to be .500 ps. The short carrier lifetime is attributed to defects in the material created by the impact of energetic He atoms in the near-surface region of the growing sample. This response time may be suitable for some ultrafast all-optical switching devices. Furthermore, our preliminary results with Be-doped samples grown by He-plasma-assisted growth indicate that substantially shorter carrier lifetimes can be obtained. Using a standard Z-scan technique,13 we measured the average index changes of the He-plasma sample during the 1-ps pump pulse and compared the results with that of the standard sample over a range of photon f luences and wavelengths (Fig. 3). An index change Dn of as much as 20.077 was observed at l ­ 1.57 mm for the He-plasma sample. Under high photon f luence, the nonlinearity of the He-plasma sample induced by the short pulses is comparable with that of the standard sample at 1.57 mm. At 1.61 mm, the index change of the He-plasma sample is actually larger than that of the standard sample, presumably because of the additional absorption in the band tail.

Fig. 2. Standard pump – probe measurement of change in transmission through the 2-mm-thick InGaAsP sample grown with He plasma using 1-ps pulses at 1.55 mm showing a 15-ps carrier lifetime.

110

OPTICS LETTERS / Vol. 22, No. 2 / January 15, 1997

cations wavelength region near 1.55 mm. Its short lifetime also makes this material a promising candidate for saturable absorbers for mode-locking semiconductor lasers. Because the He-plasma growth can be combined with previous and subsequent normal growth steps, the realization of complex heterostructures for optoelectronic integrated circuits should be more feasible with this technique than with lowtemperature techniques for growing ultrafast materials. We are currently exploring the optical properties of samples grown under various He-plasma conditions and doping levels in order to produce samples with substantially shorter carrier lifetimes.

Fig. 3. Comparison of index changes induced in the Heplasma-assisted growth and standard growth InGaAsP samples by 1-ps laser pulses of varying intensity and at a series of wavelengths that span the band edge. Points are actual data; curves are guides for the eye.

An index change as large as 0.030 was obtained at 1.61 mm, where the absorption is significantly lower [see Fig. 1(b)]. The large index change with low absorption has much practical significance in the application of photonic devices. We have also compared the measured carrierinduced index changes of the standard sample at ´ 1.57 mm with that calculated by the Banyai–Koch model.14 Here we convert the photon f luence to carrier density for comparison by measuring the absorption of pump light during the experiment. The large carrier-induced index changes under high photon f luence are in agreement with the model. At low carrier densities (,1.0 3 1018 cm23 ), our measurement shows larger index changes than the model calculates. Comparison of the experimental measurements with ´ the calculation by the Banyai– Koch model is not suitable at longer wavelengths (1.59 and 1.61 mm), where there is significant two-photon-absorption contribution to the index change that is not included in the model. In summary, we have observed large, ultrafast nonlinearities in an InGaAsP sample grown by Heplasma-assisted molecular beam epitaxy. The fact that large and rapidly recovering optically induced index changes can be induced in InGaAsP produced by He-plasma-assisted growth indicates that this material holds promise for future ultrafast switching components operating in the important telecommuni-

The authors thank J. E. Ehrlich for his involvement during the early stage of this research and S. A. McMaster for molecular beam epitaxy machine operation. This research was funded by the Ontario Center for Material Research, the Ontario Laser and Lightwave Research Center, and the Natural Sciences and Engineering Research Council. References 1. See, for example, Y. Silberberg and P. W. E. Smith, in Nonlinear Photonics, H. M. Gibbs, G. Khitrova, and N. Peyghambarian, eds. (Springer-Verlag, Berlin, 1990), p. 185. 2. See, for example, C. H. Lee, A. Antonelli, and G. Mourou, Opt. Commun. 21, 158 (1977). 3. See, for example, Y. Silberberg, P. W. Smith, D. A. B. Miller, B. Tell, A. C. Gossard, and W. Wiegmann, Appl. Phys. Lett. 46, 701 (1985). 4. See, for example, F. E. Doany, D. Grischkowsky, and C. C. Chi, Appl. Phys. Lett. 50, 460 (1987). 5. S. Gupta, J. F. Whitaker, and G. A. Mourou, IEEE J. Quantum Electron. 28, 2464 (1992). 6. E. S. Harmon, M. R. Melloch, J. M. Woodall, D. D. Nolte, N. Otsuka, and C. L. Chang, Appl. Phys. Lett. 63, 2248 (1993). 7. S. D. Benjamin, A. Othonos, and P. W. E. Smith, Electron. Lett. 30, 1704 (1994). 8. Y. Kostoulas, L. J. Waxer, I. A. Walmsley, G. W. Wicks, and P. M. Fauchet, Appl. Phys. Lett. 66, 1821 (1995). 9. S. D. Benjamin, H. S. Loka, A. Othonos, and P. W. E. Smith, Appl. Phys. Lett. 68, 2544 (1996). 10. D. B. Mitchell, B. J. Robinson, D. A. Thompson, L. Qian, S. D. Benjamin, and P. W. E. Smith, Appl. Phys. Lett. 69, 509 (1996). 11. R. Takahashi, Y. Kawamura, T. Kagawa, and H. Iwamura, Appl. Phys. Lett. 65, 1790 (1994). 12. L. Qian, S. D. Benjamin, and P. W. E. Smith, Opt. Commun. 127, 73 (1996). 13. M. Sheik-Bahae, A. A. Said, T. H. Wei, D. J. Hagan, and E. W. VanStryland, IEEE J. Quantum Electron. 26, 760 (1990). 14. L. Banyai ´ and S. W. Koch, Z. Phys. B 63, 283 (1986).