Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

Decay time extension of terahertz electromagnetic waves emitted from coherent longitudinal optical phonons in GaAs epitaxial layers with the use of fast atom bombardment

Open Access Open Access

Abstract

We succeeded in extending the decay time of terahertz electromagnetic waves from coherent longitudinal optical (LO) phonons in GaAs epitaxial layers with the use of fast atom bombardment, a treatment method for introducing defects and/or disorders at the surface. The decay time becomes long, up to 4.81 ± 0.15 ps, with the bombardment time of 4.0 min. This value is 2.4-times larger than the decay time of 2.04 ± 0.04 ps of the reference sample (untreated sample). We attribute the origin of the present phenomenon to the fact that the reduction of the photogenerated carrier scattering effect on the coherent LO phonon, which is caused by the presence of the surface defects and disorders formed by the fast atom bombardment, results in the extending of the decay time of the terahertz wave from the coherent LO phonon.

© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

1. Introduction

Recent progress with the development of femtosecond pulse laser technology brings to generating terahertz electromagnetic waves from coherent longitudinal optical (LO) phonons [1,2]. The coherent LO phonon is an oscillating longitudinal polarization leading the terahertz-wave emission. The frequency of the LO phonon is a semiconductor specific value, and, hence, the coherent LO phonon is suitable for monochromatic terahertz emitters. We, however, note that the terahertz frequency band has a finite bandwidth connecting with the decay time. Accordingly, if the decay time is extendable, the ultranarrow terahertz band is realized. In the generation of the coherent LO phonon, a dominant trigger is the phenomenon that an ultrafast photogenerated carrier flow (referred as an ultrafast photocurrent) screens the surface built-in electric field. This releases the initial amplitude corresponding to the polarization of the coherent phonon and the LO phonon oscillation starts, which emits the terahertz wave [3,4]. Taking account of the generation mechanism, the ultrafast photocurrent scattering of the coherent LO phonon is inevitable. We, here, reach a proposal; namely, the suppression of the ultrafast photocurrent scattering leads to the longer decay time of the coherent LO phonon. We reported that the reduction of the pump power density corresponding to the photogenerated carrier density causes the decay-time extension of the coherent LO phonons [5]. The above work, however, contains an issue that the coherent LO phonon signal is weakened as the pump power is reduced. In order to settle this issue, we explore how to suppress the ultrafast photocurrent scattering effect on the coherent LO phonon. In the research field of the optical reflection-type pump-probe spectroscopy, Hase et al. reported that, in an n-type GaAs crystal with the doping concentration of 1.4 × 1018 cm-3, the dephasing time of the coherent LO phonon slightly increases from 2.1 to 3.0 ps after the He ion implantation [6,7], They attributed the origin of this phenomenon to a weakening of the electron-phonon scattering on the grounds of the Raman scattering spectroscopic study [8]. Accordingly, the introduction of the defects has an ability of extending the decay time of the coherent LO phonon. We, on the other hand, point out the fact that, in the n-type GaAs crystal with the high background electrons, the surface depletion region is quite small: the depletion-region length is of the order of several nanometers. In addition, the free electrons strongly absorb the terahertz wave. Consequently, it is hopeless to obtain the intense terahertz wave in such material systems. The above consideration is supported by the earlier work [9]. In this work, we found that the GaAs-based epitaxial structures consisting of the top undoped layer (i-layer) and bottom n-type layer (n-layer) is effective to obtain the intense terahertz wave. We also note that, in a Bi thin film, the decay time of the coherent A1g phonon is shortened by the ion implantation [7]. The decay-time shortening of the coherent phonon is also observed in graphite [10] These facts imply that it is difficult to control of the decay time of the coherent phonon with the use of the ion implantation. Here, we focus our attention on the surface treatment of semiconductors. Fast atom bombardment onto the semiconductor surfaces, which is performed with the use of accelerated noble gas atoms such as Ar in a vacuum, introduces a disordered thin layer with the thickness of nanometer order [11]. Consequently, the bulk region of the epitaxial layer is hardly damaged. We mention that the present idea is in contrast to the well-known earlier work on Raman scattering spectroscopy [12]. In this work, the LO phonon Raman band is broadened by the presence of defects leading to the finite size effect: the relaxation of the wave-vector conservation.

 figure: Fig. 1.

Fig. 1. Potential structure of the present sample. The solid and dashed lines denote the conduction band bottom and Fermi energy, respectively.

Download Full Size | PDF

Tables Icon

Table 1. Correspondence table of the sample name and bombardment time

 figure: Fig. 2.

Fig. 2. AMF images of the sample surfaces in the area of 1.0 µm × 1.0 µm of sample Ref (a) and sample A4 (b).

Download Full Size | PDF

In the present paper, we demonstrate that the fast atom bombardment is effective to extending the terahertz-wave decay time of the coherent LO phonon in GaAs epitaxial structures consisting of the top i-layer and bottom n-layer. The decay time becomes long to be 4.81 ± 0.15 ps with the bombardment time of 4.0 min. This value is larger than the decay time of 2.04 ± 0.04 ps of the sample without the bombardment. We also discuss the decay time of the coherent LO phonon comparing with earlier works, and point it out that the decay time does not necessarily depend on the crystal quality but also depend on the sample pre-treatment.

2. Samples and experimental procedure

The present samples, which were grown by metal-organic vaper phase epitaxy, were GaAs-based epitaxial structures consisting of the top i-layer with the thickness of 50 nm and the bottom n-layer with the thickness and doping concentration of 3.0 µm and 3 × 1018 cm-3. The substrate was a (001)-oriented semi-insulating GaAs wafer. Figure 1 shows the potential structure of the present sample. The parameters used in the potential structure calculation were the same as those used in Ref. 13. In the i-layer, the linear potential slope is formed, which produces the initial displacement (the initial polarization) of the coherent LO phonon. We note that the electric filed strength is 140 kV/cm in the i-layer. Accordingly, the i-layer has an ability emitting the terahertz wave from the coherent LO phonon. The fast atom bombardment was performed with the use of Ar gas. The flow rate was 8.0 sccm. The acceleration voltage and current, which were applied to the generator of the fast atom beam, were 1.5 kV and 20 mA, respectively, and the incidence angle of the beam to the sample and the nominal accelerated Ar beam energy were 45° and 1.0 keV, respectively. The correspondence table of the sample name and bombardment time is listed in Table 1. The terahertz time-domain signals were measured using an optical gating method. The detector was a photoconductive dipole antenna formed on a low-temperature-grown GaAs epilayer. The pump and probe light source was a Ti:sapphire laser with pulse duration of 40 fs and a repetition rate of 90 MHz. The photon energy and pump-beam power were 1.55 eV and 100 mW, respectively. The beam profile was a Gaussian type. The diameter of the spot was about 100 µm on the sample surface. The terahertz time-domain measurements were performed at room temperature. We applied a nitrogen-gas purge to reducing the humidity less than about 5% for avoiding water-vapor absorption. The time-delay range was set from -1.0 ps to 7.0 ps.

3. Experimental results and discussion

In advance to measuring time-domain terahertz wave, we observed the surfaces of samples Ref and A4 using an atomic force microscope (AFM). The AFM images are shown in Figs. 2(a) and 2(b). The values of the root mean square (RMS), which correspond to the surface roughness, were estimated to be 0.22 nm for sample Ref and 0.23 nm for sample A4. This indicates that the present fast atom bombardment hardly causes the disorder of the crystal structure. Accordingly, the relaxation of the wave-vector conservation, which can become a trigger of reducing the decay time of the coherent phonon, is negligible.

 figure: Fig. 3.

Fig. 3. Time-domain terahertz wave of each sample. For clarity, each terahertz wave is vertically shifted.

Download Full Size | PDF

Next, we measured the terahertz time-domain signals of the present samples, the results of which are shown in Fig. 3. The bipolar signal, which is located at the time delay of 0 ps, originates from ultrafast photocurrent generated by the pump pulse [14]. The bipolar pulse follows the oscillation pattern with the period of 114 fs, which corresponds to the LO phonon frequency of 8.8 THz. The oscillation pattern remains up to more than 7 ps (the delay stage limit). The decay time of the terahertz wave from the coherent phonon apparently becomes longer as the fast atom bombardment time is longer. After the fast atom bombardment, it is reasonable to consider that the ultrafast photocurrent is suppressed owing to defects and/or disorders at the surface. Consequently, the scattering effect on coherent LO phonon becomes less, leading to the longer decay time of the coherent LO phonons. We analyzed the time-domain signals with the use of Fourier transform. The Fourier power spectra are shown in Fig. 4. The terahertz frequency band, which originates from the ultrafast photocurrent, is located at the frequency of 1.5 THz. The intensity of this band is reduced after the fast atom bombardment, meaning that the ultrafast photocurrent tends to be weakened. This result coincides with the above discussion on the longer decay time of the coherent LO phonon. The band at 8.8 THz is attributed to the coherent LO phonon. After the bombardment, the coherent LO phonon band widths become narrow. The band width is 0.19 THz in sample Ref, while the band width becomes narrow to be 0.11 THz in sample A4. In addition, the peak value of the band is increased. This reflects the decay time characteristics of the coherent LO phonon terahertz wave shown in Fig. 3. In the Fourier power spectrum of sample Ref without the bombardment, the band labeled by LOPC(-), which originates from coherent LO-phonon plasmon coupled (LOPC) mode of the lower branch [15,16], clearly appears at 4.8 THz. In contrast, the coherent LOPC(-) band disappears in the samples with the bombardment. In terahertz time-domain spectroscopy, only the coherent signal is observed in principle. The disappearance of the coherent LOPC mode band implies that the fast atom bombardment reduces the coherency of the coherent LOPC mode; namely, the plasmon (plasma oscillation) loses coherency. This results in the coherency of the coherent LOPC mode, the coupled oscillation, is suppressed though the decay time of the coherent LO phonon is extended.

 figure: Fig. 4.

Fig. 4. Fourier power spectra of the terahertz waveforms shown in Fig. 3.

Download Full Size | PDF

For precisely estimating the decay time τ of the terahertz wave from the coherent LO phonon, we performed inverse Fourier transform to extract only the oscillatory pattern of the coherent LO phonon. Figure 5 shows the waveform A(t) obtained with the use of the inverse Fourier transform. The solid line is the obtained waveform, which is fitted to the following equation:

$$A(t) = {A_0}\textrm{exp} ( - t/\tau )\sin (\omega t + \phi ).$$

Here, the quantities A0, ω, and ϕ are the initial amplitude, frequency, and initial phase, respectively. In Fig. 5, we indicate the decay time τ. The decay time is extended as the bombardment time is increases. In sample Ref, the decay time is 2.04 ± 0.04 ps, while the decay time becomes long to be 4.81 ± 0.15 ps. According to Ref. 17, the decay time of the GaAs LO phonon ranges from 2.0 ps to 4.0 ps. Cho et al. [18] reported that the decay time is about 2.0 ps, using ultrafast optical reflection-type pump-probe spectroscopy. We note that this value is the same as the decay time of the terahertz wave from coherent LO phonon in sample Ref without fast atom bombardment treatment. The present findings indicate that the decay time is controllable and extendable with the sample pre-treatment.

 figure: Fig. 5.

Fig. 5. Waveform component of the coherent LO phonon as a function of time delay obtained with the use of the inverse Fourier transform. The open circles denote the fitting results with the use of Eq. (1).

Download Full Size | PDF

Finally, we briefly describe the dependence of the decay time on the fast atom bombardment. We plotted the decay time of the terahertz wave from the coherent LO phonon as a function of bombardment time in Fig. 6. The decay time of the coherent LO phonon shows a sharp rise and tend to saturate. At present, we assume that the suppression of the ultrafast photocurrent scattering has a tendency of the saturation in the fast atom bombardment and that the effect of the present fast atom bombardment is limited at the surface of the GaAs epilayer. We could not perform cross-sectional transmission electron microscope (TEM) measurement. It is quite impossible to observe the region from the surface with the limited depth of the order of several nanometers because of the difficulty in the appropriate sample preparation for the TEM observation. We, therefore, could not assign the species of the defects at the surface to the reduction of the ultrafast photocurrent scattering. This is a remained issue in the present work.

 figure: Fig. 6.

Fig. 6. Decay time of the coherent LO phonon plotted as a function of fast atom bombardment time (open circle). The solid line is a guide for eyes.

Download Full Size | PDF

4. Summary

We have demonstrated that the decay time of the terahertz wave emitted from the coherent longitudinal optical (LO) phonons is extendable with the use of fast atom bombardment. We conclude that the present phenomenon results from the reduction of the ultrafast photocurrent scattering effect on the coherent LO phonon owing to the presence of the surface defects and/or disorders formed by the fast atomic bombardment. The present strategy provides a guideline to realizing ultranarrow terahertz frequency band.

Disclosures

The authors declare no conflicts of interest.

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

References

1. 1. P. Gu and M. Tani, “Terahertz radiation from semiconductor surfaces,” in Terahertz Optoelectronics, K. Sakai, Ed. (Springer, 2005), pp.63–97, and relevant references therein.

2. A. V. Kuznetsov and C. J. Stanton, “Coherent phonon oscillations in GaAs,” Phys. Rev. B 51(12), 7555–7565 (1995). [CrossRef]  

3. T. Dekorsy, G. C. Cho, and H. Kurtz, “Coherent phonon in condensed media,” in Light Scattering in Solids VIII, M. Cardona and G. Güntherodt, eds. (Springer, Berlin: 2000), Chapter 4, and relevant references therein.

4. T. Pfeifer, T. Dekorsy, W. Kiitt, and H. Kurz, “Generation mechanism for coherent LO phonons in surface-space-charge fields of III-V-compounds,” Appl. Phys. A 55(5), 482–488 (1992). [CrossRef]  

5. H. Takeuchi, T. Sumioka, and M. Nakayama, “Effects of photogenerated carrier scattering on the decay process of coherent longitudinal optical phonons in an undoped GaAs/n-type GaAs epitaxial structure investigated by terahertz time-domain spectroscopy,” J. Vac. Sci. Technol. A 35(4), 04D104 (2017). [CrossRef]  

6. M. Hase, K. Ishioka, and M. Kitajima, “Ultrafast carrier and plasmon-phonon dynamics in ion-irradiated n-GaAs,” Appl. Phys. Lett. 82(21), 3668–3670 (2003). [CrossRef]  

7. M. Hase and M. Kitajima, “Interaction of coherent phonons with defects and elementary excitations,” J. Phys.: Condens. Matter 22(7), 073201 (2010). [CrossRef]  

8. N. Dharmarasu, B. Sundarakkannan, R. Kesavamoorthy, K. G. M. Nair, and J. Kumar, “Investigations on H+ and He+ implantation effects in n-InP using Raman scattering,” Physica B Condens. Matter 262(3-4), 329–335 (1999). [CrossRef]  

9. H. Takeuchi, J. Yanagisawa, T. Hasegawa, and M. Nakayama, “Enhancement of terahertz electromagnetic wave emission from an undoped GaAs/n-type GaAs epitaxial layer structure,” Appl. Phys. Lett. 93(8), 081916 (2008). [CrossRef]  

10. K. Ishioka, M. Hase, M. Kitajima, and K. Ushida, “Ultrafast carrier and phonon dynamics in ion-irradiated graphite,” Appl. Phys. Lett. 78(25), 3965–3967 (2001). [CrossRef]  

11. E. Higurashi, K. Okumura, K. Nakasuji, and T. Suga, “Surface activated bonding of GaAs and SiC wafers at room temperature for improved heat dissipation in high-power semiconductor lasers,” Jpn. J. Appl. Phys. 54(3), 030207 (2015). [CrossRef]  

12. K. K. Tiong, P. M. Amirtharaj, and F. H. Pollak, “Effects of As+ ion implantation on the Raman spectra of GaAs: ‘Spatial correlation’ interpretation,” Appl. Phys. Lett. 44(1), 122–124 (1984). [CrossRef]  

13. H. Takeuchi, Y. Kamo, Y. Yamamoto, T. Oku, M. Totsuka, and M. Nakayama, “Photovoltaic effects on Franz-Keldysh oscillations in photoreflectance spectra: application to determination of surface Fermi level and surface recombination velocity in undoped GaAs/n-type GaAs epitaxial layer structures,” J. Appl. Phys. 97(6), 063708 (2005). [CrossRef]  

14. X.-C. Zhang and D. Auston, “Optoelectronic measurement of semiconductor surfaces and interfaces with femtosecond optics,” J. Appl. Phys. 71(1), 326–338 (1992). [CrossRef]  

15. I. Yokota, “On the coupling between optical lattice vibrations and carrier plasma oscillations in polar semiconductors,” J. Phys. Soc. Jpn. 16(10), 2075 (1961). [CrossRef]  

16. A. Mooradian and G. B. Wright, “Observation of the interaction of plasmons with longitudinal optical phonons in GaAs,” Phys. Rev. Lett. 16(22), 999–1001 (1966). [CrossRef]  

17. D. K. Ferry, “Non-equilibrium longitudinal optical phonons and their lifetimes,” Appl. Phys. Rev. 8, 021324 (2021). [CrossRef]  The decay time range is indicated in Fig. 24 in this reference.

18. G. C. Cho, W. Kütt, and H. Kurz, “Subpicosecond time-resolved coherent-phonon oscillations in GaAs,” Phys. Rev. Lett. 65(6), 764–766 (1990). [CrossRef]  

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (6)

Fig. 1.
Fig. 1. Potential structure of the present sample. The solid and dashed lines denote the conduction band bottom and Fermi energy, respectively.
Fig. 2.
Fig. 2. AMF images of the sample surfaces in the area of 1.0 µm × 1.0 µm of sample Ref (a) and sample A4 (b).
Fig. 3.
Fig. 3. Time-domain terahertz wave of each sample. For clarity, each terahertz wave is vertically shifted.
Fig. 4.
Fig. 4. Fourier power spectra of the terahertz waveforms shown in Fig. 3.
Fig. 5.
Fig. 5. Waveform component of the coherent LO phonon as a function of time delay obtained with the use of the inverse Fourier transform. The open circles denote the fitting results with the use of Eq. (1).
Fig. 6.
Fig. 6. Decay time of the coherent LO phonon plotted as a function of fast atom bombardment time (open circle). The solid line is a guide for eyes.

Tables (1)

Tables Icon

Table 1. Correspondence table of the sample name and bombardment time

Equations (1)

Equations on this page are rendered with MathJax. Learn more.

A ( t ) = A 0 exp ( t / τ ) sin ( ω t + ϕ ) .
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.