APPLIED PHYSICS LETTERS 93, 262102 共2008兲

Drift current dominated terahertz radiation from InN at low-density excitation K. I. Lin,1,a兲 J. T. Tsai,1 T. S. Wang,1 J. S. Hwang,1,b兲 M. C. Chen,2 and G. C. Chi3 1

Department of Physics, National Cheng Kung University, Tainan 701, Taiwan Institute of Nuclear Energy Research, Longtan, Taoyuan 325, Taiwan 3 Institute of Optical Science, National Central University, Jhongli, Taoyuan 320, Taiwan 2

共Received 2 November 2008; accepted 5 December 2008; published online 30 December 2008兲 This letter investigates the polarity of terahertz radiation from indium nitride 共InN兲 excited by femtosecond optical pulses wherein a central wavelength of around 790 nm is measured. The InN epilayers are grown by metalorganic chemical vapor deposition on sapphire and silicon substrates. The polarity of the terahertz radiation field from InN is opposite to that from p-InAs whose radiation mechanism is dominated by the photo-Dember effect indicating that the dominant radiation mechanism in InN is the drift current induced by the internal electric field at low-density excitation below 590 nJ/ cm2. The internal electric field consists of the surface accumulation field and the spontaneous polarization-induced electric field. In addition, since no azimuthal angle dependence of the terahertz radiation is observed, the optical rectification effect is ruled out. By comparing the wave forms of terahertz radiation from the front and the back of the InN sample grown on sapphire in reflection geometry, the N polarity of the InN sample is confirmed. © 2008 American Institute of Physics. 关DOI: 10.1063/1.3056635兴 Indium nitride 共InN兲 is an important component of the group-III nitride system and has recently received considerable attention due to the lowest effective mass, the highest mobility, and the highest saturation velocity of the group-III nitrides.1 The exact value of the fundamental band gap of InN is still under debate. In contrast to the previous value of 2 eV, a much smaller value close to 0.7 eV has been suggested.1–3 These properties make InN a candidate for terahertz sources of narrow band gap semiconductors.3–6 In addition to the band gap, many important properties of InN are not well established to date. For example, the terahertz emission from InN has been demonstrated by Ascázubi et al.,3 who attributed the terahertz generation mechanism to transient photocurrents but did not specify the exact mechanism. Recently, many reports indicate that the transient photocurrents in InN are dominated by the photo-Dember effect.4–6 Even the resonance-enhanced optical rectification in InN under very high-excitation fluence 共⬎2 mJ/ cm2兲 is reported.7 It has also been reported that semi-insulating InP generates terahertz waves inducted by a built-in surface field 共drift current兲 at low-excitation density and conversely dominated by the photo-Dember effect 共diffusion current兲 at highexcitation density.8 More recently, a significant enhancement in terahertz emission from a-plane InN, whose radiation mechanism is attributed to the acceleration of photoexcited carriers under the polarization-induced in-plane electric field, is observed.5 Unlike conventional III-V semiconductor compounds, wurtzite III-nitride semiconductors possess a significant spontaneous polarization and an even larger piezoelectric polarization under strain, leading to an internal electric field.9 Up to now, there has been no report of terahertz radiation in c-plane InN showing the mechanism of drift currents at low-density excitation. Therefore, the understanding of the terahertz mechanism in InN at low-density excitation is necessary and important. a兲

Electronic mail: xlightគ[email protected]. Electronic mail: [email protected].

b兲

0003-6951/2008/93共26兲/262102/3/$23.00

In this letter, InN epilayers are used as terahertz emitters grown by metalorganic chemical vapor deposition 共MOCVD兲 on sapphire and silicon substrates. Surface morphologies and optical properties of the InN epilayers are characterized by atomic force microscope 共AFM兲 and photoluminescence 共PL兲 measurements. A femtosecond pulse laser with a central wavelength of around 790 nm is utilized as the excitation source of terahertz radiation. The polarity of the terahertz radiation field from InN is opposite to that from p-InAs whose radiation mechanism is dominated by the photo-Dember effect. This indicates that the dominant radiation mechanism in InN is the drift current induced by the surface accumulation field and the spontaneous polarization field at low-density excitation below 590 nJ/ cm2. Since no azimuthal angle dependence of the terahertz radiation is observed, an optical rectification effect is ruled out. In addition, the N polarity of the InN epilayer is confirmed. The InN epilayers are grown on c-plane sapphire and Si共111兲 substrates by MOCVD with a vertical reactor at atmospheric pressure. For the InN-A sample, a 3 nm lowtemperature 共LT兲 grown gallium nitride 共GaN兲 共growth temperature TG = 550 ° C兲 layer and a 5 nm high-temperature 共HT兲 grown GaN 共TG = 1120 ° C兲 layer are deposited on sapphire substrates as a buffer layer. An In thin layer is predeposited at 450 ° C prior to the growth of InN layer. The In layer thickness of about 1 nm is estimated based on the deposition rate. Finally, an InN layer with a total thickness of approximately 220 nm is grown at 500 ° C. For the InN-B sample, a 3 nm LT grown InN 共TG = 450 ° C兲 layer is deposited on Si共111兲 substrates as a buffer layer. After the LT InN growth, the temperature is raised to 550 ° C to grow HT InN. The thickness of the HT InN layer is approximately 150 nm. Detailed growth procedures are the same as those previously described.10,11 The critical thickness of an InN well on an In0.7Ga0.3N barrier, whose lattice mismatch is smaller than that of sapphire or Si共111兲, is about 1 nm.12 This implies that the InN epilayers of our samples are fully relaxed. Surface morphologies of the InN samples are characterized by AFM.

93, 262102-1

© 2008 American Institute of Physics

Downloaded 06 Jan 2009 to 117.32.153.167. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

262102-2

Lin et al.

FIG. 1. Time-domain wave forms of terahertz emission from p-InAs, SIN-InAlAs, p-GaAs, InN-A, and InN-B samples with 310 nJ/ cm2 excitation density. The wave forms of p-InAs and InN-B samples are multiplied by 1 / 3 and 5, respectively.

Both InN samples show typical grainlike morphology often observed in N-polar InN.13 To further confirm the polarity of InN, each InN epilayer is etched by a 10 mol/ l KOH solution for 80 min at room temperature. Both are easily etched and the observed surface morphologies after etching are typical of N-polar samples reported on the literature.13,14 The p-type InAs sample 共p-InAs兲 is a 共100兲 oriented wafer with a doping concentration at 3 ⫻ 1016 cm−3. The In0.52Al0.48As surface intrinsic-n+ 共SIN-InAlAs兲 structure consists of a 100 nm undoped layer on the top of a 1 ␮m Si-doped, n-type InAlAs buffer layer that has been grown previously on an Fe-doped semi-insulated 共100兲 InP substrate.15 The doping concentration in the buffer layer is approximately 1 ⫻ 1018 cm−3. A p-type GaAs wafer 共p-GaAs兲 is also used for comparison. A standard experiment setup using a free-space copropagating electro-optic sampling system is employed in this study.15,16 Terahertz radiation is detected in reflection geometry. A mode-locked Ti:sapphire laser operated at 82 MHz is used to generate 80 fs pulses with a central wavelength around 790 nm. The incident angle of the pump laser on the sample is 45° from normal. The pump beam is uniformly focused on the sample surface in s-polarization over an area with a radius of 0.5 mm. The maximum excitation laser power is 400 mW. The terahertz radiation is collected by a pair of parabolic mirrors and focused on a 2 mm thick ZnTe crystal. The electric field of the terahertz wave form in the far field is given by Eterahertz ⬀ dJ / dt, where J is the transient current induced by the optical pulse excitation.8 The polarity of the terahertz wave form is determined by the direction of the photocurrent. Figure 1 depicts time-domain wave forms of terahertz emission from p-InAs, SIN-InAlAs, p-GaAs, InN-A, and InN-B samples. The excitation density is 310 nJ/ cm2. The terahertz wave forms from p-InAs and SIN-InAlAs have the same polarity, and that from p-GaAs, InN-A, and InN-B are in the reverse direction. It is well known that the radiation mechanism of p-InAs is dominated by the photo-Dember effect. The diffusion current 共or the

Appl. Phys. Lett. 93, 262102 共2008兲

FIG. 2. Peak amplitudes 共solid circles兲 of the terahertz emission from the InN-A sample vs pump fluence. The solid line is a linear fit to the data. The lower and the upper insets show the terahertz wave form and the Fourier transformed spectrum of InN-A, respectively, excited at 590 nJ/ cm2.

photo-Dember field兲 flows towards the surface of p-InAs. For SIN-InAlAs, the drift current 共or the built-in electric field兲 also flows towards the surface of the sample.15 Using the direction of the photo-Dember field in p-InAs that points outward to the surface as a reference to determine the field orientation, we find that the internal electric field in both the InN layers is antiparallel to the electric field in p-InAs and SIN-InAlAs and parallel to the surface depletion field in p-GaAs.16–18 This indicates that the dominant radiation mechanism in InN is the drift current induced by the internal electric field at low-density excitation. However, electron accumulation layer at the surface of InN is very thin 共⬍10 nm兲 and its contribution to terahertz generation is small although the surface accumulation field is parallel to the electric field in p-GaAs.4–6 Unlike conventional III-V semiconductor compounds, wurtzite III-nitride semiconductors, such as InN, GaN, and AlN, possess a significant spontaneous polarization 共polarization at zero strain兲 and an even larger piezoelectric polarization under strain, leading to an internal electric field.9 Since the InN layers of our samples are fully relaxed, there is no piezoelectric polarization, and we conclude that the internal electric field in the InN layers results from spontaneous polarization. In N-polar InN, the direction of the spontaneous polarization is outward to the surface and the induced electric field has opposite direction in definition.9 This direction for the induced electric field is consistent with the foregoing inferences from the terahertz measurements. Peak amplitudes of the terahertz emission from the InN-A sample versus pump fluence are shown in Fig. 2. The solid line is a linear fit to the data. The peak amplitude of the terahertz emission increases linearly with the pump fluence and saturation of the emission is not observed below 590 nJ/ cm2. This suggests that our pump fluence is in the region of transient photocurrent generation.3 The lower and the upper insets show, respectively, the terahertz wave form and the Fourier transformed spectrum on a logarithmic scale of InN-A excited at 590 nJ/ cm2. The terahertz emission from InN-A essentially shows no azimuthal angle dependence, so an optical rectification effect is ruled out for this low-density excitation.3,5,8

Downloaded 06 Jan 2009 to 117.32.153.167. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

262102-3

Appl. Phys. Lett. 93, 262102 共2008兲

Lin et al.

FIG. 3. PL spectra of the InN-A and InN-B samples at 293 K. The solid lines are the Gaussian fit to the data.

PL measurements are used to characterize the optical properties and the free electron concentrations of the InN samples. Figure 3 displays the PL spectra of the InN-A and InN-B samples at 293 K. The solid lines show the Gaussian fit to the data. The PL peak positions are 0.730 and 0.811 eV and the full width at half-maximum values are 152 and 199 meV for InN-A and InN-B, respectively. The free electron concentrations are estimated as 1.4⫻ 1019 and 3.9⫻ 1019 cm−3 according to Ref. 2. This suggests that the weaker terahertz emission from InN-B shown in Fig. 1 is due to the higher free carrier absorption. The terahertz amplitude from InN-A is about six times smaller than for p-InAs due to large screening from a high density of background electrons. It is expected that the terahertz intensity from InN will improve with lower n-type bulk carrier concentration.4,17 InN grown on sapphire substrates with GaN buffer layers allows optical excitation at the InN surface or at the GaN / InN interface since both the substrates and the buffer layers are transparent at 790 nm and in the terahertz frequency range.3 A laser pulse with a wavelength of 790 nm penetrates InN to a depth of about 200 nm.3 Figure 4 shows the terahertz wave forms of the InN-A sample measured from the front 共the InN surface兲 and the back of the sample, respectively, using reflection geometry. There is no terahertz emission observed from the back of the InN-B sample due to the absorption of laser pulses by the Si substrate. The structural quality of InN is improving with increasing thickness. As a consequence the structural quality of the sample surface is much better in comparison to the InN layers close to the

FIG. 4. Terahertz wave forms of the InN-A sample measured from the front 共the InN surface兲 and the back side in reflection geometry, respectively. The arrows indicate the peaks for comparison.

InN-substrate interface. The gradient in structural perfection causes higher terahertz radiation by illuminating the surface in comparison to excitation through the substrate. In Fig. 4, the time delay between the two terahertz wave forms is mainly caused by the optical path difference in the sapphire substrate. Evidently, the polarity of the terahertz wave form from the back is opposite to that from the front. From the back, the direction of the spontaneous polarization points inward to the InN surface for N-polar InN and is just opposite to that in the front case and therefore causes the inverse terahertz polarity.8,9 This establishes that the InN-A sample is N polar and that the dominant mechanism of terahertz emission is drift current at low-density excitation. By extension, this technique is expected to allow determination of the direction of electric field in In-polar InN or the polarity of InN at low-density excitation. In conclusion, terahertz radiation is measured from InN excited by femtosecond optical pulses at 790 nm. The polarity of the terahertz radiation field from InN is opposite to that from p-InAs whose radiation mechanism is dominated by the photo-Dember effect. This indicates that the dominant radiation mechanism in InN is the drift current induced by the internal electric field at low-density excitation below 590 nJ/ cm2. The internal electric field consists of the surface accumulation field and the spontaneous polarization field. By comparing the wave forms of terahertz radiation from the front and back side of the InN-A sample, the InN epilayer is confirmed as of N polarity. This work was supported by the National Science Council of Taiwan under Contract Nos. NSC96-2112-M-006-020MY2 and NSC96-2112-M-006-017-MY3. 1

A. G. Bhuiyan, A. Hashimoto, and A. Yamamoto, J. Appl. Phys. 94, 2779 共2003兲. 2 S. P. Fu, T. T. Chen, and Y. F. Chen, Semicond. Sci. Technol. 21, 244 共2006兲. 3 R. Ascázubi, I. Wilke, K. Denniston, H. Lu, and W. J. Schaff, Appl. Phys. Lett. 84, 4810 共2004兲. 4 G. D. Chern, E. D. Readinger, H. Shen, M. Wraback, C. S. Gallinat, G. Koblmüller, and J. S. Speck, Appl. Phys. Lett. 89, 141115 共2006兲. 5 H. Ahn, Y. P. Ku, C. H. Chuang, C. L. Pan, H. W. Lin, Y. L. Hong, and S. Gwo, Appl. Phys. Lett. 92, 102103 共2008兲. 6 V. Cimalla, B. Pradarutti, G. Matthäus, C. Brückner, S. Riehemann, G. Notni, S. Nolte, A. Tünnermann, V. Lebedev, and O. Ambacher, Phys. Status Solidi B 244, 1829 共2007兲. 7 X. Mu, Y. J. Ding, K. Wang, D. Jena, and Y. B. Zotova, Opt. Lett. 32, 1423 共2007兲. 8 M. Nakajima, Y. Oda, and T. Suemoto, Appl. Phys. Lett. 85, 2694 共2004兲. 9 M. T. Hasan, A. G. Bhuiyan, and A. Yamamoto, Solid-State Electron. 52, 134 共2008兲. 10 K. J. Chang, J. Y. Chang, M. C. Chen, S. M. Lahn, C. J. Kao, Z. Y. Li, W. Y. Uen, and G. C. Chi, J. Vac. Sci. Technol. A 25, 701 共2007兲. 11 M. D. Yang, J. L. Shen, M. C. Chen, C. C. Chiang, S. M. Lan, T. N. Yang, M. H. Lo, H. C. Kuo, T. C. Lu, P. J. Huang, S. C. Hung, G. C. Chi, and W. C. Chou, J. Appl. Phys. 102, 113514 共2007兲. 12 S. B. Che, T. Mizuno, X. Wang, Y. Ishitani, and A. Yoshikawa, J. Appl. Phys. 102, 083539 共2007兲. 13 X. Wang, S. B. Che, Y. Ishitani, A. Yoshikawa, H. Sasaki, T. Shinagawa, and S. Yoshida, Appl. Phys. Lett. 91, 081912 共2007兲. 14 D. Muto, T. Araki, H. Naoi, F. Matsuda, and Y. Nanishi, Phys. Status Solidi A 202, 773 共2005兲. 15 J. S. Hwang, H. C. Lin, K. I. Lin, and X. C. Zhang, Appl. Phys. Lett. 87, 121107 共2005兲. 16 J. S. Hwang, H. C. Lin, Y. C. Huang, K. I. Lin, J. W. Chang, and T. F. Guo, Electrochem. Solid-State Lett. 11, H63 共2008兲. 17 K. Liu, J. Xu, T. Yuan, and X.-C. Zhang, Phys. Rev. B 73, 155330 共2006兲. 18 J. S. Hwang, H. C. Lin, C. K. Chang, T. S. Wang, L. S. Chang, J. I. Chyi, W. S. Liu, S. H. Chen, H. H. Lin, and P. W. Liu, Opt. Express 15, 5120 共2007兲.

Downloaded 06 Jan 2009 to 117.32.153.167. Redistribution subject to AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp