Free carrier dynamics of InN nanorods investigated by time-resolved terahertz
spectroscopy
H. Ahn, C.-H. Chuang, Y.-P. Ku, and C.-L. Pan
Citation: Journal of Applied Physics 105, 023707 (2009); doi: 10.1063/1.3068172
View online: http://dx.doi.org/10.1063/1.3068172
View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/105/2?ver=pdfcov Published by the AIP Publishing
Articles you may be interested in
Terahertz electroluminescence of surface plasmons from nanostructured InN layers Appl. Phys. Lett. 96, 183106 (2010); 10.1063/1.3425897
Carrier and phonon dynamics of wurtzite InN nanorods Appl. Phys. Lett. 94, 071911 (2009); 10.1063/1.3086888
Time resolved measurements of spin and carrier dynamics in InAs films J. Appl. Phys. 103, 064318 (2008); 10.1063/1.2899091
Time-resolved spectroscopy of recombination and relaxation dynamics in InN Appl. Phys. Lett. 83, 4984 (2003); 10.1063/1.1633973
Carrier relaxation and recombination in GaAs/AlGaAs quantum heterostructures and nanostructures probed with time-resolved cathodoluminescence
J. Appl. Phys. 81, 3186 (1997); 10.1063/1.364148
Free carrier dynamics of InN nanorods investigated by time-resolved
terahertz spectroscopy
H. Ahn,a兲C.-H. Chuang, Y.-P. Ku, and C.-L. Pan
Department of Photonics and Institute of Electro-optical Engineering, National Chiao Tung University, Hsinchu 30010, Taiwan
共Received 13 August 2008; accepted 3 December 2008; published online 23 January 2009兲 Ultrafast time-resolved terahertz spectroscopy is employed to investigate the carrier dynamics of indium nitride 共InN兲 nanorod arrays and an epitaxial film. Transient differential transmission of terahertz wave shows that hot carrier cooling and defect-related nonradiative recombination are the common carrier relaxation processes for InN film and nanorods. However, the electrons confined in the narrow structure of nanorods are significantly affected by the carrier diffusion process near the surface, which causes the abnormally long relaxation time for nanorods. © 2009 American Institute
of Physics.关DOI:10.1063/1.3068172兴
I. INTRODUCTION
Due to its intrinsic narrow bandgap and remarkably large energy difference between the conduction band minimum and the next local minimum, indium nitride 共InN兲 inspires potential applications in the terahertz range. Recently, low-dimensional InN nanomaterials in the forms of nanowires, nanorods, nanotubes, etc., have received great attention due to their importance in near-infrared optoelectronics and pho-tovoltaic applications. Unambiguously, carrier dynamics in these nanomaterials is one of the most important aspects of material properties for optimized performance of optoelec-tronic applications. Carrier dynamics in semiconductors is typically investigated by using optical pump-probe tech-nique, of which the photon energy of the probe is of the order of the bandgap energy of the samples. The photon en-ergy of terahertz probe is very small共1 THz=4 meV兲 com-pared to that of optical probe so that it has a unique capabil-ity of examining the ultrafast relaxation mechanisms of free electrons, which can provide the essential information for device design and optimization. Several results on the ul-trafast carrier dynamics of InN epilayers have been reported by using time-resolved optical1–4or terahertz spectroscopy,5 but no systematic study has been reported on carrier dynam-ics of InN nanostructures. In the present work, we report on the carrier dynamics of InN nanorods compared to that of InN epilayer measured by optical pump–terahertz probe technique.
Previously, we have reported a significant enhancement of terahertz emission from InN nanorod arrays compared to that from InN epilayer.6 The major terahertz emission mechanism of InN film has been proposed to be the photo-Dember effect, which is driven by the transient current due to the difference of diffusion velocities of electrons and holes. Furthermore, a modified photo-Dember effect is sug-gested to explain the enhancement mechanism of terahertz emission from the nanorods, where terahertz emission de-pends on the size and aerial density of the nanorods. A
tera-hertz time-domain spectroscopy共TDS兲 measurement showed that InN nanorods have much shorter Drude scattering time constant than InN film, which may be due to the less perfect crystalline quality of the nanorods as well as the geometrical confinement of mobile carriers in the rods.7 Since terahertz emission is based on the transient behavior of the carriers, it is essential to characterize the carrier dynamics of InN films and nanorods in order to understand terahertz emission mechanism from these samples. Time-resolved terahertz transmission measurement allows us to study the transient optical responses and electron transfer mechanism of verti-cally aligned InN nanorod arrays, which cannot be explained by the static measurement techniques such as terahertz TDS. Distinctively fast decay of photoexcited carrier density is ob-served for the InN nanorod arrays compared to the InN films, which is attributed to the defect-related electron trapping and the increased interaction with the boundaries of nanorods. Additionally for nanorods, reduced diffusion rate in the vi-cinity of the surface is found to play an important role in the carrier relaxation dynamics at the long time delay.
II. EXPERIMENT
In these measurements, ultrafast optical pump was pro-vided by a Ti:sapphire regenerative amplifier laser system, which delivers⬃50 fs optical pulses at a center wavelength of 800 nm with a repetition rate of 1 kHz. The terahertz probe beam was generated from a photoexcited 共100兲 InAs surface and detected by free-space electro-optic sampling in a 2-mm-thick ZnTe crystal. In the optical pump–terahertz probe experiment, the transient behavior of the photoexcited carriers was monitored by measuring the transmitted peak amplitude of terahertz waveforms at normal incident angle as a function of delay time between the terahertz probe and optical pump pulses. The static electrical properties of the samples were separately measured by a terahertz-TDS sys-tem based on low-sys-temperature-grown GaAs photoconductive dipole antennas, which were excited and probed by a Ti:sap-phire laser at a repetition rate of 82 MHz. All the measure-ments were done under dry nitrogen purge.
a兲Author to whom correspondence should be addressed. Electronic mail: [email protected].
0021-8979/2009/105共2兲/023707/4/$25.00 105, 023707-1 © 2009 American Institute of Physics
For this work, a wurtzite InN epitaxial film and verti-cally aligned InN nanorod arrays were grown on Si共111兲 sub-strates by plasma-assisted molecular-beam epitaxy. The InN epilayer was grown on Si共111兲 using the epitaxial AlN/-Si3N4double-buffer layer technique.8The InN nano-rods were grown at a sample temperature of 330 ° C on -Si3N4/Si共111兲 without the AlN buffer layer. The scanning electron microscopy共SEM兲 image of the hexagonal-shaped nanorods exhibits nanorods with a uniform diameter of ⬃130 nm and an average aspect ratio 共height/diameter兲 of ⬃6. The nanorod arrays have an aerial density of ⬃5 ⫻109 cm−2. The thicknesses of the InN epilayer and nano-rods are about 1.0 and 0.75 m, respectively. The morphol-ogy and size distribution of InN nanorods were analyzed using SEM. Electrical characteristics were measured by tera-hertz TDS,7and the characteristics are as follows: InN epil-ayer film has a carrier density Ne=共2.5⫾0.2兲⫻1018 cm−3 and a carrier mobility = 1217⫾58 cm2/V s, while nano-rods have Ne=共4.9⫾0.2兲⫻1019 cm−3 and = 80⫾5 cm2/V s. The corresponding plasma frequencies of InN film and nanorods are 52⫾1.2 and 199⫾3 THz, re-spectively. Figure 1 shows near-infrared reflectivity of InN film and nanorods measured at the wavelength covering from 800 to 2400 nm. The oscillation of reflectivity of InN film is due to the light interference within the thin film in the trans-parent region below the bandgap energy. Meanwhile, the de-crease in reflectivity of InN nanorods at the wavelength be-low 1500 nm 共⬇199 THz兲 is attributed to the collective behavior of electrons and holes above the plasma frequency. The reflectivity responses at the pump wavelength共800 nm兲 show that InN nanorods absorb more light than InN film, which is due to the increased surface-to-volume ratio for nanorods.
III. DATA AND DISCUSSION
Figure2shows the time-dependent differential transmis-sion signals of InN nanorods共blue兲 and the epilayer 共black兲: ⌬Tterahertz/⌬Tterahertz0 , where⌬Tterahertz0 is the transmitted inten-sity of terahertz probe through the unexcited sample. Each sample is excited at the laser fluence of 1.1 mJ/cm2. As soon as the pump pulse arrives, transmission responses of both samples instantaneously drop and the sample-independent sharp fall time is measured to be 0.6–0.7 ps. Due to our relatively broad pulsewidth of terahertz probe
共⬃0.6 ps兲 and the slow detector response time, sample-dependent fall time cannot be monitored. The transmission response of InN film gradually recovers from 70% to⬃22% within 200 ps, while that of nanorods quickly recovers to its steady value within 2 ps and persists at this value over 200 ps.
The solid lines in Fig.2共a兲are obtained from a biexpo-nential fit,
⌬Tterahertz/Tterahertz 0
=⌬Tterahertzmax 关− Ae−t/1+共A − 1兲e−t/2兴, 共1兲 where ⌬Tterahertzmax is the maximum transmission change, A is the weighting factor, and i are relaxation time constants. Single exponential recovery function results in poor fitting for each sample. According to the best fit parameters listed in Table I, the initial fast relaxation time 共1兲 of nanorods is 2.6⫾0.5 ps and that for InN film is 30.7⫾0.9 ps. The slow relaxation time共2兲 of the InN film is 194⫾2.5 ps, but that of nanorods is much longer共⬎7 ns兲 and its accurate value cannot be determined by our system with a limited scanning range. From the measured⌬Tterahertz/Tterahertz
0
data, the time-dependent carrier density in a photoexcited sample can be calculated by the relation9
N共t兲 = 1 + n Z0e␦
冋
冉
1 +⌬Tterahertz Tterahertz0冊
−1 − 1册
, 共2兲where n = 3.3 is the refractive index for silicon substrate in the terahertz range,10 ␦= 133 nm is the optical penetration depth of 800 nm light,11and Z0= 377 ⍀ is the impedance of free space. With the electron mobilities measured by tera-hertz TDS, N共t兲 is calculated for nanorods and film and FIG. 1. 共Color online兲 Near-infrared reflectivity of InN film and nanorods
measured at normal incidence.
0 50 100 150 200 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0 10 20 -0.2 0.0 (a) T /T 0 nanorods InN film nanorods 0 50 100 150 0.0 0.5 1.0 1.5 2.0 2.5 Carrier density (10 19 cm -3 ) Time delay (ps) (b) nanorods InN film
FIG. 2. 共Color online兲 共a兲 Differential terahertz transmission dynamics, ⌬Tterahertz/Tterahertz
0 due to 800 nm excitation for nanorods共blue兲 and the InN epilayer 共black兲. The excitation fluence is 1.1 mJ/cm2. Inset: ⌬Tterahertz/Tterahertz0 of nanorods in the expanded scale. The solid lines are the results of biexponential fitting as described in the text.共b兲 The photoexcited carrier density of InN film and nanorods calculated by Eq.共2兲.
023707-2 Ahn et al. J. Appl. Phys. 105, 023707共2009兲
shown in Fig. 2共b兲. The peak values of the carrier density near zero time delay indicate that during the pump pulse, photoexcited carriers of ⬃1.0⫻1019 cm−3 is generated for InN film, which is higher than its unintentionally doped car-rier concentration关共2.5⫾0.2兲⫻1018 cm−3兴. On the contrary, pump pulse generates⬃2.3⫻1019 cm−3of photoexcited car-rier density for nanorods, which is of the same order to or lower than its free electron concentration.
Observed biexponential relaxation of InN film agrees with other experimental results, in which the carrier dynam-ics is found to be due to the hot carrier cooling through phonon emission followed by the defect-related nonradiative recombination process.1,2,5The fast relaxation time of 30 ps measured for our InN film is consistent with that of 48 ps for InN film 共Ne= 1.2⫻1019 cm−3兲 measured by the optical pump-probe technique.1 Meanwhile, previously measured near-infrared photoluminescence 共PL兲 response shows that PL efficiency for nanorods is two orders of magnitude lower than that for InN epilayer,12 indicating that nanorods have considerable amount of structural defects. Then the localized electrons within nanorods experience a preferential backward scattering caused by the increased structural defects12 or a Coulombic restoring force from charged defects.13The car-rier scattering 共or damping兲 time 0 related to the carrier conduction mobility through the Drude model is indeed shorter共13 fs兲 for nanorods than that for film 共52 fs兲.7 The observed shorter scattering time of nanorods suggests a faster capture rate to the defect states, which further supports the existence of the high defect concentration of nanorods. Therefore, observed shorter initial relaxation time constant of nanorods compared to film can be explained by the in-verse relation between the carrier lifetime and the free elec-tron density due to the increased elecelec-tron trapping by the defects.
It is well known that nonradiative defect-related recom-bination has the carrier-density-independent lifetime. To identify the nature of the recombination in InN nanorods, differential transmission is measured for the pump fluence range of 0.32– 0.96 mJ/cm2共see Fig.3兲. The general behav-ior of transmission trace for each sample is the same and the average fast relaxation time constant of nanorods is 2.1⫾0.3 ps. The slow relaxation components for both samples also do not show any observable pump fluence de-pendence. The pump-fluence-independent carrier lifetime suggests that the defect-related nonradiative recombination rather than Auger recombination is the common recombina-tion process for nanorods and InN film. At high pump flu-ence 共⬎0.6 mJ/cm2兲, the maximum negative change in transmission⌬Tterahertzmax of InN film in Fig.3共b兲 shows satu-ration, while that of nanorods scales linearly with the pump
fluence. The saturation of ⌬Tterahertzmax for InN film can be de-scribed as trap saturation due to the high photoexcited carrier density exceeding free electron concentration. However, the photoexcited carrier density of nanorods at pump fluence be-low 1 mJ/cm2 is still smaller than the free electron density so that photoexcited carrier density and subsequently ⌬Tterahertzmax increase linearly with pump fluence.
Here we need to pay attention to the extremely slow relaxation time constant 2 observed for nanorods. In addi-tion to the significant amount of defects, the physical prop-erties of InN nanorods strongly depend on the geometrical nature of nanorods, such as high surface-to-volume ratio. Therefore carrier dynamics of nanorods can be sensitive to the carrier relaxation near the surface, which is significantly different from that of the bulk. In particular, it is known that the carrier diffusion reduces diffusion rate significantly near the surface.14 Since the lattice heating takes place on the extent region in which the energy transfer occurs by diffu-sion, reduced diffusion rate near the surface corresponds to the reduced heated volume. We assume that the effective diffusion distance has the form of a conventional diffusion length 共Dt兲1/2 where t is the time for a hot carrier to diffuse before losing its energy by phonon emission.14With the am-bipolar diffusion coefficient of 2.0 cm2/s near the InN TABLE I. Summary of the biexponential fitting results for InN film and nanorods. The carrier scattering time
0and electron mobilities are obtained from a separate terahertz-TDS measurement共Ref.7兲. Sample ⌬Tterahertzmax A
1 共ps兲 共ps兲2 0 a 共fs兲 a 共cm2/V s兲 pa 共THz兲 InN film 0.7 0.29 30.7⫾0.9 194⫾2.5 52⫾2.5 1217⫾58 52⫾1.2 Nanorods 0.26 0.12 2.6⫾0.5 ⬎7000 13⫾0.2 80⫾5 199⫾3 aReference7. 0 10 20 30 40 50 -0.3 -0.2 -0.1 0.0 0.96 mJ/cm2 0.64 mJ/cm2 T/T 0 (a) nanorods 0.32 mJ/cm2 0 10 20 30 40 50 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 T /T0 Time delay(ps) (b) InN film
FIG. 3. 共Color online兲 Pump fluence dependence of differential terahertz transmission for共a兲 InN nanorods and 共b兲 InN film excited at the pump fluences of 0.32共blue兲, 0.64 共red兲, and 0.96 mJ/cm2共black兲.
surface,2 the corresponding diffusion length is about 45 nm in 10 ps, during which the lattice heating is achieved. This diffusion length is of the same order of the radius 共average radius= 65 nm兲 of nanorods, indicating that carriers within nearly the whole volume of nanorods participate in the car-rier relaxation process. The carcar-rier lifetime near the InN sur-face is particularly slow and is known2 to be 5.4 ns. This value agrees well with our abnormally long relaxation time constant 共ⱖ7 ns兲 of nanorods, confirming that the carrier dynamics of InN nanorods is mainly governed by the carrier diffusion near the surface. Meanwhile, the same long relax-ation of transient transmission response was observed for another InN nanorod sample6 with much higher aerial den-sity 共8⫻109 cm−2兲 of rods 共not shown here兲. For this sample, the aerial density is so high that the gaps between the rods are very small and thus the excitation light cannot propagate through the gaps to excite the carriers in Si sub-strate. Therefore, the nanosecond-order relaxation observed for both nanorod films is the intrinsic property of the nano-rod layer, not due to the slow relaxation of photoexcited carriers in Si substrate.
IV. SUMMARY
We have investigated the carrier relaxation dynamics in InN film and InN nanorods using optical pump–terahertz probe spectroscopy. Biexponential relaxation fit to the differ-ential transmission traces shows that the carrier dynamics of InN film and nanorods is due to the fast hot carrier cooling through phonon emission and the defect-related nonradiative recombination process. The pump-fluence-independent re-laxation dynamics rules out nonlinear recombination process such as Auger recombination. The extremely long relaxation
time constant observed for nanorods indicates that the carrier relaxation process is dominated by the slow carrier diffusion in the vicinity of the surface, which is due to the geometrical characteristics of nanorods.
ACKNOWLEDGMENTS
The authors are grateful to Professor S. Gwo for the InN samples and helpful discussions and H.-Y. Chen for provid-ing optical reflectivity data. This work was supported by the National Science Council 共NSC兲 through NSC Grant No. 96-2112-M-009-016-MY3.
1F. Chen, A. N. Cartwright, H. Lu, and W. J. Schaff,J. Cryst. Growth269, 10共2004兲.
2F. Chen, A. N. Cartwright, H. Lu, and W. J. Schaff,Appl. Phys. Lett.87, 212104共2005兲.
3Y.-C. Wen, C.-Y. Chen, C.-H. Shen, S. Gwo, and C.-K. Sun,Appl. Phys.
Lett.89, 232114共2006兲.
4T.-R. Tsai, C.-F. Chang, and S. Gwo,Appl. Phys. Lett.90, 252111共2007兲. 5V. Pačebutas, G. Alecksejenko, A. Krotkus, J. W. Ager III, W.
Waluk-iewicz, H. Lu, and W. J. Schaff,Appl. Phys. Lett.88, 191109共2006兲. 6H. Ahn, Y.-P. Ku, Y.-C. Wang, C.-H. Chuang, S. Gwo, and C.-L. Pan,
Appl. Phys. Lett.91, 132108共2007兲.
7H. Ahn, Y.-P. Ku, Y.-C. Wang, C.-H. Chuang, S. Gwo, and C.-L. Pan,
Appl. Phys. Lett.91, 163105共2007兲.
8S. Gwo, C.-L. Wu, C.-H. Shen, W.-H. Chang, T. M. Hsu, J.-S. Wang, and J.-T. Hsu,Appl. Phys. Lett.84, 3765共2004兲.
9K. P. H. Lui and F. A. Hegmann,J. Appl. Phys.93, 9012共2003兲. 10M. van Exter and D. Grischkowsky,Phys. Rev. B41, 12140共1990兲. 11H. Ahn, C.-H. Shen, C.-L. Wu, and S. Gwo,Appl. Phys. Lett.86, 201905
共2005兲.
12C.-H. Shen, H.-Y. Chen, H.-W. Lin, S. Gwo, A. A. Klochikhin, and V. Yu. Davydov,Appl. Phys. Lett.88, 253104共2006兲.
13J. B. Baxter and C. A. Schmuttenmaer, J. Phys. Chem. B 110, 25229 共2006兲.
14E. J. Yoffa,Appl. Phys. Lett.36, 37共1980兲.
023707-4 Ahn et al. J. Appl. Phys. 105, 023707共2009兲
