Letter pubs.acs.org/NanoLett
Room-Temperature Photodetection Dynamics of Single GaN Nanowires F. González-Posada,*,† R. Songmuang,‡ M. Den Hertog,‡ and E. Monroy† †
CEA-CNRS Group “Nanophysique et Semiconducteurs”, INAC-SP2M, CEA-Grenoble, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France ‡ CEA-CNRS Group “Nanophysique et Semiconducteurs”, Institut Néel-CNRS, BP 166, 38042 Grenoble Cedex 9, France ABSTRACT: We report on the photocurrent behavior of single GaN n−i−n nanowires (NWs) grown by plasmaassisted molecular-beam epitaxy on Si(111). These structures present a photoconductive gain in the range of 105−108 and an ultraviolet (350 nm) to visible (450 nm) responsivity ratio larger than 6 orders of magnitude. Polarized light couples with the NW geometry with a maximum photoresponse for polarization along the NW axis. The photocurrent scales sublinearly with optical power, following a I ∼ Pβ law (β < 1) in the measured range with β increasing with the measuring frequency. The photocurrent time response remains in the millisecond range, which is in contrast to the persistent (hours) photoconductivity effects observed in two-dimensional photoconductors. The photocurrent is independent of the measuring atmosphere, either in the air or in vacuum. Results are interpreted taking into account the effect of surface states and the total depletion of the NW intrinsic region. KEYWORDS: Nanowire, GaN, photocurrent, ultraviolet
N
along a nonpolar axis.20 This dominant role of the surface levels in the photodetection mechanisms can cause variations of the photoresponse as a function of the environment, as reported by some authors,21,22,20,23 but in disagreement with other observations.11 In this paper, we present a comprehensive study of the photocurrent phenomena in single defect-free n−i−n GaN NWs. Specifically, we analyze the effect of the excitation power, light polarization, measuring frequency, and environment. Comparing these results with the ones available in the literature, we identify the NWs intrinsic photodetection mechanisms and discuss the origin of the reported deviations. Materials and Methods. N-polar n−i−n GaN NWs were grown by plasma-assisted molecular-beam epitaxy (PAMBE) on Si(111) substrates under N-rich atmosphere at ∼790 °C without using any external catalyst.24−26 The NWs have a length of approximately 1.2 μm and a diameter of about 50−80 nm. They consist of two Si-doped n-type edges and an undoped (i) middle section with a nominal length of 400 nm. The residual doping in our PAMBE chamber is in the range of 1−5 × 1017 cm−3 for N-polar GaN. As a result, taking the NW diameter into account, the undoped region of the wires should be fully depleted of carriers due to the surface band bending introduced by surface states.11,16,27 For n-type doping, the Si cell temperature was tuned to the value that yields a Hall
ew applications for nanophotonics and nanoelectronics are been inspired by the development of defect-free onedimensional (1D) nanostructures. To assess their promised capabilities is particularly important to uncover their new and interesting characteristics and properties concerning electronic transport, light emission and photodetection. The low dimensionality and the high surface-to-volume ratio play a fundamental role in the overall photodetection performance. The photodetector capabilities of single nanowire (NW) and vertical NW arrays are under study for a variety of materials, such as Si,1,2 Ge,3 GaAs,4 ZnO,5−7 ZnTe,8 CdS,9 or SnO210. Regarding the properties of GaN NWs as photodetectors, some aspects have been investigated;11−17 however, systematic measurements should still be performed to understand and differentiate the intrinsic behavior related to the 1D geometry. There is a general agreement on the presence of an ultrahigh photocurrent gain (G ∼ 105−108).14 Often (but not always), the photoconductivity is associated to persistent photoconductivity (PPC) in the range of tens to hundreds of seconds.16,18 There is contradictory information in terms of linearity with the excitation power9,4,16 and data on the spectral response remain vague. Experimental results are often explained using the traditional photoconductor models, where gain arises from the increase of the carrier lifetime in comparison to the electron transit time.19 The enhanced lifetime of the photogenerated carriers, consistent with the long photocurrent decay time, is explained by the localization of the holes at surface trap states in NWs grown along a polar axis11 or to the charge separation induced by polarization in the case of NWs grown © 2011 American Chemical Society
Received: September 19, 2011 Revised: November 2, 2011 Published: December 5, 2011 172
dx.doi.org/10.1021/nl2032684 | Nano Lett. 2012, 12, 172−176
Nano Letters
Letter
Figure 1. Side-view SEM image of n−i−n GaN NWs on Si(111) substrate. Insets: Top panel is the top-view SEM image of a contacted single NW; bottom panel shows the measurement scheme for characterization of a single NW device.
electron concentration of ∼2 × 1019 cm−3 in two-dimensional (2D) GaN layers. The length of each section is estimated from the vertical growth rate multiplied by the deposition time, without taking Si diffusion into account. At the growth temperature under consideration, Si from the bottom part of the nanowires might penetrate into the undoped region, but the residual doping level (1017 cm−3) should be reached at maximum 20 nm after shuttering the Si cell,28 keeping an i region of at least 380 nm. Figure 1 shows a side-view scanning electron microscopy (SEM) image of the n−i−n GaN NWs on the Si(111) substrate. To perform the characterization of single objects, the NWs were mechanically dispersed on heavily doped Si substrates capped with 280 nm thick SiO2 layer. Some NWs were individually contacted with Ti/Al/Ti/Au (5/25/15/100 nm) electrodes using e-beam lithography and lift-off technique. A rapid thermal annealing at 700 °C for 1 min was performed to improve the contact characteristics. The top right inset of Figure 1 shows a top-view SEM micrograph of a contacted wire. The current−voltage (I−V) characteristics of the bonded NWs were tested using an Agilent 4155C semiconductor parameter analyzer. For photocurrent measurements, the NWs were biased (Vb) and connected in series with a load resistance (RL = 100−180 kΩ), following the scheme shown in the inset of Figure 1. RL was connected in parallel to a TDS2022C oscilloscope or to an SR830 lock-in amplifier (input resistance = 10 MΩ). The spectral response was characterized using a 150 W Xe-arc lamp coupled to a Gemini 180 monochromator. At the exit, the light was directed with an optic fiber and focused onto the contacted NW. The photocurrent as a function of the optical power was analyzed using an Ar+ laser (wavelength λ = 244 nm, 488 nm) as excitation source. All the measurements presented in this work were performed at room temperature. Results and Discussion. Typical I−V characteristics of single NWs in the dark and under ultraviolet (UV) illumination are displayed in Figure 2. NWs with symmetrical ohmic contacts located on the Si-doped sections present a fairly symmetric I−V curve in a bias range from −4 to 4 V, as illustrated in Figure 2c. Under UV exposure, the NW
Figure 2. Top-view SEM images and I−V characteristics at room temperature of single NWs (a,c) with symmetrical contacts on the ntype regions of the wire and (b,d) with a misalignment leading to one of the contacts touching the undoped region. Solid lines correspond to the dark current; red dashed lines correspond to illumination with ∼0.2 W/cm2 of UV light (λ = 244 nm). (e) Double logarithmic plot of the I−V characteristics of a symmetric NW measured in the dark and under UV illumination. For comparison, I ∼ V and I ∼ V4 behaviors are given in dashed lines and solid lines are for eye-guide.
conductivity increases, and recovers to its initial value when switching off the excitation source. In the case that one of the metal pads gets in contact with the intrinsic region due to misalignment, the I−V characteristics present an asymmetry (see Figure 2d) due to the alteration of the surface potential of the i-region under the metallization. This perturbation induces an electric field along the wire that enhances the collection of photogenerated carriers resulting in a higher photoresponse (see Figure 2d). Note that the magnitude of the dark current and photocurrent is consistent with reports by other groups measuring NWs with comparable dimensions.11,13,16,17 This agrees with the observation of a reaction to UV light by more than 1 order of magnitude if the contacts are deposited directly 173
dx.doi.org/10.1021/nl2032684 | Nano Lett. 2012, 12, 172−176
Nano Letters
Letter
the photocurrent increases sublinearly following approximately a power law I ∼ Pβ with β < 1 for all the measurement frequencies, although both the photocurrent magnitude and the value of β depend on the frequency. The photodetector gain, G, defined as the number of electrons detected per absorbed photon, can be estimated from the photocurrent, I, via the equation
on undoped GaN, and a significantly smaller photocurrent when the contacts are deposited on n-type doped sections. Figure 2e presents a double logarithmic plot of the I−V characteristic of a symmetric wire. The I ∼ Vα behavior, with α > 1 and increasing with bias, is characteristic of space-chargelimited transport, as expected in intrinsic or fully depleted NWs. This situation appears when the NW resistance exceeds the contact resistance,2,4,9 that is, for an intrinsic NW section with ohmic contacts as presented in this work. Space-chargelimited current was first studied in insulators, where it is the dominant mechanism when the charge injected by the electrodes exceeds the free carrier density in the intrinsic material. In ideal insulators, the I−V curve present a characteristic I ∼ V 2 dependence (Mott−Gurney law).29,30 However, in low dimensional materials such as NWs, the I ∼ Vα behaves with an α > 2 due to the presence of surface states.9 In the case of GaN NWs, space-charge-limited current is observed for NW diameters below 200 nm,11,16 which is consistent with theoretical calculations that predict that NWs with a residual ntype doping in the 1017 cm−3 range should be fully depleted by surface states. The coupling of the light into the n−i−n GaN NW can be confirmed by analyzing the photoresponse as a function of the polarization angle, θ, when exciting with linearly polarized light.31,32 The inset of Figure 3 presents the photocurrent from
I hc P λe where h is Plack’s constant, c is the speed of light, and e is the electron charge. Note that this equation provides an underestimation of G, since it assumes the detector internal quantum efficiency to be unity, that is, the light reflected or transmitted through the NW is neglected. Taking into account the NW size, G reaches values of 106−107 for an irradiance of 0.2 mW/cm2. A sublinear response with the optical power has been previously reported in NWs based on other materials, like Si,1 Ge,3 ZnO,5,6 SnO2,10 ZnTe,8 or m-axial GaN NWs14. This strong nonlinearity and high gain, also comparable to previous observations in GaN 2D photoconductors,33 is generally associated to PPC with a strong photoresponse to the excitation energy below the GaN bandgap. This behavior was ascribed to a modulation of the semiconductor conductivity induced by the photogenerated carriers from extended defect states inside the bandgap. In fact, semiconductor surface is also considered as an extended defect. In PAMBE grown GaN NWs as the ones described in this paper, the presence of structural defects, if any, is restricted to the first hundred nanometers close to the substrate. However, because of the large surface-tovolume ratio, NWs contain an extremely high density of surface states. Thus, the presence of a nonlinear photocurrent gain in NWs has been assigned to an increase of the photogenerated electron lifetime due to the hole trapping at surface states.11,19 This interpretation, which will be further discussed below, implies a spatial separation of the electron and hole with a timedependent potential barrier for carrier recombination, associated to the surface band bending. As a result, the dynamic response of the NW should be very slow (seconds or even minutes). The GaN intrinsic nature of the NW photoresponse is confirmed by spectrally resolved measurements. Figure 4 G=
Figure 3. Photocurrent variation from a single NW as a function of the excitation power (λ = 244 nm) and frequency. Inset: Photocurrent dependence with the polarization angle of a linearly polarized white light source (0.2 mW/cm2 of the Xe lamp) and simulated angle dependence for a NW, following the I ∼ cos2 θ proportionality. The measurement was performed under 3 V bias, using a lock-in amplifier with a chopping frequency of 69 Hz.
a single NW exposed to the Xe lamp through a linear polarizer. The typical 8-shape observed in the plot in polar coordinates, with negligible photocurrent for θ = 30° and θ = 210°, confirms the extinction of the light absorption perpendicular to the NW longitudinal axis, as theoretically expected.31 The strong anisotropy of the optical absorption in the NW, that is, 8shape, is a direct consequence of the photocurrent varying as I ∼ cos2 θ which yields a good fit to the measured data (dotted line in the inset of Figure 3).19 The quantitative analysis of the photoresponse requires an assessment of the linearity of the device with the impinging optical power, P. For this purpose, the photocurrent was measured as a function of P using an Ar laser (λ = 244 nm) as excitation source. Measurements were systematically carried out from low to high excitation power to prevent artifacts related to persistence or temperature transients. The characterization was performed under continuous-wave (CW) illumination, as well as at different light chopping frequencies. As shown in Figure 3,
Figure 4. Spectral response of a single NW under 3 V bias at a measurement frequency of 2−1000 Hz. The spectrum was corrected by the lamp intensity taking the linearity measurement in Figure 3 into account. 174
dx.doi.org/10.1021/nl2032684 | Nano Lett. 2012, 12, 172−176
Nano Letters
Letter
measurements. The decay becomes faster and increasingly nonexponential for higher excitation power. The time scale of the NW response is in any case much shorter than the PPC effects observed in 2D GaN photoconductors, where the transients are in the range of hours.33 This fact suggests that the slow photocurrent components in the NWs are not related to extended defects, consistent with their defect-free nature. However, the millisecond time response is orders of magnitude slower than the subnanosecond photoluminescence decay times measured in GaN NWs,27,36 which indicates that the photocurrent time response is not governed by the lifetime of the photogenerated carriers, but to the redistribution of charge at the surface. This redistribution is relatively fast in comparison to NWs with a larger diameter or 2D layers, thanks to the situation of total depletion of the NW active region, which implies a reduction of the potential barrier associated to the surface states. Intuitively, the involvement of surface states in the photodetection mechanism might result in a sensitivity of the photoresponse to the environment. To assess this issue, the I− V characteristics and photocurrent performance of n−i−n GaN NWs were analyzed both in the air and in vacuum at 2−4 V bias. In vacuum, the dark current increases by about 60% in comparison to the value obtained by the measurements in air, which is in agreement with the reports by other groups.20,22 This drop is consistent with a decrease of the carrier lifetime in presence of oxygen adsorbed at the NW surface.27 However, the ratio of the photocurrent measured in the air and in vacuum is 0.95 ± 0.2, that is, the effect of the atmosphere on the photocurrent response is almost negligible under UV excitation power in 0.2−100 mW/cm2 range, in agreement with Calarco et al.11 This insensibility to the environment is associated to both a nonlinear photoresponse behavior and a total carrier depletion situation. First, the variation of the carrier lifetime has a linear effect on the density of photogenerated carriers, but the photocurrent scales sublinearly with the carrier density, which attenuates the effect. On the other hand, the total carrier depletion leads to a reduction of the NW surface band bending forcing almost flat-band conditions. As a result, changes in the surface potential do not induce critical changes in the NW radial potential profile. This latter mechanism is no longer valid in the case of NWs with a larger diameter, where the changes in the surface potential translate into a variation of the width of the surface depletion region with drastic effects on the NW conductivity. In the case of NWs grown along a nonpolar axis, the asymmetry of the polar sidewalls amplifies the effect of surface states, and as a consequence the sensitivity to the chemistry of the environment is enhanced.20 This insensitivity of depleted polar NWs to the environment is an attractive feature for reliable and robust photodetector applications. Further improvement of the photodetector time response, going beyond the gain × bandwidth limitation, requires either an adjustment of the device design (e.g., replacing the ohmic contacts by Schottky barriers,37 which enhance the collection of photogenerated carriers) or a modification of the surface properties via core−shell structure,38 surface treatment or passivation.39 Conclusions. Defect-free GaN n−i−n NWs present high photocurrent gain in the range of 105−108. The photocurrent increases sublinearly with the excitation power, approximately as a function of Pβ with β increasing with the measuring frequency. The spectral response is relatively flat for excitation above the GaN bandgap and presents a visible rejection of
presents the spectral photoresponse of a single n−i−n GaN NW at 3 V bias measured by the lock-in technique at various frequencies. Data are corrected by the Xe-lamp emission spectrum taking the sublinear power dependence of the NW into account.34 The properly corrected photocurrent spectra present a flat spectral response for wavelengths above the GaN bandgap (λ < 350 nm), while the response to λ > 450 nm is below the resolution limit of the system. In order to verify the NW blindness to visible light, they were exposed to 0.5 W/cm2 of the 488 nm line of an Ar laser. The NWs showed no sensitivity to this illumination neither under CW illumination nor at different chopping frequencies (2−100 Hz). Taking into account the power density of the Xe lamp and the resolution of our system, this measurements confirm a UV(350 nm)/ visible(488 nm) contrast of more than 6 orders of magnitude. This result sets a critical difference in performance with GaN 2D photoconductors, whose visible rejection ratio decreases markedly when decreasing the measuring frequency, to the point of being lower than one decade for CW measurements.33 This poor visible rejection of GaN 2D photoconductor in orders of magnitude worse than expected from the spectral variation of the GaN absorption coefficient is explained by the presence of charge states within the bandgap associated to the extended defects that generate the photocurrent gain. In the case of GaN NWs, the huge UV/visible contrast is understood as a result of the absence of structural defects, except for the surface dangling bonds. Furthermore, the NW sidewalls are {1−100} m-planes, whose dangling bonds do not create occupied states within the band gap.35 To get further insight in the photoresponse dynamics, the variation of the photodetector gain was measured as a function of the chopping frequency, as presented in Figure 5. The
Figure 5. Variation of the photodetector gain of a GaN NW under 3 V bias as a function of the measuring frequency for various excitation powers. Dotted lines are guides for the eye. Inset: Photocurrent decay from a single NW after UV (λ = 244 nm) illumination at 80 and 0.2 mW/cm2. The light was chopped at f = 2 Hz to minimize thermal effects.
frequency dependent behavior confirms the larger bandwidth for higher excitation powers, associated to a decrease of the photodetector gain. The inset of Figure 5 shows the photocurrent decays from a single NW after 0.2 and 80 mW/ cm2 of UV (λ = 244 nm) illumination. By fitting the initial decay with a single exponential function, the time constants are in the order of 4−10 ms, consistent with the bandwidth 175
dx.doi.org/10.1021/nl2032684 | Nano Lett. 2012, 12, 172−176
Nano Letters
Letter
(18) Dhara, S.; Lu, C. Y.; Wu, C. T.; Hsu, C. W.; Tu, W. S.; Chen, K. H.; Wang, Y. L.; Chen, L. C.; Raj, B. J. Phys. Chem. 2011, 114, 15260− 15265. (19) Soci, C.; Zhang, W. J.; Bao, X. Y.; Lo, Y. H.; Wang, D. J. Nanosci. Nanotechnol. 2010, 10, 1. (20) Chen, R.-S.; Lu, C.-Y.; Chen, K.-H.; Chen, L.-C. Appl. Phys. Lett. 2009, 95, 233119. (21) Aluri, G. S.; Motayed, A.; Davydov, A. V.; Oleshko, V. P.; Bertness, K. A.; Sanford, N. A.; Rao, M. V. Nanotechnology 2011, 22, 295503. (22) Kang, M.; Lee, J.-S.; Sim, S.-K.; Kim, H.; Min, B.; Cho, K.; Kim, G.-T.; Sung, M.-Y.; Kim, S.; Han, H. S. Jpn. J. Appl. Phys. 2004, 43, 6868−6872. (23) Teubert, J.; Becker, P.; Furtmayr, F.; Eickhoff, M. Nanotechnology 2011, 22, 275505. (24) Yoshizawa, M.; Kikuchi, A.; Mori, M.; Gujita, N.; Kishino, K. Jpn. J. Appl. Phys. 1997, 36, L459. (25) Sanchez-Garcia, M. A.; Calleja, E; Monroy, E; Sanchez, F. J.; Calle, F.; Munoz, E.; Beresford, R. J. Cryst. Growth 1998, 183, 23. (26) Songmuang, R.; Landré, O.; Daudin, B. Appl. Phys. Lett. 2007, 91, 251902. (27) Pfüller, C.; Brandt, O.; Grosse, F.; Flissikowski, T.; Chèze, C.; Consonni, V.; Geelhaar, L.; Grahn, H. T.; Riechert, H. Phys. Rev. B 2010, 82, 045320. (28) Furtmayr, F.; Vielemeyer, M.; Stutzmann, M.; Arbiol, J.; Estradé, S.; Peirò, F.; Morante, J. R.; Eickhoff, M. J. Appl. Phys. 2008, 104, 034309. (29) Rose, A. Phys. Rev. 1955, 97, 1538. (30) Lampert, M. A. Phys. Rev. 1956, 103, 1648. (31) Maslow, A. V.; Bakunov, M. I.; Ningb, C. Z. J. Appl. Phys. 2006, 99, 024314. (32) Ruda, H. E.; Shik, A. Phys. Rev. B 2005, 72, 115308. (33) Monroy, E.; Omnès, F.; Calle, F. Semicond. Sci. Technol. 2003, 18, R33. (34) González-Posada, F.; Songmuang, R.; Den Hertog, M.; Monroy, E. Phys. Status Solidi 2011. (35) Van de Walle, C. G.; Segev, D. J. Appl. Phys. 2007, 101, 081704. (36) Schlager, J. B.; Bertness, K. A.; Blanchard, P. T.; Robins, L. H.; Roshko, A.; Sanford, N. A. J. Appl. Phys. 2008, 103, 124309. (37) Monroy, E.; Calle, F.; Pau, J. L.; Muñoz, E.; Omnès, F.; Beaumont, B.; Gibart, P. J. Cryst. Growth 2001, 230, 537−543. (38) Gallo, E. M.; Chen, G.; Currie, M.; McGuckin, T.; Prete, P.; Lovergine, N.; Nabet, B.; Spanier, J. E. Appl. Phys. Lett. 2011, 98, 241113. (39) Decorby, R. G.; Macdonald, R. I.; Beaudoin, M.; Pinnington, T.; Tiedje, T.; Gouin, F. J. Electron. Mater. 1997, 26, L25−28.
more than 6 orders of magnitude, independent of the measuring frequency. The photocurrent time response is in the millisecond range, significantly slower than the subnanosecond carrier lifetime, but far from the PPC effects in the range of seconds, or up to minutes, reported for larger NWs or 2D layers. The above-described results indicate that the photoresponse is dominated by the redistribution of charge at the surface levels. Nevertheless, the condition of total depletion of the NW active region reduces the surface band bending, thus preventing PPC effects and providing a certain insensitivity to the chemical environment. Furthermore, the m-plane crystallographic orientation of the NW sidewalls is characterized by the absence of occupied states within the band gap, which explains their excellent spectral response.
■
AUTHOR INFORMATION
Corresponding Author *E-mail:
[email protected].
■
ACKNOWLEDGMENTS Partial support from the French National Research Agency within “COSNI” project (ANR-08-BLAN-0298-02) is acknowledged. We thank J. Dussaud, Y. Curé, and Y. Genuist for their technical support.
■
REFERENCES
(1) Zhang, A.; You, S.; Soci, C.; Liu, Y.; Wang, D.; Lo, Y.-H. Appl. Phys. Lett. 2008, 93, 121110. (2) Alvarez, J.; Ngo, I.; Gueunier-Farret, M.-E.; Kleider, J.-P.; Linwei, Y.; Cabarrocas, P. R.; Perraud, S.; Rouvière, E.; Celle, C.; Mouchet, C.; Simonato, J.-P. Nanoscale Res. Lett. 2011, 6, 110. (3) Kim, C.-J.; Lee, H.-S.; Cho, Y.-J.; Kang, K.; Jo, M.-H. Nano Lett. 2010, 10, 2043. (4) Schricker, A. D.; Davidson, F. M. III; Wiacek, R. J.; Korgel, B. A. Nanotechnology 2006, 17, 2681. (5) Soci, C.; Zhang, W. J.; Xiang, B.; Dayeh, S. A.; Aplin, D. P. R; Park, J.; Bao, X. Y.; Lo, Y. H.; Wang, D. Nano Lett. 2007, 7, 1003. (6) Chen, M.-W.; Chen, C.-Y.; Lien, D.-H.; Ding, Y.; He, J.-H. Opt. Express 2010, 18, 14836. (7) Peng, S.-M.; Su, Y.-K.; Ji, L.-W.; Young, S.-J.; Tsai, C.-N.; Wu, C.Z.; Chao, W.-C.; Cheng, W. B.; Huang, C.-J. Electrochem. Solid-State Lett. 2011, 14, J13. (8) Cao, Y. L.; Liu, Z. T.; Chen, L. M.; Tang, Y. B.; Luo, L. B.; Jie, J. S.; Zhang, W. J.; Lee, S. T.; Lee, C. S. Opt. Express 2011, 19, 6100. (9) Gu, Y.; Lauhon, L. J. Appl. Phys. Lett. 2006, 89, 143102. (10) Hu, L.; Yan, J.; Liao, M.; Wu, L.; Fang, X. Small 2011, 8, 1012. (11) Calarco, R.; Marso, M.; Richter, T.; Aykanat, A. I.; Meijers, R.; Hart, A; v.d.; Stoica, T.; Lüth, H. Nano Lett. 2005, 5, 981−984. (12) Polenta, L.; Rossi, M.; Cavallini, A.; Calarco, R.; Marso, M.; Meijers, R.; Richter, T.; Stoica, T.; Lüth, H. ACS Nano 2008, 2, 287− 292. (13) Richter, T.; Lüth, H.; Meijers, R.; Calarco, R.; Marso, M. Nano Lett. 2008, 8, 3056. (14) Chen, R.-S.; Chen, H.-Y.; Lu, C.-Y.; Chen, K.-H.; Chen, C.-P.; Chen, L.-C.; Yang, Y.-J. Appl. Phys. Lett. 2007, 91, 223106. (15) Chen, R.-S.; Wang, S.-W.; Lan, Z.-H.; Tsai, J. T.-H.; Wu, C.-T.; Chen, L.-C.; Chen, K.-H.; Huang, Y.-S.; Chen, C.-C. Small 2008, 4, 925−929. (16) Sanford, N. A.; Blanchard, P. T.; Bertness, K. A.; Mansfield, L.; Schlager, J. B.; Sanders, A. W.; Roshko, A.; Burton, B. B.; George, S. M. J. Appl. Phys. 2010, 107, 034318. (17) Rigutti, L.; Tchernycheva, M.; De Luna Bugallo, A.; Jacopin, G.; Julien, F. H.; Zagonel, L. F.; March, K.; Stephan, O.; Kociak, M.; Songmuang, R. Nano Lett. 2010, 10, 2939−2943. 176
dx.doi.org/10.1021/nl2032684 | Nano Lett. 2012, 12, 172−176