Suitability of Au- and Self-Assisted GaAs Nanowires for Optoelectronic

LETTER pubs.acs.org/NanoLett

Suitability of Au- and Self-Assisted GaAs Nanowires for Optoelectronic Applications Steffen Breuer,* Carsten Pf€uller, Timur Flissikowski, Oliver Brandt, Holger T. Grahn, Lutz Geelhaar, and Henning Riechert Paul-Drude-Institut f€ur Festk€orperelektronik, Hausvogteiplatz 5-7, 10117 Berlin, Germany

bS Supporting Information ABSTRACT: The incorporation of Au during vapor-liquid-solid nanowire growth might inherently limit the performance of nanowire-based devices. Here, we assess the material quality of Au-assisted and Au-free grown GaAs/(Al,Ga)As core-shell nanowires using photoluminescence spectroscopy. We show that at room temperature, the internal quantum efficiency is systematically much lower for the Au-assisted nanowires than for the Au-free ones. In contrast, the optoelectronic material quality of the latter is comparable to that of state-of-the-art planar double heterostructures. KEYWORDS: Nanowire, vapor-liquid-solid, heteroepitaxy, photoluminescence, minority carrier lifetime, deep recombination center

S

emiconductor nanowires are believed to be promising components for future optoelectronic devices.1,2 Most frequently, they are fabricated with the help of Au droplets which induce vapor-liquid-solid growth of nanowires beneath each droplet.3,4 Yet, the presence of Au at the growth front might result in the incorporation of traces of Au into the nanowire.5 This contamination could have grave consequences, since Au is known as an efficient nonradiative recombination center in several semiconductors, drastically reducing the internal quantum efficiency η of the material.6-8 Recently, an alternative (selfassisted) nanowire growth method emerged for III-V compounds, in which droplets of the group III element replace those of Au.9 Here, we examine the potential of GaAs/(Al,Ga)As core-shell nanowires produced by both techniques for actual optoelectronic applications. To this end, continuous-wave and time-resolved photoluminescence spectroscopies are employed. Both techniques reveal a drastically higher value of η for the selfassisted nanowires. The internal quantum efficiency η is the central figure-of-merit expressing a material’s potential for optoelectronic applications. For example, the value of η determines the threshold current density of injection laser diodes, the luminous efficacy of lightemitting diodes, and the power conversion efficiency of solar cells. Experimentally, η is proportional to the spectrally integrated intensity of the spontaneous emission of the semiconductor and is thus accessible by recording its steady-state photoluminescence (PL) spectrum. Since η may, in general, be written as η = τ/τr with the minority carrier lifetime τ and the radiative lifetime τr, it can be measured independently by timeresolved PL upon pulsed excitation. In contrast to the steadystate intensity, τ is affected neither by volume nor by the coupling of light into and out of the structure. Thus, the relevant quantities are attainable by all-optical measurements, independent of r 2011 American Chemical Society

contacts of any kind, and avoiding the ambiguities inherent in measurements relying on them. GaAs nanowires were fabricated by molecular beam epitaxy (MBE) using both the Au- and the self-assisted growth mechanism in the same growth chamber. Au-assisted nanowires were grown at 500 °C on deoxidized Si(111) substrates covered with Au nanodroplets and self-assisted nanowires were fabricated at 580 °C on Si(111) with remaining native oxide (see Supporting Information for details).10,11 The GaAs cores were surrounded with (Al,Ga)As shells in order to disable the dominant nonradiative recombination at the free GaAs surface.12 Figure 1a,b shows scanning electron micrographs of one representative sample for each growth mechanism. The Auassisted nanowires (a) are shorter and thinner but stand closer than the self-assisted ones (b). The Au-assisted nanowires have a pencil shape, which arises from comparably small Au droplet diameters (around 10 nm, below the resolution of the micrograph) in conjunction with radial growth of the nanowires.13,14 Along the lower 3/4 of their lengths of 5 ( 1 μm, the Au-assisted wires have uniform diameters of 68 ( 12 nm, and above they are obviously tapered. Their average number density is 5 μm-2. The Au-assisted core diameters were determined to be 42 ( 7 nm. In contrast, Ga droplets are clearly observable at the tip of the selfassisted nanowires, which show uniform diameters of 150 ( 25 nm along their entire lengths of 9 ( 1 μm. Their average number density is 1 μm-2. The diameters of the self-assisted GaAs cores were determined to be 106 ( 18 nm. Error values show the standard deviation between different nanowires on the same sample. Received: December 10, 2010 Revised: January 18, 2011 Published: February 14, 2011 1276

dx.doi.org/10.1021/nl104316t | Nano Lett. 2011, 11, 1276–1279

Nano Letters

LETTER

Figure 1. Morphology and luminescence for both growth mechanisms. (a, b) Scanning electron micrographs of (a) Au-assisted and (b) self-assisted GaAs/(Al,Ga)As core-shell nanowires on Si(111), insets show magnifications of the top and foot region. (c) μ-PL spectra of the Au-assisted and the self-assisted nanowires at room temperature. The spectra were recorded under identical excitation conditions. The integrated intensity for the selfassisted nanowires is larger than that for the Au-assisted ones by more than 2 orders of magnitude. Band-to-band transition energies of ZB and WZ GaAs as well as WZ AlxGa1-xAs with x = 0.06 are indicated. (d) μ-PL transients for the Au-assisted and the self-assisted nanowires at room temperature. The inset shows a magnification of the Au-assisted nanowire transient. Solid lines represent fits to the data. The obtained minority carrier lifetimes differ by more than 2 orders of magnitude.

Figure 1c shows the room-temperature micro-PL (μ-PL) spectra of the two representative samples measured side-by-side under identical conditions (see Supporting Information for details). The integrated PL intensity I of the self-assisted nanowires is larger than that of the Au-assisted nanowires by more than 2 orders of magnitude. The spectral positions of the bands at 1.429, 1.446, and 1.522 eV correspond to radiative recombination in GaAs in the zinc blende (ZB) phase for the self-assisted nanowires, as well as in GaAs and AlxGa1-xAs (x = 0.06 ( 0.01) in the wurtzite (WZ) phase for the Au-assisted nanowires, respectively.15,16 The observation of PL emission from the Al0.06Ga0.94As barrier in the Au-assisted case indicates a higher internal quantum efficiency of the shell compared to the GaAs core for this type of growth mechanism. The fact that different polytypes are obtained by the Au- and self-assisted growth techniques is typical for nanowire growth under similar conditions.10,17 Figure 1d displays PL transients of the two representative samples obtained by time-resolved PL spectroscopy at the energy of the band-to-band recombination of the respective polytype of GaAs (see Supporting Information for details). The minority carrier lifetimes τ are extracted by least-squares fits using the general expression for recombination of free carriers (see Supporting Information).18 For the self-assisted nanowires, a carrier lifetime of τS = (2.5 ( 0.1) ns is found, while τA = (9 ( 1) ps results for the Au-assisted nanowires. The latter value is essentially identical to the one reported earlier for Au-assisted GaAs/(Al,Ga)As core-shell nanowires grown by metal-organic vapor phase epitaxy (MOVPE).19 In order to compare the lifetimes obtained here with those for planar heterostructures, we recall that the effect of interface

recombination on τ is described by 1/τ = 1/τb þ nS/d, where τb is the bulk minority carrier lifetime, S is the interface recombination velocity, n equals 2 for planar geometry and 4 for nanowires with square or circular cross section, and d is the nanowire core diameter or layer thickness, respectively.20,21 Figure 2 depicts our measurements in the context of literature data for planar GaAs/(Al,Ga)As double heterostructures (DH) representing the state of the art of planar growth by MBE. The lifetimes for homoepitaxially grown DHs on GaAs(001) as well as heteroepitaxially grown DHs on Si(001) using step-graded Ge/(Ge,Si) buffer layers are shown together with fits (solid lines) using the above equation, n = 2, and the values for S and τb that were extracted by the original authors.22,23 Substitution with n = 4 yields extrapolations (dashed lines) for nanowires with identical interface and bulk quality. It follows that lifetimes between 0.2 and 10 ns should be expected for GaAs/(Al,Ga)As core-shell nanowires on Si with d between 40 and 100 nm. It is the central result of this study that only the self-assisted nanowires exhibit a minority carrier lifetime comparable to that expected for material of state-of-the-art quality. Moreover, the lifetime obtained for the self-assisted nanowires, τS, is larger by more than 2 orders of magnitude than the one measured for the Au-assisted nanowires, τA. This factor directly reflects the ratio of the internal quantum efficiencies η and also corresponds to the ratio of the PL intensities obtained in steady state. This drastic difference in minority carrier lifetime between nanowires synthesized using the Au-assisted and the self-assisted growth modes is observed for all analyzed samples without exception. 1277

dx.doi.org/10.1021/nl104316t |Nano Lett. 2011, 11, 1276–1279

Nano Letters

Figure 2. Comparison of GaAs/(Al,Ga)As core-shell nanowires and planar double heterostructures (DH). Minority carrier lifetimes for core-shell nanowires on Si substrates (filled circles, this work), Auassisted core-shell nanowires (open circle),19 planar DH on GaAs substrates (squares),22 and planar DH grown on step-graded Ge/(Ge, Si) buffer layers on Si substrates (diamonds)23 are shown as a function of the GaAs dimension, i.e., the nanowire core diameter or the layer thickness without barriers. Fits (solid lines) for planar DH are shown as well as extrapolations (dashed lines) to core-shell nanowires with identical bulk and surface recombination parameters. While our selfassisted nanowires do not yet reach the quality of planar DH on GaAs, they are already better than the extrapolation for planar DH on Si. The optoelectronic material quality at room temperature of Au-assisted nanowires is considerably lower.

Figure 3. Thermal activation energies. Temperature dependence of the integrated PL intensities I of self-assisted and Au-assisted GaAs/(Al, Ga)As core-shell nanowires plotted in a double logarithmic representation. The inset presents the same data in a form suitable for the determination of activation energies. Linear fits to the low-temperature side of the inset yield activation energies of 4.3 meV, which correspond to the exciton binding energy in GaAs. A linear fit to the hightemperature side of the inset is possible only for the Au-assisted sample. The obtained energy of 77 meV is attributed to the thermal activation of carrier capture by a deep center.

Figure 2 reveals furthermore that nonradiative recombination at the interfaces influences the minority carrier lifetime even of state-of-the-art GaAs structure as long as d < 10 μm. Evidently, an additional, highly efficient nonradiative recombination channel exists in the Au-assisted nanowires that drastically reduces the internal quantum efficiency of these structures. The obvious question is about the nature of this additional channel. Its effect is far too large to be explained by the 80 °C lower growth temperature.24 As a matter of fact, self-assisted nanowires grown at 540 °C exhibit the same long lifetimes and the same high PL intensity as those grown at 580 °C (a common growth temperature does not exist for the two growth modes used in this study).

LETTER

The most straightforward and plausible explanation for our finding is that (i) Au is in fact incorporated into the nanowires during Au-assisted VLS growth and (ii) acts as nonradiative recombination center. Concerning incorporation, Au is of course present in abundance at the growth front during Au-assisted growth. The solid solubility of Au in GaAs is 2.5  1016 cm-3 at 900 °C.25 This value may appear low, but a concentration of 1016 cm-3 still corresponds to one Au atom every 80 nm of length for a 40 nm diameter GaAs nanowire core. Given that the ambipolar diffusion length in GaAs is 0.69 μm,26 all photoexcited carriers could indeed reach a Au center. Considering the behavior of Au or a Au-related complex as a nonradiative recombination center, we recall that such centers can often be recognized by their characteristic thermal activation energy EA. Figure 3 shows the temperature (T) dependence of the integrated PL intensities IS and IA for the self- and the Auassisted nanowires, respectively (see Supporting Information for measurement details). For low temperatures, both IS and IA saturate, indicating that η is basically unity at 10 K. This conclusion is further corroborated by the essentially identical = 5.0 ns and PL decay times at low temperature, namely, τ10K S τ10K A = 4.9 ns for the self- and Au-assisted nanowires, respectively. Nanosecond PL decay times at low T, which have already been reported for Au-assisted GaAs-(Al,Ga)As core-shell nanowires grown by MOVPE (τ10K = 1.1 ns),27 demonstrate that the decay is dominated by the radiative lifetime of free excitons.28,29 Our even longer PL decay times confirm the high quality of the interface between the GaAs core and the (Al,Ga)As shell for both nanowire types. As illustrated in Figure 2, even state-of-the-art planar heterostructures show nonradiative recombination at the interfaces. Assuming that the interface recombination is independent of T, this leads to a variation of η with 1/τr and thus approximately as T-3/2, which is close to what we observe for the self-assisted nanowires.29,30 For the Au-assisted nanowires, IA falls off more rapidly with T, indicating a strongly temperature-activated nonradiative recombination channel. Close to room temperature, IA is, as already reported above, more than 2 orders of magnitude smaller than IS. In the inset of Figure 3, the data are presented such that the activation energy EA can be determined from a linear slope similar to an Arrhenius plot.31,32 Toward lower T, both samples show the same behavior with an EA corresponding to the exciton binding energy in GaAs (4.2 meV). At higher T, however, only the Au-assisted nanowires exhibit an activated quenching with EA = 77 ( 2 meV. A nonradiative recombination center present only in the Au-assisted nanowires is thus identified. The microscopic nature of this center, i.e., whether it is due to substitutional Au, a Au-related complex, or a Au-induced secondary center, cannot be clarified in the present work. Specifics on Au-related nonradiative recombination centers in GaAs are not presently available and are probably rather complex.25,33 Furthermore, we cannot distinguish whether the center is situated in the GaAs core or at the core-shell interface. However, for all practical purposes, these details are quite inconsequential. Instead, it is most important that the detrimental center is not present in the self-assisted nanowires, since we invariably observe a high internal quantum efficiency for nanowires grown by this technique. At the same time, all Au-assisted nanowires we have investigated, independent of the details of the growth conditions, suffer from very strong nonradiative recombination at room temperature. 1278

dx.doi.org/10.1021/nl104316t |Nano Lett. 2011, 11, 1276–1279

Nano Letters To summarize and conclude, only the self-assisted nanowires exhibit a room-temperature minority carrier lifetime (and thus an internal quantum efficiency) comparable to the expectations for state-of-the-art material quality. The minority carrier lifetime for the Au-assisted nanowires is more than 2 orders of magnitude shorter than that, and this difference is directly reflected in the PL intensity of these structures. We interpret this drastic quenching as being due to the non-negligible incorporation of Au, an efficient nonradiative center, into nanowires grown under the assistance of Au droplets. Thus, the use of Au-assisted GaAs nanowires for optoelectronic applications appears questionable, while their self-assisted counterparts do have promise.

’ ASSOCIATED CONTENT

bS

Supporting Information. Experimental details of the nanowire growth by MBE and their characterization by PL spectroscopy. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: steff[email protected].

’ ACKNOWLEDGMENT We are indebted to Claudia Herrmann for thoroughly calibrating the MBE machine and to Anne-Kathrin Bluhm for highquality SEM imaging. Critical reading of the manuscript by Uwe Jahn is greatly appreciated.

LETTER

(20) Ahrenkiel, R. K. In Semiconductors and Semimetals; Ahrenkiel, R. K..Lundstrom, M. S.., Eds.; Academic Press: Boston, 1993; Vol. 39, Chapter 2, p 39. (21) Schlager, J. B.; Bertness, K. A.; Blanchard, P. T.; Robins, L. H.; Roshko, A.; Sanford, N. A. J. Appl. Phys. 2008, 103, No. 124309. (22) Wolford, D. J.; Gilliland, G. D.; Kuech, T. F.; Klem, J. F.; Hjalmarson, H. P.; Bradley, J. A.; Tsang, C. F.; Martinsen, J. Appl. Phys. Lett. 1994, 64, 1416–1418. (23) Sieg, R. M.; Carlin, J. A.; Boeckl, J.; Ringel, S. A.; Currie, M. T.; Ting, S. M.; Langdo, T. A.; E A Fitzgerald, G. T.; Keyes, B. M. Appl. Phys. Lett. 1998, 73, 3111. (24) McAfee, S. R.; Tsang, W. T.; Lang, D. V. J. Appl. Phys. 1981, 52, 6165. (25) Milnes, A. G. Deep Impurities in Semiconductors; Wiley-Interscience: New York, 1973. (26) Zarem, H. A.; Sercel, P.; Lebens, J.; Eng, L. E.; Yariv, A.; Vahala, K. Appl. Phys. Lett. 1989, 55, 1647. (27) Perera, S.; Fickenscher, M. A.; Jackson, H. E.; Smith, L. M.; Yarrison-Rice, J. M.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Jagadish, C.; Zhang, X.; Zou, J. Appl. Phys. Lett. 2008, 93, No. 053110. (28) ’t Hooft, G. W.; van der Poel, W. A. J. A.; Molenkamp, L. W.; Foxon, C. T. Phys. Rev. B 1987, 35, 8281. (29) Rappel, W. J.; Feiner, L. F.; Schuurmans, M. F. H. Phys. Rev. B 1988, 38, 7874. (30) Wolford, D. J.; Gilliland, G. D.; Kuech, T. F.; Smith, L. M.; Martinsen, J.; Venkatasubramanian, R.; Ghandi, S. K.; Hjalmarson, H. P. J. Vac. Sci. Technol., B 1991, 9, 2369. (31) Chiari, A.; Colocci, M.; Fermi, F.; Li, Y.; Querzoli, R.; Vinattieri, A.; Zhuang, W. Phys. Status Solidi B 1988, 147, 421. (32) Titova, L. V.; Hoang, T. B.; Jackson, H. E.; Smith, L. M.; Yarrison-Rice, J. M.; Kim, Y.; Joyce, H. J.; Tan, H. H.; Jagadish, C. Appl. Phys. Lett. 2006, 89, No. 173126. (33) Deep centers in semiconductors: a state-of-the-art approach, 2nd ed.; Pantelides, S. T., Ed.; Gordon and Breach Science Publishers: Philadelphia, PA, 1992.

’ REFERENCES (1) Yan, R.; Gargas, D.; Yang, P. Nat. Photonics 2009, 3, 569. (2) Li, Y.; Qian, F.; Xiang, J.; Lieber, C. M. Mater. Today 2006, 9, 18. (3) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89. (4) Messing, M. E.; Hillerich, K.; Johansson, J.; Deppert, K.; Dick, K. A. Gold Bull. 2009, 42, 172. (5) Putnam, M. C.; Filler, M. A.; Kayes, B. M.; Kelzenberg, M. D.; Giuan, Y.; Lewis, N. S.; Eiler, J. M.; Atwater, H. A. Nano Lett. 2008, 8, 3109. (6) Chen, J. W.; Milnes, A. G. Annu. Rev. Mater. Sci. 1980, 10, 157. (7) Hiesinger, P. Phys. Status Solidi A 1976, 33, K39. (8) Yan, Z. X.; Milnes, A. G. J. Electrochem. Soc. 1982, 129, 1353– 1356. (9) Fontcuberta i Morral, A.; Colombo, C.; Abstreiter, G.; Arbiol, J.; Morante, J. R. Appl. Phys. Lett. 2008, 92, No. 063112. (10) Tchernycheva, M.; Harmand, J. C.; Patriarche, G.; Travers, L.; Cirlin, G. E. Nanotechnology 2006, 17, 4025. (11) Colombo, C.; Spirkoska, D.; Frimmer, M.; Abstreiter, G.; Fontcuberta i Morral, A. Phys. Rev. B 2008, 77, No. 155326. (12) Ehrhardt, A.; Wettling, W.; Bett, A. Appl. Phys. A: Mater. Sci. Process. 1991, 53, 123–129. (13) Plante, M. C.; LaPierre, R. R. J. Cryst. Growth 2008, 310, 356. (14) Breuer, S.; Hilse, M.; Trampert, A.; Geelhaar, L.; Riechert, H. Phys. Rev. B 2010, 82, No. 75406. (15) Pavesi, L.; Guzzi, M. J. Appl. Phys. 1994, 75, 4779–4842. (16) Hoang, T. B.; Moses, A. F.; Zhou, H. L.; Dheeraj, D. L.; Fimland, B. O.; Weman, H. Appl. Phys. Lett. 2009, 94, No. 133105. (17) Spirkoska, D.; et al. Phys. Rev. B 2009, 80, No. 245325. (18) Brandt, O.; Kanamoto, K.; Gotoda, M.; Isu, T.; Tsukada, N. Phys. Rev. B 1995, 51, 7029. (19) Parkinson, P.; Joyce, H. J.; Gao, Q.; Tan, H. H.; Zhang, X.; Zou, J.; Jagadish, C.; Herz, L. M.; Johnston, M. B. Nano Lett. 2009, 9, 3349. 1279

dx.doi.org/10.1021/nl104316t |Nano Lett. 2011, 11, 1276–1279