Academia.eduAcademia.edu

AlGaN/AlN quantum dots for UV light emitters

2013, physica status solidi c

We report on the growth and properties of two types of AlGaN/AlN nanostructures: Stranski‐Krastanow quantum dots (SK‐QDs) and nanodisks (NDs) created by nanowire heterostructuring. In both cases, the emission wavelength can be tuned in the range of 240‐350 nm at room temperature by varying the flux ratio and the nominal amount of AlGaN in the nanostructures. The efficient carrier confinement in these nanostructures leads to an internal quantum efficiency around 0.5. However, the emission spectra of AlGaN/AlN NDs show broader linewidth than those of SK‐QDs, which is attributed to inhomogeneities in height and in chemical composition. (© 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

AlGaN/AlN quantum dots for UV light emitters solidi status pss physica Phys. Status Solidi C 10, No. 3, 285–288 (2013) / DOI 10.1002/pssc.201200679 c www.pss-c.com current topics in solid state physics C. Himwas*,1, M. den Hertog1, F. Donatini1, Le Si Dang1, L. Rapenne2, E. Sarigiannidou2, R. Songmuang1, and E. Monroy3 1 CEA-CNRS Group “Nanophysique et Semiconducteurs”, Institut Néel-CNRS, 25 rue des Martyrs, 38042 Grenoble Cedex 9, France INP-Grenoble/Minatec, 3 parvis Louis Néel, BP257, 38016 Grenoble, France 3 CEA-CNRS Group “Nanophysique et Semiconducteurs”, INAC-SP2M, CEA-Grenoble, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France 2 Received 11 September 2012, revised 20 December 2012, accepted 20 December 2012 Published online 12 February 2013 Keywords ultraviolet, quantum dot, nanowire, GaN * Corresponding author: e-mail chalermchai.himwas@cea.fr We report on the growth and properties of two types of AlGaN/AlN nanostructures: Stranski-Krastanow quantum dots (SK-QDs) and nanodisks (NDs) created by nanowire heterostructuring. In both cases, the emission wavelength can be tuned in the range of 240-350 nm at room temperature by varying the flux ratio and the nomi- nal amount of AlGaN in the nanostructures. The efficient carrier confinement in these nanostructures leads to an internal quantum efficiency around 0.5. However, the emission spectra of AlGaN/AlN NDs show broader linewidth than those of SK-QDs, which is attributed to inhomogeneities in height and in chemical composition. © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1 Introduction The development of solid-state ultraviolet (UV) emitters would imply a mayor advance for applications like water/air purification, bio-detection, or phototherapy. In the recent years, AlGaN-based UV lightemitting diodes (LEDs) [1, 2] have appeared as promising candidates to replace mercury lamps due to their advantages of compactness, portability, long lifetime and environmental friendliness. Although AlGaN/AlN quantum wells (QWs) with high internal quantum efficiency (IQE) have been reported [3-5], the fabrication of electrically-driven UV LEDs must face the challenges of p-type doping and contacting high-Al-content AlGaN layers, which result in extremely low external quantum efficiency [6] (EQE < 1% for UV LEDs emitting below 250 nm). Electron-pumped UV (EPUV) sources incorporating a miniaturized electron source [7] have been introduced as an alternative approach to circumvent the p-doping issue. A key requirement for EPUV devices is an active media with high IQE at room temperature. This fact motivates the use of AlGaN/AlN nanostructures, where the threedimensional (3D) carrier confinement grants certain insensitivity to non-radiative recombination processes. In this work, we present the growth and properties of two types of AlGaN/AlN nanostructures, namely Stranski- Krastanow quantum dots (SK-QDs) and nanodisks (NDs) created by nanowire (NW) heterostructuring. The assessment and comparison of these nanostructures are relevant for the EPUV application. Nominally, both approaches present similar 3D carrier confinement, but SK-QDs provide a better QD size uniformity along the growth axis, whereas the ND geometry ensures better light extraction efficiency. 2 Experimental AlGaN/AlN 100-period QD superlattices (SLs) were synthesized on 1-µm-thick AlN-onsapphire templates via plasma-assisted molecular-beam epitaxy (PAMBE). The flux of active nitrogen was fixed at ΦN = 0.32 ML/s, and the substrate temperature (TS) was controlled to be in the range of TS = 720-745 °C. The growth of AlGaN SK-QDs was performed under N-rich conditions, since the nitrogen excess reduces the mobility of adsorbed species during growth, resulting in a high density of small QDs [8-10]. The Al-to-metal flux ratio, ΦAl/(ΦAl + ΦGa), was varied from 0.12 to 0.42. Under these conditions, AlGaN growth starts two dimensionally until the deposition of a ~2-ML-thick wetting layer, and further growth results in the formation of 3D islands (StranskiKrastanow growth mode). The N-rich deposition of 4© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim solidi physica c status pss 286 C. Himwas et al.: AlGaN/AlN quantum dots for UV light emitters Table 1 Description of the QD samples under study. Table 2 Description of the ND samples under study. Sample QD-1 QD-2 QD-3 QD-4 QD-5 QD-6 QD-7 TS 720°C 720°C 720°C 720°C 720°C 745°C 745°C ΦAl/(ΦAl+ΦGa) 0 0.14 0.24 0.33 0.42 0.12 0.21 Amount of AlGaN in each layer Emission wavelength at 300 K 4 ML 5 ML 6 ML 6 ML 7 ML 4 ML 4 ML 340 nm 287 nm 280 nm 274 nm 268 nm 256 nm 245 nm 7 ML of AlGaN was followed by a 10 s growth interruption in vacuum, before capping with ~4 nm of AlN grown under slightly Al-rich conditions, in order to achieve a good planarization before the next QD layer. The most relevant growth parameters for the QDs [TS, ΦAl/(ΦAl + ΦGa), and amount of AlGaN in each QD layer] are summarized in Table 1. To confirm the presence of QD structures via atomic force microscopy (AFM) characterization, an additional QD plane was deposited on the sample surface. AFM measurements were performed in Veeco Dimension 3100 system using the tapping mode. The ND samples under study consist of 15-30 stacks of AlGaN/AlN NDs on the top of GaN NWs grown on Si(111) by PAMBE [11]. Initially, a thin AlN buffer layer was grown at high temperature [11]. Then, GaN NWs were grown at ~790°C under N-rich conditions [12] with ΦGa= 0.17 ML/s and ΦN = 0.34 ML/s for 2.5 hours. Afterwards, AlGaN NDs with AlN barriers were grown on the top of the GaN NWs. For the AlGaN NDs, the Al-to-N flux ratio (ΦAl/ΦN) was adjusted from 0.21 to 0.50 for the various samples in the series. This ΦAl/ΦN ratio sets a lower limit of Al content in the AlGaN NDs. The precise Al mole fraction cannot be easily determined since the Ga incorporation in AlGaN NDs is not only due to the impinging Ga flux (ΦGa) but also to the Ga diffusion along the NW sidewalls [13], and there is a significant Ga desorption at this growth temperature which should also be taken to account. The nominal height of AlGaN NDs and AlN barriers are 12 nm and ~4 nm respectively, which are estimated by ΦN⋅t, where t is the deposition time [14]. The most relevant growth parameters for the NDs (TS, ΦAl/ΦN, and nominal ND thickness) are summarized in Table 2. After the growth, the structural properties were characterized by high-resolution scanning transmission electron microscopy (HR-STEM) and scanning electron microscopy (SEM). Photoluminescence (PL) measurements were performed by using a continuous-wave frequency-doubled argon laser (wavelength λ = 244 nm) as an excitation source. The PL signal was collected by a Jobin Yvon HR460 monochromator equipped with an ultraviolet-enhanced chargecoupled device (CCD) camera. Cathodoluminescence (CL) measurements were performed in a FEI quanta 200 CL system, using an acceleration voltage of 10 kV and a cur© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim * Sample TS ΦAl/ΦN Nominal ND thickness Peak emission at 5 K ND-1* ND-2 ND-3 ND-4 790°C 790°C 790°C 790°C 0.00 0.45 0.35 0.45 2 nm 2 nm 1 nm 1 nm 330 nm 295 nm 258 nm 241, 259 nm** Sample of GaN/Al0.37Ga0.63N NDs. Multiple peaks with similar intensity. ** rent of 100-200 pA. The CL emission is collected by a parabolic mirror which focuses it into a CCD camera. In the case of time-resolved PL, samples were excited using a frequency-tripled Ti:sapphire laser (λ = 270 nm) with pulse width of 200 fs and repetition rate of 76 MHz. The luminescence was dispersed into a Triax320 monochromator and was detected by a steak-camera using a 2.2 ns window, giving a system response of about 10 ps. In all the cases, the excitation power density was kept low enough to avoid any screening of the internal electric field. 3 Results and discussion The formation of 3D nanostructures during the growth of the SK-QD samples is verified in-situ by reflection high-energy electron diffraction. To confirm ex-situ the presence of QDs, we have performed AFM characterization with the results illustrated in the inset of Fig. 1. The QD height is 1-2 nm above the wetting layer, with the base diameter of ~7 nm, confirmed by cross-sectional high-resolution transmission electron microscopy (not shown). The QD density is in the range of 1011-1012 cm-2. Figure 1 Room-temperature PL and CL spectra from AlGaN/AlN QDs. Inset: AFM image of AlGaN/AlN QDs synthesized by deposition of 6 ML of AlGaN with a metal flux ratio ΦAl/(ΦAl + ΦGa) = 0.35. Figure 1 shows the emission spectra of the QD samples obtained at room temperature, measured by PL (QD-2, -3, -4, and -5 in Table 1) or CL (QD-1, -6, and -7 in Table 1). For PL measurements, the excitation energy is below the AlN band gap, and therefore the emission spectra presents a superimposed oscillation due the Bragg interference associated to the nitride layer thickness. In contrast, for CL www.pss-c.com Contributed Article 287 Phys. Status Solidi C 10, No. 3 (2013) measurements electron-hole pairs are generated both in the QDs and in the surrounding AlN matrix, resulting in Guassian-like spectra. By varying the deposition parameters as described in Table 1, the peak emission wavelength can be tuned to cover the range of 245-320 nm with full width at half maximum (FWHM) in the range of 0.19-0.26 eV. Table 2) in Fig. 3(b) shows 2 nm-thick AlGaN NDs and 4 nm-thick AlN barriers with abrupt hetero-interfaces. Figure 3 (a) SEM image of GaN NWs with AlGaN/AlN NDs on Si(111) and (b) HR-STEM image of 2 nm-AlGaN/4nm-AlN NDs (ND-2 in Table 2). Figure 2 Evolution of the integrated PL intensity from AlGaN/AlN SK-QDs with temperature. Inset: Evolution of the PL decay time as a function of the measuring temperature. Figure 2 shows the normalized integrated PL intensity as a function of temperature. The internal quantum efficiency (IQE) can be calculated by [8]: IQE = IPL(T = 300 K) / IPL(T = 5 K) (1) This equation is only valid if the integrated PL intensity is stable at low temperature, i.e. if non-radiative processes can be neglected at 5 K. For all the SK-QD samples, the integrated PL intensity remains stable below 75 K, and thus we can estimate an IQE around 40%-60%. The activation energy (Ea) extracted from the temperature-dependent PL measurements [8] is in the range of 50-80 meV. The high values of IQE and Ea are comparable to previously reported values for GaN/AlN QDs and significantly higher than those of GaN/AlN QW samples [8,10], which confirm the efficient 3D carrier confinement in AlGaN/AlN QDs. The radiative recombination efficiency of the QD samples was also assessed by time-resolved PL. The PL decay [15] is mono-exponential at low temperature (T = 5 K), with a decay time that decreases from 495 to 359 ps when ΦAl/(ΦAl + ΦGa) increases from 0.14 to 0.42. The longer decay time at low ΦAl/(ΦAl + ΦGa) is attributed to the larger separation of electron-hole wavefunctions induced by the higher polarization-induced internal electric field. The decay time decreases by less than 30% between T = 5 K and T = 300 K (inset of Fig. 2), confirming that the 3D confinement in QDs prevents the carriers escape towards to non-radiative recombination centers. Figure 3(a) shows a typical SEM image of GaN NWs with AlGaN/AlN NDs. The length of the NWs is about 900 nm and their diameter ranges from 30 to 60 nm. The depicted HR-STEM image of AlGaN/AlN NDs (ND-2 in www.pss-c.com The CL spectra of ND samples measured at 5 K are presented in Fig. 4. By varying the growth conditions as described in Table 2, the ND emission peak can be tuned down to 240 nm. The luminescence from GaN/Al0.37Ga0.63N NDs (ND-1) peaks at 330 nm. To extend the emission to shorter wavelengths, the ND heterostructure is changed to AlGaN/AlN (ND-2, -3, and -4). In the case of ND-2, with the same nominal ND thickness as ND-1 (2 nm), the main CL spectral peak blue shifts to 295 nm, resulting from the higher fundamental energy gap of AlGaN. To further reduce the emission wavelength, the nominal ND thickness was decreased from 2 nm to 1 nm (ND-3, -4), hence increasing the quantization energy in the NDs. The inset of Fig. 4 shows the CL spectra of ND-2 measured at 5 K and 300 K. At low temperature, besides the ND luminescence appearing at 295 nm, an intense peak at around 380 nm was observed, which is assigned to the emission from the GaN NW base. Using Eq. (1), we deduced that the IQE of this AlGaN NDs was ~25%, significantly higher than that of the emission from GaN NW base (IQE < 1%) because of the efficient carrier confinement in the NDs. Comparative measurements of the integrated CL intensity at 5 K and 300 K of other AlGaN NDs in our studies reveal IQEs in the 25%-40% range. The similar values of IQE of SK-QDs and NDs are consistent with comparable integrated emission intensity, which indicate that both approaches provide similar 3D carrier confinement. Considering the spectral shape, the luminescence of AlGaN NDs is structured in multiple peaks and presents broader linewidth than that of GaN NDs and SK-QDs. From HR-STEM measurements (not shown here), we expect that these multi-peak spectra (ND-2,-3, -4) result from © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim solidi physica c status pss 288 C. Himwas et al.: AlGaN/AlN quantum dots for UV light emitters a phase separation in AlGaN NDs which seems to be enhanced by decreasing the ND thickness. This alloy inhomogeneity, not observed in the SK-QD layers, might originate from the adatom surface kinetics of the NW growth. In addition, the shadow effect during the deposition of the upper layers can results in a variation of the ND height along the structure [14]. [8] Ž. Gačević, A. Das, J. Teubert, Y. Kotsar, P. K. Kandaswamy, Th. Kehagias, T. Koukoula, Ph. Komninou, and E. Monroy, J. Appl. Phys. 109, 103501 (2011). [9] J. Renard, P. K. Kandaswamy, E. Monroy, and B. Gayral, Appl. Phys. Lett. 95, 131903 (2009). [10] F. Guillot, E. Bellet-Amalric, E. Monroy, M. Tchernycheva, L. Nevou, L. Doyennette, F. H. Julien, Le Si Dang, T. Remmele, M. Albrecht, T. Shibata, and M. Tanaka, J. Appl. Phys. 100, 044326 (2006). [11] R. Songmuang, O. Landré, and B. Daudin, Appl. Phys. Lett. 91, 251902 (2007). [12] E. Calleja, M. A. Sánchez-Gracía, F. J. Sánchez, F. Calle, F. B. Naranjo, E. Muñoz, U. Jahn, and K. Ploog, Phys. Rev. B 62, 16826 (2000). [13] R. Songmuang, T. Ben, B, Daudin, D. González, and E. Monroy, Nanotechnology 21, 295605 (2010). [14] R. F. Allah, T. Ben, R. Songmuang, and D. González, Appl. Phys. Express 5, 045002 (2012). [15] The decay time is defined by an exponential fit of the PL decay in the range where the PL intensity decreases from 90% to 10% of its maximum value. Figure 4 CL spectra at 5 K from GaN/Al0.37Ga0.63N NDs (ND-1) and AlGaN/AlN NDs (ND-2, -3, and -4). Inset: CL spectra of ND-2 measured at 5 K and 300 K. 4 Conclusions We report on the growth and properties of AlGaN/AlN SK-QDs and AlGaN/AlN NDs in NWs. In both cases, the nanostructures emitting in the 240350 nm spectral range are demonstrated. Nominally, both approaches present similar 3D carrier confinement. In comparison with the luminescence from SK-QDs, NDs present broader emission spectra, in most cases consisting of multiple peaks, which are attributed to alloy inhomogeneities and ND height fluctuations. However, the better light extraction efficiency of the NDs in NWs is envisioned. Acknowledgements This work is supported by the French National Research Agency project UVLamp (BLAN081_323691). References [1] Y. Taniyasu, M. Kasu, and T. Makimoto, Nature 441, 325 (2006). [2] H. Hirayama, T. Yatabe, N. Noguchi, T. Ohashi, and N. Kamata, Appl. Phys. Lett. 91, 071901 (2007). [3] A. Bhattacharyya, T. D. Moustakas, L. Zhou, D. J. Smith, and W. Hug, Appl. Phys. Lett. 94, 181907 (2009). [4] W. Zhang, A. Y. Nikiforov, C. Thomidis, J. Woodward, H. Sun et al., J. Vac. Sci. Technol. B 30, 02B119 (2012). [5] H. M. Huang, C. Y. Chang, Y. P. Lan, T. C. Lu, H. C. Kuo et al., Appl. Phys. Lett. 100, 261901 (2012). [6] A. Khan, K. Balakrishnan, and T. Katona, Nature Photon. 2, 77 (2008). [7] Y. Shimahara, H. Miyake, K. Hiramatsu, and F. Fukuyo, Appl. Phys. Express 4, 042103 (2011). © 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.pss-c.com