Nitride heterostructure optimization by simulation

Nitride heterostructure optimization by simulation

Journal of Crystal Growth xx (xxxx) xxxx–xxxx Contents lists available at ScienceDirect Journal of Crystal Growth journal homepage: www.elsevier.com...

1MB Sizes 3 Downloads 117 Views

Journal of Crystal Growth xx (xxxx) xxxx–xxxx

Contents lists available at ScienceDirect

Journal of Crystal Growth journal homepage: www.elsevier.com/locate/jcrysgro

Nitride heterostructure optimization by simulation ⁎

O.I. Rabinovich , S.A. Legotin, S.I. Didenko National University of Science and Technology, MISIS,Moscow, Russian Federation

A R T I C L E I N F O

A BS T RAC T

Communicated by Satoshi Uda

In current paper nanoheterostructure optimization for LED and phototransistor usage is discussed. Special doping into quantum wells and barriers by Indium atoms was investigated. By simulation improved quantum sized active region was detected which increases quantum efficiency and sensitivity upto 10%. Photoluminescence spectral curve and Peak lambda of the InGaN/GaN nanoheterostructure with different Indium concentration across wafer were investigated.

Keywords: A1. Computer simulation B1. Nitrides B2. Semiconducting materials B3. Light emitting diodes

1. Introduction AIIIBV and AIIBVI nanoheterostructures (NH) and their solutions attract attention due to their unique properties. Such materials are very interesting and useful as materials for nanoelectronic devices production, e.g. Light emitting diodes (LED), photodetectors and transistors. Today, the problem of limited color range and white LЕDs lack that previously prevented LЕD usage for general lighting have been solved, but LED/photodetectors efficiency and sensitivity need to be further improved. One of the ways to do this is the heterostructure active region optimization. For complex optoelectronic materials and devices, especially, materials with quantum wells (QW), for LEDs and photodetectors the basic parameters that determine their quality, such as currentvoltage characteristics (I–V), the quantum efficiency (QE) and photodetector sensitivity can be investigated and improved via computer simulations taking into account major structural, physical, and technological NH parameters. 2. Experimental details At present, well-developed software packages are widely used for semiconductor materials and devices computer simulation and optimization. Some representative examples for III-Nitride devices simulation are presented in [1,2]. We, in this work, used the well known simulation package Sim Windows [3] because of its ready availability and the ease of adding new materials constants files and of additional equations files. The contribution of polarization fields was considered, as it should be in III-Nitrides. This is, of course, necessary for IIINitrides devices due to the importance of the presence of considerable



strain in the structures, the impact of polarization fields, and the importance of the quantum confined Stark effect (QCSE) [4–12]. Briefly, the main points should be noted: in the Schrödinger-Poisson equation the charges of the ionised impurity atoms, the free charge carriers and the associated charge carriers in QW are taken into account together with the polarization field and QCSE. The materials parameters for these calculations were taken from [13–20]. The growth was assumed to take place in the polar c-direction [0001]. The general view of heterostructure is shown in Fig. 1. In the program, the expression for the three current types are used: the drift-diffusion current in the areas of the device, the thermionic emission current contacts for quantum wells with bulk materials, and the thermionic and tunneling currents at abrupt boundaries between the two bulk materials. For the charge carriers there are different recombination mechanisms: spontaneous and stimulated ones, radiative “band-band” and the non-radiative recombination mechanisms in the models of Shockley-Read-Hall and Auger are taken into consideration in the program. In addition, for the simulation of current transport and recombination in the QW structures, LED, photodetectors special files were created. These files included the geometric dimensions of the emitters, quantum wells, and barriers; the number of quantum wells and barriers; the solid solution concentration; the conductivity type; concentration; and the impurity activation energy in each of the NH areas. In the materials file for solid solutions, more than 25 parameters such as the band gap, refractive index, optical absorption, thermal conductivity, mobility, and lifetime of charge carriers, electron affinity, and the coefficients of radiative and non-radiative recombination were included.

Correspondence to: NUST MISiS, P.O. Box 409, Moscow 119313, Russian Federation. E-mail address: [email protected] (O.I. Rabinovich).

http://dx.doi.org/10.1016/j.jcrysgro.2016.11.067

Available online xxxx 0022-0248/ © 2016 Elsevier B.V. All rights reserved.

Please cite this article as: Rabinovich, O., Journal of Crystal Growth (2016), http://dx.doi.org/10.1016/j.jcrysgro.2016.11.067

Journal of Crystal Growth xx (xxxx) xxxx–xxxx

O.I. Rabinovich et al.

Fig. 1. Investigated InGaN/GaN heterostructure.

predominate when η→1 at a low injection level (J=1–20 A/cm2) and η→2 with increasing injection level (J=20–500 A/cm2). The QW presence begins to affect the form of the I–V plot at values X=0.05– 0.1, especially in the range of Х=0.1–0.35. For Х > 0.1 the value of η gradually increases over the entire range J=0.1–500 A/cm2, reaching values of η > 2 and even higher for X > 0.15. The nonideality coefficient decreases with increasing donor impurity concentration in the barrier for the same values of X and j (Fig. 3b). It was detected that the optimum impurity concentration in barriers between QWs is Nd=1018 cm−3. The QE and sensitivity increaseat the optimumAl atom concentration is about 10%. This doping shifts the I– V plot to the lower-voltage region and increases the QE (Fig. 4). This effect is due to reducing potential barrier (additional carriers injection) between QWs and barriers among them so j increases at a constant voltage. By varying Indium and doping concentrations in AlGaInN heterostructure active region, it is possible to increase QE and sensitivity at the same voltage. Fig. 5 showssatisfactoryagreement between the simulation andexperimental results and it is evident despite the fact thatthe simulation resultsare obtained withoutany additionalapproximations (abovethe basephysical models). For comparison, it was considered the experimental current-voltage characteristics measured at the LED manufacturing company Cree Inc. for the type C460MB290E1000. It was found that the doped barriers in the quantum-well active region of the heterostructure at Nd=1018 cm−3, and further doping of In atomsatthe 5% levelsubstantiallyreduce the nonidealitycoefficient characteristics of theLEDs and increase QE. Additionally it was found that such structure shifts I-V to a lower-voltage region. Based on simulation recommendations AlGaInN heterostructures were grown by MOCVD on SiC and Al2O3 substrates in the polar cdirection [0001]. Analyzing growth results (e.g. photoluminescence spectral curve and Peak lambda) it was detected that the characteristics had similar trend. The spectral mapping for heterostructure obtained by photoluminescence mapping system for peak lambda is shown in Fig. 6. Fig. 6a shows different Indium atoms concentration and distribution over wafer. Fig. 6b shows the investigation results of NH which was grown according simulation results. It can be seen the improved plane Indium atoms distribution and signal value more smooth according special doping by indium atoms. For QE and sensitivity rise it is need to improve NH quality to more stable indium atoms distribution in QW/barriers and GaN substrates usage for reduction defects quantity. Then suggested active area structure was checked by using for photodetector.

In current paper based on previous simulation results such as – the selected NH structure should be p+GaN/p+Al0.2Ga0.8N/4 (nInXGa1−XN-n-GaN)/n+GaN with four 3.5 nm QWs and that it was also indicated that the largest contribution to the characteristics were from two central QWs and two edge QWs that act as quasi-buffers simultaneously accumulate and further inject the charge carriers [21–25] investigation for NH optimization due to impurity and Indium atoms influence in quantum active region, was carried out. 3. Results and discussion The p–Al0.2Ga0.8N-emitter inclusion into NH is due to the needfor eliminationthe electrons injectionfrom theactive region, which is especially importantin devices simulation with low content In (X) atoms (Fig. 2). Next, based on the optimized heterostructure, the effect of the impurity and In atoms doped into the barriers between quantum wells in the heterostructure active region was studied. This effect was represented by the nonideality coefficient dependence (Fig. 3). Without QWs in the active area (X=0), the I-V dependence was standard. At low current densities, up to J=0.1 A/cm2, η > 1, a significant influence on the current processes value of electrons and holes recombination in the space charge region (SCR) is observed. Then, the above-barrier-current carrier injection increasingly begins to

Fig. 2. Al atoms influence in the p-emitter on the InGaN LED QE.

2

Journal of Crystal Growth xx (xxxx) xxxx–xxxx

O.I. Rabinovich et al.

Fig. 3. Nonideality coefficient (η) vs the current dependence: a) with different contents of In atoms in quantum-sized holes; b) type of characteristic dependence with different doping barriers between quantum wells: 1–1017 cm−3, 2–5·1018 cm−3; X=20: 3–1017 cm−3, 4–5·1018 cm−3; X=10: 5–1017 cm−3, 6–5·1018 cm−3.

thickness – 0.3 µm. Acceptor concentration in p-Al0.3Ga0.7N collector was 1017 cm−3, donors in n- Al0.3Ga0.7N emitter was 1017 cm−3, the acceptor concentration in the p-GaN base – 1017 cm−3 or 1018 cm−3. For the dark current density value JPh, A/cm2 determination at different voltage U between the PT emitter and collector (plus on pAl0.3Ga0.7N collector), the voltage range was U=(1–9) V. For the photocurrent density value Jf (A/cm2) determination at various lighting power density values P (W/cm2) was P=(10−6–10−1) and voltage between the PT emitter and the collector (plus on pAl0.3Ga0.7N collector) was U=(1−9) V. The energy of quanta during lighting was E=3.5 eV. Additionally to special indium doping the Al mole fraction in the collector and emitter was varied from X=0.2 to X=0.3. Dark current density JPh (A/cm2) dependence versus the voltage PT is shown in Fig. 7. It is seen that in the PT at an acceptors concentration in the base Na=1017 cm−3, dark current begins to rise sharply at a voltage greater than 8 V, which is obviously corresponds with start of the clamping of the emitter and collector (Early's effect). When the acceptor concentration in the base Na=1018 cm−3, this effect occurs at higher voltages. At voltages less than 8 V dark current density does not exceed 10−8 A/cm2. This value limits the minimum magnitude of photocurrent density and, consequently, the minimum value of the recorded radiation power. Fig. 8 shows that at the acceptors concentration in the base Na=1018 cm−3, the PT sensitivity has a large value in a wide range of lighting, especially in the region of small power values P. At the same time, at the acceptor concentration in the base Na=1017 cm−3 the PT sensitivity is maximum at relatively larger values of the power P, which makes them very promising in a plenty of applications. Fig. 9 shows the PT sensitivity dependence versus voltage in the range from 1 to 9 V at a power density of light 1 mW/cm2, which is typical in many applications of UV photodetectors. Conclusion of this dependence is quite obvious - to obtain a high sensitivity versus voltage, applied to the PT, it must be in the range from 6 to 9 V. At the end of the discussion of UV phototransistor characteristics simulating Fig. 10 shows data of the PT sensitivity spectral dependence. It is clearly seen that the PT sensitivity is very high in the range of photon energies from 3.5eV to 4 eV (wavelength range from 354 nm to 309 nm), which allows to use them as selective photodetectors. PT selectivity can be increased by reducing the aluminum concentration in the p-Al0.3Ga0.7N collector up to 20%. Sensitivity also can be increased by improving MOCVD technology for more high quality AlGaN multilayer structures (with minimum defect concentration). If the sapphire substrate could be replaced by a gallium nitride substrate grown on

Fig. 4. I–V at different In atoms concentration in barriers between QWs.

Fig. 5. The current-voltage characteristics: 1- experimental data, 2- simultion results.

A symmetric structure n-AlGaN (In)/p-GaN/n-AlGaN (In) for photodetector sample was used. During heterostructure simulation for photodetector (phototransistor (PT)) with high efficiency and sensitivity its structure consisted of the p-Al0.3Ga0.7N collector, nAl0.3Ga0.7N emitter with an aluminum content of 30% and a p-GaN base. The emitter and collector thickness was 0.875 µm, the base 3

Journal of Crystal Growth xx (xxxx) xxxx–xxxx

O.I. Rabinovich et al.

Fig. 6. Photoluminescence spectral curve and Peak lambda of the InGaN/GaN Het. with different Indium concentration across wafer: a- standard growth, b- according simulation results.

Fig. 7. Dark current density jph vs the PT voltage.

Fig. 8. Phototransistor sensitivity vs light power @ U=9 V.

sapphire, the lifetimes of nonradiative recombination in the PTs base will be increased significantly.

teristics of theLEDs, increase QE and phototransistor sensitivity. It was also detected that by doping I-V could be shifted to a lower-voltage area. The AlXGa1−XN heterostructures characteristics computer simulation for the most important types of UV photodetectors (photodiode with p-n junction and the phototransistor) was carried out. The possibility of obtaining their high sensitivity and selectivity in a wide wavelength range was shown.

4. Conclusion It was foundthat thebarriers doped byIndium atomsatthe 5–7% level and additional inclusion impurity at Nd=1018 cm−3 in the quantum-well active region reduce the nonideality coefficient charac4

Journal of Crystal Growth xx (xxxx) xxxx–xxxx

O.I. Rabinovich et al.

Fig. 9. Phototransistor sensitivity vs voltage, @ the power illumination density of 1 mW/cm2.

Fig. 10. Phototransistor sensitivity vs quanta energy E, eV @ P=10−3 W/cm2 and U=9 V. [10] M.F. Schubert, E.F. Schubert, Appl. Phys. Lett. 96 (2010) 131102–1131103. [11] Qi Dai, Q. Shan, J. Wang, S. Chhajed, J. Cho, E.Fred Schubert, A.Mary Crawford, D.D. Koleske, Min-ho Kim, Y. Park, Appl. Phys. Lett. 97 (2010) 133507–133511. [12] M. Auf der Maur, J. Comput. Electron. 14 (2) (2015) 398–408. [13] O. Ambacher, B. Foutz, J. Smart, J.R. Shealy, N. Weimann, K. Chu, M. Murphy, A.J. Sierakowski, W.J. Schaff, L. Eastman, J. Appl. Phys. 87 (2000) 334–339. [14] J. Vurgaftman, J.R. Meyer, J. Appl. Phys. 94 (2003) 3675–3683. [15] T. Takeuchi, H. Amano, I. Akasaki, Jpn. J. Appl. Phys. 39 (1) (2000) 413–416. [16] B. Monemar, G. Pozina, Prog. Quantum Electron. 24 (2000) 239–290. [17] C. Wetzel, T. Takeuchi, H. Amano, I. Akasaki, III-Nitride Semiconductors: Optical Properties, Taylor & Francis, New York, 2002. [18] O. Ambacher, J. Majewski, C. Miskys, A. Link, M. Hermann, M. Eickhoff, M. Stutzmann, F. Bernardini, V. Fiorentini, J. Phys.: Condens. Matter 14 (2002) 3399–3434. [19] Hadis Morkoc, Handbook of Nitride Semiconductors And Devices: Materials Properties, Physics and Growth, Wiley-VCH Verlag GmbH and Co. KGaA, Weinheim, 2009. [20] X. Huang, Ch Du, Y. Zhou, Ch Jiang, X. Pu, W. Liu, W. Hu, H. Chen, Zh Lin Wang, ACS Nano 10 (5) (2016) 5145–5152. [21] O.I. Rabinovich, V.P. Sushkov, Semiconductors 43 (4) (2009) 524–527. [22] O.I. Rabinovich, S.G. Nikiforov, V.P. Sushkov, A.V. Sishov, SPIE. Physics and Simulation of Optoelectronic Devices 6468, 2007, pp. 64680U1–64680U10. [23] O.I. Rabinovich, N.V. Romanov, S.S. Sizov, Light Eng. 4 (2009) 92–96. [24] O. Rabinovich, S. Didenko, S. Legotin, Adv. Mater. Res. 1070–1072 (2015) 600–603. [25] O.I. Rabinovich, S.I. Didenko, S.A. Legotin, I.V. Fedorchenko, U.V. Osipov, J. Nano– Electron. Phys. 7 (4) (2015) 040351–040353.

A new phototransistor structure with disabled base on the basis of a symmetrical heterostructure with special indium atoms doping n Al0.3Ga0.7N (In) – p GaN – n Al0.3Ga0.7N (In) was suggested. Simulation shows the possibility of creating a device with a sensitivity of 500–700 A/W based on structures with electrical and structural parameters, typical for serial production. References [1] Simulation package, 〈http://www.str-soft.com/products/SILENSE〉. [2] Simulation package, 〈http://www.str-soft.com/products/FETIS/〉. [3] D.W. Winston, Physical simulation of optoelectronic semiconductor devices, University of Colorado, Colorado (1999). [4] J. Piprek, Nitride Semiconductor Devices. Principles and Simulation, Wiley-VCH Verlag GmbH and Co. KGaA, Weinheim, 2007. [5] S.Yu. Karpov, J.I. Chyi, Y. Nanishi, H. Morkoç, J. Piprek, E. Yoon, SPIE. Gallium Nitride Materials and Devices VI 7939, 2011, pp. 79391C1– 79391C12. [6] W. Tian, J. Zhang, Zh Wang, F. Wu, Y. Li, Sh Chen, J. Xu, J. Dai, Y. Fang, Zh Wu, Ch Chen, IEEE J. Photonics 5 (6) (2013) 55–63. [7] J.R. Chen, Y.C. Wu, S.C. Ling, T.S. Ko, T.C. Lu, H.C. Kuo, Y.K. Ku, S.C. Wang, Appl. Phys. B 98 (2010) 779–789. [8] M.H. Kim, M.F. Schubert, Q. Dai, J.K. Kim, E.F. Schubert, J. Piprek, Y. Park, Appl. Phys. Lett. 91 (18) (2007) 183507–183510. [9] K.A. Bulashevich, V.F. Mymrin, S.Y. Karpov, I.A. Zhmakin, A.I. Zhmakin, J. Comput. Phys. 213 (1) (2006) 214–238.

5