GaN multilayers: Effect on crystalline quality of a-plane GaN

Materials Science in Semiconductor Processing 105 (2020) 104700

Contents lists available at ScienceDirect

Materials Science in Semiconductor Processing journal homepage: www.elsevier.com/locate/mssp

Positioning of periodic AlN/GaN multilayers: Effect on crystalline quality of a-plane GaN

T

Anas Kamarudzaman∗∗, Ahmad Shuhaimi Bin Abu Bakar∗, Adreen Azman, Al-Zuhairi Omar, Azzuliani Supangat, Noor Azrina Talik Low Dimensional Materials Research Centre, Department of Physics, Faculty of Science, University of Malaya, 50603, Kuala Lumpur, Malaysia

A R T I C LE I N FO

A B S T R A C T

Keywords: a-plane GaN Non-polar GaN AlN/GaN multilayers MOCVD

We report the effect of 40 pairs of periodic AlN/GaN multilayers on the a-plane undoped-gallium nitride (udGaN) grown on r-plane flat sapphire substrate via metal-organic chemical vapor deposition. The influence of the position of the periodic AlN/GaN multilayers was seen to enhance the crystalline quality and surface morphology of the a-plane ud-GaN. The surface morphology analysis via atomic force microscopy has shown that the surface roughness was as low as 1.01 nm upon the insertion of the 40 pairs of periodic AlN/GaN multilayers with an optimum position. The on- and off-axis x-ray ω-scan rocking curves illustrate the enhancement in crystalline quality with a reduction of the full width at half maximum from 0.34° to 0.25° along [0001] direction and 0.91°–0.47° along [1–100] direction. The grown a-plane GaN with periodic AlN/GaN multilayers was seen to exhibit different relaxation strain states at different position as seen from the Raman spectroscopy. Roomtemperature photoluminescence spectra shows that the sample with optimum periodic AlN/GaN multilayers position exhibits the lowest yellow and blue luminescence band.

1. Introduction

effectiveness [15,16]. However, growing a-plane GaN is very challenging as the growth on foreign substrates suffers from poor crystal quality and morphological properties. Several groups reported that the growth of a-plane GaN on r-sapphire exhibited high densities of basal stacking faults (BSF) (105-106 cm−1) and threading dislocations (TD) (109-1010 cm−2) [17–19]. This is mainly caused by a huge lattice mismatch, different in thermal expansion coefficients and different growth rate within c- and m-plane direction of the a-GaN growth. Epitaxial lateral overgrowth (ELOG) or lateral epitaxial overgrowth (LEO) has been widely used to overcome the challenge in growing aplane GaN. However these techniques involve more than one processes as they require a patterning of SiNx nano-mask outside the epitaxial growth process [3,6,20]. The use of periodic multilayers (MLs) seems to be a better option as MLs and GaN growth can be done in a single epitaxial process. In addition to its simple one step process, the insertion of MLs contributes to huge crystal improvement especially in strain management in which having a role as effective dislocation filter. Xu et al. reported that AlN/AlGaN superlattices are able to reduce the strain within the growth by providing a thin layer filter with significantly high lattice mismatch before the overgrowth of thick a-plane GaN [21]. Thus far, the impact of the periodic AlN/GaN multilayers

III-Nitride semiconductors have been attracting major attentions due to their large tunable bandgap in optoelectronic devices. The conventional nitride based optoelectronics devices are mostly grown on c-plane crystal direction, which has reached remarkable performance over the past decades [1–3]. However, the polar plane within the wurtzite crystal structure suffers a huge built-in electric field from the spontaneous and piezoelectric polarization, which causes an energy band bending in the energy band diagram. This would result in the reduction of recombination rate due to a reduced electron-hole overlap in the quantum well [4–8]. Motivated to avoid the undesirable polarization effect, the growth of non-polar (11–20) a-plane GaN was explored. This is mainly due to its position that is perpendicular to polar plane which make it free from polarization effect and possess the highest indium incorporation in InGaN growth for the light emitting device's active layer [9,10]. There are several alternative in substrate selection for growing aplane GaN including r-plane flat sapphire substrate, r-plane patterned sapphire substrate and a-plane SiC substrate [11–14] where r-plane flat sapphire substrate can be considered as the best selection in term of cost

∗

Corresponding author. Corresponding author. E-mail addresses: [email protected] (A. Kamarudzaman), [email protected] (A.S.B. Abu Bakar).

∗∗

https://doi.org/10.1016/j.mssp.2019.104700 Received 23 April 2019; Received in revised form 31 July 2019; Accepted 27 August 2019 1369-8001/ © 2019 Published by Elsevier Ltd.

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

Fig. 1. (a–c) Schematic diagram for sample A-C and (d–f) process flow and structure comparison between samples A,B and C.

The microstructural properties of the grown a-plane GaN samples were explored using Hitachi SU8220 field emission scanning electron microscopy (FESEM). The scanning transmission electron microscopy (STEM) ready samples were prepared using the in situ FIB lift out technique on a FEI Strata 400 Dual Beam FIB/SEM. The samples were capped with sputtered Ir and e-Pt/I-Pt prior to milling. The STEM lamella thickness was around 100 nm. The topographic of the surface were examined using Bruker Multimode 8 atomic force microscopy (AFM). The crystalline quality of the grown a-plane GaN was studied with Rigaku SmartLab high resolution X-ray diffraction (HR-XRD) using Cu Kα1 X-ray target source (λ = 1.5406 Å). Two types of X-ray rocking curves (XRC) omega (ω) scan were carried out including on-axis (11–20) GaN and off-axis (101), (102), (100), (200) and (300) plane.

positioning on the a-plane GaN quality has not been reported. In this paper we demonstrate the impact of AlN/GaN periodic multilayers position towards the enhancement of surface morphology and crystal quality of a-GaN grown on r-plane sapphire. In addition, the influence of periodic AlN/GaN multilayers towards the enhancement of the relaxation state is studied. Our evaluation shows that there is significant improvement in surface roughness, crystalline quality and strain relaxation state in the grown a-GaN epitaxial layer with an optimized position.

2. Experimental methods Non-polar (11–20) a-plane GaN was grown on semi-polar (1–102) rplane sapphire substrate by Taiyo Nippon Sanso SR2000 metal organic chemical vapor deposition (MOCVD) system with three horizontal laminar flow. In this study, three a-plane GaN films with different positioning of periodic AlN/GaN MLs were prepared as shown in Fig. 1(a–c). Trimethylgallium (TMGa), Trimethylaluminum (TMAl) and ammonia (NH3) were used as precursor for Ga, Al and N sources respectively, with hydrogen (H2) was adapted as gas carry along the process. Low growth pressure at 13.3 kPa was employed for the whole process. The growth of a-plane GaN started with hydrogen cleaning at 1125 °C followed by 10 min of nitridation step as a surface treatment. The GaN nucleation layer (GaN-NL) was deposited at low temperature of 700 °C with a thickness of 90 nm. 20 SLM of TMGa flow were released into reactor at high temperature of 1050 °C for 2 μm thick of aGaN growth, with the first 1 μm using high NH3 flow followed by lower NH3 flow for the next 1 μm to have a V/III ratio variation. The initial sample A of 2 μm thick a-plane GaN film was grown without applying any periodic MLs layer into the structure. The second sample B was grown using the same growth condition of a-plane GaN with the insertion of 40 pairs of periodic AlN/GaN MLs that were placed at the bottom layer before the growth of 2 μm a-plane GaN. For the third sample C, 40 pairs of periodic AlN/GaN MLs was grown in between of 1 μm a-plane GaN at the top and 1 μm a-plane GaN at the bottom. AlN/ GaN MLs used in this study possess the same growth condition with a thickness of 5 nm and 20 nm respectively, for all the samples. Overall epitaxial process of sample A, B and C are represented in . 1(d–f).

3. Results and discussions To explore the morphology of the surface, AFM measurements were carried out. Fig. 2(a–c) shows 5 × 5 μm2 AFM images for all samples. The root mean square (RMS) measured for sample A, B and C is 2.72 nm, 2.73 nm and 1.01 nm, respectively. These results imply that the position of MLs in the structure affects the surface morphology and by growing MLs in the middle of GaN layer, the surface roughness of the grown a-GaN can be reduced to almost ∼62%. Sample A and B exhibit approximately the same surface roughness, which may attribute to the two-step GaN grown in the structures. H.-C. Hsu et al. obtained high crystalline quality of a-plane GaN via two-step method. First stage was by adapting high temperature and high V/III ratio to form 3D GaN-NL followed by low temperature with lower V/III ratio in second stage [22]. Using lower V/III at the final stage growth is found to promote island coalescence which results in a smoother surface [23,24]. Via this method, we also managed to get the low surface roughness for sample A, which comparable with the literature [22]. In order to further improve the crystal quality, we then adapted MLs on top of NL in sample B. Balakrishan's group shows that by inserting AlN/GaN, the crystal quality of their m-plane was improved [25]. It can be seen that there is almost no changes in surface roughness of sample B. However, there is a big difference in crystalline quality from XRD result, which will be 2

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

Fig. 2. (a–c) 5 × 5 μm AFM image of sample A-C, (d) FESEM cross-sectional views of sample A, (e–f) STEM cross-sectional views of sample B and C, (g) FESEM plane-view image of GaN-NL, (h) close-up view of STEM cross-sectional for sample B, (i) close-up view of STEM cross-sectional for sample C with SEM plane-view image of a-plane GaN before the growth of AlN/GaN MLs in sample C in the inset.

can be observed on all a-GaN sample, regardless the type of buffer layer used for the epitaxy [6,15,26]. It is also noted that there is nanometerscale pits exist on the film surfaces with the stripe-like growth feature along c-plane. This condition corresponds to the coalesced non-polar GaN attributed to the anisotropic growth, adatom diffusion kinetics and growth rate during the process [27]. Fig. 2(d–f) illustrates the cross-sections of sample A, B and C, respectively. It can be clearly seen that for all the samples, a propagation of threading dislocation (TD) is parallel to the (11–20) growth direction. The TDs is one of the main issue in non-polar GaN growth in which started from the initial epilayers and propagate along (11–20) direction [28,29]. In order to reduce the initial propagation of TDs within the epilayer, we use a thin layer of GaN-NL layer in which have a rough and island-like surface as in Fig. 2(g). Samples with periodic AlN/GaN MLs reduce the density of TDs as shown in Fig. 2(h) and (i). It is observed that the propagation of TDs originate from the bottom layer tend to reduce in its size once it passes AlN/GaN MLs. However, the positioning of the MLs resulting in different morphological structure of the MLs. Sample B shows a wavy structure especially in the bottom layer in comparison to MLs in sample C in which possess a very abrupt layer of AlN/GaN. This corresponds to the position of MLs in sample B which on top of GaN-NL that have a rough and island-like surface as in Fig. 2(g). On the other hand, AlN/GaN MLs for sample C is able to form an abrupt structure when placed on top of 1 μm of fully coalesce GaN which has a better surface roughness as in the inset of Fig. 2(i). This helps the AlN/ GaN MLs to have a stable growth surface for a thin layer growth and enable the growth of AlN to form an abrupt structure from the first AlN/ GaN layer of the MLs. Fig. 3 shows the XRD 2θ-ω profile for sample A, B and C. It can be seen that there are two primary peaks detected for all samples, which located at ∼53° and 57.6°. These peaks correspond to the diffraction of

Fig. 3. HR-XRD 2θ-ω scan of a-plane GaN for sample A, B and C.

explained later. On the other hand, the AFM images show all samples have a stripes feature along c-direction and surface undulation along mdirection. Most of the a-plane GaN growth reported that these features 3

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

direction. Minimum of two rocking curve scans are required to obtain a proper analysis of the crystalline quality [21]. At azimuth angle, φ = 0°/[0001] plane, FHWM values are 1239 (0.34°), 959 (0.27°) and 925 (0.25°) arcsec for sample A, B and C, respectively. The values are broaden when incident beam was projected at φ = 90° where samples A, B and C exhibit 3290 (0.91°), 2039 (0.56°) and 1700 (0.47°) arcsec, respectively. The ratios of 90° to the minimum 0° are 2.66, 2.13 and 1.84 for sample A, B and C, respectively; indicating the improvement of isotropy by positioning MLs in between GaN layers. Xu et al. showed that by adding ten pairs of AlGaN/AlN MLs, the anisotropy of their device reduced to almost half by reducing the dislocation density and also managed the strain [21]. It is worth mentioning that XRC FWHM of sample C of both directions are much lower compared to sample A and B. In fact, the FWHM values between sample C and B were found to be ∼25% less along [0001] direction and ∼48% less along the [1–100] direction. The large FWHM in sample A and B suggested that these samples exhibits high density defect and thus suffer from higher strain within the crystal structure [31]. In addition to that, it is also noted that the grown GaN suffer from huge twist tilt mosaicity along φ = 90 than φ = 0, in which explain the anisotropic orthogonal structures of the non-polar GaN [32]. Different from c-plane GaN, non-polar GaN exhibit anisotropic in-plane strain states causing from low in-plane symmetry, in which explain ω-scan at single azimuth angle unable to fully evaluate crystalline quality for the growth of a-plane GaN [20,31,33]. The reduction in FWHM value for sample B and C have a correlation in the reduction on surface roughness from the AFM results. This indicate when the periodic AlN/GaN MLs is positioned in the optimize position the epitaxial growth of non-polar aplane GaN tends to give a higher crystalline quality with a better surface roughness. It is reported that non-polar a-plane GaN microstructures are dominated by Basal Stacking Faults (BSFs) [34]. BSF appear very often

Fig. 4. XRC FWHM values of on-axis (11–20) ω-scan as a function of azimuthal angle.

r-plane sapphire substrate (2–204) and GaN (11–20) plane, respectively. For the samples with AlN/GaN MLs, the MLs-related satellites are obviously observed around 55°–62° for both samples. These satellites are also called Pendellosung fringes, which implies good crystalline quality [30]. This shows the MLs inserted have a good crystalline quality and very abrupt structure to provide suitable lattice point for the subsequent a-plane GaN growth. Using Rigaku Global Fit, the spacing for the satellites peak suggested that the thickness for each AlN/ GaN periods would be 20 nm. To study the anisotropic characteristics and defects of the grown GaN, the value of full width at half maximum (FWHM) from on-axis (11–20) ω-scan is assessed base on azimuthal dependence as in Fig. 4. Few azimuthal angles were measured due to the antistrophic mosaicity of a-plane GaN. At azimuth angle of 0° the incident beam projection was parallel to c-direction while at azimuth angle of 90° parallel to m-

Fig. 5. (a) Off-axis XRC FWHM values for (101), (112) and (100) ω-scan. (b) Off-axis XRC FWHM values for (n0n0) ω scan and (c) modified WH plots of ω-scan FWHMs from the (n0n0) series of diffractions (n = 1, 2, and 3; ϕ = 0°, χ = 30°) for sample A, B and C. The dotted lines derive the trends set by off-axis ω scan along (1010) and (2020) diffractions. 4

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

during GaN growth on non-polar and semi-polar planes and predominantly formed at the GaN epilayer/substrate interface [29]. Three possible mechanisms for the formation of BSF in non-polar are: i) growth error ii) mismatch relaxation and iii) coalescence of nuclei [19,35]. Other than BSF, mosaic tilt that originates from misalignment of the crystal is also a crucial part in GaN epilayers growth. Mosaic tilt typically related with TD and domain boundaries [35]. Thus, to further understand the crystalline characteristics of the grown a-plane GaN and the effect of the defects mentioned, XRD off-axis omega scan was adapted. Fig. 5 (a) shows FWHM of XRC ω-scan with respect of three inplane directions, which are (101), (112) and (100) axis. The (101) offaxis peak measures the ‘twist’ mosaic shows that sample C exhibits the lowest FWHM among all the samples with 2068 arcsec (0.57°), while sample A and B exhibits off-axis of 2403 (0.67°) and 2392 (0.66°) arcsec, respectively. Large FWHM shown here are in agreement with high dislocation density normally observed in a-plane GaN [28]. This can be explained from the huge lattice mismatch between the substrate and grown a-plane GaN along c- and m-direction from the asymmetric structure at the interface. Typically, thin films grown on different lattice coefficient exhibits a mosaic structure, implies that the films consist of slightly mis-oriented atomic arrangement with respect to each other and the foreign substrate [8]. It is also worth mentioning that off-axis FWHM for (100) and (112) reflections also decreased when MLs was grown in the middle. This decrease signified the dislocation reduction in the structure. It is reported that the dislocation in a-plane GaN are related to type I1- basal staking fault- Frank-Shockley partial dislocations [4,36]. It is reported that (11–20) diffraction is insensitive to BSF defect, thus the broadening of the XRC is mainly due mosaic tilt or domain size, thus Williamson-Hall (WH) plot analysis was adapted in skew symmetric geometry along the (h0h0) series of diffractions [8,36]. From the modified WH graph, the off-axis XRC peak broadening is a result from two main factors in which mosaic tilt as well as the lateral size of the coherently scattering domain and the angular extent of the reciprocal lattice points can be approximated as;

Δωmeasured = Δωmosaic +

λ 2Lsin (θhkl )

Table 1 Off-axis XRC data and WH analysis for a-GaN plane. Sample

Sample A Sample B Sample C

WH plots

100

200

Δω(°)

BSF density (cm−1)

c-axis LCL (nm)

2941 2714 2623

2427 2360 2196

0.8625 0.8297 0.7966

1.65 × 105 1.37 × 105 1.23 × 105

60.75 72.97 81.40

significant reduction in BSF densities of 1.23 × 105 cm−1 for sample C. This value is ∼25% lower compared to sample A (1.65 × 105 cm−1) and sample B (1.37 × 105 cm−1). The correlation between surface roughness, crystal quality and BSF densities prove the effectiveness of periodic AlN/GaN MLs as the growth of AlN and GaN cause a different types of stress within the two materials. Since GaN has a larger lattice constant compared to AlN, GaN layer will experience compressive stress while AlN layer will experience tensile stress when grown on each other [38,39]. Kuckuk et al. claims that the compressive stress within GaN layer in AlN/GaN interface give rise to the energy barrier for Al–Ga exchange resulting to lowering the mismatch of periodic AlN/GaN MLs with the subsequent thick GaN overgrowth [39,40]. In order to study the strain in the a-plane GaN [11–20] asymmetric XRD RSM at two different azimuth angles 0° and 90° has been measured. Fig. 6(a–f) depict RSM recorded at different azimuth angles Φ = 0 and Φ = 90 along (11–20) plane. Two different of azimuth angle correlate to the sample positioning with Qx and Qy direction lying in scattering plane. All maps in Fig. 6 are presented in reciprocal lattice units, where Qx representing the horizontal axis while Qy representing the vertical axis in both (11–20) and (0001) directions respectively. Merely one prominent peak on maps of sample A is clearly observed, which corresponds to the a-plane GaN. Whilst for sample B and C, there are two discrete peaks are noted, which corresponds to the peak for GaN and AlN/GaN MLs. The distance of these two peaks depicts the position of MLs. For sample B, as MLs was grown under 2 μm GaN, the position for second peak is increases compared to that of sample C. From Fig. (6), (11)–(20) gave highest peak reflection along [-1-100] plane of −0.027 in Qx component and 4.624 in Qz component. Meanwhile (11–22) shows highest peak reflection along [11–22] plane of −0.004 in Qx component and 4.624 in Qz component. The diffraction peak of MLs is aligned with that of GaN film along the same Qx for both [-1-123] and [1 100] maps, which indicate that MLs are coherently strained to the GaN film [41]. In other words, MLs layer is almost strained or in pseudomorphic state and has high strain state in both m- and c-directions. Hexagonal GaN should have in total of six main reflection in wave vector Γ. Those include A1 (LO), A1 (TO), E1 (LO), E1 (TO), E2 (low) and E2 (high) modes. However, within x (y, y) x geometry only E2 (low), A1 (TO) and E2 (high) mode peaks are able to observe in the room temperature Raman spectra. Among those modes, E2 (high) peak was the most sensitive towards stress and have been widely used to characterize residual stress in GaN epilayer. Thus, we focused on the E2 (high) mode, which normally appeared at 567.6 cm−1 for strain/stress-free GaN [42–44]. A shift on E2 (high) phonon peak would suggest the type of stress in the structure. From Fig. 7, the frequency E2 observed are 566.74 cm−1, 569.87 cm−1 and 567.993 cm−1 for sample A, B and C, respectively while the stress calculated are −0.20 GPa, 0.54 GPa and 0.09 GPa for sample A, B and C. The stress of the epilayers was obtained according to the equation;

(1)

where Δω is the value of FWHM from (h0h0) series, λ is the wavelength of the X-ray used, θhkl is the Bragg angle for the hkl reflection while L is the value of lateral coherence length (LCL) which is measured via [5,37];

LCL=0.9/2yo

Off-axis XRD ω-scan FWHM

(2)

where yo is the y-intersection from Fig. 5 (c). It is reported that BSFs might be the main cause for off-axis XRC width anisotropy in m-plane GaN films [24]. Fig. 5 (b) and (c) illustrates the FWHM and modified WH plots from off-axis (h0h0) series ω-scan FWHMs for all samples used in this work. It could be seen from Fig. 5 (b) the off-axis XRC FWHM reduced monotonically from sample B to C, which adding another proof for the crystal quality improvement. It also observed that (3030) ω-scan exhibits the lowest FWHM than that of (1010) and (2020), which is consistent with the literatures reported that (3030) diffraction is insensitive to BSFs [5,24,36]. The data extracted from Fig. 5 (b) and (c) is tabulated in Table 1. It is calculated that yo from the linear fits for samples are in agreement with off-axis result, where yo for sample without MLs are the highest. This calculation method used here is merely the approximation method. Kim et al. showed that this calculation method is reliable in comparing the BSF density trend [19]. Caxis LCL calculated for GaN samples used in this study are 60.7, 72.9 and 81.4 for sample A, B and C, respectively. The trends for LCL values are consistent with off-axis FWHM trend discussed hitherto. The longer LCL in sample C structure can be explained by the reduction of BSF density due to the MLs position. From WH plots, the BSF density can be obtained from the reciprocal of the LCLs values [5,37]. The effect of MLs in suppressing the formation of BSF is shown in Table 1. There is

σ xx ∘ = Δω/K

(3)

where Δω is the difference in the E2 (high) phonon peaks between the stressed and unstressed GaN epilayers which the E2 (high) peak at 5

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

Fig. 6. (a–c) HR-XRD RSM scan for sample A, B and C at plane (11–20) azimuth angles φ = 0 and (d–f) HR-XRD RSM scan for sample A, B and C at plane (11–20) azimuth angle φ = 90.

the compressive stress is believe to be at the highest point. Since the periodic AlN/GaN MLs unable to form abrupt structure, this led to the introduction of different plane facets other than (11–20) towards the nominal. These different plane facets contributes to high compressive stress compared to sample with abrupt and flat surface [46,47]. On the other hand for sample C, periodic AlN/GaN MLs inserted on top of the strained thick GaN possess two-dimensional structure that is able to counter balance the tensile stress from the sapphire/GaN interface, promoting near relaxation stress towards the a-plane overgrowth GaN. It is reported that periodic AlN/GaN MLs exhibits compressive stress due to huge lattice mismatch between AlN and GaN [40]. The effect of periodic AlN/GaN MLs towards the optical properties of grown a-plane GaN was examined using room temperature photoluminescence (RT-PL) measurement. Fig. 8 shows the PL spectra for sample A, B and C, in which they are normalized towards the intensity of near-band-edge emission (IBE) at the centered of 3.4 eV. Two broad band, yellow luminescence (YL) and blue luminescence (BL) can be Fig. 7. Raman spectroscopy for sample A, B and C.

567.6 cm−1, and K (4.2 cm−1 GPa−1) is the strain coefficient [43,45]. The right shift in Raman spectra for sample with periodic AlN/GaN MLs suggested the insertion of MLs contribute to the compressive strain within the growth of a-plane GaN as in sample B and C. However the position of periodic MLs play important role in controlling the amount of strain within the a-plane GaN. Sample B have a higher compressive stress due to its position at the top of GaN-NL. This is regards to the rough and island-like surface of GaN-NL that produce non-uniform lattice point for periodic AlN/GaN MLs resulting to wavy structure of the MLs that suffer from higher amount of strain. Direct growth of nonpolar GaN on sapphire substrate will experience tensile stress due to the difference in lattice constant as in sample A. The insertion of periodic AlN/GaN MLs at the bottom prior to 2 μm GaN overgrowth layer induced a high compressive stress as in sample B. Since the MLs was inserted directly on GaN-NL, the initial pairs of the MLs experienced the three-dimensional growth similar to the GaN-NL structure. This attributes to the wavy layers as shown with the red arrow in Fig. 2(h) where

Fig. 8. RT-PL spectra for sample A, B and C. 6

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

clearly seen in all samples at the centered of 2.2 eV and 2.9 eV respectively. Since the PL measurement was performed at perpendicular degree towards the sample surface, a periodic oscillation clearly appears in BL and YL band which attributed to microcavity effect. This arise from the interface of different refractive index between air, GaN and/or AlN film and sapphire substrate, [48,49]. Sample A shows the highest relative intensity of both YL and BL band. However, the relative intensity of YL and BL in a-plane GaN reduces tremendously with the insertion of periodic AlN/GaN MLs as in sample B. Furthermore, with the optimized positioning of periodic AlN/GaN MLs, it was seen that the PL properties has improved as the relative intensity of YL and BL band demonstrated the lowest value in sample C. As previously mentioned in the literature, these specific radiative transitions could be attributed to formation of gallium and nitrogen vacancy as well as deep level impurities [50]. These results are in good agreement with the Raman spectroscopy measurement, whereby the highest strain in sample A exhibits the highest YL and BL intensity while sample with a strain relaxation state promotes the lowest intensity of YL and BL emission. In addition, several publications have concluded a correlation between YL and BL band with carbon impurities [50–53]. In this work, the methyl decomposition of TMGa and TMAl could act as unintentional carbon doping source. Growth conditions of low pressure and low V/III ratio used in the process have led to the enhanced incorporation of unintentionally carbon doping in the GaN structure [54]. Moreover, the density of defects especially N-vacancies make a significant contribution towards carbon incorporation in the GaN structure [55]. This explain the correlation of YL and BL band intensity towards the unintentional carbon doping into the sample. Furthermore, this results are well correlated with the on- and off-axis XRC measurements, in which sample A possess the highest FWHM value.

[2] J.-S. Son, Y. Honda, H. Amano, Growth of low-defect-density nonpolar a-plane GaN on r-plane sapphire using pulse NH3 interrupted etching, Opt. Express 22 (2014) 3585–3592, https://doi.org/10.1364/OE.22.003585. [3] A. Chakraborty, S. Keller, C. Meier, B.A. Haskell, S. Keller, P. Waltereit, S.P. DenBaars, S. Nakamura, J.S. Speck, U.K. Mishra, Properties of nonpolar a-plane InGaN∕GaN multiple quantum wells grown on lateral epitaxially overgrown a-plane GaN, Appl. Phys. Lett. 86 (2005) 031901, https://doi.org/10.1063/1.1851007. [4] A.Z. Omar, A. Shuhaimi Bin Abu Bakar, A.H.A. Makinudin, M.I.M.A. Khudus, A. Azman, A. Kamarundzaman, A. Supangat, Effect of low NH3flux towards high quality semi-polar (11-22) GaN on m-plane sapphire via MOCVD, Superlattice Microstruct. 117 (2018) 207–214, https://doi.org/10.1016/j.spmi.2018.03.038. [5] M.B. McLaurin, A. Hirai, E. Young, F. Wu, J.S. Speck, Basal plane stacking-fault related anisotropy in X-ray rocking curve widths of m-plane GaN, Jpn. J. Appl. Phys. 47 (2008) 5429–5431, https://doi.org/10.1143/JJAP.47.5429. [6] X. Ni, U. Özgür, Y. Fu, N. Biyikli, J. Xie, A.A. Baski, H. Morkoç, Z. Liliental-Weber, Defect reduction in (1120) a-plane GaN by two-stage epitaxial lateral overgrowth, Appl. Phys. Lett. 89 (2006) 262105, https://doi.org/10.1063/1.2423328. [7] H. Song, J. Suh, E. Kyu Kim, K. Hyeon Baik, S.M. Hwang, Growth of high quality aplane GaN epi-layer on r-plane sapphire substrates with optimization of multibuffer layer, J. Cryst. Growth 312 (2010) 3122–3126, https://doi.org/10.1016/j. jcrysgro.2010.08.004. [8] H. Yu, M. Ozturk, P. Demirel, H. Cakmak, E. Ozbay, MOCVD growth and optical properties of non-polar (11–20) a-plane GaN on (10–12) r-plane sapphire substrate, J. Cryst. Growth 312 (2010) 3438–3442, https://doi.org/10.1016/j.jcrysgro.2010. 08.052. [9] S. Yamamoto, Y. Zhao, C.P. Roy, B.C. Kenji, F. Junichi, S. Steven, P.D. Shuji, Highefficiency single-quantum-well green and yellow-green light-emitting diodes on semipolar (20\bar{2}1) GaN substrates, Appl. Phys. Express (2010) 18–21, https:// doi.org/10.1143/APEX.3.122102. [10] Hisashi Yamada, Kenji Iso, Makoto Saito, Kenji Fujito, Steven P. Denbaars, James S. Speck, Shuji Nakamura, Impact of substrate miscut on the characteristic of mplane InGaN/GaN light emitting diodes, Jpn. J. Appl. Phys. 46 (2007) 1117–1119, https://doi.org/10.1143/JJAP.46.L1117. [11] D. Jinno, S. Otsuki, S. Sugimori, H. Daicho, M. Iwaya, T. Takeuchi, S. Kamiyama, I. Akasaki, Characterization and optimization of sputtered AlN buffer layer on rplane sapphire substrate to improve the crystalline quality of nonpolar a-plane GaN, J. Cryst. Growth 480 (2017) 90–95, https://doi.org/10.1016/j.jcrysgro.2017.10. 018. [12] D. Jinno, S. Otsuki, S. Sugimori, H. Daicho, M. Iwaya, T. Takeuchi, S. Kamiyama, I. Akasaki, Characterization of nonpolar a-plane GaN epi-layers grown on highdensity patterned r-plane sapphire substrates, J. Cryst. Growth 484 (2018) 50–55, https://doi.org/10.1016/j.jcrysgro.2017.12.036. [13] Z. Sun, P. Song, S. Nitta, Y. Honda, H. Amano, A-plane GaN growth on (11-20) 4H–SiC substrate with an ultrathin interlayer, J. Cryst. Growth (2017), https://doi. org/10.1016/j.jcrysgro.2017.01.031. [14] M.D. Craven, F. Wu, A. Chakraborty, B. Imer, U.K. Mishra, S.P. DenBaars, J.S. Speck, Microstructural evolution of a -plane GaN grown on a -plane SiC by metalorganic chemical vapor deposition, Appl. Phys. Lett. (2004), https://doi.org/ 10.1063/1.1650545. [15] H.K. Dong, J.K. Su, J.C. Dong, W.Y. Ji, I.S. Jae, G.K. Tae, M.H. Sung, Improved Crystal Quality, Surface Morphology, of nonpolar a -plane GaN grown on r -plane sapphire substrates, J. Korean Phys. Soc. 58 (2011) 873, https://doi.org/ 10.3938/jkps.58.873. [16] X. Wang, W. Hou, L. Zhao, S. You, T. Detchprohm, C. Wetzel, A-Plane GaN light emitting diodes on self-assembled Ni nano-islands, 2012 Lester Eastman Conf. High Perform. Devices, LEC 2012, 2012, pp. 5–7, , https://doi.org/10.1109/lec.2012. 6410990. [17] V.N. Bessolov, E.V. Konenkova, S.A. Kukushkin, A.V. Osipov, S.N. Rodin, Semipolar gallium nitride on silicon: technology and properties, Rev. Adv. Mater. Sci. 38 (2014) 75–93. [18] M.D. Craven, S.H. Lim, F. Wu, J.S. Speck, S.P. Denbaars, Structural characterization of nonpolar (112¯0) a-plane GaN thin films grown on (11¯02) r-plane sapphire, Appl. Phys. Lett. 81 (2002) 469–471, https://doi.org/10.1063/1.1493220. [19] J.H. Kim, S. Hwang, K.H. Baik, J.H. Park, Effect of Basal-plane Stacking Faults on Xray Diffraction of Non-polar (11-20) a-plane GaN Films Grown on (1-102) r-plane Sapphire Substrates vol. 14, (2014), pp. 557–565 https://doi.org/10.5573/JSTS. 2014.14.5.557. [20] H. Wang, C. Chen, Z. Gong, J. Zhang, M. Gaevski, M. Su, J. Yang, M.A. Khan, Anisotropic structural characteristics of (112¯0) GaN templates and coalesced epitaxial lateral overgrown films deposited on (101¯2) sapphire, Appl. Phys. Lett. 84 (2004) 499–501, https://doi.org/10.1063/1.1644054. [21] S.R.Ã. Xu, Y. Hao, J.C. Zhang, X.W. Zhou, L.A. Yang, J.F. Zhang, H.T. Duan, Z.M. Li, M. Wei, S.G. Hu, Y.R. Cao, Q.W. Zhu, Z.H. Xu, W.P. Gu, Improvements in a -plane GaN crystal quality by AlN/AlGaN superlattices layers, J. Cryst. Growth 311 (2009) 3622–3625, https://doi.org/10.1016/j.jcrysgro.2009.05.028. [22] H. Hsu, Y. Su, S. Huang, Y. Wang, C. Wu, M. Chou, Direct growth of a-plane GaN on r-plane sapphire by metal organic chemical vapor deposition, Jpn. J. Appl. Phys. 49 (2010), https://doi.org/10.1143/JJAP.49.04DH05 04DH05. [23] M.A. Moram, C.F. Johnston, M.J. Kappers, C.J. Humphreys, The effects of film surface roughness on x-ray diffraction of nonpolar gallium nitride films, J. Phys. D Appl. Phys. 42 (2009) 135407, , https://doi.org/10.1088/0022-3727/42/13/ 135407. [24] M.A. Moram, C.F. Johnston, J.L. Hollander, M.J. Kappers, C.J. Humphreys, M.A. Moram, C.F. Johnston, J.L. Hollander, M.J. Kappers, C.J. Humphreys, Understanding x-ray diffraction of nonpolar gallium nitride films, J. Appl. Phys. 105 (2009) 113501, https://doi.org/10.1063/1.3129307.

4. Conclusions Non-polar (11–20) a-plane GaN was grown on semi-polar (1–102) rplane flat sapphire substrate. We show that the insertion of periodic AlN/GaN MLs effectively improves crystalline quality and anisotropic as well as reduces surface roughness and stress within the a-GaN growth. By positioning MLs in the middle, we significantly improve the crystal quality of non-polar GaN and significantly reduce the strain. Raman spectra shows that the strain induced within sample with periodic AlN/GaN MLs sandwiched by 1 μm GaN is the lowest in comparison with sample without periodic AlN/GaN MLs and MLs at other position. It is also shown that the BSF density reduced as much as 25% when MLs was grown in between GaN layers. Even though the BSF is not totally eliminated, this work shows an early demonstration of reducing the defects via epitaxial strain engineering and this study may be of considerable contribution to produce better a-plane GaN for excellent optoelectronics device in the future. Acknowledgement This research work is supported by Malaysia Ministry of Education (MOE) under Long-Term Research Grant (LRGS) project no LR0012016A (um.0000021/HME.LS), MYPhD (B) scholarship, University Malaya Postgraduate Research Grant (PG116-2015B) and UMRG Programme (RP039A-18AFR). Author would like to acknowledge CREST for GaN-on-GaN Collabration Program (PVD15-2015) and continues technical support. We appreciate UTM Industry Research Laboratory and Rigaku for technical support related to XRD measurement. References [1] S. Nakamura, Background story of the invention of efficient InGaN blue-lightemitting diodes (nobel lecture), Angew. Chem. Int. Ed. 54 (2015) 7770–7788, https://doi.org/10.1002/anie.201500591.

7

Materials Science in Semiconductor Processing 105 (2020) 104700

A. Kamarudzaman, et al.

[40] P.J. Lin, S.Y. Huang, W.K. Wang, C.L. Chen, B.C. Chung, D.S. Wuu, Controlling the stress of growing GaN on 150-mm Si (111) in an AlN/GaN strained layer superlattice, Appl. Surf. Sci. 362 (2016) 434–440, https://doi.org/10.1016/j.apsusc. 2015.11.226. [41] H.M. Huang, S.C. Ling, J.R. Chen, T.S. Ko, J.C. Li, T.C. Lu, H.C. Kuo, S.C. Wang, Growth and characterization of a -plane Al x Ga 1 À x N alloys by metalorganic chemical vapor deposition, J. Cryst. Growth 312 (2010) 869–873, https://doi.org/ 10.1016/j.jcrysgro.2009.12.064. [42] H. Harima, Properties of GaN and related compounds studied by means of Raman scattering, J. Phys. Condens. Matter 14 (2002) R967–R993, https://doi.org/10. 1088/0953-8984/14/38/201. [43] T.C. Shibin Krishna, N. Aggarwal, G.A. Reddy, P. Dugar, M. Mishra, L. Goswami, N. Dilawar, M. Kumar, K.K. Maurya, G. Gupta, Probing the correlation between structure, carrier dynamics and defect states of epitaxial GaN film on (1120) sapphire grown by rf-molecular beam epitaxy, RSC Adv. 5 (2015) 73261–73267, https://doi.org/10.1039/c5ra10099b. [44] T. Jiang, S. Xu, J. Zhang, Y. Xie, Y. Hao, Spatially resolved and orientation dependent Raman mapping of epitaxial lateral overgrowth nonpolar a-plane GaN on rplane, Sci. Rep. 6 (2016) 19955, https://doi.org/10.1038/srep19955. [45] J.N. Dai, Z.H. Wu, C.H. Yu, Q. Zhang, Y.Q. Sun, Y.K. Xiong, X.Y. Han, L.Z. Tong, Q.H. He, F.A. Ponce, C.Q. Chen, Comparative study on MOCVD growth of a -plane GaN films on r -plane sapphire substrates using GaN, AlGaN, and AlN buffer layers, J. Electron. Mater. 38 (2009) 1938–1943, https://doi.org/10.1007/s11664-0090847-7. [46] M. Mishra, A. Gundimeda, S. Krishna, N. Aggarwal, B. Gahtori, N. Dilawar, V.V. Aggarwal, M. Singh, R. Rakshit, G. Gupta, Wet chemical etching induced stress relaxed nanostructures on polar & non-polar epitaxial GaN films, Phys. Chem. Chem. Phys. 19 (2017) 8787–8801, https://doi.org/10.1039/c7cp00380c. [47] T. Wei, Q. Hu, R. Duan, J. Wang, Y. Zeng, J. Li, Y. Yang, Y. Liu, Mechanical deformation behavior of nonpolar GaN thick films by berkovich nanoindentation, Nanoscale 4 (2009) 753–757, https://doi.org/10.1007/s11671-009-9310-1. [48] A. Billeb, W. Grieshaber, D. Stocker, E.F. Schubert, R.F. Karlicek, Microcavity effects in GaN epitaxial films and in Ag/GaN/sapphire structures Microcavity effects in GaN epitaxial films and in Ag/GaN/sapphire structures, Appl. Phys. Lett. 70 (1997) 2790–2792, https://doi.org/10.1063/1.119060. [49] S.J. Xu, G. Li, S.J. Chua, W. Wang, Observation of optically-active metastable defects in undoped GaN epilayers observation of optically-active metastable defects in undoped GaN epilayers, Appl. Phys. Lett. 72 (1998) 2451–2453, https://doi.org/10. 1063/1.121379. [50] G. Santana, O. de Melo, J. Aguilar-Hernández, R. Mendoza-Pérez, B. Marel Monroy, A. Escamilla-Esquivel, M. López-Sópez, F. de Moure, L.A. Hernández, G. ContrerasPuente, Photoluminescence study of gallium nitride thin films obtained by infrared close space vapor transport, Materials 6 (2013) 1050–1060, https://doi.org/10. 3390/ma6031050. [51] M.A. Reshchikov, H. Morkoç, Luminescence properties of defects in GaN, J. Appl. Phys. 97 (2005) 5–19, https://doi.org/10.1063/1.1868059. [52] L.T. Tung, K.L. Lin, E.Y. Chang, W.C. Huang, Y.L. Hsiao, C.H. Chiang, Photoluminescence and Raman studies of GaN films grown by MOCVD, J. Phys. Conf. Ser. 187 (2009) 012021, , https://doi.org/10.1088/1742-6596/187/1/ 012021. [53] F. Liang, D. Zhao, D. Jiang, Z. Liu, J. Zhu, P. Chen, Carbon-related defects as a source for the enhancement of yellow luminescence of unintentionally doped GaN, Nanomaterials 8 (2018), https://doi.org/10.3390/nano8090744. [54] T.F. Kuech, J.M. Redwing, Carbon doping in metalorganic vapor phase epitaxy, J. Cryst. Growth 145 (1994) 382–389. [55] M. Qinghua, L. Junlin, W. Xiaoming, Z. Jianli, X. Chuanbing, M. Chulan, Z. Meng, J. Fengyi, Influence of growth rate on the carbon contamination and luminescence of GaN grown on silicon, J. Semicond. 36 (2015) 093003, , https://doi.org/10. 1088/1674-4926/36/9/093003.

[25] K. Balakrishnan, M. Lachab, H.C. Chen, D. Blom, V. Adivarahan, I. Ahmad, Q. Fareed, M.A. Khan, MOCVD growth of semipolar AlxGa1-xN on m-plane sapphire for applications in deep-ultraviolet light emitters, Phys. Status Solidi 208 (2011) 2724–2729, https://doi.org/10.1002/pssa.201127078. [26] X. Shengrui, H. Yue, D. Huantao, Z. Jincheng, Z. Jinfeng, Z. Xiaowei, L. Zhiming, N. Jinyu, Surface morphology of [112¯0] a -plane GaN growth by MOCVD on [11¯02] r -plane sapphire, J. Semicond. 30 (2009) 043003, , https://doi.org/10.1088/16744926/30/4/043003. [27] J. Zhang, W. Tian, F. Wu, Q. Wan, Z. Wang, J. Zhang, Y. Li, J. Dai, Y. Fang, Z. Wu, C. Chen, J. Xu, X. Li, The effects of substrate nitridation on the growth of nonpolar a-plane GaN on r-plane sapphire by metalorganic chemical vapor deposition, Appl. Surf. Sci. 307 (2014) 525–532, https://doi.org/10.1016/j.apsusc.2014.04.069. [28] A. Chakraborty, K.C. Kim, F. Wu, J.S. Speck, S.P. Denbaars, U.K. Mishra, A. Chakraborty, K.C. Kim, F. Wu, J.S. Speck, S.P. Denbaars, U.K. Mishra, Defect reduction in nonpolar a -plane GaN films using in situ Si N x nanomask Defect reduction in nonpolar a -plane GaN films using in situ SiN x nanomask, Appl. Phys. Lett. 89 (2006) 041903, https://doi.org/10.1063/1.2234841. [29] J.H. Kim, S. Hwang, Y.G. Seo, K.H. Baik, Structural and electrical anisotropies of Sidoped a -plane (11 – 20) GaN films with different SiN x interlayers, Semicond. Sci. Technol. 28 (2013) 085007, , https://doi.org/10.1088/0268-1242/28/8/085007. [30] Y. Jianchang, W. Junxi, L. Naixin, L. Zhe, R. Jun, L. Jinmin, High quality AlGaN grown on a high temperature AIN template by MOCVD, J. Semicond. 30 (2009) 103001, , https://doi.org/10.1088/1674-4926/30/10/103001. [31] C. Roder, S. Einfeldt, S. Figge, T. Paskova, D. Hommel, C. Roder, S. Einfeldt, S. Figge, T. Paskova, D. Hommel, Stress and wafer bending of a-plane GaN layers on r-plane sapphire substrates Stress and wafer bending of a -plane GaN layers on r -plane sapphire substrates, J. Appl. Phys. 100 (2006) 103511, https://doi.org/10. 1063/1.2386940. [32] M.A. Moram, M.E. Vickers, X-ray diffraction of III-nitrides, Rep. Prog. Phys. 72 (2009) 036502, , https://doi.org/10.1088/0034-4885/72/3/036502. [33] T. Paskova, R. Kroeger, S. Figge, D. Hommel, V. Darakchieva, B. Monemar, E. Preble, A. Hanser, N.M. Williams, M. Tutor, High-quality bulk a-plane GaN sliced from boules in comparison to heteroepitaxially grown thick films on r-plane sapphire, Appl. Phys. Lett. 89 (2006) 051914, https://doi.org/10.1063/1.2236901. [34] M. Moram, C.F. Johnston, M.J. Kappers, C.J. Humphreys, Measuring dislocation densities in nonpolar a -plane GaN films using atomic force microscopy, J. Phys. D Appl. Phys. 43 (2010) 055303, , https://doi.org/10.1088/0022-3727/43/5/ 055303. [35] Y.S. Lee, T.H. Seo, A.H. Park, K.J. Lee, S.J. Chung, E. Suh, Influence of controlled growth rate on tilt mosaic microstructures of nonpolar a -plane GaN epilayers grown on r -plane sapphire, Electron. Mater. Lett. 8 (2012) 335–339, https://doi. org/10.1007/s13391-012-2060-8. [36] Q. Sun, T.-S. Ko, C.D. Yerino, Y. Zhang, I.-H. Lee, J. Han, T.-C. Lu, H.-C. Kuo, S.C. Wang, Effect of controlled growth dynamics on the microstructure of nonpolar a -plane GaN revealed by X-ray diffraction, Jpn. J. Appl. Phys. 48 (2009) 071002, , https://doi.org/10.1143/JJAP.48.071002. [37] T. Metzger, R. Höpler, E. Born, O. Ambacher, M. Stutzmann, R. Stömmer, M. Schuster, H. Göbel, S. Christiansen, M. Albrecht, H.P. Strunk, Defect structure of epitaxial GaN films determined by transmission electron microscopy and triple-axis X-ray diffractometry, Philos. Mag. A 77 (1998) 1013–1025, https://doi.org/10. 1080/01418619808221225. [38] J. Su, E.A. Armour, B. Krishnan, S.M. Lee, G.D. Papasouliotis, Stress engineering with AlN/GaN superlattices for epitaxial GaN on 200 mm silicon substrates using a single wafer rotating disk MOCVD reactor, J. Mater. Res. 30 (2015) 2846–2858, https://doi.org/10.1557/jmr.2015.194. [39] A.V. Kuchuk, V.P. Kladko, T.L. Petrenko, V.P. Bryksa, A.E. Belyaev, Y.I. Mazur, M.E. Ware, E.A. Decuir, G.J. Salamo, Mechanism of strain-influenced quantum well thickness reduction in GaN/AlN short-period superlattices, Nanotechnology 25 (2014) 245602, , https://doi.org/10.1088/0957-4484/25/24/245602.

8