The mosaic structure parameters of the AlN epilayer, dislocation densities (edge and screw dislocations) and the strain behavior in the AlN epilayers grown on 6H-SiC substrate with LT-AlN(NL) were investigated using HR-XRD measurements. We observed that the lateral and vertical coherence lengths tilt and twist angle, and heterogeneous strain values of the mosaic blocks in the AlN epilayers were slightly affected by the LT-AlN nucleation layers growth times. However, the dislocation densities in the AlN layer are affected by the growth times of the LT-AlN(NL) layer. The highest densities for both screw (6.8×1010 cm-2) and edge (8.0×109 cm-2) type dislocation were measured for sample with 240 sec. growth times of LT-AlN(NL). On the other hand, the in-plane and out-of-plane strain in the AlN epilayer depends on the LT-AlN(NL) growth times showed that the in-plane strain values decrease with the LT-AlN(NL) growth times but the out-of-plane strain values increase. Furthermore, we detected that the biaxial stress in the AlN epilayer changes from compressive to tensile form when the LT-AlN NL growth times is greater than 120 sec.
The influence of the LT-AlN(NL) growth times on the mosaic structure parameters of the AlN layer grown on the LT-AlN(NL)/6H-SiC structures and also the dislocation densities and the strain behaviors in the AlN epilayers has been investigated using XRD measurements. The growth times of the LT-AlN(NL) were changed to 0, 60, 120, 180, and 240 sec. The results, we observed that the mosaic structure parameters of the AlN epilayers were slightly affected by the LT-AlN(NL) growth times. However, the dislocation densities in the AlN layer are affected by the growth times of the LT-AlN(NL) layer. The highest values densities were measured for both screw (6.8×1010 cm-2) and edge (8×109 cm-2) type dislocation density in the sample in which 240 sec grown LT-AlN(NL) were used. Strain calculation results show that the samples without LT-AlN(NL) suffered maximum compressive in-plane strain (-10.9×10-3), which can be suppressed by increasing the LT-AlN(NL) growth times. On the other hand, out-of-plane strain has a compressive character and its values increase with LT-AlN(NL) growth times between 120 and 240 sec. Furthermore, the form of the biaxial stress in the AlN epilayer changed from compressive to tensile when the LT-AlN(NL) growth times were greater than 120 sec. The aluminum nitride (AlN) is an important member of the III-nitride semiconductor materials. Because of its unique properties, such as its very wide and direct band gap of 6.2 eV and high thermal conductivity of 285 W/mK [1], they have chosen for high-power, high-frequency electronic device applications and also for optical devices in the ultraviolet region [1-4]. The application of AlN layers to highly efficient ultraviolet solid state light sources and mobile phone Radio Frequency filters are typical future applications [2-4]. It is well known that the crystalline quality and residual stress in layered structures is a key issue that greatly influences the performance of the optoelectronics devices [5-9]. The AlN layers have especially been grown on foreign substrate such as sapphire [10, 17], SiC [18-20], or Si [14, 15]. There is a large lattice constant mismatch and difference in the thermal expansion coefficient between the AlN layers and the foreign substrates. In the optoelectronic devices application, it is very important to grow defect free high quality AlN film on this type of foreign substrate for optoelectronic applications. For this reason, some buffer design and growth techniques have been developed by researcher. Most researchers have used a low temperature growth AlN (LT-AlN) layer as a nucleation layer between the AlN layer and substrates to improve the crystal quality and reduce the residual stress of the III-nitride layers on the sapphire, SiC, and Si substrate [11-13, 19]. Because of the low lattice mismatch between AlN and SiC (approx. 0.9%), very close thermal expansion coefficients, and high thermal conductivity of SiC [10], compared to the other substrates, the properties of the SiC have been the most suitable among these substrates for the heteroepitaxial growth of AlN films [18-20]. Many more studies were published on the effects of the LT-AlN nucleation layers on the crystal quality of the AlN layers grown on SiC substrate. Previous studies indicate that the growth condition of the AlN nucleation layer (NL), such as growth temperature, recrystallization temperature, recrystallization time, growth pressure, and molar V/III ratios have a strong influence on the crystal quality and the stress form in the AlN layer and grown III-nitride materials on that AlN layer [11, 13, 17-20, 24]. However, there is no systematic investigation about the influence of LT-AlN(NL) growth times on the mosaic structures parameters and strain behavior of the AlN epilayer grown on 6H-SiC substrate.
Therefore, our intention is to investigate the influence of the LT-AlN(NL) growth times on the mosaic structure parameters of the AlN epilayers and strains in the AlN epilayers grown on the 6H-SiC substrate. The influence of the LT-AlN(NL) growth times on the mosaic structures parameters, such as vertical and lateral coherence lengths (average size of the mosaic blocks) tilt and twist angle, heterogeneous strain, and dislocation densities (edge and screw dislocations) and also the LT-AlN (NL) growth times effects on the strains in the AlN epilayers were calculated using high resolution X-ray diffraction (HR-XRD) measurements. Moreover, the surface morphology were imagined of the AlN epitaxial layers was examined using Veeco CP atomic force microscopy (AFM) imaging study. In the growth process, double polished 2-inch-diameter 6H-SiC(0001) wafers were used as substrates material. All of the samples were grown in a low-pressure metalorganic chemical vapor deposition (MOCVD) reactor (Aixtron 200/4 HT-S) using the source gases of trimethylaluminum (TMAl) and ammonia (NH3), and the carrier gas of hydrogen (H2) and nitrogen (N2). The surface oxides in the 6H-SiC substrates were removed using a solution of H2SO4/H2O2(4:1). Samples were held in the H2SO4/H2O2(4:1) solution approx.. 30 s and were then rinsed in DI water again for a prolonged period. After cleaning process, the substrate was loaded into the reactor and the surface of the substrates was baked at 1175°C in H2 ambient for 15 min to remove the oxide layer. The baking process was continued with the growth of five samples. The samples were named as sample A, B, C, D, and E after that. The sample A contains a 150 nm AlN epilayer grown on 6H-SiC substrate without LT-AlN NL. Other four samples were grown with a common structure of LT-AlN NL and 150-nm-thick AlN epilayers. In the growth process, the LT-AlN NL deposition times for the sample A, sample B, sample C, sample D, and sample E were changed as 0 (without NL), 60, 120, 180, 240 sec., respectively. And the growth parameters of growth temperature (650°C), recrystallization temperature (1130°C), growth pressure (50 mbar), recrystallization time (2 min) and molar V/III ratios (2500) of the LT-AlN NL were kept identical for all samples. In addition, the same growth parameters for the AlN layers were used for all the samples and taken as 1130°C, 25 mbar, 150 nm and 640 for the growth temperature, growth pressure, layer thickness and the molar V/III ratios, respectively. The identical growth parameters for the LT-AlN and AlN layers give a constant growth rate. For this reason, the LT-AlN(NL) layer thickness in the all samples can be taken as same.
The high resolution X-ray diffraction (HR-XRD) measurements were done using a Rigaku Smart Lab. high-resolution diffractometer system; delivering CuKα1 (1.544Ã…) radiation and the samples’ surface morphology imaging study were conducted using commercial VEECO CPII Atomic Force Microscopy (AFM) in contact mode. In order to assess the surface quality, AFM imaging was done over a 4.55×4.55 µm2 scan size. The AFM imaging of the AlN epilayers grown on LT-AlN (NL)/6H-SiC structures is shown in Fig. 1 for all of the samples. The root-mean-square (RMS) roughness values are tabulated in Table 1. The values for the samples are changed between 0.76 nm (for sample C) and 1.88 nm (for sample D).
The crystal phase of the AlN/LT-AlN(NL)/6H-SiC structures were investigated by ω-2θ scans of the X-ray diffraction for all samples. The results are given in Fig.2. The (0002) plane reflection peaks of the wurtzite AlN epilayers and (0006) plane reflection peaks of the 6H-SiC are clearly observed for all samples. There is a shift in the (0002) plane reflection peaks of the wurtzite AlN epilayers (Fig. 2). These shifts can be attributed to the strains in the AlN epilayer. On the other hand, (0006) plane reflections peak of the relaxed 6H-SiC(0001) were observed at the nearly same 2θ angle values of 35.557º.
Calculation of Strain and Stress Values in the AlN epilayers
The hexagonal III-nitride based materials grown on foreign substrates have a mosaic structure [31-35]. The crystallographic c-axis of III-nitride based epilayer mosaic columns and the c-direction of the substrate coincided with a small angle, thus, the lattice constants of a and b of the hexagonal epilayer are oriented parallel to the z- plane of the substrate [27-30]. In our case, therefore, the AlN epilayer exhibits in-plane isotropic elastic properties, and its in-plane lattice deformation states can be defined with one strain parameter. The in-plane ( ) and out-of-plane ( ) strain components in the AlN epilayers can be calculated using the crystal lattice constants a and c, respectively [27-30, 36, 37].
The crystal lattice constants (a and c) of the hexagonal crystal structures can be precisely calculated using symmetric and skew symmetric ω-2θ scans of the XRD ( ) with combined Bragg’s law ( ) [27-30]. There are two unknowns in these Equations. For this reason, we need at least two different plane reflections ( ) measurements in the calculation study of a and c parameters [27-30].
Generally, lattice constants c of the layer perpendicular to the interface are calculated from the one or two high-angle symmetric (000l) plane reflections, such as (0004), (0006), and (0008) plane [27-30]. On the other hand, the lattice constant a is parallel to the layer interface and can be derived using the one or more diffraction peaks of the high-angle asymmetrical plane reflections, such as (10-14), (11-24), (10-15), and (20-24) [27-30].
In our study, the c and a values were calculated from the (0004), (0006) symmetric plane reflections, and (10-12), (10-13), (10-14), and (20-21) asymmetric plane reflections measurements, respectively. The calculated lattice parameters (a and c) for all the samples were given in Table 1.
The and components in the AlN epilayers were calculated using the and values for the strain free AlN [26] and shown in Table 1.Moreover, Fig. 3(a) shows the in-plane and out-of-plane strain components in the AlN epilayers grown on LT-AlN(NL)/6H-SiC structures as a function of LT-AlN(NL) growth times. The results show that the samples without LT-AlN(NL) suffered maximum compressive in-plane strain (-10.9×10-3), which can be suppressed by increasing the LT-AlN(NL) growth times. The minimum strain values were obtained when the LT-AlN(NL) growth times reached 180 sec. in sample C (-8.2×10-3). On the other hand, we observed a different behavior for the out-of-plane strain in the AlN epilayers. Out-of-plane strain has a compressive character and its values increase with LT-AlN(NL) growth times between 120 and 240 sec. Minimum out-of-strain values (-1.7×10-3) have been measured in the sample B of 60 sec. LT-AlN(NL) growth times.
In cases when the III-nitride materials that are grown on the foreign substrate, such as Al2O3, SiC, and Si by the epitaxial growth techniques, contains a high density of point defects that cause a considerable contraction or expansion in the crystal lattice constant of the layer (depending on the type and concentration level of point defects). Because of this reality, the and strain components in the epilayer are determined by the superposition of biaxial ( in the c- direction and in the a-direction) and hydrostatic strains ( ) in the epilayers [27-30, 31, 32]. The following equations were used in the calculation of the biaxial and hydrostatic strains components in the epilayers [29, 30]; The is the Poisson ratio and can be calculated using the elastic constants of the layer ( and ) in Eq. (2b). The elastic constants and values for the AlN epilayer that was obtained by Brillouin scattering measurements, were used in Eq. (2b) as c13=120 GPa and c33=395GPa [39] and values of 0.210 calculated for ν parameters. After substitutions of data for the Poisson’s ratio and the measured strains and into Eq. (1a), (1b) and (2a), the , and were calculated for the AlN epilayers grown on the LT-AlN(NL)/6H-SiC structures. The calculated results are tabulated in Table 2 and also shown in Fig.3 (b) and Fig. 4 as a function of LT-AlN(NL) growth times. The biaxial strains in the a-direction for the AlN epilayers grown on LT-AlN(NL) with 0, 60, and 120 sec. growth times (sample A, B and C) are compressive but in samples D and E, the biaxial strain in the a-direction shows a tensile form. However, biaxial strains in the c-direction show a completely opposite behavior. We obtained positive biaxial strains in the c-direction for sample A, B, and C, and negative biaxial strains in the c-direction for samples D and E. As can be seen from Fig. 4, the behavior shows compressive character for all samples. The obtained values changes between -4.5×10-3 (sample B) to -9.2×10-3 (sample E). The biggest values obtained for sample E.
The stress in the AlN epilayers grown on LT-AlN(NL)/6H-SiC structures originating from the mismatch between the lattice constant of the epilayers and the substrate are biaxial [29, 30]. The in-plane biaxial stress ( ) in the AlN epilayer can be calculated using the relation given below [29, 30];In Eq. (3), the biaxial elastic modulus of the materials which have a hexagonal crystal lattice structures strained in the [0001] crystallographic direction can be calculated using the Eq. of . The biaxial stress component in the c-direction equals to zero but the other components in the crystallographic b- and a-direction are equals to each other. The biaxial elastic modulus value have been calculated using the 345, 125, 120, and 395 GPa values for the c11,c12, c13, c33 parameters, respectively. The calculated biaxial elastic modulus values found as 478.5 GPa [39]. Furthermore, the Eq. (3) have been used in the calculation of by substituting the values of biaxial strain in the a-direction and the biaxial elastic modulus value. The obtained results for the are given in Table 1and shown in Fig.4 as a function of LT-AlN(NL) growth times. The biaxial stress in the AlN epilayer decreases with LT-AlN(NL) growth times. For sample E, minimum biaxial stress values of 0.12 GPa was measured.
Calculation of Mosaic Structure Parameters of the AlN Epilayers
When the AlN and/or other III-nitride materials are grown on foreign substrate (such as Al2O3, SiC or Si) are usually highly defective and faulted, due to the large mismatch of lattice constants and large difference in thermal expansion coefficients between III-nitride based epilayers and substrates. These imperfect layers manifest a mosaic structures consisting of many small hexagonal grains and the mosaic structures of the layers can be characterized by means of mean tilt ( ), mean twist ( ) angles, and the average size of the mosaic blocks with lateral ( ) and vertical ( ) coherence length [27, 28, 31-35]. The of the mosaic blocks are defined as the rotation of the mosaic blocks out of the blocks perpendicular to the surface normal, and the as the in-plane rotation around the surface normal [27, 28, 35-39]. The mosaic structure model of the crystals has been applied several times to III-nitride based material films [27, 28, 31-35]. The degree of mosaicity expressed by lateral, vertical coherence length, heterogeneous strain ( ), tilt and twist angle are important parameters in characterizing the quality of the epitaxial films [31-35].
The average absolute values of the and are directly related to the different dependences of broadening caused by limited grain size and tilt or strain on the reflection order [35, 36]. The mosaic structure parameters of , , and along the c-axis are usually determined by Williamson–Hall (W-H) plot obtained from XRD measurements and the can be determined from direct measurements [31, 32]. Specifically in triple-axis diffractometer measurements, the broadening of the rocking curve (angular-scan or -scan) of the symmetric (0002), (0004), and (0006) plane reflections for the epilayers are influenced only by the and short coherence length parallel to the substrate surface [28, 31, 32].
The and parameters can be calculated using the W-H plot of the versus functions. The slope of the linear dependence of the vs. gives the and can be calculated from the inverse of the y-intersection of the fitted line with the ordinate. In the function expression, (FWHM)ω is in the angular unit, is the Bragg reflection angle, and is the x-ray wavelength. On the other hand, the radial-scan ( scan) directions of the symmetric reflections are affected by the small and along the c-axis and cause a broadening in the reflections. The W-H plot of the vs. for ω–2θ scans of the (0002), (0004), and (0006) plane reflections should yield a straight line. The slope of the line is equal to the and, also, can be calculated from the y-intersection y0 [28, 31, 32].
The W-H plots of the triple-axis (a) ω-scan and (b) ω-2θ-scan of (0002), (0004), and (0006) symmetric plane reflections for the AlN epilayers grown on LT-AlN (NL)/6H-SiC structures are shown in Fig. 5a and 5b. The expected linear behaviors of the graphs are experimentally well confirmed for the function of vs. , which gives the rather accurate mean tilt angle values. The larger values were found for sample E as12.3×10-3±2×10-4 degree. On the other hand, minimum values were measured for sample D (5.8×10-3±2×10-4 degree). There is no systematic relationship between the mean tilt angle and the LT-AlN(NL) growth times. The calculated values for the AlN epilayers are shown in Table 2. The values of the samples change between 39.2 and 710 nm. The maximum values were observed for sample B and the minimum values for sample A.
However, the W-H plots of vs. functions were showed in the Fig. 5b. The linear behaviors with negative slope values were observed for all samples. The values of the samples changed between -0.6×10-4±1×10-5 (sample C) and -26.4×10-4±1×10-5 (sample E). The much higher values for in the AlN epilayer for sample B and sample E were measured as -22.0×10-4±1×10-5 and -26.4×10-4±1×10-5. This measured negative slope of the plots indicates the compressive strain experienced in smaller grain size in the AlN epilayers for all the samples [40]. From Table 2, the highest value (30.3 nm) for AlN epilayers was measured on layers with the LT-AlN(NL) growth times of 120 sec (sample C). On the other hand, the obtained lowest value is 4.8 nm for sample E with the highest LT-AlN(NL) growth times (240 sec.).
The mosaic structures parameter of mean tilt, twist angle, the average size of the sub-grains, and the inhomogeneous strain causes some broadening in the FWHMs of the rocking curve of an imperfect films. The mean of the mosaic blocks can be obtained using FWHMs of ω-scan or Φ-scan of the XRD measurements [24, 30, 31-35].
The values were determined using direct measurements or extrapolation methods in which methods a geometrical model were used and that considers the simultaneous presence of tilt and twist in the structures [31, 34, 35, 41-44]. The value was determined using some complicated calculations and fitting methods in which the functions are fitted to the data obtained from the measurement of ω-scans in skew geometry from reflections with increasing lattice plane inclination. On the other hand, some authors proposed a simple empirical approach to obtain the value directly without falling into a complicated computation and/or fitting procedure [34, 35, 41-43].
In order to completely eliminate broadening due to the domain size and inhomogeneous strain effects a slit of 0.6 mm in front of the detector was used in double-axis -scans. In the direct measurement methods, the intrinsic width of reflection for the crystal and the apparatus broadening for all of the experimental reflections can be neglected because of a small amount of effects (only a few arcsec). Furthermore, the triple-axis scans of the (0002) and (hk(-h-k)l) plane reflections, with either an h or k non-zero orientation, of the AlN epilayers exhibit a small broadening. For this reason, only measurements of the broadening that was caused by the twist were analyzed using (hk(-h-k)l) reflections in skew geometry gives [31, 34, 35].
The rocking curves measurements were done for the ω-scans and Φ-scans of the (10-15), (10-14), (10-13), (10-12), (20-23), (11-22), (20-21), and (12-31) plane reflections with increasing χ angle and FWHMs of the scans were calculated using a fit of Pseudo-Voigt function to the rocking curves. The can be extrapolated from a fit to the measured FWHMs of the ω-scans and Φ-scans data for different (hkl) plane reflections in a skew symmetric diffraction.
The calculated values of the mean in the AlN epilayers grown on LT-AlN(NL)/6H-SiC structures are tabulated in Table 2. The founded values changes between 0.785º (sample D) and 1.194º (sample C). We found different mean values for each of the samples. Based on the observation from Table 2, it can be argued that there is no systematic behavior between the mosaic parameter of in the AlN epilayers grown on LT-AlN(NL)/6H-SiC structure and growth times of the LT-AlN(NL) layer in our case.
Calculation of Dislocation Density in the AlN Epilayers
The mismatch between the lattice constants of the epilayers and the substrates causes the creation of dislocations (edge, screw, and mixed type) in the epilayers. In our case, there is a high lattice mismatch between the AlN epilayer and 6H-SiC substrate that exhibits high dislocation densities [8, 14, 15, 28, 45, 46].
Dislocations in the epilayers can be mainly classified as the pure screw dislocation, pure edge dislocation, and the mixed dislocations [8, 14, 15, 28, 45, 46]. The edge (Dedge) and screw (Dscrew) type dislocation densities in the epilayers layers can be calculated using the relationships:
In the Eq. (4) and (5), the parameter of the Ω is the FWHM of the symmetric (0002) plane reflections peak and (12-31) asymmetric plane reflections peak measured by XRD rocking curves, and b is the Burgers vector length for the AlN epilayer (bscrew=0.49807 nm, bedge=0.31113nm) [26].The total dislocation densities (Ddis) in the epilayer is equal to the summation of the Dedge and screw Dscrew type dislocation densities.
The edge and screw dislocation densities in the AlN epilayers grown on LT-AlN (NL)/6H-SiC structures are shown as a function of LT-AlN NL growth times in Figure 6. The Dedge values in the AlN epilayers change between 1.8×108 cm-2 (sample C) and 5.8×109 cm-2 (sample E). On the other hand, two order higher values for the Dscrew obtained and the changes between 1.4×1010 cm-2 (sample D) and 6.8×1010 cm-2 (sample E). The highest dislocation density for both screw and edge types were measured for sample E. As can be seen in Figure 6, 240 sec. of growth time for the LT-AlN (NL) causes many more dislocation of the edge and screw types in the AlN epilayers.
...(download the rest of the essay above)