Abstract

We investigate the Raman scattering of the AlN films prepared by pulse laser deposition. The Raman spectrum and the X-ray diffraction (XRD) patterns of the AlN films were compared to find out the influence of the crystal texture on the Raman scattering. The (high) and (TO) scattering modes were observed in Raman spectra. The results show that the orientation and the crystal quality of the AlN films have a great impact on these Raman scattering modes. The deterioration of (002) orientation and the appearance of other orientations in the XRD patterns lead to the weakening of the (high) mode and strengthening of the (TO) mode in the Raman spectrum. In addition, the (high) peak is broadened with the increasing of the width of the X-ray rocking curve. The broadening of the Raman peaks can be associated with degeneration in crystal quality. Furthermore, by combining the energy shift of (high) mode with the measured residual stress in the films, the Raman-stress factor of the AlN films prepared by pulse laser deposition is −4.45 cm−1/GPa for the (high) mode.

1. Introduction

With the tremendous progress of III-nitrides research in terms of both fundamental understanding as well as devices, applications, Aluminum nitride (AlN) attracts increasing interest due to the band gap of approximately 6 eV and some excellent properties including high thermal conductivity, good stability at high temperatures, and high acoustic velocity. These excellent properties make AlN thin films useful in many applications, such as hard coatings [1], electroacoustic devices [2], buffer layers for epitaxial growth [3], buried dielectric layers in future silicon-on-insulator devices [4] and Micro/Nano-Electro-Mechanical Systems (MEMS/NEMS) [5, 6]. Recently AlN-based deep ultraviolet light-emitting diodes (LEDs) [7, 8] have been reported, which demonstrates the practicability of AlN as an active deep ultraviolet material for future optoelectronic device applications. Many methods have been used to grow AlN film including molecular beam epitaxy (MBE) [911], metal-organic chemical vapor deposition (MOCVD) [12, 13], pulse laser deposition (PLD) [1416], and reactive sputtering [17, 18]. In particular, PLD is suited to grow AlN films for many applications because it is rapid and it is capable of producing high-quality, stoichiometric films at low deposition temperature [16].

In the studies of thin films, X-ray diffraction (XRD) is a strict and widely used method to measure the crystal texture. However, a good film texture is not a guarantee for high-quality devices in many applications. For example, the residual stress in the film is also important because it may alter the energy band structure and influence the device performance, particularly in LEDs and MEMS/NEMS. On the other hand, Raman scattering is widely used to detect the phonon vibration properties in materials and has been proven to be a practical method for studying stress of III-V semiconductor. At present, some reports have focused on the Raman scatterings of the AlN films grown by MBE [9, 19] or MOCVD [12]. However, there are few reports concerning the scattering characteristics of the AlN films prepared by PLD [20].

In this work, we prepared AlN films using PLD and measured their X-ray patterns and Raman spectra. These results were compared to find out the influence of the texture on the phonon mode in the AlN films. Furthermore, we study the correlation between the shift of (high) mode and the residual stress in the films.

2. Experimental Details

The AlN films were prepared using a PLD system (LMBE-II, SKY Techno. Co., Ltd.). The target was a circular, stoichiometric AlN slab with a purity of 99.99% fabricated by hot pressing (Omat Sputtering Target Co.). Before the growth of AlN films, the chamber was evacuated until the base pressure decreases to less than 2 × 10−5 Pa. The target was placed 6 cm from the substrate. A KrF excimer laser ( nm,  Hz, and  ns) with an energy density of 1–5 J/cm2 was used to ablate the target. The laser beam was brought into the chamber through a UV fused-silica window at a 60° angle relative to the target surface normal. During the film preparation, N2 (99.9999% purity) was introduced into the chamber and maintained at a pressure of 1 Pa. The polished 1-inch-diameter (001) sapphire was used as the substrate. The substrates were heated to 500°C and kept at this temperature during the preparation process. The thicknesses of all films were controlled to about 1 μm.

The crystal structure and the orientation of the films were identified by XRD (BRUKER-AXS) at wavelength of 0.15418 nm (Cu Kα). The morphologies of the films were observed by field effect scanning electron microscope (SEM, FEI SIRION 200) under the operating voltage of 10 kV. The Raman measurement was performed at room temperature in the backscattering configuration at the film surface using the 488 nm line from Ar+ laser, and a double monochromator (Spex 1404) equipped with a charge-coupled device detector was employed to analyze the scattered light. The profilometer (Veeco DekTak-6) was used to measure the radius of the substrate curvature before and after the growth of AlN films, from which we obtained the residual stress according to Stoney’s equation. More than five positions were measured to reduce the error of the stress across the wafer.

3. Results and Discussion

3.1. The XRD Patterns of the AlN Films

Figure 1 shows the XRD patterns of the AlN films prepared at four levels of the laser energy density. All the samples are wurtzite hexagonal structure. The films prepared at 1.3 J/cm2 (sample A) and 2.4 J/cm2 (sample B) have two diffraction peaks around 33° and 36°, corresponding to the AlN (100) and (002) orientation, respectively. The (002) peaks become stronger and sharper with increasing the laser energy density, while the (100) peaks become weaker. This means that the (002) texture of the AlN films is improved. Moreover, sample C, which is prepared at the energy density of 3.5 J/cm2, exhibits a single and sharp (002) diffraction peak with the full width at half maximum (FWHM) of 0.21°, indicating a preferred -axis orientation. By Scherrer’s formula, the calculated average gain size of sample C is about 42 nm. As shown in Figure 2, the highly parallel columnar crystals are observed from the cross-sectional morphology. The film prepared at 4.8 J/cm2 (sample D) has also a (002) orientation, but the FWHM of the (002) peak is about 0.35°, which is a little larger than that of sample C.

In order to describe the crystal texture of the samples in detail, the (002) plane X-ray rocking curves of the AlN films were measured as shown in Figure 3. The FWHM value of sample C (0.75°) is observed much smaller than that of sample D for AlN (002) plane, which accords with the results of XRD. The influence of the laser energy density on the film textures can be explained through the Al-N complex kinetic energy in the laser plasma. Generally, the increasing of the laser power not only enhances the sputtering yield, but also raises the kinetic energy of the particles in the laser plasma. As a result, during the higher energy plasma, the Al-N complex particles have sufficient kinetic energy to attach themselves to the substrate, hence providing the atoms with more opportunities to move to the lowest energy state and to form a highly (002) oriented crystal structure. On the other hand, the substrate may be strongly attacked if the particle kinetic energy is too large (exceeds 3.5 J/cm2 in this experiment), resulting in the surface damage and deterioration in the (002) texture as shown in sample D.

3.2. The Raman Spectra of the AlN Films

The Raman spectrum of the samples was measured to find out the correlation between the crystal texture and the Raman scattering mode. According to the group theory, wurtzite AlN structure belongs to the space group (C63mc) with two formula units per primitive cell. Theory analysis [19] predicts the zone-center optical modes in the representation as . The , , and two modes are Raman active, while the mode is inactive. The and modes are each splited into the longitudinal optical (LO) and transverse optical (TO) components, thus creating (LO, TO) and (LO, TO) modes. The theoretical Raman scattering mode of the AlN film is summarized in Table 1.

The Raman spectra of the AlN films are shown in Figure 4. All samples exhibit two major Raman peaks around 660 cm−1 and 610 cm−1, which correspond to (high) mode and (TO) mode. There are no obvious sapphire modes in the spectrum because the AlN films may be thick enough. In order to characterize the (high) peak and (TO) peak in detail, these two peaks were fitted by Lorentzian functions, whose fitting parameters are listed in Table 2. With the increase of the laser energy density, the (TO) peak becomes weak and the (high) peak becomes narrow. The spectrum of sample C has the smallest (TO) peak and the narrowest (high) peak.

3.3. The Correlation between the Raman Spectra and XRD Patterns

In order to find out the influence of the texture on the phone vibration in AlN films, the XRD and the Raman spectrum of each sample are compared as shown in Figure 5. For comparison, here the degree of (002) orientation of the films was defined as the ratio of the integrated area of (002) peak to the whole area in the XRD patterns. As for the Raman spectra, the FWHM of (high) peaks and the relative integrated areas of (high) mode were obtained from the fitting results. Some other samples prepared by 1–5 J/cm2 laser were also involved. As shown in Figure 5, the correlation between the (high) mode and the (002) orientation degree is very clear. It is seen that the area of the (high) mode is larger for the films with higher (002) orientation. With the appearance of other orientation, the (high) area goes down while the (TO) mode is enhanced.

The close relationship between the Raman scattering modes and the film orientation can be explained using the Raman scattering rule as follows. Generally, the -axis of the wurtzite unit cell is defined as direction, and and are perpendicular to axis. According to the Raman scattering rule of AlN [9], the (high) vibration mode is excited by an electric field perpendicular to direction, while the (TO) mode is excited by an electric field parallel to direction. In this experiment, the direction of propagation of Ar+ laser is parallel to -axis ( direction) and the backscattering geometry is equivalent to in the grains with the perfect (002) orientation. So the (high) mode is allowed for the high (002) orientated AlN films and thus the (TO) mode is forbidden. However, for the grains with multiple orientations, the electric field has the components parallel and perpendicular to -axis, which excites both of the (high) and the (TO) modes. Consequently, the deterioration of (002) orientation and the appearance of other orientations in the films lead to the enhancement of the (TO) mode.

In addition, the broadening of the Raman peaks is the result of the phonon scattering caused by small grains, point defects, interfaces, and stress gradients [21]. In this experiment, we can reasonably assume that there is no strong stress gradients present in the films. The FWHM of the (high) peaks of the AlN films prepared by PLD varied from 10.1 to 32.1 cm−1, which are intermediate between the 3 cm−1 reported by Kuball et al. [22] for the AlN bulk crystals and the 50 cm−1 reported by Perlin et al. [23] for highly defective crystals. In order to explain the broadening of Raman peaks, the FWHM of the (high) peaks as the function of the FWHM of XRD rocking curve is shown in Figure 6. The (high) peak is found to be broadened with the increase of the width of the X-ray rocking curve, suggesting that the broadening of the Raman peaks in the AlN films is associated with degeneration in crystal quality.

3.4. The Dependence of the Raman Spectra on the Residual Stress

The residual stress in the prepared film distorts the crystal unit cell, resulting in the variations on the energy of phonon vibration. Some authors [9, 21, 22, 24] have reported that the Raman peak spectrum can be used to monitor the stress in the films. Figure 7 shows the dependence of the measured (high) mode energy on the residual stress in the AlN films prepared by PLD. The residual stresses were calculated from the radius of the substrate curvature according to Stoney’s equation. For the undoped AlN films under biaxial stress () perpendicular to the growth -axis, the shift in the phonon energy can be written as where is the Raman-stress factor. Tensile (compressive) stress corresponds to ().

As shown in Figure 7, the (high) energy goes down nearly linearly with the decrease of the residual stress in the AlN films. Consequently, the Raman-stress factor can be obtained by combining the energy shift of (high) mode with the residual stress. The slope of the line provides the Raman-stress factor of −4.45 cm−1/GPa in this experiment. This value is in accordance with the data reported in other papers [9, 21]. In addition, the zero-stress energy of 657.1 cm−1 is also in good agreement with the value listed in Table 1.

4. Conclusions

We measured Raman spectra and XRD patterns of the AlN films prepared by PLD to study the influence of the crystal texture on the Raman scattering mode. The (high) mode and the (TO) mode were observed in Raman scattering. The orientation and the crystal quality of AlN films have a great impact on the scattering modes. The deterioration of (002) orientation and the appearance of other orientations lead to enhancement of (TO) mode in the films. The broadening of the Raman peaks can be associated with degeneration of the crystal quality. Additionally, by combining the energy shift of (high) mode with the measured residual stress, the Raman-stress factor of the AlN films prepared by PLD is found to be −4.45 cm−1/GPa for the (high) mode. These results indicate that Raman scattering is a feasible method to detect the crystal texture and residual stress of AlN films. In the further experiment, the luminous and electrical characteristics of the AlN films will be studied for their optical-electrical applications.

Acknowledgments

This work was supported by the National Science Foundation of China no. 61101027, Qingdao Science and Technology Program of Basic Research Projects (11-2-4-4-(6)-jch, 12-1-4-6-(8)-jch), Shandong Province Young and Middle-Aged Scientists Research Awards Fund (BS2012DX032), and SDUST Research fund (2011KYJQ101).