1 Introduction

Due to a larger exciton binding energy (60 meV) [1], ZnO has been identified as the great material for light emitter in the UV spectral ranges. Moreover, the ZnO generates more efficient excitonic emissions at room temperature [2]. The surface acoustic-wave devices [3] and varistors [4] are the main devices produced by the ZnO thin films. The ZnO grown on silicon substrate meets the customer application due to the low cost and maturity of silicon technology. The growth of buffer layer between two different bandgap ZnO and silicon substrate helps to reduce the mismatch problem and continuously improve the device performance [5]. Besides, the aluminium nitride (AlN) is widely used as a buffer layer for ZnO layer due to small lattice mismatch (5%) interfaces [6]. The great things arise between ZnO and AlN are both materials have similar lattice constant c between ZnO and AlN and having compatible thermal expansion coefficient, k [7]. AlN has the thermal expansion (4.2 × 10–6 K−1) coefficient between ZnO (4.75 × 10–6 K−1) and Si sample (2.6 × 10–6 K−1) [8]. Additionally, by using deposition conditions on the substrate, such as post-annealing, in situ annealing, rapid thermal annealing (RTA), annealing under the environments of O2, N2, and Ar, annealing under ultrahigh vacuum (UHV), and infrared light ray (NIR) annealing, physical features of ZnO thin films, including their structural, optical, and electronic properties, could be improved [9,10,11,12,13]. Thermal annealing is typical external treatment use to improve the metallic contact and structural, optical and electrical characteristics of the deposited ZnO using different growth methods. The thermal annealing has been proven to enhance the ZnO crystallinity and halting the defect density in their centre. There are few reports discussed the effect of thermal annealing on ZnO/AlN hetero-structure [14]. The groups presented the effect of thermal annealing on the structural, optical and electrical characteristics of ZnO thin film grown by various methods. Various growth methods such as sputtering, molecular beam epitaxy, wet deposition and MOCVD were used to grow ZnO/AlN structures. In this work, AlN/GaN/AlN heterostructure layers were first grown on silicon substrate by radio frequency molecular beam epitaxy (RF-MBE).Then, the ZnO thin film was grown on AlN/GaN/AlN/Si sample by RF sputtering machine. The samples were than annealed in nitrogen ambient for 1 h at 600 °C and 800 °C. The results indicate that ZnO film on AlN shows excellent structural properties. Surface morphological and optical Raman properties are further improved by thermal annealing treatment. The related emission process is discussed.

2 Methodology

The ZnO layer was deposited on the III-nitride template at 100 W powers in room temperature. The growth temperature and argon flow were set at room temperature and 1%. The growth time for ZnO on the III-nitride template was recorded as 4 h and the measured thickness of film was approximately 400 nm. The samples were annealed in the vacuum tube furnace for 600 °C and 800 °C in the nitrogen ambient. The surface morphological, structural and optical properties of samples were investigated by field emission scanning electron microscopy (FESEM) (Fig. 1), atomic force microscopy (AFM), high resolution X-ray diffraction (HR-XRD), and Raman spectroscopy, respectively.

Fig. 1
figure 1

Structure of ZnO/AlN/GaN/AlN/Si substrate (111)

Full size image

3 Results and discussion

The surface analysis of the samples before and after the thermal annealing was introduced (refer Fig. 2a–c). The dimples and mild grains were spotted on the samples surface. The average number of grain sizes existed on the samples surface. Control sample contains grains size almost 27.5 counts for 100 nm less. No thermal treatment was introduced to the control sample and the rougher surface occurred due to the lattice mismatch between the ZnO and AlN interfaces. The 3-dimensional (3-D) surface roughness is generated by the AFM measurement (refer to a Fig. 3). Surface roughness of control, 600 °C, and 800 °C samples were measured as 0.0559 µm, 0.0558 µm and 0.0553 µm, respectively. The AFM results are found to be in a good agreement with the SEM results, in term of the surface morphology and grain size conditions. No cracks or pores were found from the AFM results, suggesting that the deposited layers were free from defect and granule forms. Various common granule ZnO sizes can be found consistently for the low-cost deposition process like sol–gel spin and dip-coating methods [15].

Fig. 2
figure 2

SEM and AFM images of a control, b 600 °C, and c 800 °C samples

Full size image
Fig. 3
figure 3

AFM images of a control, b 600 °C, and c 800 °C samples

Full size image

Figure 4a–c shows the thermal annealing mechanism on the ZnO/AlN/GaN/AlN/Si substrate. The surface has grooves since ZnO was not effective at first when it was applied on AlN (see Fig. 4a). As a result, it is hypothesised that the formation mechanism of the ZnO grooves can be accounted for by the Volmer–Weber-like nucleation because there is a low likelihood of ZnO adhering to AlN. Figure 4b indicates that the size of ZnO grooves becomes lower after exposing the thermal annealing at 600 °C. We propose that the heat annealing method flattens certain ZnO grooves and decreases microstructure flaws. Figure 4c shows that after applying heat annealing, the ZnO grooves are still decreased, and their size is getting smaller.

Fig. 4
figure 4

The top layer (first layer from the top) is the zinc oxide (ZnO) layer fabricated by a radio frequency (RF) sputtering machine. The second layer is the aluminum nitride (AlN) layer, followed by the third layer, the gallium nitride (GaN) layer, and the last fabricated layer is the aluminum nitride (AlN) layer. All the fabricated layers were explained in the methodology section. The ZnO grooves formed on the ZnO surface are the result of the sputtering method

Full size image

The existences of all desired elements with respect to the particular deposited layers were confirmed by the EDS analysis, as shown in Fig. 5. The inset shows the distribution of all elements on the surface sample. Impurities or foreign elements are not detected in the samples, indicating high-quality layers of the as-grown and annealed samples.

Fig. 5
figure 5

EDS spectra of a control, b 600 °C, and c 800 °C samples

Full size image

The high-resolution XRD peaks of ZnO film grown on AlN/GaN/AlN substrates are shown in Fig. 6. The annealed samples in nitrogen ambient at 600 °C and 800 °C are also provided in this work. The hexagonal structure of ZnO (002) was observed in all samples at peak 34.60° and 34.61°. Besides, among all the samples, the 2θ value is closest to standard ZnO phase value, indicating that tensile stress becomes the smallest as the sample is annealed at 600 °C. Diffraction peaks with lower intensity found at: 36.95, 37.8, 38.58, 48.05, 53.89, 55.06, 62.69, 68.76, 70.31, 75.03 and 76.02 (see Table 1) (JCPDS: 03–065-3411). Additionally, they are closely related to the phase of anatase [16]. The lower diffraction peaks located at 36.95°, and 62.69° correspond to the (101) and (103) reflection planes of hexagonal structure of ZnO (JCPDS: 03–065-3411).

Fig. 6
figure 6

XRD analysis of samples annealed at b 600 °C and c 800 °C. Sample a un-anneal is attached as a control sample

Full size image
Table 1 Summary of XRD peaks on samples control, 600 °C and 800 °C
Full size table

The increased thermal annealing dramatically increases the preferred C-axis orientation degree and continuously improves the ZnO crystallization structure [17]. In this work, thermal annealing temperatures are set to be higher than 300° because the poorer reaction process occurs for the excess Zn metal at 300 °C thermal annealing treatment due to higher melting point of Zn [17]. The growth of ZnO layer using pulsed laser deposition (PLD) and RF magnetron sputtering machine also require thermal annealing above 300 °C [18]. The different directions of ZnO crystals could be due to the weaker crystallization process producing polycrystal ZnO structure.

Bragg's law is applied to the relationship between the diffraction angle and the interplanar spacing (d) as follows,

$$n\lambda =2d\mathrm{sin}\theta ,$$
(1)

where n is an integer (n = 1) and stands for the half-diffraction angle. The distance d200 of each film is inversely proportional to the diffraction angle. Table 1 contains the measured lattice parameters, full width at half maximum (FWHM), and crystallite size for each sample. Using the Debye–Scherrer equation, the crystallite size (D) for the (002) and (102) planes of thin films were computed.

$$D=\frac{k\lambda }{\beta cos\alpha },$$
(2)

the Scherrer constant of 0.90, the average crystallite size, and the X-ray wavelength (0.15406 nm), respectively, are represented by the letters k, D, and λ. α is Bragg’s angle, and β is the full width half maximum (FWHM) in radian. The crystallite size for the (002) plane of all samples were almost the same (58.88 nm). However, 61.36 nm, 61.38 nm, and 36.83 nm were recorded at control, 600 °C, and 800 °C, respectively, in the ZnO (102) plane, which provides a distinct result for crystallite size (refer to a Table 2). The length of the dislocation lines per unit volume at preferential orientation of the crystal structure is used to define the dislocation density (δ) using the formula (3) [19].

Table 2 Summary of XRD analysis on the control, 600 °C and 800 °C samples
Full size table
$$\delta =\frac{1}{{D}^{2}}.$$
(3)

With increasing annealing treatment temperatures from 600 °C to 800 °C, the dislocation densities obtained rise from 2.66 × 10–4 nm2 to 7.37 10–4 nm2. We may state that there are obvious structural changes with this increment.

Figure 7 a–c indicates the Raman spectra of the ZnO/AlN/GaN/AlN heterostructures fabricated on silicon (111) substrate for sample control, 600 °C, and 800 °C. Si modes are shown by the peaks at 300 cm−1 and 521 cm−1 for all samples. The GaN and AlN peaks identified in this paper are all in agreement with Davydov et al. [20]. The most important findings of this work is that all Raman-active modes, i.e., E2M, ZnO-like E2 (H), ZnO-like E1 (LO), ZnO-like 2TO, ZnO-like A1 (2LO), ZnO-like 3E2 (H), and ZnO-like A1 (4TO) were observed in the spectra of ZnO/AlN/GaN/AlN heterostructure films. We found that the XRD results and the Raman data show good agreement. Table 3 displays the phonon frequencies of ZnO/AlN/GaN/AlN heterostructure samples, as well as unstrained AlN, GaN, and ZnO at room temperature.

Fig. 7
figure 7

Raman spectrums of ZnO/AlN/GaN/AlN Heterostructure on Silicon Substrate annealed at b 600 °C, and c 800 °C. Sample a un-anneal is attached as a control sample

Full size image
Table 3 Summary of Raman peaks on the control, 600 °C and 800 °C samples
Full size table

4 Conclusions

In conclusion, AlN/GaN/AlN was successfully used as a buffer layer to generate high-quality ZnO films by RF sputtering on Si(1 1 1) substrates. Thermal annealing's effects on structural and optical characteristics have been researched. According to the results of optical and structural evaluation, the ideal thermal annealing temperature is determined to be 600 °C, resulting in the least amount of dislocation density in the films. The acquired results further demonstrate that there is good agreement between the ZnO/AlN/GaN/AlN values from XRD and Raman analysis.