OPTICAL PROPERTIES OF POWDER AND CERAMICS OF ALUMINIUM OXYNITRIDE OBTAINED BY SELF-PROPAGATING HIGH-TEMPERATURE SYNTHESIS

1. Introduction

Aluminium oxynitride (AlON) is a  well-known compound since 1959 when for the  first time Yamaguchi and Yanagida reported the  possibility of a  spinel-type phase in the  Al₂O₃–AlN system [1]. Nowadays, a high interest in AlON ceramic is caused by its excellent optical [2–4] and mechanical [5] properties. For instance, a  combination of thermomechanical properties and transparency in UV, VIS, and IR of AlON is regarded as a challenge to tempered glass and sapphire single crystals in many applications [2]. Moreover, this material is also promising for high-temperature applications in aerospace engineering and power industry [6].

The classical methods of AlON synthesis are (i) the solid state reaction of AlN and Al₂O₃ [7], and (ii) the carbothermal reduction of Al₂O₃ [8] at temperatures higher than 1700°C for a few hours. However, these methods have serious disadvantages such as a high cost of initial materials (for example, AlN), carbon impurities (which cause pores in sintered ceramics) and a low energy efficiency and complexity of the technology. In this study, we propose to utilize the self-propagating high-temperature synthesis (SHS) of AlON ceramics which is a highly efficient method to produce powders and ceramics, especially nitride ceramics [9]. This method is based on the combustion processes – the exothermic reaction of initial reagents and using only internal energy of the  reacting systems without external heating. The  SHS method allows one to synthesize a  large number of high-temperature materials and ceramics based on individual nitrides or composite materials useful for electronics, aerospace engineering, nuclear industry, metallurgy, etc. There are currently no reports about production of transparent AlON ceramics from SHS powders.

2. Experiment

2.1. AlON powder synthesis

Commercial powders of Al (АSD-4 brand, d₅₀  = 7.5 µm), Al₂O₃ (α-Al₂O₃, 99.9% pure, d₅₀ = 55 µm), Mg(ClO₄)₂ and KClO₄ (99.9% pure, reagent grade) were used as starting materials. Magnesium perchlorate Mg(ClO₄)₂ or potassium perchlorate KClO₄ were added to the  mixture as an oxidizing agent. These powders taken in amounts to achieve single phase AlON powder (35.7 mol.% AlN/64.3 mol.% Al₂O₃) were dry-mixed in a ball mill for 4 h. The reactive mixture (40  g) was ignited by heating nichrome 80/20 wire at the top of the sample in a BN crucible under 20 MPa nitrogen pressure. The SHS reaction was completed in ~90  s. After cooling, the  products were crushed and screened by using a 120 μm mesh. The  combustion products were investigated using the X-ray phase analysis (Dron-3M), chemical analysis, scanning electron microscopy (SEM) for the  material microstructure and particle morphology (a LEO-1450 scanning electron microscope with a  built-in IN-CA ENERGY 350 (EDS) X-ray analyzer). These methods are traditionally used for the characterization of the phase content and structure parameters of ceramic samples [10–13]. Rietveld refinement was performed on XRD patterns t o calculate the lattice parameters of analysed phases. More information about combustion experiments is presented elsewhere [14, 15].

2.2. AlON ceramics sintering

AlON powder was ball-milled in isopropyl alcohol at 300 rpm for 6 h using Al₂O₃ balls of 2 mm diameter. 0.5 wt.% of Y₂O₃ was added to the milling process as a sintering additive. Dried powder was screened trough a 40 μm mesh. 6 g of the obtained AlON powder were pressed into a billet of 20 mm in diameter under 50 MPa. The samples were heated to 1930°C with a heating rate of 40°C/min and held for 2, 4 and 6  h. Then the  sintered samples were ground and polished on both sides to 1.2 mm thickness for optical transmittance measurements.

2.3. Optical measurements

Optical parameters were measured using a UV–Vis–NIR spectrophotometer Cary-5000 with a  UMA accessory (Agilent Technologies). The  control of the optical transmittance and reflectance of AlON ceramics is essential for determining the  quality of the  final product as transparency material. The measurement of reflection spectra allows us to estimate the values of the refractive indices of these ceramics according to the  spectrophotometric method of reflection at near-normal incidence angle (6 degrees in our case) [16].

We have measured the diffuse reflection spectra R of AlON powders and calculated the absorption coefficient using the  method for the  quantitative analysis of various powder substances based on the Kubelka–Munk theory [17]. The band-gap (E_g) was determined within the framework of the wellknown formalism for the  dependence of the  absorption coefficient [18, 19].

3. Results and discussion

3.1. AlON powder synthesis

The XRD diagrams of powders synthesized using 14 wt.% KClO₄ (denoted as AlON K) and 11.4 wt.% Mg(ClO₄)₂ (denoted as AlON M) are presented in Fig. 1. As shown, this method allows one to produce aluminium oxynitride powder with the additive of KClO₄. One of the decomposition products of KClO₄ is KCl, and small traces of KCl are also detected in Fig. 1 (top). It should be noticed that potassium chloride is easily soluble in water. The lattice parameter of γ-AlON powder obtained using a mixture containing KClO₄ is 0.7945–0.7951 nm. This lattice parameter corresponds to γ-AlON with the following ratio Al_2.85O_3.45N_0.55.

The XRD diagram of powder synthesized using Mg(ClO₄)₂ (denoted as AlON  M) is presented in Fig. 1 (bottom). This approach is capable to produce single phase aluminum oxynitride powder if the additive of Mg(ClO₄)₂ is 10–11.4  wt.%. The  absence of magnesium chloride in experiments with magnesium perchlorate is explained by a lower boiling point than that of potassium chloride. The  lattice parameter of AlON M powder is 0.7948–0.7953 nm. This lattice parameter also corresponds to γ-AlON having an atomic ratio Al_2.85O_3.45N_0.55. Taking into account that AlON compound is a solid solution of AlN and Al₂O₃, significant differences in the lattice parameters observed in AlON M and AlON K can be explained by a wide homogeneity zone of γ-AlON.

3.2. AlON ceramics sintering

This is our first attempt of the sintering of AlON ceramics from the  powder obtained by the  SHS method. The SEM images of the microstructure of AlON K and AlON M powders before the milling process are presented in Fig.  2(a,  b). The  particle size of as-synthesized AlON  M powder is about 1–3 μm, while the microstructure of AlON K powder represented by significantly bigger agglomerates seems like melted/sintered. After ball milling with 0.5 wt.% of Y₂O₃ the average particle size of AlON K and AlON M powders was d₅₀ = 1.8 μm. Then the  powders were uniaxially pressed under 50 MPa and the obtained green bodies were sintered at 1930°C in a graphite furnace under 0.1 MPa nitrogen pressure for 2, 4 and 6 h. After the sintering process all ceramic samples were grounded/polished on both sides to 1.2 mm thickness for further optical measurements.

3.3. Optical measurements of AlON ceramic samples

Keeping in mind practical applications of this material, it is important to measure the transmission of ceramics. Figure 3 exhibits the transmission spectra of AlON  M and AlON  K ceramics which were sintered for 2 and 6 h in each case. It is clearly shown that the most transparent sample is obtained from AlON M powder sintered for 6 h and its transparency is higher than 20% in the visible spectral range.

The reflection spectra (Fig. 4) of ceramic samples allowed us to estimate the values of refractive indices using the  spectrophotometric method of reflection at a near-normal incidence angle (6 degrees in our case), and this method is described thoroughly in [16]. The dispersion of refractive indices is summarized in Table 1.

Due to the fact that the reflection spectra of ceramics have differences, possibly due to the difference in the surface quality, it was decided to measure the diffuse reflection of AlON M and AlON K powders (Fig. 5).

Based on the  Kubelka–Munk theory, methods for the quantitative analysis of powder materials are successfully developed in [17]. This theory is based on the possibility of analysing the absorption spectra using the diffuse reflection spectra. The spectral dependences of the absorption coefficient k_Abs were calculated from the  measured diffuse reflection spectra R using the Kubelka–Munk formula (Eq. 1)

where R is the normalized measured diffuse reflection coefficient, and s is the  scattering coefficient taken in this case as 1.

The estimation of the band-gap E_g was carried out within the framework of the well-known formalism for the dependence of the absorption coefficient in the fundamental edge region for the direct allowed transitions (Eq. 2)

where hυ is the photon energy, eV.

To estimate the  band-gap E_g of our samples, the values of F(R) were calculated and plotted in the Tauc coordinates hυ-(αhυ)² [18, 19]. The coefficient r differs for direct-gap (r = 1) or indirectgap (r = 4) materials. According to [18], the coefficient for aluminium oxynitride powders AlON is r  =  1 and Eq. (2) is transformed into Eq.  (3), taking into account the Kubelka–Munk function in Eq. (4):

To the  inflection point on the  curve hυ(hυF(R_∞))² a tangent line is drawn, and the intersection point of the tangent line and the horizontal axis is the  value of the  band-gap E_g (Fig.  6). The values of the band-gap E_g for aluminium oxynitride powder AlON K are about 5.7 eV, while for the AlON M powder they are about 5.85 eV.

The values of band-gap energies demonstrate that the  quality of AlON  M samples is better than that of AlON  K. One of the  reasons why the  AlON  K band-gap value obtained from optical measurements is smaller than the AlON M band-gap value is a high concentration of intrinsic defects (shallow traps) in the  material. It means that the AlON M powder, in addition to a better crystallinity and a  more refined microstructure, reveals a lower concentration of the intrinsic defects. All these facts clearly demonstrate the  advantage of the  Mg(ClO₄)₂ additive over the  KClO₄ additive in the  AlON synthesis using the SHS method. However, it is worth noting that the differences in crystallinity and microstructures between the AlON M and the AlON K series are not crucial for the significantly distinguished transparency of ceramics observed for the corresponding samples. Therefore, we suggest that the intrinsic defects play the main negative role in the transparency of the  AlON ceramics. During the  sintering process, the  intrinsic defects usually migrate to the  grain boundaries forming defect interfaces among them that cause the  light scattering and obviously the  deсrease of ceramics transparency. The  smaller concentration of intrinsic defects in AlON M powder is one of the reasons of the significantly better AlON  M ceramics transparency compared to that of the AlON K ceramic sample.

4. Conclusions

We have developed the synthesis method of aluminium oxynitride powder based on the  selfpropagation high-temperature synthesis technique. Using the Mg(ClO₄)₂ additive, single phase AlON powder was obtained, while the  KClO₄ additive leads to small traces of KCl in the  final AlON powder. Using the Kubelka–Munk function we have obtained the band-gap values for aluminium oxynitride powders synthesized with the different additives. Our results show that the AlON powder synthesized with the Mg(ClO₄)₂ additive has a bigger band-gap energy value than the AlON powder produced with the  KClO₄ additive (5.85 versus 5.7 eV, respectively). This fact can indicate that the  AlON  M powder has a  smaller amount of defects (shallow traps). We suggest that these defects play a negative role in the transparency of sintered AlON ceramics. The  best optical transmission was obtained with AlON M ceramics and it is transparent up to 20% in the visible spectral range.

Acknowledgements

The reported study was funded by RFBR according to the  Research Project No.  19-08-00655. V.P. acknowledges the  State Research Program ‘Augstas enerģijas fizika un paātrinātāju tehnoloģijas’ (Projekta Nr.  VPP-IZM-CERN-2020/1-0002). The Institute of Solid State Physics, University of Latvia as the Center of Excellence has received funding from the H2020-WIDESPREAD-01-2016-2017-Teaming Phase2 under Grant Agreement No. 739508, Project CAMART2.

References