Structural and Mechanical Properties of Radiofrequency Ar-N 2 Plasma Nitrided Aluminium

nitriding its comprehensive description of plasma nitriding mechanisms and physiochemical interaction of species phases the species Aluminium (Al) samples were nitrided in 13.56 MHz radiofrequency (RF) discharge of 70:30 vol.% Ar:N 2 mixture at substrate temperature of 400 °C and RF power of 300 W for different fill pressures and treatment times. The nitrided samples were then characterized for their structural and mechanical properties using x-ray diffraction (XRD), scanning electron microscopy (SEM), energy dispersive x-ray spectroscopy (EDS), and Vickers micro hardness testing. XRD peaks (110), (111) and (200) of aluminium nitride (AlN) and their downshift confirmed the nitride formation together with nitrogen diffusion. The metastable cubic phase (c-AlN) shows texture growth and re-orientation dependent on fill pressure and treatment time. SEM micrographs showed the surface morphology with micro-pores. EDS results showed that the nitrogen content was significantly increased with nitriding time, along with the removal of oxides. Vickers hardness of the Al surface was increased more than thrice for 30 hours nitriding.


Introduction
Nitriding is a thermochemical process involving the deposition of nitride layer and the diffusion of nitrogen into metallic materials. The understanding and control of the nitride layer formation is of industrial interest owing to the improvement of surface properties such as resistance to wear, corrosion, and fatigue of metals and alloys 1,2 . AlN layers have attractive properties such as high hardness, electrical resistance and thermal conductivity 3,4 . AlN has been extensively investigated as a wide bandgap semiconductor material expected for many industrial applications such as short-wavelength light emitters, high-speed micro-electric devices, and surface acoustic devices 5,6 . A stable hexagonal phase (α-AlN) has a band gap of 6.2 eV, high chemical and thermal stability, electric resistance, and acoustic properties 7,8 . The metastable cubic-phase (c-AlN) is also attractive with a higher thermal conductivity, ballistic velocity, and acoustic velocity due to its better symmetry than α-AlN 9 .
In addition, c-AlN films are expected as buffer layers for polarity control in the growth of semiconductor films such as GaN and ZnO 10 . However, the synthesis of c-AlN is more tricky than α-AlN 11,12 . Owing to fabrication difficulties of such films and experimental verifications, applications to industrial devices have poorly been implemented yet. Only the growth of ultra-thin films or small crystallites have been reported in recent past 10,[13][14][15] .
Plasma nitriding is one of the most favored methods having diversity in selecting process parameters including plasma composition, discharge frequency, fill pressure, input electrical power, substrate temperature and processing time. It is more economical and requires shorter treatment times and less gas consumption as compared to other commercial nitriding processes 3,4,[16][17][18] . Plasma nitriding obviates the requirement of high temperature due to activation of plasma species responsible for nitriding. Therefore, there is an active ongoing research in the field of plasma nitriding of Al and its alloys [18][19][20][21][22][23][24] . A comprehensive description of plasma nitriding mechanisms is an extremely difficult task without understanding discharge and physiochemical characteristics of plasma, interaction of active nitrogen species with substrate surface and evolution of various phases in the nitrided layer. To achieve quality nitriding, it is essential to optimize the production of active species by a precise control on process parameters including input power, gas fill pressure, composition, substrate temperature, etc. 25,26 . For nitriding of metals, mixing of argon is advantageous regarding the enhancement and control of nitrogen active species concentration and consequent evolution of nitride growth 24,26,27 .
In this communication, we report on nitriding of Al in Ar-N 2 mixture plasma energized by 13.56 MHz RF source and explore the influence of fill pressure and processing time on the microstructure, surface hardness and chemical composition in the nitrided layer. Particularly, the metastable AlN phase growth/re-orientation as a function processing time is investigated.

Experimental Setup
The Ar-N 2 mixture plasma was generated with 13.56 MHz 300 W RF source and commercially available polycrystalline Al (purity 99.4%) samples having 5mm×5mm×3mm dimensions were exposed for different fill pressures (0.1, 0.2, 0.3, 0.4 and 0.5 mbar), at fixed gas composition of 70 vol.% Ar in the mixture. The selection of this particular gas composition is based on early investigations of the same system by using optical emission spectroscopy (OES) for production and optimization of active species of nitrogen (N 2 + and N) responsible for the plasma nitriding 26 . Prior to the treatment, the samples were mechanically polished by using different grit size silicon carbide papers and then mirror polished with 100 and 10 µm powder by using Metkon GRIPO 2V machine. The samples were ultrasonically cleaned in ethanol for 15 minutes and were placed on the bottom electrode for surface treatment. Prior to nitriding, the samples were cleaned for 20 minutes in Ar discharge at input power of 300 W and a pressure of 0.5 mbar in order to remove surface contaminants. To facilitate nitriding process, the samples were heated upto 400 °C temperature monitored by a thermocouple.
The samples were treated for 5 hours at 400 °C for various fill pressures in order to investigate the optimum pressure. Subsequent nitriding was accomplished at the optimum pressure of 0.5 mbar for various nitriding times (3, 6, 9… 30 hours) at fixed substrate temperature of 400 °C. The details of experimental set up are provided elsewhere 28 . The nitrided samples were analyzed by a JEOL JDX-3532 x-ray diffractometer (XRD) operated in detector scan mode at 40 kV, 30 mA using CuKα radiation (λ = 1.540598 Å). The surface morphology and elemental composition of the nitride layer were investigated by a Jeol-JSM-5910 scanning electron microscope (SEM) equipped with energy dispersive x-ray spectroscopy (EDS) attachment (INCA 200/ Oxford Instruments, U.K). The surface hardness was measured by using a Vickers micro hardness tester (Wilson Instrument 401 MVA) at applied loads of 10, 25, 50 and 100gf.  Figure 1 show respectively typical (111) and (200) peaks of AlN. Al 2 O 3 (012) peak was also observed at 25.54° in the spectrum of the sample nitrided at 0.4 mbar (JCPD-ICDD database 461212), which reports the possibility of significant production of active oxygen species under these conditions. The presence of oxygen content owes to the background air impurities at the working pressure as well as the native oxide layer on Al surface. The nitriding proceeds at a certain rate dictated by the removal of native oxide layer on the Al surface. However, the background air in the chamber supplies some oxygen content in parallel that re-oxides the Al. Both nitriding and oxidation take place with their rate depending upon the plasma conditions e.g. gas pressure, composition, substrate temperature, time etc. The effect of pressure on oxide removal and nitride formation is depicted by the presence of Al 2 O 3 for 0.4 mbar. At this pressure electron energy distribution function is favorable for efficient oxygen active species production. The crystallinity of AlN (200) peak was maximum for the sample nitrided at 0.5 mbar pressure. It is well established that AlN formation is not the sole process during nitriding of Al owing to much less chemical affinity of nitrogen with Al. In addition to the compound formation, nitrogen diffusion occurs significantly at substrate temperature of 340 to 460 °C 29 .

XRD analysis
The XRD spectra of the samples nitrided for various treatment times are shown in Figure 2. All the spectra show the appearance of (110), (111) (110) and (111) and (200) peaks.
Al to form AlN along with nitrogen diffusion in the near surface region. We observe that AlN grows preferentially in (110) and (111) plane orientations for 12 hours and in (200) plane for 21 hours treatment. This can be associated with the optimum activation energy of AlN in a particular plane under these conditions. The metastable c-AlN phase evolves with plane orientations that change with the treatment time. Ogata et al. 18 investigated the crystalline orientation control of AlN films prepared by ion-beam-assisted technology as a function of ion energy. The semi-quantitative analysis of the composition of AlN can be accomplished from the XRD peak intensities. The weight fraction of AlN in the sample surface W AlN can be estimated using the following Formula 1 30 : where A Al and A AlN represent the x-ray integrated intensities of Al and AlN peaks, respectively. It is observed that W AlN varies from 0.59 to 0.88 respectively for 12 hours and 21 hours nitriding in (110) and (111) plane orientations when estimated against Al (111) plane. This is due to the texture re-orientation during nitriding process. The crystallite size of AlN is measured using the Scherrer's Formula 2: where K = 0.99 is the numerical constant, λ = 1.540598 Å is the x-ray wavelength, β is the broadening (in radians) of the diffraction peak and θ is the Bragg's angle. The variation of crystallite size of AlN estimated from a typical (110) plane at different fill pressures is shown in Figure 3a. With the increasing pressure, the crystallite size increases with a typical value of ~100 nm for 0.5 mbar pressure. This corresponds to a maximum N 2 + active species predominantly responsible for AlN formation because the nitriding processes correlates with emission intensity of . The variation of crystallite size of Al matrix as a function of treatment time is also displayed in Figure 3a. A slight reduction in crystallite size of Al is caused due to stress incorporation by ion diffusion. Ion bombardment of noble gases also causes grain refinement and defects, as well as, increases the diffusion of active nitrogen species 32 . The diffusional transport during ion nitriding of Al is also accompanied by the diffusion of Al from the bulk to the surface that helps in AlN layer growth 33 .
The rate of reaction in ion-nitriding can be enhanced by ion-or electron-impact-induced dissociation of physisorbed molecules 34 having a similar effect as the activation in the gas phase. During the reaction of low and medium energy ions (such as nitrogen) with solid surfaces 35 increasing the energy above a certain threshold leads to ion implantation i.e. nitrogen diffusion. Ochoa et al. 32 have investigated that noble gas bombardment generates stresses in the material. Moreover, nitrogen diffusion also produces stresses in the nitrided layer as well as in the near surface Al. We observe a shift in Al (111) peak having standard position at 2θ = 38.50 o . The down-shifting of peak indicates tensile stresses caused by diffusion of interstitial nitrogen into the Al lattice that generates lattice distortion. The stresses are generally tensile in nature and can be estimated by the Formula 3: where d OBS and d STA represent respectively the d-spacing (A˚) of the stressed and the stress-free diffraction planes parallel to the film surface. E is the Young's modulus and v is the Poison's ratio. Using the above expression, the tensile stress corresponding to strong Al (111) peak at 2θ=38.50 o varies from 0.83 GPa to 2.00 GPa by varying the treatment time and is presented in Figure 3b. It is observed that increasing treatment time initially increases tensile stresses, approaching maximum value (2.0 GPa) for 21 hours and then decreases again. This is explained by the fact that nitrogen diffusion interstitially causes texture re-orientation in the Al lattice as well as the nitride layer, leading to the phase growth in particular plane orientation depending upon the activation energy required. The texture re-orientation/phase growth is accompanied by the stress relaxation. The saturation of nitrogen diffusion for treatment time above 21 hours may be the reason of slight decrease in tensile stress. All the diffraction peaks show broadening and shifting. The broadening increases with treatment time and approaches to the maximum value for 21 hours treatment, and then decreases owing to saturation of nitrogen diffusion accompanied by stress relaxation. The peak broadening not only reflects the crystallite size reduction but also includes the microstrain effect resulted by crystal defects 36 .

SEM/EDS analysis
A visual examination reveals that the nitrided surfaces are dark grayish in color. Figure 4 provides the SEM micrographs of an untreated sample and samples nitrided for 3, 15, 21 and 30 hours. The surface of the untreated specimen (Figure 4a) appears smooth but having polishing marks. The surface of sample nitrided for 3 hours (Figure 4b) has sparsely distributed shallow ion impact craters/micropores formed due to nitrogen bombardment. The surfaces of sample nitrided for time upto 12 hours (not shown) are found to exhibit similar morphology having micropores of almost identical size and distribution. Such micropores are resulted by the ion-induced impact, restructuring and defect migration during auxiliary heating of the substrate. However, for 15 hours treatment (Figure 4c), the surface is rough owing to significant surface damage by ions bombardment, but covered with shallow layer of nitride.
For 21 hours nitriding (Figure 4d), nitride layer grows with comparatively smooth surface but having some micropores. The spatial distribution of the micropores is, however, small and the layer appears denser. By increasing the nitriding time to 30 hours (Figure 4e), the nitride surface appears granular in nature. The micrograins have uniform dense spatial distribution, clearly patterning a continuous layer. The nitride pattering is accelerated with treatment time since the nitrogen diffusion in Al and the reaction rate of nitrogen with Al substrate at 400 o C is significantly enhanced for higher treatment times having higher total density of active species. SEM investigations show that the processing time significantly affects the nitriding process, and eventually the surface morphology of the nitride layer.
The compositional analysis of the nitrided samples shows that nitrogen content gradually increases with nitriding time (from 3-30 hours). Table 1 provides the EDS data showing the variation of nitrogen and oxygen concentration with treatment time. For 3 hours nitriding, the oxygen content in the nitrided surface is reduced due to removal of oxides by ion bombardment. But the oxygen content again increases for 6 hours treatment due to enhanced oxidation in the plasma environment compared with the nitriding. However, by increasing processing time above 6 hours, the nitriding process overcomes the oxidation. Eventually, the nitride growth proceeds along with the oxide removal. This can be explained as follows: The nitriding process is accompanied by the oxidation natively present in the background environment. Even at a favorable discharge condition for nitriding i.e. 0.5 mbar pressure, the effect of oxidation due to background air cannot be eliminated for a certain time (9 hours in the present experiment). The nitride dominates oxidation initially upto 3 hours. However, for 6 hours treatment the substrate temperature is such that thermal energy of the plasma-substrate system facilitates poorly the oxide removal. But as the time increase (9 hours and above) further nitriding efficiently removes the oxides.
It has been observed that the nitrogen and oxygen contents in the nitrided Al are strongly dependent on nitriding time. Figure 5 shows the variation of surface hardness of the samples nitrided for different treatment times at fill pressure of 0.5 mbar as a function of indentation depth. All the nitrided samples with increasing treatment time show an increase in surface hardness as compared to that of untreated one. The maximum surface hardness of 56 HV is estimated at the indentation depth of 1 μm for 30 hours nitriding. The enhanced surface hardness can be attributed to the formation of AlN compound and the nitrogen diffusion interstitially. It has also been strongly supported by the linear relationship between the hardness and nitrogen concentration investigated by Ochoa et al. 32 . It is obvious that up to ~2 µm depth, the surface hardness is doubled. Noticeably, the hardened zone also extends below the top layer owing to nitrogen diffusion. The surface hardness decreases with indentation depth and saturates below 4 μm depth.

Conclusions
Plasma nitriding of Al samples at temperature of 400 o C was performed in Ar-N 2 (70 vol.% Ar) 300 W RF plasma at various fill pressures and treatment times. The structural analysis of the nitrided Al samples confirmed AlN formation having (110), (111) and (200) plane reflections as well as nitrogen diffusion into the substrate surface. The crystallinity of AlN is maximum with the crystallite size of ~100 nm for nitriding at 0.5 mbar pressure. This is owing to the much efficient production of active nitrogen species. For 0.4 mbar, the formation of Al 2 O 3 supersedes the nitriding process owing to the possibility of pronounced production of active oxygen species. The stress relieving occurs for nitriding above 21 hours. Nitrided samples depict smooth but micro-porous surface profile for lower treatment times (3-12 hours). For nitriding above 15 hours, the AlN layer smoothly grows with sparse micro-pores distributed on the surface. For higher treatment time, nitriding process dominates over oxidation. The metastable c-AlN shows texture growth and re-orientation dependent on fill pressure and treatment time. The surface hardness of nitrided Al has been increased more than thrice.