Effect of Hf on structure and age hardening of Ti–Al-N thin films

Abstract

Protective coatings for high temperature applications, as present e.g. during cutting and milling operations, require excellent mechanical and thermal properties during work load. The Ti_1 − xAl_xN system is industrially well acknowledged as it covers some of these requirements, and even exhibits increasing hardness with increasing temperature in its cubic modification, known as age hardening. The thermally activated diffusion at high temperatures however enables for the formation of wurtzite AlN, which causes a rapid reduction of mechanical properties in Ti_1 − xAl_xN coatings. The present work investigates the possibility to increase the formation temperature of w-AlN due to Hf alloying up to 10 at.% at the metal sublattice of Ti_1 − xAl_xN films. Ab initio predictions on the phase stability and decomposition products of quaternary Ti_{1 − x − y}Al_xHf_yN alloys, as well as the ternary Ti_1 − xAl_xN, Hf_1 − xAl_xN and Ti_1 − zHf_zN systems, facilitate the interpretation of the experimental findings. Vacuum annealing treatments from 600 to 1100 °C indicate that the isostructural decomposition, which is responsible for age hardening, of the Ti_{1 − x − y}Al_xHf_yN films starts at lower temperatures than the ternary Ti_1 − xAl_xN coating. However, the formation of a dual phase structure of c-Ti_1 − zHf_zN (with z = y/(1 − x)) and w-AlN is shifted to ~ 200 °C higher temperatures, thus retaining a film hardness of ~ 40 GPa up to ~ 1100 °C, while the Hf free films reach the respective hardness maximum of ~ 38 GPa already at ~ 900 °C. Additional annealing experiments at 850 and 950 °C for 20 h indicate a substantial improvement of the oxidation resistance with increasing amount of Hf in Ti_{1 − x − y}Al_xHf_yN.

Highlights

► Ab initio calculations enable for the prediction of as-deposited structures in Ti_{1 − x − y}Al_xHf_yN. ► Additions of 5 mol% HfN in Ti_{1 − x − y}Al_xHf_yN raise the formation temperature of w-AlN by ~ 200 °C. ► 10 at.% Hf at the metallic sublattice of Ti_{1 − x − y}Al_xHf_yN protect from full oxidation at 950 °C for 20 h.

1. Introduction

Phase transformations in ternary transition metal (TM = e.g. Ti, V, Cr, Zr, Hf) aluminum nitrides, TM_1 − xAl_xN, are the key phenomena for tailoring material properties for industrial applications, such as mechanical engineering. Especially the transition from cubic (c; Fm-3m space group, B1-NaCl prototype) to wurtzite (w; P6₃m, B4-ZnS prototype) structure in TM_1 − xAl_xN thin films has been extensively studied in terms of size effects, lattice strain and deposition kinetics [1–7], as physical and chemical properties of TM_1 − xAl_xN are known to strongly scale with the Al content, x. Among other TM of group IVb–VIb, Ti_1 − xAl_xN has obtained a dominant role as protective coating for cutting tools over the last decades. One reason is the limited solubility of AlN in TiN, resulting in a metastable solid solution after deposition and isostructural decomposition during thermal load, which in turn results in a hardness increase at temperatures of ~ 900 °C. This phenomenon, called age hardening, present in Ti_1 − xAl_xN [8–10] involves the formation of c-TiN-rich and c-AlN-rich domains during heat exposure, which increases the coherency strains during decomposition, before the c-AlN-rich domains transform into their thermodynamically stable w-AlN modification and the mechanical properties quickly deteriorate within this dual phase regime at higher temperatures.

Today, the limits of commercially available Ti_1 − xAl_xN coatings are often exceeded during tooling operations of high strength materials, e.g. high speed steels, nickel- or molybdenum based alloys, which raises the need for smart alloying concepts in order to enhance the thermal stability of ternary Ti_1 − xAl_xN. This in turn requires a thorough understanding of the development of microstructure and thermal stability of quaternary Ti_{1 − x − y}Al_xTM_yN alloys as a function of composition and temperature.

In general, the thermal stability, but also the driving force for decomposition of TM_1 − xAl_xN [5,8–12], increase with increasing Al content within the cubic phase field. Previous works on quaternary Ti_{1 − x − y}Al_xHf_yN thin films were dealing with either relatively low Al contents [13] or did not focus on mechanical properties [14]. In the present work, the impact of Hf on structure of Ti_{1 − x − y}Al_xHf_yN, and especially its effect on thermal stability and mechanical properties is emphasized in a combined theoretical and experimental approach. We thereby concentrate on the region close to the cubic solubility limit, where the biggest effect of age hardening is obtained for the ternary Ti_1 − xAl_xN.

2. Methodology

All coatings investigated were deposited in a lab-scale unbalanced magnetron sputter deposition plant (modified Leybold Heraeus A400) from powder-metallurgically produced Ti–Al–Hf targets (Plansee SE) with increasing amount of 0, 2, 5, and 10 at.% Hf, and Ti:Al ratios of 1:1 (for 0 at.% Hf) or 1:2 (for the targets with 2, 5, and 10 at.% Hf). The process parameters, such as the substrate temperature of 500 °C, a bias voltage of − 50 V, a nitrogen partial pressure of ~ 40%, and the target power density of ~ 9 W cm^− 2, were kept constant during all depositions. In order to synthesize Ti_{1 − x − y}Al_xHf_yN films with comparable AlN mole fractions, x, additional Ti-platelets (∅ 5 × 3 mm³) were placed on the race track of the targets with a Ti:Al ratio of 1:2. These preliminary investigations were performed on single crystalline Silicon substrates ([001]-orientation, 20 × 7 × 0.38 mm³). The resulting chemical compositions were measured by energy-dispersive X-ray analysis (EDX) in a Zeiss Evo 50 scanning electron microscope (SEM), using a TiN standard for the correction of the nitrogen content, according to Ref. [15]. All coating compositions exhibit a slight nitrogen overstoichiometry of ~ 51 ± 2 at.%, and were normalized to 50 at.% for the further presentation, as summarized in Fig. 1a. Furthermore, polycrystalline Al₂O₃ (20 × 7 × 0.5 mm³) and low-alloy steel sheets (∅ 75 × 0.2 mm³) were used as substrates for selected compositions in order to perform thermal annealing treatments on coatings with comparable Al:(Ti + Al) ratios of 0.60 ± 0.01 on the metallic sublattice, e.g. Ti_0.41Al_0.59N, Ti_0.38Al_0.60Hf_0.02N, Ti_0.38Al_0.57Hf_0.05N, Ti_0.35Al_0.55Hf_0.10N. Chemical etching of the low-alloy steel substrate by 10 mol% nitric acid provided substrate free coating material, which was used for the determination of the coating structure thereby effectively avoiding substrate interference during annealing and X-ray analysis.

Fig.1.

(a) Quasiternary system of TiN–HfN–AlN, including theoretical and experimental data on the metasolubility limits of Ti_{1 − x − y}Al_xHf_yN. (b) Energies of formation, E_f, for Ti_{1 − x − y}Al_xHf_yN with y = 0 and 0.11 as a functions of the Al content, x.

The annealing treatments were performed in a vacuum furnace (HTM Reetz, base pressure < 5 × 10^− 4 Pa), using a heating rate of 20 K min^− 1, and a cooling rate of 50 K min^− 1, which prevents from structural changes during the cooling period. Individual samples (films on Al₂O₃ and free standing film material) were kept for 1 min at different annealing temperatures (T_a = 800, 900, 1000, 1100 °C for both and 1200, 1300 and 1400 °C only for film powder). Subsequently, the film hardness was determined by nanoindentation from the films on Al₂O₃ at room temperature. A minimum of 20 indents, ranging from 6 to 25 mN according to a plateau test [16] enabled for reasonable statistics on film hardness. The coating structure was identified before and after annealing from free standing coating material by X-ray diffraction (XRD; Cu K_α-radiation) using a Bruker D8 diffractometer in Bragg–Brentano geometry between 20 and 85°. Additionally, the as deposited films with varying Al-content on Si-substrates were structurally characterized by XRD.

The oxidation resistance of the Ti_{1 − x − y}Al_xHf_yN films was estimated by annealing separate coatings on Al₂O₃ in ambient air at T_a of 850 and 950 °C for 20 h, respectively. Subsequent SEM examination of fracture cross-sections of the coatings, enabled for the determination of the remaining unoxidized film thickness and the thickness of the oxide scale. The unaffected nitride layer thickness was normalized to the as deposited thickness of the individual films for a better comparability.

To support our experimental findings, we carried out quantum mechanical calculations employing Vienna Ab-initio Simulation Package (VASP) [17,18] and PAW-GGA pseudopotentials [19]. We used akin parameters as in our previous studies [11,20–22], which guarantee the accuracy of several meV at^− 1. The alloys were modeled using special quasi-random structures consisting of 36 (cubic B1) and 32 (wurtzite B4) atoms. Their exact configurations can be found in Ref. [20]. The number of atoms in the cell determines the available compositional steps. In this case the composition can be adjusted with 5.56 at.% (i.e. 1/18) and 6.25 at.% (i.e. 1/16) steps on the metallic sublattice of the cubic and hexagonal structure, respectively. While a recently proposed methodology by Lind et al. [23] uses an akin approach throughout the whole ternary phase field for the exploration of decomposition phenomena in c-Ti_xCr_yAl_zN thin films, this work focuses especially on moderate Hf-contents (up to 12.5 at.% at the metallic sublattice) and high Al-contents (close to the metastable cubic solubility limit).

3. Results and discussion

3.1. Structure and phase stability of Ti_{1 − x − y}Al_xHf_yN after deposition

The (quasi)ternary phase field of TiN–HfN–AlN, including some previously published data, is presented in Fig. 1a. Filled and half-filled symbols refer to a single phase cubic and dual phase (cubic + wurtzite) structures, respectively, while the filled hexagons correspond to a single phase wurtzite structure. A full solubility in the cubic crystal structure is obtained along the tie-line of TiN–HfN [24,25], while in the TiN–AlN [26,27] and HfN–AlN [11] border systems only a limited range for the cubic single phase exists, before a dual phase regime is entered at sufficiently high Al contents, followed by a single phase wurtzite phase field.

The reported solubility limits of the ternary Ti_1 − xAl_xN [11,26] and Hf_1 − xAl_xN [11] systems are connected via dashed orange lines. Results from Cremer et al. [14] and Kutschej et al. [28] are in good agreement with the interpolated cubic and wurtzite solubility limits. The Ti_{1 − x − y}Al_xHf_yN films on Si investigated in this work, all cross the cubic solubility limit due to different Al contents (see diamond symbols in Fig. 1a). It can be seen that the experimentally observed upper solubility limit for single phase cubic Ti_{1 − x − y}Al_xHf_yN reasonably agrees with the ab initio predictions in this work, and derived from the interpolation of the ternaries in Ref. [11].

In order to corroborate these findings, the energy of formation, E_f, of Ti_{1 − x − y}Al_xHf_yN as a function of AlN mole fraction, x, with constant HfN amounts of y = 0 and 0.11, is plotted in Fig. 1b. The crystal structure for a given alloy composition is given by the lowest E_f [4,29], defined as a difference between the energy of the alloyed crystal and the energies of individual forming elements in their respective single-crystal structures or molecule in the case of N, weighted by the alloy composition. In the cross-over region, characterized by similar values of E_f, a dual phase structure is likely to be formed.

Calculations of the density of states (DOS) of Ti_{1 − x − y}Al_xHf_yN in the cross-over region have shown that Hf has no significant influence on the electronic structure of the quaternary alloy. This presumably stems from the isovalent nature of Hf and Ti (both group IVb transition metals), and indicates that additions up to ~ 11 at.% Hf at the metallic sublattice do not influence the cubic meta-solubility limit. It is however surprising that the effect of Hf on the magnitude of E_f is only in the range of a few meV at^− 1 for all AlN mole fractions, although the atomic radius of Hf (155 pm) is 10.7% larger than Ti (140 pm). The same conclusions have recently been reported for the ternary Ti_1 − xAl_xN and Hf_1 − xAl_xN systems [11]. There, the cubic-phase solubility limit is the same for both systems (x ~ 0.7), but a broad dual-phase transition region between ~ 0.45 < x < 0.7 is predicted for Hf_1 − xAl_xN (in contrast to a practically single-valued transition region between single-phase cubic and wurtzite fields in the case of Ti_1 − xAl_xN).

The higher melting point of HfN (T_m^HfN ~ 3387 °C) compared to TiN (T_m^TiN ~ 3290 °C) and AlN (T_m^AlN ~ 2800 °C) [30] nonetheless promises increased thermal stability of the quaternary Ti_{1 − x − y}Al_xHf_yN. For each Hf amount in the target (0, 2, 5, 10 at.%) a chemical composition close to the cubic solubility limit was chosen for further investigations of the thermal stability of Ti_{1 − x − y}Al_xHf_yN. These coatings are reflected by green asterisks in Fig. 1a.

3.2. Phase stability and decomposition products of Ti_{1 − x − y}Al_xHf_yN

XRD patterns of the selected free standing Ti_{1 − x − y}Al_xHf_yN thin films in the as deposited state are presented in Fig. 2a. A single phase cubic solid solution is obtained for all Ti_{1 − x − y}Al_xHf_yN films, exhibiting a significant peak shift towards lower angles due to the incorporation of larger Hf atoms on Ti or Al lattice sites. The stress free lattice constants of 4.148, 4.174, 4.183, and 4.199 Å for Ti_0.41Al_0.59N, Ti_0.38Al_0.60Hf_0.02N, Ti_0.38Al_0.57Hf_0.05N, and Ti_0.35Al_0.55Hf_0.10N, respectively, were determined by the Cohen–Wagner method [31]. A positive deviation in the order of 0.5% from the ab initio obtained values was observed for all films and is discussed in detail in Ref. [21] for a similar system. The peak intensity first increases with increasing amount of Hf in the films (from 0 to 5 at.% at the metal sublattice), indicating an increasing feature size (grain or column diameter) from ~ 14 nm (y = 0) to ~ 24 nm (y = 0.05), while the film with y = 0.1 exhibits broadened peaks and a smaller feature size of ~ 10 nm, as determined by single line profile analysis [32]. The Hf incorporation in the films goes along with a steady increase of the residual stresses from ~−1.6 ± 0.1 GPa (y = 0) to ~−1.85 ± 0.1 GPa (y = 0.03) to ~−2.95 ± 0.1 GPa (y = 0.05) to ~−3.4 ± 0.1 GPa (y = 0.1), as determined by the substrate curvature method from the corresponding films on Si substrates.

Fig. 2-rev.

XRD patterns of free standing Ti_{1 − x − y}Al_xHf_yN films with increasing amount of Hf (a) in a single phase cubic structure after deposition and (b) after decomposition into the stable dual phase structure after annealing to 1400 °C. (c) Mixing enthalpies, H_mix, of the ternary boundary systems TiN–HfN, TiN–AlN and HfN–AlN from Fig. 1a.

Vacuum-annealing of the free standing film material to 1400 °C for 1 min results in the formation of a dual phase structure, as shown in Fig. 2b. It can be assumed, that the observed structural features mirror the thermodynamically stable phases after the thermally activated decomposition of the metastable Ti_{1 − x − y}Al_xHf_yN films. Only the cubic peaks shift to lower angles, indicating the incorporation of Hf in the fcc-TiN lattice, whereas the w-AlN peaks are at their standard positions [33]. Hence, a solid solution of Ti_1 − zHf_zN (with z ≈ y/(1 − x)) has formed in addition to pure w-AlN. The thermodynamic argument for this behavior is found in the mixing enthalpies, H_mix, of the ternary boundary systems, see Fig. 2c. For completeness, the compositions of the respective ternary boundary systems corresponding to the investigated Ti_{1 − x − y}Al_xHf_yN films are indicated by pale-color rectangles.

3.3. Phase evolution and mechanical properties of metastable Ti_{1 − x − y}Al_xHf_yN during decomposition

The formerly discussed age hardening phenomenon of Ti_1 − xAl_xN, due to spinodal decomposition [34,35], strongly correlates with the limited solubility of metastable c-AlN in c-TiN. As a prerequisite, the second derivative of the Gibbs free energy curve has to be positive and the same crystal structure has to be maintained [34,35]. This also implies that H_mix is strongly positive, indicating a strong driving force for decomposition. This is the case throughout the whole composition range for Ti_1 − xAl_xN, as shown in Fig. 2c. The solution of TiN in AlN and vice versa is thus unlikely for equilibrium conditions. A similar situation is obtained for Hf_1 − xAl_xN, as reported in Refs. [5,11] and this work, however exhibiting an even higher driving force for decomposition. In a strong contrast, a solid solution of Ti_1 − zHf_zN reaches almost negligible mixing enthalpies over the whole composition range, and hence almost no driving force for phase separation, see Fig. 2c (green triangles). This result is consistent with literature reports on the existence of ternary Ti_1 − zHf_zN cubic solid solutions [24,25].

A comparison of the experimentally observed and ab initio predicted lattice constants (positive deviation of ~ 1%) also supports the argument above. The stress free lattice parameter of Ti_1 − zHf_zN shifts from 4.239 to 4.244 to 4.258 to 4.283 Å for z = 0, 0.04, 0.11, and 0.21, respectively. In addition, an increasing peak width can be observed with increasing Hf amount in the coatings, which points towards a finegrained microstructure which is retained even after the decomposition into the thermodynamically stable compounds.

Additional annealing treatments in vacuum were performed on free standing films as well as films on Al₂O₃ substrates, in order to determine the impact of Hf alloying on the decomposition behavior of Ti_{1 − x − y}Al_xHf_yN. Fig. 3 presents the structural evolution as a function of annealing temperature and alloy composition. The above mentioned isostructural decomposition of the ternary Ti_0.41Al_0.59N film when annealed to 900 °C is accompanied by the formation of c-TiN-rich and c-AlN-rich domains, visible by a slight peak broadening of the cubic peaks, in agreement with the literature [8–10,36]. This peak broadening becomes even more evident for the Ti_0.38Al_0.57Hf_0.05N and Ti_0.35Al_0.55Hf_0.10N films and suggests the formation of c-Ti(Hf)N-rich and c-AlN-rich domains, similar to the ternary film.

Fig. 3-rev.

Structural development of (a) Ti_0.41Al_0.59N, (b) Ti_0.38Al_0.57Hf_0.05N, and (c) Ti_0.35Al_0.55Hf_0.10N as a function of T_a, measured at room temperature after annealing.

The XRD results indicate that the formation of the w-AlN phase is retarded to higher T_a for the Hf containing films. While Ti_0.41Al_0.59N exhibit traces of w-AlN (first appearing at a 2 Θ angle of ~ 33°) at 1000 °C, the formation temperature for w-AlN increases to 1200 °C and 1100 °C for the films with 5 and 10 at.% Hf at the metallic sublattice, respectively. This result goes in line with observations on the impact of Zr on Ti_1 − xAl_xN, where additions of > 5 at.% Zr to the metallic sublattice have proven to be less effective in retarding the appearance of w-AlN than smaller amounts up to 5 at.% [20]. Akin investigations on the impact of Cr.

The development of film hardness, measured after annealing on individual samples for each annealing treatment, is shown in Fig. 4. A Hf induced solid solution hardening is observed for all Ti_1 − x-yAl_xHf_yN films after deposition, where the hardness gradually increases from ~ 29.8 ± 1.2 GPa (y = 0) to ~ 37.5 ± 2 GPa (y = 0.1). During the isostructural decomposition up to ~ 900 °C a hardness increase by roughly ~ 25%, to ~ 38.2 ± 1.5 GPa, is observed for Ti_0.41Al_0.59N. Higher T_a result in the formation of w-AlN and subsequent grain growth, thus reducing the mechanical strength of the material and decreasing hardness with increasing amount of w-AlN, compare Fig. 3a. A similar but less pronounced age hardening behavior, with a peak hardness of ~ 37.9 ± 2.6 GPa, was found for the films with 2 at.% Hf at the metallic sublattice. Although the hardness of Ti_0.38Al_0.60Hf_0.02N increases already for T_a = 700 °C, the temperature for peak hardness as well as the formation temperature for w-AlN is similar to Ti_0.41Al_0.59N. Increasing the amount of Hf in the film to ~ 5 at.% at the metallic sublattice however induces a shift of the peak hardness (~ 40.3 ± 1.6 GPa) to T_a ~ 1000 °C, before over-aging occurs. Although the hardness of Ti_0.35Al_0.55Hf_0.10N also slightly increases with annealing temperature to ~ 40.0 ± 1.5 GPa, the extent of the age hardening is only ~ 7% and thus less pronounced than for the less Hf containing coatings investigated. This behavior can be explained, taking into account the microstructural starting point of the different films, since a smaller initial grainsize of Ti_0.35Al_0.55Hf_0.10N compared to Ti_0.38Al_0.57Hf_0.05N also implies shorter diffusion length necessary for the development of c-Ti(Hf)N-rich and c-AlN-rich domains during annealing. The subsequent precipitation of w-AlN can thus occur at lower T_a which is in agreement with the XRD results, compare Fig. 3b and c, respectively. Another argument can be found in the higher residual stresses of the Hf containing coatings, which in turn increase the driving force for spinodal decomposition and thus the hardness increases already at lower T_a. However, the Hf induced compressive stresses can also be sustained at high temperatures in the c-Ti(Hf)N-enriched domains, which effectively hinders the formation of w-AlN with larger volumes (~+24%) [10,37].

Fig. 4-rev.

Evolution of film hardness of Ti_{1 − x − y}Al_xHf_yN as a function of T_a, measured at room temperature after the respective annealing treatment.

3.4. Oxidation resistance of metastable Ti_{1 − x − y}Al_xHf_yN during decomposition

As discussed above, Hf retards the isostructural decomposition and thus the formation of w-AlN to higher T_a more likely due to its bigger size and higher activation energy for diffusion, than due to electronic effects as reported for e.g. Tantalum [38]. The Ti_{1 − x − y}Al_xHf_yN coatings on Al₂O₃, oxidized in ambient air at T_a = 850 and 950 °C for 20 h, exhibited unaffected nitride layer thicknesses (normalized to the corresponding as deposited thickness) as presented in Fig. 5 (top right). While the Hf free Ti_0.41Al_0.59N films fully oxidized during both annealing treatments, all Ti_{1 − x − y}Al_xHf_yN films with y = 0.02, 0.05, and 0.1 retained a dense unoxidized structure with only thin oxide scales on top of the coatings after annealing at T_a = 850 °C. The oxidation resistance thus increases with increasing Hf content in the coatings. Increasing T_a to 950 °C yields however a fully developed oxide scale throughout the Ti_{1 − x − y}Al_xHf_yN coatings with y = 0, 0.02 and 0.05, while the only film with an intact remaining nitride layer (with a normalized thickness of ~ 61%) was observed for Ti_0.35Al_0.55Hf_0.10N. This coating was polished and examined by EDX elemental mapping, as shown in Fig. 5 for the respective elements. The formation of a multilayered oxide scale has been observed, consisting of alternating Al-oxide rich and a mixed Ti–Hf–oxide rich layers. The improved oxidation protection due to Hf alloying of Ti_1 − xAl_xN is in general agreement with the literature [14,39], where the beneficial effect is related to the higher activation energy for diffusion of Hf, compared to Ti or Al, however lower oxidation temperatures were used. In the present case, the finer grain size of Ti_0.35Al_0.55Hf_0.10N, compared to the other coatings investigated, implies shorter diffusion distances for e.g. Al outdiffusion of the grains, which thus explains also the appearance of the layered structure with alternating Al-rich and Ti–Hf-rich oxides.

Fig. 5.

Polished Ti_0.35Al_0.55Hf_0.10N coating after oxidation at 950 °C for 20 h in electron backscatter contrast (top left). The other images correspond to the respective elements obtained from EDX mapping. The remaining unoxidized nitride thicknesses of Ti_{1 − x − y}Al_xHf_yN, normalized to the initial coating thickness, as a function of HfN mole fraction and oxidation temperature (exposure time each 20 h) are given in the top right diagram.

4. Conclusions

In summary, the use of DFT calculations enables for reasonable prediction of phase stabilities in the as deposited state of quaternary Ti_{1 − x − y}Al_xHf_yN alloys and can further be used to predict their thermodynamic equilibrium phases formed during vacuum annealing. Solid solution hardening with increasing amount of Hf in the coatings can be observed, as well as a beneficial effect on the mechanical properties of Ti_{1 − x − y}Al_xHf_yN as a function of T_a, as long as the initial grain size of the films is comparable. The increasingly limited kinetics with increasing amount of Hf in the films retard the formation of w-AlN to ~ 200 °C higher T_a in the case of 5 at.% Hf at the metallic sublattice and yield a film hardness of ~ 38 ± 2 GPa after annealing to 1100 °C. Moreover, the oxidation resistance at T_a = 850 °C for 20 h in ambient air is significantly increased with increasing Hf content, where the Hf free Ti_0.41Al_0.59N coating fully oxidized. Increasing the oxidation temperature to 950 °C only yields an effective oxidation protection for the Ti_{1 − x − y}Al_xHf_yN films with y = 0.1. The used approach, combining theory and experiments, thus provides a methodology capable for the development of application tailored quaternary alloys with optimized thermal properties.