The high-temperature oxidation resistance of Ti1-xSixN films with Si content varying in wide range, 0.13 <= x <= 0.91, is evaluated. Films are grown in Ar/N-2 atmospheres using a hybrid high-power impulse and dc magnetron sputtering (HiPIMS/DCMS) configuration with Si target powered by HiPIMS and Ti target operated in DCMS mode. The substrate bias is synchronized to the Si+-rich portions of the HiPIMS pulses in order to promote solid solution formation. A combination of X-ray photoelectron spectroscopy, elastic recoil detection analysis, and cross-sectional scanning electron microscopy reveals a sharp increase in the oxidation resistance for layers with x > 0.50. The thickness of the oxide layer, following 1 h anneal at 800 degrees C in air, is in the range 150-200 nm for 0.13 <= x <= 0.50 and decreases to only 4 nm with x = 0.91, which is similar to 30 times lower than for the best performing Ti1-xAlxN film (x = 0.64) tested under the same conditions. The oxide forming on top of Ti1-xSixN films with x = 0.41 consists of a SiO2-TiO2 two-phase mixture with a molar ratio given by Si/Ti ratio. In Ti1-xSixN layers with x <= 0.31, the presence of grain boundaries, which act as diffusion paths facilitates Si diffusion towards the bulk of the film resulting in that TiO2, the thermodynamically more stable oxide, terminates the surface. Ti0.09Si0.91N films, are essentially unaffected by the anneal and exhibit a hardness of 23 GPa, which is similar to 30% higher than for the reference SiNz film. Moreover, we demonstrate that 25 nm thick Ti0.09Si0.91N capping layers successfully prevent Ti(0.36A)l(0.64)N oxidation at 800 degrees C. Such approach provides superior oxidation protection compared to alloying TiAlN with Si. Our results suggest that multilayers including nm thin layers of high Si-content TiSiN is a most effective approach to improve high-temperature oxidation resistance of functional ceramic thin films.
Funding Agencies|Knut and Alice Wallenberg Foundation Scholar Grant [KAW2016.0358]; Competence Center Functional Nanoscale Materials (FunMat-II) VINNOVA grantVinnova [2016-05156]; Swedish Research Council VR Grant [2018-03957]; VINNOVA grantVinnova [2019-04882]; Carl Tryggers Stiftelse for Vetenskaplig Forskning [CTS 17:166, CTS 15:219, CTS 14:431]; Swedish Research Council VR-RFISwedish Research Council [2017-00646_9]; Swedish Foundation for Strategic ResearchSwedish Foundation for Strategic Research [RIF14-0053]