Thin Solid Films 544 (2013) 515–520

Contents lists available at ScienceDirect

Thin Solid Films

j ourna l homepage: www.e lsev ie r .com/ locate / ts f

A comparative study of the oxidation behavior of Cr2N and CrN coatings

Z.B. Qi a, B. Liu a, Z.T. Wu b, F.P. Zhu a, Z.C. Wang a,⁎, C.H. Wu c

a College of Chemistry and Chemical Engineering, Xiamen University, Xiamen, Chinab College of Materials, Xiamen University, Xiamen, Chinac China National R&D Center for Tungsten Technology, Xiamen, China

⁎ Corresponding author. Tel./fax: +86 592 2180738.E-mail address: zcwang@xmu.edu.cn (Z.C. Wang).

0040-6090/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.tsf.2013.01.031

a b s t r a c t

a r t i c l e i n f o

Available online 25 January 2013

Keywords:Oxidation behaviorCrN and Cr2N coatingsPhase transformationOxidation kineticsOxidation mechanism

A comparative study of the oxidation behavior of Cr2N and CrN coatings was investigated in air at tempera-tures ranging from 700 °C to 1000 °C including phase structure, surface and cross-sectional morphologies,oxidation kinetics and mechanism. During oxidation, the dense Cr2O3 scales form both on the Cr2N andCrN coatings. The Cr2N phase transforms to CrN phase, which is accompanied with 20% volume expansionin transverse. This volume expansion induces the porous and non-columnar cross-sectional morphology ofthe Cr2N coating. The worse oxidation resistance of the Cr2N coating compared with the CrN coating is attrib-uted to its higher chromium content and phase transformation induced porous structure. The oxidation acti-vation energy values are 139 and 190 kJ/mol for the Cr2N and CrN coatings respectively, which is comparableto the literature results. Marker experiments prove that the growth of Cr2O3 scales is controlled by both theoutward diffusion of chromium and inward diffusion of oxygen and the outward diffusion of chromium isdominated.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Comparedwithwidely used TiN coatings, CrN coatings exhibit lowerfriction coefficient, higher toughness, better corrosion and oxidation re-sistance, whichmake them promising candidates as protective coatingsformachining tools and engineering components [1,2]. In such industri-al applications, the coatings are often exposed to an oxidizing atmo-sphere at high temperature, inducing oxidation of the coatings whichsignificantly deteriorates mechanical and tribological properties of thecoatings. Therefore, it is necessary to look into the oxidation behaviorof the CrN coatings under high temperature.

Two kinds of chromium nitride Cr2N and CrN coatings can be de-posited by controlling the nitrogen flow [2–4]. Despite the fact thatthe hardness of the Cr2N coating is higher than that of the CrN coat-ing, the oxidation resistance of the Cr2N coating is worse than thatof the CrN coating according to the comparative studies of oxidationbehavior of the two coatings [5–8]. Such studies were carried out byusing the thermogravimetric analysis (a measurement of weightgain) and Rutherford backscattering spectrometry (RBS) (a measure-ment of scale thickness) to explore the oxidation kinetics and mech-anism of the Cr2N and CrN coatings. However, surface and cross-sectional morphological changes of the Cr2N and CrN coatings duringoxidation have been less dealt with, while the morphologies are con-sidered to be important for understanding the oxidation behavior ofthe hard coatings [9]. Therefore, this study aims to investigate thesurface and cross-sectional morphologies of the two coatings after

rights reserved.

oxidation. Phase transformation, oxidation kinetics and mechanismare discussed as well based on this study and previous reportedresults.

2. Experiment details

The Cr2N and CrN coatings were deposited on silicon (111) wafersusing aDCmagnetron sputtering technique in anAr andN2 atmosphere.Chromium target of 99.95% puritywas used and the diameter of the tar-get was 76 mm. The base pressure was below 6.0×10−4 Pa. During de-position, the working pressure was kept at 0.3 Pa and the total gas flowwas 60 SCCMwith nitrogen flow percentage N2/(N2+Ar) fixed at 20%(Cr2N) and 60% (CrN). The DC power was adjusted to 250 W and depo-sition temperature was set at 300 °C. No bias was applied to the sub-strates. The working distance was fixed at 90 mm. The depositiontimewas set at 30 and 60 min for the Cr2N and CrN coatings tomaintainthe thickness of ~2.0 μm. The ZrN coating was deposited at the samechamber, and the zirconium target of 99.99% purity was used. Duringdeposition, the working pressure was kept at 0.5 Pa with nitrogenflow percentage fixed at 20%. The deposition time was set at 90 minfor the ZrN coating with the thickness of ~1.7 μm. Other deposition pa-rameters of the ZrN coating were the same with those of the Cr2N andCrN coatings. The detailed preparation process of the ZrN coating canalso be obtained in our previous literatures [10,11]. After deposition,the Cr2N and CrN coatings were oxidized in air for 2 h in a quartz tubefurnace at temperatures ranging from 700 °C to 1000 °C at 5 °C/min,while the oxidation temperature of the ZrN coating was 650 °C.

Crystal structure of the coatings before and after oxidation was in-vestigated by Panalytical X'pert PRO grazing incidence X-ray diffraction

Fig. 1. The GIXRD patterns of the Cr2N (a) and CrN (b) coatings before and after oxidationin air for 2 h at various temperatures ranging from 700 °C to 1000 °C.

516 Z.B. Qi et al. / Thin Solid Films 544 (2013) 515–520

(GIXRD, Philips, Netherlands) usingCuKα radiationwith a grazing angleof 5°. Raman spectra of the coatings were acquired using a XploraRaman system (HORIBA, France). The excitation wavelength used wasa 532 nm (20 mW) He–Ne laser. Field emission scanning electron mi-croscopy (FESEM, LEO-1530, Germany) operated at 20 kV was used toobserve the surface and cross-sectional morphologies.

In order to understand the oxidation mechanism of the Cr2N andCrN coatings, marker experiments were performed by depositing a5 nm thick Au film on the surface of the coating using electric beamphysical vapor deposition. The coatings with Aumarkers were oxidizedat 900 °C for 2 h in air in the quartz tube furnace. After oxidation, thecontent and depth profile of Au were investigated using X-ray photo-electron spectroscopywith sputter etching. The sputtering ratewas cal-ibrated using standard SiO2 films. If the oxide scale growth wascontrolled only by outward diffusion of metal cations, the Au markerafter oxidation would be at the oxide/nitride interface. If, on the otherhand, the scale growth was controlled exclusively by inward diffusionof oxygen, the Au markers would remain on the surface of the coatings[6].

3. Results and discussion

3.1. Thermodynamic analysis

During oxidation of the Cr2N and CrN coatings at the temperaturesranging from 700 °C to 1000 °C, several reactions occur and areormulated by:

23Cr2Nþ O2 ¼ 2

3Cr2O3 þ

13N2 ð1Þ

43CrNþ O2 ¼ 2

3Cr2O3 þ

23N2 ð2Þ

Cr2Nþ 12N2 ¼ 2CrN: ð3Þ

The Gibbs free energy changes of the reactions are given by

ΔG Tð ÞR 1ð Þ ¼ ΔG0 þ RT ln pN21=3

� �= PO2ð Þ

h i¼ 0:137T−670:7 ð4Þ

ΔG Tð ÞR 2ð Þ ¼ ΔG0 þ RT ln pN22=3

� �= PO2ð Þ

h i¼ 0:089T−600:0 ð5Þ

ΔG Tð ÞR 3ð Þ ¼ ΔG0 þ RT ln pN2−2=3

� �¼ 0:0794T−106 ð6Þ

where, the values of PO2 and PN2 are 0.21 atm and 0.78 atm in air, R isthe gas constant (8.314 JK−1 mol−1), T is absolute temperature. Thestandard Gibbs free energy changes ΔG0 are taken from the NIST-JANAF thermochemical tables [12]. As predicted by thermodynamics,reactions (1) and (2) are thermodynamically favorable. The corre-sponding temperature at which Cr2N is in equilibrium with CrN is1060 °C calculated from Eq. (6). Namely, under non-oxidative condi-tions, the Cr2N coating is thermodynamically stable above 1060 °C,whereas the CrN coating is stable below that temperature.

3.2. Phase structure

Fig. 1 illustrates the GIXRD patterns of the Cr2N [Fig. 1(a)] and CrN[Fig. 1(b)] coatings before and after oxidation. The JCPDS cards used inthis study are 76-2494 (CrN), 35-0803 (Cr2N) and 38-1479 (Cr2O3) re-spectively. The Cr2O3 peaks show up after oxidation at 700 °C. The inte-grated intensity of the strong diffraction peaks Cr2O3 (012) and (110)increases with increasing oxidation temperature, while the integratedintensity of Cr2N (110), (111) and CrN (111), (220) decreases. Further-more, in Fig. 1(a), besides diffraction peaks of Cr2N and Cr2O3, CrN (200)

peak is also observed above 700 °C. Further increasing oxidation tem-perature results in an increase intensity of the (200) CrN peak and con-versely a decrease of the (111) Cr2N peak. Thus, it is concluded thatCr2N transforms to CrN phase under high temperature following reac-tion (3). As the oxidation temperature above 900 °C, CrN becomes thedominated nitride phase. In Fig. 1(b), no peak ascribed to Cr2N isobtained in the whole temperature range. These observations are con-sistent with the above thermodynamic prediction and previous studies[8,13–15]. The occurrence of reaction (3) is attributed to the behavior ofCr2O3 to act as a diffusion barrier for N2 releasing and an increasing Crdepletion of the Cr2N coatings during oxidation [8,13].

The Raman spectra of the Cr2N and CrN coatings before and after ox-idation are exhibited in Fig. 2(a) and Fig. 2(b). As seen in Fig. 2(a), theoxidation products of the Cr2N coating show peaks at 553 and610 cm−1, characteristic peaks of Cr2O3 [15], illustrating that the Cr2Nbegins to oxidize at 700 °C. The intensities of these two peaks increaseand correspondingly more Cr2O3 scales form with increased oxidationtemperature. The peak ~530 cm−1 shows up at 700 °C in Fig. 3(a), char-acteristic peak of CrN [16,17]. It means that CrN forms during oxidationfollowing reaction (3). This is consistent with the GIXRD results. Similarto the Raman spectra of Cr2N after oxidation, the CrN also begins to ox-idize at 700 °C at 553 and 610 cm−1, while the Cr2O3 553 cm−1 peak isoverlapped with the CrN 530 cm−1 peak. More Cr2O3 scales form at ahigher oxidation temperature. After oxidation at 1000 °C, the Ramansignal of silicon (~522 cm−1) appears. At 1000 °C, the peeling off of

Fig. 2. The Raman spectra of the Cr2N (a) and CrN (b) coatings before and afteroxidation.

517Z.B. Qi et al. / Thin Solid Films 544 (2013) 515–520

the CrN coating is observed, which is due to the large difference of ther-mal expansion coefficient between Cr2O3 (9.55×10−6 K−1) and CrN(2.3×10−6 K−1) [15,18]. Thus, the Raman signal of silicon must comefrom the exposed silicon substrate. It is worth noting that the peaks at309, 350 and 398 cm−1 are found both in the Cr2N and CrN coatings.

Fig. 3. The surface morphologies of the Cr2N an

These peaks belong to both CrN and Cr2O3 [15–17]. However, since theintensities of these peaks increase with increasing temperature, thesepeaks are more likely origin from the Raman vibration modes of Cr2O3.

3.3. Surface and cross-sectional morphologies

Fig. 3 illustrates the surface morphologies of the Cr2N and CrN coat-ings before and after oxidation. During oxidation, the surface morphol-ogies change from as-deposited taper and triangular sheet structure toCr2O3 granular and faceted structure. The Cr2O3 grain size increaseswith increasing oxidation temperature and could reach as large as1 μm at 1000 °C due to the grain coalescence effect [15].

Fig. 4 displays the cross-sectional morphologies of the Cr2N and CrNcoatings before and after oxidation (red line is an oxide/nitride inter-face). The as-deposited Cr2N and CrN coatings exhibit similar columnarstructure with almost the same thickness of 2.0 μm. Dense and protec-tive oxide scales are observed after oxidation above 700 °C for Cr2N and800 °C for CrN. The thickness of the Cr2O3 scales increases with increas-ing oxidation temperature. At the same oxidation temperature, thethickness of Cr2O3 scale formed on the Cr2N coating is higher thanthat on the CrN coating, which further confirms the poorer oxidation re-sistance of the Cr2N coating with respect to the CrN coating. Higherchromium content is proposed to account for the poorer oxidation re-sistance of the Cr2N coating, compared with the CrN coating [5–8].The molar volume ratio VCr2O3

:VCr2N and VCr2O3:VCrN are 1.61 and 1.36,

which are comparable to VZrO2:VZrN 1.47, VTiO2

:VTiN 1.74 calculated byusing the data from the handbook [18]. Under the large volume expan-sion, however, dense oxide scales formed on the Cr2N and CrN coatingsare obtained, which is different from porous oxide scales formed on theTiN and ZrN coatings after oxidation [19,20]. This is attributed to the dif-ferent oxide scale growth modes. The growth of TiO2 and ZrO2 scales iscontrolled by inward diffusion of oxygen [20,21], whereas the growth ofCr2O3 scales ismainly dominated by outward diffusion of chromium (asdiscussed below). Furthermore from Fig. 4, some pores behind theoxide/nitride interface are observed for the Cr2N coating after oxidationat 700–1000 °C and the CrN coating at 1000 °C. This behavior was alsofound during oxidation of bulkmetals [22]. It is supposed that theN2 re-leased by reactions (1) and (2) is accumulated behind the oxide/nitrideinterface due to the nitrogen barrier of Cr2O3 [8,20].

For the CrN coatings after oxidation, un-oxidized nitride layers canstill keep the dense columnar structure similar with as-deposited coat-ing. However, for the Cr2N coatings after oxidation, columnar structureis destroyed and un-oxidized nitride layers become porous. Thereforeit is expected that even though the Cr2N has almost completelytransformed to the CrN above 900 °C (from GIXRD and Raman results),this CrN layer formed through phase transformation cannot sustain

d CrN coatings before and after oxidation.

Fig. 4. The cross-sectional morphologies of the Cr2N and CrN coatings before and after oxidation (the red line is the oxide/nitride interface).

518 Z.B. Qi et al. / Thin Solid Films 544 (2013) 515–520

similar oxidation resistance with the CrN (as-deposited) coating as a re-sult of the porous structure. The explanation of the porous structure isthat under high temperature, the pores accumulated behind theoxide/nitride interface may further diffuse into the coatings [8]. Here,another probable reason is proposed. For the Cr2N coatings during oxi-dation, reaction (3) occurs. The molar volume ratio VCrN:VCr2N is ~1.20calculated by using the data from the handbook [18]. Moreover, the N2

released by the oxidation reactions will diffuse inward into the nitridelayer, since the outward releasing is inhibited by the Cr2O3 scales. There-fore, the 20% volume expansion in transverse would cause the porousstructure of the Cr2N coatings. This explanation is confirmed by the re-sults of XRD (Fig. 1(a)), Raman (Fig. 2(a)) and cross-sectional morphol-ogies (Fig. 4). At 700 °C, only a small amount of Cr2N transforms to CrNand the cross-sectional morphology can still keep columnar structure.At 800 °C, a majority of Cr2N transforms to CrN and meanwhile the ni-tride layer becomes porous and the columnar structure almost disap-pears. In case of the CrN coatings, reaction (3) does not occur andtherefore the columnar structure is preserved. Additionally, this volumeexpansion may also induce the formation of pores behind the oxide/nitride interface. So, that is why the amount of pores of the Cr2N coatingis muchmore than that of the CrN coating after oxidation. In conclusion,besides the higher chromium content (proposed by previous studies[5–8]), phase transformation induced porous structure could be anotherimportant factor accounting for the worse oxidation resistance of theCr2N coatings compared with the CrN coatings.

3.4. Oxidation kinetics and mechanism

It has been reported that the oxidation of the Cr2N and CrN hardcoatings follows a parabolic relation by thermogravimetric, Augerelectron spectroscopy and RBS analysis [5–8,15,23,24]. The parabolic

rate constant kp(T) can be calculated from Wagner's parabolic oxida-tion theory:

d ¼ 2ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffikp Tð Þ � t

qð7Þ

where d is the thickness of the Cr2O3 scales from the cross-sectionalobservation, and t is oxidation time (2 h, here). The parabolic rateconstant of the Cr2N coating is about one order of magnitude higherthan that of the CrN coating at the same oxidation temperature (asshown in Fig. 5(a)), stating clearly worse oxidation resistance of theCr2N coating compared with the CrN coating.

The temperature dependence of the parabolic rate constant is de-duced from the Arrhenius equation:

kp Tð Þ ¼ kp0 exp − EaRT

� �ð8Þ

where kp0 is a constant, R is gas constant, T is absolute temperature, Eais the activation energy of oxidation. Fig. 5(a) shows the Arrheniusplot ln kp(T) vs. 1/T for the Cr2N and CrN coatings in this study anddata from previous studies. From the slope of the linear-regression,the values of the activation energy of the Cr2N and CrN coatings arecalculated to be 139 and 190 kJ/mol respectively. Our values are com-parable to previous reported results, 116, 141, 195 [25], 110 [15], 150[23], 163 [26], 225 [27], 131.7 [7], 162.7 [9] 186 kJ/mol [28] for CrN,123 kJ/mol for Cr2N [24] and 157 kJ/mol for mixture CrN and Cr2N[29]. Irrespective of the different oxidation conditions and measure-ment methods, these deviations may be related to differences in mor-phology and stoichiometry of the chromium nitride coatings, whichstrongly depend on the deposition process [7,15,26]. In addition, wesuggest the phase transformation between Cr2N and CrN as anotherimportant reason accounting for the deviations.

Fig. 5. (a) Arrhenius plot ln kp(T) vs. 1/T for the Cr2N and CrN coatings based on thisstudy and previous studies, (b) the Au atomic content with the distance from the sur-face of the Cr2N, CrN and ZrN coatings.

519Z.B. Qi et al. / Thin Solid Films 544 (2013) 515–520

Several studies proposed that the growth of Cr2O3 scales is con-trolled by outward diffusion of chromium, since the oxidation activa-tion energies of the Cr2N and CrN coatings are similar with those ofthe self-diffusion of chromium in Cr2O3 [7]. However, such suggestionis questionable due to the scattering results of oxidation activationenergies (from 116 to 225 kJ/mol as listed above) [7]. Therefore theAu marker experiments were performed on the Cr2N, CrN and ZrNcoatings to provide direct evidences of the oxidation mechanism.Here, we used the ZrN coating as a reference, since it is generally ac-cepted that the oxide scale growth of the ZrN coating is controlled byinward diffusion of oxygen [21,30]. The Au atomic content with thedistance from the surface is shown in Fig. 5(b). For the ZrN coating,the Au atomic content on the surface is ~40%, and the atomic contentsignificantly decreases with increasing depth from the surface. As thedepth higher than 20 nm, no Au is obtained. This indicates that theoxide scale growth of the ZrN coating is controlled by inward diffu-sion of oxygen. For the Cr2N and CrN coatings, the atomic content ofAu on the surface is ~5%, the atomic content of Au slightly decreasewith increasing depth from the surface. More than 1 at.% Au stillcan be obtained as the depth higher than 50 nm. Although the Au

marker remains on the surface after oxidation, the amount is muchless than the ZrN coatings. Furthermore, inward diffusion of Au intothe Cr2O3 scales is also obtained. As a result, the oxide scale growthof the Cr2N and CrN coatings is controlled by both the outward diffu-sion of chromium and inward diffusion of oxygen and the outwarddiffusion of chromium is dominated.

4. Conclusions

A comparative study of the oxidation behavior of Cr2N and CrNcoatings was investigated by using grazing incidence X-ray diffraction(GIXRD), Raman scattering, field emission scanning electron micros-copy (FESEM) and gold marker experiments. The GIXRD and Ramanreveal that the oxidation of the Cr2N coating produces two phases,Cr2O3 and CrN, while only Cr2O3 formed for the CrN coating, sincethe formation of Cr2O3 and CrN is thermodynamically favorable. Thediffusion barrier of Cr2O3 for N2 releasing also contributes to thephase transformation from Cr2N to CrN. During oxidation the surfacemorphologies change from as-deposited taper and triangular sheetstructure to Cr2O3 granular and faceted structure, accompanyingwith Cr2O3 grain growth at higher temperature. The Cr2O3 scales ex-hibit dense structure and the oxide thickness of the Cr2N coating isthicker than that of the CrN coating. The poorer oxidation resistanceof the Cr2N coating compared with the CrN coating is attributed toits higher chromium content and porous structure induced by phasetransformation accompanying with 20% volume expansion. The acti-vation energy values during oxidation are 139 and 190 kJ/mol forthe Cr2N and CrN coatings respectively, which is comparable to theliterature results. Marker experiments prove that the growth ofCr2O3 scales is controlled by both the outward diffusion of chromiumand inward diffusion of oxygen and the former is dominated.

Acknowledgment

The authors thank for the financial support from the NationalKey Technology R&D Program of China (2007BAE05B04), theScience and Technology Foundation of Fujian (2012H6019) andXiamen (3502Z20113003).

References

[1] H.C. Barshilia, N. Selvakumar, B. Deepthi, K. Rajam, Surf. Coat.Technol. 201 (2006)2193.

[2] R. Cecchini, A. Fabrizi, M. Cabibbo, C. Paternoster, B. Mavrin, V. Denisov, N.Novikova, M. Haīdopoulo, Thin Solid Films 519 (2011) 6515.

[3] E. Forniés, R. Escobar Galindo, O. Sánchez, J. Albella, Surf. Coat.Technol. 200(2006) 6047.

[4] H.N. Shah, R. Jayaganthan, D. Kaur, R. Chandra, Thin Solid Films 518 (2010) 5762.[5] T. Kacsich, K.-P. Lieb, Thin Solid Films 245 (1994) 4.[6] T. Kacsich, K. Lieb, A. Schaper, O. Schulte, J. Phys. Condens. Matter 8 (1996) 10703.[7] D. Lee, Y. Lee, S. Kwon, Surf. Coat.Technol. 141 (2001) 227.[8] P. Mayrhofer, H. Willmann, C. Mitterer, Surf. Coat.Technol. 146 (2001) 222.[9] P. Hones, C. Zakri, P.E. Schmid, F. Levy, O.R. Shojaei, Appl. Phys. Lett. 76 (22)

(2000) 3194.[10] Z.B. Qi, F.P. Zhu, Z.T. Wu, B. Liu, Z.C. Wang, D.L. Peng, C.H. Wu, Surf. Coat. Technol.

(2012), http://dx.doi.org/10.1016/j.surfcoat.2012.01.056.[11] Z.B. Qi, P. Sun, F.P. Zhu, Z.C. Wang, D.L. Peng, C.H. Wu, Surf. Coat.Technol. 205

(2011) 3692.[12] M.W. Chase Jr., NIST-JANAF, Thermochemical Tables, Journal of Physical and

Chemical Reference Data Monograph 9, 4th ed., American Chemical Society,Washington, DC, 1998.

[13] H.Y. Chen, F.H. Lu, J. Vac. Sci. Technol. A 21 (2003) 695.[14] F.H. Lu, H.Y. Chen, C.H. Hung, J. Vac. Sci. Technol. A 21 (2003) 671.[15] H.Y. Chen, F.H. Lu, Thin Solid Films 515 (2006) 2179.[16] J. Mougin, T. Le Bihan, G. Lucazeau, J. Phys. Chem. Solids 62 (2001) 553.[17] H.C. Barshilia, K. Rajam, J. Mater. Res. 19 (2004) 3196.[18] H. Warlimont, Springer Handbook of Condensed Matter and Materials Data,

Springer Verlag, 2005.[19] F.H. Lu, W.Z. Lo, J. Vac. Sci. Technol. A 22 (2004) 2071.[20] H.Y. Chen, F.H. Lu, J. Vac. Sci. Technol. A 23 (2005) 1006.[21] G. Reddy, J. Ramana, S. Kumar, S.V. Kumar, V. Raju, Appl. Surf. Sci. 253 (2007)

7230.[22] N. Birks, G.H. Meier, F.S. Pettit, Introduction to the High-temperature Oxidation of

Metals, Cambridge Univ Pr, 2006.

520 Z.B. Qi et al. / Thin Solid Films 544 (2013) 515–520

[23] S. Hofmann, H.A. Jehn, Mater. Corros. 41 (1990) 756.[24] E. Huber, S. Hofmann, Surf. Coat.Technol. 68 (1994) 64.[25] J. Lin, N. Zhang, W.D. Sproul, J.J. Moore, Surf. Coat.Technol. 206 (2012) 3283.[26] Y. Otani, S. Hofmann, Thin Solid Films 287 (1996) 188.[27] H. Ichimura, A. Kawana, J. Mater. Res. 9 (1994) 151.

[28] P. Panjan, B. Navinšek, A. Cvelbar, A. Zalar, I. Milošev, Thin Solid Films 281 (1996)298.

[29] W.P. Hsieh, C.C. Wang, C.H. Lin, F.S. Shieu, J. Electrochem. Soc. 149 (2002) B234.[30] L. Krusin-Elbaum, M. Wittmer, Thin Solid Films 107 (1983) 111.