Magnetron Sputtered NbN Films with Nb Interlayer onMild Steel

The aim of the study is to extend the NbN coating on MS with Nb interlayer to explore the benefits of hard nitride coatings on low-cost structural material and to compare the coating with NbN monolithic coating on SS. NbN on MS and SS was deposited by reactive d.c. magnetron sputtering at various N2/Ar flow ratios and substrate bias. Deposition rate decreased from 20 to 10 nm/min (without biasing) and from 16 to 8 nm/min (−50 V biasing) when FN2 /FAr ratio was varied from zero to 70%. Deposition rate decreased with the increase in bias voltage. Coatings showed hexagonal β Nb2N, cubic δ NbN, and hexagonal δ′ NbN as major phases with the increasing N2 flow. Surface hardness reached a maximum of 2040 HK25 at a FN2 /FAr of 20%. Critical loads, for cohesive and adhesive failure for coating on MS, were between 6–8 N and 9–12 N respectively; for coating on SS, the values were between 7–15 N and 12–25 N respectively. Duplex coatings were studied for hardness by Knoop microindentation, adhesion by scratch tester, and corrosion by potentiodynamic polarization technique. Hardness, adhesion, and corrosion resistance all improved when NbN coating was incorporated with Nb interlayer on MS.

Mild steel (MS) is widely used as a structural material due to its low cost.Properties of mild steel such as hardness, wear resistance, and corrosion resistance are not adequate.Application of hard nitride coatings on mild steel can modify the surface of mild steel necessary for practical applications.Hard nitride coatings deposited by physical vapor deposition (PVD) techniques exhibit high hardness, good wear resistance, and excellent chemical inertness.However, corrosion behavior of these coatings is often insufficient because of the presence of pin-hole porosity and microcracks inherent in PVD coatings.The corrosive media can attack the steel through these pores and microcracks.Various interlayers have been used to address the problem [29,30].NbN International Journal of Corrosion coatings with interlayer, however, have not been investigated much.Duplex coatings involving TiN top coat, deposited by sputtering, with interlayer of chromium (Cr) or nickel (Ni) deposited by electroplating and electroless nickel (EN) deposited by electroless technique have been studied to explore the benefits of hard nitride coatings on cheaper mild steel (MS) substrates [29][30][31][32][33].These duplex coatings were found to improve the hardness, corrosion resistance, and other properties when extended on to MS substrates.
The aim of the present study is to propose a prospective coating combination on mild steel for versatile application.
In the present study, NbN coatings have been extended on to MS substrate with niobium (Nb) interlayer deposited by sputtering and the results have been compared with the monolithic NbN coating on SS.For this NbN, coatings were deposited on MS and stainless steel (SS) substrates.Coatings were studied for their thickness, structure, hardness, and adhesion.Process parameters were optimized, and then, NbN coating with Nb interlayer was deposited on to MS substrate.These duplex coatings were studied for the improvement with respect to adhesion by scratch tests, surface hardness by Knoop microindentation, and corrosion performance by potentiodynamic polarization technique.Open circuit potentials were also measured.

Experimental Procedure
NbN films were deposited using reactive DC magnetron sputtering on SS-, MS-, and Nb-coated MS substrates.An Nb (99.99% purity) metallic target, 160 mm diameter and 4 mm thick, was mechanically clamped to a planar sputter source mounted horizontally on the base of the chamber evacuated to a base pressure of 2 × 10 −6 mbar.The distance between the target and the substrate was 60 mm.The sputtering pressure was kept at 5 × 10 −3 mbar by admitting a stream of mixed gas of argon (Ar) and nitrogen (N 2 ) into the chamber.Flow of Ar gas was fixed at 20 sccm, while N 2 flow was varied from 0-14 sccm.Substrate biasing was kept constant at −50 V for coatings deposited at different N 2 /Ar flow ratios.Coatings were deposited without externally heating the substrate.The power to the target was supplied through a stabilized d. c. power supply of 0-1000 V (6 Amperes maximum).Substrate biasing was varied from 0 to −150 V in a step of 25 V (keeping the N 2 /Ar flow ratio constant at 20%) by means of a stabilized d. c. power supply of variable voltage (0-300 V) and current (0-500 mA).The samples were polished, cleaned thoroughly, and degreased in alkaline solution prior to deposition.The flow sheet used for the cleaning of the substrate samples is presented in the Figure 1.All the cleaning steps were performed ultrasonically for the specified duration.The MS samples were also deposited with Nb interlayer and NbN as top coatings.The deposition parameters for Nb and NbN coatings were similar (Table 1) except that no N 2 was introduced for the Nb coating.Substrate biasing for Nb deposition was kept at −50V.Thickness of Nb interlayer was 2 μm.
Weight gain of the samples was recorded and thickness of the coatings was calculated using bulk density values.Actual coating thicknesses were studied by microabrasion using   Tests were performed in a linearly progressive mode from 1 N start load to a predefined maximum load.The maximum load was varied from 10 N to 60 N.The results displayed are for the maximum load of 40 N, since the coatings were found to delaminate beyond this load.Electrochemical evaluation of the coated samples was carried out using the standard potentiodynamic measurement technique with a computer controlled Santronic Electrochemical Analyzer.Tests were carried out using a threeelectrode cell.Coated samples were soldered (with indium) to a copper wire coated with enamel.The samples were masked by Shailmask 800 lacquer (proprietary) to get the 1 cm 2 surface area exposed.All potentials were measured with respect to a saturated calomel electrode (SCE).The auxiliary or counter electrode was platinum.The anodic and cathodic electrochemical polarization curves of all the samples were obtained in N 2 de-aerated 1N H 2 SO 4 electrolyte at room temperature.Open circuit potentials (OCPs) were measured in deaerated 1 N H 2 SO 4 solution for 2 hrs.Before potentiodynamic measurements, the samples were allowed to reach equilibrium potential (E ocp ).This potential was reached after 30-40 minutes, and the electrochemical measurements were started when the potential did not change by more than 1 mV/min.The solution was replaced after each sweep run.Polarization resistance was determined in the ±15 mV domain of E ocp potential using the linear polarization method at a scan rate of 0.1 mV/sec.For potentiodynamic studies, a potential sweep range of −1.000 V to +1.000 V was applied with a scan rate of 0.5 mV/sec.The plot of E measured against SCE v/s log I was plotted.The corrosion potential (E corr ) was determined from the intersection of Tafel slopes and the corrosion current density (I corr ) was calculated using the anodic and cathodic Tafel slopes (β a and β c ) and polarization resistance (R p ).

Thickness. A variation of 5%-10% in the thickness of the coatings was found between the calculated values (weight gain method) and the actual values (Calotest technique).
This was due to the density of the coatings being lower than the bulk values.In the deposition using biasing, there was a continuous ion bombardment at the substrate, which reduced the effective deposition rate.Therefore, more time was required to get the same coating thickness for coatings deposited at higher bias voltages.NbN coating thicknesses of 1.8 μm ± 10% was obtained for coatings deposited at various parameters.Thickness of Nb interlayer was 2 μm.Total coating thickness on the MS substrate with interlayer was about 4 μm.N 2 /Ar flow ratio.Deposition rate decreased from 20 nm/min to 10 nm/min (without biasing) and from 16 nm/min to 8 nm/min (with biasing at −50 V) with the increase in N 2 /Ar flow ratio from 0% to 70%.Deposition rate reduced almost linearly in both the cases with every increase in N 2 flow.The decrease in deposition rate, with every increase in nitrogen flow, was due to the increased nitridation of the target well known as the poisoning of the target surface.Further, it was seen that the reduction in deposition rate followed three different linear paths in both the cases-with biasing or without biasing.Transition in linearity in deposition rates was due to the transition in phases as revealed by X-ray diffraction (Section 3.3).The transition in deposition rates corresponded to the transition of phases from Nb to β-Nb 2 N, β-Nb 2 N to cubic δ-NbN, and cubic δ-NbN to hexagonal δ -NbN.

Effect of Substrate
Biasing. Figure 3 shows the deposition rate plotted against the substrate bias voltage keeping the N 2 /Ar flow ratio constant at 20%.It was observed that deposition rate decreased with the increase in substrate biasing (negative) voltage.The deposition rate decreased from 15.9 to 6.0 nm/min when the bias voltage was increased from zero to −150 V in a step of 25 V.The decrease in deposition rate with the increase in substrate biasing was due to the resputtering effect at the substrate.With every increase in substrate biasing (negative voltage) increased resputtering at the substrate with higher energy ions takes place.This causes removal of entrapped gas atoms, densification of the grain boundaries, and resputtering of depositing particles, resulting in fall of deposition rates.Increased ion bombardment improves adhesion, hardness, and density of the coatings.showed hexagonal β-Nb 2 N as the major phase with (101) preferred orientation.With the increase in N 2 /Ar flow ratio to 10%, the major phase became cubic δ-NbN with preferred orientation of (111).At 30% N 2 /Ar flow ratio hexagonal δ -NbN phase appeared though the major phase was still cubic δ-NbN but now with preferred orientation of (200).With the further increase in N 2 /Ar flow ratio, the hexagonal δ -NbN phase increased and became major phase at 70% N 2 flow.In all the coatings, substrate peaks were identified as the stronger peaks.

X-Ray
Figure 5 shows the X-ray diffraction patterns of Nb and NbN with Nb interlayer on MS substrate; NbN coatings were deposited at N 2 /Ar flow ratio of 20%.Nb (110) was found to be stronger than Nb (200) peak.The intensity of Nb peaks reduced when NbN top coat was given.MS sample showed the peak of Fe (110) in the 2θ range performed.NbN (111) peak was found to be stronger than NbN (200) or NbN (220) peaks.the softer substrate such as steel.Lichinchi et al. [34] have shown by means of finite element method (FEM) modeling that indentation into TiN on high-speed steel (HSS) causes plastic deformation in the substrate when the indentation depth reaches 15%.
To get the true hardness of the NbN coating, a coating thickness of about 4 μm was required.However, in the present study, the coating thickness was ≤2 μm; therefore, the hardness values obtained show the composite hardness of the substrate-coating combination and not the true hardness of the films.
Grain size also influences the hardness of the materials.Hardness changes as per the Hall-Petch relationship.Grain size has been found to increase with the increase in the partial pressure of nitrogen [35][36][37].The decrease in hardness at higher partial pressures of nitrogen could also be attributed partly due to the increase in grain sizes.
Roughness of the coated samples was found to replicate the values of the polished surface (0.06-0.08 μm), and no difference in the roughness of the samples was observed after the coatings.Therefore, roughness of the coating does not seem to play a role in the hardness effects observed in the present study.

Effect of Substrate
Biasing. Figure 7 shows the surface hardness of NbN coatings at 25 gf deposited at various substrate bias voltages, keeping the N 2 /Ar flow ratio constant at 20%.Hardness increased continuously with the increase in substrate bias voltage.Hardness increased from 1692 HK 25 for coatings deposited without biasing to 2346 HK 25 for coatings deposited at −150 V substrate biasing.Kim et al. [38] observed the increase in hardness similarly with the increase in substrate bias voltage up to −200 V.The hardness enhancement by energetic ion bombardment is due to a complex, synergistic effect involving a decrease of crystallite size, densification of the grain boundaries, formation of radiation damage and other point defects, and built-in biaxial compressive stress [39].The development of the morphology and microstructure of sputtered films bombarded by energetic ions during their growth is described by the Thornton diagram [40].It is well known that ion bombardment due to biasing (the substrate) generally causes a reduction in the grain size [40,41].The increased ion flux with higher kinetic energy provides increased ion bombardment on the growing film, which leads to a reduction in the grain size.Hardness increases with the refinement of grain structure.
Hardness is found correlated directly to the compressive stresses also.Researchers have reported similar enhancement of the hardness and compressive stress for a variety of hard coatings deposited by magnetron sputtering [42] and vacuum arc evaporation [43].

Duplex Coatings on MS.
The surface hardness of MS substrate, Nb, NbN, and duplex coating of NbN with Nb interlayer on MS are given in Table 2. Hardness values obtained show the composite hardness of the substratecoating combination.With the incorporation of Nb interlayer, the surface hardness of NbN coating increased from 1084 HK 25 to 1436 HK 25 due to the load support provided by the interlayer.

Scratch Adhesion Test.
Friction force and depth of indentation for all the scratch tests were recorded online along with the indenter movement to confirm the critical loads for cracks, chipping, delamination, coating failure, or other such observations.Tests performed at different loading rates were observed to give almost similar results, and variation in loading rate had little impact; therefore, loading rate for scratch tests was kept at 30 N/min.
Several types of observations were revealed as the scratch progressed, such as upper mono layers removal, pile-up on the sides, visibility of small cracks to long wide cracks within the coatings, pores, chipping, and partial or complete delamination of the coating.Figure 8 shows the scratch patterns for NbN coating on MS sample; at 9 N load, the scratch pattern showed the chipping and cracks at few places, while at 28 N load, the coating was delaminated almost completely.
Figure 9 exhibits the scratch patterns for NbN coating on SS taken at various loads.Start of the scratch is shown at 1 N load; at 12 N load, segregation, cracks within coating and pores were visible though coating was still intact; at 22 N load, along with cracks, pores, chipping, and delamination at few places occurred; at 32 N load, coating was found to delaminate at many places.The coatings showed much better  performance on SS samples than on MS samples.However, when NbN coated MS sample with Nb interlayer was scratchtested (Figure 10), it performed equally or even better than NbN coating on SS.Comparing the scratch patterns of NbN/SS (Figure 9) and NbN/Nb/MS (Figure 10) at the same loads reveal less damage for NbN coatings with Nb interlayer with respect to chipping, cracks, pores, and delamination.

Critical Loads.
Two critical loads, Lc 1 and Lc 2 , have been defined for the failure of the coatings.Lc 1 , the first critical load, corresponds to initial cohesive failure of the coating such as appearance of first cracks within the coating.Lc 2 , the second critical load, corresponds to initial adhesive failure of the coating, that is, first observation of adhesive failure such as chipping, partial delamination, pores, or some such phenomena, where substrate beneath coating gets exposed.Lc 1 and Lc 2 for coating on MS samples were observed to be between 6-8 N and 9-12 N loads, respectively.For coatings deposited on SS substrates Lc 1 varied between 7-15 N and Lc 2 between 12-25 N. NbN coatings on SS showed better performance than NbN coatings on MS.However, for Nb interlayered NbN coatings on MS, the critical loads improved; Lc 1 was found to be between 8-14 N and Lc 2 between 10-24 N, showing almost equal or sometimes slightly better performance than NbN coatings on SS.

Coefficient of Friction.
Coefficient of friction (μ), as observed in the scratch adhesion test, increased with the increase in scratch load.μ represents the value for coatingsubstrate combination.NbN coating, being hard, expected to show lower μ value, while MS substrate is expected to show higher μ value.The increase in the μ value with the increase in scratch load was due to the increasing effect from the substrates.For coatings deposited on SS samples, the value was found to vary within a narrow range of 0.22-0.25 at 30 N load irrespective of coatings deposited at different N 2 /Ar flow ratios.At 60 N, load the value increased to 0.40.Table 3 lists the μ value at different loads for the two types of substrates along with Nb interlayered NbN coating on MS substrate.NbN coating on MS samples showed higher coefficient of friction than on SS samples at the same applied scratch loads.However, with Nb interlayer, the μ decreased to 0.24, 0.30, and 0.38 at 30, 40, and 60 N loads, respectively, thus clearly demonstrating the improvement in μ value for Nb interlayered NbN coatings.

Depth of Penetration.
Depth of penetration increased with the increase in applied load.At 30 N load, on an average, MS samples had 20-30 μm depth of penetration, while SS samples had 12-25 μm depth of penetration.NbN coating on MS with Nb interlayer reduced the depth of penetration to 14-24 μm at the same applied loads.Depth of penetration includes elastic as well as plastic deformation of the coating substrate combination during loading.Besides, there could be error factors such as natural slopes of the samples (thickness variation in sample) and mounting errors.

Effect of Loading Rate.
The effect of loading rate was studied for NbN coatings on MS and SS substrates.The effect of loading rate on Lc 1 and Lc 2 was found to have little impact.The Lc 1 value was found to shift from 7 N at 10 N/min to 7.5 N at 30 N/min and further to 8.2 N at 50 N/min.Similarly, Lc 2 value shifted from 8.2 N to 9 N and further to 9.5 N with similar increase in applied loading rates.
For SS samples, the Lc 1 changed from 15 to 17 N and Lc 2 changed from 22 N to 24 N when the applied loading rate was increased successively from 20 N/min to 80 N/min.11 N, respectively.The values at various substrate biasing are shown in Table 4. Decreased adhesion of the films at higher bias is likely due to the greater stresses introduced into the film, as it is bombarded with more energetic ions [44,45].
3.5.6.Effect of N 2 Flow.Nb-N coatings on SS samples deposited at N 2 /Ar flow ratio ranging from 10% to 70%, (keeping the biasing voltage constant at −50 V) were tested for scratch adhesion.Results with respect to critical loads are shown in Table 5. Coatings deposited at 20% and 30% of N 2 /Ar flow ratio showed better adhesion with higher critical loads showing Lc 1 at 11-12 N and Lc 2 24-25 N. Lc 1 and Lc 2 values decreased successively to 7 N and 14 N with the further increase in N 2 /Ar flow to 70%.This could be attributed to more gaseous impurities, less dense structure, and the presence of pores in the coatings [40].

Corrosion Resistance.
It is difficult to deposit the hard coatings by physical vapor deposition (PVD) techniques without any micro porosity.Thus, when a PVD coated sample is exposed to the corrosive environment, the electrochemical behavior of the coated sample is the combined behavior of the coating and the substrate.The polarization curve of such a specimen may be considered as a combination of two curves-one representing the base material and the other the coating.The equilibrium value was found to be quite similar to MS substrate without coating, thus indicating the presence of pores in the coating, resulting in corrosion taking place beneath NbN coating.For NbN coatings on MS substrate with Nb interlayer, the OCP shifted to −154 mV in the beginning from −252 mV for coating without interlayer.The value decreased to about −443 mV after 40 min and remained constant thereafter.

Potentiodynamic Tests.
A high E corr and a low I corr values are indicative of good corrosion resistance.Table 6 lists the E corr and I corr values for MS substrate: NbN, Nb, and NbN coating with Nb interlayer on MS substrate.Plain MS substrate had E corr value of −496.4 mV and I corr value of 1440 μA/cm 2 .Coating the MS sample with NbN improved the corrosion resistance by decreasing the corrosion current (I corr ) to 150 μA/cm 2 and increasing the E corr (less negative) to −412 mV.However, the improvement was constrained due to the presence of pin-hole porosity inherent in PVD coatings.Due to the presence of pin-hole defects in the NbN coatings, rapid pitting corrosion of MS substrate takes place at these defects; even leading to partial debonding of the coating during the tests.The potentiodynamic curves for NbN coatings, therefore, mimic the behavior of MS substrate.
Nb interlayer was found to improve the corrosion resistance of NbN coated MS substrates effectively.For Nb interlayered NbN coating, E corr and I corr was found to be −396.2mV and 14.8 μA/cm 2 , respectively.With the duplex coating, the adherence also improved, because duplex coating did not delaminate for the full potential sweep of −1000 mV to +1000 mV during the potentiodynamic corrosion test.Potentiodnamic curves for substrate, NbN coating, and NbN coating with Nb interlayer on MS substrates are shown in Figure 12.Potentiodynamic curve for the duplex coatings showed a remarkable improvement in corrosion improvement for MS substrate.In general, it was observed that the duplex coating is much superior to NbN coatings alone.

Conclusions
NbN coatings were deposited on MS and SS substrates by reactive DC magnetron sputtering.F N2 /F Ar was varied from 0% to 70% and substrate biasing from zero to −150 V in a step of 25 V. Deposition rate decreased from 20 to 10 nm/min (without biasing) and from 16 to 8nm/min (with biasing at −50 V) with the increase in N 2 flow from 0% to 70%.Deposition rate decreased from 15.9 to 6 nm/min when the bias voltage was increased from zero to −150 V successively.Coatings showed presence of hexagonal β Nb 2 N, cubic δ NbN, and hexagonal δ NbN as the major phases with increasing N 2 flow.
Surface hardness on SS reached a maximum of 2040 HK 25 at a N 2 /Ar flow ratio of 20% and then decreased slowly with the further increase in N 2 /Ar flow ratio.Surface hardness increased from 1084 HK 25 to 1618 HK 25 when incorporated with Nb interlayer.Critical loads for cohesive (Lc 1 ) and adhesive (Lc 2 ) failures during scratch test for coatings on MS samples were observed to be between 6-8 N and 9-12 N loads, respectively; while for coatings on SS substrates, the values were between 7-15 N and 12-25 N, respectively.For duplex coating on MS, the values were between 8-14 and 10-24 N, respectively.Coefficient of friction (μ) during scratch test was 0.22-0.25 at 30 N, 0.30, at 40 N and 0.40 at 60 N loads for coatings deposited on SS samples.For coatings on MS sample, the μ was 0.28 at 30 N, 0.35 at 40 N, and 0.45 at 60 N loads.For NbN coating on MS with Nb interlayer, the μ was 0.24, 0.30 and 0.38 at 30, 40, and 60 N loads, respectively, demonstrating the improvement in μ value.Increase in loading rate had little impact on critical loads.Increase in substrate biasing up to −75 V improved the Lc 1 and Lc 2 to 12 and 25 N, respectively; further increase in substrate biasing caused the critical loads to decrease to 7 and 14 N, respectively.Coatings deposited at 20% and 30% N 2 /Ar flow ratio showed better adhesion with higher critical loads showing Lc 1 11-12 N and Lc 2 24-25 N.
Open circuit potential for NbN coated MS sample decreased from an initial value of −252 mV to −478 mV in 30 min and to −498 mV in 40 min after which it stabilized.The equilibrium value was found to be quite similar to MS substrate, thus indicating the presence of pores in the coating, resulting in corrosion taking place beneath coating.For duplex coating, OCP shifted to −154 mV in the beginning and stabilized to −443 mV after 40 min.With duplex coating the corrosion resistance increased remarkably.I corr decreased from 150.2 μA/cm 2 to 14.8 μA/cm 2 .

FFigure 5 :
Figure 5: X-ray diffraction patterns of Nb/MS and NbN/Nb/MS (Nb and NbN coatings were deposited in pure Ar and F N2 /F Ar = 20%, respectively; biasing was −50 V for both).

Figure 8 :
Figure 8: Scratch test for NbN coating (deposited at N 2 /Ar flow = 20%) on MS at (a) 9 N load, showing chipping and cracks and (b) 28 N load, showing nearly complete delamination.

3. 5 . 5 .
Effect of Biasing.Effect of biasing was studied on NbN coated SS samples.Increase in biasing voltage from zero to −75 V (in a step of 25 V), keeping other factors constant, led to successive increase in the values of Lc 1 and Lc 2 .At −75 V, biasing the Lc 1 and Lc 2 values was found to be 15.6 N and 26 N, respectively.However, at −100 V, the coating became brittle, and Lc 1 and Lc 2 values dropped drastically to 7 and

Figure 11
shows the changes in open circuit potential (OCP) with immersion time for NbN and NbN with Nb interlayer coated on MS sample.For NbN coating, potential decreased from an initial value of −252 to −478 mV in 30 min and to −498 mV in 40 min after which it remained nearly constant indicating the establishment of equilibrium between metal and the solution.

Figure 11 :
Figure 11: E OCP versus time for NbN/MS with and without Nb interlayer in 1 N H 2 SO 4 .

Table 1 :
Deposition parameters for Nb and NbN coatings.

Table 2 :
Surface hardness of MS, and Nb, NbN, and duplex coatings on MS.

Table 4 :
Effect of biasing on critical loads during scratch tests for NbN coatings on SS.

Table 5 :
Effect of N 2 flow on critical loads during scratch tests for NbN coatings on SS.

Table 6 :
E corr and I corr values for single and duplex coatings on MS.
Coatings were characterized for their thickness by weight gain and Calotest technique, hardness by Knoop microhardness tester, phase analysis by X-ray diffraction technique, and adhesion by scratch adhesion tester.The effect of N 2 flow and substrate biasing was evaluated.After the optimization of process parameters, NbN coatings were deposited on MS substrate with 2 μm thick Nb interlayer.Effect of interlayer on NbN coating on MS substrate was studied for the improvements in surface hardness by Knoop microindentation, adhesion by scratch test, and corrosion by potentiodynamic polarization technique in 1 N H 2 SO 4 solution at room temperature.Open circuit potentials were also measured.The following conclusions were drawn.