Microstructure and adhesion characteristics of duplex coatings with different plasma‐nitrided layers and a Cr‐Al‐Ti‐B‐N physical vapor deposition coating

Abstract

A duplex treatment consisting of plasma nitriding (PN) and physical vapor deposition (PVD) significantly improves the thermal, tribological, and corrosion resistance of forming tools, and especially if they are intended for applications subject to high mechanical loads. This study investigates the influence of nitriding on the properties of a conventionally heat‐treated AISI D2 tool steel, coated with a Cr‐Al‐Ti‐B‐N layer, while the effect of the presence or absence of a compound layer is discussed. PN is performed at 510–520°C using different N₂‐H₂ gas mixtures. The Cr‐Al‐Ti‐B‐N layers are deposited at 480°C using a combination of cathodic arc evaporation and magnetron sputtering. The samples are characterized using electron probe microanalysis with wavelength‐dispersive X‐ray spectroscopy, optical and scanning electron microscopy, glancing‐angle X‐ray diffraction, surface hardness measurements, profilometry, Rockwell indentation, and scratch tests. These techniques reveal relationships between the depth gradient of the chemical composition and microstructure of the nitrided interlayer, the adhesion of the PVD coating, and the hardness of the tool steel. Although the PVD process induces a structural transformation in the compound layer, this transition does not have a negative influence on the adhesion of the PVD coating.

INTRODUCTION

Duplex surface engineering involves the application of two sequential surface treatment technologies for the generation of surface composites with combined and improved properties that cannot be achieved by the respective individual processes. An essential criterion for successful duplex treatment is the compatibility of the individual techniques with respect to the metallurgical properties of the substrate used. In recent decades, the combination of plasma nitriding (PN) with physical vapor deposition (PVD) has been established as an industrial process particularly suitable for tailoring the physical properties of high-alloyed forming tool steels.^1-6

PVD coating systems are regularly applied as protective coatings for forming tools in order to influence the tool's ability to withstand the tribological conditions and frictional forces they are subjected to, and thus play a crucial role in tool service life.⁷ Both (Ti,Al)N-^8-11 and (Cr,Al)N-based coatings^11-14 can be applied to reduce the level of friction between the tool and the workpiece and, therefore, to protect the sliding contact areas from wear and adhesive interaction.¹⁵ Nitride-based hard protective coatings are established industrial technologies for cutting and forming applications due to their high oxidation resistance, thermal stability, and high hardness, with CrN-based coatings considered particularly suitable for sheet and bulk metal forming.^16-19 The boron doping of (Cr,Al)N coatings can further improve the coating properties through the formation of nanocomposites, which enhance hardness.^20,21 In addition, a reduction in the coefficient of friction can be achieved.²¹ In order to fully exploit the potential of hard coatings, their support and adhesion to steel substrates must be ensured, for example, by a thermochemically generated nitrided layer.^22-24

Nitrided layers can have different morphologies. They can appear in the form of a diffusion layer, in which nitrogen is interstitially dissolved and nitrides are precipitated, or can exhibit an additional compound layer formed of iron nitrides such as γ′-Fe₄N (space group (SG): $\mathrm{Pm}\bar{3}\mathrm{m}$) and/or ε-Fe_2-3N (SG: P6₃22) on top of the nitrogen diffusion zone. A specific modification of the nitrided surface—such as the presence or absence of a compound layer and its chemical composition—influences the adhesion of the hard coating and, thus, determines the functionality of the duplex system.^2,22,25-28 However, there is a range of information available in the literature regarding the creation of a coating-compatible pre-nitrided layer structure, particularly with regard to the adhesion of the hard coating and to the compatibility of the compound layer with the coating process.

In the past, compound layers were generally considered to be unfavorable for the adhesion strength of duplex coatings, which was mainly due to the decomposition of the hard compound layer into α-Fe and nitrogen.^29-31 Sun and Bell²⁹ report that the formation of a soft “black layer” on a low-alloy steel is the result of the high temperatures applied during the PVD process. In addition to the thermal sequence of the PVD process, the ion bombardment of the nitrided steel substrates performed prior to the deposition is identified as a factor contributing to the formation of this black layer.^30-32 However, while the influence of ion bombardment is still frequently reported, consistent statements are not available in the literature due to the high variety of process parameters applied.

Nevertheless, even without the transformation of the compound layer into a black layer, the negative influence of the compound layer on the adhesion of the hard coating is reported by D'Haen et al.³³ For a two-phase compound layer generated on AISI D2 steel, the weak bonding between ε-Fe_2-3N and γ′-Fe₄N within the compound layer and the difference between the coefficients of thermal expansion of the two nitride phases are considered to be responsible for their low adhesion strength.³³ Investigations into duplex treatments on H13 hot-worked tool steel attribute the reduction in the adherence of a TiN coating deposited onto a compound layer to premature brittle fracture between the hard coating and the substrate.³⁴

Contrary to the above findings, positive effects of a compound layer on the resulting duplex layer properties are also found in the literature. Podgornik et al³⁵ report improved hard coating adhesion and increased wear resistance with a dense, 5 μm-thick γ′-compound layer in comparison to the compound layer-free conditions. A black layer did not form in the course of their investigation, however, details of the coating process are not specified in the study. Scratch and impact tests carried out on duplex systems consisting of a PVD TiN coating on nitrided H13 hot-worked steel substrates both without a compound layer and with an approx. 1 μm-thick compound layer revealed increased impact resistance without any loss of adhesion strength for the duplex systems with the compound layer.³⁶ This behavior is attributed to the improved support of the PVD coating by the compound layer which, moreover, acted as a barrier between the harder, more brittle PVD layer and the softer, more ductile diffusion layer of the nitrided substrate. Recent studies report on the beneficial influence of compound layers on the adhesion strength of the coating to the substrate, and further point out the importance of the strengthening effect of the compound layer. For example, γ′-Fe₄N provides heterogeneous nucleation sites for epitaxial growth for nitrides, for example, CrN, thus enhancing the adhesive strength.^37,38 Even if the above-mentioned investigations exhibit differences with regard to the steel substrates and PVD coating systems, it can be assumed that the presence of a compound layer in duplex systems has a positive influence on the performance of the duplex coatings.

For cold-worked tool steels in particular, however, apart from the investigation of the influence of a white layer on a post-oxidizing process,²⁴ systematic investigations regarding the behavior of a compound layer or a compound layer converted to a black layer in a duplex-layer composite are not found in the literature. This contribution analyzes the influence of a nitrided layer, which was produced either with or without a compound layer, on the adhesion of a Cr-Al-Ti-B-N hard coating on AISI D2 cold-worked steel. The adhesion of the hard coatings to the compound layer was compared with the adhesion of the hard coatings deposited on nitrided substrates after the compound layer was mechanically removed. The effect of the PVD coating process on the thermal stability of the nitrided layers both with and without a compound layer is investigated.

EXPERIMENTAL METHODS Material and preparation

As substrates, round samples made from a standard X155CrVMo12-1 tool steel (AISI D2) with the chemical composition given in Table 1 were used. Their diameter was 30 mm, and their thickness 7 mm. The plane-parallel samples were hardened and subsequently tempered in three steps in order to achieve a substrate hardness of approx. 62 ± 1 HRC (hardness Rockwell-C). Before PN, the samples were polished and cleaned in ethanol.

TABLE 1 Chemical composition of X155CrVMo12-1 according to the manufacturer

Element	Cr	C	Mo	V	Mn	Si	Fe
Composition (wt%)	11.9	1.5	0.8	0.8	0.3	0.5	Balance

Plasma nitriding

The plasma treatments of the polished steel substrates were carried out in a commercial direct current glow discharge facility under different conditions, as described in Table 2, in order to generate nitrided layers without (PN0) and with a compound layer (PN1). In this respect, the gas composition was changed in order to produce nitrided layers with nitrogen diffusion depths in the range of 30–50 μm. Subsequently, one sample with a compound layer was mechanically polished to remove the compound layer in a defined manner. Such a procedure is typically used in industrial processes if a compound layer has formed during nitriding. In this study, the uppermost 8 μm of the initial nitrided layer was removed, and the corresponding sample was labeled as PN1-8.

TABLE 2 Plasma nitriding conditions, post-processing of the sample surfaces and substrate hardness

Sample	Nitrogen contenta (%)	Nitriding time (h)	Temperature (°C)	Nitrogen diffusion depth (μm)	Compound layer thickness (μm)	Post-processing	Core hardness (HRC)
PN0	25	4	520	30	0	None	57 ± 1
PN1	50	2	510	50	4 μm	None	61 ± 1
PN1-8	50	2	510	50	4 μm	8 μm surface removal	61 ± 1

a The balance of the atmosphere consisted of hydrogen. Coating deposition

In a second step, heat-treated (BM) and plasma nitrided substrates (PN0, PN1 and PN1-8) as well as cemented carbide substrates consisting of WC (tungsten carbide) and 6 wt% Co were coated with a PVD Cr-Al-Ti-B-N coating. The deposition was carried out in an industrial-size π411 PVD facility from PLATIT, which facilitated the deposition of protective coatings by a hybrid PVD technique that combined cathodic arc evaporation from three outer arc sources and magnetron sputtering from a central sputter cathode. During the deposition of the Cr-Al-Ti-B-N, Cr and Al were arc evaporated from the respective metal cathode, while the central rotation cathode was used for the magnetron sputtering of TiB₂.

The substrates were cleaned with ethanol before loading them into the deposition chamber. Prior to the deposition, the chamber was evacuated to a base pressure of 7.4·10⁻³ Pa and heated to 480°C. The substrates were then etched by argon glow discharge and by Cr metal ions in the deposition facility. In order to improve the adhesion of the Cr-Al-Ti-B-N coating to the substrate, a CrN adhesion layer was applied. This in turn was followed by a Cr-Al-N gradient layer and by the main Cr-Al-Ti-B-N coating. The deposition chamber was maintained at 480°C during the entire PVD process. The main coating was prepared in an argon/nitrogen atmosphere at a working pressure of 3·10⁻² mbar and at a bias voltage of −70 V. During deposition, a 2-fold rotation was applied. The deposition time for the main coating was 210 min, resulting in a total coating thickness of approximately 4.3 μm. Hereinafter, the coating systems are denoted as BM-C, PN0-C, PN1-C, and PN1-8-C according to the notation of the substrates.

Microstructural analysis

The methods of microstructural analysis included optical microscopy (OM), electron probe microanalysis with wavelength-dispersive X-ray spectroscopy (EPMA/WDS), and glancing-angle X-ray diffraction (GAXRD). For the visualization of the thickness of the PVD coating and the nital-etched (5%) nitrided layers, OM was carried out using a Carl Zeiss Neophot 30 optical light microscope equipped with a JVC TK C1381CCD camera. EPMA/WDS revealed the depth-concentration profiles of selected elements. The JXA8230 SuperProbe (Jeol) electron probe microanalyzer was equipped with a tungsten hairpin-type cathode. The measurements were performed at an acceleration voltage of 12 kV and at a beam current of 40 nA using a spot size = 1 μm. The nitrogen diffusion profiles of the PN0 and PN1 samples were measured along line scans on polished cross-sections perpendicular to the sample surface. The length of each line scan was 100 μm and the step size was 1 μm. The composition of the Cr-Al-Ti-B-N coating was measured at 40 points on the surface of the sample deposited onto the cemented carbide substrate. The quantification comprised the elements present in the coating and in the substrate. Si, Al, Cr, Fe, V, Ti, Mo, and W were quantified using pure element standards. Compound standards such as SiC, AlN, and B₄C were used for the quantification of C, N, and B, respectively. The measured intensities of the spectral lines were corrected for atomic number, absorption, and fluorescence (ZAF correction).

GAXRD experiments were performed using a D8 Advance diffractometer (Bruker AXS) equipped with a sealed X-ray tube that incorporated a copper anode (λ = 0.15418 nm) and a parabolic Goebel mirror, which produces a parallel primary beam. The diffracted beam passed a Soller collimator with an acceptance angle of 0.12°, which defines the angular resolution of the diffractometer. A flat LiF monochromator was mounted between the Soller collimator and the scintillation detector. Behind the monochromator, the Cu K_α2/K_α1 intensity ratio was 0.1. The GAXRD experiments were performed with an angle of incidence of the primary beam of 10°. This incidence angle reduced the penetration depth of the Cu K_α radiation in the uncoated samples below 0.7 μm (as calculated for α-Fe with the linear absorption coefficient of 2378 cm⁻¹³⁹) or below 0.8 μm (as calculated for Fe₃N and Fe₄N with the linear absorption coefficients 2019 cm⁻¹ and 2060 cm⁻¹, respectively). The penetration depth is defined as the distance from the sample surface that delivers, after absorption, an intensity that is equal to ∼37% of the non-absorbed intensity.⁴⁰ In the Cr-Al-Ti-B-N coating with the linear absorption coefficient of 495 cm⁻¹, the penetration depth of the X-rays was below 3 μm at an angle of incidence of 10°. In order to reduce the penetration depth of the Cu K_α radiation in the duplex coatings, GAXRD was also performed at an incidence angle of 3°. The GAXRD patterns were collected in the 2θ range between 20° and 150°.

Surface characterization and mechanical testing

Profilometric measurements were carried out to determine the arithmetic mean roughness R_a and the roughness depth R_z in all samples. These measurements were performed using a monochromatic interferometer from Breitmeier Messtechnik GmbH, which was equipped with a BOP 140 sensor. Three line scans with a length of 4.8 mm were carried out on each sample, using a scanning rate of 250 points per mm and a frequency of 50 Hz, which resulted in a measuring speed of 0.2 mm·s⁻¹.

Hardness-depth profiles of the plasma nitrided samples in the as-nitrided as well as in the PVD-coated state were measured according to the Vickers method HV_0.1 on polished cross sections using a LECO M-400-G3 hardness testing apparatus. In addition, the surface hardness was measured according to the Vickers method using loads between 500 g and 10 kg (HV_0.5, HV₁, HV₁₀) in order to determine the load-bearing capacity of the substrate. The core hardness of the steel substrates subjected to different treatments was measured by Rockwell hardness testing, HRC. Additionally, the hardness of the Cr-Al-Ti-B-N coatings was determined by instrumented nanoindentation experiments using an NHT² nanohardness tester from Anton Paar. The experiments were performed with a Berkovich indenter using a load of 30 mN in order to reduce the penetration depth below 10% of the total coating thickness. The indentation hardness H_IT was determined according to the method of Oliver and Pharr⁴¹ from at least 20 indents. The indentations were carried out on polished sites of the coating surface.

Rockwell-C indentation and scratch testing

To evaluate the adhesion of the Cr-Al-Ti-B-N coating on the pretreated steel substrates, Rockwell (HRC) indentation testing was carried out according to the ISO 26443 standard, which facilitated the assignment of an adhesion class to each duplex-coating state.⁴² Subsequently, scratch tests were performed on a CSEM Revetest RST device to obtain more detailed information about the adhesion behavior. For the scratch test according to DIN EN ISO 20502 using a continuously increasing load, a scratch length of 2 mm and a scratch speed of 4.04 mm min⁻¹ were applied. During the tests, the acoustic emission signal and the tangential force were recorded. The critical load L_C was defined as the tangential force at which layer chipping, as registered by the acoustic emission signal, occurs for the first time. In addition, scanning electron microscopy (SEM) analysis of the scratch tracks was performed in a Mira Tescan 3 XMU at an acceleration voltage of 15 kV in order to assess the characteristic layer damage in the region of L_C.

RESULTS AND DISCUSSION Studies on the nitrided layer prepared by PN

Figure 1 shows optical images of the cross sections of the as-nitrided samples in the compound layer-free state (PN0, Figure 1(A)) and with the compound layer (PN1, Figure 1(C)). The areas that appear dark after nital etching indicate the diffusion depth of nitrogen within the respective substrate. EPMA/WDS performed before the nital etching on sample regions similar to those shown in the optical micrographs in Figure 1 revealed a nitrogen diffusion depth of 30 μm and a maximum nitrogen content at the surface of about 8 at% (Figure 1(B), left) for PN0. In PN1, the nitrided layer separated into a compound layer (white layer) located at the substrate surface and an underlying diffusion layer of approximately 50 μm in thickness (Figure 1(C)). For this sample state, the maximum nitrogen concentration at the surface reached 26 at%, which facilitated the formation of a compound layer with a thickness of approx. 4 μm as measured by OM. Furthermore, larger primary carbides typical for this tool steel⁴³ as well as small, finely distributed, globular secondary carbides were visible in the optical micrographs (Figure 1) and in the element concentration profiles measured by EPMA/WDS.

The formation of a compound layer in the PN1 sample was confirmed by GAXRD (Figure 2), which revealed that the uppermost part of the nitrided substrate consisted of ε-Fe₃N (SG: P6₃22), γ′-Fe₄N (SG: $\mathrm{Pm}\bar{3}\mathrm{m}$), and fcc-CrN (SG: $\mathrm{Fm}\bar{3}\mathrm{m}$). The CrN-based precipitates were also detected by the EPMA/WDS line scan in the diffusion zone (see Figure 1(B)) indicating a transformation of carbides due to nitriding treatment. In contrast to PN1, the BM non-nitrided sample contained only Cr-based carbides, which were mainly of the M₇C₃ type (M = Cr, Fe) (SG: Pnma), α-Fe (SG: $\mathrm{Im}\bar{3}\mathrm{m}$), and traces of Fe₃C (SG: Pnma).^43,44 The presence of Cr-based carbides in the steel substrate was also indicated by local EPMA/WDS line scans in the PN0 and PN1 samples (see Figure 1(B)), where Fe-Cr-C carbides were detected below the nitrogen diffusion zone, for example at ∼68 μm for PN0 and at 52 μm and 76 μm for PN1.

GAXRD of the PN1-8 sample confirmed that the compound layer was removed by polishing, as diffraction lines of ε-Fe₃N and γ′-Fe₄N were not present in the corresponding diffraction pattern. However, the presence of fcc-CrN was still visible in the diffraction pattern of PN1-8, because this phase also formed below the compound layer. Additional diffraction lines in the diffraction pattern of PN1-8 arose from Fe₃C and Cr carbides, which are typical for this type of tool steel. GAXRD of the PN0 sample verified the absence of a compound layer. However, the nitriding conditions of the PN0 sample led to the formation of fcc-CrN due to the reaction between nitrogen and chromium from the ferritic steel. The interstitial solution of nitrogen in Fe was indicated by the peak shift of the diffraction line 110 of α-Fe (marked by the red line below the diffraction patterns in Figure 2) to lower diffraction angles.

Properties of duplex coatings

When the roughness values R_a and R_z of the untreated substrate and the substrate after nitriding were compared with the roughness of the PVD-coated samples, clear differences were evident (Figure 3). As reported by other authors,^29,35 the nitrogen uptake significantly increased the surface roughness of the originally polished substrates (Figure 3(A)). The arithmetic mean roughness value R_a increased by a factor of between 8 and 11, while the roughness depth R_z increased by a factor of 3.5 to 4. Thus, the nitrided layer with a compound layer exhibited higher roughness values in each case. However, the PVD coating of the as-nitrided samples PN0 and PN1 caused a significant reduction in the roughness values R_a and R_z, that is, by a factor of two (Figure 3(B)). For the non-nitrided base material BM-C, an increase in roughness was detected after PVD coating. Sample PN-1-8-C, on which the compound layer was removed prior to PVD coating, exhibited roughness values comparable to those of the coated substrate BM-C.

Figure 4 presents OM images of the cross-sections of the Cr-Al-Ti-B-N coated substrates, and shows all investigated coated or duplex layer systems. The thickness of the hard coating was approx. 4.3 μm, and was comparable for all substrates. The dark etched areas in the OM micrographs (Figure 4(B)-(D)) indicated the diffusion zones (DZ) of the pre-nitrided substrates. For the nitrided sample with a compound layer (Figure 4(C)), a thin black layer (BL) appeared directly beneath the Cr-Al-Ti-B-N coating. For the sample state in which the compound layer was removed prior to the deposition of the PVD coating (Figure 4(D)), a black layer was not visible under the hard coating. The analysis of the formation of the black layer is discussed in detail below.

According to EPMA/WDX, the PVD coating consisted of 15.1 ± 0.5 at% Cr, 30.6 ± 1.1 at% Al, 1.5 ± 0.1 at% Ti, 1.7 ± 0.4 at% B, and 51.1 ± 1.8 at% N. GAXRD performed with an incidence angle of 3° revealed that the Cr-Al-Ti-B-N coating was a nanocomposite coating composed of fcc-(Cr,Al)N (SG: $\mathrm{Fm}\bar{3}\mathrm{m}$) and wurtzitic (w) AlN (SG: P6₃mc), cf. Figure 5(A). The crystallites of both phases were of sizes that measured below 10 nm. The Rietveld analysis yielded crystalline phase fractions of ∼60 vol% fcc-(Cr,Al)N and ∼40 vol% w-AlN. It was assumed that boron was primarily accommodated in an amorphous phase of BN.²⁰ The analysis of the positions of the GAXRD lines from fcc-(Cr,Al)N using the routine described by Rafaja et al⁴⁵ revealed a stress-free lattice parameter of (0.4125 ± 0.0003) nm and a compressive residual stress of (−4.9 ± 0.5) GPa assuming a rotationally symmetric biaxial stress state, a Young's modulus of 425 GPa and a Poisson's ratio of 0.22.⁴⁶ Due to the nanocomposite character of the Cr-Al-Ti-B-N coating, a high coating hardness was expected—as verified by instrumented nanoindentation experiments. The indentation hardness (H_IT) of the Cr-Al-Ti-B-N coating deposited onto the cemented carbide substrate was 32.2 ± 1.2 GPa and the indentation modulus E_IT was 385 ± 9 GPa. Similar hardness values of the PVD coating were obtained for the samples BM-C, PN0-C, PN1-C and PN1-8-C, see Table 3.

TABLE 3 Duplex sample states and resulting layer properties

		Core hardness (HV10)
Sample #	PVD coating hardness HIT (GPa)	Before PVD	After PVD	Surface hardness (HV10)	Black layer thickness (μm)
BM-C	32.0 ± 1.3	720 ± 10	725 ± 10	750 ± 10	-
PN0-C	34.9 ± 1.6	685 ± 10	695 ± 10	930 ± 30	-
PN1-C	31.8 ± 1.1	720 ± 10	715 ± 10	1120 ± 20	2–4
PN1-8-C	Not measured	720 ± 10	715 ± 10	1020 ± 30	-

The substrate hardness (core hardness) of the duplex samples measured on their cross sections as well as their surface hardnesses (HV₁₀) are summarized in Table 3. The surface hardness corresponds to a composite hardness that is determined by the hard coating, nitrided layers and the substrate. During PVD, the nitrided substrates were exposed to elevated temperatures. The time–temperature cycle of the PVD treatment did not influence the core hardness values of the steel samples, as they exhibited similar values before and after the coating process. In contrast, the surface hardness values HV₁₀ showed clear differences for the four sample states.

Furthermore, the depth profiles of the hardness values measured perpendicular to the surface revealed different hardness gradients beneath the surface in the as-nitrided and in the duplex samples (Figure 6). The comparison of the nitrided states before and after the PVD process shows clearly that there was no decrease in hardness within the diffusion layer. Thus, the diffusion layer in the nitrided sample did not experience a change in hardness during the PVD process at 480°C. Furthermore, it was proven that due to the different degrees of nitrogen incorporation into the diffusion layer, the values of the resulting case hardness differed significantly. For those samples without the compound layer (PN0/PN0-C), which contained approx. 4 at% N at a distance of 20 μm from the surface (cf. Figure 2(B)), hardness values of 1100 HV_0.1 were measured at a depth of 20 μm (Figure 6, gray lines). In contrast, the pre-nitrided samples with a compound layer (PN1, PN1-C) attained a maximum case hardness of 1350 HV_0.1 due to their higher nitrogen content of ∼10 at% at the same distance from the surface (cf. Figure 1(B)). For sample PN1-8-C, the displacement of the hardness-depth curve to the left (Figure 6, red line) corresponded to a value of ∼8 μm, which resulted from the mechanical removal of the compound layer prior to the deposition of the PVD Cr-Al-Ti-B-N coating.

It was obvious that the measurement of the hardness HV_0.1 carried out on the cross sections of the samples could not be accomplished immediately below the PVD coating (cf. Figure 6) because of the large diagonal length of the hardness indentation that resulted from the test load used. In order to obtain the hardness of the uppermost part of the nitrided substrate, which is a measure of the load-bearing capacity of the substrate beneath the PVD hard coating, the surface hardness was measured instead. The surface hardness measurements were performed with loads of between 500 g and 10 kg. Due to the application of different loads, individual contributions to the total measured hardness including the hardness of the Cr-Al-Ti-B-N coating, the hardness of the uppermost part of the nitrided substrate, the hardness of the nitrogen diffusion zone, and the hardness of the underlying substrate were weighted in different ratios. At the lowest load of 500 g, the indentation depth was approx. 5–6 μm (Figure 7) for all samples. Thus, the Cr-Al-Ti-B-N coating and the top part of the non-nitrided or nitrided substrate were the main layers probed (cf. Figure 7). When measuring HV₁₀, the indentation depth increased to approx. 30 μm, and the measured hardness value was dominated by the hardness of the nitrided substrate.

Independent of the compound layer formation, PN caused an increase in the compound hardness in comparison to the hardness of the non-nitrided substrate. All test loads (Figure 7) verified this general statement. In addition, the differences in the hardness values of the pre-nitrided substrates were relatively small. Still, the ‘near-surface’ compound hardness value HV_0.5 measured in the PN1-C duplex system, in which a compound layer was present in the as-nitrided state, was lower than the equivalent hardness value of the PN1-8-C duplex system in which the compound layer was completely removed before PVD coating. Since the two samples differed only in terms of the presence or absence of the compound layer, the difference in their hardness values could only have resulted from changes within the nitrided substrate area directly below the PVD coating, as shown in Figures 4(C),(D) and 5(B).

It was already mentioned above that the PVD process performed at elevated substrate temperatures implies an additional thermal treatment of the nitrided substrates. Although this thermal cycle did not have any effect on the substrate hardness, there was a significant influence on the structure of the former compound layer, which is reported in earlier work for low-alloy steels.^29-31 In the cross-sectional micrograph of the as-nitrided substrate, the compound layer was clearly visible as a white layer (Figure 5(B)). In the PVD-coated sample PN1-C, this zone appeared as a black layer (Figures 4(C) and 5(C)). GAXRD performed with two different incidence angles (3° and 10°) and, thus, at two different penetration depths of the X-rays (0.9 and 2.6 μm) (Figure 4(A)) revealed that the black layer contained α-Fe at the interface between the nitrided substrate and the Cr-Al-Ti-B-N coating. The diffraction pattern of the uncoated sample PN1 (gray pattern in Figure 5(A)) contained strong diffraction lines from ε-Fe₃N, γ′-Fe₄N and fcc-CrN. In the diffraction patterns of the PVD-coated sample PN1-C, the strongest diffraction lines stemmed from fcc-(Cr,Al)N. The diffraction line 110 from α-Fe became the most intense diffraction line from the substrate, while the diffraction lines from ε-Fe₃N and γ′-Fe₄N disappeared. It is worth noting that the intensity of the diffraction line from α-Fe was higher at the incidence angle of 10° than at the incidence angle of 3°, because at 3° the X-rays were strongly absorbed in and diffracted by the Cr-Al-Ti-B-N coating. Thus, the diffraction signal from the substrate was weak at an incidence angle of 3°. The formation of a black layer was in accordance with the literature,²⁹ and could be explained by the decomposition of the iron nitrides from the compound layer and by the formation of α-Fe. In the investigated coatings, the formation of the black layer in PN1-C was responsible for the reduction of the compound hardness value HV_0.5 of the PN1-C duplex system in comparison to the compound hardness value of the PN1-8-C duplex system (see Figure 7). Dingremont et al³⁰ provide two different temperature values that influence the compound layer structure during the PVD process. First, a temperature of 470°C is defined as a maximum temperature of pre-nitrided substrates, up to which the formation of a black layer does not take place. However, it is pointed out that this temperature, which is determined for a nitrided low-alloy steel, is material-specific and, thus, somewhat variable. Second, Dingremont et al state, that the temperature of an applied Ar ion etching step during the PVD process should not exceed 350°C in order to avoid the decomposition of the white layer (i.e., the compound layer). The formation of the black layer that was proven by GAXRD in the present study could be attributed to the thermal impact of the PVD process on the pre-nitrided layers—either by too high a deposition temperature or by a temperature increase caused by ion bombardment during the etching or sputter cleaning process prior to coating deposition.

Adhesion analysis of the PVD hard coating

The hardness indents of the Rockwell-C indentation tests for selected treatment conditions are shown in Figure 8. The PVD-coated base material (BM-C, Figure 8(A)) was the only material state that exhibited detachments around the entire circumference of the hardness indent, indicating poor load-bearing capacity. According to the HF classification (HF-Haftfestigkeit, German abbreviation defining the adhesive strength classification^42,47), such a pattern is assigned to the inadmissible adhesion strength class HF5. In contrast, all duplex conditions exhibited HF1 adhesion strength (Figure 8(B)-(D)). Even if the compound layer was removed (PN1-8-C) prior to the PVD process and the load-bearing capacity was thus reduced, no detachment or chipping of the hard coating was detected around the hardness indent. The poor adhesion strength of the hard coating deposited on the non-nitrided steel substrate (Figure 8(A)) was attributed to the absence of a hardness increase introduced by the nitrided layer, which would have provided an increased load-bearing capacity for the hard coating, as confirmed in the study carried out by Torres et al²³ on an H13 tool steel. In addition, Ge et al⁴⁸ report on the improved adhesion of a CrTiAlN layer in a duplex-treated H13 steel in contrast to a coated but non-nitrided state. Peng et al²⁷ compare the adhesion of an AlCrN hard coating deposited on a nitrided substrate with a compound layer with the adhesion of an AlCrN coating deposited on a nitrided substrate in which the compound layer is removed by polishing, and conclude that both duplex states exhibit the adhesion strength class HF1.

The results of the Rockwell-C indentation tests carried out in this study show that the adhesion of hard coatings deposited on all nitrided substrates with or without mechanical finishing prior to the PVD process was very good. Since the Rockwell indentation test only facilitates the making of qualitative statements with respect to adhesive strength and since all duplex systems were assigned to adhesive strength HF1, this test was not suitable for differentiating between the adhesive strengths of the various duplex layer systems. In order to be able to evaluate the influence of the black layer on the adhesive strength, additional scratch tests were carried out on selected treatment conditions.

Figure 9 shows images of selected scratch tracks of different sample treatment conditions as well as the corresponding curves of the acoustic emission (AE) signal and the tangential force F_T. The first significant increase in the acoustic emission signal was registered at the respective critical load L_C1, at which the first cracks in the hard coating occurred according to DIN 1071-3.2005. The critical load L_C was defined as the load at which the first occurrence of layer chipping was observed in the border area of the scratch track. At the critical load L_C, the tangential force F_T declined for the first time while, simultaneously, the acoustic signal increased rapidly in all investigated samples. The PVD-coated but not nitrided substrate (BM-C) exhibited smaller critical loads in the range of L_C = 30 ± 1 N (Figure 9(A)) than the duplex-treated samples (Figure 9(B)-(D)). This could have been due to the lower surface hardness of the non-nitrided substrate and its lower load-bearing capacity for the hard coating. Sample PN0-C, which was compound layer-free in the as-nitrided state, exhibited a critical load of L_C = 45 ± 1 N (Figure 9(B)). This critical load was about 50% higher than the critical load measured for the PVD-coated non-nitrided substrate BM-C. Moreover, for sample PN1-C, which contained a compound layer in the as-nitrided state, an increased critical load of L_C = 53 ± 1 N (Figure 9(C)) was determined that corresponded to an increase of approximately 70% in comparison to the PVD-coated non-nitrided substrate. These significantly increased critical loads clearly showed the improvement in the load-bearing capacity of the substrate as a result of the increase in hardness achieved by means of PN. Sample PN1-8-C, in which the compound layer was mechanically removed before the PVD process, exhibited a significantly lower critical load of L_C = 40 ± 1 N in comparison to the coated nitrided sample PN1-C. This value was also lower than the critical load of sample PN0-C, which was nitrided without forming a compound layer. Since sample PN1-8 exhibited a similar hardness gradient to the one in the compound layer-free nitrided sample PN0, the observed reduction of the critical load in sample PN1-8-C could only be explained by a change in the interface quality between the hard coating and the nitrided layer resulting from the mechanical polishing before the PVD process.

Figure 10 shows the SEM images of the scratch tracks in three PVD-coated states BM-C, PN0-C, and PN1-C that were recorded in the range of the critical load L_C. In all of the investigated states, bending cracks and the material penetration caused by the plastic deformation of the substrate were visible. These features combined with the characteristic adhesive spallation at the edge areas of the cracks to act as failure mechanisms.

The duplex state with the black layer (PN1-C, Figure 10(E)) showed the highest critical load and, thus, the best adhesive strength. In contrast to the data available in the literature,^34,49 the mechanical polishing of the compound layer in sample PN1-8 led to a reduction of the critical load and to a deterioration in adhesive strength. Although the compound layer that formed during nitriding completely transformed into a black layer during the PVD process, the decomposition of the compound layer did not yield a reduction in the adhesive strength of the Cr-Al-Ti-B-N PVD coating. In accordance with the results obtained in this work, He et al² observe good adhesive strength in duplex layer systems with a black layer. According to their results, the adhesion of a deposited hard coating depends much more on the reduction of the substrate hardness as a result of the PVD coating process than on the formation of a black layer.

CONCLUSIONS

In this study, the influence of PN and a subsequent PVD process on selected properties of a duplex-treated AISI D2 tool steel was investigated. PN promoted the formation of a diffusion layer in which nitrogen was interstitially dissolved in the steel as well as precipitated as nitrides, and an outer compound layer consisting of γ′-Fe₄N and ε-Fe₃N.

The additional thermal treatment of the nitrided substrates during the PVD process that was carried out at 480°C led to the decomposition of the iron nitrides in the compound layer and to the formation of a black layer composed primarily of α-Fe. For the PVD-coated states the compound hardness HV_0.5 of the sample with the black layer was lower than that of both the nitrided substrate without the compound layer and the nitrided substrate with mechanically removed compound layer. However, the presence of the black layer improved the adhesion of the PVD coating, as concluded from the highest critical load measured by scratch testing. Mechanical removal of the compound layer prior to the deposition of the hard coating led to a reduction of the critical load in the scratch test.

ACKNOWLEDGMENTS

The authors wish to thank Dr. H. Großmann from Anton Paar for facilitating their use of the NHT² device for the performance of instrumented hardness measurements of the PVD coating. The support of Dr. A. Engel (University of Applied Science, Mittweida) in the realization of the scratch tests is kindly appreciated.

The authors acknowledge the support of the project “Saxon Alliance for Material- and Resource-Efficient Technologies (AMARETO)” (Project No. 100291457), which is funded by the European Union (European Regional Development Fund) and by the Free State of Saxony.

PEER REVIEW INFORMATION

Engineering Reports thanks Yang Li and other anonymous reviewers for their contribution to the peer review of this work.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

AUTHOR CONTRIBUTIONS

Anke Dalke: Conceptualization; writing-original draft. Tom Weinhold: Data curation. Alexander Schramm: Investigation. Christina Wüstefeld: Conceptualization; investigation; writing-original draft. Ulrike Ratayski: Conceptualization; data curation. Vjaceslav Sochora: Investigation. Horst Biermann: Supervision. David Rafaja: Funding acquisition; supervision; writing-review and editing.