Corrosion and Wear-Resistant Composite Zirconium Nitride Layers Produced on the AZ91D Magnesium Alloy in Hybrid Process Using Hydrothermal Treatment

Abstract

The aim of the study was to investigate the possibility of an effective improvement in performance properties, including corrosion and wear resistance of magnesium AZ91D alloy using a surface engineering solution based on zirconium nitride composite surface layers produced on AZ91D alloy in a hybrid process using hydrothermal final sealing. Research results show that the formation of a composite ZrN-Zr-Al-type zirconium nitride layer on zirconium and aluminum sublayers results in a significant increase in resistance to corrosion and wear. The decrease in chemical activity of the sealed zirconium nitride composite layer on AZ91D, expressed by the displacement of the corrosion potential in the potentiodynamic test, reaches an outstanding value of ΔE_corr = 865 mV. The results of the SIMS chemical composition analysis of the layers indicate that the sealing of the composite layer occurs at the level of the aluminum sublayer. The composite layer reduces wear in the Amsler roll on block test by more than an order of magnitude. The possibility of effective sealing of zirconium nitride layers on the AZ91D alloy demonstrated in this study, radically increases the corrosion resistance and combined with the simultaneous mechanical durability of the layers, is of key importance from the point of view of new perspectives for application in practice.

1. Introduction

Magnesium alloys, with the lowest density (1.84 g/cm³) of all metallic materials, have a number of advantageous performance properties, such as high specific strength, vibration damping ability, effective shielding of electromagnetic fields, and biocompatibility, as well as advantageous technological properties, especially casting properties and good thermal conductivity predestining them for die casting, are of growing interest in modern technology in a wide range of applications beyond their original dominance in the aerospace and automotive industries [1,2]. In practice, however, the use of magnesium alloys is limited by their often insufficient corrosion resistance, as well as their resistance to wear. In industrial practice, the predominant way to obtain satisfactory corrosion resistance of products is their surface treatment by anode oxidation in a classic variant, and in particularly demanding applications, the modern variants of Plasma Electrolytic Oxidation (PEO) [3], which also offer increased mechanical properties, including hardness and wear resistance. Potential new specialized applications of magnesium alloys necessitate the search for adequate, effective surface engineering solutions. Among such presently investigated modern solutions one could mention Ni-diamond micro-composites coatings [4], laser-cladded coatings [5] and new PEO treatment variants [6]. The production of surface layers of nitrides on magnesium alloys of such metals as aluminum [7], chromium [8,9], titanium [8,10,11], zirconium [12,13,14], and their composites [15,16], and other advanced nitride-based solutions [17,18,19], due to the high corrosion and tribological resistance, seems to be also a promising direction, systematically explored by various researchers. One of the new promising areas of application of magnesium alloys is in biomedical applications, which have recently attracted the attention of researchers and engineers. This is due to the advantageous properties of magnesium such as biocompatibility and mechanical properties, in particular its bone-like stiffness, hence the concept of using suitable magnesium alloys for implants, including resorbable implants. Due to the high activity of magnesium, which can also be subject to more or less intense corrosion processes in body fluids, the key issue for the success of this concept in practice is the development of adequate controlled corrosion resistance of the magnesium alloy forming the implant. This can be achieved on the one hand by selecting the alloy forming the implant and, on the other hand, most effectively by surface engineering methods, producing suitable surface layers controlling corrosion processes. Recently, there has been a growing interest among researchers in using zirconium nitride-based layers in biomedical applications [12,13,14], due to their advantages [20], in particular biocompatibility, which is key in such applications. It should be noted that the surface nitride layers on magnesium alloys, including zirconium nitride layers, modify the corrosion properties of the magnesium alloy to a relatively small extent, particularly as regards changes in chemical activity, as measured by the corrosion potential (E_corr) [11,14]. This is due to structural defects typical of Physical Vapor Deposition (PVD) methods used in their production processes, such as droplets and craters, which are the source of potential micro-leaks, through which the corrosive medium can penetrate through the layer to the substrate, and consequently, due to the conductive, cathodic nature of the nitride layers on magnesium alloys, leads to the formation of corrosion cells and accelerated galvanic corrosion [21]. This unfavorable behavior of the layers, more or less effectively, may be counteracted by applying various solutions limiting the harmful effects of defects [22]. In this work, in order to obtain high performance properties of the AZ91D alloy, in particular corrosion resistance, as well as resistance to wear, a solution was proposed based on the production of a composite three-zone zirconium nitride, zirconium and aluminum surface layer of the ZrN-Zr-Al type produced by a hybrid process using earlier developed [21] hydrothermal final sealing in boiling water bath. In particular, the main objective of the study was to verify the susceptibility of zirconium nitride-based layers to hydrothermal sealing. The origin of the concept of final sealing of zirconium nitride composite layers was in earlier works on the hybrid production of tight, highly resistant to corrosion and wear, TiN-Ti-Al titanium nitride composite layers, which showed outstanding effectiveness [21,23,24].

2. Materials and Methods

All layers tested in this work were produced on a substrate of pressure die casting magnesium alloy AZ91D containing 9.0 wt.% Al, 0.7 wt.% Zn and 0.1 wt.% Mn. Surface composite layers of the ZrN-Zr-Al type, consisting of an outer zirconium nitride zone, zirconium in the middle zone and an aluminum zone near the substrate, were produced on the substrates initially prepared by mechanical polishing, using diamond suspension with sequential powder gradation from 9 to 1 µm. A hybrid method was used, consisting of a combination of PVD processes, sequentially Magnetron Sputtering (MS) for Al and Zr and Arc Evaporation (AE) for ZrN, with the final hydrothermal sealing by immersion in a boiling water bath for 30 min [21]. A scheme describing the above method is shown in Figure 1. The scheme additionally indicates the presence of typical defects arising during the deposition of subsequent sublayers by PVD methods, i.e., droplets and craters that are crucial from the point of view of corrosion resistance. The parameters of the methods used to produce the successive zones of the tested layer and the control samples are included in

Table 1. Variants of the investigated materials.

Variant Denotation	Type of Coating Type of PVD Method Coating Thickness [µm]			Treatment
	Al	Zr	ZrN
	MS 1	MS 1	AE 2
AZ91D	–	–	–	As delivered
ZrN-Zr-Al	9	1	2	As deposited
ZrN-Zr-Al_S	9	1	2	Hydrothermal sealing
ZrN-Zr *	–	1	2	As deposited

¹ Magnetron Sputtering, ² Arc Evaporation, * reference material.

and

Table 2. PVD processes parameters used to produce the investigated layers.

Processing Parameters	Deposition of Al Magnetron Sputtering	Deposition of Zr Magnetron Sputtering	Deposition of ZrN Arc Evaporation
Working atmosphere	Ar	Ar	N2
Pressure	5 × 10−3 mbar	5 × 10−3 mbar	1.2 × 10−2 mbar
Gas flow	200 mL/min	200 mL/min	72 mL/min
Substrate temperature	<200 °C	<200 °C	<230 °C
Bias voltage	−100 V	−50 V	−120 V
Bias current	65 A	65 A	4 × 65 A

. Table 1 additionally presents the designations of the samples used in the further part of the article for their identification.

First, the microstructure, chemical and phase composition as well as surface morphology of the produced layers were examined. Embedded in resin, ground, polished metallographic specimens were observed in a reflection metallographic microscope at magnifications up to 1000 times. X-ray diffraction analysis (XRD) was performed using a Rigaku SmartLab SE diffractometer using a Cu Kα lamp. The surface morphology was analyzed using a Scanning Electron Microscope (SEM) Hitachi SU8000. The accelerating voltage was 10 kV. The analysis of the chemical composition was carried out using Secondary Ion Mass Spectrometry (SIMS) on a Cameca IMS6F spectrometer (Cameca, Gennevilliers, France). SIMS measurement was conducted using a cesium (Cs⁺) primary beam and secondary ions as measured, and secondary ions as MeCs⁺ clusters were analyzed. The method of measuring nitrogen as NCs⁺ and oxygen as OCs⁺ clusters has been described elsewhere [25].

Corrosion resistance is a key aspect of suitability for use of the produced composite layers of the ZrN-Zr-Al type on the magnesium alloy AZ91D. It was tested using the potentiodynamic method on the AutoLab PGSTAT100 device from EcoChemic B.V. The range of potentials used in the study was from −1600 mV to −1000 mV at room temperature with a potential change rate of 0.1 mV/sec. The tests were carried out in a 0.5 M NaCl solution. The potential before sample polarization was stabilized by immersing the sample in the tested solution under electroless conditions for 60 min. The reference electrode was a saturated calomel electrode (Hg/Hg₂Cl₂/KCl) with a potential of +240 mV relative to the hydrogen electrode. The auxiliary electrode was a platinum electrode.

The mechanical durability of cathodic anti-corrosion layers on magnesium alloys, which include zirconium nitride layers, is crucial for maintaining corrosion resistance during the operation of machine parts and devices made of magnesium alloy in conditions of exposure to mechanical puncture damage, scratches or tribological wear. Any violation of the continuity of the layer under the influence of mechanical factors, similar to the natural, leaking defects of the structure, will result in the formation of corrosion microcells and, as a result, accelerated destruction of the layer as a result of galvanic corrosion. Therefore, in order to ensure the service life of products made of magnesium alloys exposed to corrosion, apart from corrosion resistance, high mechanical resistance is also required [23]. Therefore, a number of mechanical tests were performed. The resistance of the layers to concentrated point loads was tested using a Vickers hardness indentation at a load of 9.81 N (HV1) in order to detect the layer’s eventual cracking or exfoliation. Scratch resistance in the scratch test was performed on a Micro Combi Tester (CSM Instruments SA, Puseux, Switzerland) device. A Rockwell indenter was used. The load on the indenter increased from 1 N to 20 N, and the feed rate was 5 mm/min. The length of the scratch was 6 mm. On the basis of the recorded acoustic emission and microscopic observations after the test, the critical forces Lc₁, Lc₂ and Lc₃ breaking the continuity of ZrN-Zr sublayers system were determined. The resulting scratches were observed in a Scanning Electron Microscope Hitachi S-3500N. The accelerating voltage was 15 kV. The distribution of elements in the scratch area was also studied using Energy Dispersive Spectroscopy EDS on the same microscope. Wear resistance by the roll on block method was performed in sliding wear conditions on a Amsler A-135 apparatus. Heat-treated C45 steel with a hardness of 35 HRC was used as the roll (counter-body). The test time was 60 min. Three tests were performed with loads of 10, 25 and 50 N.

3. Results and Discussion

3.1. Microstructure, Chemical and Phase Composition of the Layer

The sealed composite layer of the ZrN-Zr-Al type produced on the AZ91D alloy (ZrN-Zr-Al_S) was characterized by a macroscopically homogeneous appearance and no visible defects in the form of exfoliation or cracks. The microstructure of this layer is shown in Figure 2. In cross-section, the zirconium nitride layer has a characteristic golden color, the zirconium intermediate layer is dark grey, and the aluminum is light grey.

The results of the phase composition tests using X-ray phase analysis confirm that in the process of producing the outer surface layer of zirconium nitride, a nitride with the stoichiometry of ZrN was obtained.

Within the range of available magnifications in optical microscopy, no defects in the form of discontinuities or cracks, or decohesion of the layer from the substrate or between the sublayers were observed (Figure 2). The occurrence of defects typical for layers produced by PVD methods in the form of droplets and craters are mainly confined to the outer nitride zone and ZrN-Zr zirconium sublayer. As it can be assumed, by analogy to the previously developed TiN-Ti-Al composite titanium nitrides on the AZ91D alloy [24], the composite layer in this study has a diffusion character. As shown [24] in the hybrid manufacturing process, due to the increased temperature of the substrate in the process of deposition of the ZrN coating by evaporation in the arc and its hydrothermal sealing, diffusion processes should occur between the aluminum sublayer and the substrate, and probably also in a thin nanometric zone between the aluminum and zirconium sublayers.

The thicknesses of individual component layers of the composite layer—outer zirconium nitride, intermediate zirconium and aluminum layers (Figure 2), were, respectively, ca. 2, 1 and 9 µm, which gives a total layer thickness of ca. 12 µm.

A view of the surface of the composite zirconium nitride layer ZrN-Zr-Al is shown in Figure 3. The surface shows a morphology typical for coatings deposited by arc evaporation with characteristic defects formed in this process, in the form of the aforementioned droplets and craters formed after some droplets chipping off (Figure 3).

It should be noted that the morphology of the outer surface of the zirconium nitride layer, as a result of the hydrothermal sealing treatment in a boiling water bath, does not seem to change, which indicates that the hydrothermal treatment may not result in covering the surface with a homogeneous oxide film sealing it and its defects (Figure 3a,c), as was observed in our previous studies for the sealing of composite titanium nitride layers of the TiN-Ti-Al type [24]. Moreover, the gaps separating the droplets from the nitride layer remain open (Figure 3d). Therefore, the SEM observation does not unambiguously settle whether a sealing oxide coating is formed on the surface of the zirconium nitride layer and its discontinuities.

A comparison of the distribution of selected elements in the unsealed layer (ZrN-Zr-Al), and in the hydrothermally sealed layer (ZrN-Zr-Al_S) as determined by the SIMS method (Figure 4), shows that the hydrothermal treatment results in an increased concentration of oxygen and magnesium at the interface of the zirconium sublayer and aluminum (Figure 4c). It leads to the conclusion that the sealing zone, most likely of magnesium hydroxide, is formed at the level of the interface between the zirconium and aluminum sublayers, in the aluminum sublayer, where magnesium diffusing from the substrate through the aluminum sublayer reacts with the environment of the boiling water bath, resulting in the formation of magnesium hydroxide, building discontinuities. The SIMS analysis results also shows a magnesium diffusion profile in the aluminum sublayer (Figure 4d) that confirms the diffusion character of the ZrN-Zr-Al composite layer.

3.2. Corrosion Resistance

Figure 5 shows the results of corrosion resistance tests using the potentiodynamic method in the form of the polarization curves of the tested variants of the AZ91D magnesium alloy, i.e., the variant with a composite layer in the initial state (as deposited) (ZrN-Zr-Al), in the state after sealing with the hydrothermal method (ZrN-Zr-Al_S) and for comparative variants, i.e., alloy without layer (AZ91D) and alloy with the reference zirconium nitride layer without aluminum sublayer (ZrN-Zr).

Table 3. Values of potentials and corrosion current densities in tests in 0.5M NaCl using the potentiodynamic method.

	AZ91D	ZrN-Zr-Al	ZrN-Zr-Al_S	ZrN-Zr
Potential E [mV]	−1530 ± 1.4	−1147 ± 0.7	−665 ± 0.9	−1577 ± 0.3
Corrosion current density j [µA/cm2]	99.72 ± 2.39	5.17 ± 0.59	3.24 ± 0.26	22.48 ± 0.61

, on the other hand, contains the corresponding values of corrosion parameters, i.e., corrosion potentials and currents.

Analyzing the layout of the polarization curves, it can be seen that the formation of a layer of zirconium nitride on the zirconium sublayer on the AZ91D magnesium alloy (reference variant ZrN-Zr) causes the corrosion potential E_corr to shift in the negative direction by ΔE_corr = −47 mV to the value of E_corr = −1577 ± 0.3 mV, which indicates an increase in chemical activity, and therefore a deterioration in corrosion resistance. The corrosion current, on the other hand, decreases nominally by almost five times, which seemingly indicates a significant slowdown of corrosion processes; however, taking into account that the damage to the coating occurs locally, by creating a pit, the actual corrosion kinetics are probably significantly higher. A similar effect of nitride coatings on the corrosion behavior of the AZ91D alloy, manifested by the deterioration of corrosion resistance, was already observed in earlier works on titanium nitrides [24]. The reasons for such unfavorable behavior of the alloy with nitride coatings, including zirconium nitrides, should be sought in the lack of tightness of the nitride coating associated with the occurrence of typical defects in the form of droplets and craters, characteristic for PVD methods used for their production, and even more so in the case of possible discontinuities of the layer in the form of micro-cracks or micro-flakes. Layer discontinuities, such as deep gaps between the surface of the droplets and the layer when they pass through the layer to the substrate, allow the corrosive environment to access the magnesium alloy substrate. Due to the conductive nature of most nitride coatings, including zirconium, and their cathodic nature in relation to the magnesium alloy, this leads to the formation of local corrosion microcells between the highly active magnesium alloy and the relatively nobel nitride coating [16,21,23], and as a result, galvanic corrosion causing local perforation of the layer by the mechanism of pitting corrosion. It should be noted that the occurrence of defects, and consequently leak thickness, is statistically unavoidable in the case of layers produced by PVD methods, therefore, by nature, such layers must be susceptible to galvanic corrosion. As a result, in order to obtain the absolute tightness necessary to eliminate the risk of galvanic corrosion and accelerated degradation, these layers require final sealing [22]. The developed three-layer composite zirconium nitride layer in a deposited state, on the contrary, significantly increases the corrosion resistance (Figure 5, ZrN-Zr-Al curve), which is manifested by the shift of the corrosion potential in the positive direction by almost ΔE_corr = 400 mV to the value E_corr = −1147 ± 0.7 mV, while the corrosion current decreases by about an order of magnitude (about 5 µA/cm²). What is more, the character of the polarization curve changes, on which a clearly marked, stable passive region appears, about 250 mV wide, with a breakdown potential of about E_p = −905 ± 0.5 mV. The reason for such favorable behavior as observed earlier in the case of composite layers of the TiN-Ti-Al type [21,24] is certainly consistent with the assumption of a separation in the structure of the composite layer, the outer layer of zirconium nitride from the active substrate by a relatively thick, corrosion-resistant intermediate, sealing aluminum sublayer. The key sealing hydrothermal treatment of the composite layer of zirconium nitride in a boiling water bath, which ends the hybrid process, results in a further significant reduction in chemical activity (Figure 5, ZrN-Zr-Al_S curve), and consequently an improvement in corrosion resistance expressed by a shift of the corrosion potential in the positive direction to the value E_corr = −665 ± 0.9 mV, i.e., further approx. ΔE_corr = 865 mV. This represents a radical reduction in chemical activity, unprecedented in the literature, and, as a result, a significant improvement in corrosion resistance. The exception is our previous works on analogous composite layers of titanium nitride of the TiN-Ti-Al type [21,23], in the case of which an even greater increase in corrosion resistance was obtained with a shift of the corrosion potential to positive values. The reasons for the differences in the behavior of the composite layers of titanium nitride and zirconium nitride with an aluminum sublayer are probably due to the different layer sealing mechanism. In the case of titanium nitride layers, it was shown to occur at the level of the titanium sublayer, hence the recorded corrosion potential, as for titanium, is positive [24]. Sealing of the composite zirconium nitride layer with the aluminum sublayer, as indicated by the similar values of corrosion potentials recorded for the ZrN-ZR-Al_S composite layer and for the hydrothermally sealed and unsealed aluminum layers on the magnesium alloy AZ91D, occurs on the level of the aluminum sublayer [24]. Moreover, this is supported by the results of the SIMS study (Figure 4), which reveal an increased concentration of oxygen and magnesium at the interface between the zirconium and aluminum sublayers in the case of the composite zirconium nitride layer subjected to hydrothermal treatment, while the oxygen levels in the outer zirconium nitride layer and the zirconium sublayer remain at a similar level. This effect suggests the formation of magnesium hydroxide sealing the interface between the two sublayers. It should be mentioned that due to the fact that earlier works [24] showed the ineffectiveness of attempts to seal the layers of TiN-Ti titanium nitrides with the hydrothermal method, i.e., without an aluminum sublayer, in this study similar attempts with regard to the analogous layers of zirconium nitride of the ZrN-Zr type were considered groundless.

3.3. Mechanical Properties

The results of mechanical damage resistance tests are shown in Figure 6, Figure 7 and Figure 8. From the result of the Vickers hardness indentation test (Figure 6), it can be concluded that the composite layer does not show any tendency to cracking or exfoliating of layer fragments in the area of the indentation, which proves a qualitatively good connection of the component layers with each other and the composite layer with the substrate.

Similarly, in the case of the scratch test (Figure 7a), until the outer zirconium nitride layer on the ZrN-Zr zirconium sublayer is completely removed, no damage to the layer in the vicinity of the scratch in the form of cracks or exfoliation of the nitride layer fragments is observed. On the other hand, in the scratch trace, the nitride layer, starting from the critical load Lc₁ = 1.88 ± 0.019 N (

Table 4. Values of critical forces Lc₁, Lc₂, Lc₃ in the scratch test.

Layer Type		Lc1	Lc2	Lc3
ZrN-Zr-Al	Road [mm]	0.28 ± 0.01	0.53 ± 0.01	1.47 ± 0.01
	Force [N]	1.88 ± 0.019	2.69 ± 0.027	5.66 ± 0.057
ZrN-Zr	Road [mm]	0.53 ± 0.01	1.34 ± 0.01	2.56 ± 0.01
	Force [N]	2.64 ± 0.026	5.11 ± 0.051	8.93 ± 0.089

), successively cracks radially as the load increases and is dented into the relatively plastic aluminum sublayer, but without exfoliating it. The first decohesion of the fragments of the zirconium nitride layer between the cracks, resulting in the exposure of the aluminum sublayer, is observed under the critical load Lc₂ = 2.69 ± 0.027 N, and the complete removal of the nitride layer in the trace of the crack for the load Lc₃ = 5.66 ± 0.057 N. It should be noted that the composite layer of the ZrN-Zr-Al type is damaged at lower critical load values than the reference layer of zirconium nitride ZrN-Zr produced directly on the AZ91D alloy (Table 4), but these damages, as shown in Figure 7 (EDS), do not lead to the exposure of a highly chemically active magnesium substrate and, in case of potential contact with a corrosive environment, pose a serious risk of accelerated galvanic corrosion. The aluminum sublayer, as the scratching progresses and the load increases, is not subject to cracking or exfoliation of its fragments, but due to its plasticity, it is gradually abraded, protecting the magnesium alloy substrate from the corrosive environment for a relatively long and effective time. Its complete removal from the surface of the AZ91D alloy occurs for a critical load value of approx. Lc_3” = 12.5 ± 0.125 N, significantly exceeding the value of the critical force for the comparative variant ZrN-Zr, which is Lc₃ = 8.93 ± 0.083 N, for which actually the first layer crack at Lc₁ = 2.64 ± 0.026 N creates a critical risk of galvanic corrosion. As long as mechanical damage to the ZrN-Zr-Al type composite layer is localized in the outer zirconium nitride layer and the zirconium sublayer, without affecting the cohesion of the aluminum sublayer with the magnesium alloy substrate, it leads only to a decrease in corrosion resistance, and not to accelerated galvanic corrosion, as in the case of ZrN-Zr layers without an aluminum sublayer. It can therefore be assumed that the sealed zirconium nitride composite layer of the ZrN-Zr-Al type on the AZ9D alloy has a good prognosis in terms of service life of products made of this alloy, also in conditions of simultaneous corrosion and mechanical hazards.

The result of the production of the ZrN-Zr-Al-type zirconium nitride composite layer on the AZ91D magnesium alloy on the intermediate zirconium and aluminum sublayers in the process is more than a 1.5-fold increase in surface hardness from 84 ± 7 HV0.05 for the alloy to 132 ± 4 HV0.05. Hardening the surface of the AZ91D alloy with a layer of zirconium nitride results in a significant, more than one order of magnitude, increase in resistance to wear in the load range up to 50 N in the modified Amsler roll on block test (Figure 8). It should be noted that during the test, the outer layer of zirconium nitride with a thickness of approx. 2 µm does not wear through. As can be seen, the wear of the ZrN-Zr-Al type layer in the roll on block test is nearly 2.5 times lower than for the ZrN-Zr layer without an aluminum sublayer.

4. Conclusions

• The effect of the zirconium nitride composite layer of the ZrN-Zr-Al type with zirconium and aluminum intermediate layers produced on the magnesium alloy AZ91D using PVD methods sealed in the final hydrothermal treatment process is a significant improvement in corrosion resistance in the 0.5M sodium chloride environment, manifested in the potentiodynamic test shift of the corrosion potential towards positive values by ΔE_corr = 865 mV in comparison to the alloy without the layer (E_corr = −1530 ± 1.4 mV). The unsealed ZrN-Zr-Al layer improves the corrosion resistance to a much lower extent (ΔE_corr = 385 mV). On the other hand, the zirconium nitride layer of the ZrN-Zr type, produced directly on the AZ91D alloy, reduces the corrosion resistance of the alloy with a negative shift of the corrosion potential to the value of E_corr = −1577 ± 0.3 mV. The unfavorable effect of the ZrN-Zr-type zirconium nitride layer on the corrosion resistance of the AZ91D alloy is the result of the presence of inevitable defects in the layer structure, typical for PVD methods, in the form of droplets and craters, which are the source of micro-discontinuities of the nitride layer. This leads to the formation of corrosion cells between the cathodic zirconium nitride coating and the highly chemically active magnesium alloy, and consequently results in accelerated galvanic corrosion. Hence, the key role of the hydrothermal final sealing treatment of the composite layer of zirconium nitride with an intermediate layer of zirconium and aluminum plays a key role in achieving maximum tightness.
• The production of a composite layer of zirconium nitride on the intermediate zirconium and aluminum sublayer of the ZrN-Zr-Al type on the AZ91D magnesium alloy resulted in a more than two-fold increase in surface hardness. The effect of hardening the surface of the AZ91D alloy with a layer of zirconium nitride results in a significant, approximately more than one order of magnitude increase in friction wear resistance in the load range up to 50 N in the modified Amsler roll on block test, while the outer layer of zirconium nitride is not worn through. The increase in resistance observed for the ZrN-Zr layer without the aluminum sublayer is almost 2.5 times lower. The composite zirconium nitride layer of the ZrN-Zr-Al type in the Vickers hardness indentation test carried out using the Vickers HV1 indentation showed resistance to mechanical damage under the influence of concentrated loads. This layer also showed favorable behavior in the scratch test, because damage to the external layer of zirconium nitride during scratching, in the form of cracks, its local or even complete exfoliation, did not expose the substrate made of magnesium alloy AZ91D, but only the aluminum sublayer. During the scratching progresses and the load increases, the sublayer is not subject to cracking or exfoliation of its fragments, but due to the plasticity of aluminum, it is only gradually worn, protecting the magnesium alloy substrate from the corrosive environment effectively. Mechanical damage to the outer layer of zirconium nitride in the ZrN-Zr-Al layer only leads to a decrease in corrosion resistance, and not to accelerated galvanic corrosion, as in the case of ZrN-Zr layers without an aluminum sublayer. As a result, a composite layer of the ZrN-Zr-Al type promises prospectively well in terms of its service durability of AZ91D magnesium alloy products, not only in corrosive, but also mechanical conditions.