Effect of Preliminary Irradiation of 321 Steel Substrates with High-Intense Pulsed Ion Beams on Scratch Test Results of Subsequently Deposited AlN Coatings

: The paper presents the effect of irradiation of 321 steel substrates with a high-intense pulsed ion beam (HIPIB) on changes in functional properties of the surface layers and tribological characteristics of AlN coatings subsequently deposited above by the reactive magnetron sputtering method. The morphology of the modiﬁed surface layers, their microhardness and free surface energy levels are presented for different HIPIB energy densities. HIPIB irradiation of the substrates caused variations in the results of scratch tests combined with the acoustic emission signal processing. Their analysis has enabled concluding that the crack initiation threshold could be at least doubled for the studied coating/substrate system due to preliminary HIPIB irradiation. Finally, the obtained data were discussed, and future research directions were proposed. methodology, V.T., S.P., E.S., V.U. and G.R.; validation, V.T., S.P., V.U., E.S. and G.R.; analysis, V.T., S.P., V.U., M.S. and and data and

One of the effective ways to modify the surface layers of metals, alloys and steels is irradiation with a high-intense pulsed ion beam (HIPIB). In this case, a surface is cleaned, a specific microrelief is formed, and the microstructure and phase composition of a material is changed [68][69][70][71][72][73][74][75][76][77][78][79][80]. In addition, internal stresses in the surface layers can be varied due to different HIPIB parameters. Accordingly, matching in magnitude and sign of stresses in types of coatings and deposition methods. The aim of these studies has been their partial filling in relation to tribological characteristics of the AlN coatings deposited on the 321 steel substrates, including preliminary HIPIB irradiated ones, by the reactive magnetron sputtering procedure. Moreover, dependences of residual stresses in trial coatings on silicon substrates were initially assessed.
It should be noted that the samples were irradiated at room temperature of about 20-25 °C and did not heat up additionally. During HIPIB irradiation, the surface layers were re-melted at the energy densities of 1.5 and 2.0 J/cm 2 , while their temperatures were below the melting points at the lower levels. The heated layer depths corresponded to the stopping range of ions within few micrometers. After one pulse, the entire sample temperature (T) could be increased by about 1-3 °C according to the following relationship: where EHIPIB was the energy densities per pulse, c and ρ were heat capacity and density of the 321 steel (c = 462 J/kg°C, ρ = 7670-8000 kg/m 3 [82]), V was the sample volume. Respectively, the samples could be heated by no more than 3-9 °C in three pulses. Firstly, by analogy with the paper [35], trial AlN coatings were deposited on silicon substrates 0.38 mm thick by the reactive magnetron sputtering method to assess internal It should be noted that the samples were irradiated at room temperature of about 20-25 • C and did not heat up additionally. During HIPIB irradiation, the surface layers were re-melted at the energy densities of 1.5 and 2.0 J/cm 2 , while their temperatures were below the melting points at the lower levels. The heated layer depths corresponded to the stopping range of ions within few micrometers. After one pulse, the entire sample temperature (T) could be increased by about 1-3 • C according to the following relationship: where E HIPIB was the energy densities per pulse, c and ρ were heat capacity and density of the 321 steel (c = 462 J/kg • C, ρ = 7670-8000 kg/m 3 [82]), V was the sample volume. Respectively, the samples could be heated by no more than 3-9 • C in three pulses. Firstly, by analogy with the paper [35], trial AlN coatings were deposited on silicon substrates 0.38 mm thick by the reactive magnetron sputtering method to assess internal stresses. Then, the optimal mode was applied to deposit them on the HIPIB irradiated substrates from the 321 steel for evaluating tribological characteristics of the coatings. The deposition parameters were the following: a magnetron power of 1.5 kW, the substrate-totarget distance of 150 mm and partial pressure of reactive gases of 0.16 Pa. Various residual stresses were achieved due to different substrate temperatures of 120, 280, and 500 • C. This prerequisite was based on the results reported in [83], according to which residual stresses in the AlN coatings changed from tensile to compressive ones as the substrate temperatures rose.
The thicknesses of the coatings were determined using an 'MII-4 microinterferometer. Average residual stresses (σ f ) in the trial ones were estimated using the Stoney formula [84]: where E s was the silicon elastic modulus of 169 GPa [85]; d s was the substrate thickness; R was the substrate bending radius; ν S was the silicon Poisson's ratio of 0.28 [85]; d f was the coating thickness. The substrate bending radiuses R were measured by a 'Micro Measure 3D Station (STIL)' non-contact profilometer. X-ray diffraction analysis was performed using a 'Rigaku Ultima IV' diffractometer (Rigaku Analytical Devices, Inc., Wilmington, MA, USA) Microstresses were calculated by the Williamson-Hall method [86]. A beam incidence angle was 5 • to exclude the influence of the substrates.
Roughness was measured by an 'NT-MDT Integra Prima' atomic force microscope on a 100 µm base (NT-MDT Spectrum Instruments, Moscow, Russia).
Vickers microhardness was determined using a 'PMT-4M' device (Matsuzawa Co.,Ltd, Akita, Japan) at a load of 40 g. The indentation depth was 2.0-2.4 µm. The measured diagonals of ten diamond indenter imprints were averaged for each HIPIB irradiation mode.
The elastic modus values of the modified layers were assessed by a 'Micro-Scratch Tester MST-S-AX-0000 facility (CSEM, Neuchatel, Switzerland).
On the metal substrates, the tribological characteristics of the coatings were investigated by scratch tests combined with processing of acoustic emission (AE) signals by the 'Micro-Scratch Tester MST-S-AX-0000 setup with a diamond indenter, similar to [47,51,[87][88][89][90]. The indenter radius was 100 µm.
All quantitative data were statistically processed using MS Excel software (Microsoft, Redmond, WA, USA) for the most illustrative visualization of the dynamics of changes. Table 1 presents the deposition conditions and properties of the trial coatings on the silicon substrates. According to their lattice parameters calculated by the Rietveld method using the X-ray diffraction data, the trial coatings included the single AlN phase with the hexagonal close-packed lattice (the P63mmc space group). At the minimum studied substrate temperature, the average internal stresses were tensile of about 0.1 GPa, while they changed the sign to compressive and increased in amplitude with enhancing the temperature up to 280 and then up to 500 • C. The coating deposited at 120 • C had the lowest microstrains of 0.184%, while they were the highest (0.920%) after deposition at 280 • C. This coating possessed the AlN (100) most intense diffraction peak, but it was the AlN (002) one for the other two samples.  Figure 2 shows SEM images of the 321 steel surfaces (both before and after HIPIB irradiation at different modes). Their comparison enabled tracing changes in their microreliefs depending on the energy density: from the initial surfaces with clearly distinguishable grinding traces to the smooth recrystallized ones at 2.0 J/cm 2 through partial melting of irregularities at both levels of 0.6 and 1.0 J/cm 2 . After HIPIB irradiation using the re-melting modes, typical microcraters were found on the sample surfaces. Thicknesses of the modified layer were not determined in these studies due to both high labor intensity and cost, but they were typically a few micrometers for such cases [91].

Characteristics of the Trial AlN Coatings on the Silicon Substrates
This coating possessed the AlN (100) most intense diffraction peak, but it was the AlN (002) one for the other two samples.  Figure 2 shows SEM images of the 321 steel surfaces (both before and after HIPIB irradiation at different modes). Their comparison enabled tracing changes in their microreliefs depending on the energy density: from the initial surfaces with clearly distinguishable grinding traces to the smooth recrystallized ones at 2.0 J/cm 2 through partial melting of irregularities at both levels of 0.6 and 1.0 J/cm 2 . After HIPIB irradiation using the remelting modes, typical microcraters were found on the sample surfaces. Thicknesses of the modified layer were not determined in these studies due to both high labor intensity and cost, but they were typically a few micrometers for such cases [91].

Properties of the Modified Layers on the 321 Steel Substrates
Roughness Ra of the HIPIB irradiated sample surfaces increased by three times (from 20 up to 60 nm) with raising the energy density ( Figure 3) because of the formation of the crater, which was a characteristic for this surface treatment method [79,91].   Roughness Ra of the HIPIB irradiated sample surfaces increased by three times (from 20 up to 60 nm) with raising the energy density ( Figure 3) because of the formation of the crater, which was a characteristic for this surface treatment method [79,91]. This coating possessed the AlN (100) most intense diffraction peak, but it was the AlN (002) one for the other two samples.  Figure 2 shows SEM images of the 321 steel surfaces (both before and after HIPIB irradiation at different modes). Their comparison enabled tracing changes in their microreliefs depending on the energy density: from the initial surfaces with clearly distinguishable grinding traces to the smooth recrystallized ones at 2.0 J/cm 2 through partial melting of irregularities at both levels of 0.6 and 1.0 J/cm 2 . After HIPIB irradiation using the remelting modes, typical microcraters were found on the sample surfaces. Thicknesses of the modified layer were not determined in these studies due to both high labor intensity and cost, but they were typically a few micrometers for such cases [91].

Properties of the Modified Layers on the 321 Steel Substrates
Roughness Ra of the HIPIB irradiated sample surfaces increased by three times (from 20 up to 60 nm) with raising the energy density ( Figure 3) because of the formation of the crater, which was a characteristic for this surface treatment method [79,91].   The results of X-ray diffraction analysis are shown in Figure 4 and Table 2 for both initial and HIPIB irradiated samples. In the initial state, the 321 steel contained two phases: the α-Fe and γ-Fe ones. Modification of the surface layers caused changes in their phase composition. With an increase in the energy density, the α-Fe phase content reduced with a corresponding increase in the γ-Fe amount, which was the only one found at 2.0 J/cm 2 . Moreover, a decrease in the lattice parameters of both phases was observed with raising the energy density levels ( Table 2). Lowering the α-Fe lattice parameter indicated the formation of compressive macrostresses in the surface layers under HIPIB irradiation. The γ-Fe lattice parameter also decreased with an increase in the energy density up to 1.0 J/cm 2 but slightly enhanced (up to 0.3587 nm) at 1.5 and 2.0 J/cm 2 . It should be noted that microstrains rose much faster in the α-Fe phase than in the γ-Fe one. The results of X-ray diffraction analysis are shown in Figure 4 and Table 2 for both initial and HIPIB irradiated samples. In the initial state, the 321 steel contained two phases: the α-Fe and γ-Fe ones. Modification of the surface layers caused changes in their phase composition. With an increase in the energy density, the α-Fe phase content reduced with a corresponding increase in the γ-Fe amount, which was the only one found at 2.0 J/cm 2 . Moreover, a decrease in the lattice parameters of both phases was observed with raising the energy density levels ( Table 2). Lowering the α-Fe lattice parameter indicated the formation of compressive macrostresses in the surface layers under HIPIB irradiation. The γ-Fe lattice parameter also decreased with an increase in the energy density up to 1.0 J/cm 2 but slightly enhanced (up to 0.3587 nm) at 1.5 and 2.0 J/cm 2 . It should be noted that microstrains rose much faster in the α-Fe phase than in the γ-Fe one.  After HIPIB irradiation, the microhardness of the surface layers increased slightly at the minimum studied energy density of 0.6 J/cm 2 and then decreased by 25% (compared to the initial values) at the maximum level of 2.0 J/cm 2 (Figure 5a). At the same time, the elastic modulus values showed directly opposite dependences (Figure 5b). The most probable reason for this phenomenon was reducing the work-hardening effect for the rolled 321 steel plates upon re-melting.
Significant changes in the energy and chemical activity characteristics of the surface layers were found after HIPIB irradiation. The FSE levels monotonically increased over the entire studied range of the energy densities (Figure 6). At the maximum FSE value, its polar component also reached the highest level, which reflected, among other things, the chemical activity of the surfaces. At the same time, re-melting of the surface layers resulted in an almost complete suppression of the dispersed component, which could have both a positive effect due to decreasing residual stresses at the coating/substrate interfaces and a negative impact on adhesion because of reducing Van der Waals forces.  After HIPIB irradiation, the microhardness of the surface layers increased slightly at the minimum studied energy density of 0.6 J/cm 2 and then decreased by 25% (compared to the initial values) at the maximum level of 2.0 J/cm 2 (Figure 5a). At the same time, the elastic modulus values showed directly opposite dependences (Figure 5b). The most probable reason for this phenomenon was reducing the work-hardening effect for the rolled 321 steel plates upon re-melting.
Significant changes in the energy and chemical activity characteristics of the surface layers were found after HIPIB irradiation. The FSE levels monotonically increased over the entire studied range of the energy densities (Figure 6). At the maximum FSE value, its polar component also reached the highest level, which reflected, among other things, the chemical activity of the surfaces. At the same time, re-melting of the surface layers resulted in an almost complete suppression of the dispersed component, which could have both a positive effect due to decreasing residual stresses at the coating/substrate interfaces and a negative impact on adhesion because of reducing Van der Waals forces.

The Scratch Test Results of the AlN Coatings on the 321 Steel Substrates
In order to assess the effect of HIPIB irradiation of the substrates on the tribological behavior of the AlN coatings, scratch tests were carried out, combined with the AE signal processing. In this research method, the AE signal peaks were associated with the initiation and propagation of cracks in the coatings under indenter forces (Fn) above limited levels [32,37,51,[87][88][89][90]. The intensity of the AE signal peaks was proportional to (i) the number of formed cracks and (ii) the number of broken interatomic bonds (the fractured coating volume). The scratch test results and images of the sample surfaces with the indenter traces are shown in Figures 7 and 8, respectively. Unfortunately, the tester software

The Scratch Test Results of the AlN Coatings on the 321 Steel Substrates
In order to assess the effect of HIPIB irradiation of the substrates on the tribological behavior of the AlN coatings, scratch tests were carried out, combined with the AE signal processing. In this research method, the AE signal peaks were associated with the initiation and propagation of cracks in the coatings under indenter forces (Fn) above limited levels [32,37,51,[87][88][89][90]. The intensity of the AE signal peaks was proportional to (i) the number of formed cracks and (ii) the number of broken interatomic bonds (the fractured coating volume). The scratch test results and images of the sample surfaces with the indenter traces are shown in Figures 7 and 8, respectively. Unfortunately, the tester software

The Scratch Test Results of the AlN Coatings on the 321 Steel Substrates
In order to assess the effect of HIPIB irradiation of the substrates on the tribological behavior of the AlN coatings, scratch tests were carried out, combined with the AE signal processing. In this research method, the AE signal peaks were associated with the initiation and propagation of cracks in the coatings under indenter forces (Fn) above limited levels [32,37,51,[87][88][89][90]. The intensity of the AE signal peaks was proportional to (i) the number of formed cracks and (ii) the number of broken interatomic bonds (the fractured coating volume). The scratch test results and images of the sample surfaces with the indenter traces are shown in Figures 7 and 8, respectively. Unfortunately, the tester software did not enable any dimension markers to be obtained, but it was possible to proceed from the indenter radius of 100 µm when evaluating image sizes.
Coatings 2021, 11, x FOR PEER REVIEW 7 of 14 did not enable any dimension markers to be obtained, but it was possible to proceed from the indenter radius of 100 µm when evaluating image sizes.  For deposition on the metal substrates, the mode at the temperature of 120 °C was chosen with minimal resulting tensile stresses on the silicon ones. It was assumed that their adhesion to elastic substrates would be higher than that for coatings with compressive stresses. Moreover, they were minimal in value.
For the coating on the initial substrate, the fracture threshold was 3 N. HIPIB irradiation caused a shift at the beginning of the coating fracture up to 5 and 6 N at the energy densities of 2.0 and 1.0 J/cm 2 , respectively. The lower Fn value in the case of 2.0 J/cm 2 could be associated with the decrease in the substrate hardness and, as a consequence, its greater strains under the load. As a result, the AE signal intensity was significantly lower for the  did not enable any dimension markers to be obtained, but it was possible to proceed from the indenter radius of 100 µm when evaluating image sizes.  For deposition on the metal substrates, the mode at the temperature of 120 °C was chosen with minimal resulting tensile stresses on the silicon ones. It was assumed that their adhesion to elastic substrates would be higher than that for coatings with compressive stresses. Moreover, they were minimal in value.
For the coating on the initial substrate, the fracture threshold was 3 N. HIPIB irradiation caused a shift at the beginning of the coating fracture up to 5 and 6 N at the energy densities of 2.0 and 1.0 J/cm 2 , respectively. The lower Fn value in the case of 2.0 J/cm 2 could be associated with the decrease in the substrate hardness and, as a consequence, its greater strains under the load. As a result, the AE signal intensity was significantly lower for the For deposition on the metal substrates, the mode at the temperature of 120 • C was chosen with minimal resulting tensile stresses on the silicon ones. It was assumed that their adhesion to elastic substrates would be higher than that for coatings with compressive stresses. Moreover, they were minimal in value.
For the coating on the initial substrate, the fracture threshold was 3 N. HIPIB irradiation caused a shift at the beginning of the coating fracture up to 5 and 6 N at the energy densities of 2.0 and 1.0 J/cm 2 , respectively. The lower Fn value in the case of 2.0 J/cm 2 could be associated with the decrease in the substrate hardness and, as a consequence, its greater strains under the load. As a result, the AE signal intensity was significantly lower for the coatings deposited on the modified substrates, which indicated an increase in the mechanical resistance of the coating/substrate system. Accordingly, the crack initiation threshold could be at least doubled for the studied coating and substrate system due to preliminary HIPIB irradiation.

Discussion
According to the authors, discussion of the improved scratch test results after HIPIB irradiation should be started with the most probable root causes of this phenomenon. It is known that many factors affected adhesion at the ceramic/metal interfaces, among which the key ones were the microstructure, topography and chemical composition of the surface layers, plastic and elastic properties of both materials, loading conditions, presence of defects and their sizes, as well as residual internal stresses [26][27][28][29][30][31][32]92]. All of the above parameters, together with many other ones that have less effect, caused the change in the FSE levels, which were typically considered as a measure of the 'unsatisfied bond energy', arising from the 'broken bonds' exposed on the material surfaces, that affected wetting [93]. This relationship was confirmed, as the scratch test results correlated with the FSE levels ( Figures 6 and 7, respectively). However, any scratch tests showed only qualitative patterns. Despite the widespread applications of this method, it was difficult to quantify adhesion because the critical load was affected by many factors such as substrate hardness, coating thickness, interfacial bonding and parameters related to the test conditions. For example, the effect of the indenter radius on adhesion of both TiN and CrN coatings deposited on different metal substrates by arc spraying was reported in [94]. The critical load depended on the indenter radius, while groove depths in the coatings were determined only by the sample strain degree and not by the substrate materials. For both coatings, fracture modes were mainly related to their cohesive strength. In addition, the authors of [51] reported that fracture of the WC coating on pure tungsten did not depend on the scratching rate, since its enhancing did not change shapes of forming grooves, but caused earlier initiation of cracks and their higher density, as well as raising the AE signals.
The next nuance that should be discussed is the identification of the factor that had the greatest impact on the increase in the FSE levels of the 321 steel after HIPIB irradiation. As indicated above, the morphology and roughness of the surface layers (Figures 2 and 3, respectively), as well as their phase composition ( Figure 4 and Table 2) and the mechanical properties ( Figure 5), changed simultaneously.
According to previously published data, there was no clear correlation between surface roughness and coating adhesion. For example, surface roughness minimization was recommended by the authors of [28,53,57,92]. However, there were other suggestions, such as high-energy surface processing, that caused its raising. As an instance, ion irradiation cleared the substrate surfaces from most contaminants and broke chains of surface bonds, exposing active centers on which the deposited coatings were covalently bound [66]. The substrate roughening had to result in extremely effective adhesion combined with ion irradiation that activated the interface chemical activity. On some inhomogeneous substrates, ion bombardment caused the formation of rough surfaces, which strengthened the interface formed due to its fracture toughness, as well as its increased clean contact area. At the same time, plasma pre-treatment allowed surfaces of aluminum, copper and AlN ceramics to possess hydrophilic behavior due to removing oxides and impurities, which could enhance the surface reactivity [64]. The adhesion strength of the Al/AlN and Cu/AlN combinations increased with rising both plasma power and treatment duration. However, the adhesion of titanium and porcelain was not improved by either sandblasting or electric discharge surface processing [61].
Changes in the mechanical properties of the surface layers also had a dual effect. For instance, cracks in ceramics adjacent to substrates typically propagated towards the metal and passed through the interface if it was poorly bonded. In other cases, cracks propagated into a brittle reaction layer or an embrittled intermediate one if the interface was strong [29]. As an example, preliminary sandblasting of the 17-4PH steel substrates before deposition of the Cr 3 C 2 -NiCr coatings was ambiguous [58]. Despite the positive effect on adhesion due to a change in roughness, the microstructure and mechanical properties of the surface layer were also varied. An increase in hardness at the interface by more than 20% was observed compared to that in the bulk steel. The influence of the deposition parameters was close to zero in this case. The potential adverse effects of sandblasting were confirmed by a local investigation of an embedded residual alumina particle. An ultra-fine grain microstructure was found associated with a very local increase in hardness (40% higher than that for the bulk steel). Cracks initiated from the particle edges and fractures mainly occurred through the surface layer but not at the coating/substrate interface [58]. It should be noted that preliminary sandblasting also resulted in the cracking of glass-ceramic coatings deposited on the 441 stainless steel substrates [59]. Typically, the interface fracture did not occur in the case of high-strength ductile metals firmly bonded to ceramics [29]. In this case, the crack could propagate in the ceramic coatings near the interface, causing plastic metal strains, or cross the interface repeatedly. The results depended on the specific materials, load conditions and interface strength.
According to the concept proposed by the authors of [29], it was difficult to characterize the fracture energy and resistance of the metal/ceramic interfaces because of their different elastic constants and complex stress states. The work of adhesion actually accounted for only a few percent of the total fracture energy of most metal/ceramic interfaces, and the remaining energy was largely dissipated by plastic metal strains. The fracture toughness of the metal/ceramic interfaces could be maximized by forming a strong bond with a high-strength metal that absorbed energy. This concept was confirmed by the authors [67], who investigated the adhesion of the CrN x coatings on various types of steel and polycrystalline copper. In the steel cases, two-layer structures formed within the coating, which was more pronounced if the substrate was implanted with metal ions prior to deposition. Four-point bending tests showed plastic strains of the CrN x coating deposited on the soft copper substrate, but its brittle delamination from the hard tool steel was observed [67]. Accordingly, the decrease in microhardness and the increase in the elastic modulus of the surface layers due to the change in their phase composition could also have a positive effect on the scratch test results in these studies.
It should also be considered that the best deposition mode estimated from the three investigated ones on the silicon substrates was probably not the optimum for the 321 steel since there could be two types of residual stresses in the coatings [92]. The first type arose from discontinuities formed in the deposition process (their levels could be reduced by an increase in the substrate temperature). Another reason for residual stresses was associated with a mismatch between the thermal expansion coefficients of the substrate and the coating. Levels of such stresses depended on their ratio, as well as the substrate dimensions. Since these characteristics varied for the silicon and 321 steel substrates, resulting residual stresses were also different. In this regard, additional studies (both theoretical and experimental) are required for a more accurate understanding of the essence of these variations. This is one of the most urgent research areas since, for example, residual stresses were the dominant factor affecting the bond strength of the Si 3 N 4 -WC/TiC/TaC coatings and nickel substrates [54]. In this case, cracks were initiated near the interface edges and then propagated into the ceramics.
Finally, it could be concluded on the basis of the foregoing that the obtained results were rather primary and needed additional justification both from the point of view of the root causes of this phenomenon and the possibility of their extrapolation to other coatings/substrate systems, as well as irradiation modes and coating methods. Moreover, additional research methods are required for assessing both mechanical and tribological properties, as well as the microstructure of the modified layers, despite their high labor intensity and cost. The authors plan to report such results in their following papers.

Conclusions
The obtained results enabled the following conclusions to be drawn:

1.
HIPIB irradiation of the 321 steel surfaces caused partial melting of irregularities at the energy densities of 1.0 and 1.5 J/cm 2 but full recrystallization of the surface layer at 2.0 J/cm 2 . As a result, roughness Ra of the modified surfaces increased by three times (from 20 up to 60 nm).

2.
With an increase in the energy density, the α-Fe phase content in the surface layers reduced with the corresponding increase in the γ-Fe amount, which was the only one found at 2.0 J/cm 2 . The γ-Fe lattice parameter decreased with an increase in the energy density up to 1.0 J/cm 2 but slightly enhanced (up to 0.3587 nm) at 1.5 and 2.0 J/cm 2 . Microstrains rose much faster in the α-Fe phase than in the γ-Fe one.

3.
After HIPIB irradiation, the microhardness of the surface layers increased slightly at the minimum studied energy density of 0.6 J/cm 2 and then decreased by 25% compared to the initial values, at the maximum level of 2.0 J/cm 2 . The elastic modulus values showed directly opposite dependences due to the reduced work-hardening effect. 4.
The FSE values and their polar components monotonically increased over the entire studied range of the energy densities, which resulted in the chemical activity of the surfaces. However, suppression of the dispersed component was observed that decreased both residual stresses at the coating/substrate interfaces and Van der Waals forces. 5.
During the scratch tests, the fracture threshold was 3 N for the AlN coating on the initial 321 steel substrate. HIPIB irradiation caused a shift at the beginning of the coating fracture up to 5 and 6 N at the energy densities of 2.0 and 1.0 J/cm 2 , respectively. The lower fracture threshold value in the case of 2.0 J/cm 2 could be associated with the decrease in the substrate hardness and, as a consequence, its greater strains under the load.