Improved Adhesion and Tribological Properties of AlTiN-TiSiN Coatings Deposited by DCMS and HiPIMS on Nitrided Tool Steels

: Hard coatings, such as AlTiN-TiSiN, deposited by Physical Vapor Deposition (PVD) techniques are widely used in industrial applications to protect and increase the lifetime of industrial components, such as cutting tools, dies, and forming tools. Despite their great properties, such as high hardness and wear and oxidation resistance, they are limited in cases of severe conditions due to the poor adhesion between the coating and the substrate. Duplex treatments have commonly been used to improve the adhesive properties of PVD coatings, especially those of the cathodic arc evaporation type. The purpose of this study is to achieve coatings with the good properties of the Magnetron Sputtering processes but with higher adhesion than that achieved with these techniques, thus achieving coatings that can be used under the most severe conditions. In this work, an AlTiN-TiSiN coating was deposited by a combination of DC Magnetron Sputtering (DCMS) and High-Power Impulse Magnetron Sputtering (HiPIMS) after a gas nitriding pretreatment on 1.2379 and Vanadis 4 tool steels. Mechanical (ultra-microhardness and scratch tests) and tribological tests were carried out to study the improvement in the properties of the coating. Duplex-treated samples showed improved adhesion between the coating and the substrate, with second critical load (Lc2) values greater than 100 N. Furthermore, they showed great toughness and wear resistance. These results show that this type of coating technique could be used in the most extreme applications and that they can compete with other techniques and coatings that to date they have not been able to compete with.


Introduction
Hard coatings have been widely used to protect and increase the lifetime of industrial components, such as cutting tools, dies, and forming tools. Their excellent properties, such as high hardness, toughness, good tribological properties, and wear and oxidation resistance, make them ideal for applications in which industrial components are under high and constant loads [1][2][3]. This family contains different coatings (e.g., AlTiN, TiN, CrN, and AlCrN), and the addition of metallic or non-metallic elements, such as Ti, Si, Al, and C, and the deposition parameters have a strong influence on their physical, chemical, and mechanical properties [4][5][6]. TiN-based films are very hard and have good tribological properties; however, they present problems at high temperatures due to their poor thermal The use of duplex PVD coatings for improving the adhesion of hard coatings has been reported in many studies [48][49][50][51]. Other novel techniques, such as HiPIMS with positive pulses [52] and closed-field unbalanced magnetron sputtering ion plating (CFUMSIP) [53], have also been used to improve the adhesion of PVD coatings in steel substrates. As reported in [49,50], duplex treatments based on plasma nitriding and magnetron sputtering showed critical scratch loads (Lc3) of about 60 N and 50 N, respectively. On the other hand, high-velocity oxygen fuel (HVOF) spraying and PVD cathodic vacuum arc duplex treatments showed second critical scratch loads (Lc2) of around 55 N on steel substrates [54] and nitrocarburized and PVD cathodic multi-arc ion plating duplex treatments showed Lc2 values of about 50-57 N [4].
The aim of this study is to improve the adhesion and tribological properties of duplex treatments formed by gas nitriding and hard coatings deposited by magnetron sputtering techniques. To achieve the objective, AlTiN-TiSiN coatings were deposited by a combination of DC Magnetron Sputtering and HiPIMS after a gas nitriding pretreatment on 1.2379 and Vanadis 4 powder metallurgical (PM) tool steels. These materials were chosen due to their good properties and their great industrial applicability. This way, hard coatings with a good surface finish and the superior mechanical and tribological properties of magnetron sputtering techniques together with an improved adhesion that can compete with cathodic arc duplex treatments, which normally present better adhesion to the substrate, were studied. The adhesion of the coating to the substrate is essential to extend the useful life of tools, especially under conditions of extreme loads. So, the novelty of and motivation for this work is the improvement of the adhesion of the coatings using the HiPIMS technique and combining it with the gas nitriding pretreatment, thus improving the adhesion normally obtained with magnetron sputtering techniques and obtaining coatings that can compete with those deposited by the cathodic arc technique but that have the advantages of the magnetron sputtering techniques, e.g., a better surface finish.

Reference Substrates
The 1.2379 and Vanadis 4 (a brand-specific steel) tool steels with flat geometries and 30 mm diameters were used as reference substrates. The samples were polished (Ra < 0.2 µm) and thoroughly cleaned before the surface treatments. The latter process consisted of ultrasonic washing with alkaline detergents (1% Tickopurr R33) followed by rinsing with deionized water and isopropanol cleaning and finally air drying, respectively. The chemical composition of the materials is shown in Table 1. Table 1. Chemical composition of the steels studied in this work (wt.%). Both steels have extraordinary mechanical properties and are widely used in applications such as cutting, stamping, and forming tools, where the materials are subjected to high repetitive stresses; therefore, they must be hard and resistant to compression and they must have high wear resistance [55]. Vanadis 4 steel is of a higher quality than conventional steels since it is a powder metallurgical steel, which makes it have good dimensional stability and a finer carbide distribution than conventional steels, and it is free of segregations [55]. All these characteristics make it especially suitable for PVD treatments.

Film Deposition Techniques
The gas nitrocarburizing process was carried out at 500-520 • C in an atmosphere rich in ammonia gas (NH 3 ) and nitrogen (N 2 ) and with the pulsed addition of carbon dioxide (CO 2 ). The total process time (heating up in a nonreactive atmosphere, nitrocarburizing, and cooling down) was 22 h. The nitrocarburizing time was 12 h.
After nitriding, the samples were sand blasted and polished to avoid a white layer on top of the surface. The white layer (also called the compound layer) shows low toughness, poor adhesion to PVD coatings, and a lower oxidation temperature than the PVD layer.
The PVD coating depositions were carried out in a CC 800/9ML HiPIMS system (CemeCon, Horseheads, New York). A pretreatment was carried out on the substrates before depositing the coatings. The sequence of operations and parameters used in the PVD deposition process are summarized in Table 2.

Thickness, Structural Properties, and Profile Composition
The chemical composition profiles and the resultant thickness of the layers were analyzed by glow discharge optical emission spectrometry (GD-OES). The instrument employed was a JOBIN YVON 100000RF GD-OES system (HORIBA Instruments, Kyoto, Japan) [56]. To confirm the obtained results on thickness in the previous tests, a CSM Calotest instrument (CSM Instruments, Needham, MA, USA) was used. It was equipped with a 30 mm stainless-steel ball and a superfine (0.25 µm) diamond water suspension was used as the abrasive medium.
Finally, SEM cross-section images were obtained using a HITACHI S4800 cold cathode system (HITACHI, Tokyo, Japan).

Mechanical and Tribological Tests
In order to obtain information about the adhesion between the substrate and the coating, adhesion tests were carried out with a CSM REVETEST Scratch tester (CSM instruments, Needham, MA, USA) equipped with a diamond Rockwell indenter with a tip radius of 200 µm.
In this test, the indenter moves forward at a given speed while the load is progressively increasing at a constant rate. The indenter then introduces energy into the system and, at a given load, some cracks or delamination are produced to dissipate some of the energy (events). The different loads at which the different modes of failure or events occur are registered with the help of various signals such as the penetration of the indenter within the substrate, the acoustic emission, and the variation in the coefficient of friction. Finally, the optical microscope allows us to obtain an image of the spot where the event occurs and helps to define the different critical loads.
The first critical load (Lc1) corresponds to the first cohesive failure observed in the coating, usually perpendicular cracks on the borders of the track; the second critical load (Lc2) is registered when the first adhesive failure is appreciated, i.e., chipping on the edges of the track; and the third critical load (Lc3) corresponds to the total delamination of the coating or when a critical defect is observed in the substrate. The second critical load (Lc2) is commonly defined as the adhesion strength due to the fact that it is considered a sign of the coating adhesion's failure [57].
The test parameters for the samples coated only with the AlTiN/TiSiN PVD layers were a load rate of 100 N/min, up to a final load of 100 N, a speed of 9.58 mm/min, and a total test length of 10 mm. For the duplex treatment specimens, a final load of 200 N and a total length of 20 mm were applied while keeping the same loading rate and speed.
A Microtest MT series instrument (Microtest S.A., Madrid, Spain) in a ball-on-disk configuration was used for the tribomechanical tests. For the tests, 6 mm alumina balls with a maximum roughness Ra max = 0.050 µm and hardness values of around 1650 HV were used, and the different coated and uncoated samples were used as disks. The test parameters for the coated samples were as follows: a load of 20 N, 200 rpm, and 20,000 cycles, and the test was repeated two times for each sample at 10 mm and 12 mm track radii. For the uncoated samples, these parameters were too extreme; so, the tests were carried out at 200 rpm for 6000 cycles and a load of 10 N and 5 N (for the 8 mm and 6 mm track radii, respectively). Confocal microscopy using a Confocal Smart microscope (Sensofar, Barcelona, Spain) and optical microscopy were used to measure the corresponding wear tracks. The volume loss and the wear coefficient were determined in two ways: straight from the confocal measurements of volume loss and according to the ASTM G99 standard [58]. This standard defines an equation that assumes that there is not significant ball wear to calculate the volume loss in the disk, and it is necessary to measure the width of the wear track. With that measurement and the information about the wear track and ball radii, the equation is applied and the value of the volume loss is obtained.
Ultra-microhardness measurements with the aim of assessing the mechanical properties were performed using a FISHERSCOPE H100 VP/X-Y micro durometer (Helmut Fischer Instruments, Sindelfingen, Germany) fitted with a Vickers diamond pyramid. Loads ranging from 200 µN to 2 N can be applied with the indenter against the surface of the samples while the indenter position and the applied load are monitored to plot a curve that reflects the material's response to the load. In this study, two different maximum loads were used to perform the tests: 500 mN and 1000 mN.

Thickness, Structural Properties, and Profile Composition
GD-OES was used to analyze the chemical composition profiles and the resultant thickness of the coatings, and the results are shown in Figure 1. The sample coated only with the PVD coating showed an outermost layer of about 1 µm containing (as a percentage of atomic concentration) approx. 60% N, 30% Ti, 5% Si, and 2% Al. This layer is followed by another layer of around 2 µm with concentrations in percentages of about 50% N, 30% Al, and 15% Ti. After these two layers, an anchoring layer-aiming to improve the adhesion to the substrate-of around 1 µm of chromium can also be observed. As the PVD coating is the same in both the PVD coated sample and the duplex samples, the first two layers are identical. The main difference can be seen in the N percentage of the duplex samples after the first two layers as can be seen in Figure 1. While in the PVD sample the percentage of N decreases to nearly 0%, in the duplex samples this percentage remains at 10% and 7% in the 1.2379 and Vanadis 4 duplex samples, respectively. This is due to the nitriding process of the duplex samples, which makes them have a thin film of iron nitride. the first two layers are identical. The main difference can be seen in the N percentage of the duplex samples after the first two layers as can be seen in Figure 1. While in the PVD sample the percentage of N decreases to nearly 0%, in the duplex samples this percentage remains at 10% and 7% in the 1.2379 and Vanadis 4 duplex samples, respectively. This is due to the nitriding process of the duplex samples, which makes them have a thin film of iron nitride.
As has already been mentioned, this layer is meant to increase the hardness, to increase the load support effect of the substrate, and to eliminate the stress concentration at the substrate-coating interface, thus improving the adhesion between the hard coating and the substrate. Calotest measurements that were carried out to determine the resultant thickness of the coatings showed different circumferences ( Figure 2) corresponding to different depth levels that were caused by the coatings' layers with different compositions rolling the ball with the abrasives during the tests. The thickness was calculated using the relationship between these dimensions. As can be observed in Table 3, the total thickness of each coat-  As has already been mentioned, this layer is meant to increase the hardness, to increase the load support effect of the substrate, and to eliminate the stress concentration at the substrate-coating interface, thus improving the adhesion between the hard coating and the substrate.
Calotest measurements that were carried out to determine the resultant thickness of the coatings showed different circumferences ( Figure 2) corresponding to different depth levels that were caused by the coatings' layers with different compositions rolling the ball with the abrasives during the tests. The thickness was calculated using the relationship between these dimensions. As can be observed in Table 3, the total thickness of each coating was about 5 µm, with an outer layer of nearly 1.65 µm. A middle layer of 2.70 µm was observed in the case of the 1.2379 + PVD sample, while in the duplex samples this layer is a little bit thicker (2.90 µm). In the three samples, an anchoring layer of about 0.60 µm can also be observed. These results are consistent with those reported in the GD-OES tests, although it is important to note that the measurements in this case are approximate, and therefore it can be observed that there are slight differences between the measured values. Calotest measurements that were carried out to determine the resultant thickness of the coatings showed different circumferences ( Figure 2) corresponding to different depth levels that were caused by the coatings' layers with different compositions rolling the ball with the abrasives during the tests. The thickness was calculated using the relationship between these dimensions. As can be observed in Table 3, the total thickness of each coating was about 5 µm, with an outer layer of nearly 1.65 µm. A middle layer of 2.70 µm was observed in the case of the 1.2379 + PVD sample, while in the duplex samples this layer is a little bit thicker (2.90 µm). In the three samples, an anchoring layer of about 0.60 µm can also be observed. These results are consistent with those reported in the GD-OES tests, although it is important to note that the measurements in this case are approximate, and therefore it can be observed that there are slight differences between the measured values.     Figure 4 illustrates an example of the ultra-microhardness load-displacement curves for the 1.2379 + PVD sample with a maximum load of 1000 mN. With these curves, it is possible to represent the material's response to the load. From these curves, it is possible to determine the Universal Hardness (HU), elastic modulus (Er), percentage of the work of elastic deformation (% We), maximum hardness, and plastic hardness of the samples. The values of these parameters obtained for each sample and the relations between hardness and elastic modulus, HU/Er and HU 3 /Er 2 , for the tests carried out with a maximum load of 500 and 1000 mN are summarized in Tables 4 and 5, respectively. The ratio H/E represents the elastic strain to failure and H 3 /E 2 the resistance to plastic deformation; thus, it is expected that an improvement in the elastic recovery of the coating would imply an increase in the H 3 /E 2 ratio [59][60][61]. Because of this, this parameter is used to determine the toughness and the wear resistance of the coating [60,62]. Different studies have concluded that these ratios are more relevant than just the hardness to determine the wear resistance of a material and they have highlighted the influence of toughness in tribological applications. Moreover, the relation between H 3 /E 2 and wear resistance has been studied, showing that the resistance to various forms of wear is related to the H 3 /E 2 ratio [59,60,62].     Figure 4 illustrates an example of the ultra-microhardness load-displacement curves for the 1.2379 + PVD sample with a maximum load of 1000 mN. With these curves, it is possible to represent the material's response to the load. From these curves, it is possible to determine the Universal Hardness (HU), elastic modulus (Er), percentage of the work of elastic deformation (% We), maximum hardness, and plastic hardness of the samples. The values of these parameters obtained for each sample and the relations between hardness and elastic modulus, HU/Er and HU 3 /Er 2 , for the tests carried out with a maximum load of 500 and 1000 mN are summarized in Tables 4 and 5, respectively. The ratio H/E represents the elastic strain to failure and H 3 /E 2 the resistance to plastic deformation; thus, it is expected that an improvement in the elastic recovery of the coating would imply an increase in the H 3 /E 2 ratio [59][60][61]. Because of this, this parameter is used to determine the toughness and the wear resistance of the coating [60,62]. Different studies have concluded that these ratios are more relevant than just the hardness to determine the wear resistance of a material and they have highlighted the influence of toughness in tribological applications. Moreover, the relation between H 3 /E 2 and wear resistance has been studied, showing that the resistance to various forms of wear is related to the H 3 /E 2 ratio [59,60,62].  Observing the results summarized in Tables 4 and 5, it can be seen how the Universal Hardness (HU) and the elastic modulus (Er) are higher in the coated samples; the highest values were obtained for the duplex coated Vanadis 4 sample. Despite the differences, the values obtained are similar for both duplex coated samples, with HU values of around 9-   Observing the results summarized in Tables 4 and 5, it can be seen how the Universal Hardness (HU) and the elastic modulus (Er) are higher in the coated samples; the highest values were obtained for the duplex coated Vanadis 4 sample. Despite the differences, the values obtained are similar for both duplex coated samples, with HU values of around 9-10 GPa. The elastic modulus, however, is more similar in the case of 1.2379 specimens coated with PVD and duplex treatments, with values of approximately 300 GPa, than in the case of the specimen with Vanadis 4 as a substrate, which showed values of 328-329 GPa. The elastic deformation work percentage (% We) and the ratios between hardness and elastic modulus (HU/Er and HU 3 /Er 2 ) are similar in all of the coated samples, showing greater values than those obtained for the uncoated 1.2379 specimen. If we go into detail, it can be said that the duplex samples showed the highest HU/Er and HU 3 /Er 2 values, and the greatest resistance to plastic deformation was obtained in the case of the Vanadis 4 + Duplex sample.

Adhesion Tests
The adhesion between the coatings and the substrates and their mechanical response were evaluated by scratch testing. The results obtained for each sample are shown in Figure 5. The failure modes along the scratches were observed and the critical loads (Lc1, Lc2, and Lc3) where these failures occurred were spotted through the sudden changes in the evolution of the COF and sharp rises in the AE signal. Pictures of the different spots where the different features are related to either Lc1, Lc2, or Lc3 are also shown in Figure 5.
As was mentioned above, each critical load represents a different failure mechanism (cohesive (Lc1), adhesive (Lc2), or the appearance of the substrate (Lc3)) and with their values it is possible to evaluate the differences in the adhesion properties of the samples. Apart from the adhesion properties, the value of Lc1 and the difference between Lc2 and Lc1 (Lc2-Lc1) can provide information about the toughness of the coating. Zhang et al. [63] proposed a new parameter to evaluate the toughness of the coating: the scratch crack propagation resistance (CPR) as shown in Formula (1). Higher CPR values are related to a greater capacity of the coating to resist crack propagations [64].
The adhesion between the coatings and the substrates and their mechanical response were evaluated by scratch testing. The results obtained for each sample are shown in Figure 5. The failure modes along the scratches were observed and the critical loads (Lc1, Lc2, and Lc3) where these failures occurred were spotted through the sudden changes in the evolution of the COF and sharp rises in the AE signal. Pictures of the different spots where the different features are related to either Lc1, Lc2, or Lc3 are also shown in Figure  5. As was mentioned above, each critical load represents a different failure mechanism (cohesive (Lc1), adhesive (Lc2), or the appearance of the substrate (Lc3)) and with their values it is possible to evaluate the differences in the adhesion properties of the samples. Apart from the adhesion properties, the value of Lc1 and the difference between Lc2 and Lc1 (Lc2-Lc1) can provide information about the toughness of the coating. Zhang et al. [63] proposed a new parameter to evaluate the toughness of the coating: the scratch crack propagation resistance (CPR) as shown in Formula (1). Higher CPR values are related to a greater capacity of the coating to resist crack propagations [64].  Table 6 summarizes the results of the critical loads for each sample. The results show very promising values since high critical values were reached, especially for duplextreated specimens. As can be observed, the samples coated with duplex presented higher critical load values (Lc1, Lc2, and Lc3) than the sample coated only with PVD in all cases. Furthermore, although the two duplex-coated samples presented similar Lc2 and Lc3 values, it should be noted that the best results were obtained in both cases for the Vanadis 4 sample. The adhesion values obtained in this study are higher than those reported in other studies where conventional PVD techniques were used. References [65,66] reported Lc2 values of around 80-90 N for high-speed steel substrates coated by cathodic arc evaporation PVD (AlTiN and AlCrN, respectively), while [67] presented Lc2 values of 30 N for multilayer Ti/TiAlN/TiAlCN coatings deposited on tool steels by magnetron sputtering. Many other studies showed Lc2 values from 50 N to 80 N for duplex (nitriding + PVD) hard coatings deposited on tool steels [1,[68][69][70]. Furthermore, the values obtained in this study are similar, and even higher in some cases, to those reported for novel PVD techniques such as hybrid PVD [25,71] and closed-field unbalanced magnetron sputtering ion plating (CFUMSIP) [53,61], which presented adhesion strength values up to 116 N and 143 N, respectively. These values can even be compared with those obtained in studies using carbides as substrates, such as [22,23], that showed Lc2 values of approximately 100 N. Finally, it can be said that these results show how this duplex treatment improves the adhesion properties of the conventional PVD magnetron sputtering techniques, making them capable of competing with other types of techniques that to date they have not been able to compete with. The values of the first critical load (Lc1), the adhesion strength (Lc2), and the CPR of each sample are presented in Figure 6. As can be seen, the toughness of the coatings, represented by the CPR value, is much higher in the case of the duplex samples. The 1.2379 + PVD specimen presented a value of 1334 N 2 , while the sample with the same substrate but coated with the duplex treatment showed a CPR value of 2108 N 2 . Furthermore, the toughness of the duplex-treated Vanadis 4 sample was even greater than the 1.2379 duplex's, with a CPR value of 2928 N 2 . These values indicate the great capacity of the coatings to resist crack propagation, and are higher than those reported in [53], which reported CPR values of approximately 1468 N 2 for NiCrN coatings deposited on M2 tool steel, and in [66], which showed CPR values of around 1900 N 2 for AlCrN coatings deposited on HS6-5-2 steel. However, similar results were reported in [61], with CPR values between 1500 and 4300 N 2 depending on the Ni content of the CrAlNiN coatings deposited onto AISI M2 high-speed tool steel substrates. The CPR values, which are related to the toughness of the coating, present a similar trend to the values obtained for the relation between HU 3 and Er 2 , which is also related to the toughness. The comparison of the trends of the values obtained for these two parameters is shown in Figure 7 where it can be seen how the highest values in both cases were obtained for the Vanadis 4 + duplex sample.

Friction and Wear Tests
The friction and wear properties of the coatings as well as those of the reference substrate were studied by a pin-on-disc tribometer. The experimental results on the coefficients of friction (COFs) are shown in Figure 8. COF values of around 0.65 were shown by all specimens. The fluctuations in the COF curves are inherent to the test itself during the run-in period, while, during steady state, the appearance of different tribooxidation bodies is the cause of this phenomenon. The fact that all the samples have nearly the same slightly high COF value can be explained because the coating used in this study is not a low-COF coating but rather a hard coating, and the top layer is the same for all the samples. However, as will be seen later, these coatings improve the wear resistance of the uncoated samples. The COF values are similar to those reported by Warcholinski et al. [66], where COF values between 0.62 and 0.68 were reported for AlCrN coatings deposited using Cathodic Arc Evaporation on HS6-5-2 steel substrates, and in [49], where values of around 0.5-0.75 were reported for duplex PVD nanocomposite TiAlN coatings deposited on different steels. References [67,68], however, showed lower COF values, around 0.3-0.4, for Ti/TiAlN/TiAlCN multilayer coatings and graded TiAlN coatings deposited on tool steels, respectively. but coated with the duplex treatment showed a CPR value of 2108 N 2 . Furthermore, the toughness of the duplex-treated Vanadis 4 sample was even greater than the 1.2379 duplex's, with a CPR value of 2928 N 2 . These values indicate the great capacity of the coatings to resist crack propagation, and are higher than those reported in [53], which reported CPR values of approximately 1468 N 2 for NiCrN coatings deposited on M2 tool steel, and in [66], which showed CPR values of around 1900 N 2 for AlCrN coatings deposited on HS6-5-2 steel. However, similar results were reported in [61], with CPR values between 1500 and 4300 N 2 depending on the Ni content of the CrAlNiN coatings deposited onto AISI M2 high-speed tool steel substrates. The CPR values, which are related to the toughness of the coating, present a similar trend to the values obtained for the relation between HU 3 and Er 2 , which is also related to the toughness. The comparison of the trends of the values obtained for these two parameters is shown in Figure 7 where it can be seen how the highest values in both cases were obtained for the Vanadis 4 + duplex sample.

Friction and Wear Tests
The friction and wear properties of the coatings as well as those of the reference substrate were studied by a pin-on-disc tribometer. The experimental results on the coefficients of friction (COFs) are shown in Figure 8. COF values of around 0.65 were shown by all specimens. The fluctuations in the COF curves are inherent to the test itself during the run-in period, while, during steady state, the appearance of different tribo-oxidation bodies is the cause of this phenomenon. The fact that all the samples have nearly the same slightly high COF value can be explained because the coating used in this study is not a low-COF coating but rather a hard coating, and the top layer is the same for all the samples. However, as will be seen later, these coatings improve the wear resistance of the uncoated samples. The COF values are similar to those reported by Warcholinski et al. [66], where COF values between 0.62 and 0.68 were reported for AlCrN coatings deposited using Cathodic Arc Evaporation on HS6-5-2 steel substrates, and in [49], where values of around 0.5-0.75 were reported for duplex PVD nanocomposite TiAlN coatings deposited on different steels. References [67] and [68], however, showed lower COF values, around 0.3-0.4, for Ti/TiAlN/TiAlCN multilayer coatings and graded TiAlN coatings de- The width of the wear tracks obtained by optical microscopy and confocal microscopy is shown in Figure 9, where the differences are evident. The width of the wear tracks is between 560 and 906 µm for the uncoated sample, but it must be remembered that the tests on this sample were carried out under less-demanding conditions. The PVD-coated sample, instead, presents values of around 1156-1564 µm, while the duplex-coated samples are approximately 840-1120 µm. As expected, the coated samples improved the wear resistance of the substrate, and the duplex samples showed lower values of the width of the wear track than the PVD-coated sample. Moreover, the wear mechanisms present in each coated sample were very similar. Observing the images shown in Figure 9, the three coated samples show very little abrasive wear, with few microscratches, while the non-nitrided PVD-coated sample shows a little bit more abrasive wear, which is an indicator of good wear resistance. The volume loss and the wear coefficient were evaluated using two different methods: the standard ASTM G99 and directly using the results obtained by confocal microscopy. For the first method, the width of the wear track was measured by optical The volume loss and the wear coefficient were evaluated using two different methods: the standard ASTM G99 and directly using the results obtained by confocal microscopy. For the first method, the width of the wear track was measured by optical microscopy, and then we followed the standard ASTM G99 to determine the value of the volume loss for the entire track. Afterward, the calculated value was used to determine the wear coefficient through a formula that relates it to the applied load in the test and the total sliding distance, thus obtaining a normalized value. In the second method, the volume loss of the focused part of the wear track is calculated directly using a confocal microscope and the Sensoview program, and then that value is extrapolated for the entire wear track using the equation shown in Formula (2). Then, the same equation used in the first method is used to calculate the wear coefficient.
V loss con f ocal (m 3 ) wear track length (m) Table 7 and Figures 10 and 11 show a summary of the results on the friction coefficient, volume loss, and wear coefficient calculated with both the standard ASTM G99 and confocal microscopy. The values shown are the average values of the results obtained for the tests that were carried out with different track radii. A difference of approximately an order of magnitude can be observed between the values by both methods. In the calculations carried out following the standard ASTM G99, it was assumed that the shape of the wear track is that of a perfect sphere, while by confocal microscopy the real shape of the wear track was used for calculating the volume loss. Anyway, both methods show the same trend, and the wear coefficient is lower for the duplex-coated samples in both cases, which indicates a better wear resistance. Values of (2.49 ± 1.11) × 10 −4 and (1.30 ± 0.18) × 10 −5 mm 3 /Nm were obtained for the uncoated samples using the ASTM G99 and the confocal measurements, respectively. For the 1.2379 + PVD sample, nonetheless, the values were lower: (2.23 ± 0.133) × 10 −4 mm 3 /Nm for the ASTM G99 measurements and (4.10 ± 0.63) × 10 −6 mm 3 /Nm for the confocal ones. As mentioned above, the lowest wear coefficient values were presented by the duplex samples, with values of around 6.35 × 10 −5 mm 3 /Nm in the ASTM G99 results and between 1.75 × 10 −6 and 1.96 × 10 −6 mm 3 /Nm in the confocal measurements. As expected, the duplex treatment improved both the adhesion of the coatings and their wear resistance. Different studies showed that the nitrides and carbonitrides formed during the nitriding process contribute to the improvement of the wear resistance [72,73]. These wear coefficient results are similar to those reported in reference [53], ranging between 0.23 × 10 −6 and 9.5 × 10 −6 mm 3 /Nm for NiCrN coatings deposited on M2 tool steel, reference [67], showing values of 3.7624 × 10 −15 m 3 /Nm, and reference [49], with values of around 12 × 10 −16 -400 × 10 −16 m 3 /Nm. On the other hand, lower wear coefficient values were reported in [66,74], with wear coefficient values of around 2.5 × 10 −7 m 3 /Nm for Cr/CrN/CrTiN and AlCrN coatings deposited on tool steels, respectively. Figures 10 and 11 present the wear coefficient values for each sample calculated by both methods (ASTM G99 ( Figure 10) and confocal microscopy ( Figure 11)). As can be seen, they show good repeatability in the tribological test. Additionally, a difference of about an order of magnitude between the results obtained by the different methods is evident in these graphs.

Discussion
Once all the tests have been carried out on all the samples, it is possible to comment on and relate some of the results to discuss the properties of the different coatings. First of all, GD-OES spectrometry showed the chemical composition profiles and the resultant thickness of the coatings, where it was possible to observe that the duplex coatings had higher nitrogen content due to the nitriding process. The compressive stresses formed in the nitrided region increase the load bearing capacity, plastic deformation, and the adhesion of the PVD coatings. The PVD coating, meanwhile, improved the hardness and tribological properties of the substrates. Related to this, the differences observed in the ultramicrohardness results were very small, since the top layer is the same in all the coated specimens. Thus, values of around 8.5-10 GPa were obtained for all the coated specimens.

Discussion
Once all the tests have been carried out on all the samples, it is possible to comment on and relate some of the results to discuss the properties of the different coatings. First of all, GD-OES spectrometry showed the chemical composition profiles and the resultant thickness of the coatings, where it was possible to observe that the duplex coatings had higher nitrogen content due to the nitriding process. The compressive stresses formed in the nitrided region increase the load bearing capacity, plastic deformation, and the adhesion of the PVD coatings. The PVD coating, meanwhile, improved the hardness and tribological properties of the substrates. Related to this, the differences observed in the ultramicrohardness results were very small, since the top layer is the same in all the coated specimens. Thus, values of around 8.5-10 GPa were obtained for all the coated specimens.

Discussion
Once all the tests have been carried out on all the samples, it is possible to comment on and relate some of the results to discuss the properties of the different coatings. First of all, GD-OES spectrometry showed the chemical composition profiles and the resultant thickness of the coatings, where it was possible to observe that the duplex coatings had higher nitrogen content due to the nitriding process. The compressive stresses formed in the nitrided region increase the load bearing capacity, plastic deformation, and the adhesion of the PVD coatings. The PVD coating, meanwhile, improved the hardness and tribological properties of the substrates. Related to this, the differences observed in the ultra-microhardness results were very small, since the top layer is the same in all the coated specimens. Thus, values of around 8.5-10 GPa were obtained for all the coated specimens.
It should also be noted that in this study the toughness of the coatings was estimated by two different parameters: the relation between hardness and elastic modulus (HU 3 /Er 2 ), which is related to the resistance to plastic deformation, and the CPR value, which is considered to represent the scratch crack propagation resistance by some authors. Higher H 3 /E 2 and CPR values suppose a greater capacity of the coatings to resist plastic deformation and crack propagation, respectively. As shown in Figure 8, the duplex-treated samples showed higher values than the non-nitrided sample for both parameters. The CPR results are especially good, with values up to 2928 N 2 for the Vanadis 4 duplex-treated sample. On the other hand, as already mentioned, the toughness of the coating has an influence on the tribological properties, so that a greater toughness indicates a greater wear resistance. This correlation can clearly be seen in this study, since the duplex-treated samples showed higher toughness and greater wear resistance than the non-nitrided samples.
The slight discrepancies of approximately an order of magnitude that were observed in the wear coefficient measurements with both different methods-following the ASTM G99 standard and from the confocal microscope directly-may be associated with the shape of the wear tracks. During the pin-on-disc tests, the alumina balls suffer a small amount of wear, so they do not have a perfect spherical shape; therefore, the shape of the wear tracks is not that of a perfect sphere either. The ASTM G99 standard defines the calculations of the volume loss taking the shape of the wear track as that caused by a perfect sphere, while the measurements made by confocal microscopy consider its real shape. This could be the reason why the values obtained by confocal measurements are lower but more realistic.
The objective of this study was focused on improving the adhesion properties of the coatings and, as seen in the results, nitriding the samples led to a clear improvement in their adhesion strength. This is demonstrated by the critical load (Lc1, Lc2, and Lc3) results obtained for each sample, where the duplex samples showed higher values than the non-nitrided samples for all samples. Comparing the results of this study with those observed in other studies, it can be seen that the critical load values obtained in the scratch tests are higher than those reported in studies that used cathodic arc techniques [65,66] or duplex treatments [68][69][70]. Furthermore, the adhesion properties observed in this work are even higher than those reported in studies that use novel PVD techniques such as hybrid PVD or closed field unbalanced magnetron sputtering ion plating (CFUMSIP) [25,61,71]. Thereby, it is possible to highlight the improvement in adhesion achieved in this study. As mentioned above, the duplex treatment also improves other properties, such as wear resistance and toughness. The values obtained in the wear tests are similar to those reported in other studies that evaluated the performance of hard coatings deposited on tool steels [49,53,67], and the CPR values, which are related to the toughness of the coating, are similar and even higher than those reported in references [53,61,66].
In summary, combining the nitriding process with the AlTiN-TiSiN coating deposited by HiPIMS and DCMS does not improve only the adhesion between the coating and the substrate, but it also leads to an improvement in other properties, such as toughness and wear resistance. This combination of the nitriding process and the AlTiN-TiSiN coating formed by first depositing a bonding layer by HiPIMS and then other layers by DCMS optimizes the deposition process, achieving coatings with greater properties. The nitrided layer can improve the adhesion and the load bearing capacity of the coatings, whereas the highly ionized plasma of the HiPIMS process allows us to have a higher ion bombardment of the growing film, which translates to denser coatings with improved mechanical properties, providing them with better adhesion to the substrate.
Finally, it should be noted that small differences between the substrates were observed. Comparing the results obtained for the duplex-coated Vanadis 4, which is a powder metallurgical steel, and for the duplex-coated 1.2379, which is a conventional steel, it can be seen that the value of the adhesion strength (Lc2) is higher in the case of the Vanadis 4 sample. Furthermore, the toughness of the coating is also greater for this sample, both for the HU 3 /Er 2 value and for the CPR value. The higher quality of the powder metallurgical steels, where a better dimensional stability, a greater homogeneity of the carbides, and greater mechanical properties can be highlighted, could be the reason for this better behavior. Although it would be necessary to carry out more studies to provide the reason for this better behavior, there are some aspects that could explain it. First of all, a lower strain on the interface between the substrate and the coating due to a more homogeneous distribution of the residual stresses could lead to a greater capacity to withstand the loads transmitted through the coating. On the other hand, the higher vanadium content in powder metallurgical steels could be a reason for their greater yield strength than conventional steel, which allows them to have a better behavior under load.

Conclusions
In this study, nitrided high-quality tool steels were coated with AlTiN-TiSiN coatings deposited by a combination of HiPIMS and DCMS. The aim of this study was to improve the adhesion properties of this type of hard coating deposited by magnetron sputtering PVD techniques. To achieve this objective, the deposition process was optimized by combining the nitriding process with magnetron sputtering techniques. Adhesive and tribological properties of great interest were obtained with this technique, such as great toughness and wear resistance and improved adhesion between the coating and the substrate. Our analysis of the results obtained from the experimental procedure yielded the following conclusions: • The ultra-microhardness of the coatings was about 8.5-10 GPa for all the coated samples, with the higher values obtained for the duplex ones. • Coefficient of friction (COF) values of around 0.65 were shown by all the specimens, something that can be explained because this coating is not a tribological coating but a hard coating. • Duplex-coated samples presented greater toughness than the non-nitrided PVD-coated sample, since the HU 3 /Er 2 ratio, which is associated with the resistance to plastic deformation, and CPR, which is related to the resistance to crack propagation, values are higher for those samples. This is correlated with a greater resistance to wear of those samples, which is a sign that hardness is not the only important parameter in determining the wear behavior of coatings.

•
The coatings improved the wear resistance of the substrate considerably, though the best results were obtained for the duplex-coated samples.

•
The coatings showed a great adhesion to the substrate, with Lc2 values up to 75 N for the PVD-coated sample and 96 N and 109 N for the duplex-treated 1.2379 and Vanadis 4 samples, respectively. This shows that the combination of the nitriding process with the bonding layer deposited by HiPIMS and the DCMS-deposited layers improves the adhesion properties of these coatings.

•
Differences between the different substrates were observed, with greater adhesive and mechanical properties shown by the powder metallurgical steel compared with the conventional steel. This could be because of the higher quality of these steels.
The duplex treatment, which is a combination of the nitriding process with the HiPIMS and DCMS PVD techniques, showed great results, where the improvement in toughness and adhesion can be highlighted. Moreover, these results show that this type of coating technique could be used in the most extreme and demanding applications and competes with other techniques and coatings that to date it has not been able to compete with.