Performance Evaluation of Different Coating Materials in Delamination for Micro-Milling Applications on High-Speed Steel Substrate

The objective of the present work is to carry out analytical and finite element analysis for commonly used coating materials for micro-milling applications on high-speed steel substrate and evaluate the effects of different parameters. Four different coating materials were selected for micro-milling applications: titanium nitride (TiN), diamond-like carbon (DLC), aluminium titanium nitride (AlTiN) and titanium silicon nitride (TiSiN). A 3D finite element model of coating and substrate assembly was developed in Abaqus to find the Hertzian normal stress when subjected to normal load of 4 N, applied with the help of a rigid ball. The radius of the rigid ball was 200 µm. For all the coating materials, the length was 3 mm, the width was 1 mm, and the thickness was 3 µm. For the high-speed steel substrate, the length was 3 mm, the width was 1 mm, and the thickness was 50 µm. Along the length and width, coating and substrate both were divided into 26 equal parts. The deformation behaviour of all the coating materials was considered as linear–elastic and that of the substrate was characterized as elastic–plastic. The maximum normal stress developed in the FEA model was 12,109 MPa. The variation of the FEA result from the analytical result (i.e., 12,435.97 MPa is 2.63%) which is acceptable. This confirms that the FEA model of coating–substrate assembly is acceptable. The results shows that the TiSiN coating shows least plastic equivalent strain in the substrate, which serves the purpose of protecting the substrate from plastic deformation and the TiSiN of 3 micron thickness is the most optimum coating thickness for micro-milling applications.


Introduction
Micro-manufacturing processes have been used extensively in the aerospace, biomedical and defense industries. Presently, photolithography-based manufacturing techniques are used for selective materials such as Ni, Cu, Si and polymers to produce high aspectratio components. Micro-machining processes like micro-milling are able to generate three-dimensional surfaces in ceramics, metals and polymers [1][2][3][4][5][6]. Micro-machining is becoming popular because of its ability to produce three-dimensional parts of different sizes varying from a few micrometers to a few millimeters across various materials [7]. The In the conventional milling process, the work pieces act as isotropic and homogeneous materials, whereas in micro-milling, the smaller grains in the work piece are comparable to the size of the tool. In view of the above, the micro-milling process is very complicated [14]. The chips, which adhere to the tool, block the path at the cutting zone, and this results in an increase in the cutting forces and leads to a catastrophic failure of the tool because of its low flexural stiffness. Moreover, the small size of the micro-milling cutters makes the tip very weak because of its low stiffness value. Diamond-coated cutters are promising because of their ability to improve tool stiffness and tool life [16].
In micro-machining, the uncut chip thickness (h) is less than the cutting-edge radius (re) due to the negative effective rake angle (−α) influencing the ploughing effect in the work piece [5]. Therefore, the ratio of uncut chip thickness to cutting-edge radius is an important parameter in micro-machining [5,8]. A sharper cutting edge is required to remove the smallest amount of undeformed chip thickness [17].
The major limitations of the micro-milling process are unpredictable tool life and premature tool failure, deterioration of the cutting edge and tool wear leading to high friction generation [18][19][20][21][22][23][24][25][26][27][28][29]. It has been reported that coated micro-cutting tools have longer tool life and improved cutting performance [25]. Many researchers have used TiAlN-coated micro tools in cutting tests [19,20,23,24,28,30]. It has also been reported that CrTiAlN-coated micro-end mills provide great advantage in tool wear reduction and smooth surface finish [31,32]. The small size of micro-milling tools makes coating deposition difficult around the cutting edge. The desirable properties of the coatings for micro-machining tools are high hardness, toughness, chemical/erosive and abrasive wear resistance as well as dense and fine microstructure. Coating also provides smooth To prevent tool failure, a few methods have been proposed in previous research work. The first method is using cutting fluids to dissipate heat and provide lubrication. This facilitates the minimization of friction at the tool-workpiece interface and, in turn, reduces tool wear. However, this method is not very successful in micro-milling machining. The main reason is that the application of the cutting fluid to the cutting zone and the toolworkpiece interface is very difficult because of high cutting speeds and the small size of the contact zone [12]. The second method is the use of coatings on the micro-tool to reduce wear. This enhances tool life [1]. However, very little has been reported in toolwear studies of coated micro-tools. The aim of the present work is to study the wear of micro-machining. Therefore, later in the introduction section, a brief description of wear is presented. There are several types of wear phenomena occurring in the field of mechanics, such as adhesive, abrasive, fatigue, corrosive and fretting wear as well as erosive wear by solid particles, fluids and cavitation and electric arc-induced and delamination wear. Wear of non-lubricated metal pairs sliding in a dusty environment may be termed as dry sliding wear and abrasive wear. The classification of wear processes is based on the type of wearing contacts, such as single-phase and multiple-phase.
In micro-machining, smaller tools are used. Researchers have reported that low flexural stiffness and strength causes huge bending of the tool, hampering the cutting process and leading to tool failure [5,13,14]. This is avoided by minimizing the cutting forces below a certain critical value in order to ensure that the uncut chip thickness remains sufficiently small. During the machining of steel, titanium, nickel alloys, etc. the maximum permissible chip thickness is on the order of or less than the cutting-edge radius [14,15].
In the conventional milling process, the work pieces act as isotropic and homogeneous materials, whereas in micro-milling, the smaller grains in the work piece are comparable to the size of the tool. In view of the above, the micro-milling process is very complicated [14]. The chips, which adhere to the tool, block the path at the cutting zone, and this results in an increase in the cutting forces and leads to a catastrophic failure of the tool because of its low flexural stiffness. Moreover, the small size of the micro-milling cutters makes the tip very weak because of its low stiffness value. Diamond-coated cutters are promising because of their ability to improve tool stiffness and tool life [16].
In micro-machining, the uncut chip thickness (h) is less than the cutting-edge radius (re) due to the negative effective rake angle (−α) influencing the ploughing effect in the work piece [5]. Therefore, the ratio of uncut chip thickness to cutting-edge radius is an important parameter in micro-machining [5,8]. A sharper cutting edge is required to remove the smallest amount of undeformed chip thickness [17].
The major limitations of the micro-milling process are unpredictable tool life and premature tool failure, deterioration of the cutting edge and tool wear leading to high friction generation [18][19][20][21][22][23][24][25][26][27][28][29]. It has been reported that coated micro-cutting tools have longer tool life and improved cutting performance [25]. Many researchers have used TiAlN-coated micro tools in cutting tests [19,20,23,24,28,30]. It has also been reported that CrTiAlNcoated micro-end mills provide great advantage in tool wear reduction and smooth surface finish [31,32]. The small size of micro-milling tools makes coating deposition difficult around the cutting edge. The desirable properties of the coatings for micro-machining tools are high hardness, toughness, chemical/erosive and abrasive wear resistance as well as dense and fine microstructure. Coating also provides smooth machined surfaces with a reduced coefficient of friction compared with uncoated tools [33].
TiN, TiCN, TiAlN and Al 2 O 3 are coatings that have been frequently used in industry. Earlier studies reveal that an increase in tool life is due to an increase in hardness, greater bonding energy of the coating elements and lower friction coefficients. Due to oxidation resistance and wear resistance at higher temperatures, TiAlN coating improved cutting performance. These properties make TiAlN an appropriate coating for cutting abrasive work pieces at high speeds [34]. It has been found that the coating on a micro-milling cutting tool fails due to delamination, which was confirmed by the SEM images and EDS spectra of the worn tools [35]. Delamination wear occurs due to surface layer deformation, crack nucleation and propagation of cracks parallel to the surface. Cracks finally shear the surface, resulting in long and thin wear sheets.
Earlier it was reported in the literature through the four-point bending test that thick coatings, usually more than 2 µm, delaminate from the surface of the substrate because of their high summary toughness. Along the interface of the coating and substrate, there is a difference in material properties which facilitates delamination. Finally, the coating fails due to the buckling and spalling of the delaminated portion [36].

Research Objective
From the literature review, it was found that delamination in the coating of micromilling cutting tools is confirmed only through SEM images and EDS spectra. However, coating material which is suitable for micro-milling applications and of appropriate thickness has not been reported in the literature. The delamination of coating from the substrate from the mechanics point of view has not been reported in the literature. Moreover, the factors on which delamination depend need to be examined. It was also found in the literature review that the coatings which are commonly used for micro-milling applications are titanium nitride (TiN), diamond-like carbon (DLC), aluminium titanium nitride (Al-TiN) and titanium silicon nitride (TiSiN). It has been reported that the thickness of these coatings usually ranges from 2 to 4 microns [37][38][39][40][41][42][43][44][45]. Authors have also studied various metal deposition techniques in the past 5 years and studied their performance to reduce delamination [46][47][48][49][50][51][52][53][54][55].
The objective of the present work is to carry out finite element analysis of commonly used coatings for micro-milling applications on high-speed steel substrates and evaluate the following.

•
Objective 1: to model delamination of the coating from the substrate for micro-milling applications and find out factors on which delamination depends. • Objective 2: to evaluate the performance of different coating materials for delamination and report the best coating material for micro-milling applications and their corresponding thickness.
The above mentioned objectives can be achieved by carrying out finite element analysis with high speed steel as a substrate with different coating materials of different thicknesses. In the present study, three-point bending was examined to simulate the practical conditions of the micro-milling tools during machining. The FEA results were validated using analytical results.

Designing of Coating and Substrate Assembly
A 3D finite element model of coating and substrate assembly was developed in Abaqus to find the Hertzian normal stress when subjected to normal load of 4N, applied with the help of a rigid ball, as shown in Figure 2. To validate the FEA results, it was compared with the analytical result of Hertzian normal stress. Earlier, the ball on a flat coating-substrate assembly was used to simulate the scratch test [37]. The dimensions of the FEA model are as follows: the radius of rigid ball is 200 µm. For the TiN coating, the length is 3 mm, the width is 1 mm, and the thickness is 3 µm. For the high speed steel substrate, the length is 3 mm, the width is 1 mm, and the thickness is 50 µm. Table 1 elaborates the material properties of the coating and substrate.

Mechanical Properties of Different Coatings and the Substrate
The deformation behaviour of all the coating materials is considered as linear-elastic and that of the substrate is characterized as elastic-plastic, which was already mentioned earlier. The substrate material for all four coating materials was taken as high-speed steel (HSS). Material properties of different coating materials and substrate are given in Table

Mechanical Properties of Different Coatings and the Substrate
The deformation behaviour of all the coating materials is considered as linear-elastic and that of the substrate is characterized as elastic-plastic, which was already mentioned earlier. The substrate material for all four coating materials was taken as high-speed steel (HSS). Material properties of different coating materials and substrate are given in Table 2 [37,[43][44][45]. The following values were considered for the analysis.

Assumptions in the Present Study
In the present study, the following assumptions are made based on which analysis was carried out.

•
The deformation behaviour of the substrate was characterized as elastic-plastic with isotropic hardening.

•
The deformation behaviour of the coating materials was modelled as linear-elastic.

•
The hemispherical ball was modelled as analytical rigid.

•
For the surface interaction between the ball and coating material, the outer surface of the hemispherical ball was considered as the master surface and the top surface of the coating material was considered as the slave surface. This was done because the hemispherical ball was modelled as analytical rigid [37]. • For the surface interaction between the coating material and substrate, the bottom surface of the coating was considered as the master surface and the top surface of the substrate was modelled as the slave surface as the coating material was harder than the substrate.

•
The interaction property for the junction of the coating and substrate was modelled as a 'tie' such that there was no slip, separation and penetration. • Brick elements were taken for both the coating and substrate, and the element type was taken as the quadratic with the hybrid formulation and reduced integration. The structured type of mesh control is used.

Dimensioning and Boundary Conditions
• Along the length and width, the coating and substrate both were divided into 26 equal parts. Along the thickness, the coating material was divided into two equal parts and the substrate was divided into five parts with single bias with the bias ratio as five.
Along the length, the mesh size of the coating material and substrate was 115 µm; along the width, the mesh size of the coating material and substrate was 38 µm; along the thickness, the mesh size for the coating material was 1.5 µm; and for the substrate, the mesh size ranged from 4 µm to 19 µm. The interaction property for the junction of the coating and substrate was modelled as a 'tie' such that there was no slip, separation and penetration.

•
The bottom of the substrate was given a rigid support and a normal load of 4 N was applied at the reference point of the rigid ball. The stress contour thus obtained is shown in Figure 3a,b.
equal parts. Along the thickness, the coating material was divided into two equal parts and the substrate was divided into five parts with single bias with the bias ratio as five. Along the length, the mesh size of the coating material and substrate was 115 µm; along the width, the mesh size of the coating material and substrate was 38 µm; along the thickness, the mesh size for the coating material was 1.5 µm; and for the substrate, the mesh size ranged from 4 µm to 19 µm. The interaction property for the junction of the coating and substrate was modelled as a 'tie' such that there was no slip, separation and penetration.

•
The bottom of the substrate was given a rigid support and a normal load of 4 N was applied at the reference point of the rigid ball. The stress contour thus obtained is shown in Figure 3a,b. Now the above FEA model of coating-substrate assembly is treated as simply supported beam and is subjected to a normal load of 4 N centrally throughout the width of the coating-substrate assembly and analysed for three-point bending test. From previously reported results on micro cutting tool it was found experimentally that Now the above FEA model of coating-substrate assembly is treated as simply supported beam and is subjected to a normal load of 4 N centrally throughout the width of the coating-substrate assembly and analysed for three-point bending test. From previously reported results on micro cutting tool it was found experimentally that cutting force varies between 1.5 N to 2 N. Therefore, in view of the above, for the present analysis a load of 4 N is taken considering extreme conditions. Four different coating materials are used for analysis, which are commonly used for micro-milling applications. Dimensions of each coating materials and substrate used in the FEA model are tabulated in Table 3.

Surface-Based Cohesive Behaviour
To simulate the delamination of the coating, the interface of different coatings and the substrate was modelled with cohesive surface in Abaqus. Surface-based cohesive behaviour was used to simulate the interface adhesion between the different coatings and substrate since the interface thickness was negligibly small. Cohesive behaviour is defined by means of the surface interaction property between the surfaces of the coating material and substrate material coming in contact with each other. The surface of the coating interacting with substrate was considered as the master surface since all the coating materials are harder than the substrate. The surface of the substrate interacting with coating was considered as the slave surface. Cohesive behaviour is defined by specifying the stiffness coefficients: Knn, Kss and Ktt for uncoupled traction-separation behaviour, where Knn represents the stiffness coefficient for the cohesive behaviour-enabled surface interaction in the normal direction and Kss and Ktt represent the stiffness coefficients for the cohesive behaviour-enabled surface interaction in the shear directions. However, it is advisable to keep the same stiffness value for the stiffness coefficients Knn, Kss and Ktt [47].
The stiffness value of the interface of the coating material and substrate is given by the relation given in Equation number (1) [48].
In equation number (1), 'i' represents the interface, 's' represents the substrate, 'c' represents the coating, E represents the modulus of elasticity and H represents the Vickers hardness. The hardness of the interface between the coating material and substrate is assumed as the average hardness of the coating material and substrate. The interface hardness, Hi, is given by Equation number (2).
The stiffness coefficients of the interface of the different coating-substrate assemblies Ei or Knn, Kss and Ktt are calculated by using Equation number (2) and are given in Table 4. These values are used for the analysis. All the required inputs are fed into the FEA model of the different coating-substrate assemblies with three different coating thicknesses of 2, 3 and 4 µm with the corresponding mesh size of the coating material along the thickness as 2, 1.5 and 2 µm, respectively. All the models run for three-point bending load condition to evaluate desirable outputs such as the von Mises stress, the plastic equivalent strain and the deformation in the coating material and substrate for each case. Consolidated results of the desirable outputs of all FEA models are shown in the results section.

Analytical Calculation for Hertzian Normal Stress
When a spherically shaped summit of radius R is brought into contact with a flat surface with a load L, as shown in Figure 4, the surfaces deform to create the contact zone of radius a. According to Hertz's equations for the elastic deformation of a sphere on a flat surface, the radius of the contact zone is given by where Ec is the composite elastic modulus of the two contacting materials with the elastic modulus E 1 and E 2 and the Poisson's ratio ν 1 and ν 2, respectively. The value of Ec is given by the relation as given below.
Micromachines 2022, 13, x FOR PEER REVIEW 9 of 21 The mean normal stress, pm, is given by or, The maximum normal stress, po, is given by po = When a rigid ball of radius R = 200 µm and elastic modulus E1 = ∞ is pressed with a normal load of 4 N against a flat plate of TiN with elastic modulus E2 = 300 GPa and Poisson's ratio ν = 0.22, then the maximum normal stress obtained, po, is given in Table 5. The above loading conditions of a load of 4 N through a rigid ball pressed against a flat plate of TiN is simulated in Abaqus 6.11-1. The maximum normal stress developed in FEA model was 12,109 MPa. The variation of the FEA result from the analytical result, i.e., 12,435.97 MPa, is 2.63%, which is acceptable. This confirms that the FEA model of the coating-substrate assembly is acceptable.
The plastic strain curve of the high-speed steel material used as a substrate is shown in Figure 5 [46]. With reference to Table 6, for compressive strain, v/s compressive stress For this geometry, the real area of contact A is given by The mean normal stress, pm, is given by or, The maximum normal stress, po, is given by When a rigid ball of radius R = 200 µm and elastic modulus E 1 = ∞ is pressed with a normal load of 4 N against a flat plate of TiN with elastic modulus E 2 = 300 GPa and Poisson's ratio ν = 0.22, then the maximum normal stress obtained, po, is given in Table 5. The above loading conditions of a load of 4 N through a rigid ball pressed against a flat plate of TiN is simulated in Abaqus 6.11-1. The maximum normal stress developed in FEA model was 12,109 MPa. The variation of the FEA result from the analytical result, i.e., 12,435.97 MPa, is 2.63%, which is acceptable. This confirms that the FEA model of the coating-substrate assembly is acceptable.
The plastic strain curve of the high-speed steel material used as a substrate is shown in Figure 5 [46]. With reference to Table 6, for compressive strain, v/s compressive stress values of the HSS five data point values are extracted for the plastic behaviour of high-speed steel corresponding to a 20 • C curve, as shown in Figure 5. This plastic behaviour data were added into the Abaqus 6.11-1 software as the plastic behaviour of the high-speed steel.
Micromachines 2022, 13, x FOR PEER REVIEW 10 of 21 behaviour data were added into the Abaqus 6.11-1 software as the plastic behaviour of the high-speed steel.

Results and Discussion: Performance Evaluation of Different Coating Materials in Delamination (Objective 2)
All the FEA models of the different coating-substrate assemblies with three different coating thicknesses of 2, 3 and 4 µm were run for the three-point bending load condition to evaluate the von Mises stress, the plastic equivalent strain and the deformation in the coating materials and substrate. Contour plots of the desirable outputs are given in the Appendix A. The consolidated results of the desirable outputs of all FEA models are shown in this section. Figure 6 shows the maximum von Mises stress acting on the coating material at the junction of the coating and substrate, situated below the loading region, where AlTiN_3 represents the coating of aluminium titanium nitride material of 3 micron thickness.
Micromachines 2022, 13, x FOR PEER REVIEW 11 o where AlTiN_3 represents the coating of aluminium titanium nitride material of 3 mic thickness. It is quite clear from Figure 6 that as the thickness of the coating material increa the stress developed on the coating material decreases since the section modulus of coating increases. Additionally, for a given thickness of coating material, the str It is quite clear from Figure 6 that as the thickness of the coating material increases, the stress developed on the coating material decreases since the section modulus of the coating increases. Additionally, for a given thickness of coating material, the stress developed in the diamond-like carbon coating material is minimum (DLC_4) and that in the titanium silicon nitride coating material is maximum (TiSiN_4), which means the stress bearing capacity of the DLC coating is minimum and that of the TiSiN is maximum. Figure 7 shows the maximum von Mises stress acting on the substrate at the junction of the coating material and substrate, situated below the loading region with different coating materials of different thicknesses. It can be observed from Figure 7 that the stress developed on the substrate material decreases with the increase in the coating thickness due to the increase in the section modulus of the coating-substrate assembly. Additionally, for a given thickness of the coating material, the stress developed on the substrate is maximum in the case of the DLC coating and minimum in the case of the TiSiN coating. This shows that the TiSiN coating material bears most of the stress developed due to the application of the load preventing the substrate from experiencing high stress, unlike the other coating materials. Figure 8 shows the differential stress at the junction of the coating and substrate, which is nothing but the difference of the stresses experienced by the coating material and substrate. It is clear from Figure 8 that for a given thickness of the coating material, the differential stress between the coating material and substrate at the junction is maximum for DLC and minimum for TiSiN. As the thickness of the coating material increases, the differential stress at the junction of the coating material and substrate increases. It can be observed from Figure 7 that the stress developed on the substrate material decreases with the increase in the coating thickness due to the increase in the section modulus of the coating-substrate assembly. Additionally, for a given thickness of the coating material, the stress developed on the substrate is maximum in the case of the DLC coating and minimum in the case of the TiSiN coating. This shows that the TiSiN coating material bears most of the stress developed due to the application of the load preventing the substrate from experiencing high stress, unlike the other coating materials. Figure 8 shows the differential stress at the junction of the coating and substrate, which is nothing but the difference of the stresses experienced by the coating material and substrate. It is clear from Figure 8 that for a given thickness of the coating material, the differential stress between the coating material and substrate at the junction is maximum for DLC and minimum for TiSiN. As the thickness of the coating material increases, the differential stress at the junction of the coating material and substrate increases. The differential stress at the junction of the substrate and coating material cau some plastic strain at the substrate surface due to the difference in the hardness of coating material and substrate material. Figure 9 shows the plastic equivalent str developed on the substrate with different coating materials of different thicknesses. It c be seen from Figure 9 that for a given coating material, as the thickness of the coat increases, the plastic equivalent strain in the substrate decreases as the stress develop in the substrate decreases. Additionally, for a given thickness of the coating, the plas equivalent strain in the substrate is maximum for the DLC coating and minimum for TiSiN coating. The differential stress at the junction of the substrate and coating material causes some plastic strain at the substrate surface due to the difference in the hardness of the coating material and substrate material. Figure 9 shows the plastic equivalent strain developed on the substrate with different coating materials of different thicknesses. It can be seen from Figure 9 that for a given coating material, as the thickness of the coating increases, the plastic equivalent strain in the substrate decreases as the stress developed in the substrate decreases. Additionally, for a given thickness of the coating, the plastic equivalent strain in the substrate is maximum for the DLC coating and minimum for the TiSiN coating. Figure 10 shows the deformation in the substrate at one of the corners at the free edge having coatings of different materials with varying thicknesses. It is very clear from Figure 10 that as the thickness of the coating increases, the deformation in the substrate decreases. It can be also seen ( Figure 13) that for a given thickness of coating, the deformation in the substrate is maximum for the DLC coating and minimum for the TiSiN coating. Figure 11 shows the deformation in the different coating materials of varying thicknesses at the junction of the coating material and the substrate at one of the corners of the free edge. It can be observed from Figure 13 that for a given thickness of coating, the deformation is maximum for the DLC coating and minimum for the AlTiN coating. This is because the deformation is inversely proportional to the elastic modulus. As the thickness of coating material increases, its deformation decreases due to the increase in its section modulus.  Figure 10 shows the deformation in the substrate at one of the corners at the free edge having coatings of different materials with varying thicknesses. It is very clear from Figure 10 that as the thickness of the coating increases, the deformation in the substrate decreases. It can be also seen ( Figure 13) that for a given thickness of coating, the deformation in the substrate is maximum for the DLC coating and minimum for the TiSiN coating.    Figure 10 shows the deformation in the substrate at one of the corners at the free edge having coatings of different materials with varying thicknesses. It is very clear from Figure 10 that as the thickness of the coating increases, the deformation in the substrate decreases. It can be also seen ( Figure 13) that for a given thickness of coating, the deformation in the substrate is maximum for the DLC coating and minimum for the TiSiN coating.  thicknesses at the junction of the coating material and the substrate at one of the corners of the free edge. It can be observed from Figure 13 that for a given thickness of coating, the deformation is maximum for the DLC coating and minimum for the AlTiN coating. This is because the deformation is inversely proportional to the elastic modulus. As the thickness of coating material increases, its deformation decreases due to the increase in its section modulus.  Figure 12 shows the difference in the deformation between the different coating materials of varying thicknesses and the substrate located at one of the corners at the free edge, thereby showing the extent of the delamination. It can be said from Figure 12 that for a given thickness of coating, the difference in the deformation between the coating material and substrate is the maximum for the AlTiN coating and minimum for the DLC coating. From Figure 13 it can be concluded that as the thickness of coating material increases, the difference in the deformation between the coating material and substrate, i.e., delamination, increases.  Figure 12 shows the difference in the deformation between the different coating materials of varying thicknesses and the substrate located at one of the corners at the free edge, thereby showing the extent of the delamination. It can be said from Figure 12 that for a given thickness of coating, the difference in the deformation between the coating material and substrate is the maximum for the AlTiN coating and minimum for the DLC coating. From Figure 13 it can be concluded that as the thickness of coating material increases, the difference in the deformation between the coating material and substrate, i.e., delamination, increases.    Because of the non-availability of FEA results in the literature, the results of the present study were not compared. However, it is clear from the result that aluminium titanium nitride coating material performs better than titanium nitride, which was reported earlier in the literature. Because of the non-availability of FEA results in the literature, the results of the present study were not compared. However, it is clear from the result that aluminium titanium nitride coating material performs better than titanium nitride, which was reported earlier in the literature.

Conclusions
From the different consolidated results of the desired outputs, the following conclusions can be made:

1.
Using a 3D modelling approach, the coating and substrate assembly was modelled and studied for delamination. Various factors were evaluated which impacted the delamination in micro-milling.

2.
On the basis of the plastic equivalent strain occurring in the substrate material, the TiSiN and AlTiN coatings are the best since the plastic equivalent strain occurring in the substrate with the TiSiN and AlTiN coating materials is less (Appendix A). Between these two, the TiSiN coating shows the least plastic equivalent strain in the substrate, which serves the purpose of protecting the substrate from plastic deformation. The plastic equivalent strain in the substrate decreases with increase in the thickness of the coating material.

3.
On the basis of the difference in deformation, i.e., delamination, the TiSiN coating is better than the AlTiN coating since delamination in the coating-substrate assembly with the TiSiN coating is less than that of the AlTiN coating. The delamination of the coating material from the substrate increases with increase in the thickness of the coating material.

4.
By combining the plastic equivalent strain in the substrate and the delamination of the coating from the substrate, we can conclude that the TiSiN coating of 3 micron thickness is the most optimum coating thickness for micro-milling applications. These results fulfill the requirements of objectives 1 and 2.

5.
The higher the interface stiffness coefficient of the coating-substrate assembly is, the less the delamination of the coating material from the substrate is.

6.
The delamination depends on the Young's modulus and the hardness of both the coating material and substrate.
Suggestions for future work: In order to validate the FEA results obtained, experimental studies are essential. Therefore, these values will help in comparing the FEA results. An experimental setup is required to see the delamination process. Different coatings on the substrate are required to be generated to conduct the experiment. The mechanical properties of the coatings need to be evaluated. SEM images can be taken to evaluate the extent of the delamination occurring on different coating materials.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A
The contour plots of the desirable outputs of the FEA model with titanium silicon nitride coating of 3 µm thickness over high-speed steel substrate are as follows.