Cracking analysis of HVOF coatings under Vickers indentation.

Abstract The fracture strength of five HVOF coatings, which are made of hard metals, Tribaloy alloy, and superalloys, respectively, coated on 1018 low carbon steel substrate, is studied under Vickers identation, associated with FEA stress computation. The cross sections of the coating specimens are examined on a Hitachi Model S-570 scanning electron microscope (SEM), which investigates the quality and measures the geometry of the coatings. The mechanical properties of the coatings and the substrate are determined in the cross sections using the nano-identation technique. The cracking behavior of the coatings under different indentation loads is investigated using a Vickers hardness tester. Three-dimensional finite element analysis (FEA) simulation of the Vickers indentation test is conducted to determine the stress fields in the coating/substrate systems in order to understand the fracture mechanisms of the coatings under the indentation loads using the ABAQUS software package. The FEA stress results are in good agreement with the experimental observation of Vickers identation.

Keywords HVOF coating, Vickers indentation, Crack, Fracture, FEA, Stress

Introduction

HVOF coating is emerging as one of the most cost effective ways to combat wear and corrosion. It is reported to be a versatile technology and has been adopted by many industries due to its flexibility, cost effectiveness, and the superior quality of coatings produced. (1) HVOF coating provides very dense (low porosity), high bond strength to substrate, and very smooth as-sprayed textures. The products show low residual tensile stress or in some cases compressive stress, which enables very much thicker coatings to be applied than previously possible with the other processes. As with any thermal spray process, the objective of HVOF technology is to transfer energy, both kinetic and thermal, to the powder particles with a high degree of efficiency. The physical and chemical conditions of the particle that impinge on the surface are dependent on a large number of parameters such as the preparation of the part surface, composition, morphology, size distribution and feed rate of the powder, and the precise control of gas flow, torch design, the gas jet conditions (temperature, velocity, pressure), and particle injection method. (2) The properties of HVOF coatings rely highly on these factors.

Adhesion of thermal sprayed coating to a substrate has been a primary concern to engineers ever since thermal spray processes were introduced to various industries. This is because the process cannot be effectively employed for engineering applications if a coating does not bond well to a substrate. Therefore, investigation on the bonding mechanisms of coating-substrate, which is an important element of the theory of thermal spray technology, has attracted a great deal of attention in the past decades.

However, currently, mechanical characterization of coating/substrate systems is rather lagging behind the coating development, which to some degree hinders the application of many coatings. The importance of coating evaluation has been emphasized by many investigators, and several techniques have been developed to address this issue. A simple test to assess the adhesive strength is the direct pull test, as described in ASTM D4541 (3) and ASTM C633-79. (4) Non-standard test methods are also developed such as laser spallation (5) and ultracentrifugal tests. (6) There are also quality control test methods such as peel test, (7) hammering test, (8) and so on. Instead of measuring the adhesive strength, many testing methods have also been devised to measure the fracture energy of the coating/substrate interface--a fundamental measure of the adhesion. These include the pull test of chevron-notched sandwich specimens, (9), (10) the bending test of notched delamination specimens. (11) and the pull test of radial-notched cylindrical specimens. (12)

In recent years, indentation/scratch technique has been widely employed to study the fracture behavior of coatings. (13-18) Historically, indentation technique is used as a testing method of hardness, but it has become a standard technique to assess elastic modulus and hardness of materials. (19-21) Meanwhile, attempts have been made to determine the fracture toughness of coatings and the coating-substrate interface strength via indentation. (13), (15) The particular attraction of this technique lies in its unique simplicity as a nondestructive means for producing regions of high stress intensity in a specimen.

In the present research, five coating specimens, made of different materials and produced by HVOF process, were studied. The coating strengths were evaluated under Vickers indentation through investigating the cracking response of the coatings to the applied load. The indented surfaces of the specimens were examined using SEM to obtain the cracking profiles. Three-dimensional finite element analysis (FEA) was employed to simulate the indentation test and thus to determine the stress distributions in the coating/substrate systems under the indentation loads. The fracture mechanisms of the coatings under Vickers indentation were discussed in terms of the stress results.

Coating materials 

Five plate coating specimens (~ 100 x 50 mm) were fabricated with the Jet Kote HVOF process at Deleoro Stellite Inc., designated as WC-Co hard metal (WC/17Co), WC-Co-Cr hard metal (WC/10Co/4Cr), Tribaloy T-800 (Co/18Cr/28Mo/3Si), superalloys Nistelle C (Ni/16Cr/16Mo/4Fe/4W), and Nistelle Super C (Ni/23Cr/18Mo). The substrate is 1018 low carbon steel.

The WC hard metals have exceptional hardness and toughness. WC has low wear rate, especially when it has metallic binder in its cemented matrix. The common metallic substance used in WC is cobalt (Co) and is normally written as WC-Co. Cobalt can increase the hardness and strength as the binder, but causes the low corrosion resistance. (22), (23) Addition of chromium to WC-Co matrix can increase both the hardness and erosive/corrosive resistance.

T-800 is a cobalt-based Tribaloy alloy, containing chromium, molybdenum, and silicon. Chromium is added to enhance corrosion resistance, whereas molybdenum and silicon are used to impart wear resistance by forming the Laves phase. T-800 has exceptional oxidation/corrosion and wear-resistance due to the high Cr and Si contents. It is harder and has better abrasive wear-resistance than either Tribaloy alloys T-400 or T-700. (24) The allotropic nature of cobalt can cause either the face centered cubic (fee) or hexagonal close packed (hcp) crystal structures, or both, to be present in Tribaloy alloys, depending on the thermal treatment. The hard primary phase is a ternary Laves phase of the C-14 ([MgZn.sub.2]) type having a melting point of about 1560[degrees]C, and its compositions are approximately [Co.sub.3][Mo.sub.2]Si or CoMoSi for the Co-Mo-Cr-Si Tribaloy family. The primary dendrites of the hard intermetallic phase are in a eutectic matrix of smaller intermetallic particles embedded in a Co solid solution; there are also regions of Co solid solution which are free from the secondary Laves phase. (25-27) T-800 is a typical Tribaloy alloy for corrosion and wear applications, with the overall hardness of Laves phase and cobalt solid solution of 54-60 HRC: the Laves phase content is 55%. However, the Laves phase limits ductility and impact strength. (25)

Nistelle C and Nistelle Super C are nickel-based superalloys, which are very important group of engineering materials for high strength and high corrosion/wear resistance applications at high temperatures. Nickel, as the base metal, possesses high hardness, which contributes to superior wear resistance. It is also a good base metal for corrosion-resistant materials because it readily alloys with other corrosion-resistant metals. Chromium is included because it is an element known to promote the formation of passive films in oxidizing acids. Molybdenum is an element known to enhance the resistance of nickel to reducing acids such as hydrochloric. Another important contribution of molybdenum to these alloys is its solid-solution strengthening effect, owing to its large atomic volume in the matrix to pin dislocations. When alloyed with both Cr and Mo, Ni-based alloys display resistance to complex corrosive media. (28), (29)

Nistelle C is one of the popular Ni-Cr-Mo alloys. In its chemical composition tungsten has the same solid solutioning effect as molybdenum, and iron are present as less expensive raw materials. (28), (30) However, for some specific applications, structural or thermal stability requires minimization or removal of tungsten and iron from the alloy, because both of the elements have a negative effect on thermal stability. (28) Iron is also vulnerable to corrosive attacks. Furthermore, because most industrial environments contain oxidizing impurities, such as ferric ions and dissolved oxygen, the chromium content in the alloy must be increased. To this end, Nistells Supper C was developed at Deloro Stellite. Inc. enhanced corrosion resistance. The microstructure analyses show that Nistelle C and Nistelle Super C all consist of a Ni solid solution matrix with a eutectic of intermetallic compounds, [Ni.sub.3] (Mo, Cr) and [Ni.sub.4] (Mo, Cr), embedded that are well known to enhance the hardness, strength, and mechanical properties. Nistelle Super C has superior corrosion resistance to both oxidizing acid and reducing acid, due to the increased Cr and Mo contents and minimized Fe content. Both of the alloys possess good ductility with the high elongations of 12.5% and 14% for Nistelle C and Nistelle Super C, respectively. (31) The higher ductility of Nistelle Super C is attributed to the lower volume fractions of he brittle intermetallic compounds. Since Nistelle Super C contains more Mo and Cr, it is goof for applications in highly corrosive and acidic, high-temperature environments. However, Nistelle C has the advantage of having a larger volume fraction of eutectic intermetallic phase, which contributes to tremendous reduction in deformation. (31)

Coating specimen fabrication 

In the HVOF process of the coating specimen fabrication, the Jet Kote system developed high-particle velocities through combustion of a fuel gas and oxygen. The high-velocity combustion gases heated and propelled surfacing powders onto the workpiece. All powder materials were mixed thoroughly prior to use and applied directly to the substrate with no bond coat required. The powder feed featured a superior powder delivery unit to ensure repeatable powder flow. The gas stream, produced by internal combustion of oxygen and fuel gas, had an initial velocity of 1350 m/s at a temperature of 3000 [degrees] C. The pressure and flow of gases were monitored by a central console. The powder particle size was about 40-100 [micro]m. The coating thickness of the specimens was between 0.3 and 0.4 mm. The fuel gas used was hydrogen.

Cross-section examination

The coating specimens were then sectioned into small pieces 3/4 x 3/4 in. for indentation tests using the Diamond Cut Saw machine with the cutting speed of 1000 rpm. Before the tests, the cross section of each specimen was examined on a Hitachi Model S-570 SEM machine. The purposes of doing this were twofold: one was to investigate the quality of the coatings (for example, porosity, cracking, and defects or flaws); the other was to measure the geometry of the coatings such as coating thickness and interface profile.

The cross-section surface of each specimen was prepared by grinding with grit papers from #180 to #600, and then polishing with abrasive cloth plus 1 [micro]m alumina powders. After cleaned with an ultrasonic cleaner, the surfaces were then etched with the solution of the mixture of 15 mL [HNO.sub.3], mL acetic acid, 60 mL HCI, and 15 mL [H.sub.2] O, in order to develop the features of the coating, interface, and the substrate under SEM for cross-section examination and under the optical microscope in nano-indentation test.

The polished and etched specimens were then carbonized and examined with SEM. The obtained SEM images of cross section of the specimens, with sharp contrast between the substrate and the coating, are presented in Fig. 1. It is observed that the coatings have a high-density structure and no obvious cracking and porosity are found. It is also noticed that the coatings do not have obvious diffusion zone at the interfaces and the interfaces are not smooth but rugged. This may be due to the high temperature (3000 [degrees] C) in the HVOF process which melted the substrate surfaces, resulting in rugged interfaces. But since the velocity of the spray is very high, there is almost no time for diffusion; as a result, no obvious diffusion range is formed at the interfaces. Figure 1 shows that the coating thickness is not uniform over the specimen surface. It is measured that the average thickness of coating of specimens WC-Co, WC-Co-Cr, and T-800 is about 0.3 mm; that of specimens Nistelle C and Nistelle Super C is about 0.4mm.

Nano-indentation test

In order to obtain the mechanical properties of the coatings and the substrate, respectively, which were required for the following FEA simulation, nanoindentation test was conducted on each component of cross section of each specimen, using a load and displacement sensing indentation technique on a CSM nano hardness machine. The load frame compliance is 0.2 [micro]/N (0.2 nm/mN) and the indenter tip radius is 100 nm. The nano-indentation effective Young's modulus E* = E/(1-[v.sup.2]) and hardness H are calculated based on the loading/unloading curves which are measured with a Berkovich indenter using the Oliver-Pharr method, (32) where E and v are the Young's modulus and Poisson's ratio of the tested material, respectively. The load-displacement curves of loading and unloading are recorded automatically during the indentation. The area enclosed by the loading and unloading curves represents the dissipated energy due to plastic deformation. (33)

The maximum load applied in each indentation was 300 mN with the loading and unloading speed of 600mN/min in the present test. The load level was selected depending on the consideration in the material being tested. If the overall properties of the material are of interest, the load should be higher because the contact area of indentation would be larger under a high load than under a low load, otherwise the load should be lower when the individual phases of the material are considered. In this article, the overall properties of each component of the coating systems. that is, the overall properties of the coating layer. interface, the substrate, were concerned; therefore, a high load was used. Poission's ratio was assumed to be 0.3 The Young's modulus values of the coatings and the substrate obtained from this test would be used for the following FEA simulation.

Three indentations were preformed on each component in different locations along the coating thickness of the specimens. It was observed that the three testing results were very close, the data error is between 2% and 3%; therefore the average of the three data is taken as the final result. The loading unloading curves obtained from the nano-indentation test are presented in Fig. 2. Since the loading/unloading curves of the substrate measured on each specimen are almost identical because the substrate material is the same for all the specimens, only one of them is presented as a representative in Fig. 2. It is clear that under the indentation load of 300 mN. the substrate material experienced the maximum deformation of about 2200 nm. Comparing the coating material among the specimens, WC-Co has the minimum deformation, followed by WC-Co-Cr, then T-800; Nistelle C and Nistelle Super C have similar deformation, which is larger than the other specimens.

[FIGURE 2 OMITTED]

The hardness and Young's modulus and Poisson's ratio. obtained from the nano-indentation test, are presented in Table 1. The data of Young's modulus and Poisson's ratio of the coating materials and the substrate material would be used in the following FEA simulation. In the loading/unloading curves, the area encolosed by the loading curve represents the total deformation energy end they enclosed by the unloading curve is the elastic deformation energy. Define [[eta].sub.e] as the ratio of the elastic deformation energy to the total deformation energy and [[eta].sub.p] as the ratio of the plastic deformation energy to the total deformation energy, then [[eta].sub.e] + [[eta].sub.p] = 1. Thus, the maximum deformation. residual deformation. and [[eta].sub.e] and [[eta].sub.p] values can be obtained from the loading/unloading curves; the data are summarized in Table 2.

Comparing these data between the specimens, WC-Co-Cr possesses the highest Young's modulus, which means that it has the highest resistance to elastic deformation, followed by WC-CO. The other specimens have close values of Young's modulus, Regarding hardness. the WC-Co hard metal is the hardest: another WC-CO-Cr hard metal is the second hardest; Tribaloy alloy T-800 is harder than the superalloys, Nistelle C and Nistelle Super C. There is almost no difference in hardness between Nistelle C and Nistelle Super C. As concerns the [[eta].sub.e] and [[eta].sub.p] values. the WC-CO hard metal exhibits the highest elasticity ([[eta].sub.e]), but the lowest ductility ([[eta].sub.p]). WC-CO-Cr and T-800 have the similar elasticity and ductility. The superalloys have better ductility but lower elasticity.

Table 1: Mechanical properties of the coating materials

Specimen                          Property

                       Young's modulus,  Hardness, H
                          E (GPa)           (GPa)

WC-Co                        212             21.4
WC-CO-Cr                     297             14.8
T-800                        195             11.4
Nistelle C                   196              6.7
Nistelle Super C             201              6.8
1018 low carbon steel        248              3.1

Table 2. Indentation deformations of the coating materials

Specimen                               Property

                Maximum      Residual     [[eta].sub.e]  [[eta].sub.p]
              deformation  deformation          (%)           (%)
                  (nm)         (nm)

WC-Co             1078          619              43            57

WC-Co-Cr          1170          842              28            72

T-800             1368          963              30            70

Nistelle C        1661         1301              22            78

Nistelle          1634         1303              20            80
Super C

1018 low          2200         2037               7            93
carbon steel

Vickers indentation test

To evaluate the bonding strengths. the coating specimens were tested under Vickers indentation. The fracture behaviors of the coatings were investigated through the response of coating cracking to the indentation load. In order to remove the loose or porous structure in the top layer of the coating specimens. the coating surfaces were polished before the Vickers indentation test. First, the coating surfaces were ground with grit papers from # 180 to # 1200. After ultrasonic cleaning for 10 min to remove debris, the specimens were polished using Texmet 1500 cloth with 3 [micro]m abrasive fluid and GBuehler Metadi fluid. Final Polishing used Fourspoc cloth with 30 mL Buehler Mastermet Colloidal Silica polishing suspension.

The indentation test was performed on the polished surface of each specimen using a standard Vickers machine with a diamond tip of 136 [degrees] angle. The test was started with a low indentation load of 10 N for each specimen and the load was then gradually increased up to a high level of 700 N to investigate the dynamic cracking/fracture response of the coating to the load increasing. When total fracture was observed in a coating under a load level. The indentation test was stopped on this specimen: otherwise, the load was continuously increased up to the machine capability of 700 N. Five indentations were taken on each specimen under each load level.

The morphologies of the indented coating surfaces were examined using SEM after the indentation test. The morphologies of indentation marks left on the coating surfaces could be observed under SEM. For the hard metal WC-CO, cracks were observed along the indentation edges under the low load of 10 N but no cracks were found at the corners of the indentation, as shown in Fig. 3. Under the higher load of 35 N, it was observed that the edge cracks grew deeper and longer; meanwhile, shallow and short cracks were found at some corners of the indentation, as shown in Fig. 4. With the increase in the indentation load to 100 N, the corner cracks propagated and at the same time cracks were initiated at all the corners of the indentation (see Fig. 5). Under the higher load of 500 N. the cracks were further extended and connected with each other, leading to fracture of the coating surface. as shown in Fig. 6.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

Similar to specimen WC-CO, cracks were also found in the WC-CO-Cr specimen at the edges of the indentation under the indentation load of 10 N. but not at every edge of the indentation. only some of cracks grew to longer and deeper cracks, and in the meanwhile small cracks were found at the corners of the indentation. The corner cracks were extended in both radial direction and depth direction under a high load of 300 N. as shown in Fig. 7d.

[FIGURE 7 OMITTED]

According to the cracking profiles of specimens WC-Co-Cr under different load levels. one may summarize the steps of the crack development process as:

Step 1: Cracks were initiated in the area around the indentation edges.

Step 2: The edge cracks grew along the indentation edged and also developed deeper.

Step 3: Cracks were initiated at the vertices (corners) of the indentation.

Step 4: The corner cracks grew in both radial direction and the depth direction.

It should be noticed that Steps 2 and 3 may occur simultaneously, depending on the load level and the coating material. Since hard metals, WC-Co and WC-Co-Cr. are typical brittle materials, their fractured surfaces exhibited obvious brittle failure features.

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

However, for the Tribaloy alloy T-800 coating, no obvious cracking was observed in the coating surface, whether the load was low or high. Even micro or small cracks were not observed in the indented surface under the load below 500 N, as shown in Figs. 8-10. Under high loads only micro or small cracks were observed in the areas of the indentation edge and indentation corner (see Figs. 11 and 12). Therefore, compared with the WC-containing hard metal coatings. the Tribaloy alloy T-800 coating exhibited higher resistance to cracking/fracture.

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

Similar to the Tribaloy coating, the Nistelle C and Nistelle Super C coatings did not exhibit the brittle fracture behavior as specimens WC-Co and WC-Co-Cr did. It was observed that the indented surfaces of the two superalloy coatings had same features and no significant difference between them was found. As seen in Figs. 13-16, they have the following two common features:

[FIGURE 13 OMITTED]

[FIGURE 14 OMITTED]

[FIGURE 15 OMITTED]

[FIGURE 16 OMITTED]

1.Whether under a low load or under a high load, no significant cracking or fracture was observed, even in the indentation edge region and in the indentation corner region.

2. Micro or small cracks were found in the regions of indentation edge and indentation corner under high loads.

Based on the observation of the indented surfaces in the Vickers indentation test, the cracking/fracture behavior of the five coating specimens can be summarized as below. The five specimens behaved differently under the indentation loads. The hard metal coatings exhibited lowest fracture strength; even under low loads the coatings totally fractured. Cracking occurred first in the indentation edge area; the indentation corner area was cracked with increasing of the indentation load. The crack at the indentation edge propagated along the indentation edge and that at the indentation corner grew radially with an increased load. The Tribaloy alloy coating exhibited the highest resistance to cracking/fracture among the five coating specimens; no obvious cracking was observed in this specimen even under the highest load. Only small or micro cracks were found at the indentation edges and the indentation corners in the coating surface under high fracture strength, compared to the hard metal coatings. They behaved similarly to the Tribaloy alloy coating; no significant cracking occurred in these coatings, but only micro or small cracks were found in the indentation edge and indentation corner areas when the specimens were highly loaded.

FEA simulation of Vickers indentation test

In order to investigate the fracture mechanisms of the coating specimens under Vickers indentation test and associate the cracking profiles with the stress fields in the coating systems, three-dimensional FEA modeling were conducted simulating the Vickers indentation process. The model was generated by a square-based pyramid shape indenter loaded on the coated surface, corresponding to the indentation test. The model consisted of a rigid indenter and a deformable-coated specimen (coating and substrate). The indenter was loaded on the coated surface under a normal force and an indentation mark was created, corresponding to the magnitude of the applied load in the Vickers indentation test. The stress field in the indentation area was determined, which was related to the cracking profile observed in the indentation test. The FEA results were expected to predict the locations where cracks might occur and also the critical loads for the cracks.

To simplify the FEA models, the following important assumptions were made in the FEA simulation:

1. The coating surface was flat and smooth; roughness was ignored.

2. The coating layer and the substrate were solid cohesively joint with a zero-thickness interface; that is, the interface was a plane without thickness. This was based on a cross-section examination under SEM (see Fig. 1).

3. The interface between the coating and the substrate was a flat plane, although, in reality, the interface was not exactly flat or smooth but rugged, as shown in Fig. 1.

4. Based on linear elastic fracture mechanics, the coating materials were considered to deform elastically only; plastic deformation was not considered. But the substrate material. AISI 1080 low carbon steel, deformed elastically and plastically, because it is more ductile compared with the coating materials.

The Vickers indentation test on the coating specimens was simulated with three-dimensional models using the commercial FEA software package ABA-QUS/CAE version 6.5.1. Both the coating and substrate materials were considered solid and homogenous. Due to the geometrical and loading symmetry, the problem was solved by modeling a quarter of the coating/substrate system, as shown in Fig. 17. The 1-, 2-, and 3-axes intersect at the indenter tip, which correspond to the x-, y-, and z-coordinate axes, respectively. Three planes were also defined, as shown in Fig. 17.

[FIGURE 17 OMITTED]

The volume of the substrate was fixed by 0.6 x 0.6 x 3 mm (length, width, thickness) for all of the coating specimens. This dimension was determined such that, compared with dimension of the indenter, the boundary constraint of the volume would not affect the stress results in the region of the indenter. The indenter tip was modeled as a truncated four-sided pyramid. The five coating specimens shared the same volume of the substrate, but the thickness of the coatings are not the same. As measured under SEM in Fig. 1, the coating thickness varied from approximately 0.3 mm for WC-Co, WC-Co-Cr. and T-800 to approximately 0.4 mm for Nistelle C and Nistelle Super C.

The mesh of the model was constructed using a twenty-node quadratic brick, reduced integration element. The substrate contained 81,773 nodes with 18,671 hexahedral elements, whereas the coating contained 36,102 nodes and 7201 hexahedral elements. The meshing element size at the localized region of the indentation on the coating surface was very small (0.010 mm), while for the meshing element that crossed the coating thickness, the size was only one to the sixth of the coating thickness; that is, 0.050 mm for 0.3 mm coatings and 0.067 mm for 0.4 mm coatings. The element size was increased gradually from the indenter region to the coating boundary in 1-direction and 3-direction, and from the interface to the substrate bottom in 2-direction (see Fig. 17).

In this model, there are two surface contacts: (1) between the indenter tip and the coating surface and (2) the interface between the coating and substrate surface. The special surface treatments were employed to simulate the real contact behaviors of the surfaces during the indentation test. The contact between the indenter and the coating surface was treated as a non-linear "contact" problem in the FEA modeling. Symmetry boundary conditions were defined for the cutting section of the coating/substrate system under a normal load. The indenter tip and the coating surface should be allowed contacting dynamically during an indentation test; that is, relative motions between the two surfaces should be allowed. The indenter was modeled as a rigid body and placed at the end-corner of the coating/substrate specimen with only 1/4 of the indenter tip's area contacting the coating surface. A static stress analysis was used as no velocity was involved during the displacement. The indenter was rotationally constrained in all directions and moved only in 2-direction. For each specimen model, different load levels were applied, consistent with the loads applied in the Vickers indentation test.

For linear elastic mechanics analysis, only Young's modulus and Poission's ratio are required for the material properties. These data were obtained from the nano-indentation test, as given in Table 1. The yield stress of the substrate material, AISI 1080 low carbon steel, was obtained from a tensile test, and it is about 373 MPa. (34)

Stress results

For brittle fracture analysis, maximum principal stress is the major consideration. According to the indentation test results, the hard metal WC-Co and WC-Co-Cr specimens fractured at low loads while the Tribaloy alloy T-800 and the superalloy Nistelle C and Nistelle Super C specimens only cracked but not totally fractured even under high loads. The FEA stress results are presented in two groups: low-load group for the hard metal specimens and high-load group for the other specimens. Figures 18 and 19 illustrate the isometric view of topographical distributions of the maximum principal stress (in Pa) for the low-load group and the high-load group, respectively. It is clear that in the coating surface the highest compressive stress occurs in the area right beneath the indenter tip and the highest tensile stress occurs in the area of the indentation edges and in the area of the indentation vertices or corners.

To better understand the fracture mechanisms of the coating specimens under the Vickers indentation test, the variation of the maximum principal stress along the indentation edge from the center A to end B, as illustrated in Fig. 20, is plotted for each specimen. The plotted curves are shown in Fig. 21. It is demonstrated that the highest tensile stress occurs in the center of the indent edge, point A in Fig. 20, and the maximum principal stress decreases along the indentation edge toward its vertex, point B in Fig. 20.

[FIGURE 18 OMITTED]

It is also interesting to investigate the variation of the maximum principal stress along the outward diagonal of the indentation in the indentation corner area from corner B to point C, as illustrated in Fig. 20. It is shown in Fig. 22 that the highest stress occurs around the corner, but not right at the corner of point B in Fig. 20, and the stress decreases along the outward diagonal of the indentation to point C in Fig. 20. Comparing the stress at the indentation edge with the at the indentation corner for each specimen, it is found that the former is larger.

[FIGURE 19 OMITTED]

[FIGURE 20 OMITTED]

[FIGURE 21 OMITTED]

[FIGURE 22 OMITTED]

Furthermore, the variation of the maximum principal stress from the indented area, point D in Fig. 20, to the indentation-free area, point P in Fig. 20, across the center A of the indentation edge, and the variation of the maximum principal stress from the identation corner B along the normal of the identation diagonal to point S in Fig. 20, are plotted in Figs. 23 and 24, respectively. It is seen that the indented area experiences very high compressive stresses and the maximum principal stress becomes tensile at the indentatin edge with the largest magnitude; it then decreases gradually from the indentation edge in the indentation-free area (see Fig. 23). From the curves in Fig. 24, along the normal of the indentation diagonal the highest stress occurs at the indentation corner and the maximum principal stress decreases away from the indentation corner.

[FIGURE 23 OMITTED]

[FIGURE 24 OMITTED]

Three load levels were used in the simulation for each specimen to investigate the response of the stress field to the indentation load. For the low-load group, the three loads are 10, 35, and 100 N, and 100, 300, and 500 N are for the high-load group. The maximum principal stress of the center of indentation edge versus the indentation load and the maximum principal stress of the indentation corner versus the indentation load are plotted for each specimen in Figs. 25 and 26, respectively.

[FIGURE 25 OMITTED]

[FIGURE 26 OMITTED]

Discussion

Comparing the cracking profiles of the five coatings under the Vickers indentation test, the hard metals WC-Co and WC-Co-Cr exhibited typical brittle features in tensile testing. (25) The significant difference between the hard metals and the Tribaloy alloy is that the hard phase WC in the hard metal powders used for the HVOF process exists as separate particles while the brittle Laves phase in the Tribaloy alloy powder is alloyed with the cobalt solid solution. In the HVOF coating fabrication process, although the fuel temperature was very high (about 3000 [degrees]C), due to the high velocity (1350 m/s), the powder was deposited on the substrate surface, forming the coating in a very short time. In this case, the WC particles had no time melting or dissolving into the cobalt matrix; they were just embedded in the matrix as separate particles. Therefore, the bond between the WC particle and the cobalt matrix in the coatings was not very strong. However, unlike in the hard metal coatings, for the Tribaloy alloy T-800 coating the interface between the Laves phase and the Co solution matrix in T-800 was already formed in the powder form, leading to stronger interface bonding in the coating after the HVOF process, compared with the WC-Co and WC-Co-Cr hard metal coatings. As a result, T-800 coating exhibited much higher resistance to cracking than WC-Co and WC-Co-Cr coatings in the Vickers indentation test.

For the superalloy specimens, both Nistelle C and Nistelle Super C possess high ductility, as demonstrated by the nano-indentation test. Therefore, they did not behave as the brittle WC-Co and WC-Co-Cr did in the Vickers indentation test. According to the microstructures of Nistelle C and Nistelle Super C, they consist of a Ni solid solution matrix with an eutectic of intermetallic compounds, [Ni.sub.3](Mo, Cr) and [Ni.sub.4] (Mo, Cr), embedded. Although the intermetallic compounds are brittle, their volume fraction in these alloy is very small (<20%), so that Nistelle C and Nistelle Super C exhibit ductile features. In the indented surfaces of these two alloys, no severe fracture was observed even under the maximum indentation edges and in the area of the indentation corners where stress concentration occurred. Owing to the good ductility, the microcracks in these alloys may not propogate unless a fatigue load is applied. As reported by previous research, comparing Nistelle C(Ni/16Cr/16Mo/4Fe/4W) with Nistelle Super C (Ni/23Cr/18Mo), it is found that these two alloys contain almost the same amount of Ni, but the latter has better ductility, owing to the reduced volume fraction of the brittle intermetallic compounds [Ni.sub.3] (Mo, Cr) and [Ni.sub.4](Mo, Cr). (31) However, this difference was not exhibited obviously by the indentation tests in the present research. The reason for this may be dual:

1. The nano-indentation was performed on nano scale, which means that it covered only a small area of the microstructure. Since the volume fraction of the brittle intermetallic compounds in the alloys was small, the main contribution to the mechanical properties measured by the indentation test was the Ni solid solution rather than the intermetallic compounds. At the same time, the Ni solid solutions of Nistelle C and Nistelle Super C had very similar mechanical properties.

2. The Vickers indentation was performed on macro scale, but as the volume fraction of the intermetallic compounds in the alloys is small, the Ni solid solutions dominate the overall performance of the alloys, which are ductile. Also, the brittle intermetallic compounds were alloyed with the Ni solid solution, similar to the Laves phase in Tribaloy alloys, but unlike the WC in hard metals present as separate particles. Thus, the bond between the intermetallic compounds and the solid solution in Nistelle C and Nistelle Super C would be much stronger than in hard metals. Therefore, the damage of Nistelle C and Nistelle Super C in the indentation test might not be caused mainly by the interface debonding, but the cracking or failure of the solid solution matrices. As a result, one may hardly see the difference in the cracking profiles between the Nistelle C and Nistelle Super C coatings.

The FEA simulation determined the stress fields of the coating/substrate systems under the Vickers indentation, which provided the fundamentals of the cracking analysis of the coatings under the indentation load. The FEA results are in good agreement with the experimental observation, and thus help understand the fracture mechanisms of the five coating specimens under the Vickers indentation. First, the distributions of the maximum tensile principal stress demonstrate the highest stress occurs in the center of the indentation edge for all the specimens. In accordance with the observation in the Vickers indentation test, cracks were induced initially in the central region of the indentation edges. Since the stress is tensile over the entire indentation edge, the edge crack would propagate along the indentation edge toward the vertex or corner of the indentation, as observed in the indentation test. Second, the FEA results show that the maximum principal stress at the indentation corner is also tensile and high. This explains why cracks were also generated at the indentation corners. However, since the stress at the indentation corner is also tensile and high. This explains why cracks were also generated at the indentation corner is lower than that at the indentation edge, the corner crack requires an increased load. As demonstrated by the indentation test, edge crack was initiated first and corner crack was induced with further loading. Third, based on the experimental observation that the corner crack was extended radially along the outward diagonal of the indentation, the variation of the maximum principal stress from the indentation corner along the outward diagonal of the indentation and the variation of the maximum principal stress from the indentation corner along the normal of the indentation diagonal were plotted in Fig. 22. It is seen that the stress is tensile over the outward diagonal, but it decreases from the indentation corner to zero at a distance. This is why the corner crack propagated to a certain length and then stopped under a given indentation load. From Fig. 24, the largest magnitude of the maximum tensile principal stress occurs at the indentation corner along the normal of the indentation diagonal. Due to the symmetry of the model geometry and loading, the highest stresses must occur over the diagonal of the indentation, which leads to the corner crack growing radially along the indentation diagonal.

Furthermore, as seen in Fig. 21a, at the center of the indentation edge the maximum principal stress is about 320 MP a for the WC-Co-Cr specimen. There are no available data for the tensile strengths of the coating materials (WC/170Co and WC/10Co/4Cr) of specimens WC-Co and WC-Co-Cr, but it is;reported that the tensile strength of WC hard metals is approximately 200 MPa. (35) Thus, according to the FEA results, the maximum principal stress in the center of the indentation edge under the load of 10 N is greater than the tensile strength of the coating material for both the WC-Co-Cr specimen and the WC-Co-Cr specimen; therefore, the coating materials should fail under this load. As observed in the indentation test, cracks were initiated in the central region of the indentation edge. Under the same indentation load, the stress in the WC-Co-Cr specimen is higher than in the WC-Co specimen (see Fig. 21a). However, the latter exhibited higher strength against cracking in the indentation test. This implies that the WC-Co-Cr material has higher tensile strength than the WC-Co material. Previous research has reported the tensile strength of Tribaloy T-800, which is about 700 MPa. (25) From Fig. 21b; the maximum principal stress in the T-800 specimen under the load of 100 N, which is about 440 MPa, is lower than the tensile strength of the T-800 material. In accordance with the experimental observation, no cracks were found in the T-800 specimen under the 100 N load. For the Nistelle C and Nistelle Super C coatings. the stress values under the same indentation load are very close for the two specimens, as presented in Fig. 21 b; the highest stresses are about 385 MPa. The tensile strengths of Nistelle C and Nistelle Super C are very close and about 530 MPa, which were provided by previous research. (31) Since the maximum principal stress is lower than the tensile strength of the coating material for the Nistelle C and Nistelle Super C specimens under the 100 N indentation load, cracking would not occur in the specimens at this level, as observed in the indentation test.

For the hard metal coatings, the maximum principal stresses at the indentation corner are 160-170 MPa, which are lower than the tensile strengths of the coating materials under the 10 N load, as seen in Fig. 22a; therefore, corner crack would not occur in these specimens, at this load level, which accords with the experimental observation. For the Tribaloy specimen and the superalloy specimens, the maximum principal stresses induced in the coating surfaces under the 100 N load are all below the tensile strengths of their coating materials so that both edge crack and corner crack would not occur under this load.

With increasing the indentation load to 35 N, the maximum principal stresses along the indentation edge are larger than the tensile strength of the hard metals; therefore, the edge crack would be developed over the entire indentation edge in specimens WC-Co and WC-Co-Cr. Meanwhile, the maximum principal stress at the indentation corner also reaches the tensile strength of the coating material, see Fig. 26a, which would lead to the corner crack. However, as the stress decreases rapidly along the outward diagonal of the indentation, the corner crack would not grow at this load level. The experimental observation is consistent with the FEA results for specimen WC-Co, but for specimen WC-Co-Cr, the indentation test showed that no corner cracks occurred under the 35 N load. This discrepancy may be due to the coating material property. As indicated above, there are no available data for the tensile strength of the WC/10Co/4Cr hard metal; 200 MPa is only generally or approximately for WC hard metals. Therefore, it can be a case that the real tensile strength of WC/10Co/4C is greater than 200 or 250 MPa. Under the 100 N indentation load, however, the maximum principal stresses over the entire indentation edge and in a certain region of the indentation corner are greater than the tensile strength of the coating material for the hard metal specimens (see Figs. 25a and 26a). As a result, the edge crack would grow over the entire indentation edge; at the same time, the corner crack would propagate along the outward diagonal of the indentation to a distance, as observed in the indentation test.

For the high-load group specimens, the maximum principal stress in all the specimens under the load of 300 N is below the tensile strength of their coating materials (see Figs. 25b and 26b); therefore, no cracking would occur in the specimens at this load level, same as observed in the Vickers indentation test. With increasing the load to a level of 500 N, the maximum principal stress at the indentation edges in the T-800 specimen is 665 MPa, which is below the tensile strength of the coating material, 700 MPa, so that cracking is not expected to occur in this specimen under this load either. However, this stress level is very close to the tensile strength of the coating material, and also defects or flaws were possibly introduced during the coating fabrication process, which may result in microcracks occurring under the lower load. As observed in the indentation test, microcracks were generated at the indentation edges under the load of 500 N and the load of 700 N.

As reported by previous research, Nistelle C and Nistelle Super C have almost the same tensile strength, which is about 530 MPa. When the indentation load is below 500 N, the maximum principal stresses in these coating specimens are all less than 530 MPa, as seen in Fig. 25b and 26b, which implies that cracks would not occur in this loading case. This is consistent with the indentation test, but a higher load of 500 N would increase the stress level in the specimens, leading to cracking, as demonstrated by the indentation test. Under the load levels being discussed, the maximum principal stresses in the indentation corner area of the Nistelle C and Nistelle Super C specimens are lower than the tensile strengths of the coating materials; therefore, no cracking would occur at the indentation corner in these coatings under this load, but further loading may induce corner cracks in the specimens. In addition, under a same indentation load the stress levels in specimens Nistelle C and Nistelle Super C are very close and they are lower than the stress in the T-800 coating. As reported in Table 1, T-800 has a similar Young's modulus to Nistelle C and Nistelle Super C so that the difference in the stress may be due to the different coating thickness.

Conclusions

Among the five coatings being studied, WC-Co and WC-Co-Cr hard metals have the lowest fracture strength, characterized by significant cracking under a low indentation load: 10 N. This is due to their low tensile strength and brittleness. T-800 Tribaloy alloy has the highest resistance to fracture. It was not seriously cracked even under high indentation loads: 500 and 700 N. This is because it possesses high tensile strength. Nistelle C and Nistelle Super C have good ductility so that they did not exhibit obvious cracking even under high indentation loads: 500 700 N. Only micro or small cracks were observed in the indented surfaces under high loads.

The FEA stress results are in good agreement with the Vickers indentation test. In consideration of the maximum principal stress in the coating/substrate systems, the highest tensile stress occurs in the center of the indentation edge and the stress decreases along the indentation edge toward the indentation corner under the indentation load, which induced the edge crack first in the coating surface and promoted the edge crack growing along the indentation edge. The area of the indentation corner also has high tensile stress, which induced the corner crack and caused the crack to propagate radially along the outward diagonal of the indentation.

Acknowledgment The authors are grateful for financial support from the Natural Science & Engineering Research Council of Canada (NSERC), in-kind support form National Research Council Canada, and both financial and in-kind support from Deloro Stellite Inc.

Referecnes

(1.) Wielage, B.Wank, A, Pokmurska, H, Grund. T, Rupprecht, C, Reisel. G, Friesen, E, "Development and Trands in HVOF Spraying Technology." Surf. Coat. Technol., 201 2032 (2006)

(2.) Wang, YY, Li, CJ, Ohmori, A, "Examination of Factors Influencing the Bond Strength of High Velocity Oxy-Fuel Spraved Coatings." Surf. Coat. Technol., 200 2923 (2006)

(3.) ASTM D4541, "Standard Test Method for Pull-Off Strength of Coatings Using Portable Adhesion Tester," ASTM International 

(4.) ASTM C633-79, "Standard Test Method for Adhesion or Cohesive Strength of Flame-Sprayed Coatings," ASTM International

(5.) Kuwahara, K, Hirota, H, Umemeto N, "Adhesion Measurement of Thin Evaporated Films." In: Mittal, KL (ed.) Adhesion Measurement of Thin Films, Thick Films, Bulk Coatings. p. 122. STP 640, ASTM, Philadelphia. PA (1978)

(6.) Rosa, G, Oltra, R. "Evaluation of the Coating-Substrate Adhesion by Laser-Ultrasonics: Modeling and Experiments." J. Appl. Phys., 91 (10) 6744 (2002)

(7.) Aravas, N, Kim, KS, Loukis, MJ, "On the Mechanics of Adhesion Testing of Flexible Films." Mater. Sci, Eng. A, 107 159 (1989)

(8.) Arai. T, Fujita, H, Watanabe, M, "Evaluation of Adhesion Strength of Thin Hard Coatings." Thin Solid Films, 154 387 (1987)

(9.) Shaw, LL, "Measurement of the Fracture Engery of Interfaces in Composites through Sandwich-Type Chevron-Notched Specimens." J. Text. Eval., 25 (4) 391 (1977)

(10.) Shaw, LL, Barber, B, Jordan, EH, Gell, M, "Measurements of the Interfacial Fracture Energy of Thermal Coatings." Scr. Mater., 39 (10) 1427 (1998)

(11.) Charalambides, PG, Lund, J, Evans, AG, McMeeking, RM, "A Test Specimen for Determining the Fracture Resistance of Biomaterial Interfaces." J. Appl. Mech., 56 77 (1989)

(12.) Shaw, LL, Ren, R. Goberman, D, "Measurements of the Fracture Energy of the Coasting/Specimens." Surf. Coat. Technol., 130 74 (200)

(13.) Malzbender, J, "Comparison of Surface and Cross-Sectional Indentation in a Coating." Surf. Coat. Technol., 201 3797 (2006)

(14.) Lima, MM, Godoy, C, Modenesi, PJ, Avelar-Batista, JC. Davison, A, Matthews, A, "Coating Fracture Toughness Determined by Vickers indentation: An Important Parameter in Cavitation Erosion Resistance of WC-Co Thermally Sprayed Coatings." Surf. Coat. Technol., 177-178 489 (2004)

(15.) Rother, B, Dietrich, DA, "Evaluation of Coating-Substrate Interface Strength by Differential Load Feed Analysis of Load-Indentaiton Measurements." Thin Solid films, 250 181 (1994)

(16.) Steinmann, PA, Tardy, Y, Hintermann, HE, "Adhesion Testing by the Scratch Test Method: The Influence of Intrinsic and Extrinsic Parameters on the Critical Load." Thin Solid Films, 154 333 (1987)

(17.) Marshall, DB, Evans, AG, "Measurement of Adherence of Residually Stressed Thin Films by Indentaiton, I. Mechanics of Interface Delamination." J. Appl. Phys., 56 (10) 2632 (1984)

(18.) Rossington, C, Evans, AG, Marshall, DB, Khuri-Yakub, BT. "Measurements of Adherence of Residually Stressed Thin Films by Indentation. II. Experiments with ZnO/Si." J. Appl. Phys., 56 (10) 2639 (1984)

(19.) Gao, XL, Jing, XN, Subhash, G, "Two New Expanding Cavity Models for Indentation Deformations of Elastic Strain-Hardening Materials." Int. J. Solids Struct., 43 2193 (2006)

(20.) Lawn, BR, Swain, MV, "Microfracture Beneath Point Indentations in Brittle Solids." J. Mater. Sci,m 10 113 (1975)

(21.) Cook, RF, Pharr, GM, "Direct Observation and Analysis of Indentation Cracking in Glasses and Ceramics." J. Am. Ceram. Soc., 73 (4) 787 (1990)

(22.) Suzuki, H, Cemented Carbides and Sintered Hard Materials. Maruzen, Tokyo (1986)

(23.) Imasato, S, Tokumoto, K, Kitada, T, Sakaguchi, S, "Properties of Ultra-Fine Binderless Cemented Carbide 'RCCFN'." Int. J Refract. Met. Hard Mater., 13 305 (1995)

(24.) Davis, JR, ASM Specialty Handbook: Nickel, Cobalt and Their Alloys. ASM International, Materials Park, OH (2000)

(25.) Cameron, CB, Ferriss, DP, "Tribaloy Intermatallic Materials: New Wear - and Corrosion-Resistant Alloys." Anti-Corros. Methods Mater., 22 (4) 5 (1975)

(26.) Mason, SE. Rawlings, RD. "Fracture Behavior of Two Co-Mo-Cr-Si Wear Resistant Alloys." J. Mater. Sci., 20 (4) 1248 (1985)

(27.) Schmidt. RD, Ferriss, DP, ";New Materials Resistant to Wear and Corrosion to 1000[degrees] C." Wear, 32 (3) 279 (1975)

(28.) Crook, P, "Corrosion Characteristics of the Wrought Ni-Cr-Mo Alloys." Mater. Corros.. 56 (9) 606 (2005)

(29.) Brill, U. Klower. J. Nickel Alloys. Marcel Dekker, New York (1998)

(30.) Dong, JX. Zhang, MC. Zeng, YP, Xie, XS "Freckle Formation and Characterization of a Ni-Cr-Mo-Fe-W Alloy." Acta Metall. Sin., 17 (4)414 (2004)

(31.) Huang, P. Liu, R, Wu, XJ. Yao, MX. "Mechanical Properties and Corrosion Resistance of a Novel Ni-Cr-Mo Alloy." Adv. Eng. Mater., 9 (1-2) 60 (2007)

(32.) Oliver, WC, Pharr, GM, "An Improved Technique for Determining Hardness and Elastic Modulus Using Load and Displacement Sensing Indentation Experiments." J. Mater. Res., 7 1564 (1992)

(33.) Musil, J, Kune, F, Zeman, H. Polakova, H. "Relationships Between Hardness, Young's Modulus and Elastic Recovery in Hard Nanocomposite Coatings." Surf. Coat. Technol., 154 304 (2002)

(34.) Benham, PP, Crawford, RJ, Armstrong, CG. Mechanics of Engineering Materials (2nd ed.). Longman, UK (1996)

(35.) Lu. S-P, Kwon, O-Y, "Microstructure and Bonding Strength of WC Reinforced Ni-Base Alloy Brazed Composite Coating," Surf. Coat. Technol., 153 40 (2002)

A. P. Buang, R. Liu

Department of Mechanical and Aerospace Engineering, Carleton University, 1125 Colonel By Drive, Ottawa, ON, Canada KIS 5B6

e-mail: [e-mail]

X. J. Wu

Institute for Aerospace Research, National Research

Council Canada, 1200 Montreal Road, Ottawa, ON, Canada KIA 0R6

M. X. Yao

Deloro Stellite Inc., Belleville, ON, Canada K8N 5C4