Development of Near Net Shaped Coatings for Reduced Postprocessing Costs in Valves

Thermally sprayed WC/CoCr coatings are established in the valve industry for wear protection. However, conventional coatings have to be cost-intensive postprocessed. Therefore, the aim of this study is to develop near net shaped (nns) WC/CoCr coatings with a high wear resistance in order to avoid the expensive grinding postprocess. For the development of the nns coatings, a parameter study is used to investigate the influence of the stand-off distance and hydrogen volume flow rate in the HVAF process. The parameter study indicates the influence of the hydrogen volume flow and the stand-off distance on the porosity, microhardness and wear resistance of the coatings. The developed coatings exhibit a low porosity and high wear resistance. Through the correlation of the coating properties with the process parameters, promising parameter ranges for a further development of HVAF-sprayed nns coatings of WC/CoCr can be identified. With these results, first benchmarks for HVAF process parameters, hydrogen flow rate and stand-off distance could be delivered, advancing the overall goal of reducing manufacturing costs of valves.


Introduction
Valves are used in almost every industry to regulate various media flows. Wear and corrosion can cause a leakage of the valve. Such a leakage must be avoided, especially in the case of media that are harmful to the environment or health. To ensure this, thermally sprayed coatings are conventionally used in the valve industry. In particular, thermally sprayed coatings consisting of WC/CoCr are the most commonly used coating in many valve applications (Ref 1). Due to the high wear protection of thermally sprayed WC/CoCr coatings, they are used, for example, in high-erosive stressed shut-off valves in the oil industry as it can be seen from the different literature sources ( Ref 2,3). In order to have a sufficient barrier function, coating thicknesses of t = 300 lm are typically used. In the study (Ref 4), various coatings were developed for the use in the valve industry. They showed, on the basis of HVOFsprayed coatings, that wear can be significantly reduced by the integration of solid lubricants. By integrating hexagonal boron nitride into WC/CoCr, the wear volume was reduced by 75% in a pin-on-disk test against an Al 2 O 3 counterbody. However, since such coatings are too rough in as-sprayed conditions, the surfaces must be ground before they can be used in valves.
By reducing the surface roughness of the coating in assprayed condition, the required surface roughness for the application could be achieved by direct polishing without grinding. Fedrizzi et al. (Ref 5) achieved this goal by using Cr 3 C 2 /NiCr feedstock material with a particle size smaller than dp \ 10 lm and an achieved as-sprayed surface roughness of Ra = 2.2 lm. Other references report a necessary surface roughness in the as-sprayed condition of Ra \ 2 lm in order to avoid the grinding process ( Ref 6). To achieve coatings with high contour accuracy and low surface roughness, feedstock materials with a particle size dp \ 15 lm are recommended in ( Ref 7). Such feedstock materials with fine grain size distribution were applied with different thermal spraying processes. Baumann et al. (Ref 8) applied WC-Co powders with a grain size distribution of f = -10 ? 2 lm by means of warm spraying. They investigated the influence of the stand-off distance in a range of s = 10 -50 mm on the surface roughness in assprayed condition and on the porosity. They showed that with a stand-off distance of s = 50 mm, a reduction in the as-sprayed surface roughness to Ra = 1.65 lm and porosity to U = 0.23% is possible. However, the formation of the undesired Co 6 W 6 C phase was also observed in this publication. With smaller stand-off distance, the smaller particles can be deflected in the gas jet by the impingement flow, resulting in a higher surface roughness and porosity.
In (Ref 9), WC/Co feedstock materials with a grain size distribution of f = -8 ? 1 lm were applied with a hydrogen supported HVOF system. Due to the use of the fine feedstock material, the microhardness of the coating could be increased by 30% compared to the coating applied with a conventional grain size distribution of f = -45 ? 15 lm. The researchers attribute this to the finer distribution of the finer carbides and the higher density of the coating. Tillmann et al. (Ref 10) analyzed the wear behavior of HVOF-sprayed WC-Co coatings that were applied with a conventional grain size distribution of f = -45 ? 15 lm and a fine grain size distribution of f = -10 ? 2 lm. They demonstrated in a pin-on-disk test against an Al 2 O 3 counterbody that the wear coefficient could be reduced by 81% for the coating, which was applied with the fine grain size distribution. In case of the coating, which was applied with the conventional grain size distribution, there is a high Hertzian contact pressure on just a few carbides during the wear test. This causes some carbides to break into smaller fragments, which break out of the coating during the wear test, resulting in an increased wear volume. In the coating, applied with the fine feedstock material with finer carbides, the load is distributed over several smaller carbides, which prevents the carbides from breaking. In addition, the reduced mean free path of the binder phase reduces the wear of the matrix, which in turn allows the carbides to be held better in the matrix. The results show that feedstock materials with fine grain size distribution can be used to apply dense and thin coatings with high wear resistance. However, due to the higher specific surface area and lower mass, fine feedstock materials can interact stronger with the hot gas jet. As a result, the particles are heated up more, which can lead to undesired phase transformations (Ref 8). Lanz et al. (Ref 11) compared WC-CoCr fine feedstock material coatings that were applied by an HVOF and HVAF system. They showed that coatings with a high density and suitable phase composition can be applied with the HVAF process. In the sand/rubber wheel test, the HVAF-sprayed coatings showed a wear volume that was lower by a factor of 2 than the wear volume of the HVOF-sprayed coating. Undesired phase transformations in the HVOF-sprayed coatings are the reason for the higher wear volume. The phases, formed during the spraying process, cause a more brittle material behavior in the wear test, resulting in breakouts. A more ductile material behavior is necessary to provide a sufficient wear protection without particle breakouts. However, only a few studies exist in this field of research so far, but they demonstrate that the HVAF process can be a good candidate for the application of thin, dense and near net shaped coatings by the use of fine feedstock materials.
Thin and near net shaped WC/CoCr coatings can reduce the deposition times and can avoid cost-intensive grinding postprocesses. Therefore, such coatings can be an economical alternative to conventional coatings in the valve industry. However, to ensure tightness and prevent undercorrosion, the coatings must be extremely dense. HVAFsprayed fine feedstock material coatings could be the solution to meet such requirements, but parameter studies are necessary to apply suitable coatings for the valve industry. In this work, such a parameter study is conducted. The developed coatings are characterized in terms of their coating properties and are correlated with the results of the tribological investigations by means of pin-on-disk tests.

Experimental
The stainless steel 1.4404, which is conventionally used in the valve industry, is used as substrate. Before the application of the coatings, the substrates were roughened with corundum F16 at a pressure of p = 0.4 MPa. The AK7 HVAF system from Kermetico Inc., Oregon, USA, was used for the application of the coatings. In total, 20 passes were used for each coating, with a cooling cycle after 10 passes. WC/CoCr with a grain size distribution of f = -15 ? 5 lm was used as a feedstock material. The coating structure was evaluated on cross sections using the Axio Imager light microscope from Zeiss, Oberkochen, Germany. To evaluate possible decarburization and oxidation processes, the qualitative distribution of oxygen and carbon was analyzed by scanning electron microscopy (SEM) and energy-dispersive x-ray spectroscopy (EDX).
The REM/EDX-system Phenom XL from Thermo Fisher, Waltham, USA, was used. The determination of the porosity was conducted on the basis of cross sections using the program Fiji on the basis of machine learning algorithms, WEKA. More detailed information is found in (Ref 12). A total of 10 images in 500 9 magnification were produced for each coating. The machine learning algorithms were trained on one image in each coating. After the training phase, the porosity was determined on the remaining images using the same algorithm. The surface roughness was determined in the as-sprayed condition using the VK-X210 confocal laser microscope from Keyence, Osaka, Japan. Fine feedstock materials have a higher specific surface-area-to-volume ratio compared to conventional grain size distributions, so they can be heated up more in flight, and undesired phases can be formed during the spraying process. Therefore, the phase composition was determined in the as-sprayed condition by x-ray diffraction (XRD) in the range of 2H = 15-100°. The diffractometer system XRD 3000 from Waygate Technologies, Hürth, Germany, with a Cu Ka radiation was used for this purpose. A radiation time of t = 120 s and a step size of x = 0.05°/s was applied. The microhardness was determined with the microhardness system Micromet 1 from Buehler, Lake Bluff, USA, using a Vickers indenter at 20 locations in each cross section according to DIN E-N ISO 6507. To reduce the influence of the substrate and the embedding medium, all indentations were performed in the middle of the coatings. A load of F = 0.98 N (m = 100 g) and a holding time of t = 15 s were used for these tests. The fracture toughness of the coatings was determined by the indentation method as described in (Ref 13). The cracks were created using a Vickers indenter at a load of m = 5 kg for a conventional HVOF-sprayed coating which was applied with a conventional particle size distribution of f = -45 ? 15 lm and a coating thickness of t & 300 lm. For the most promising thinner nns coating P--, a load of m = 2 kg is used. The Young's modulus E, which is necessary for the determination of the K IC value, was determined from the indentation modulus E IT and the Poisson's ratio m according to Kan et al. (Ref 14). The indentation modulus E IT was determined using fisherscope HM2000 from Helmut Fischer GmbH, Sindelfingen, Germany. The crack lengths were measured using the Axio Imager light microscope from Zeiss, Oberkochen, Germany.
Due to impurities in the medium, particles can get between the sealing surfaces of a valve, resulting in a strong abrasive wear. In order to analyze the wear protection of the coatings against an abrasive load, pin-on-disk tests were conducted at a load of F = 15 N, a wear distance of l = 1000 m and a relative velocity of v = 100 mm/s. An Al 2 O 3 ball with a diameter of d = 6 mm was used as a counterbody. The wear volume after the test was analyzed by confocal laser microscopy. For this purpose, the same laser microscope was used as for the determination of the surface roughness. The wear area AV was measured at 9 locations in the wear track. Each test was repeated twice. The wear coefficient K was determined from the average wear surface, the wear track radius r, the load F and the wear distance l according to (Ref 15). In order to evaluate the wear mechanisms, all wear tracks were analyzed in more detail by means of optical microscopy and laser microscopy.

Results
In Fig. 1, the constant and variable process parameters are shown in form of a full factorial experimental design. Based on the coating at the center point, a smaller (-) and larger (?) stand-off distance and hydrogen volume flow were investigated. The stand-off distance is defined in the following as the first factor and the hydrogen volume flow as the second factor. In the following, the developed coatings are named according to the applied factors in the process point P, whereby the first factor is named first. For example, the coating applied at the process point P with small (-) stand-off distance and large (?) hydrogen volume flow has the notation P-?. The coating in the center point has the notation C. The particle temperature and velocity are both significantly influenced by the hydrogen volume flow. Higher hydrogen volume flows result in hotter and faster gas jets as well as in a higher heat transfer coefficient, which consequently accelerates and heats the injected particles more strongly during the spraying process. The dwell time of the particles in the gas jet is influenced by the stand-off distance. Higher stand-off distances allow the particles to be accelerated and heated for a longer time. However, if the stand-off distance is too high, the particles can also cool down again, causing them to flatten less upon impact with the substrate. With this full factorial experimental design, the correlations between the process parameters and coating characteristics should be determined. Figure 1 also depicts the cross sections of the developed coatings. The comparison of the coatings with a small stand-off distance shows that spalling can be observed at the higher hydrogen volume flow. It is known from the literature that high velocity flame spraying processes generally induce residual compressive stresses in the coating (Ref 16,17). A certain degree of residual compressive stresses can be useful in tribological and corrosive applications, as they can suppress the crack propagation. However, excessive residual compressive stresses can initiate spalling ( Ref 18). Finer particles are more strongly accelerated in the gas jet and reach a higher kinetic energy in flight, which induces higher residual compressive stresses. In the case of the high hydrogen volume flow, this effect can be observed. With smaller hydrogen volume flow, such spalling in the coating P--can hardly be observed. The ratio of particle temperature and velocity seemed to be suitable in this process point to produce thin and dense coatings without spalling.
In the case of the coating P?? with a high stand-off distance and high hydrogen volume flow, spalling can be detected as observed in coating P-?. Residual compressive stresses can also be suspected as the reason in this coating. In coating P?-, spalling can hardly be detected. However, more pores can be identified compared to the coating P--, which was deposited at the same hydrogen volume flow. The pores are very small and circular. A possible explanation is that particles are atomized and cooled down in flight, due to the long dwell time in the gas jet. This causes the particles to flatten less upon impact, resulting in a decreased bonding of these particles within the coating structure and in particle breakouts during the cross section preparation. Since such atomized and cooled down particles have a spherical shape in flight, the size and shape of the observed pores can be explained. Higher hydrogen flows provide higher particle velocities, resulting in denser coating structures. However, the higher particle velocities can also lead to spalling. A higher stand-off distance leads to a premature cooling of the particles during the flight, which creates less bonded particles in the coating structure. At the center point, both factors seem to influence the coating structure, since both spalling and pores can be detected in coating C.
To study the coating structures in more detail, investigations were conducted using a scanning electron microscope (SEM) with integrated energy-dispersive x-ray spectroscopy (EDX). Spalling can be observed in the coatings P-? and P?? that were applied with increased hydrogen volume flow. SEM/EDX images show that fine oxides are present at some splat boundaries in these coatings, see Fig. 2(b). Such areas may favor spalling. Particle breakouts can be observed in the coatings P?? and P?-, which were applied with high stand-off distance. SEM/ EDX images show that circular particles are present in these coatings. At the edge of such particles, oxides can be identified that favor breakouts during metallographic preparation, see Fig. 2(c). In the coating P--, oxygen can hardly be identified, see Fig. 2(a). The carbon distribution appears to be homogeneous.
To quantify the porosity, a program that uses machine learning algorithms was trained and applied for the analysis of the light microscopic images. The results of this analysis are shown in Fig. 3(a). In order to provide a high density of the coatings, a maximum coating porosity of U B 0.5% was set in consultation with valve manufacturers. This target could only be reached with the coating P--. The observed small circular pores were found in coating C and P?-. This can also be seen from the higher porosity values of the coatings C and P?-. In coating C, small cracks and spalling could also be detected by the program, which explains the higher porosity compared to coating P?-. The coatings P-? and P?? at high hydrogen   volume flows show no or only a few pores. However, the coatings P-? and P?? are characterized by spalling, which is why cracks were detected in the coatings, resulting in higher porosity values. Grinding postprocesses are a major cost factor for the construction of valves. Low surface roughnesses in the assprayed conditions should eliminate the need for such costintensive grinding postprocesses. To achieve this goal, a target value of Ra B 1.5 lm in the as-sprayed condition was defined. Coating P--showed that the lowest porosity and spalling could hardly be observed. This is also shown in the low surface roughness, see Fig. 3(b). With coating P--, the target value could be almost achieved. Direct polishing of this coating was possible. In coating P-?, the particles seemed to cool down in flight, causing them to flatten less upon impact. In the cross section of this coating, this effect was visible by small circular pores. A higher surface roughness can be measured in this coating, due to the decreased flattening of the particles. The coatings with the high hydrogen flow had a higher surface roughness than the coatings with the low hydrogen flow. The reason for this is the presence of spalling on the surface. At process point C, a lower hydrogen volume flow was used than for the coatings P-? and P??, which results in cooler and slower particles before impact. These particles flatten less upon impact with the substrate, which also explained the observed pores in the cross section. However, the kinetic energy of the particles was still sufficient that spalling was observed in this coating. Both effects are the reason for the high-measured surface roughness.
The microhardness values of the coatings are presented in Fig. 4 in form of box plots. In this representation, 75% of all hardness values are above the first quartile and below the third quartile. The median of the measured values are also presented. All coatings have an average microhardness of over 1300 HV0.1. In particular, the coatings applied with short stand-off distances show higher microhardness values. Especially for the coating P--, a median microhardness of 1550 HV0.1 is achieved. The high microhardness of the fine feedstock material coatings can be explained by the denser coating structure and the induction of residual compressive stresses (Ref 10). Increased residual compressive stresses can be negative and reduce the microhardness again. The small difference between the third and first quartile shows that the microhardnesses are homogeneously distributed in the coatings. Only coating C shows a larger difference between the first and third quartile. In this coating, spalling and pores were observed. Both effects can explain the local differences of the microhardness in this coating.
The phase composition was determined by XRD. The results of the investigations are shown in Fig. 5. The main peaks of the diffractogram can be assigned to the WC phase. The undesired g-phase could be detected in each coating. However, the low intensity of the peaks also shows that only a small amount of this phase is present in the coating. The main peak of the W 2 C phase at 2h = 40°s hows that the intensity increases slightly with increasing hydrogen volume flow. Due to the higher particle temperatures, a stronger phase transformation in flight can be given as a reason for this observation. Before the wear test, all coatings were polished to a surface roughness of Ra = 0.02 lm. Preliminary tests of coating P--showed that this coating could be polished directly. However, since the other coatings had a higher assprayed surface roughness, the coatings were first ground on a 10 lm diamond disk and subsequently polished on a 6 lm disk before the start of the wear test. This procedure was done for all coatings in order to have the same surface conditions before the test. The determined wear coefficients after the wear test are shown in Fig. 6. Coating P-has the lowest wear coefficient. Due to the dense coating structure, the particles are held effectively in the coating structure, which prevents breakouts during the wear test. The coating P-?, which was applied at the same stand-off distance, but with a higher hydrogen flow, demonstrated a 47% higher wear coefficient than P--. The coating exhibited some spalling in the cross section. Due to the tribological loading, particle breakouts were initiated in the coating. Such particles remain in the wear track and act as abrasive particles, whereby the wear increases. The coatings at high stand-off distance show a comparable wear coefficient, taking the standard deviation into account. The mean value of the wear coefficients reveals a slightly higher wear for coating P-?, which was applied with small hydrogen flow, compared to coating P??. Coating C in the center point has the highest wear coefficient. Spalling and pores were identified in this coating. Both observations suggest that some particles were less bonded in the coating structure, allowing particles to break out in the wear track. The goal of a higher abrasion protection compared to a coating, which was sprayed with a conventional grain size distribution of f = -45 ? 15 lm, could be achieved with all coatings as shown by the lower wear coefficient compared to the target value.
In order to analyze the wear behavior of the coatings in more detail, laser microscopic images of the wear tracks were conducted for all coatings, see Fig. 7. Abrasive wear is the dominant wear mechanism for all coatings. The coating P--exhibits the lowest wear coefficient. This can also be seen in the wear track. Only a few carbide breakouts can be identified. Such carbides remain in the wear track and cause micro-grooving. The coating P-?, which was applied at the same stand-off distance and high hydrogen volume flow, shows significantly bigger particle outbreaks. It can be assumed that such particles break out due to the combination of the tribological load and the residual compressive stresses in the coating. Such particles can also remain in the wear track and cause abrasion, as it can be seen in the wear track. The coatings with the high stand-off distance show breakouts in the wear track, mostly carbides. In coating P??, the combination of residual compressive stresses and tribological loads can be named as reason for the observed breakouts. In coating P?-, circular pores were found in the cross section, which were attributed to less bonded particles. The circular breakouts in the wear track correlate with the observations in the cross section. In comparison with the coating P??, more of these breakouts can be observed in the coating P?-, which in turn correlates with the observed higher wear coefficient. The coating in the center point exhibits bigger breakouts in the wear track. This also correlates with the higher wear coefficient. The goal of a higher protection against abrasion compared to a coating, which was sprayed with a conventional grain size distribution of f = -45 ? 15 lm, could be achieved with all coatings.
In order to determine the influence of the hydrogen volume flow and the stand-off distance on the characteristic coating properties microhardness, porosity and surface roughness and to demonstrate the correlations between these properties, interpolation graphs were made using the program Origin Pro from Origin Lab Corporation, Northampton, USA. These graphs are shown in Fig. 8. For the microhardness, a clear increase can be observed with lower hydrogen volume flow and lower stand-off distance. This trend also correlates with a lower porosity and lower surface roughness. The microhardness cannot be directly correlated with the wear resistance, since besides the hardness of the coating, a certain toughness is necessary to provide a high wear protection. In combination with the lower porosity, the splats stay in the coating even under the applied tribological load, whereby this trend also correlates with a lower wear coefficient. The particle temperature and velocity upon impact seems to be sufficient to create strongly flattened splats, resulting in a lower surface roughness. This observation also explains the strong correlation to the porosity, the microhardness and the wear Fig. 6 Wear coefficients of the WC/CoCr coatings sprayed by HVAF after the pin-on-disk test J Therm Spray Tech (2023) 32:681-692 687 coefficient in the range of low hydrogen flows and standoff distances. In the area of low stand-off distances and high hydrogen volume flows, a decrease in microhardness and an increase in porosity and surface roughness can be observed. The coating in this area has some spalling, which can also account for the higher surface roughness. Cracks in the coating are responsible for the increase in porosity. Spalling in this area can be named as reason for the higher wear coefficient. In the area with high stand-off distance and low hydrogen volume flow, the strongest decrease in microhardness as well as some increase in porosity and surface roughness can be observed. Light microscopic images of the cross section of the coating in this area indicate breakouts of smaller circular particles in the coating structure. The less bonded particles lead to breakouts in the wear track and explain the higher wear coefficient of the coating in this area. In the center of the interpolation graphs, there is a low microhardness as well as high porosity and surface roughness. This can be explained by less bonded particles and spalling. The wear coefficient is also the highest at this point. At high hydrogen volume flow and high stand-off distance, comparable porosity and microhardness can be observed compared to the coating in the center point C, but the surface roughness and wear coefficient are slightly lower. Compared to the coating C, particles increasingly flatten upon impact on the substrate, resulting in fewer poor bonded particles in the coating structure. This is also evident from the lower surface roughness of coating P?? compared to coating C and can be seen as a reason for the slightly higher wear resistance of coating P?? compared to coating C. The observations show that the characteristic coating properties microhardness, porosity and surface roughness can be correlated with the wear coefficient. In particular, the combination of high microhardness, low porosity and low surface roughness provides high wear protection against abrasion. However, the investigations also show that residual stresses and crack In addition to a high wear resistance, valve coatings must also provide a sufficient protection against corrosion. Since the coating P--already exhibited promising coating and wear properties, the electrochemical corrosion behavior was investigated for this coating in a 5% NaCl solution with a pH value of pH = 7. For the polarization experiments, the commercially available potentiostat Reference 600 ? and measuring cell ParaCell from Gamry, Warminster, USA, were used. The reference electrode was a Hg|Hg 2 Cl 2 |KCl electrode. A graphite plate functioned as counter electrode and the investigated samples were used as working electrode. The corroding surfaces of the investigated samples were ground to a surface roughness of Ra = 0.1 lm. In order to investigate a defined corrosion area, a masking foil with an exposed area of A = 1 cm 2 was applied. For the potentiodynamic measurement of the current density, a potential feed of _ E = 0.5 mV/s in a range of E = -250 ? 1000 mV against the open current potential was investigated. To evaluate the corrosion behavior of the coatings after the corrosion test, cross sections of the corroded areas were analyzed. For this purpose, the corrosion areas were embedded with an epoxy resin after the test to avoid a detachment of corrosion products during the metallographic preparation. In order to evaluate the corrosion resistance, an in the valve industry conventionally used coating made of WC/CoCr, which was applied with of a conventional powder size distribution of f = -45 ? 15 lm, was also investigated. Results regarding the coating properties and the wear behavior of this coating can be found in (Ref 4). In addition, the substrate made of 1.4404, which was used for the conventionally applied WC/CoCr coating and for the coating P--, was also investigated. The results of these investigations are shown in Fig. 9. The current density curve of the developed coating P-exhibits a current density that is almost one order of magnitude lower than of the conventional WC/CoCr coating over the entire potential range. This higher corrosion protection can be attributed to the denser coating structure of the novel developed coating. A comparison of the coating structures of the conventional WC/CoCr coating and the coating P--shows that the corrosion mechanism of the conventional WC/CoCr coating is based on the infiltration of the electrolyte by means of cracks and pores. In the case of the coating P--, the electrolyte can hardly infiltrate into the coating, which is the reason why the corrosive attack occurs much closer to the surface. Only a few fine cracks can be identified in which the electrolyte can infiltrate. Conventional WC/CoCr coatings are used in valves with a coating thickness of s = 300 lm.
The investigations show that with this coating, a sufficient corrosion protection can be ensured in the investigated medium. As the results show, with the new developed coating P--, such a corrosion protection is also possible with significantly thinner coating thickness. The formula F1 as described in (Ref 13) and shown in Fig. 10 was used to calculate the fracture toughness K IC . Bolelli et al. (Ref 19) used this formula in a previous study to investigate the fracture toughness of WC/CoCr coatings applied with feedstock materials with larger particle size distributions compared to this study. Figure 10 also illustrates the fracture toughness of a conventional HVOFsprayed WC/CoCr coating compared to the coating P--. Both coatings are in a comparable range as described in the study of Bolleli et al. (Ref 19). The coating P--, which was applied with the finer feedstock material, shows a slightly reduced fracture toughness compared to the conventional HVOF-sprayed WC/CoCr coating. Due to the smaller particles, there is a higher number of splat boundaries in the coating P--, which can favor crack propagation.

Summary
Thin and dense coatings with low as-sprayed surface roughness represent an economical alternative to conventional coatings in the valve industry. In this study, a WC-CoCr feedstock material with a grain size distribution of f = -15 ? 5 lm was used to develop such coatings. Previous studies have shown that a precise process control is necessary to avoid undesirable phase reactions during the spraying process. The HVAF system represents a promising process to apply such thin, dense and near net shaped coatings without undesired phase transformations. However, parameter studies are necessary to develop coatings that are suitable for valve construction. In this work, such a parameter study was conducted. The most important results are summarized in the following: Fig. 9 Current density-potential curves and cross section images of the developed coating P-in comparison with a conventional coating made of WC/CoCr and substrate made of 1.4404 after the corrosion test in a 5% NaCl solution at pH = 7 Fig. 10 Fracture toughness of a conventional HVOF-sprayed WC/CoCr coating compared to coating P--and the formula used for K IC calculation • The stand-off distance and the hydrogen volume flow have a major influence on the coating structure. High hydrogen volume flows cause spalling and particle breakouts in the wear track. High stand-off distances increase the porosity. • With coating P--, it was possible to develop a coating that was almost pore-free and showed hardly any spalling. The surface roughness of this coating was so low that direct polishing was possible without prior grinding. • The wear coefficient in the pin-on-disk test against Al 2 O 3 of all developed coatings was below the wear coefficients of a conventionally used WC/CoCr coating in valve industry. • Microhardness, porosity and surface roughness are promising parameters to determine the protection against abrasion. Nevertheless, the results also demonstrate that the residual stress state and the fracture toughness are further important coatings properties to determine the wear behavior more precisely. • Due to the dense coating structure of the developed coatings, sufficient corrosion protection can be achieved with significantly reduced coating thicknesses.
• The fracture toughness is slightly reduced compared to a conventional HVOF-sprayed WC/CoCr coating. It can be assumed that the increased number of splat boundaries favor the crack propagation.

Outlook
With the coating P--, a promising process point was identified which should be used in future to develop new types of coatings for the valve industry. In previous studies (Ref 4,20,21), it was shown that the integration of solid lubricants reduces the wear of conventionally used valve coatings. In subsequent studies, the integration of solid lubricants into HVAF-sprayed fine feedstock material coatings will be investigated. In addition, fine feedstock material coatings will be developed using the novel UHVOF process.