Effect of HVOF processing parameters on Cr₃C₂-NiCr hard coatings deposited on AISI 4140 steel

Effect of HVOF processing parameters on Cr3C2NiCr hard coatings deposited on AISI 4140 steel S. Sauceda a, ⇑ , S. Lascano a, ⇑⇑ , J. Núñez a , C. Parra b , C. Arévalo c , L. Béjar d a Departamento de Ingeniería Mecánica, Universidad Técnica Federico Santa María, Avda. Vicuña Mackenna N°3939, San Joaquín, Santiago, Chile b Laboratorio de Nanobiomateriales, Departamento de Ingeniería Mecánica, Universidad técnica Federico Santa María, Avenida España 1680, Valparaíso, Chile c Departamento de Ingeniería y Ciencia de los Materiales y del Transporte, E.T.S. de Ingeniería-Escuela Politécnica Superior, Universidad de Sevilla, Camino de los Descubrimientos, s/n, 41092 Seville, Spain d Facultad de Ingeniería Mecánica, Universidad Michoacana de San Nicolás de Hidalgo, Ciudad Universitaria, Morelia 58000, Mexico article info Article history: Received 22 June 2022 Revised 7 November 2022 Accepted 17 January 2023 Keywords: Thermal spray coating HVOF Experimental design Hard coatings abstract High Velocity Oxygen Fuel (HVOF) thermal spray coating is a deposition process that is widely employed at present to improve surfaces and the remanufacturing of components subject to severe wear, offering an alternative to reduce the cost of spare parts and used as a substitute for electro-chromium, also generating greater thickness than chemical vapor deposition and physical vapor deposition. This study evaluates the influence of the projected application parameters for Cr 3 C 2 20ðNi20CrÞcoatings on AISI 4140 steel using the HVOF technique. To do so, a 2 4 factorial experimental design was used to evaluate the effect on the thickness, porosity, surface hardness, and microhardness in the cross-section, produced by four of the main operational factors in the HVOF process: the fuel/oxygen (F/O) ratio entering the combustion chamber, powder flow, the roughness of the substrate before deposition, and spray gun speed. The morphology of the Cr 3 C 2 20ðNi20CrÞpowders and the microstructure of the coatings were studied through a scanning electron microscope and energy dispersive spectroscopy, while the composition were analyzed by X-ray diffraction. It was determined that: (i) the F/O ratio is the operation parameter with the most significant influence on the properties studied; (ii) the gun speed has a significant effect on the thickness generated and porosity; and (iii) the powder flow and roughness of the substrate have a greater effect on the surface hardness and microhardness. Finally, it was determined that coatings with great thickness, low porosity, and great hardness, ideal for the recovery of parts, can be achieved from a F/O ratio of 0.45 and a powder flow with the system feeder rotating at 12 rpm, applied to a substrate with a roughness of R a ¼18 lm, combined with a spray gun speed of 5 mm/s. Ó2023 Karabuk University. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). 1. Introduction There is currently an increasing trend towards products with shorter useful life and industrial waste which adds further complexity to the task of protecting the environment [1]. Surface engineering and additive manufacturing techniques can provide solutions aimed at reusing components with a substrate at a lower cost and reduced environmental impact, through the application of a coating with sufficient thickness and quality to recover the initial geometry lost due to severe wear and tear, which also helps reduce time, costs, and energy consumption, in comparison to the manufacturing of a new component [2,3]. There are different techniques to produce functional coatings that are resistant to wear and tear and corrosion, which can be applied to different shaped objects. In particular, the high velocity oxygen-fuel (HVOF) technique projects a powder at high supersonic speed generated by an increase in pressure from the combustion of kerosene and oxygen, adhering to the substrate through plastic deformation [4,5]. This process achieves highly adherent coatings, with low porosity (<2%), high microhardness (>800HV)[6,7], good resistance to corrosion [8–10], and great thickness (>200 l m) [11–14], which is why it is currently used in the industry to recover mechanical components such as tools, valves, turbine blades, axles, or decorative surfaces [15–19]. This technique is complex by nature and involves a large number of operating parameters that affect the formation of the coating, its microstructure, and mechanical properties. The effect of these parameters must be known in order to maximize thickness https://doi.org/10.1016/j.jestch.2023.101342 2215-0986/Ó2023 Karabuk University. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ⇑ Corresponding author. ⇑⇑ Principal corresponding author. E-mail addresses: [email protected] (S. Sauceda), [email protected] (S. Lascano). Engineering Science and Technology, an International Journal 39 (2023) 101342 Contents lists available at ScienceDirect Engineering Science and Technology, an International Journal journal homepage: www.elsevier.com/locate/jestch [3,20,21], decrease the presence of porosity in the coating, and obtain an adequate adherence to the substrate and cohesion among the sprayed layers. This helps avoid nucleation and the development of cracks that compromise quality [22–24] and guarantees that the mechanical properties are preserved during the processing cycle and use of the component [25,26]. Increase surface hardness to avoid wear and tear [26–28]. The microhardness in the cross-section of the coating helps us to know the state of some mechanical properties, such as Young’s modulus and Poisson’s ratio, since they follow the same trend [29]. In addition, it is a technique that allows us to achieve the coating quality without the difficulties of large sample sizes required in conventional mechanical tests on cylindrical parts applied in the industry [30]. This justifies the study of processing parameter’s effect on the coating quality using process mapping methodologies [31,32] and experimental design [33–35] to achieve an application protocol, control model, and process optimization. The effect of the variables associated with HVOF application on the physical and mechanical properties of coatings has been studied in literature with different experimental designs; Clavijo et al. [36] studied the influence of the fuel/oxygen (F/O) ratio, gun spray distance, and powder feed rate (PFR), on the speed and temperature of the TiO 2 particles projected, using 2 3 factorial designs; the first to study the effect on the speed and temperature of the particles, and the second with the same parameters at different values for the study of splats formed on the coating surface. In addition to these three input variables, Khan et al. [37] also analyzed the effect of thickness in a 2 4 factorial design on properties such as microhardness, contact angle, and cross-section porosity in WC 12Co coatings over AISI 4340 steel. In particular, the microhardness of the coating depends on the placement of splats and dispersion of phases in the coating, since a better distribution of splats and phases generates greater microhardness and better quality of the coating [38,39]. Meanwhile, J. Singh et al. [40] used Taguchi orthogonal factorial designs to study the behavior of coating wear and tear, for both WC 10Co 4Cr and Ni 20Cr 2 O 3 over SS304 steel. In the case of Cr 3 C 2 25ðNi20CrÞcoatings, the optimization method has been studied by Varis et al. [41] to understand the behavior of the stoichiometry parameters in combustion on the fatigue of these coatings applied to S355J2G3 structural steel. The studies of this coating have been oriented to optimize its wear and fatigue resistance, proving that this material is suitable for high-wear applications. However, further studies on the optimization of manufacturing based on remanufacturing and component recovery capacity, evaluating the densification and thicknesses are essential to achieved industrial applications. Nickel–chromium based coatings are widely used in the industry for the remanufacturing of components [3,42] thanks to the excellent resistance to corrosion, higher than that seen for tungsten carbide coatings [8,43], and their resistance to wear and tear exceeds that of electro-chromium. It is also possible to apply these to different types of steel. The wide acceptance of AISI 4140 steel makes it one of the most widely used steels. It uses in manufacturing rollers, turbines, and transmission shafts for the mining, energy, automotive, and oil industries [44,45]. However, despite numerous studies related to the properties of Cr 3 C 2 ðNiCrÞcoatings over stainless [46–48] and low-carbon steels [7,22,38,49], there are no optimization studies by experimental design for nickel–chromium based coatings on this type of steel. This study proposes to evaluate the effect of the process variables in the deposition of Cr 3 C 2 20ðNi20CrÞcoatings over an AISI 4140 steel substrate using the HVOF technique. We aim to generate hard coatings with thicknesses over 100 l m, low porosity (<2%), and high microhardness (>800 HV) to obtain dense and good-quality coatings that can be used for recovering components of cylindrical geometries with large dimensions, such as shafts, rods, piston rings, among others. For this, a 2 4 factorial design is proposed to evaluate the main effects and combined effects of the powder feed rate, gun speed, fuel/oxygen ratio, and roughness of the substrate on porosity, adherence, hardness, and final thickness. 2. Materials and methods This study used Cr 3 C 2 20ðNi20CrÞpowder with a size distribution of 45 ± 15 l m supplied by Sulzer Metco, USA (Woka 7102). The initial powder was characterized using Scanning Electron Microscopy (SEM), FEI QuantaTM 250 FEG SEM,equipped with energy-dispersive X-ray spectroscopy (EDS) for the analysis of local distribution and area of metallic and non-metallic elements in a powder particle. The images obtained from the powder surfaces were taken using 20 kV accelerating voltage. The phases found in the powder were obtained through X-ray Diffraction (XRD) using a STOE STADI MP diffractometer, a radiation of CuK a _1 (k¼0:15406 nm), one step from 0.02°, and a permanence time ranging from 30°to 70°at 2h. 2.1. Experimental design This study used a 4-variable and two-level (2 4 ) factorial design, to evaluate the effect of each of the variables on (i) thickness, (ii) internal porosity, (iii) surface hardness, and (iv) microhardness in the cross-section of the coatings deposited. For this, the following variables were taken: the fuel/oxygen ratio (A), the powder flow rate (B), the roughness of the substrate before application of the coating (C), and gun speed (D). The two levels used, as well as the parameters that remained constant, are shown in Table 1. The fuel/oxygen ratio and flow rate were chosen thanks to their strong influence on quality, residual tension, and coating thickness [50,51]. Roughness, on the other hand, is closely related to coating adherence and the formation of splats of the projected particles [52], while the longitudinal gun speed was considered because the goal is to reduce coating application times without sacrificing quality, therefore this must be studied in order to propose an optimal level. For all statistical analysis, the Statgraphics Ò Centurion XVIII software was used. A prediction model was prepared for each of the output variables using the linear first order equation, with interactions: ^ y¼b 0 x i þX K i¼1 b i x i þX i>i X i¼1 b ii x i x i þð1Þ where ^ yrepresents the output variable to be predicted, x i represents the respective variable of each regression coefficient bobtained, and is a term of random error. Table 1 Summary of the experimental design and results of HVOF coatings. Level Variables Low (1) High (+1) Fuel/oxygen ratio (F/O) 0.3 0.45 Powder flow [rpm]1220 Surface roughness Ra [ l m]18 28 Gun speed [mm=s]515 S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 2 2.2. Coating application conditions The deposition was performed using a JET4L-100 liquid fuel HVOF spray gun. The deposition conditions were established through the experimental design with the levels shown in Table 1. The F/O was obtained through the ratio: F=O¼Kerosene flow ½ml=min oxygen flow ½l=minð2Þ In the case of powder flow, a Metallisation Ò model 2007MF-PF powder feeder was used. The powder was fed volumetrically using a disc with perforated holes around the circumference. An inverter controls the revolutions per minute (rpm) of the disc and a mass flow meter controls the gas carrier flow that transports the powder to the chamber. Both devices are connected to a programmable logical controller. For powder parameters, the powder flow regulation variable used was rpm of the feed disc such that, for a single gas carrier flow, a greater feed disc speed would generate a greater powder flow. It should be noted that for all applications, the powder used came from the same commercial container to avoid variability in the system. Prior to the coating process, the surface was shot peened with a hardened grade 4 steel grit. Using a roughness tester PCE-RT 10, measurements were made at 0  ;120  , and 240  around the circumference of the cylindrical tube to ensure homogeneous values. Finally, the fourth variable is the longitudinal speed of the gun as it moves longitudinally through the lathe with the cylinder mounted. In the applications, all other variables remained constant at the characteristic values recommended by the machine manufacturer, such as, gas carrier flow and spray distance, maintaining an initial substrate temperature of 50 °C, and projecting 10 layers at an angle of 90°to generate thicknesses of around 100 l m. 2.3. Microstructural characterization of the coatings The powder was deposited over AISI 4140 steel cylinders with a diameter of 110 mm and 10 cm long. Samples of 1 cm 3 were obtained with a coated curved surface. This is due to the objective of manufacturing and remanufacturing curved surfaces found in shafts, rods, or pistons [53,54]. Due to the angle of impact of the particles of the deposited material, they behave differently when deposited on the substrate, affecting its final properties [55]. The coating cross-section surface was prepared for metallography analysis using a grinding and polishing sequence according to the ASTM E3-11 standard. The treated section was analyzed through the SEM using images of backscattered electrons (Circular Backscatter Detector, CBS) and an accelerating voltage of 20 kV to characterize the status of the coating and substrate and to obtain a better contrast between the phases present in the coating. Micro X-ray Diffraction ( l XRD) analysis was then performed on the coated section (surface and cross section) with a capillary of 1 mm, with the aim to study changes in coating composition after deposition. The l XRD patterns were obtained with a D8 DISCOVER diffractometer equipped with a 2D detector model VANTEC-500 (Bruker, Germany), using CuK a 1 radiation (k¼0:15406 nm) and a step size of 20°; the scan was recorded in the 2hrange comprised from 20°to 60°and a time per step of 10 min, with tube conditions of 40 kV and 40 mA. Thus, three frames were collected, each frame being about 20 min. Each frame allows seeing a range of about 30° 2h. After integrating the frames, diffractograms in the range 30°to 70°were obtained for analysis. 2.4. Characterization of thickness To ensure a representative measurement of the coating thickness, 3 images were taken of different sections of each sample obtained in each experiment, using SEM images at 500x to determine the coating end edges. In each image, 3 different measurements were obtained, and measurement were repeated 3 times to take the average and calculate the standard deviation. 2.5. Characterization of porosity To estimate the porosity, 3 SEM images at 1000x and 3 images at 2000x of each sample were used, taken from 3 different sections of the coating. The images of the sections were converted to binary format using the ImageJ Ò software [56], obtaining the percentage of totally black area in the coating section in relation to the total surface of the selected section, and this was defined as the porosity percentage. This process was repeated 3 times for each measurement, and all measurements of a sample were averaged to calculate the standard deviation. The defects were characterized qualitatively with the same images. 2.6. Characterization of surface hardness To evaluate the surface hardness on the Rockwell C (HRC) scale, a Rockwell TH320 (Times, China) Durometer was used with a load of 150 kg for 10 s, according to the ASTM E18-17 standard, taking the total average of 6 indentations by sample linearly along a plane tangential to the coated curved surface. Based on the results, the effect on interface adhesion, density, thickness, and cohesion over the substrate-coating ensemble was studied indirectly. 2.7. Characterization of microhardness Performing microhardness in the cross-section of the sample allows us to evaluate the intrinsic characteristics of the isolated coating, consisting of hard particles and the binder material. On the other hand, surface hardness evaluates coating-substrate assembly behavior indirectly by assessing adhesion, density, thickness, and cohesion [57]. When a spherical indenter is placed on the surface, the generated stress at the interface, if sufficient, can overcome adhesion resistance or cause damage due to low cohesion [58]. The microhardness in the cross-section of the coated samples was analyzed through a microindentation profile from the coating surface to the substrate, using a Vickers (MVK-HVL, Akashi) microdurometer with loads of 200 and 25 gf for F/O ratios of 0.45 and 0.3 respectively, with a 10 s maintenance time. This difference in load is because coatings with a F/O of 0.3 were not thick enough to perform indentations with a higher load. Three indentation profiles were prepared for each coating, obtaining the average measurements. Moreover, an indentation profile was prepared in the interface of the substrate coating with loadings of 50;100;200;300;500 and 1000 gf to study adhesion, and this process was repeated 3 times for each sample. The distributions of micro-hardness were plotted using individual readings measured for the coating with the same interval along three parallel lines spaced (i.e. 0:5 mm) from one another which, as such, they will represent the averaged distribution of the micro-hardness of the coating, measured from surface layer of the coating up to substrate material. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 3 3. Results and discussion 3.1. Characterization of the powder The powders used present a spherical morphology and are composed of a subset of agglomerate particles, as observed in the CBSSEM images (Fig. 1). The analysis performed using the compositional map (mapping) in Fig. 1d confirms the chemical composition of the powder and evidence the presence of Ni in the clear gray zone as an agglomerating phase in the irregular sections that join the C and Cr particles corresponding to chromium carbide, which are seen as dark gray zones. The XRD obtained for the CrC 20ðNi20CrÞpowder used for the coatings is shown in Fig. 2. The diffraction peaks of Cr 3 C 2 (space group symmetry Pnma (No. 62)) and Ni (space group symmetry Fm-3m (No. 225)) were the main phases identified. This is consistent with the compounds indicated by the manufacturer, and phases were found in a lower proportion such as in Cr 7 C 3 (space group of PMCM (No. 51)) and Cr 23 C 6 (space group symmetry Fm3m (No. 225)), according to literature [49,6]. The summary of the properties evaluated (porosity, thickness, surface hardness, and microhardness) corresponding to the application performed in the 2 4 experimental design (16 samples), can be found in Table 2. In the case of thickness and porosity, both measured through SEM, their values vary between 20 and 274 l m and 1.4 and 33%, respectively, while the surface hardness and microhardness ranges from 14 to 43 HRC and 495 to 930 HV, respectively. 3.2. Microstructural characterization of the coatings Based on the CBS-SEM images obtained from the cross-section of coatings in 16 samples (Fig. 3), it was determined that, in all Fig. 1. CBS-SEM images of Cr 3 C 2 20ðNi20CrÞpowder (a) 500, (b) 8000, (c) 30,000and (d) compositional mapping at 5000. Fig. 2. Cr 3 C 2 20ðNi20CrÞPowder X-ray diffraction pattern. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 4 cases, there is a homogeneous distribution of phases, where the light gray area is made up of Ni and the dark gray section of Crcarbides. It should be noted that samples 4, 5, 6, 9, 10, 14, 15, 16, corresponding to a F/O ratio = 0.45, presented a more homogeneous coating as shown in Fig. 3. The other samples with a F/O ratio = 0.3 have poor quality, evidenced by discontinuous and insufficient thickness, as well as particles breaking off from the coating as a sign of lack of cohesion, as seen in samples 1, 2, 7, 8, 11 and 12, as well as the presence of cracks in samples 3 and 13, which, in some cases, caused the layer to detach. Despite this, there is no evidence of cracks in the interface of any of the samples, which shows a good potential adhesion of the powder to the steel. It is observed that the substrate used has a high level of porosity in all specimens. In some samples, such as sample 10, it can be seen that the coating, despite achieving an adequate level of thickness, has interTable 2 Summary of the experimental design and results of HVOF coatings. Application F/O Powder flow [rpm] Substrate roughness Ra [ l m] Gun speed [mm=s] Coating thickness [ l m] Porosity [%] Surface hardness [HRC] Microhardness [HV 0:2 ] 1 0.3 12 18 15 29.48 ± 9.68 31.38 ± 4.08 26.1 ± 2.27 599 ± 35.23 2 0.3 12 28 15 23.20 ± 2.04 8.16 ± 2.20 24.8 ± 1.66 495 ± 23.69 3 0.3 20 28 5 21.54 ± 4.02 27.77 ± 6.61 29.4 ± 1.63 509 ± 21.59 4 0.45 20 18 15 99.28 ± 6.59 1.63 ± 0.49 30.6 ± 3.44 764 ± 34.51 5 0.45 12 28 5 132.76 ± 13.01 1.62 ± 0.33 33.7 ± 2.46 895 ± 27.86 6 0.45 12 28 15 63.40 ± 8.70 2.09 ± 0.49 27.8 ± 4.00 875 ± 17.11 7 0.3 12 18 5 25.95 ± 1.99 24.47 ± 3.43 25.4 ± 0.77 611 ± 61.61 8 0.3 12 28 5 28.89 ± 2.58 21.94 ± 0.90 22.6 ± 2.00 499 ± 13.93 9 0.45 20 28 15 38.80 ± 8.08 2.34 ± 0.37 26.2 ± 5.01 699 ± 59.56 10 0.45 20 28 5 126.92 ± 11.47 4.22 ± 0.61 41.5 ± 1.20 741 ± 50.40 11 0.3 20 18 5 20.01 ± 3.01 27.33 ± 8.57 26.5 ± 3.61 614 ± 14.39 12 0.3 20 28 15 35.35 ± 0.96 32.71 ± 4.37 31.8 ± 1.39 482 ± 28.00 13 0.3 20 18 15 48.12 ± 2.34 19.16 ± 4.20 27.2 ± 3.56 570 ± 63.07 14 0.45 20 18 5 273.77 ± 10.26 1.44 ± 0.12 42.8 ± 0.78 910 ± 34.38 15 0.45 12 18 5 147.80 ± 29.22 1.91 ± 0.20 32.9 ± 3.32 881 ± 23.70 16 0.45 12 18 15 63.18 ± 2.05 3.74 ± 0.33 14.0 ± 1.57 832 ± 83.22 Fig. 3. CBS-SEM images of coating cross-section at 2000. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 5 nal porosity greater than 2% due to a lack of cohesion among the particles, which could affect the mechanical performance of the substrate-coating compound. Through SEM images and the summary of results in Table 2, it is clear that the variations in the HVOF operation parameters have different effects on the coatings, making it necessary to statistically investigate the effects related to these parameters. The results obtained from the statistical analysis are discussed in the following subsections. 3.3. Influence of application parameters on thickness This work was carried out to obtain coatings of 100 l m thickness or more; however, achieving a greater thickness indicates that this is a more efficient process by having a greater number of powder particles deposited on the coating. The ANOVA study performed for the purpose of maximizing thickness is shown in Table 3 for the F/O ratio (A), powder flow (B), surface roughness (C), and gun speed (D) factors. The analysis reveals that the F/O ratio (A) and gun speed (D) are statistically significant factors because they have a P v alue <0:05. This indicates that the variation in these two factors has a greater influence on thickness, with factor A being the most significant with a lower P v alue . This can be better appreciated in the Pareto chart shown in Fig. 4a, where the red line represents the line of significance, showing that the factor with the greatest importance is the F/O ratio, followed by the combination of the AD factors. This interaction indicates that the A and D factors are not independent, as shown in the combined effects graph in Fig. 4b. When the F/O ratio is at its lower level, a greater thickness is achieved with a low gun speed. However, there is no difference in this value, as occurs when changing the F/O to the higher level, where it is seen that a lower gun speed achieves a more positive effect in the coating thickness. Modifying the F/O ratio helps to generate a combustion process that will deliver greater kinetic energy and raise the temperature of the sprayed powder [59], enabling the particles to reach a semi-melted state. This facilitates the plastic deformation upon impact with the substrate, obtaining greater adhesion and cohesion of the coating. A process that is richer in fuel (F/O = 0.45) allows for a greater thickness. In the case of speed, a lower speed means that the gun will spend more time in the same area of projection, depositing a larger amount of powder over the coating and, therefore, generating a greater thickness. Fig. 5 shows the response surfaces obtained from the prediction model with a R 2 adjusted ¼82:28%, where, based on Eq. (1), the following was obtained: ^ y¼73:65 þ44:6x 1 þ9:3x 2 14:8x 3 23:6x 4 13x 1 x 3 28:5x 1 x 4 12:5x 2 x 3 ð3Þ Fig. 5a exhibits the interaction of the F/O ratio with powder flow, where there is a clear upward trend in thickness with a powder flow of 20 rpm. It also evidence that this factor is completely independent from the F/O ratio, since there is no curvature on the surface, as opposed to surface roughness (Fig. 5b), where there is an interaction between A and C, and it shows a tendency to maximize thickness when there is a low roughness and high F/O ratio. This same interaction of factors is found between A and D (Fig. 5c) and between B and C (Fig. 5d). In this last case, there is a tendency to maximize thickness when the B factor is at the highest value and C at the lowest, coinciding with the previous graphs. Fig. 5e and f indicate that the gun speed is independent from powder flow and surface roughness, where a lower gun speed tends to maximize thickness. Table 3 ANOVA results for thickness. Factors SS DF MS F P v alue A 31806.00 1 31806.00 38.22 0.0003 B 1389.98 1 1389.98 1.67 0.2323 C 3502.57 1 3502.57 4.21 0.0743 D 8875.05 1 8875.05 10.66 0.0114 AC 2692.83 1 2692.83 3.24 0.1097 AD 13016.00 1 13016.00 15.64 0.0042 BC 2510.26 1 2510.26 3.02 0.1206 Error 6657.63 8 832.20 Total 70450.30 15 SS: sequential sum of squares, DF: degree of freedom, MS: mean square, F: statistical test. (a) (b) Fig. 4. (a) Pareto chart for thickness, and (b) Graph of the interaction between speed and F/O ratio for thickness. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 6 Fig. 5. Thickness response surface based on: (a) F/O and powder Flow, (b) F/O and surface roughness, (c) F/O and gun speed, (d) powder flow and surface roughness, (e) powder flow and gun speed, and (f) surface roughness and gun speed. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 7 3.4. Influence of application parameters on porosity The CBS-SEM images of the cross-section in Fig. 3 show a microporosity found in all coatings with the exception of samples 4, 10, 15 and 15. This is possibly caused by the segregation of carbides without an agglomerating matrix Ni to promote the formation of cavities between particles. Moreover, all coatings show a low level of particle cohesion in the area closest to the coating surface, due to the lack of deformation of the last particles projected during the spray process as a result of the absence of a hammering action of other particles that would have a later effect [60]. The result of the ANOVA study (Table 4) exhibits that the F/O ratio is the only independent factor with a significant effect (P v alue <0:05). It is also the case that some combined effects such as BC, ABC, BCD, and ABCD are statistically significant. As mentioned, the F/O ratio directly influences the energy given to the powder and, therefore, the degree of deformation of the particles, making a denser coating. In this instance, the roughness creates initial conditions for the first sprayed layer to achieve good adhesion, with stable support for the following layers. The quantity of mass deposited in the system depends on the gun speed and powder flow, producing, as a result, along with other variables, a significant effect. The Pareto chart in Fig. 6a exhibits that the main C and D variables are not significant on their own and that, despite the effect of combined parameters, the F/O ratio is the value with the most significant importance to obtaining a reduction in porosity, since the ratio of 0.45 gives the powder particle greater energy for these to adhere to the coating and get higher thickness values and lower porosity. This indicates that there is better cohesion and adhesion of the coating generated. The powder flow, despite having a P v alue >0:05, is more significant in the porosity than in the thickness and appears in all significant combined effects, therefore it is considered for the prediction model obtained according to Eq. (1): ^ y¼13:210:9x 1 þ1:3x 2 þ2:8x 2 x 3 2:1x 1 x 2 x 3 þ2:1x 2 x 3 x 4 2:2x 1 x 2 x 3 x 4 ð4Þ with a R 2 adjusted ¼94:72%. The BC effect is shown as the combination of parameters with a higher prediction coefficient (2.8). The graph in Fig. 6b shows that, with a high surface roughness value, the lower powder flow generates lower porosity. However, the surface roughness at the lower level results in a denser coating combined with a higher powder flow, which matches the plot in Fig. 5d, and indicates that this combination also produces thicker layers. Fig. 7a–c show the response surface by analyzing the F/O ratio in combination with other parameters. These results confirm that this factor, at the upper value, produces a reduction in porosity in all cases and has a much greater effect than the others. Fig. 7d evidence the above mentioned combined effect, where there is a green area on the response surface that is larger for powder flow (negative) and substrate roughness (positive), but also, a green area for powder flow (positive) with a surface roughness at its lower value. Then, Fig. 7e and f show nearly no slope on the surfaces, indicating that these parameters in combination have no relevant effect on porosity. Table 4 ANOVA results for porosity. Factors SS DF MS F P v alue A 1890.73 1 1890.73 230.37 0.0000 B 28.33 1 28.33 3.45 0.1055 C 6.52 1 6.52 0.79 0.4025 D 5.63 1 5.63 0.69 0.4349 BC 127.52 1 127.52 15.54 0.0056 ABC 73.57 1 73.57 8.96 0.0201 BCD 68.43 1 68.43 8.34 0.0234 ABCD 74.43 1 74.43 9.07 0.0196 Error 57.45 7 8.21 Total 2332.61 15 SS: sequential sum of squares, DF: degree of freedom, MS: mean square, F: statistical test. (a) (b) Fig. 6. (a) Pareto chart for porosity, and (b) Graph of the interaction between surface roughness and powder flow for porosity. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 8 Fig. 7. Porosity response surface based on: (a) F/O and powder Flow, (b) F/O and surface roughness, (c) F/O and gun speed, (d) powder flow and surface roughness, (e) powder flow and gun speed, and (f) surface roughness and gun speed. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 9 fuel (HVOF) sprayed Cr3C2-NiCr coating, Surf. Coat. Technol. 225 (2013) 85–91 (jun 2013). doi:10.1016/j.surfcoat.2013.03.020. URL: https://linkinghub. elsevier.com/retrieve/pii/S0257897213002727. [49] D. Poirier, J.-G. Legoux, R.S. Lima, Engineering HVOF-Sprayed Cr3C2-NiCr Coatings: The Effect of Particle Morphology and Spraying Parameters on the Microstructure, Properties, and High Temperature Wear Performance, J. Therm. Spray Technol. 22 (2–3) (2013) 280–289 (mar 2013). doi:10.1007/ s11666-012-9833-3. [50] C. Li, X. Gao, D. Zhang, H. Gao, X. Han, B. Zhang, Numerical Investigation on the Flame Characteristics and Particle Behaviors in a HVOF Spray Process Using Kerosene as Fuel, J. Therm. Spray Technol. 30 (3) (2021) 725–738 (feb 2021). doi:10.1007/s11666-021-01165-3. [51] A.G.M. Pukasiewicz, H.E. de Boer, G.B. Sucharski, R.F. Vaz, L.A.J. Procopiak, The influence of HVOF spraying parameters on the microstructure, residual stress and cavitation resistance of FeMnCrSi coatings, Surf. Coat. Technol. 327 (2017) 158–166 (oct 2017). doi:10.1016/j.surfcoat.2017.07.073. URL: https:// linkinghub.elsevier.com/retrieve/pii/S0257897217307752. [52] M. Abbas, G.M. Smith, P.R. Munroe, Microstructural study of HVOF sprayed Ni particles on a grit-blasted stainless-steel substrate, Surf. Coat. Technol. 409 (2021) 126832 (mar 2021). doi:10.1016/j.surfcoat.2021.126832. URL: https:// linkinghub.elsevier.com/retrieve/pii/S025789722100005. [53] M.A. Javed, A.S. Ang, C.M. Bhadra, R. Piola, W.C. Neil, C.C. Berndt, M. Leigh, H. Howse, S.A. Wade, Corrosion and mechanical performance of hvof wc-based coatings with alloyed nickel binder for use in marine hydraulic applications, Surf. Coat. Technol. 418 (2021) 127239 (7 2021). doi:10.1016/J. SURFCOAT.2021.127239. [54] T. Sahraoui, N.E. Fenineche, G. Montavon, C. Coddet, Alternative to chromium: characteristics and wear behavior of hvof coatings for gas turbine shafts repair (heavy-duty), J. Mater. Process. Technol. 152 (2004) 43–55 (10 2004). doi:10.1016/J.JMATPROTEC.2004.02.061. [55] H. Wu, X. Xie, M. Liu, C. Chen, H. Liao, Y. Zhang, S. Deng, A new approach to simulate coating thickness in cold spray, Surf. Coat. Technol. 382 (2020), https://doi.org/10.1016/j.surfcoat.2019.125151. [56] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis, Nature Methods 9 (7) (2012) 671–675 (jul 2012). doi:10.1038/ nmeth.2089. URL: https://www.nature.com/articles/nmeth.2089. [57] Šárka Houdková, O. Bláhová, F. Zahálka, M. Kašparová, The instrumented indentation study of hvof-sprayed hardmetal coatings, J. Therm. Spray Technol. 21 (2012) 77–85 (2012). doi:10.1007/s11666-011-9677-2. [58] X. Zhang, X.-B. Tian, Z.-W. Zhao, J.-B. Gao, Y.-W. Zhou, P. Gao, Y.-Y. Guo, Z. Lv, Evaluation of the adhesion and failure mechanism of the hard crn coatings on different substrates, Surf. Coat. Technol. 364 (2019) 135–143 (4 2019). doi:10.1016/j.surfcoat.2019.01.059. URL: https://linkinghub.elsevier.com/ retrieve/pii/S0257897219300751. [59] L. Qiao, Y. Wu, S. Hong, J. Zhang, W. Shi, Y. Zheng, Relationships between spray parameters, microstructures and ultrasonic cavitation erosion behavior of HVOF sprayed Fe-based amorphous/nanocrystalline coatings, Ultrason. Sonochem. 39 (2017) 39–46 (nov 2017). doi:10.1016/j. ultsonch.2017.04.011. URL: https://linkinghub.elsevier.com/retrieve/pii/ S1350417717301694. [60] B. Song, M. Bai, K.T. Voisey, T. Hussain, Role of Oxides and Porosity on HighTemperature Oxidation of Liquid-Fueled HVOF Thermal-Sprayed Ni50Cr Coatings, J. Therm. Spray Technol. 26 (3) (2017) 554–568 (feb 2017). doi:10.1007/s11666-017-0531-z. [61] W. Tillmann, L. Hagen, C. Schaak, J. Liß, M. Schaper, K.P. Hoyer, M.E. Aydinöz, K. U. Garthe, Adhesion of HVOF-Sprayed WC-Co Coatings on 316L Substrates Processed by SLM, J. Therm. Spray Technol. 29 (6) (2020) 1396–1409 (2020). doi:10.1007/s11666-020-01081-y. [62] R.S. Paredes, S.C. Amico, A.S. d’Oliveira, The effect of roughness and preheating of the substrate on the morphology of aluminium coatings deposited by thermal spraying, Surf. Coat. Technol. 200 (2006) 3049–3055 (2 2006). doi:10.1016/J.SURFCOAT.2005.02.200. [63] A.W.Y. Tan, W. Sun, A. Bhowmik, J.Y. Lek, X. Song, W. Zhai, H. Zheng, F. Li, I. Marinescu, Z. Dong, E. Liu, Effect of substrate surface roughness on microstructure and mechanical properties of cold-sprayed ti6al4v coatings on ti6al4v substrates, J. Therm. Spray Technol. 28 (2019) 1959–1973 (12 2019). doi:10.1007/S11666-019-00926-5/FIGURES/14. [64] R.P. Garcia, T. Luz, S.C. Canobre, al, Y. Cheng, S. Choi, T. Watanabe, B. Swain, A. Behera, Effect of powder feed rate on adhesion strength and microhardness of aps niti coating: a microstructural investigation, Surf. Topogr.: Metrol. Properties 9 (2021) 025039 (6 2021). doi:10.1088/2051672X/AC0A38. S. Sauceda, S. Lascano, J. Núñez et al. Engineering Science and Technology, an International Journal 39 (2023) 101342 16