Water-droplet erosion behavior of high-velocity oxygen-fuel-sprayed coatings for steam turbine blades

2.1 Sample preparation

Three types of specimens were prepared for use in this study: an uncoated steel substrate (1Cr12Ni2W1Mo1V), a steel substrate coated with Cr₃C₂–25NiCr (5–30 μm, Amperit^® 588, H.C. Starck), and a steel substrate coated with WC–10Co–4Cr (15–45 μm, Amperit^® 558, H.C. Starck). The coatings were applied by thermal spraying using a HVOF system (JP8000, Praxair), and the spraying parameters employed are listed in Table 1. Sandblasting with Al₂O₃ powder (size ∼46 mesh) was performed prior to thermal spraying. The blasting distance and pressure were 100–150 mm and 0.4–0.5 MPa, respectively. Each sample type had six samples with dimensions of 15 × 8 × 8 mm. During the spraying process, the sample was fixed, and the surface of the sample was perpendicular to the axis of the spray gun. The thicknesses of the Cr₃C₂–25NiCr and WC–10Co–4Cr as-sprayed coatings were 210 ± 5 and 210 ± 9 μm, respectively. Dense and uniform microstructures were observed in the coatings with approximate porosity values of <0.5%. Considering that these samples are directly used in the actual application after preparation, the samples were not polished prior to the water-droplet erosion tests.

2.2 Water-droplet erosion tests

A high-speed pulsed water jet device fabricated in-house was used for the water-droplet erosion tests, as shown in Figure 2a. The device comprised an electrical motor, high/low-pressure water pump, vacuum pump, water tank, and rotating test rig. The output end of the electrical motor was connected to a gear speed-increasing box. The shaft at the output end of the gearbox was coupled to the rotating shaft, and rotated the test pieces at high speed. A high-pressure water pump was used to generate instantaneous pressure, which in turn produced a high-speed water flow. The position of the nozzle of the high-pressure water pump corresponded to the position of the sample. The shape and position of the water flow were taken at a speed of 104 frames per second using a high-speed camera. According to the calibration of the high-speed camera, the experimental system can produce jet speeds of up to 600 m s⁻¹. In addition, the jet outlet pressure reached 300 MPa, and the maximum pressure for continuous and stable work was 240–260 MPa, which can fully simulate the impact velocity of real wet-steam water droplets. The pressure and velocity of the jet flow were precisely controlled by adjusting the high-pressure water pump and regulating valve in front of the nozzle. Using this equipment, the liquid impact velocity was not only greatly improved, but also precisely controlled to simulate water erosion of the blade material at different turbine speeds. Furthermore, by varying the nozzle diameter, the diameter of the jet outlet could be adjusted to simulate erosion of the material by water droplets of different diameters. The water in the water tank can also be mixed with a corrosive medium to simulate the real steam turbine blade operating environment.

Figure 2:

Experimental setup for the water-droplet erosion tests: (a) high-speed pulsed water jet device; and (b) specimens fixed on a rotating rig with different impingement angles.

In this work, three types of samples were eroded using the described device. To obtain different impingement angles, the specimens were machined with angles of 30°, 60°, and 90° with respect to the water-droplet incidence direction, as shown in Figure 2b. The specimens were symmetrically fixed on a rotating rig to avoid mass eccentricity. The test liquid employed was tap water at a temperature of ∼20 °C, the jet velocity was 692 m s⁻¹, the pulse frequency was 50 Hz, the nozzle diameter (i.e., the droplet size) was 0.2 mm, and the specimen rotation speed was 1500 rpm. During the experiment, the mass was measured six times, and the mass loss curve was recorded. The relationship between the erosion time and material loss was obtained to provide a basis for evaluating the water-erosion resistance behavior of the turbine blade.

2.3 Microstructural characterization and measurement of the mechanical properties

Before and after the water-erosion tests, the coatings were characterized by polishing the samples with diamond abrasives and imaging by an optical microscope (Ntegra Spectra SNOM), and the changes in the elemental distribution were investigated by an energy dispersive X-ray spectroscopy (EDS, MICs F device). To evaluate the thicknesses and porosities of the as-sprayed coatings, cross-sectional scanning electron microscopy (SEM, DMI5000M device) images were evaluated by image analysis. The 2D and 3D surface morphologies of the samples after the water-erosion tests were examined using metalloscopy (DMI5000M, Leica) and laser microscopy (VK-9700, Keyence), respectively. Based on high resolution (1024 × 768) 3D microscopic analysis, the water-erosion depth could be calculated. The cumulative mass loss induced by water-erosion was measured using an analytical balance with an accuracy of 0.1 mg.

The coating hardness was determined by micro-indentation, with the consideration of negligible damage to the specimen. The Vickers hardness of the coating can be calculated according to the following equation (Tuck et al. 2000):

where HV is the Vickers hardness in gf  mm⁻², P is the load in gf, and d is the diagonal length in mm. The testing load was 300 g and the loading time was 10 s. The average value of hardness was obtained by averaging the results of five randomly selected points on the cross-section of the coated specimen.

The fracture toughness (K_Ic) of each coating was measured by micro-indentation to quantitatively characterize the resistance to crack propagation. Upon loading, as imposed by the Vickers indenter on the polished cross-section of the sample, a crack is triggered, and its length can be measured to determine the fracture toughness, according to the following equation (Anstis et al. 1981):

where P, c, E, and H are the load, half-length of the crack, Young’s modulus, and micro-hardness, respectively. These two mechanical parameters can be obtained from micro-indentation tests, as indicated above. In the fracture toughness test, a constant load of 50 N was applied.

Adhesion is one of the most important mechanical properties of a coating, and is directly related to the reliability of the repairing and strengthening parts. Moreover, it serves as an important basis for optimization of the process parameters. Adhesion tests were therefore carried out to evaluate the tensile bonding strength between the coating and substrate. A standard adhesion test, ASTMC 633-79, was used to quantitatively evaluate the bonding strength. The samples were bonded to the loading jig with the use of E-7, and heat-treated in a drying oven at 110 °C for 3 h prior to carrying out the adhesion tests. A perpendicular load was applied on the coating at the rate of 1 mm min⁻¹ using a universal testing machine. The bonding strength was determined as follows:

where δ_b is the tensile bonding strength in MPa, F is the maximum load at delamination in N, and S is the projected area of the coating in mm².

3.1 Microstructures and elemental distributions of the as-sprayed coatings

Typical optical microscope images of the polished HVOF-sprayed Cr₃C₂–25NiCr and WC–10Co–4Cr coatings with dense microstructures are shown in Figure 3. It should be noted that the HVOF coatings had a uniform thickness and a dense and homogenous microstructure, as well as compact bonding to the substrate. Based on the optical microscope images and ImageJ software (National Institutes of Health, America), the thickness and total porosity were obtained. The thickness and porosity of the Cr₃C₂–25NiCr coating were found to be 205 μm and 0.19%, while those of the WC–10Co–4Cr coating were 213 μm and 0.07%.

Figure 3:

Optical microscope images of the morphologies of (a) Cr₃C₂–25NiCr coating and (b) WC–10Co–4Cr coating.

To acquire the detailed elemental distributions of the as-sprayed coatings, the chemical compositions of the Cr₃C₂–25NiCr and WC–10Co–4Cr coatings were determined using EDS mapping. C, Cr, and Ni contents are the main components of the Cr₃C₂–25NiCr coating. Ni and part of the Cr content formed the NiCr bonding phase, which provides good toughness and strength. Meanwhile, C and part of the Cr content formed the Cr₃C₂ ceramic phase, which provides high hardness. The Cr₃C₂ ceramic phase was uniformly dispersed in the NiCr binding phase, imparting the Cr₃C₂–25NiCr coating with excellent hardness, good toughness, and good binding strength. In addition, Ni and Cr were uniformly distributed, while C was flacked in the interior of the coating, indicating that the ceramic phase particles of Cr₃C₂ were relatively large.

C, Cr, W, and Co contents are the main components of the WC–10Co–4Cr coating. W and C formed the WC ceramic phase, which provides high hardness, while Co and Cr formed the CoCr bonding phase, which provides good toughness and strength. The WC ceramic phase was uniformly dispersed in the CoCr binding phase, thereby imparting WC–10Co–4Cr with high hardness, good toughness, and good binding strength. The colored spots corresponding to W are small and present a uniform distribution in the coating interior of WC–10Co–4Cr, indicating that the WC ceramic phase was fine and evenly distributed.

3.2 Water-droplet erosion behavior

3.2.1 Uncoated steel specimens

Figure 4 shows the morphology of the 90° uncoated steel specimen after 7 h water-droplet erosion. Obvious water-erosion surface cavities were formed on the surface of the steel, and some discontinuous surface cavities were also observed in comparison to the undamaged positions on both sides. Figure 5 shows the 3D topographic evolution of the erosion areas of the 30°, 60°, and 90° uncoated steel specimens after 3 and 42 h of erosion, respectively. As indicated in Figure 5a1, after 3 h, some discrete droplet erosion surface cavities appeared on the surface of the 30° specimen, which adopted irregular rather than circular shapes due to the surface roughness. At the same time, some valleys could be observed on the profiles, with undulations measuring ∼25 μm in depth. As the water-droplet erosion process continued, these surface cavities grew at an approximately constant rate until developing into a continuous droplet erosion groove. The groove became deeper after further exposure, reaching a depth of ∼90 μm at 42 h, while the width increased due to the relatively small impingement angle (Figure 5a2).

Figure 4:

Morphology of the 90° uncoated steel specimen after 7 h water-droplet erosion.

Figure 5:

3D topography images of the erosion areas of the uncoated steel specimens with different impingement angles and after different exposure times: (a1) 30° and 3 h; (a2) 30° and 42 h; (b1) 60° and 3 h; (b2) 60° and 42 h; (c1) 90° and 3 h; (c2) 90° and 42 h.

3D topography plots were also obtained for the erosion areas of the 60° and 90° steel specimens at different erosion times (Figure 5b and c). A similar erosion behavior was observed, and could be summarized into three stages: (1) the formation of discrete erosion surface cavities on the surface; (2) the development of continuous erosion grooves from the discrete surface cavities; and (3) the broadening and deepening of the groove under the effect of water-droplet erosion. In addition, for the same exposure time, the pit width increased as the impingement angle increased.

3.2.2 Steel specimens bearing the Cr₃C₂–25NiCr coating

The Cr₃C₂–25NiCr coating presented a typical morphology of HVOF coatings, as indicated in Figure 6a and b, whereby large particles were observed. The surface morphology of the 90° coating after 42 h water-droplet erosion test is shown in Figure 6c and d, where surface damage can clearly be observed. More specifically, the coating exhibited particularly shallow water-erosion traces, and the water-erosion area was relatively smooth. To obtain a clearer understanding of the damaged surface morphology, the 3D topographic evolutions of the erosion areas of the steel specimens bearing the Cr₃C₂–25NiCr coating were obtained after 42 h of erosion for all three impingement angles, as shown in Figure 7. In this case, the 3D morphology of water-droplet erosion after 3 h was not recorded due to the rough surface, which resulted in large scattering. As illustrated in Figure 7a1, after 37 h, water-erosion surface cavities were observed on the 30° specimen, along with a water erosion groove. Comparison of the 37 and 42 h specimens indicates that the groove continued to deepen from 44.25 to 141.9 μm, and the maximum width of the groove was 440 μm. In addition, the water-erosion area was relatively smooth, which indicates that the failure mode of the coating was relatively uniform. 3D topography plots for the erosion areas of the 60° and 90° specimens at different erosion times are shown in Figure 7b and c, respectively. Although the depth of the groove continued to grow to 42 h, it grew less than in the case of the 30° specimen, which indicated that the powder coating was more brittle than the metal.

Figure 6:

SEM images of Cr₃C₂–25NiCr coating: (a) 3000× magnification before water-erosion tests, (b) 16,000× magnification before water-erosion tests, (c) 3000× magnification after 90° and 42 h water-droplet erosion tests, and (d) 16,000× magnification after 90° and 42 h water-droplet erosion tests.

Figure 7:

3D topography images of erosion areas of Cr₃C₂–25NiCr coated steel specimens with different impingement angles and after different exposure times: (a1) 30° and 37 h; (a2) 30° and 42 h; (b1) 60° and 37 h; (b2) 60° and 42 h; (c1) 90° and 37 h; (c2) 90° and 42 h.

After the same exposure time, the groove width of the 60° specimen was the largest, followed by the 90° specimen, with the 30° specimen being the narrowest. For all specimens, the width of the groove was significantly larger than the depth, which indicates that the lateral jet produced by the high-speed impact was responsible for groove formation, while the shear force produced by the high-speed liquid–solid impact caused more damage than the vertical impact.

3.2.3 Steel specimens bearing the WC–10Co–4Cr coating

The WC–10Co–4Cr coating presented a typical microstructure of HVOF coatings, as indicated in Figure 8a and b, whereby relatively fine particles could be seen. The surface morphology of the 90° coating after 42 h water-droplet erosion test is shown in Figure 8c and d, with obvious water-erosion damage being observed. In addition, the surface color of the specimen at the water erosion site was lighter, which indicates that the depth of water-erosion increased gradually. However, since the coating particles were small, the damage caused by water erosion was also considered to be relatively small, and the effect of water-erosion on the coating was not obvious. Figure 9 shows the 3D topographic evolutions of the erosion areas of the steel specimens bearing the WC–10Co–4Cr coating after 3 and 42 h of erosion for all three impingement angles. Since this coating was good at resisting water-erosion, no obvious damage was observed in the 3D images. For the 30° specimen (Figure 9a), the erosion effect on the specimen surface was relatively small; therefore, the difference in depth between the water erosion area and the surrounding area could not be clearly distinguished, even after 42 h. For the 60° specimen (Figure 9b), the erosion depth was small compared to the surface roughness after 3 h; therefore, it was difficult to measure. After 42 h, it was clear that the water erosion area was not a continuous erosion groove, but a series of discrete discontinuous erosion surface cavities. A similar erosion behavior was observed for the 90° specimen (Figure 9c).

Figure 8:

SEM images of WC–10Co–4Cr coatings: (a) 3000× magnification before water-erosion tests, (b) 16,000× magnification before water-erosion tests, (c) 3000× magnification after 90° and 42 h water-droplet erosion tests, and (d) 16,000× magnification after 90° and 42 h water-droplet erosion tests.

Figure 9:

3D topography images of the erosion areas of the WC–10Co–4Cr coated steel specimens with different impingement angles and after different exposure times: (a1) 30° and 3 h; (a2) 30° and 42 h; (b1) 60° and 3 h; (b2) 60° and 42 h; (c1) 90° and 3 h; (c2) 90° and 42 h.

3.3 Cumulative mass loss curves during water-droplet erosion

In addition to the 3D topographies, the cumulative mass losses are another important indicator of water-droplet erosion. During our experiments, the mass of each specimen was measured six times, and the mass loss curves were plotted, as shown in Tables 2–4 and Figure 10. Table 2 and Figure 10a show the cumulative mass loss of the uncoated steel as a function of the water-droplet erosion time, where the 30° specimen exhibited the smallest mass loss per unit area, while the 90° specimen had the largest. Moreover, the mass loss increased significantly more between 30° and 60° than between 60° and 90°, which indicates that the impingement angle had a nonlinear effect on the cumulative mass loss. In all three cases, the rate of mass loss increased as the erosion time increased, resulting in reduced erosion resistance for the uncoated steel in the latter stages of erosion.

Figure 10:

Cumulative mass loss curves: (a) uncoated steel specimens; (b) Cr₃C₂–25NiCr coated specimens; and (c) WC–10Co–4Cr coated specimens.

Compared to the uncoated specimens, reductions of >75% mass loss were recorded for the Cr₃C₂–25NiCr-coated specimens, as shown in Table 3 and Figure 10b, indicating that improved water-droplet erosion resistance was achieved by the application of this coating. In addition, the cumulative mass loss was proportional to the erosion time, which signified that little degradation in the erosion resistance occurred during the test. Interestingly, the mass loss was the largest in the case of the 60° impingement angle. However, further studies into the mechanical properties of the Cr₃C₂–25NiCr coating may be necessary to account for this phenomenon.

As shown in Table 4 and Figure 10c, the WC–10Co–4Cr coated specimens exhibited similar water-droplet erosion behavior to the Cr₃C₂–25NiCr coated specimens. The low mass losses recorded indicated that the WC–10Co–4Cr coated specimens exhibited excellent corrosion resistance. This was attributed to the densification of the coating surface under the effect of polishing by the high-speed water jet, which resulted in improved erosion resistance for the WC–10Co–4Cr coated specimens compared to the Cr₃C₂–25NiCr coated specimens.

3.4 Effects of the materials properties on the erosion resistance

To better understand the water-droplet erosion resistance mechanism of the coatings, a series of performance tests were conducted on the uncoated steel substrate, the steel substrate coated with Cr₃C₂–25NiCr, and the substrate coated with WC–10Co–4Cr.