The importance of phase composition for corrosion resistance and thermal oxidation behavior of gas-borided layer produced on nickel alloy

2.1 Materials

The substrate material used in this study was Inconel^®600-alloy. This Ni–Cr–Fe alloy contained chromium (15.72 wt. %) and iron (8.63 wt. %) as the main alloying elements. The detailed chemical composition of the Inconel^®600-alloy was presented in Table 1. The shape of the samples was adapted to the corrosion test conditions, therefore, flat discs were used for the tests. The dimensions of the samples were as follows: diameter of 20 mm, height of 6 mm.

2.2 Gas boriding in N₂–H₂–BCl₃ atmosphere

Gas boriding in N₂–H₂–BCl₃ atmosphere was selected as the method for production of layer on the Inconel^®600-alloy because of its high efficiency. This method results in the formation of thick layers at a lower temperature and shorter process duration in comparison to the pack boriding method. In the present study, gas boriding was carried out as a continuous process. The following process parameters were used: temperature of 920 °C (1193 K), duration of 2 h, average concentration of boron trichloride of 5.23 vol. % (in relation to hydrogen). The devices used for gas boriding, and a detailed description of the process, were presented in the previous work (Makuch et al. 2015).

2.3 Surface roughness

A white light 3D interference microscope (WYKO NT−1100) was used to evaluate the 3D surface profiles of non-borided and borided samples. During the measurements, the interferometric objective moved vertically to scan the surface at changeable heights. The following parameters were determined: R_a, R_q, S_a, and S_q. Amplitude parameters are the most important parameters characterizing the surface topography in respect of the surface deviations. The arithmetic average height R_a is defined as the average absolute deviation of the roughness irregularities from the mean line. The extension of the R_a parameter to the surface is S_a. The root mean square roughness R_q represents the standard deviation of the distribution of surface heights in the linear profile whereas S_q represents the root mean square value in relation to the surface.

The 3D surface topography and roughness parameters of the non-borided and borided samples were compared in Figure 1 and Table 2. The surface topography of non-borided (Figure 1a) and borided (Figure 1b) was uniform, without craters or cracks. Only the effects of mechanical machining were visible. The non-borided sample was characterized by a more uniform profile with a lower roughness (R_a = 1.38 µm, S_a = 1.96 µm) compared to the borided sample (R_a = 1.45 µm, S_a = 2.46 µm).

Figure 1:

3D images of surface topography of non-borided (a) and borided (b) samples.

2.4 Microstructure characterization

First, the specimens for microstructure characterization were cut out. Standard metallographic preparation included the following steps: mounting in conductive resin, grinding, polishing, etching with the use of a Marble’s reagent.

The microstructure of the boride layer was observed with the use of a Keyence VHX-6000 digital microscope. The porosity of the boride layer was determined on the cross-section of the layer with the use of a digital microscope Keyence VHX-6000. A Mira3 Tescan scanning electron microscope (SEM) equipped with an Energy Dispersive Spectrometer was used for X-ray microanalysis.

2.5 Cohesion evaluation

The bond strength between the gas-borided layer and the substrate material was determined based on the destructive method in accordance with the VDI 3198 norm (Verein Deutscher Ingenieure Normen 1991). The measurements consisted of the generation of a Rockwell indentation mark under a total load of 1470 N (scale C). During the measurements, the cone-shaped indenter caused the destruction of the layer in the form of microcracks, delamination, or flacking (Figure 2). After the Rockwell C test, the generated indentation marks with visible failures were observed with the use of an optical microscope. It is a comparative method in which the indentation craters and generated failure marks were compared with those presented in the norm as quality maps HF1–HF6 (Figure 2).

Figure 2:

The principle of the VDI 3198 cohesion test with HF1-HF6 quality maps.

2.6 Corrosion resistance evaluation

To determine the corrosion resistance of the samples, a potentiodynamic anodic polarization test was used. The investigations were carried out with the use of the Atlas Solich ATLAS 0531 EU&IA device. A three-electrode cell system containing a platinum electrode as a counter electrode, a saturated calomel electrode as a reference electrode, and a sample as a working electrode, was used for the experiments. Measurements were performed on three test areas for each sample. The tested sample (the exposed surface was 50 mm²) was immersed in a 3.5% NaCl solution at a constant temperature of 22 °C. The specimens were polarized in the anode direction from a potential −1.5 V up to 1.5 V, with a potential change rate of 0.5 mV/s. The results were plotted as a polarization curve showing the E − log I dependence. From the obtained polarization curve, the following quantities were determined: corrosion potential E_CORR, corrosion current density I_CORR, primary passivation potential E_P, Flade potential E_F, transpassive potential E_TP, secondary passivity potential E_SP, oxygen evolution potential E_OE, and passivation current density.

2.7 Oxidation test

Oxidation tests were conducted in a resistance furnace at a temperature of 1000 °C (1273 K) for a period of 24 h. First, the cleaned sample was put into the furnace. Then the heating process was started. A heating rate of 10 °C/min was applied to minimize the oxidation during heating. After the furnace reached a temperature of 1000 °C (1273 K), the oxidation test in the open-air atmosphere was started. After 24 h of oxidation, the gas-borided sample was slowly cooled in a furnace to minimize the risk of thermal shock. The oxidized sample was examined by scanning electron microscopy (Mira3 Tescan Vega). The reaction products formed on the surface during oxidation were identified by X-ray diffraction (XRD) and energy dispersive spectroscopy (EDS).

3.1 Microstructure analysis

Gas boriding resulted in the formation of a compact boride layer which was characterized by a high average thickness of 86 μm. The XRD pattern of gas-borided layers formed on the Inconel^®600-alloy was analyzed in the previous study (Makuch et al. 2015). The compact boride zone included nickel borides (NiB, Ni₂B, Ni₃B, and Ni₄B₃) and chromium borides (CrB and Cr₂B). This type of multiphase microstructure did not show zonal phase distribution. It was found that nickel borides, and chromium borides, appeared in the cross-section of the layer only as a mixture. Moreover, the distribution of nickel borides and chromium borides was independent of the distance from the surface.

The high amount of porosity could be the reason for diminished mechanical properties and high brittleness of the borided layer. Therefore, the percentage of pores in the borided layer produced on the Inconel^®600-alloy was measured. The OM microstructure (Figure 3a) and the binary image of a layer with marked porosity (Figure 3b) were analyzed. The investigated area was characterized by a small amount of porosity equal to 2.912%. However, the pores were mainly detected near the top-surface region. This specific pore distribution could be the reason for the reduced cohesion of the layer with the substrate material, and the increased brittleness of the layer. In the case of the gas-borided Inconel^®600-alloy, the presence of porosity was caused by the high concentration of iron in the base material (8.63 wt. %).

Figure 3:

The OM microstructure (a) and binary image (b) of porosity of the gas-boride layer produced on Inconel^®600-alloy.

The presence of iron in the base material enabled the formation of ferrous and ferric chlorides (FeCl₂ and FeCl₃) during gas boriding in N₂–H₂–BCl₃ atmosphere. As a consequence of the reaction of boron trichloride with iron and reducer (in this process – hydrogen), the following products were formed: HCl, BHCl₂, FeCl₂, and FeCl₃. During the process, hydrochloric acid reacted with iron and caused the formation of ferrous and ferric chlorides as a consequence of the reactions (Hegewaldt et al. 1984):

According to the paper (Hegewaldt et al. 1984), the porosity appeared as a result of corrosion caused by the presence of FeCl₂.

3.2 Cohesion

Figure 4 shows an OM image of a Rockwell-C indentation crater with visible failure marks. The following failure marks were visible: long radial cracks (Figure 4d), flaking, and delamination (Figure 4b and c). These observed failure marks can be classified as acceptable failures, which indicate sufficient cohesion represented by the HF3 quality map (Figure 4e). The high hardness gradient between the borided layer and the base material could be the cause of delamination. The second important factor that strongly influenced the results of the cohesion test was the presence of porosity below the top surface and the brittleness of the boride layer.

Figure 4:

OM images of gas-borided layers produced on Inconel^®600-alloy after Rockwell C indentation test (a–d) and scheme of HF3 standard with failure marks (e).

3.3 Corrosion resistance

The corrosion resistance was determined based on the obtained polarization curves. Three curves were plotted for each borided and non-borided sample. Owing to the high similarity of the obtained curves, only one representative polarization curve was presented and described. Simultaneously, the electrochemical parameters derived from the polarization curves were presented in Table 3. The values of E_CORR and I_CORR evaluated from three measurements for each sample were shown. A representative corrosion potentiodynamic curve recorded in a 3.5% NaCl solution for gas-borided Inconel^®600-alloy is shown in Figure 5a. To determine the influence of boriding on the corrosion behavior, the results obtained for the non-borided Inconel^®600-alloy were presented (Figure 5b). The width of the corrosion resistance region (1) was higher for the gas-borided sample as compared to the non-borided sample. For this reason, the value of the corrosion potential E_CORR was also higher for the borided sample. Both samples show resistant-active–passive-transpassive behavior in the corrosive environment used. However, the passive region differed when the borided and non-borided samples were compared. First, in the case of the borided sample, the passivation current density was 2·10⁻⁵ A/cm². Whereas, in the case of the non-borided Inconel^®600-alloy, the passivation current density was 5·10⁻⁵ A/cm².

Figure 5:

Corrosion potentiodynamic curves recorded in a 3.5% NaCl solution for gas-borided Inconel^®600-alloy (a) and non-borided Inconel^®600-alloy (b); E_CORR-corrosion potential; I_CORR-corrosion current density; E_P-the primary passivation potential; E_F-the Flade potential; E_TP-the transpassive potential; E_SP-the secondary passive potential; E_OE-oxygen evolution potential.

1, corrosion resistance region; 2, active region; 3, primary passivation region; 4, passive region; 5, transpassive region; 6, secondary passivity region; 7, oxygen evolution region.

These results indicated that the presence of a borided layer on the surface of Inconel^®600-alloy resulted in reduced corrosion. Comparing the range of the passive region, the non-borided Inconel^®600-alloy was found to be passive over a wide potential range (from −0.51 to 0.77 V) before transpassive dissolution occurred at 0.77 V. The borided sample was characterized by the narrower range of the passive state, ranging from −0.52 to −0.22 V. These results indicated a higher susceptibility to passivation in a 3.5% NaCl solution. Moreover, the non-borided sample spontaneously passivated without exhibiting the distinct active to passive transition peak.

The electrochemical parameters I_CORR and E_CORR determined based on the polarization curves are presented in Table 3. The higher value of the corrosion potential (average E_CORR = −0.952 V) characterized the gas-borided Inconel^®600-alloy. The non-borided sample demonstrates a slightly lower average E_CORR value of −1.006 V. The lower corrosion current density I_CORR (average 2.4·10⁻⁶ A/cm²) was measured for the gas-borided Inconel^®600-alloy. The non-borided sample was characterized by a higher I_CORR value (2.1·10⁻⁵ A/cm²). Therefore, it can be concluded that the gas-borided sample shows a higher corrosion resistance in a 3.5% NaCl solution.

However, the multicomponent character of the gas-borided layer could be the cause of the formation of corrosive micro-cells. The privileged areas of micro-cells formation were those with different types of borides. In the gas-borided layer produced on Inconel^®600-alloy, nickel borides and chromium borides were formed. These borides differed in physical and chemical properties. As a result of these differences, the microstructure of the gas-borided layer produced on the Inconel^®600-alloy was divided into the anode region and cathode regions. Therefore, during the potentiodynamic corrosion test, micro-cells were formed between different types of borides, as shown in Figure 6. In the gas-borided layer produced on the Inconel^®600-alloy, micro-cells were created between the different types of borides designated MeI_xB_y and MeII_xB_y. Borides marked as MeI_xB_y donated valence electrons and passed into solution in the form of MeI⁺ ions. The electrons in the metal migrated to the cathode area (MeII_xB_y). Then, they combined with the depolarizer (D), i.e., an ion with the ability to add electrons (reduction).

Figure 6:

Scheme of work of corrosive micro-cells for a multiphase gas-borided layer.

3.4 Oxidation behavior at 1000 °C

SEM images of the cross-section of the non-oxidized and oxidized sample were presented in Figure 7. The appearance of the gas-borided layer was significantly different in comparison to the borided sample before the oxidation test (Figure 7a). In some regions, layer losses were observed (Figure 7b and d), and delamination of the layer from the substrate material (Figure 7c). This indicated a change in the phase and chemical composition of the borided layer. The oxidation test was carried out in an open-air atmosphere, which allowed for easy diffusion of oxygen into the sample. To confirm the phase composition of the gas-borided layer after the oxidation test, the XRD pattern was analyzed (Figure 8). It was found that the oxidation at 1000 °C for 24 h resulted in the formation of various oxides: Cr₂O₃, Cr₃O₄, NiO, B₂O₃, Fe₃O₄. However, the phases characteristic of the gas-borided layer have also been identified. Many of the peaks corresponded to nickel borides (NiB, Ni₂B, Ni₃B, and Ni₄B₃) and chromium boride (CrB).

Figure 7:

The SEM microstructure of the gas-boride layer produced on Inconel^®600-alloy: before oxidation test (a), after oxidation at 1000 °C for 24 h (b), visible delamination of borided layer after oxidation test (c), layer damage after oxidation test (d).

Figure 8:

XRD pattern of the gas-borided sample after oxidation test at 1000 °C for 24 h. 1, NiB; 2, Ni₂B; 3, Ni₃B; 4, Ni₄B₃; 5, CrB; 6, Cr₂O₃; 7, Cr₃O₄; 8, NiO; 9, B₂O₃; 10, Fe₃O₄.

It was found that after oxidation, the layer could be classified as a mixed boride-oxide layer. However, it was difficult to identify the phase distribution in the cross-section of the borided layer. Therefore, the linear EDS analysis was carried out in the cross-section of the specimen as shown in Figure 9. The concentration of boron and oxygen was higher in the area corresponding to the mixed boride-oxide layer. However, in some areas, the content of oxygen was significantly lower compared to the boron content. This may indicate the presence of chromium and nickel borides in these regions. Simultaneously, low boron concentration was detected in some areas, especially near the top surface. Chromium oxides were probably the dominant phases in this area. The presence of a compact, dense layer of boron oxide B₂O₃ after oxidation is important to prevent oxygen penetration. This layer, formed on the top surface of the boride layer, could be a barrier to further diffusion of oxygen through the substrate material (Kartal Sireli et al. 2017). In the present study, the presence of boron oxide was confirmed by XRD analysis, however, this phase was not formed as a compact layer on the top surface.

Figure 9:

SEM image of the oxidized sample with results of EDS line analysis.