1 Introduction

High-chromium white irons (HCWI) are commonly used for the slurry pump components and pipelines in the Bayer refining process of alumina refineries [1, 2] due to their high wear resistance. The Bayer process solution is highly alkaline, and there are limited studies on the corrosion behaviour of the hardfacing alloys. There are also limited studies in the use of Pourbaix diagrams to predict a trend for the corrosion behaviour of hardfacing alloys.

Nelson [2] exposed hardfacing alloys consisting of Cr7C3 primary and eutectic carbides, and eutectic austenite to sodium hydroxide (pH 14) at 90 °C and found that the eutectic austenite corroded. Cast white irons were found to have better corrosion resistance than welds. Similarly, other investigators [1, 3] also found that the corrosion of white chromium cast irons occurred in the eutectic austenite or martensite. In contrast to this, in the study conducted by Salasi [4], on cast primary chromium carbides were found to corrode in preference to the eutectic austenite or martensite matrix. Tsai et al. [5] performed corrosion experiments on sintered pure chromium carbides (Cr23C6) 0.1 N NaOH that was deaerated and saturated with H2 before each experiment in the temperatures range of 275–325 °C. It was found that dissolution of chromium carbides occurred.

Salasi et al. [6, 7] found that in an alkaline solution of pH 12, a layer of Fe3O4 can form on carbides and a passive layer forms in the matrix eutectic austenite. In the studies conducted by Nelson [2] and Mcleod [8], in alkaline solution, it was proposed that the dissolution of the matrix occurred primarily due to depletion of chromium in a matrix. They also proposed that chromium depletion hindered the formation of passive film on the surface of austenite. The corrosion of either matrix [2, 8] or carbides [6, 7] depended on the formation of passive film on the matrix. Aziz et al. [3] proposed that the corrosion attack of the matrix in an alkaline solution occurred due to micro-galvanic interaction between matrix and the carbides, and chromium depletion. Similar ideas were also proposed by other investigators [2, 6, 8, 9].

Whilst these studies have done much to improve our knowledge on corrosion behaviour of high-chromium white irons, they did not discuss a method to predict corrosion behaviour. This is where the use of Pourbaix diagrams fits in. Pourbaix diagrams can not only be used to predict the corrosion behaviour of materials so that problems in the future can be avoided, but also provide simplified explanation of complicated chemistry and thermodynamics. Pourbaix diagrams can also be used to diagnose past and present problems in a thoughtful manner. They can also be used to design suitable experiments in an electrolyte of interest. Chen et al. [10] have used the Pourbaix diagrams of metals to explain the behaviour of corrosion of metals at elevated temperatures in energy industries. With this in view, Varmaa et al. [11, 12] developed Pourbaix diagrams for carbides in order to explain the trend observed in the corrosion behaviour of carbides. In this experimental study, the corrosion of hardfacing alloys based on HCWI will be explained through the use of Pourbaix diagrams of carbides.

2 Experimental Method

Hardfacing alloys based on HCWI were deposited on low carbon steel using shielded metal arc welding (SMAW). The typical composition of the deposited hardfacing alloys is given in Table 1. Alloy 1 mainly had high carbon and chromium while Alloy 2 had a number of elements, such as niobium, molybdenum and tungsten. These elements would form carbides in preference to chromium, and hence this sample was likely to contain complex carbides. The recommended welding parameters provided by the manufacturer of the hardfacing alloys are given in Table 2. To avoid the dilution, welding was performed in four layers. Samples were cut from the weld deposits and were prepared for microstructural observations. The samples were etched in 20 ml distilled water, 10 ml HCl, 30 ml HNO3 and 5 g of FeCl3 [13].

Table 1 Chemical compositions (wt%) of hardfacing alloys
Full size table
Table 2 Welding parameters
Full size table

Corrosion tests were conducted using electrochemical methods with a Gamry Reference 3000 potentiostat and a three-electrode cell. The HCWI (sample area = 2.85 cm2) was the working electrode exposed to the electrolyte 2 M NaOH solution of pH ≈ 14. The graphite and silver/silver chloride were the counter and reference electrodes, respectively. Before measuring corrosion potential, the alkaline solution in the corrosion cell was deaerated using nitrogen gas for 30 min, as described in the ASTM Standard G5-94 [14]. The open-circuit potential (E corr) was measured after 10 min, and then the potentiodynamic test was conducted at the scan rate of 0.17 mV/s [14]. The potentiodynamic test was continued till the sample showed transpassive behaviour. Then, the microstructure was examined using an optical microscope by removing the corrosion product using rubber cork under running tap water [2]. The potentiodynamic tests were conducted three times for each specimen to determine the consistency of the curves.

On the basis of the potentiodynamic test results, potentiostatic etching was conducted at different potentials on HCWI samples to determine the effect of corrosion in the microstructures. The potentials, points 1–6, were selected based on the changes in the anodic polarization curves or approximately in the middle of an anodic curve segment if there was no change in the anodic polarization curves. The intention was to cover the anodic polarization curve reasonably. The points 1–6 are given in Table 3, and the microstructures of the samples were recorded. Tafel plots were also conducted at −150–+150 mV with the scan rate of 0.2 mV/s to determine the corrosion behaviour. The electrochemical data analysis was done using Echem Analyst software [15].

Table 3 Potentials used for potentiostatic tests from the anodic curves of alloy 1 and 2
Full size table

The scanning electron microscopy (SEM, Philips XL 30 series) was used to study the microstructure, and energy dispersive X-ray spectrometry (EDS) was used study the chemical composition of the samples before and after corrosion. The ICPMS was conducted to analyse the chemical composition of the electrolyte after the corrosion tests.

3 Results

3.1 Microstructure Characteristics of HCWI Alloys

Figure 1 shows the microstructure of alloy 1 hardfacing sample. The microstructure shows primary carbides surrounded by eutectic carbides and austenite [16, 17]. The primary carbides are polygonal and some are needle shaped depending on the direction of solidification. The primary and eutectic carbides are of the type M7C3 and M23C6, respectively [18–24]. Figure 2 shows the microstructure of alloy 2 sample. The microstructure shows again primary carbides surrounded by eutectic carbides and austenite. The microstructure also shows some light regions of niobium carbide (NbC) [21].

Fig. 1
figure 1

Microstructure of alloy 1 sample showing polygonal primary carbides surrounded by eutectic carbides and austenite

Full size image
Fig. 2
figure 2

Microstructure of alloy 2 sample showing polygonal primary carbides surrounded by eutectic carbides and austenite. a1 is NbC

Full size image

The SEM pictures of the alloy 1 and alloy 2 samples are given in Fig. 3i, ii. The point elemental analysis was done on primary carbides (a), carbide/matrix interfaces (b), matrix (c) and eutectic carbides (d), and the values are given in Table 4. In Fig. 3i, ii, the areas marked d1 were found to be molybdenum carbide. In Fig. 3ii, the area marked a1 was found to be niobium carbide. From Table 4, it can be seen that the primary and eutectic carbides (a and d, respectively) had high amounts of chromium and iron in both the samples. Area b denotes carbide–austenite interface and shows that the chromium is lower than 10 % for alloy 1 sample and 14.38 % for alloy 2 sample. Area c denotes austenite for both alloy 1 and alloy 2 samples, where the percentage of chromium is higher in alloy 2 samples than in alloy 1. This may be due to the presence of strong carbide-forming element, niobium, that allows chromium to remain in austenite [25]. The area d1 of both alloy 1 and alloy 2 samples exhibits the presence of molybdenum in eutectic carbide with a lower amount of chromium compared to area d.

Fig. 3
figure 3

SEM microstructure analysis of i alloy 1 and ii alloy 2 hardfacing samples

Full size image
Table 4 EDS analysis (wt%) for hardfacing samples
Full size table

3.2 Corrosion Behaviour of HCWI Alloys

3.2.1 Potentiodynamic and Potentiostatic Test

The potentiodynamic tests were conducted by exposing the HCWI samples in 2 M NaOH solution of pH 14. The results of the three potentiodynamic tests are given Figs. 4 and 5 for alloy 1 and alloy 2 samples, respectively. It can be seen that there was very little difference between the three tests for both the samples. The curves did not exhibit a typical active–passive behaviour but exhibited an active behaviour all through. Although the curves were very nearly the same for both alloy 1 and alloy 2 samples, there were a few minor differences. The transpassive region occurred at about 550 mV for alloy 1 samples, while it occurred at about 600 mV for the alloy 2 samples.

Fig. 4
figure 4

Anodic polarization curve for alloy 1

Full size image
Fig. 5
figure 5

Anodic polarization curve for alloy 2

Full size image

The microstructures after the potentiostatic etching are shown in Figs. 4 and 5. For the alloy 1 sample, corrosion occurs at the eutectic carbides at points 1–4 in Fig. 4 followed by the corrosion of primary carbides at points 5 and 6, as shown in the microstructures of Fig. 2. The microstructure of eutectic carbide corrosion is shown in Fig. 6, and one can see that the primary carbides are still intact. Figure 7 shows the microstructure of the alloy 1 sample at point 6 of the polarization curve in Fig. 4. It can be seen that the carbides have dissolved leaving just the eutectic austenite without corrosion.

Fig. 6
figure 6

Corrosion of eutectic carbides in alloy 1 sample

Full size image
Fig. 7
figure 7

Corrosion of eutectic and primary carbides in alloy 1 sample and SEM/EDS microstructure analysis of matrix sample (a) higher magnification of dissolved carbides

Full size image

For alloy 2 samples at point 3 in Fig. 5, niobium carbides have undergone extensive corrosion, as seen in Fig. 8. At the transpassive region in Fig. 5, it can be seen that the primary and eutectic carbides have undergone corrosion, as seen in Fig. 9. In summary, it can be said that primary carbide dissolution happens at higher oxidation potentials than the eutectic carbides and niobium carbides.

Fig. 8
figure 8

Corrosion of niobium carbide in alloy 2 sample

Full size image
Fig. 9
figure 9

Corrosion of NbC, eutectic and primary carbides in alloy 2 sample

Full size image

The dissolution of primary carbides also occurred only at potentials 5 and 6 in Figs. 4 and 5. Even at these high potentials, eutectic austenite is protected. The final potentials of the work from Nelson [2] and Salasi [6] are given in Table 5. These potentials are then placed on the potentiodynamic curves in Fig. 10. Nelson [2] found that corrosion occurred mainly on eutectic austenite. The final potential of transpassive region for Nelson [2] which was placed in Fig. 10 shows that it was closer to potentiostatic test point 4 where eutectic carbides were found to corrode. However, it must be noted that Nelson did his corrosion tests at 90 °C. Although it is expected that the Pourbaix diagram regions will shift to lower potentials at high temperatures, it is interesting to note that tests conducted by Nelson [2] did not reach transpassive potential found in the present study. This is possibly the reason why corrosion did not occur in the carbides. However, Tsai et al. [5] found that chromium carbide (Cr23C6) did corrode at 300 °C. Cr23C6 is a less stable carbide than Cr7C3 [11]. The final potential from the tests done by Salasi [6] was closely matching with that of alloy 1 in this investigation. The microstructures from the work of Salasi [6] also showed that corrosion occurred on eutectic and primary carbides at the transpassive regions.

Table 5 The final potentials used for potentiodynamic test of HCWI at alkaline solutions conducted by other literature
Full size table
Fig. 10
figure 10

Final potentials (FP) used for potentiodynamic test of HCWI at alkaline solutions conducted by other literature

Full size image

3.2.2 Tafel Plot

The Tafel analysis was done on Echem Analyst to obtain the values of the slope of tangents (β A and β C), E corr and I corr. From the derived values of β A, β C and I corr for all the HCWI samples, the polarization resistance Rp values were determined from the below equation [26]

$${\text{Polarization resistance }}(Rp) \, = \, \frac{{\beta_{\text{A}} \beta_{\text{C}} }}{{2.3I_{\text{corr}} (\beta_{\text{A}} + \beta_{\text{C}} )}}.$$

The values, E corr, I corr, Rp and corrosion rate, are given in Table 6. The alloy 2 specimen shows better corrosion resistance than alloy 1, due to its higher corrosion potential and polarization resistance and lower corrosion current. The better corrosion resistance of alloy 2 sample may be due to the higher percentage of chromium in matrix than alloy 1 sample as shown in Table 4a.

Table 6 I corr, E corr and corrosion rate obtained from Tafel plot for hardfacing samples
Full size table

3.2.3 Microstructure Analysis After Corrosion

The SEM/EDS analysis was done after the corrosion test for alloy 1 sample to understand the dissolution/corrosion of elements associated in the corrosion of carbides as shown in Fig. 11. In Tables 4 and 7, the analysis of primary and eutectic carbides of alloy 1 sample before and after corrosion shows that significant dissolution of chromium element takes place at higher potentials in alkaline solution. The ICPMS tested results for alloy 1 sample in Table 8 revealed that significant dissolution of chromium (7600 µg/L) takes place at higher potentials in 2 M NaOH solution. There were no significant differences in percentage of chromium in the matrix after the corrosion test as given in Table 7c. The increase in percentage of iron in the carbides after potentiostatic corrosion test is clearly seen in Table 7 of alloy 1 sample, and this may be due to passive film growth and low dissolution of iron in alkaline solutions [27]. The partial dissolution of iron 3330 µg/L was detected using ICPMS. The EDS and ICPMS test results also showed that there was complete dissolution of molybdenum, manganese and vanadium elements at higher potentials in 2 M NaOH solution as given in Tables 7 and 8. After corrosion, the carbides become hollow in shape, and some cracks were found as shown in Fig. 7a. Hardness test was done on the alloy 1 sample before and after corrosion as shown in Fig. 12 to ascertain how corrosion affected carbides. This will help us understand how useful is the carbide to resist wear. It was found that there was a significant drop of hardness from 2299 HV before corrosion to 41.8 HV after corrosion in the carbides. This clearly showed that the carbides had indeed corroded at higher potentials in an alkaline solution and perhaps are not useful any longer to resist wear.

Fig. 11
figure 11

SEM microstructure analysis of alloy 1 hardfacing sample after corrosion

Full size image
Table 7 EDS analysis (wt%) of alloy 1 sample after corrosion
Full size table
Table 8 ICPMS Analysis in 2 M NaOH solution after corrosion test at point 6 of alloy 1 sample
Full size table
Fig. 12
figure 12

Hardness test a before corrosion, b after corrosion

Full size image

3.2.4 Corrosion Analysis of Carbide and Matrix Samples

After the carbides had dissolved in the potentiostatic tests, the potential of the eutectic austenite alone in the alkaline solution was measured for alloy 1 sample. For this, epoxy solution was placed on the surface of the sample, and the sample was repolished after the epoxy hardened. Similarly, in order to measure the potential of the carbides, the matrix was dissolved in an acidic solution [28]. Again, epoxy was applied to the sample surface, and it was repolished. After preparing carbide and matrix sample, corrosion potential of the samples was recorded in 2 M NaOH alkaline solution as shown in Table 9. The matrix sample shows nobler potential than a carbide sample in highly alkaline solution.

Table 9 Corrosion potential for carbide and matrix alloy 1 sample in 2 M NaOH solution
Full size table

4 Discussion

4.1 Galvanic Effect

The results showed that corrosion of HCWI in alkaline solution occurs mainly on carbides. The corrosion of the HCWI can be explained by micro-galvanic corrosion between carbide and eutectic austenite [3]. It was shown in this work that carbide exhibited a less noble potential than the eutectic austenite. Galvanic couple is formed between carbides and matrix. The noble potential of eutectic austenite matrix produces a cathode, and active potential of eutectic and primary carbides produces an anode. Due to potential difference, preferential dissolution of carbides occurs in preference to the matrix. Table 7 also confirms for alloy 1 sample, after corrosion test, that there is significant reduction of chromium content in the carbides but not in the matrix. Others [2, 6, 8, 9] have found similar galvanic corrosion effect for HCWI materials in alkaline media. The study conducted by Tsai et al. [5] suggests that the dissolution rates of carbides are higher in alkaline solution while in contact with platinum. This theory is in line with the findings of Salasi et al. [4]. However, they contradict the findings of others who found that carbides did not corrode [2, 3, 8]. It is possible that these investigators did not test till the transpassive region.

The corrosion trend of HCWI hardfacing deposits can be explained through the use of Pourbaix diagrams. Since the alloys have significant amounts of elements other than chromium, the superimposition of Pourbaix diagrams of other elements on chromium carbide Pourbaix diagrams will be useful in explaining the stability of these alloys in high pH solutions.

4.2 Stability of Carbides of HCWI and Pourbaix Diagrams

From the Pourbaix diagram of chromium carbide (Cr7C3) in Fig. 13, it can be seen at pH of 14, the immune region boundary is around −1.3 V (SHE). At higher potentials, Cr7C3 forms Cr2O3 followed by CrO4 2−. The rest potentials measured for the hardfacing alloys lie in the region of Cr2O3. The presence of iron (Fe) was identified in carbides and eutectic austenite as described in Table 4. There is a sort of galvanic effect described in the previous section due to the presence of Fe in carbides and eutectic austenite. It would be useful to superimpose the Fe Pourbaix diagram [29] with Pourbaix diagram of Cr7C3 as shown in Fig. 14.

Fig. 13
figure 13

Pourbaix diagram of in pH 14 with anodic potentials of HCWI [11]

Full size image
Fig. 14
figure 14

Pourbaix diagram of Cr7C3–Fe in pH 14 with anodic potentials of HCWI [11, 29]

Full size image

At the rest potential of the carbides, the hardfacing alloys undergo corrosion as seen in Figs. 4 and 5. The Tafel analysis also indicates that corrosion can be identified at rest potentials for HCWI hardfacing alloys, and this is borne out by the fact that the anodic polarization curve exhibited active behaviour. The alloys undergo some mild corrosion until point 4 in Figs. 4 and 5. Although the alloys are in the region of corrosion as shown in Fig. 14, due to the presence of Fe and its oxide Fe2O3, the alloys undergo only mild corrosion. The presence of Fe2O3 stops extensive corrosion from happening. However, on reaching the potential 5, there is substantial corrosion of carbides as Fe is also taken to the region of corrosion FeO4 2−. Corrosion of carbides can be seen in the microstructures in Figs. 7 and 9.

However, in the case of alloy 2, NbC was found to have no Fe as seen in Table 4. This is the reason NbC was found to corrode at lower potentials than Cr7C3 primary carbides as seen in Fig. 5. On superimposing NbC and Cr7C3 Pourbaix diagrams on each other in Fig. 15, it can be seen that NbO is formed at the rest potentials. NbC has undergone extensive corrosion at potentiostatic point 3, whereas Cr7C3 was still reasonably stable until potentiostatic point 6. NbC is found to be unstable only due to the absence of iron. Strehblow et al. [30, 31], McLeod et al. [32] and Salasi [7] have found that iron forms a thicker stable passive layer than chromium in highly alkaline solution. Further at potentiostatic point 4 in the Pourbaix diagram of Fe, FeO4 2− is the stable phase.

Fig. 15
figure 15

Pourbaix diagram of Cr7C3–NbC in pH 14 with anodic potentials of HCWI [11, 12]

Full size image

The molybdenum, manganese and vanadium are not stable and exist in the corrosion region at higher potentials as explained in the Pourbaix diagram of molybdenum [33], manganese [34] and vanadium [35] at pH 14. According to these Pourbaix diagrams, the elements molybdenum, manganese and vanadium may dissolve in the form of species MoO4 2−, MnO4 − and VO4 3−. The ICPMS test results as shown in Table 7 also revealed that the dissolution of above elements occurred at higher potentials as these are located in the corrosion region.

The dissolution of carbides in the hardfacing alloy shown in this work will have an effect on the usefulness of hardfacing material against wear. The hardfacing alloys even at open-circuit potential are in the active state in high pH solutions. Although it has been shown through Pourbaix diagrams [11] that there is the formation of Cr2O3 passive layer, it was argued that in order for the passive layer to form, Cr7C3 should undergo dissociation. It was found that iron in the primary carbides provides some stability to the carbides. However, under wear conditions, the passive layers would be removed and will likely be subjected to synergistic erosion–corrosion mechanism. Although no use of cathodic protection has been done on hardfacing alloys, it may be useful to apply cathodic potentials to place the hardfacing alloys in the immune region of the Pourbaix diagram. In that instance, the loss of material would be likely erosion and not the synergistic erosion–corrosion. Work is underway in this laboratory to introduce nitrogen into the hardfacing alloys to form nitrides, carbo-nitrides or simply nitrogen as an alloying element. Nitrogen has been added to stainless steels to improve corrosion resistance [36, 37].

By developing the Pourbaix diagrams for various carbides [12], it can be seen that the formation of passive layers will result in the dissociation of carbides, and erosive particles will likely remove the oxide layer resulting in erosion–corrosion. The dissociation of carbides to form the passive layer will likely render the carbides ineffective against wear. It was found that silicon and tungsten carbides also undergo corrosion [38, 39]. All carbides seem to have this issue. It is proposed that cathodic potentials may be another avenue that needs to be explored to improve the erosion–corrosion resistance of hardfacing alloys.

5 Conclusions

The corrosion behaviour of typical HCWI alloys with different chemical compositions in alkaline solutions was investigated. The four main conclusions drawn from this study are as follows:

  1. (1)

    The preferential corrosion occurs on carbides for all the HCWI samples at higher potentials.

  2. (2)

    Eutectic austenite attains noble potential and acts as a cathode, and the carbides attain an active potential.

  3. (3)

    The analysis of SEM/EDS and Pourbaix Diagram reveals that chromium, molybdenum and carbon are not so stable as iron at higher potentials in alkaline solutions.

  4. (4)

    Application of cathodic potentials to place the carbides in the immune region is likely to improve the durability of carbides in erosive–corrosive fluids.