The effect of chromium content on the corrosion behavior of ultrafine-grained Cr_xMnFeCoNi high-entropy alloys in sulfuric acid solution

Alloy fabrication

The Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs (hereafter denoted as Cr0, Cr0.6, Cr1, Cr1.5, respectively) were prepared using commercially pure (99.9 wt.%) elemental powder of Cr, Mn, Fe, Co and Ni as the starting material. The powder mixture was put into a stainless steel vial with a ball-to-powder mass ratio of 5:1 and subjected to high-energy ball milling for 9 h using a SPEX 8000D mill at ambient temperature in an argon glovebox. The ball-milled powders were consolidated by spark plasma sintering (SPS-211Lx, Fujidempa Kogyo. Co., Ltd, Japan) at 950 °C for 5 min under a pressure of 45 MPa in vacuum condition (~ 6 Pa).

The sintered cylindrical HEAs were cut into discs using electrical discharge machining for subsequent characterization and tests. For electrochemical tests, the samples were cut into discs with a size of Φ = 10 mm × 3 mm, and sealed in epoxy with one surface exposed. The exposed surface was first mechanically ground using SiC paper down to 2400 grit, followed by vibratory polishing using a 0.06 μm alumina suspension. The polished samples were then cleaned with deionized water and alcohol and dried in air. For static immersion tests, both surfaces of the samples of the discs were ground and polished following the same procedure as those used for the electrochemical tests.

Microstructure characterization of the as-sintered HEAs

The phase of the as-sintered Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs was identified using XRD patterns recorded on a Rigaku Smartlab-9KW diffractometer with Cu-Kα radiation (45 kV, 200 mA). The data were collected with 2θ in the range of 30° to 80° with a step size of 0.02° at a scanning speed of 10°/min. The surface morphology and composition were examined using SEM (TESCAN MIRA 3, Czech Republic) coupled with energy-dispersive X-ray spectroscopy (EDS). Electron backscatter diffraction (EBSD) was performed to characterize the grain structure of the HEAs. TEM (FEI Tecnai F30, USA) and high-angle annular dark-field scanning TEM (HAADF-STEM) were used to characterize the microstructure of the as-sintered HEAs. The TEM samples were prepared using site-specific focused ion beam (FIB; FEI Helios Nano Lab^TM 600i) lift-out technique. Besides, platinum cap layers with a thickness of ~1 μm were deposited before milling the trench to protect the surface.

Electrochemical corrosion and immersion tests

Electrochemical measurements were performed in a 0.5 M H₂SO₄ solution at ambient temperature using a Princeton Applied Research Versa Studio (PARSTAT 4000A, Ametek Scientific Instruments, USA). A standard three-electrode cell system was performed during all the measurements using a platinum sheet as the counter electrode, a saturated calomel electron (SCE) as the reference electrode and the Cr_xMnFeCoNi HEAs as the working electrode. Before EIS and potentiodynamic polarization tests, the specimens were subjected to the potentiostatic polarization at -0.6 V vs. SCE for 5 min to remove the surface oxidation, and then the open-circuit potential (OCP) was measured for 2 h in order to obtain a steady or quasi-steady potential value. EIS tests were performed at OCP with the sinusoidal potential amplitude of 10 mV in the frequency range from 100 kHz to 10 mHz. The data derived from EIS tests were fitted and analyzed using the Zsimpwin software. Then, the potentiodynamic polarization tests were conducted with the scanning range from -0.25 V vs. OCP to 1.3 V vs. SCE at a scan rate of 1 mV/s. Each electrochemical test was performed three times to confirm the reproducibility of the results.

The static immersion tests were performed in 0.5 M H₂SO₄ at ambient temperature for 15 days. Three parallel experiments were performed for each HEA. The average corrosion rate by mass loss C_w in mm/y was calculated using the following Eq. (1):

where k is the constant of 8.76 × 10⁴, ∆W is the mass loss (g) after the immersion test, A is the surface area (cm²), t is the immersion time (h), and ρ is the density (g/cm³) of the sample, measured by Archimedes method.

Corroded morphology and composition characterization

The corroded surface morphology was examined using SEM and AFM (Dimension edge, Bruker Corporation, Germany). The surface composition was analyzed using EDS and XPS (Thermo Scientific K-Alpha⁺). The standard C 1s spectrum at a binding energy of 284.8 eV was employed to calibrate the original peaks, and the calibrated XPS data were analyzed and fitted by the software Avantage 5.9. TEM and HAADF-STEM coupled with EDS were used to analyze the typical corrosion micropores at the surface.

Phase, composition and microstructure of the as-sintered HEAs

The measured density of all the sintered Cr_xMnFeCoNi HEAs exceeds ~ 98%. Very few pores were located on the surface [Figure 1A-D], confirming the high density of the HEAs. Some randomly distributed tiny dark spots were found on the surface. EDS line scanning profiles [Supplementary Figure 1] show that such dark spots are rich in Cr, Mn and O. EDS elemental maps [Figure 1] show the uniform distribution of Co, Fe, Mn and Ni, while Cr shows segregation and inhomogeneity. With a higher concentration of Cr, the Cr-rich region in the HEA is more obvious. Quantitative EDS analysis [Table 1] shows that the sintered HEAs have very close compositions to the nominal input ones.

Figure 1. Surface morphology and EDS elemental maps of the as-sintered Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs. (A-A₄) Cr0; (B-B₅) Cr0.6; (C-C₅) Cr1; and (D-D₅) Cr1.5. EDS: Energy-dispersive X-ray spectroscopy; HEAs: high-entropy alloys.

XRD patterns of the Cr_xMnFeCoNi HEAs [Figure 2] show that the spark-plasma-sintered HEAs have a dominant fcc matrix phase with a trace amount of (Cr, Mn)O_x oxides owing to the relatively low vacuum during the sintering process and the high chemical activity of Cr and Mn. The oxide phase content increases with the Cr concentration. The XRD patterns confirmed that the tiny dark particles found on the surface were oxides. EBSD inverse pole figure (IPF) maps [Figure 3] show that the Cr_xMnFeCoNi HEAs have equi-axed grain structure and very similar average grain size, in the range of 312 to 360 nm. Furthermore, TEM-based techniques were used to characterize the microstructure in more detail. The HAADF-STEM image [Figure 4A] shows three contrasts. A considerable number of nano-sized precipitates in grey contrast are distributed at the grain boundaries and inside the grains of the matrix, where a few black precipitates are located only at the grain boundaries. EDS quantitative analysis [Table 2] reveals that the matrix is depleted in Cr, while the grey precipitates are rich in Cr and Mn, and the black precipitates were identified to be (Cr, Mn)O_x, as detected in the XRD pattern in Figure 2. Bright-field TEM image [Figure 4B] clearly shows the microstructure of the Cr1 HEA. Selected area electron diffraction (SAED) patterns [Figure 4C-E] show that the matrix belongs to the Fm-3m space group, while both precipitates are in Fd-3m space group, consistent with the recent study^[26] that also reported that the (Cr, Mn)O_x phase belonged to the Fd-3m space group in the Cr1 HEA fabricated using medium frequency induction melting. Combined with the EDS quantitative analysis, the black precipitates were identified to be (Cr, Mn)₃O₄.

Figure 2. XRD patterns of the Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs. XRD: X-ray diffraction; HEAs: high-entropy alloys.

Figure 3. EBSD IPF maps and the average grain size distribution of the as-sintered Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs. (A) Cr0, (B) Cr0.6, (C) Cr1 and (D) Cr1.5. EBSD: Electron backscatter diffraction; IPF: inverse pole figure; HEAs: high-entropy alloys.

Figure 4. TEM characterization of the as-sintered CrMnFeCoNi (Cr1, x = 1) HEA. (A and B) are HAADF-STEM and corresponding bright-field TEM images; (C-E) are SAED patterns of the matrix, the grey precipitate and the black precipitate in (A), respectively. TEM: Transmission electron microscopy; HEA: high-entropy alloy; HAADF-STEM: high-angle annular dark-field scanning TEM; SAED: selected area electron diffraction.

Potentiodynamic polarization

Figure 5 shows the potentiodynamic polarization curves of Cr_xMnFeCoNi HEAs in a 0.5 M H₂SO₄ at room temperature. The values of E_corr and I_corr derived from the polarization curves were included in Table 3. With the increase of the Cr concentration, the E_corr shows more negative value from -0.40 V to -0.45 V, which suggests a more severe corrosion tendency, but not the actual corrosion rate at the moment in terms of kinetics^[27,28]. The increment in the corrosion sensitivity may be due to the presence of more Cr-rich phases and the oxides precipitated in the matrix. In comparison with the CoFeMnNi HEA (Cr0) in the absence of the Cr, the I_corr of Cr0.6 HEA first increases and then gradually decreases with the increasing Cr concentration. The Cr0.6 HEA shows the highest I_corr of 5.36 μA/cm² and the Cr1.5 HEA shows the lowest I_corr of 2.75 μA/cm². Since the corrosion rate is positively correlated to I_corr, Cr0.6 has the lowest corrosion resistance, which may be attributed to the galvanic corrosion between the Cr-rich phase and the matrix as well as the lack of protection from the passive film. On the contrary, the Cr1.5 HEA has the highest corrosion resistance. However, when the applied potential reaches ~ 0.8 V, the current density of Cr1.5 increases faster than Cr0, suggesting that the Cr1.5 HEA has inferior pitting resistance.

Figure 5. Potentiodynamic polarization curves of the as-sintered Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs in a 0.5 M H₂SO₄ solution at room temperature.

Surface morphology and composition after polarization tests

After potentiodynamic polarization tests in the 0.5 M H₂SO₄ solution at room temperature, the Cr_xMnFeCoNi HEAs surfaces were covered with uniformly distributed micro-and nano-pores [Figure 6]. The average size of the pores increased, but the number reduced on the corroded surfaces of the HEAs with a higher concentration of Cr, suggesting stronger tendency towards localized corrosion. AFM height images [Figure 6A3-D3] show that the depth of pores is in the range of 170 - 360 nm for Cr0, Cr0.6, and Cr1 HEAs. The Cr1 HEA shows the smallest depth of pores of ~ 180 nm, which might be attributed to the enhanced protective effect of the passive film to inhibit further corrosion from the etching solution. However, the surface micropores of the Cr1.5 HEA show much larger depth and reach ~1 μm, which could be liable to cause galvanic corrosion in the deep regions.

Figure 6. Surface morphology (A₁-D₁, A₂-D₂) and AFM height images (A₃-D₃) of the Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs after polarization tests in a 0.5 M H₂SO₄ solution at room temperature.

Quantitative EDS analysis on the honeycomb-like surface of Cr_xMnFeCoNi HEAs [Figure 6A2-D2] is included in Table 4. Inside the pores, all the four groups of HEAs have reduced content of Fe, Co and Ni but increased O, suggesting the selective dissolution of the elements Fe, Co and Ni and the formation of oxide films occurred simultaneously during the corrosion process. It is noted that the peak positions (point 1, Figure 6D2) of the Cr1.5 HEA surface contain a much higher content of O, indicating a better passivation effect on the region outside the micropores. However, the content of O in the deep valley of the pores (point 3,Figure 6D2) is lower than the shallow ones (point 2, Figure 6D2), and much lower than the peak positions (point 1,Figure 6D2), suggesting that the formation of oxide films was inhibited in the deep micropores. Hence, the region inside the deep micropores (point 3, Figure 6D2) of the Cr1.5 HEA would have the worst pitting corrosion resistance, as also confirmed by the content of selective dissolution elements. Furthermore, the corrosion behavior would be deteriorated due to the weak sites in sufficiently deep micropores during further corrosion and could cause pitting-like corrosion^[29,30].

XPS was performed to identify the chemical state of the corrosion products on the corroded surface of Cr_xMnFeCoNi HEAs after polarization tests. The high-resolution XPS spectra and the quantitative analysis results were presented in Figure 7 and Table 5, respectively. For all the four groups of HEAs, the Co 2p_3/2 spectrum is split into four constituent peaks, including Co (778.3 eV), CoO (780.2 eV), Co₂O₃ (781.3 eV), and Co(OH)₂ (782.1 eV). In addition, a Fe LM2 peak at 783.7 eV was also detected. The Fe 2p_3/2 spectrum can be divided into four peaks as Fe at 707 eV, FeO at 709.3 eV, Fe₂O₃ at 710.8 eV, Fe(OH)₃ at 711.3 eV along with the Co LM2 at 712.7 eV, but the Fe in Cr0 HEA shows much lower intensity [Figure 7A2] compared with the Cr-containing HEAs. The Mn 2p_3/2 spectrum is separated into four peaks as Mn at 638.8 eV, MnO at 640.5 eV, Mn₂O₃ at 641.5 eV, and MnO₂ at 642.6 eV along with Ni LM2 at 640.7 eV. The relative content of Mn oxides is larger than that of Co, Fe and Ni, especially for Cr0 and Cr1.5 HEAs. The Ni 2p_3/2 spectrum is separated into four peaks with Ni at 853 eV, NiO at 854.2 eV, Ni(OH)₂ at 855.5 eV and Ni_sat at 859.4 eV. Based on the relative intensity of each element, it is concluded that the corroded surfaces are mainly composed of the metallic Ni and Co and small amounts of oxides/hydroxides. The Cr0 and Cr1.5 HEAs contain more Fe oxides/hydroxides. Moreover, the Cr 2p_3/2 spectrum for Cr-containing HEAs is separated into three peaks as Cr at 574.2 eV, Cr₂O₃ at 576.1 eV and Cr(OH)₃ at 577.1 eV. The relative contents of Cr oxide and hydroxide are obviously higher compared to those of Co, Fe and Ni for Cr1.5 HEA.

Figure 7. High-resolution XPS spectra of the corroded surface on Cr0 (A₁-A₄), Cr0.6 (B₁-B₅), Cr1 (C₁-C₅), and Cr1.5 (D₁-D₅) HEAs after polarization in a 0.5 M H₂SO₄ solution at room temperature. (A₁-D₁) Co 2p_3/2; (A₂-D₂) Fe 2p_3/2; (A₃-D₃) Mn 2p_3/2; (A₄-D₄) Ni 2p_3/2; and (B₅-D₅) Cr 2p_3/2. XPS: X-ray photoelectron spectroscopy.

Compared with Cr0 HEA in the absence of Cr, the Cr0.6 HEA shows significantly reduced content of Fe and Mn oxides/hydroxides in the oxide film, which thus shows weaker corrosion resistance. Likewise, as the Cr concentration increases from Cr 0.6 to Cr1.5, the increased amount of oxide/hydroxide of Cr, Mn and Fe provides denser and more protective oxide film for Cr1.5 HEA than Cr0.6 and Cr1 HEAs. However, the enrichment of Mn in the oxide film could degrade the stability of the passive film^[31], which could lead to the breakdown of the oxide film in the local region of the Cr1.5 HEA during severe corrosion. On the whole, the quantitative analysis results suggest that Co and Ni are preferentially dissolved, which are in agreement with the results of the EDS point analysis in Table 4. The content of Fe does not reduce obviously for Cr0, Cr0.6 and Cr1, except for Cr1.5, showing that the selective dissolution of Fe is more difficult than Co and Ni. It also demonstrates that Cr oxide (Cr₂O₃) and hydroxide (Cr(OH)₃) are indeed generated on the surfaces of Cr-containing HEAs in the nominal Cr concentration of ~13.0 at.% at least, which is close to the threshold in conventional stainless steels^[12].

In order to clarify the honeycomb-like structure of the surface, the cross-sectional microstructure of a typical corrosion micropore located on the surface of Cr1 was characterized by TEM in detail, as shown in Figure 8. The depth of the micropore is ~200 nm, in line with the result of the AFM height image in Figure 6. Furthermore, the selected cross-section can clearly be divided into two regions. The top region above the interface (the dotted line in Figure 8A) is referred to as the oxide film inside the micropore, while the region below the interface is the un-corroded matrix. High-resolution TEM image [Figure 8A] along with fast Fourier transform (FFT) patterns [Figure 8B and C] reveals that the oxide film consists of both amorphous and nanocrystalline (Cr, Mn)O_x. The boundaries between the nanocrystalline and the amorphous phases are considered to provide tunnels for species diffusion and transport, where the corrosive ions permeate preferentially, similar to the mechanism that the Cl^- ion attacks on the oxide film of the metals^[32]. It might be induced by the detrimental effect of Mn, which could be concentrated in the oxide film to greatly affect the occurrence of the nanocrystalline phase. The EDS elemental maps [Figure 8E] indicate that the oxide film is rich in Mn, Cr and O, but deficient in Fe, Co and Ni.

Figure 8. TEM characterization of a typical corrosion micropore on the surface of Cr1 HEA. (A) Bright-field TEM image; (B-D) are the high-resolution TEM image obtained from the interface region and the corresponding FFT patterns, respectively; and (E-E₇) HAADF-STEM image with the corresponding EDS elemental maps. TEM: Transmission electron microscopy; HEA: high-entropy alloy; FFT: fast fourier transform; HAADF-STEM: high-angle annular dark-field scanning TEM; EDS: energy-dispersive X-ray spectroscopy.

EIS

To understand the corrosion mechanism of Cr_xMnFeCoNi HEAs in the 0.5 M H₂SO₄ solution, the EIS measurements under OCP conditions were conducted and analyzed in Figure 9. In terms of the Nyquist plots [Figure 9A], all HEAs have similar features where a depressed capacitive semicircle covers high frequency region and some scattered points are distributed randomly near the real axis in the low frequency region. The diameter of the capacitive semicircle first decreases from Cr0 to Cr0.6 and then increases to Cr1.5, indicating the same trend for the corrosion resistance^[22,33]. It is found that only one time constant is presented in the high frequency of each Bode plot [Figure 9B], indicating the corrosion behavior of the passive film. With the decrease of the frequency, the phase angle reduces. The capacitance for the passive film increases and the corrosion resistance decreases, suggesting that the etching solution gradually sinks into the passive film. Moreover, in the wide range of the low frequency region, the phase angles are all approximately 0° and the values of |Z| are constant, which are characteristic of the pure resistance, indicating that the passive films on the HEAs are not very compact. Likewise, the value of |Z| at a fixed frequency of 0.1 Hz is usually equal to the polarization resistance related to the corrosion resistance^[34]. Therefore, the corrosion resistance follows the order of Cr1.5 > Cr0 > Cr1 > Cr0.6. On the other hand, there is no straight line representing the diffusion process or another capacitive semicircle corresponding pitting behavior present in the Nyquist plot, which is due to the simultaneous formation of the passive film at the corroded region when the etching solution reaches the matrix [Figure 6].

Figure 9. Nyquist (A) and Bode plots (B) of the Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs in a 0.5 M H₂SO₄ solution under open circuit potential condition. The inset in (A) is the electrical equivalent circuit (EEC) for fitting the EIS results. HEAs: High-entropy alloys; EIS: electrochemical impendence spectroscope.

In order to fit the results of the EIS under the OCP condition, an electrical equivalent circuit (EEC) model (inset in Figure 9A) was used. In the EEC model, a parallel combination of a constant phase element Q_p and a passivation layer resistance R_pcorresponding to one time constant is in series with a solution resistance R_s, which has good fitting reliability as the small chi-square values in the order of magnitudes of 10^-4 listed in Table 3. Herein, the ideal capacitance C_pis replaced by the constant phase element Q_pto compensate for the non-homogeneity and the capacity dispersion in the system^[35], such as an uneven corroded surface. The Q_p impendence Z_Q is given as follows,

Where Y_p is the proportionality factor, j is the imaginary unit, ω is the phase frequency, and n is the phase shift, reflecting the degree of the dispersion for an ideal capacitance. For n_p= 1, Z_Q represents a pure capacitance with C_p= Y_p; for n_p= 0.5, it presents a Warburg resistance; and for n_p= 0, it presents a pure resistance with R_p= Y_p^-1. As a result, n_p increases from Cr0 to Cr1, implying that the passive film becomes more compact on the surface. While from Cr1 to Cr1.5, n_p decreases slightly, which is ascribed to the discontinuous passive film containing the weak site in the deep micropore, as shown in Figure 6. Moreover, as the value of n_p is in the range of 0.74-0.84, which is close to 1, the value of Y_p can be approximately regarded as capacitance C_p of the passive film. Considering the Helmholtz model^[21,36], the value of Y_p is determined by the equation as follows,

where d is the thickness of the passive film, S is the exposed surface area, ε₀ is the permittivity of the vacuum, and ε is the dielectric constant of the surface which is mainly determined by the composition of the passive film. Therefore, the decrease of the value of Y_p from Cr0 to Cr1 and then the increase from Cr 1 to Cr1.5 can be affected by two factors for all samples: the change of the composition and the thickness of the passive film. Meanwhile, as Cr0.6 and Cr1 have a similar surface composition as analyzed by XPS [Figure 7], the larger Y_p of Cr1 is mainly induced by the thicker passive film than Cr0.6. Notably, the decreased Y_p for Cr1.5 should be mainly attributed to the change of the composition rather than the reduced thickness of the passive film. In terms of the resistance R_p, which reflects the corrosion resistance area of the passive film and the electron migration rate at the sample’s surface/electrolyte interface, the smaller value of R_p indicates the lower corrosion resistance. Based on this, the corrosion resistance for the four groups of HEAs first decreases and then increases with the increasing concentration of Cr, in line with the results of the polarization tests.

Long-time immersion

Figure 10 shows the surface morphology and average corrosion rate of Cr_xMnFeCoNi HEAs after immersion in a 0.5 M H₂SO₄ solution for 15 days at room temperature. The distribution of the micropores on the surface has no difference from that after polarization tests, but the micropores are much larger and the surfaces become looser. The formation of a more honeycomb-like corrosion surface, especially for Cr1.5, suggests that the HEAs suffered from more severe corrosion during the immersion of the 15 days, due to the absence of the long-term stable passive film. Likewise, the average corrosion rate [Figure 10E] increases exponentially with the Cr concentration. The Cr1.5 HEA shows the highest corrosion rate, 6 times higher than that of Cr1, which seems in contradiction with the results of the electrochemical tests. The significant change in the corrosion behavior during the long-time immersion between the four groups of HEAs could be attributed to the transformed corrosion mechanism. The dominant galvanic corrosion behavior has a direct correlation with the Cr concentration, which would be discussed in detail later.

Figure 10. Surface morphology (A-D) and the average corrosion rate (E) by mass loss of the Cr_xMnFeCoNi (x = 0, 0.6, 1, and 1.5) HEAs after immersion in a 0.5 M H₂SO₄ solution for 15 days at room temperature. HEAs: high-entropy alloys.

Honeycomb-like morphology after polarization tests

Interestingly, all samples show the honeycomb-like morphology after polarization tests, as shown in Figure 6, which means that the corroded micropores were uniformly distributed on the surfaces of Cr_xMnFeCoNi HEAs. A similar phenomenon has also been observed in the Cr1 HEA corroded in H₂SO₄ solution^[21,22]. Actually, such corrosion morphology for HEAs can be mainly affected by two factors: (i) the second phase; and (ii) the area of the grain boundaries. Considering that the four groups of HEAs show similar average grain sizes in the range of 312.5-360.6 nm (as shown in Figure 3), the difference in the area of the grain boundaries on the corrosion behavior should be negligible.

With the single fcc phase, the grain boundary has a dominant effect on the corrosion behavior for Cr0, which can act as a defect zone^[37,38] and the galvanic corrosion between the grain boundary and the intragranular zone can occur during the corrosion process. Moreover, the grain size of Cr0 is ~ 300 nm, which is close to the diameters of micropores after the polarization tests in Figure 6A. Thus, it is confirmed that the intragranular zone as the anode for Cr0 was preferentially corroded and the grain boundary as the cathode acted as the corrosion barrier at the early stage in H₂SO₄ solution, so that many small micropores were generated on the surface. However, for Cr0.6, Cr1 and Cr1.5, the Cr-rich phase such as (Cr, Mn)₃O₄ dissolved out from the matrix, causing a greater impact than the grain boundary on the corrosion behavior. Meanwhile, the corrosion for the Cr-containing HEAs occurred mainly around the (Cr, Mn)₃O₄ phase zone and the galvanic corrosion mainly occurred between the second phases and the matrix. With more segregation of the Cr and the increase of the average size of the Cr-rich zone at higher Cr concentration [Figure 1], the galvanic corrosion effect would become stronger as the increasing area ratio of cathode to anode, resulting in the larger and more dispersed micropores in the honey-like morphology [Figure 6]. The depth of the micropores was mainly increased by the galvanic corrosion effect and restrained by the protective effect of the passive film. As more severe galvanic corrosion present in Cr1.5, also indicated by the value of E_corr, much deeper micropores were formed on the surface than the other HEAs. The smallest depth of the micropores of Cr1 should be mainly attributed to its dense passive film that has not been damaged by the insufficient galvanic effect at this stage. Moreover, the tiny height fluctuation in or out of the micropores was observed in Figure 6, which may be caused by the galvanic effect between the matrix and the heterogeneous nanoscale particles, such as the (Cr, Mn)-rich phase shown in Figure 4.

Corrosion process during the polarization tests

It is found that when the four groups of HEAs were corroded in the 0.5 M H₂SO₄ solution, the selective dissolution proceeded and the element Co, Ni and Fe were preferentially released from the matrix, which may be attributed to the dissolution rate difference of the constituent elements in these HEAs in the corrosion environment^[7,39]. The hydroxides of Cr, Mn and Fe were deposited on the surface of the HEAs by the hydrolysis reactions at the same time, then some hydroxides formed early in the inner layer may dehydrate to form the oxides^[40,41], such as Cr₂O₃. Therefore, the complete passive films were formed. The major chemical reactions present in the above corrosion process are shown as follows:

Anodic reaction: M → Mⁿ⁺ + ne^- (M: constituent metal element of the HEAs)

Cathodic reaction: 2H⁺ + 2e^- → H₂

                                   O₂ + 2H₂O + 4e^- → 4OH^-

Hydrolysis reaction: Mⁿ⁺ + nH₂O → M(OH)_n + nH⁺

Figure 11 shows the schematic diagrams of the corrosion process for these four groups of HEAs during polarization in the 0.5 M H₂SO₄ solution. As discussed above, the intragranular zone near the grain boundary of Cr0 starts to corrode preferentially as the galvanic corrosion effect. With the selective dissolution proceeding, the corroded zone gradually turns into the Mn-rich zone at noble potential, and then the corrosion spreads from the nearest zone to the grain boundary into the grain interior. As the gradually increased area ratio of cathode to anode, the central zone suffers from more severe corrosion than the zone adjacent to grain boundary and the micropores are formed on the surface. Additionally, as described in Figure 11A, more metal ions would be released into the solution than the uncorroded zone by the redox reaction. This would facilitate the hydrolysis reaction to form stable metal hydroxide/oxide, such as Fe and Mn hydroxide/oxide, on the corroded zone, which can retard the corrosion process effectively. Meanwhile, when a small amount of Cr was added in Cr0.6, as shown in Figure 11B, the second phase of (Cr, Mn)₃O₄ would precipitate from the matrix, and the corrosion process then occurred near the second phase zone. As the grain boundary can be regarded as the corrosion barrier^[38], the corrosion process could end up in the grain boundary zone far away. Notably, such a corrosion process may have two routes: one is that the corrosion propagates along one direction, as the passive film in other directions is more compact; the other one is that the corrosion propagates around the second phase so that the second phase particles would locate in the micropores and peel off from the surface finally. Moreover, the passive film is not only formed in the corroded zone inside the micropores, but also out of the micropores as the slight corrosion by the effect of the nanoscale heterogeneous composition. In terms of Cr1 in Figure 11C, the alloy has the same corrosion mechanism as Cr0.6 except for the microstructure of the passive film. That is, Cr1 has a more compact and thicker passive film than Cr0.6 to provide effective protection to the matrix. However, the micropores for Cr1.5 are much larger and deeper than Cr1 and a few hydroxides/oxides are deposited in the micropores, as shown in Figure 11D. The larger micropores should be due to the larger second phase in the matrix, which would accelerate the corrosion in the local regions without enough protection from the passive film. It can be explained as follows: the dissolution rate of the anode in the micropores is so fast due to the strong galvanic effect that the concentrations of the metal ions are very high; when the diffusion rate of metal cations out of the micropore is much lower than the anodic dissolution rate in the activated region, the concentrations of the metal ions in the micropores could be maintained above a critical value, then the pH value decreases in the local regions by the excessive hydrolysis reactions, and thus causing a more aggressive environment^[42,43], which not only restrains the formation of the stable passive film in the micropores, but also induces the formation of the additional (Mn, Cr)-rich layer on the inner surface. In contrast, compared with Cr1, in the larger region outside the micropores of Cr1.5, the more compact and flat passive film following the same passivation mechanism without dissolution process is generated based on the results shown in Figures 6 and 7, which has more effective protection against the matrix, resulting in the lower corrosion rate for Cr1.5 during the polarization tests.

Figure 11. The schematic diagrams of the corrosion process during the polarization tests in 0.5 M H₂SO₄ for Cr0 (A), Cr0.6 (B), Cr1 (C) and Cr1.5 (D).

Corrosion process during the long-time immersion tests

It is interesting to find that when the HEAs were immersed in the 0.5 M H₂SO₄ solution at room temperature for 15 days, the corrosion rates increased quickly with the Cr concentration, which seems in contradiction with the results of the polarization tests. This contradiction can mainly be attributed to the competition mechanism between the passivation effect and the galvanic corrosion effect on the honeycomb-like surface. At the early period of the immersion tests, the corrosion behavior can be regarded as the same with the polarization tests^[44] and the EIS measurement after 2 h of immersion, as the EIS measurement is at a quasi-stable state and does not change the condition of the corroded surface, when the passivation effect plays an important role in the corrosion process for the samples. Therefore, with the addition of Cr, the stable Cr hydroxide/oxide would be deposited more on the surface, so that the passive film is more compact and thicker. As a result, Cr1.5 has the best corrosion resistance. With the extending time for Cr0, Cr0.6 and Cr1, the etching solution would gradually permeate the passive films in the micropores as the passive films are not very compact and stable. Meanwhile, the composition of the corroded area would convert to the Mn-rich or the (Cr, Mn)-rich region, which increases the area of the cathode. In the end, the enhanced galvanic corrosion effect is strong enough as that of Cr1.5 to increase the depth of micropores so that the passive films in the inner surfaces of the micropores would be dissolved. At this stage, the galvanic corrosion effect for all samples gradually accumulates and starts to transcend the impact of the passivation effect. Subsequently, the micropores continue to grow and the samples are corroded more severely. During this period, pitting-like corrosion is considered to proceed on the honeycomb-like surface, which is also promoted by the loose surface film. Therefore, the galvanic corrosion effect would play a dominant role during the following immersion. As mentioned, the galvanic corrosion would get more severe with the increase of the Cr concentration; hence, Cr1.5 gradually has the highest corrosion rate during the long-time immersion tests. Moreover, the porous structure is also found inside or outside of the micropores of the samples in Figure 10, which evolved from the tiny height fluctuation during the initial period of the immersion, due to the intrinsic heterogeneous nanoscale particles.

In summary, the addition of Cr in Cr_xMnFeCoNi HEAs is indeed liable to form the Cr oxide in the passive film, which has good stability in the solution and improves the corrosion resistance of the HEAs^[17,45]. On the other hand, the Cr can also precipitate from the matrix, which could accelerate corrosion by enhancing the galvanic corrosion effect. In addition, the formed honeycomb-like corroded surfaces on the HEAs would further deteriorate the corrosion resistance. Therefore, it is vital to optimize the matrix structure of the Cr-containing HEAs by changing the constituent elements as well as the processing methods to improve the corrosion resistance in future work. Generally, there are two methods to improve the corrosion resistance of the Cr-containing HEAs. One is to decrease or even eliminate the segregation of the Cr and homogenize the composition and the other is to optimize the structure and the composition of the passive film for high Cr-containing HEAs^[46] and eliminate the nanocrystalline-amorphous phase boundaries.