Effect of High Deformation without Preheating on Microstructure and Corrosion of Pure Mg

Abstract

In this study, the relationship between the extrusion ratio and the corrosion resistance of pure Mg deformed using extrusion with an oscillating die (KoBo) without preheating of the initial billet was investigated. The materials investigated in this study were extruded at high deformation ratios, R₁ 5:1, R₂ 7:1, and R₃ 10:1, resulting in significant grain refinement from the very coarse grains formed in the initial billet to a few µm in the KoBo-extruded samples at room temperature, which is not typical for hexagonal structures. Our research clearly shows that KoBo extrusion improves the corrosion performance of pure Mg, but there is no straightforward dependence between the extrusion ratios and corrosion resistance improvement. Although it was expected that the smallest grain size should provide the highest corrosion resistance, the dislocation density accumulated in the grain interiors during deformation at the highest extrusion ratio, R₃ 10:1, supports dissolution reactions. This, in turn, provides the answers for the greater grain size observed after deformation at R₂ 7:1, where dynamic recovery prevailed over dynamic recrystallization. This situation led to the annihilation of dislocation, leading to better corrosion resistance of the respective alloy. Therefore, the alloy with the greatest grain size has the best corrosion resistance.

1. Introduction

Mg alloys, due to their advantageous ratio of light weight to high strength, tend to be excellent candidates for many industrial applications. Their use is desirable to reduce the mass of engineering systems, resulting in a reduction in CO₂ emissions, battery electrodes, and, of course, load-bearing degradable materials [1]. The increasing demand for lightweight structures drives modern researchers and industry to look for solutions that allow for faster, more economical, and easier methods of Mg deformation. To avoid cracking, resulting in a lack of bulk integrity, Mg and Mg alloys are deformed at elevated temperatures, where their limited slip planes are activated [2,3]. The most common conventional processing methods of Mg and Mg alloys are rolling, forging, and extrusion; their properties after deformation using these methods have been well explored [4,5,6]. Severe plastic deformation (SPD) methods are also well known for their significant refinement of Mg and Mg alloys [7,8]. Because Mg crystallizes in a hexagonal close-packed (hcp) structure, it has only two independent, easy-slip systems at room temperature [9], so most deformation methods are conducted at high temperatures [10]. Poor formability at room temperature is a critical factor that dramatically restricts the widespread use of Mg alloys in industry [11]. There have been some trials concerning the deformation of Mg at room temperature by high-pressure torsion [12]; however, this method is only able to produce samples in the form of discs, which are not suitable for many industrial applications.

Lately, extrusion with an oscillating die (called KoBo) has been used to form materials that are hard to deform at ambient temperature. KoBo is a modified extrusion method, which, thanks to the oscillating die assembled at the end of the extruder, is able to significantly reduce the grain size of hcp-structured material [13]. The combination of the extrusion and oscillations of the bilaterally rotating die forces the extruded materials to enter into a state of plastic flow, enabling the grain size to be reduced and improving the mechanical properties of the alloys. During KoBo, even complicated shapes can be formed at high extrusion ratios [14,15], thereby minimizing costs and maximizing the possibilities of extrusion. Therefore, the scientific and industrial interest in this method is high. To examine the method’s possibilities, many materials have been extruded at high extrusion ratios under various parameters [16,17], including the hcp-structured Mg alloys investigated by our research team [18,19]. To support the results of these investigations, numerical simulations have also been conducted [20]. Nevertheless, there are very few data regarding the corrosion behavior of alloys extruded using KoBo. The corrosion mechanisms are complex, and their characterization involves many aspects (grain size, precipitations, dislocations, various environments, residual stresses, etc.) [21]. Therefore, to minimize the impact of the alloy chemistry, and to clearly define the main parameters influencing the corrosion performance of KoBo-extruded materials, in this study, the corrosion of pure Mg is evaluated. The goal of this research was to explore the possibility of extrusion at high deformation ratios of pure Mg without preheating of the initial billet using upgraded extrusion with an oscillating die at the end of the extruder. To date, there have been no detailed investigations of the effects of KoBo processing on the subsequent microstructural and corrosion changes in pure Mg. The results of this clearly show that extrusion with an oscillating die located at the end of the extruder enables the deformation of pure Mg at high deformation ratios without preheating of the initial billet. The grain refinement obtained during KoBo processing improves the corrosion resistance of the initial billet; however, no clear relationship between grain size and corrosion rate increase was observed.

2. Materials and Methods

2.1. Materials and Extrusion Method

Pure Mg in the form of a rod with an initial diameter of Ø40 mm was extruded using KoBo to a diameter of Ø8, Ø6, or Ø4 mm, giving various extrusion ratios: R₁ 5:1, R₂ 7:1, and R₃ 10:1 (Figure 1a). Mg with a technical purity of 99.9% was cast using a Balzers furnace in graphite molds at 680 °C. During KoBo processing, the material was pressed through an extruder with a bilaterally rotating die located at its end. The die moves at high speed; therefore, the material enters into a viscoplastic state (Figure 1b). The initial billet temperature was 24 °C. The KoBo parameters were the same as those used for other Mg alloys: a punch speed of 0.2 mm/s, a die oscillating angle of ±8°, and an oscillation frequency of 5 Hz [18].

2.2. Microstructural Observations

Optical microscopy observations of the microstructure of the initial billet were performed using an optical microscope (AxioVision, Zeiss, Oberkochen, Germany). The microstructural characterization of the KoBo-extruded samples employed electron backscattered diffraction (EBSD) using a field-emission scanning electron microscope (FE-SEM, Hitachi SU-70, Hitachi Ltd., Tokyo, Japan) equipped with a Bruker EBSD detector (Bruker GmbH, Karlsruhe, Germany). To prepare samples for SEM observations and EBSD measurements, the extruded rods were cut into slices and polished with 1200-grit and 2400-grit SiC papers; subsequently, the surfaces perpendicular to the extrusion direction were polished with a low-energy Ar⁺ ion-beam milling system (Hitachi IM4000 Ion Milling System, Hitachi Ltd., Tokyo, Japan) for a duration of 4 h. To avoid oxidation, measurements were performed immediately after sample preparation. The high-resolution scanning electron microscopic observations were carried out in backscattered electron mode (BSE, FE-SEM Hitachi SU-70, Hitachi Ltd., Tokyo, Japan). The EBSD data were recorded with a step size of 200 nm and processed using ATEX software (Lorraine University, Metz, France) (www.atex-software.eu) [22]. The crystallographic orientations of the grains are presented in the form of inverse pole figure (IPF) maps, where various colors distinguish the orientation of a given sample direction in a crystal frame. If the angle between two neighboring grains was distorted by more than 15°, then the boundary between those grains was described as a high-angle grain boundary (HAGB); when the grains were misoriented by less than 15° (a cut-off limit of 3° was selected), the boundary was described as a low-angle grain boundary (LAGB). Internal grain deformation is presented in the form of grain orientation spread (GOS) maps.

The dislocation density in the Mg samples was determined by X-ray line profile analysis (XLPA). The patterns were measured by a high-resolution rotating anode diffractometer (type: RA-MultiMax-9, manufacturer: Rigaku, Tokyo, Japan) using CuKα₁ radiation (wavelength, λ = 0.15406 nm). The height and width of the rectangular X-ray spot on the sample surface were 1.5 and 0.2 mm, respectively. The measured peak profiles of hcp Mg were evaluated by the convolutional multiple whole profile (CMWP) fitting procedure [23]. In this method, the diffraction pattern is fitted by the sum of a background spline and the diffraction peaks obtained as the convolution of the instrumental peaks and the theoretical line profiles related to the crystallite size and the dislocations. The instrumental peaks were measured on a NIST SRM660a LaB6 peak profile standard material. About twenty reflections of Mg appearing in the diffraction angle (2θ) range between 30 and 150° were evaluated. The dislocation density was determined by employing the CMWP fitting evaluation procedure on the diffraction patterns. During CMWP evaluation, the diffraction pattern was fitted by the sum of a background spline and the convolution of the instrumental pattern and the theoretical line profiles related to the parameters of the microstructure, i.e., to the crystallite size and the dislocations. The theoretical peak profile functions used in the CMWP fitting were presented in ref. [24]. In the calculation of the theoretical peak profiles, the crystallites were modelled by spheres with a log-normal size distribution, and the peak broadening caused by the strain field of dislocations was described by the Wilkens model [24]. The CMWP analysis provided the median and the lognormal variance of the crystallite size distribution and the dislocation density. Since, in the present experiments, the peak broadening caused by the crystallite size was much lower than the instrumental broadening, only the dislocation density values obtained by the CMWP method are reported in this study.

Microstructural observations of the analyzed materials were carried out using a transmission electron microscope (TEM, JEOL JEM 1200, Jeol Ltd., Tokyo, Japan) with an accelerating voltage of 120 kV. Samples cut from a cross-section perpendicular to the rod axis, in the form of discs with a diameter of 3 mm, were mechanically ground to a thickness of approximately 100 µm. The samples were then subjected to an electrolytic thinning process using a Struers double-jet TenuPol 5 device (Struers GmbH, Willich, Germany). The electrolyte comprised a lithium chloride solution, magnesium perchlorate, methanol, and 2-butoxy-ethanol. Electrolytic thinning was performed at −45 °C with a voltage of approximately 90 V and a current of 100 mA. Ion milling using a Gatan 691 precision ion polishing system (PIPS, Gatan Inc., Pleasanton, CA, USA) at 3 kV for 10 min was necessary to remove the surface oxidation.

Microhardness (HV_0.2) tests were performed using the Vickers method under a load of 200 g (Innovatest Falcon 500 Micro/Macro Vickers Tester, Innovatest, Wiry, Poland). Ten points in line were measured on each material.

2.3. Corrosion Tests

Electrochemical measurements were carried out in naturally aerated 0.01 M NaCl solution using a Gamry FAS1 potentiostat (Gamry Instruments Inc., Warminster, PA, USA) equipped with three electrodes: platinum as the counter electrode, Ag/AgCl as the reference electrode, and the measured sample as the working electrode. The electrolyte was made up using analytical-grade reagents and distilled water. The samples were immersed for 12 h, and electrochemical impedance spectroscopy was performed after 1, 6, and 12 h of immersion. Afterward, potentiodynamic polarization tests were recorded over a range of 0.2 V below E_OCP to 1.0 V vs. Ref (a scan rate of 5 mV/s was used), with at least three tests being conducted for each microstructural condition. The EIS and polarization curves were fitted using Gamry Echem software Version 5.58. To obtain detailed information about the corrosion damage, the corroded surfaces of the samples were observed after 1 h of immersion under open-circuit conditions using SEM (Hitachi SU-70, Hitachi Ltd., Tokyo, Japan). To remove corrosion products, the observed surfaces were chemically treated in CrO₃ solution.

The corrosion rate of the investigated materials was calculated based on hydrogen evolution [25,26], and the concentration of Mg²⁺ ions in solution after the tests was measured using atomic absorption spectroscopy (AAS, GBC instrument, GBC Scientific, Keysborough, Australia). To determine the standard deviation of the measured data, three parallel samples were immersed for each extrusion condition.

3. Results

3.1. Microstructural Observations

The KoBo extrusion performed without initial preheating of the billet and at high deformation ratios resulted in significant grain refinement (Figure 2). The initial billet had a coarse-grained microstructure, which is typical for cast materials (Figure 2a). The KoBo extrusion caused a significant grain size reduction, with the distribution of grain sizes varying with respect to the changing extrusion ratio (Figure 2b–d).

The extruded microstructures show that the recrystallization intensity varied with respect to the extrusion ratio (Figure 3a–c). Interestingly, a greater extrusion ratio did not always cause a higher grain refinement. The measured grain size distribution plots are presented in Figure 2g. Surprisingly, the largest average grain size (d_avg 3.88 µm) and the widest grain size spread were observed in the sample extruded to Ø6 mm, R₂ 7:1 (Figure 2f), while samples subjected to a lower (Ø8 mm, R₁ 5:1) or greater (Ø4 mm, R₃ 10:1) reduction exhibited greater grain refinement (d_avg 2.88 µm and d_avg 1.20 µm, respectively) with a narrow grain size distribution. This trend was confirmed by the GOS map (Figure 3e). The XRD results clearly showed that the dislocation density accumulated in the sample extruded to Ø6 mm was lower than those in the other two KoBo-extruded samples (

Table 1. Dislocation density measured using XRD (detection limit of X-ray line profile analysis is 10¹³ m⁻²).

Sample	Ø40 mm	Ø8 mm (R1 5:1)	Ø6 mm (R2 7:1)	Ø4 mm (R3 10:1)
Dislocation density	Below detection limit	(7 ± 1) × 1013 m−2	Below detection limit	(5 ± 1) × 1013 m−2

). The same trend was observed when the GOS_avg values were analyzed (Figure 3d–f). The dislocation densities were found to be (7 ± 1) × 10¹³ m⁻² and (5 ± 1) × 10¹³ m⁻², respectively, for the KoBo samples extruded to Ø8 mm (R₁ 5:1) and to Ø4 mm (R₃ 10:1). Since the detection limit of the XRD method was 10¹³ m⁻², we were not able to obtain the numerical value of the dislocation density in the sample extruded to Ø6 mm (R₂ 7:1).

The orientation of the grains was random, with the majority oriented to (10-10) and to (2110) (Figure 3a–c). Moreover, the GOS distribution maps indicated around 85% undeformed grains with a GOS value of <5° for the materials extruded at extrusion ratios R₁ 5:1 (Figure 3d) and R₃ 10:1 (Figure 3f), indicating the completion of the dynamic recrystallization (DRX) processes [27,28,29]. After deformation at R₂ 7:1, the GOS map (Figure 3e) revealed a slightly lower recrystallized fraction of about 80%, with a visible substructure (LAGB) within the larger grains. All the investigated samples exhibited a $\left\{10\overline{1}0\right\}$ fiber texture, typical for extruded HCP materials, with a similar intensity for all the samples (Figure 4).

TEM images revealed that the microstructures of the extruded samples were composed of both heavily deformed areas with dislocation tangles inside (yellow arrows, Figure 5) and nearly equiaxed grains with high-angle boundaries and a low dislocation density inside them (red arrows, Figure 5). The microstructure of the sample extruded to Ø8 mm (R₁ 5:1) consisted of grains with various sizes (Figure 5a) and well-defined boundaries. These factors confirm that DRX takes place during KoBo extrusion [30,31]. At higher deformation ratios (R₂ 7:1, Ø6 mm), defects in the crystalline structure in the form of dislocations were also observed, and heavily deformed grains were formed (Figure 5b). The differing ratios of the two areas indicated that microstructural reformation during DRX and dynamic recovery (DRV) occurred with a different intensity than in the sample extruded at R₁ 5:1. KoBo extrusion at the highest deformation ratio (R₃ 10:1) did not affect microstructural reformation significantly when compared to the sample deformed at R₂ 7:1 (Figure 5c); however, during TEM observations, a higher number of small grains was distinguishable, which is in agreement with the EBSD data (Figure 3). The microhardness results shown in Figure 6 indicated that increasing the deformation ratio from R₁ 5:1 (Ø8 mm) to R₂ 7:1 (Ø6 mm) did not change the microhardness of the materials significantly. The average values of microhardness calculated for these samples were similar: 40.1 ± 1.3 and 41.6 ± 1.2, respectively. Both materials exhibited similar microstructures, particularly with regard to the presence of areas with large, well-developed grains and smaller grains with defects inside. The highest extrusion ratio resulted in a slight increase in microhardness (about 6 HV_0.2), which may be related to the smaller grain size formed.

3.2. Corrosion Tests

The open circuit potential evolution during 12 h of immersion in 0.01 M NaCl is shown in Figure 7a. All the curves exhibited initial increases toward more positive values; however, oscillations were observed in the curves recorded for the initial billet and the samples extruded to Ø8 mm (R₁ 5:1). Such behavior indicates the strong anodic and cathodic reactions proceeding on the surface. This is consistent with the polarization curves, for which the highest values of the corrosion current were recorded for the initial billet and the samples extruded to Ø8 mm (R₁ 5:1) (Figure 7b). Both materials underwent active dissolution. Extrusion at higher deformation ratios (R₂ 7:1 and R₃ 10:1) changed the corrosion mechanism to a localized one. Wide plateau regions were visible on the anodic branches of the polarization curves for the samples with dimensions of Ø6 mm (R₂ 7:1) and Ø4 mm (R₃ 10:1). The higher value of ΔE indicated a higher resistance to localized corrosion of the sample extruded to Ø6 mm (R₂ 7:1) (

Table 2. The electrochemical parameters calculated from the potentiodynamic polarization curves recorded in 0.01 M NaCl (E_corr—corrosion potential, E_b—breakdown potential).

Material	Ecorr (V/Ref)	Eb (V/Ref)	ΔE
Initial billet Ø40 mm	−1.47	n/a	n/a
Ø8 mm	−1.44	n/a	n/a
Ø6 mm	−1.5	−0.76	0.74
Ø4 mm	−1.5	−0.88	0.62

).

The Nyquist plots of the investigated alloys obtained from the EIS measurements are shown in Figure 8. In order to analyze the corrosion mechanisms of the above alloys, the corresponding equivalent circuits and the fitting data of the EIS spectra are presented in

Table 3. Equivalent electronic circuits (EECs) used for EIS data fitting.

(a)		(b)	(c)
Sample/immersion time
Ø40 mm/1 h	Ø40 mm/6 h	Ø40 mm/12 h	Ø8 mm/6 h
Ø8 mm/1 h		Ø6 mm/1 h	Ø8 mm/12 h
Ø6 mm/6 h	Ø6 mm/12 h	Ø4 mm/1 h
Ø4 mm/6 h	Ø4 mm/12 h

and

Table 4. Electrochemical parameters fitted for EIS results with the EECs shown in Table 3.

Sample	Immersion Time	Rs (Ω∙cm2)	Rct (Ω∙cm2)	CPEct (µSsa/cm2)	a	RL (Ω∙cm2)	L (H∙cm2)	Rf (Ω∙cm2)	CPEf (µSsa/cm2)	a2
Ø40 mm	1 h	189	380	0.003	0.83	2115	7	1906	0.00002	0.86
Ø40 mm	6 h	189	320	0.004	0.88	841	8	2085	0.0004	0.82
Ø40 mm	12 h	168	340	0.000205	0.73	1388	3130	35	0.000016	0.60
Ø8 mm	1 h	232	937	0.000202	0.75	17,520	8	2114	0.000019	0.85
Ø8 mm	6 h	247	1557	0.000054	0.73	3300	247	n/a	n/a	n/a
Ø8 mm	12 h	245	1057	0.000071	0.86	3390	52,000	n/a	n/a	n/a
Ø6 mm	1 h	506	5508	0.000004	0.89	35,000	22	5318	0.000006	0.91
Ø6 mm	6 h	569	3635	0.000006	0.81	9771	22	8099	0.000007	0.87
Ø6 mm	12 h	632	4600	0.000005	0.75	6703	20	9359	0.000006	0.88
Ø4 mm	1 h	451	2929	0.000006	0.85	8959	868,000	5425	0.000006	0.88
Ø4 mm	6 h	458	3638	0.000521	0.90	3716	12	7722	0.000006	0.86
Ø4 mm	12 h	462	3521	0.000613	0.83	4840	26	8142	0.000006	0.87

.

R_s stands for the solution resistance, R_ct represents the charge transfer resistance, and R_f is the film resistance. Higher values of R_ct and R_f indicate a better corrosion resistance of the corresponding alloy. Constant phase elements (CPEs) are used instead of an ideal capacitor to compensate for non-homogeneity in the corrosion system. Importantly, R_L (resistance) and L (inductance) describe the low-frequency inductance loop, indicating that localized corrosion has been initiated. Among the investigated materials, the highest values of resistance to corrosion were observed for the materials KoBo-extruded to Ø6 mm (R₂ 7:1) and to Ø4 mm (R₃ 10:1) and were confirmed by the highest R_ct and R_f values.

The shape of the Nyquist plot is typical for materials with an oxide layer formed on the surface. It is worth noting that in the case of the sample extruded to Ø6 mm (R₂ 7:1), the corrosion resistance after 12 h was still increasing, while in the case of the sample extruded to Ø4 mm, it remained at the same level after 6 h of immersion. This is related to the passivation mechanisms occurring on the surfaces (Table 3 and Table 4). The active performance of the initial billet and the sample deformed to Ø8 mm (R₁ 5:1) caused a weakening in their corrosion resistance with extended immersion time (Figure 8a,b). An increasing trend in the corrosion resistance with extended immersion time was shown for the samples extruded to Ø6 mm (R₂ 7:1) and Ø4 mm (R₃ 10:1) (Figure 8c,d). As shown by King et al. [32], circuit simplification enables the calculation of polarization resistance (R_p). Given that the corrosion rate is inversely proportional to R_p, the dependence of corrosion resistance on immersion time for the samples was calculated, and the results are shown in Figure 9. This is clear confirmation that KoBo enhances the corrosion resistance of pure Mg, with a slight increase for the sample with a diameter of Ø8 mm (R₂ 5:1) and a significant improvement for the material deformed to Ø6 mm (R₂ 7:1), though a slightly higher corrosion rate was noted for the Ø4 mm (R₃ 10:1) sample. Clearly, the lowest corrosion rate, regardless of the immersion time, was calculated for the Ø6 mm (R₂ 7:1) sample.

Damage in the form of filiform corrosion with deep corrosion cavities was observed on the surface of the initial billet (Figure 10a). Corrosion with threads developing on the Mg surface was formed on the sample extruded to Ø8 mm (R₂ 5:1) (Figure 10b). The same type of corrosion occurred on the samples extruded at the highest extrusion ratios (R₂ 7:1 and R₃ 10:1, Figure 10c,d, respectively). The observed corrosion damage was not as deep as that observed for the initial billet. It should also be noted that the threads observed on the sample extruded to Ø8 mm had a different shape to those formed on the samples extruded to Ø6 mm and Ø4 mm. In the latter ones, corrosion damage propagated along privileged crystallographic planes; this type of corrosion is known as preferred crystallographic pitting [33,34].

4. Discussion

The corrosion of Mg is a complex process and depends on the specific thermomechanical fabrication route and the resulting microstructure formed. During the last decade, researchers have tried to characterize the corrosion rate of Mg and its alloys after plastic deformation; however, their findings did not lead to a straightforward explanation concerning which microstructural components are the key factors in the mechanisms controlling their corrosion performance. The texture, dislocation density, grain size, grain distribution, and morphology of second phases are among the relevant factors; the predominance of one over another may significantly change the corrosion mechanism and the resulting corrosion rate of the materials. To simplify the complications concerning the ambiguous (cathodic/anodic) role of second phases in the corrosion of Mg alloys [35], pure Mg was investigated after KoBo extrusion. Since a similar texture intensity was observed in all the extruded samples, we can exclude this factor from further discussion. The remaining components worth consideration are the grain size, distribution, and crystallographic orientation, along with dislocation densities.

Generally speaking, in chloride-containing solution, several SPD methods improve the corrosion resistance of pure Mg via grain size reduction. Equal channel angular extrusion (ECAE) performed on pure Mg at 200 °C reduced the grain size to ~9 µm [36], which improved the corrosion performance of the material. This result was explained by the high dislocation density, which forms more energetic crystalline defects, leading to the easier formation of a passive film on the surface. Birbilis et al. [37] showed that after equal channel angular pressing (ECAP) at 250 °C of pure Mg, the resultant more-refined grain size with a high misorientation angle promoted the formation of a more stable oxide film. Similar conditions of ECAP for pure Mg were used by Op’t Hoog et al. [38], and the results showed lowered cathodic and anodic kinetics when compared to cast material. The same pure Mg samples with subsequent annealing were described in their other study [39], where the authors stated that not only does grain size reduction play a predominant role in corrosion resistance improvement, but the surface free energy and residual stress in a material are also critical. The various factors influencing pure Mg corrosion are not solely those related to the grain size and its characteristics (such as grain boundaries and crystallographic orientation) but also include the dislocation densities and strain imposed during fabrication, which seem to be especially crucial during KoBo extrusion. The dislocation density, besides high-angle grain boundaries, is of crucial importance because it can change the surface potential [40]. Everything, therefore, leads to the consideration of recrystallization intensity, which, during KoBo, depends on the extrusion ratio, thus resulting in the formation of various microstructures. Without knowledge concerning how various microstructural components formed during KoBo extrusion affect the corrosion behavior of pure Mg, we are not able to predict the corrosion performance of other Mg-based alloys fabricated using the same method. KoBo extrusion enhances the corrosion resistance of pure Mg by microstructural refinement; however, there is no clear relationship between the grain size and corrosion behavior. The corrosion resistance improvement is strongly dependent on the influence of individual microstructural components (grain size vs. dislocation density). Also, the decreasing grain size is not itself the main factor that may improve the corrosion resistance of pure Mg extruded using KoBo. During KoBo, recrystallization processes are affected by the massive plastic strain imposed during deformation. Therefore, the respective dominance of DRX vs. DRV at various extrusion ratios results in there being no linear change in the grain refinement with increasing extrusion ratio. As a result of the extrusion, the grain sizes were reduced to 2.88 µm, 3.88 µm, and 1.20 µm in the case of materials extruded to Ø8 mm (R₁ 5:1), Ø6 mm (R₂ 7:1), and Ø4 mm (R₃ 10:1), respectively. Therefore, considering the grain size itself, the corrosion resistance should follow the order Ø4 mm (R₃ 10:1) > Ø8 mm (R₁ 5:1) > Ø6 mm (R₂ 7:1). Although the smallest grain size should lead to increased passivation behavior [40], in our case, a mismatch was caused by the high dislocation density observed in the samples extruded at R₁ 5:1 and R₃ 10:1. This, in turn, provides an explanation for the greater grain size observed after deformation at R₂ 7:1, where, during DRV, the destruction of dislocations was observed, resulting in greater grain size formation and leading to better corrosion resistance of the corresponding alloy. This observation is opposite to the previous studies, where, with increased intensity of the plastic deformation, the dislocation intensity also increased; however, this was observed for a Mg alloy wherein second phase strengthening was also observed [41,42].

5. Conclusions

An attempt was made to describe the main factors influencing the corrosion resistance of KoBo-extruded pure Mg. We conclude the following:

• The KoBo method enables significant microstructural refinement for pure Mg without preheating of the initial billet, having a positive influence on its corrosion resistance.
• During KoBo deformation, there is no clear tendency for grain size reduction with increasing extrusion ratios. The grain size decreased from coarseness typical of cast material to 2.88 µm, 3.88 µm, and 1.20 µm for R₁ 5:1, R₂ 7:1, and R₃ 10:1.
• The corrosion resistance of the KoBo-extruded samples is related not solely to grain refinement but also to other microstructural factors, particularly the dislocation density accumulated during deformation.