Effect of rolling ratios on the microstructural evolution and corrosion performance of an as-rolled Mg-8wt.%Li alloy

B.J. Wng, D.K. Xu, X. Ci, Y.X. Qio,*, L.Y. Sheng

aSchool of Environmental and Chemical Engineering, Shenyang Ligong University, Shenyang, 110159, China

b Key Laboratory of Nuclear Materials and Safety Assessment, Institute of Metal Research, Chinese Academy of Sciences, Shenyang, 110016, China

c School of Materials Science and Engineering, Jiangsu University of Science and Technology, Zhenjiang, China

d Peking University, Shenzhen Institute, Shenzhen Key Lab Human Tissue Regenerate & Repair, Shenzhen 518057, China

Received 14 January 2020; received in revised form 14 February 2020; accepted 14 February 2020

Available online 2 October 2020

Abstract

Keywords: Mg-li alloys; Wrought; Microstructure; Corrosion performance.

1. Introduction

Due to Mg-Li alloys having the high specifi strength and stiffness, ultra-lightweight and good damping resistance, they could have wide application prospects in the advanced fabrication industries [1-7]. Since element Li has a low density of 0.534 g·cm-2, the specifi weight of Mg-Li alloys will be further decreased when compared with traditional Mg alloys[4].Meanwhile,the addition of lithium could be beneficia for the enhancement in plasticity of Mg-Li based alloys [8-10].

Compared with the traditional Mg alloys[11-16],the crystallographic structure of composed phases in Mg-Li system alloys is unique and closely dependent on the amount of added Li element [17,18]. At room temperature, Mg-Li based alloys having the content of Li less than 5.5 wt% are consisted of the hexagonal close-packed (HCP) structured α-Mg phases, whereas the alloys having the content of Li higher than 10.3 wt% are composed of the body-centered cubic(BCC) structured β-Li phases. When the content of Li in Mg-Li alloys varies from 5.5 wt% to 10.3 wt%, they will have a (α-Mg+β-Li) duplex phase structure. Previous work demonstrated that the formation of β-Li phases in Mg-Li alloys can significantl improve their plasticity [19-21]. Due to the high chemical activity, element Li can readily react with air, aqueous solutions and copious hydrogen gas [22].Thus, Mg-Li alloys have the relatively poor corrosion performance. Recently, researchers found that the alloying [23-25],severe plastic deformation [26,27] and coating [28-30] could enhance the corrosion performance of Mg-Li based alloys.Among them, the hot-rolling is a traditional processing technology for metallic materials, which can optimize the microstructure [31,32], refin grain structure and then improve their mechanical strength and ductility [33-35]. Additionally,the hot rolling process can be beneficia for enhancing the corrosion performance of Mg-Li based alloys [36-41]. Since β-Li phases have higher deformation capability than α-Mg phases in the dual phase structured Mg-Li based alloys [42-45], their morphologies and spatial distribution of two types of phases will undergo different changes during the rolling process. Thus, the influenc of these microstructural changes on the corrosion performance of dual phase structured Mg-Li based alloys should be considered. However, no relevant research work has been done and the underlying mechanism is still unknown.

In this work, the target is to disclose the effects of microstructural changes especially the morphologies and spatial distribution of α-Mg and β-Li phases on the corrosion resistance by exploring the corrosion behaviors of an as-rolled Mg-8wt.%Li alloy with different rolling ratios. Moreover, the underneath mechanisms for the corrosion attacks have been explained.

2. Experimental procedures

The investigated material in this work is a dual phase structured Mg-8wt.%Li alloy, which was prepared under the protection of the flu argon by melting high-purity Mg and Li in vacuum reaction furnace. The ingot was cut into thick plates with a rectangular cross-section of 150mm × 150mm and a thickness of 15mm. After being homogenized at 573K for 30 min, the plates were respectively rolled with the rolling ratios of 3, 5 and 10 at room temperature. Then, samples were cut from the rolled plates and ground on SiC papers.Finally, sample surfaces were polished up to their roughness of 1μm using ethanol. For the phase analysis, a D/Max 2400 X-ray diffractometer (XRD) of monochromatic Cu Kα radiation with a wavelength of 0.154056nm were used. The employed step size is 0.02° and the scanning rate is 4°/min.For the microstructural observation using optical microscope(OM: VHX-900F), an enchant consisting of 4% HNO3and 96% C2H5OH was used for etching the polished surfaces.The area fractions of phases were calculated using software of image J.

For weight loss testing in 0.1mol/L NaCl solution,samples with a cross-section of 10mm(RD:rolling direction)×10mm(TD: transverse direction) and a thickness of 1.5mm were immersed for up to 48h at room temperature. Before and after immersion, the weight of samples was recorded using electronic balance scale (BSA224S) with the accuracy of 0.1mg. To determine the variations of measured data, three parallel samples were used for each condition. After testing,corrosion products were removed by immersed in a 180g/L CrO3chromic acid solution [46-51]. Then, OM observation performed to the surface morphologies. For hydrogen evolution testing, sample surfaces with a dimension of 10mm(RD)×10mm (TD) were exposed in 0.1mol/L NaCl solution for up to 48h at room temperature.The ratio between the volume of solution (ml) and exposed sample surface area (cm2)was 40. During the immersion process, no vibration or deaerating was allowed. By using a digital pH meter (PHS-25),the pH value of the solution was monitored to be 6.3±0.2.Electrochemical measurement was carried out solution using CS350H (Corrtest Co. Ltd, Wuhan) electrochemical workstation at room temperature. A classical three electrode cell consisting of the counter electrode of Pt, reference electrode of saturated calomel electrode (SCE) and the working electrode of sample surfaces with an exposed area of 10mm(RD)×10mm (TD) was used. At a scan rate of 1 mV·s-1,the potentiodynamic polarization curves were measured.Electrochemical impedance spectroscopy (EIS) spectra were measured in a frequency range varying from 100kHz to 10 mHz with a disturbed potential amplitude of 10mV and then fitte by using the ZSimDemo 3.30 software.For each electrochemical measurement, at least three times were performed. Combining with the scanning electron microscope (SEM; XL30-FEG-ESEM) and stereo optical microscope (Keyence VHX 2000), the corrosion attack and three-dimensional (3D) morphologies on the surfaces of samples being immersed respectively for 4h, 8h and 48h were observed to disclose the underneath corrosion mechanisms.

Fig. 1. XRD patterns of the as-rolled Mg-8wt.%Li alloy.

3. Results

3.1. Microstructural characterization

Fig.1 shows the XRD pattern of the as-rolled Mg-8wt.%Li alloy. It reveals that only α-Mg and β-Li phases are present in the alloy. Fig. 2 shows the microstructure of the as-rolled Mg-8wt.%Li alloy with different rolling ratios. Based on previous work [52], α-Mg and β-Li phases appear in white and black in the OM images, respectively. It indicates that with the increase of rolling ratio, the α-Mg phases were elongated along the RD (Fig. 2(b) and Fig. 2(c)) and the average thickness values of the elongated α-Mg phases in the samples with rolling ratio of 3, 5 and 10 are 50, 30 and 20μm respectively,resulting in the increase of the area fraction of exposed β-Li phases. The area fractions of β-Li phases in the alloy with rolling ratio of 3, 5 and 10 were measured to be 56%, 60%and 75%, respectively.

3.2. Hydrogen evolution

Fig. 2. Optical microstructure of the as-rolled Mg-8wt.%Li alloy with the rolling ratios of: (a) 3, (b) 5 and (c) 10, respectively.

Fig.3. Hydrogen evolution curves of the differently rolled Mg-8wt.%Li alloy plates measured in 0.1mol/L NaCl solution.

Fig. 3 shows the hydrogen evolution curves of the samples with different rolling ratios. It reveals that when the immersion time is shorter than 6h, the hydrogen evolution rates of samples increases gradually with the rolling ratios. When the immersion time is longer than 6h, the hydrogen evolution rates of the sample with rolling ratio of 5 is the highest.However,the hydrogen evolution rates of samples with rolling ratios of 3 and 10 are basically the same.

3.3. Weight loss

Fig. 4. Weight loss versus time curves of the differently rolled Mg-8wt.%Li alloy plates measured in 0.1mol/L NaCl solution.

Table 1 Fitting results from the polarization curves of differently rolled Mg-8wt.%Li alloy plates measured in 0.1mol/L NaCl solution.

Fig. 4 shows the weight loss versus time curves of the differently rolled Mg-8wt.%Li alloy plates. The measured mass loss rates of the samples with rolling ratios of 3, 5 and 10 are 1.15, 1.68 and 1.25 mg·cm-2·day-1, respectively. Fig. 5 shows the surface morphologies of different samples after immersion for 48h in 0.1mol/L NaCl solution. Compared with the other two conditions, the pits on the surface of the sample with rolling ratio of 5 is much deeper and larger. Based on the 3D measurements (Fig. 5(d), (e) and (f)), the pits on surfaces of the samples with rolling ratios of 3, 5 and 10 can reach up to 230, 330 and 250μm, respectively.

3.4. Electrochemical measurement

Potentiodynamic polarization curves of the as-rolled Mg-8wt.%Li alloy with different rolling ratios is shown in Fig. 6.It reveals that the anodic and cathodic branches are asymmetrical and the increased degree in current density of the anodic branch is much larger than that of the cathodic branch. In potentiodynamic polarization curves of Mg alloys, cathodic branches are associated with the hydrogen evolution, whilst anodic branches correspond to the Mg dissolution [53]. Since the negative difference effect and pitting corrosion can easily occur during the measurements,the anodic branch is seldomly used for fittin analysis [54,55]. For compare conveniently,the determined corrosion potential (Ecorr) and the fitte corrosion current density (icorr) are listed in Table 1. It can be seen that the corrosion resistance of the alloy increases with the rolling ratio, which is in agreement with the tendency of the initial stage in hydrogen evolution curves. The determined icorrvalues of the samples with rolling ratios of 3, 5 and 10 are 5.63, 19.2 and 63.6 μ/cm2, respectively.

Fig. 5. Surface morphologies of the as-rolled Mg-8wt.%Li alloy with the rolling ratios of: a) 3, b) 5 and c) 10 after immersion for 48h in 0.1mol/L NaCl solution. Images (d), (e) and (f) are the 3D profile of the squared areas in images (a), (b) and (c), respectively.

Fig. 6. Potentiodynamic polarization curves of the differently rolled Mg-8wt.%Li alloy plates measured in 0.1mol/L NaCl solution.

Fig. 7. Nyquist plots of the differently rolled Mg-8wt.%Li alloy plates measured in 0.1mol/L NaCl solution. The equivalent circuit is inserted.

Table 2 The fitte EIS data on the basis of the equivalent circuits.

EIS curves of the differently rolled Mg-8wt.%Li alloy plates are shown in Fig.7.For all three conditions,the spectra are composed of a high frequency capacitive loop, a mediumto-low frequency capacitive loop and a low frequency inductive loop. Previous work demonstrated that the presence of the low frequency inductive loop was ascribed to the initiation stage of localized corrosion [46,56]. To further clarify the corrosion characteristics of different samples, an equivalent circuit was proposed and used for fittin the EIS curves and the fitte data are summarized in Table 2. In the equivalent circuit, it consists of solution resistance of Rs, charge transfer resistance of Rctand electric double layer at the substrate/electrolyte interface of Qdlin the high frequency capacitance loop, fil resistance of Rfand capacity of Qfin the medium frequency capacitance loop (define as Yfand nf),resistance of RLand inductance of L in low frequency inductance loop [46,47]. Since an ideal capacitor is not suitable for accounting for the non-homogeneity in the system, a constant phase element of Qdldetermined by two values of Ydland ndlis used. Meanwhile, ndlis designated as the dispersion coeffi cient of Qdl. When ndlvalue is 1, Qdlcan be considered as a capacitor; when ndlvalue is zero, Qdlcan act as a resistance.The determined Rctvalues of samples with rolling ratios of 3, 5 and 10 are 1.18×103, 1.05×103and 0.81×103Ω·cm2,respectively.

3.5. Corrosion morphology

Corrosion morphologies of differently rolled samples after immersion for 4h in 0.1mol/L NaCl solution are shown in Fig. 8. It reveals that for all samples, the typical“filiform-li e” corrosion morphologies can be seen on the surfaces, which is related to the scenario whereby locally anodic and cathodic sites evolve dynamically [57-60]. Highmagnificatio images showed that the corrosion attack preferentially occurred in β-Li phases and the α-Mg/β-Li interfaces can suppress the development of “filiform-li e” corrosion (Fig. 8(b), (e) and (h)). Based on the 3D observations to the severely corroded areas on sample surfaces (Fig. 8(c), (f)and (i)), it indicates that the maximum pitting depths on the surfaces of samples with rolling ratios of 3, 5 and 10 were respectively measured to be 12, 15 and 20μm. Since the corrosion potential of α-Mg phase(-1.75 VSCE)is much positive than that of β-Li phase (-2.38 VSCE) in the dual phase structured Mg-Li alloys [61], β-Li phases act as anodic sites and will be preferentially corroded during immersion, whereas the α-Mg phases are protected.

Corrosion morphologies of differently rolled samples after immersion for 8h in 0.1mol/L NaCl solution are shown in Fig. 9. High-magnificatio images demonstrate that the corrosion attack can also occur in the α-Mg phases (Fig. 9(b),(e) and (h)). Moreover, compared with other two conditions,the corrosion attack of the sample with the rolling ratio of 5 is much worse. Based on the 3D observations to the severely corroded areas on sample surfaces (Fig. 9(c), (f) and (i)), it indicates that the maximum depths of corrosion pits of the samples with rolling ratios of 3, 5 and 10 are respectively measured to be 15, 40 and 25μm, which further proves that when the immersion time exceeds 6h, the corrosion performance of the sample having a rolling ratio of 5 is the lowest.

4. Discussion

It has been reported that the addition of element Li is harmful to the corrosion resistance of Mg-Li alloys[3,21].For the dual phase structured Mg-Li based alloys, the difference in potentials of α-Mg and β-Li phases can form corrosion couples [46], resulting in the preferentially anodic dissolution of β-Li phases during immersion [22]. Since the elongated α-Mg phases increase the amount of the formed α-Mg/β-Li interfaces, more micro-galvanic cells present in the sample with the rolling ratio of 10. Therefore, in the initial immersion stage (≤6h), the corrosion performance of the alloy decreases with the increase of the rolling ratio. However, when the samples were immersed for longer than 6h, the corrosion resistance reflecte by the hydrogen evolution (Fig. 3)and weight loss (Fig. 4) demonstrates that the sample with the rolling ratio of 5 has the lowest corrosion resistance. To have a deep understanding about this phenomenon, it will be discussed as follows.

Fig. 8. Corrosion morphologies of the differently rolled Mg-8wt.%Li alloy plates after immersion for 4h in 0.1mol/L NaCl solution. Images (a), (c) and (e)are respectively the SEM observations to the surface morphologies of samples with rolling ratios of 3, 5 and 10. Images (b), (d) and (f) are 3D profile of surface morphologies of samples with rolling ratios of 3, 5 and 10.

Recently, Morishige et al. reported that the typical“filiform-li e” corrosion could present in β-Li phases and the corrosion attack can continuously proceed during the subsequent corrosion process because the front area is acidic,whilst the corrosion products are mainly deposited at the back [62].Song et al. reported that the formed surface product fil of a duplex structured Mg-8Li (in wt%) alloy had a multi-layer structure consisting of MgO, Mg(OH)2, Li2O and LiOH, and the outer layer was mainly composed of lithium oxide [63].Generally, Li2CO3can be formed due to the reaction between the lithium oxide in outer layer and CO2in aqueous environments or the atmosphere,i.e.Li2O+CO2=Li2CO3,resulting in the surface fil being composed of Li2CO3, LiOH, Li2O,Mg(OH)2and MgO [64]. Moreover, the standard enthalpies for the formation of Li2CO3,LiOH,Li2O,Mg(OH)2and MgO were determined to be -1215.9, -487.46, -599.48, -924.5 and -1095.8 KJ mol-1, respectively [64]. Since the low enthalpy for the formation of a substance reflect it has the high chemical stability, the Li2CO3is the most stable during the immersion process [65-67]. Moreover, the Pilling-Bedworth Ratio (PBR) is an effective parameter for measuring the fil density of corrosion products. For the Li2CO3, LiOH, Li2O,Mg(OH)2and MgO, their PBR values are 1.35, 1.26, 0.57,1.77 and 0.88,respectively[64].Based on the ranking of PBR values,it demonstrates that Mg(OH)2has excellent density for the protection to the matrix [64,68]. However, in the solution containing with high concentration of Cl-, the Mg(OH)2can be transformed into MgCl2due to its strong solubility [69].Following this, it can be concluded that compared with other corrosion products, the protection capability of Li2CO3to the substrate ranks first Moreover, Mg(OH)2and MgO mainly present on the α-Mg phases and the surface fil formed upon β-Li phases is composed of Li2CO3, LiOH and Li2O [32].When the immersion time is longer than 6h, the accumulated Li2CO3in the fil corrosion products can effectively weaken the further corrosion of β-Li phases and then the corrosion attack is much severe in the α-Mg phases. Since the area fraction of exposed β-Li phases increases with the rolling ratio (Fig. 2), the corrosion performance of the sample having the larger rolling ratio should be higher. However, with the rolling ratio increasing, the amount of α-Mg/β-Li interfaces increases and then results in the formation of higher quantity of micro-galvanic cells. Therefore, the corrosion performance of the differently rolled Mg-8wt.%Li alloy plates is ascribed by the competition between the formation of Li2CO3fil and the micro-galvanic corrosion at α-Mg/β-Li interfaces. Since the area fraction of exposed β-Li phases was only increased 4% when the rolling ratio changes from 3 to 5, the effect of Li2CO3in the surface fil on the corrosion resistance can be omitted. Therefore, the poor corrosion performance of the alloy having a rolling ratio of 5 is mainly controlled by the increased quantity of micro-galvanic cells due to the larger amount of α-Mg/β-Li interfaces. However, for the sample with the rolling ratio of 10, the area fraction of exposed β-Li phases increased remarkably. Then, the surface of α-Mg phases can be partially covered by Li2CO3. This beneficia effect could effectively offset the detrimental effect of microgalvanic cells at the α-Mg/β-Li interfaces.Thus,the corrosion resistance of samples with rolling ratio of 3 and 10 is similar.

Fig. 9. Corrosion morphologies of the differently rolled Mg-8wt.%Li alloy plates after immersion for 8h in 0.1mol/L NaCl solution. Images (a), (c) and (e)are respectively the SEM observations to the surface morphologies of samples with rolling ratios of 3, 5 and 10. Images (b), (d) and (f) are 3D profile of surface morphologies of samples with rolling ratios of 3, 5 and 10.

5. Conclusions

In this work, through comparing the corrosion behavior of the differently rolled Mg-8wt.%Li alloy,the main conclusions can be drawn:

(1) With the increase of rolling ratio,the α-Mg phases were elongated and their thickness became thinner, resulting in the increase of area fraction of exposed β-Li phases.

(2) Since it exists the potential difference between α-Mg and β-Li phases, the formed corrosion couples at the α-Mg/β-Li interfaces accelerate the corrosion attack.

(3) The formation of Li2CO3on the surface of β-Li phases could effectively weaken the corrosion attack of β-Li phases.

(4) The corrosion performance of the as-rolled Mg-8wt.%Li alloy is mainly determined by the competition between the detrimental effect from the corrosion couples at the interfaces of α-Mg/β-Li phases and the beneficia effect due to the presence of Li2CO3on the surface of β-Li phases.

Acknowledgements

This work was supported by the National Key Research and Development Program of China under Grant[Nos. 2017YFB0702001 and 2016YFB0301105], Liaoning Province’s project of “Revitalizing Liaoning Talents”(XLYC1907062), National Natural Science Foundation of China Projects under Grant [Nos. 5207011217, 51871211 and 51701129],the Doctor Startup Fund of Natural Science Foundation Program of Liaoning Province (No. 2019-BS-200), the Strategic New Industry Development Special Foundation of Shenzhen (JCYJ20170306141749970), the funds of International Joint Laboratory for Light Alloys, Liaoning BaiQian-Wan Talents Program, the Domain Foundation of Equipment Advance Research of 13th Five-year Plan (61409220118), the Innovation Fund of Institute of Metal Research (IMR), Chinese Academy of Sciences (CAS), the National Basic Research Program of China (973 Program) project under Grant No. 2013CB632205.