Effect of Impurity Fe Concentration on the Corrosion Behaviorof Mg-14mass%Li-1mass%Al Alloy

Taiki Morishige1,+, Kumiko Ueno1, Masahiro Okano2, Takayuki Goto3,Eiji Nakamura3 and Toshihide Takenaka1

1Department of Chemistry and Materials Engineering, Kansai University, Suita 564-8680, Japan2Graduate School of Science and Engineering, Kansai University, Suita 564-8680, Japan3Santoku Corporation, Kobe 658-0013, Japan

Mg­Li system alloys with over 11mass% Li have an excellent cold-workability in Mg alloys due to the bcc-structured crystal system.However this system alloys have poor corrosion resistance caused by Li solid-solution alloy. The corrosion resistance of conventional hcp-Mgalloys is influenced by the composition of impurity elements. Particularly, minute iron element deteriorates the corrosion resistance of Mg alloyand there is a threshold concentration on the corrosion rate in Mg alloys. In this study, the effect of iron element on the corrosion resistance ofMg­Li alloy was investigated by addition of iron. The corrosion resistance of Mg­Li with no Fe addition was higher than conventional Mg alloy.The corrosion rate of Mg­Li alloy increased with increasing iron concentration. However, there is no threshold iron concentration affected to thecorrosion resistance on Mg­Li alloy. [doi:10.2320/matertrans.MAW201405]

(Received April 18, 2014; Accepted May 21, 2014; Published July 11, 2014)

Keywords: magnesium­lithium alloy, corrosion rate, impurity, exfoliation corrosion

1. Introduction

Mg alloy is subject to increasing the demand because Mgalloy is the lightest metal material for structural use with highspecific strength compared with other ferrous and non-ferrousalloys. Therefore, Mg alloys are used for aerospace andtransportation components to reduce the weight and the fuelconsumption. However, the workability of this alloy is notgood due to hexagonal closed packed crystal structure. Mgalloys are generally used as cast (including die-cast) and hot/warm formed products. In the 1960s, Mg­Li system alloyssuch as LA141 (Mg-14mass% Li-1mass% Al) have beendeveloped to improve the cold workability of Mg alloy.1) Thecrystal structure of Mg­Li alloys with over 11mass% Libecomes body-centered cubic (bcc) structure that is Li solidsolution alloy.2) As a result, Mg­Li alloys are able to severeplastic deformation at room temperature. On the other hand,Mg­Li alloys have poor corrosion resistance because Lielement has the lowest standard electrode potential in all-metalelements. Especially, cold-worked Mg­Li alloy bring theexfoliation corrosion that fractured along the rolling plane asthe progression of corrosion in the longitudinal section and thetransverse section of the cold-rolled sheet. The corrosionresistance of Mg alloys depends on the impurity traceelements with high standard electrode potential such as iron,nickel and copper.3) It causes the galvanic corrosion betweenLi solid solution matrix and precipitates. These contamina-tions of trace elements in commercial Mg metal and alloys arestrictly regulated by ASTM standards. For example, ironelement is restricted to less than 50 ppm of the compositionof Mg alloy. Further contamination of iron causes thedegradation of corrosion resistance, and more than 160 ppmof iron elements in Mg alloy severely deteriorates of thecorrosion rate.4) However, there is not reported about the effectof these trace elements on the corrosion behavior of Mg­Lialloy because previous reports are only about hcp-structured

Mg alloys. In this study, the corrosion resistance of bcc-structured Mg­Li alloy contained trace iron element and theeffect of the concentration of iron element were investigated.

2. Experimental Procedures

LA141 (major composition is Mg-14mass% Li-1mass%Al) alloy with additional trace iron was prepared by remeltingalloy and immersion of 10 © 30 © 0.5mm3 pure iron plateinto 30 g of LA141 alloy melt. Remelting Mg­Li alloy wasconducted at 973K in Ar atmosphere. The iron concentrationin LA141 alloy was controlled by varying the immersion timefrom 3.6 to 21.6 ks. The composition of as-received LA141alloy was shown in Table 1. These trace iron concentration inLA141 alloy was measured by atomic absorption spectrosco-py (AAS) using Hitachi high-technologies Z-2710 polarizedZeeman atomic absorption spectrophotometer. These ingotswere homogenized at 673K for 3.6 ks in Ar atmosphere.These alloys were cut out to the size of 10 © 10 © 50mm3

and were cold-rolled to 87.5% of the rolling reduction. Thecold-rolled specimens were annealed at 373K for 3.6 ks forthe purpose of elimination of residual strain induced by cold-rolling. The corrosion behaviors of these alloys with variousconditions were measured. In as-rolled and annealed speci-mens, corrosion behaviors were investigated on the longi-tudinal section to the rolling direction. The corrosion rate ofthese alloys were measured by potentiodynamic measurement(anodic polarization) using the electrochemical measuringsystem composed of HOKUTO denko HA-501G potentio/galvano stat and HB-111 function generator. The long-termimmersion test was performed in 5mass% NaCl aqueoussolution. The corroded surface was observed by opticalmicroscopy using Nikon Eclipse MA-100 optical microscope.

3. Results and Discussion

As a result of chemical analysis by AAS, the ironconcentrations of each specimen were 29 ppm (as-received)+Corresponding author, E-mail: tmorishi@kansai-u.ac.jp

Materials Transactions, Vol. 55, No. 9 (2014) pp. 1506 to 1509©2014 The Japan Institute of Metals and Materials RAPID PUBLICATION

and 152, 186 and 229 ppm (additional iron), respectively. Allspecimens consisted of bcc-structured Li solid solution singlephase. There were no visible precipitates from the result ofmicrostructural observations. Figure 1 shows the macro-scopic images of homogenized specimens after immersiontest for 14.4 ks. The filiform corrosion progressed in alloverthe corrosion surface. As increasing iron concentration of thespecimen, it increased the area of filiform corrosion and thewhole surface was covered with white corrosion productswhen the iron concentration increased to 229 ppm. In thisstudy, specimens could not be contaminated by otherelements such as copper and nickel because addition ofimpurity was using only pure iron. The pitting corrosion wasalso observed in the sample with high iron concentration andthe number of pitting site increased with increasing ironconcentration. The relationship between the corrosion currentdensity and the corrosion potential of homogenized specimenand the iron content was shown in Fig. 2. There was nothreshold content of iron element on corrosion sensitivity inthis experimental composition range. The corrosion potentialwas less noble with increasing the iron concentration. Also,the corrosion current density linearly increased with increas-ing iron content. The corrosion current density of other hcp-Mg metal and alloys were listed in Table 2.5­10) Thecorrosion current density of this alloy was equivalent to thatof other hcp-Mg alloys. Assuming that the major alloyingelements (Mg, Li and Al elements) were soluble inaccordance with alloy composition to NaCl aqueous solution,the corrosion rate, equivalent to the weight loss rate, wascalculated from the value of the corrosion current densityusing following equation:11)

ðcorrosion rate of LA141 alloyÞ¼ ðicorr=96500Þ � ð0:63� 1=2� 24:31þ 0:36� 6:94

þ 0:0067� 1=3� 26:98Þ � 10�1

when the major alloy components were soluble as followingelectrochemical reaction:

Mg ¼ Mg2þ þ 2e�

Li ¼ Liþ þ e�

Al ¼ Al3þ þ 3e�

where the calculated corrosion rate is expressed in units ofweight per unit area per unit time [mg cm¹2 s¹1] and thecorrosion current density, icorr, was measured in units ofAm¹2. The calculated corrosion rate of Mg-14mass% Li-1mass% Al alloy with 29 ppm iron impurity was 1.52 © 10¹5

mg cm¹2 s¹1, this value was approximately equivalent tomeasured corrosion rate of other hcp-Mg metal and alloyswith similar level of impurities as shown in Table 2. Thecalculate corrosion rate from the value of icorr has a goodagreement with experimental value of weight loss rate.8)

There was few report of the iron solubility of Mg­Li alloys orLi solid solution alloys. It is not clear the existence state oftrace iron whether solute to the matrix or fine precipitates.The threshold concentration of iron element to which thecorrosion rate rapidly increases may exist in the lower region.

The relationship between the corrosion rate and ironconcentration was shown in Fig. 3. The corrosion rate of

(a)

(b)

(c)

(d)

Fig. 1 Macroscopic images of homogenized LA141 alloys after 14.4 ksimmersion in 5mass% NaCl aqueous solution; (a) 29 ppm Fe, (b) 152ppm Fe, (c) 186 ppm Fe and (d) 229 ppm Fe.

0 100 200 300-1.85

-1.80

-1.75

-1.70

0

1

2

3

4

5

Corrosion potentialCorrosion current density

Cor

rosi

on p

oten

tial,

Eco

rr /V

vs.

Ag/

Ag

+

Iron concentration (ppm)

Cor

rosi

on c

urre

nt d

ensi

ty, i

corr /A

m-2

Fig. 2 The relationship between iron concentrations and the electro-chemical corrosion properties of homogenized LA141 alloys.

Table 1 Chemical compositions of as-received LA141 alloy. (mass%)

Li Al Ca Zn Si Fe Cu Ni Mn Mg

as-received 13.74 1.09 0.29 0.004 0.024 0.003 <0.001 0.001 0.038 bal.

0 100 200 3000

2

4

6C

alcu

late

d co

rros

ion

rate

/mg

cm-2

s-1

Iron concentration (ppm)

homogenizedas-rolledannealed

[x 10-5]

Fig. 3 The relationship between iron concentrations and the calculatedcorrosion rate of LA141 alloys.

Effect of Impurity Fe Concentration on the Corrosion Behavior of Mg-14mass%Li-1mass%Al Alloy 1507

Table 2 Corrosion properties of Mg-alloys.4­8)

Alloy Corrosion Solution Corrosion Current Density, icorr/Am¹2 Corrosion Rate (Weight loss)/mgcm¹2 s¹1 Ref.

Ultra-High purity Mg 3.5mass% NaCl 2.22 © 10¹2­2.24 N/A 5)

High purity Mg 3.5mass% NaCl 10­13 N/A 6)

AM20 5mass% NaCl N/A 7.8 © 10¹5 7)

AM60 1.7 © 10¹5­2.3 © 10¹5

AS21 9.3 © 10¹5

AS41 2.1 © 10¹5­2.8 © 10¹5

AZ51 3.5mass% NaCl 6.22 3.8 © 10¹4­6.4 © 10¹4 8)TAZ151(Mg-1Sn-5Al-1Zn)

8.37 3.4 © 10¹4­5.2 ©10¹4

TAZ551(Mg-5Sn-5Al-1Zn)

7.16 2.1 © 10¹4­2.9 © 10¹4

TAZ951(Mg-9Sn-5Al-1Zn)

9.13 2.5 © 10¹4­4.6 © 10¹4

Mg-9Al-0.6Zn-0.2Mn 3.5mass% NaCl 0.11 N/A 9)

Mg-9Al-0.8Zn-0.2Mn-0.14Ca 2.71

Mg-9Al-0.8Zn-0.2Mn-0.4Sb 2.07

Mg-9Al-0.8Zn-0.2Mn-0.4Sb-1Bi 7.62

AM50 with 23 ppm Fe(0.23mass% Mn)

5mass% NaCl N/A 1.25 © 10¹4 10)

AM50 with 57 ppm Fe(0.23mass% Mn)

1.86 © 10¹4

AM60 with 49 ppm Fe(0.21mass% Mn)

1.69 © 10¹4

LA141 with 29 ppm Fe 5mass% NaCl 1.44 1.52 © 10¹5* This work

LA141 with 152 ppm Fe 2.25 2.38 © 10¹5

LA141 with 186 ppm Fe 2.70 2.86 © 10¹5

LA141 with 229 ppm Fe 3.22 3.40 © 10¹5

*Corrosion rate of this work is calculated value.

(a) (b)

(c) (d)

Fig. 4 Microstructures of as-rolled LA141 alloys after corrosion test for 43.2 ks.

T. Morishige et al.1508

cold-rolled and annealed specimens also increased withincreasing the iron content. The corrosion rates with 29 ppmof iron concentration were comparable in all specimens,however cold-rolled and annealed specimens with high ironcontent had high corrosion sensitivity. It can be explained thistrend from the residual strain induced by cold-rolling. Themicrostructure with deformation band and high dislocationdensity was susceptible to the anodic dissolution reaction.12)

The corrosion rate of annealed specimens with high ironcontent was lower than that of as-rolled. In spite of theannealed specimen was relaxed to some extent by micro-structural restoration, the residual strain was not entirelyreleased. Figure 4 shows macroscopic images of as-rolledspecimens after 43.2 ks. The as-rolled specimen with 29 ppmiron was covered with the oxide film on the corrosion surfaceand the oxide film was partly broken in the specimen with152 ppm iron in the same immersion period. As increasesiron concentration, the breaking of oxide film occurred inearlier immersion period. There were microcracks along thegrain boundary in longitudinal and transverse sections ofas-rolled and annealed specimens with 186 and 229 ppm ofiron content. These microcracks lead to origination of theexfoliation corrosion.13) Even though the specimen has only29 ppm iron, microcracks appeared after longer corrosiontest. The exfoliation corrosion was eventually observed in as-rolled and annealed specimens regardless of iron concen-trations in LA141 alloy. The initiation time of the exfoliationcorrosion decreased with increasing the iron concentrationbecause the high iron content alloy become more sensitive tothe corrosion progression.

4. Conclusion

The effect of impurity iron element on the corrosion rate

of Mg-14mass% Li-1mass% Al alloy was investigated.The corrosion rate linearly increased with increasing ironconcentrations. The threshold content of iron was notindicated in this composition range. The absolute value ofcorrosion rate of LA141 alloy with several hundred ppmwas not remarkably high compared to commercial hcp-structured Mg metal and alloys. The cold-rolled LA141alloy has more sensitive to exfoliation corrosion with theincrease of iron concentration, although the sensitivity ofexfoliation corrosion was improved to some extent byannealing due to the residual strain induced by cold-rollingwas relieved.

REFERENCES

1) R. J. Jackson and P. D. Frost: NASA Technology Utilization SP-5068,(1967).

2) T. B. Massalski, J. L. Murray, L. H. Benett and H. Baker: Binary AlloyPhase Diagram, (ASM International, 1990).

3) H. E. Friedrich and B. L. Mordike: Magnesium Technology, (Springer-Verlag, Berlin Heidelberg, 2006).

4) J. D. Hanawalt, C. R. Nelson and J. A. Peloubet: Trans. AIME 147(1942) 273­299.

5) F. Cao, Z. Shi, J. Hofstetter, P. J. Uggowitzer, G. Song, M. Liu and A.Atrens: Corros. Sci. 75 (2013) 78­99.

6) F. Cao, Z. Shi, G. L. Song, M. Liu and A. Atrens: Corros. Sci. 76(2013) 60­97.

7) K. Nisancioglu, O. Lunder and S. Ono: J. Surf. Finish. Soc. Jpn. 44(1993) 883­889 (in Japanese).

8) K. C. Park, B. H. Kim, H. Kimura, Y. H. Park and I. M. Park: Mater.Trans. 51 (2010) 472­476.

9) Z. Shi, M. Liu and A. Atrens: Corros. Sci. 52 (2010) 579­588.10) H. Matsubara, Y. Ichige, K. Fujita, H. Nishiyama and K. Hodouchi:

Corros. Sci. 66 (2013) 203­210.11) E. McCafferty: Introduction to Corrosion Science, (Springer, 2010).12) T. Ohnishi and Y. Nakatani: J. JILM 28 (1978) 123­129 (in Japanese).13) T. Morishige, H. Doi, T. Goto, E. Nakamura and T. Takenaka: Mater.

Trans. 54 (2013) 1863­1866.

Effect of Impurity Fe Concentration on the Corrosion Behavior of Mg-14mass%Li-1mass%Al Alloy 1509