• 沒有找到結果。

中 華 大 學 工程科學研究所 博 士 論 文

N/A
N/A
Protected

Academic year: 2022

Share "中 華 大 學 工程科學研究所 博 士 論 文"

Copied!
139
0
0

加載中.... (立即查看全文)

全文

(1)

中 華 大 學 工程科學研究所 博 士 論 文

鎂鋰鋅合金之顯微結構與機械性質研究 Microstructures and Mechanical Behavior of

Processed Mg-Li-Zn Alloys

學號姓名:D09324004 邱垂泓 指導教授:吳泓瑜 博士

中 華 民 國 九 十 六 年 七 月

(2)

摘 要

本研究以真空感應爐熔煉 37mm 厚的 Mg-9%Li-1%Zn (LZ91)合

金,並以二重軋延機將其輥軋至 2mm,以此板材進行材料特性之分

析研究。此外,亦熔鑄直徑 200mm 之大尺寸鎂鋰鋅合金錠,包括

Mg-6%Li-1%Zn (LZ61)、Mg-9%Li-1%Zn 及 Mg-10%Li-1%Zn (LZ101)

等三種組成,並將圓錠材擠製及冷軋成為 0.6mm 的捲料薄板。本研

究 對 LZ91 鎂 鋰 鋅 合 金 施 以 不 同 熱 機 處 理 ( Thermo mechanical treatment)程序,並以 TEM、SEM、XRD 及微硬度計等設備,探討 LZ91 合金之微觀結構與時效析出(Precipitation)處理之關係。採用 抗拉試驗及成型極限曲線(Forming limit curve)評估前述各材料之機 械性質與成型性,亦由真應力-應變曲線探討其應變硬化指數(Strain hardening exponents, n values)。

研究結果顯示,鎂鋰鋅合金具有大量的冷加工能力,係因原HCP

之鎂合金微結構中,出現延展性佳之 BCC 相所致。析出處理及冷加

工對提升 LZ91 強度之效益均不顯著,然冷加工方式稍佳。LZ91 合

金中散佈微細Wurtzite 結構的 ZnO 與立方晶之 MgO 顆粒;ZnO 與鎂

基地具有良好的整合性(Coherence),而 MgO 之整合性則不佳。時 效析出處理之結果亦顯示,LZ91 在 100℃處理 10 小時可達尖峰時效

(3)

(Peak aging);而於 50℃析出處理時,則須 100 小時才能得到最大 硬度值。X 光繞射分析則顯示,此合金經 50℃/100 小時與 100℃/10 小時的時效處理後,α(0002)面之主繞射峰旁均出現額外之小峰,此

現象應為析出相或Spinodal 相分解所造成。

不同鋰含量之鎂鋰鋅合金的拉伸試驗結果顯示,提高鋰含量可明

顯增加材料之延展性,且 LZ61、LZ91 與 LZ101 的拉伸強度異向性

(Anisotropy)不明顯,可能利於板材之沖壓成型製程。此外,LZ91

與 LZ101 的延伸率均超過 40%,與一般常溫沖壓鋼板及 200℃之

AZ31B 鎂板的延展性相當。而應變硬化指數則隨合金之鋰含量的增

加而減少,且含鋰大於 9wt.%合金的 n 值小於 0.05。LZ91 與 LZ101

之曲線位於成型極限圖上方區塊;AZ31B 於 100℃及室溫的曲線則居

於其下方部位,顯示LZ91 與 LZ101 具有較佳之成型性。

關鍵字:鎂鋰鋅合金、析出處理、Wurtzite 結構、整合性、應變硬化 指數、成型極限曲線

(4)

ABSTRACT

An Mg-9%Li-1%Zn (LZ91) alloy was successfully cast into a 37mm thick ingot by vacuum induction melting technique, and then rolled into a thickness of 2mm. Greater size of 200 mm diameter billets with various compositions, which included Mg-6%Li-1%Zn (LZ61), Mg-9%Li-1%Zn and Mg-10%Li-1%Zn (LZ101), were also cast as well and extruded and subsequently rolled into 0.6-mm in thickness. The tests of tension and forming limit diagram (FLD) were conducted to investigate the mechanical behavior and formability of the magnesium-lithium alloys.

The strain hardening exponents, n, of the alloys were also calculated through various true stress-true strain data sets. The preceding alloys have remarkable workability with respective to the rolling process, not demonstrated by other Mg alloys. This was attributed to the presence of a ductile β phase of BCC structure, despite the coexistence of the brittle HCP α phase. Thermo and mechanical treatments were performed, and the resultant microstructures were examined to elucidate the strengthening mechanisms associated with the alloys. A distinguishing feature of this research is the employment of XRD, SEM and TEM

(5)

instruments to identify various phases involved.

The experimental results indicated that neither age hardening nor cold rolling was effective in improving the strength of the LZ91 alloy, with cold rolling modestly better. Small number of fine particles was detected in the quenched state of the alloy, and their presence may contribute to the strength of this alloy. Furthermore, the LZ91 alloy had a dual phase structure with dispersed particles of ZnO and MgO oxides.

The Wurtzite structure of ZnO was well oriented with respect to the Mg matrix, but the MgO was not. Peak aging hardness was obtained at a temperature of 100°C for 10hrs. Alternatively, maximum hardness could also be reached at 50°C for 100 hrs. In the XRD spectrum, the appearance of the extra bump next to the main peak of α(0002) after aging treatment at 50°C/100 hrs and 100°C/10 hrs, was believed to originate from a precipitate phase or a phase resulted from spinodal decomposition.

The results of tensile tests revealed that the ductility of various magnesium-lithium alloys was much improved with a higher Li content.

Furthermore, the tensile strength of Mg-Li sheets were found to show little anisotropy, which is expected to deliver great benefit to the press-forming process. The elongation of LZ91 and LZ101 were found to

(6)

exceed 40%, which were comparable to those of conventional stamping steel and AZ31 at 200℃. The calculated n value demonstrated that (a) n decreased with the increasing lithium content in LZ alloys, and (b) as the lithium content was higher than 9wt%, all the n values were lower than 0.05. The present research found that the forming limit curves of LZ91 and LZ 101 were located at higher positions in the major strain versus minor strain plots than their counterparts of AZ31 at 100℃ and room temperature. This indicates a better formability of the alloys.

Keywords: Mg-Li-Zn alloy, Precipitation hardening, Wurtzite structure, Coherence, Strain hardening exponent, Forming limit curve

(7)

謝 誌

本論文承蒙指導教授吳泓瑜博士及共同指導教授王建義博士的 悉心指導方能順利付梓,於此致上最深的敬意與謝忱。吳老師認真的 教學精神、寬宏豁達的情懷,讓學生耳濡目染,不論在學識上及待人 處事方面均有莫大長進,對我從事的研發工作更獲得諸多的啟示。在 此亦特別感謝王建義教授帶領我進入博士研究及學習的生涯,在我離 開學校、步入職場的多年後推薦及鼓勵我重拾書本,使我得到許多成 長的機會與寶貴的經驗。

感謝同事林景正博士在投稿國外期刊論文的諸多指正與建議,以 及亦師亦友的老同事陳豐彥博士在科技英文的教導與修正,促使我在 英文撰稿方面有顯著的進步。此外,本論文經由機械系林育立主任、

葉明勳教授、李雄教授及碩士班指導老師楊智富教授的悉心審核與斧 正,使得本論文內容更加詳實與完善,在此表達誠摯之謝意。

在這幾年學習的歷程中,同事吳學陞博士、陳俊沐博士、王文寬 博士、林珮君工程師及張淑芬管理師在工作上的支持與協助,是我完 成階段性學習生涯不可或缺的要素與助力。與鋁合金實驗室的同門師 兄弟渝翔、禎蔚、耿中、家宇、宏偉相處和諧融洽,並獲得各方面的 襄助亦是十分感念。

最後特別感謝我敬愛的父母、岳母及愛妻嘉惠,在上活上的分擔 與付出,讓我無後顧之憂地完成博士學位的求學目標。

(8)

TABLE OF CONTENTS

摘 要... I ABSTRACT...III 謝 誌... VI

CHAPTER 1 OVERVIEW ...1

1.1 Scope ...1

1.2 Problems and Issues...3

1.3 Research Objective ...5

CHAPTER 2 LITERATURE SURVEY ...7

2.1 Alloys Based on the Mg-Li System...7

2.2 Characterization of Mg-Li Alloys ...15

2.3 Deformation Textures and Slip in HCP Metals...24

2.4 Corrosion Control of Mg-Li Alloys...28

2.5 Mg-Li Matrix Composites ...32

CHAPTER 3 EXPERIMENTAL PROCEDURES...38

3.1 Material Preparation ...38

3.2 Thermo-mechanical Processing...44

3.3 Microstructural Characterization...47

(9)

3.4 Mechanical Properties Test...47

3.5 Formability Evolution of Mg-Li Alloys...48

CHAPTER 4 RESULTS AND DISCUSSION ...52

4.1 As-cast LZ91 Microstructures...52

4.2 As-rolled and Solution-treated LZ91 Microstructures...58

4.3 Effect of Aging Treatment on LZ91 ...64

4.4 Cold Work Hardening Phenomenon of LZ91...77

4.5 Softening of LZ91 by Annealing...80

4.6 Tensile Properties of LZ Alloys ...82

4.7 Strain Hardening Exponent of Mg-Li Alloys (the n value) ...92

4.8 Forming Limit Diagram (FLD) of Mg-Li Alloys ...95

CHAPTER 5 CONCLUSIONS ...104

FUTURE STUDIES ...107

REFERENCES ...108

LIST OF PUBLICATIONS ...116

CURRICULUM VITAE ...120

(10)

LIST OF TABLE

Table 2.1 Properties of LA141A and LS141A at room

temperature [19]...10 Table 2.2 The effect of temperature range on the coefficient of

linear thermal expansion of LA141A [28]...11 Table 2.3 The slip systems of FCC, BCC and HCP structures

[45]...27 Table 3.1 Chemical analysis results of as cast

magnesium-lithium alloys, in weight percent...41 Table 4.1 Tensile properties obtained from uniaxial tension tests

at an initial strain rate of 1×10-3S-1. ...88 Table 4.2 Comparison of n value for Mg-Li alloy and AZ31B

Magnesium alloy...94

(11)

LIST OF FIGURE

Fig. 2.1 Effect of temperature on the modulus of elasticity of

LA141A [29]...11 Fig. 2.2 Effect of temperature on the properties of LA141A [29]...12 Fig. 2.3 Effect of temperature on the tensile elongation of

LA141A [29]...12 Fig. 2.4 Effect of temperature on the specific heat of LA141A

[29]...13 Fig. 2.5 Effect of temperature on the thermal conductivity of

LA141A [29]...13 Fig. 2.6 Effect of temperature on the tensile strength of LS141A

[29]...14 Fig. 2.7 Effect of temperature on the tensile elongation of

LS141A [29]...14 Fig. 2.8 Binary magnesium-lithium phase diagram of

Mg-9wt.%Li alloy, comprising α and β phases [17]. ...20 Fig. 2.9 Stress-strain response of MgLi alloys with different Li

alloying (in wt.%) that correspond to single-phase

(12)

structures α and β as well as two-phase one (α+β) [21]:

α-lines a, b, c; (α+β)-line d; β-line e...21 Fig. 2.10 Tensile strength (UTS) of age-hardened

(α+β)-MgLiAl and (α+β)-MgLiAlZn alloys

demonstrating effect of additional Al and Zn alloying

[31]...22 Fig. 2.11 Effect of additional alloying of Mg14Li alloy on its

corrosion rate in an artificial tropical climate

environment [7]...23 Fig. 2.12 The mass-time course of various Mg-Li alloys in the

synthetic seawater [6]...31 Fig. 2.13 Model of the formation of interfacial bond in

δ-Al2O3/MgLi composites during melt infiltration process in which Li. ions are inserted topotactically into cation vacancies on the surface of δ-Al2O3

lattice [31]...37 Fig. 3.1 The overall experimental framework of this study...40 Fig. 3.2 The LZ61, LZ91 and LZ101 alloys, in the diameter of

200mm and as-cast condition, were successfully

(13)

produced for this work...41 Fig. 3.3 The optical microstructure of as-cast LZ61 alloy,

comprising α and β phases (bright and dark regions,

respectively)...42 Fig. 3.4 The optical microstructure of as-cast LZ91 alloy,

comprising α and β phases (bright and dark regions,

respectively)...42 Fig. 3.5 The optical microstructure of as-cast LZ101 alloy,

comprising α and β phases (bright and dark regions,

respectively)...43 Fig. 3.6 The Mg-Li plate of 3mm in thickness was produced by

direct extrusion at the temperature of 200℃...46 Fig. 3.7 The coiled Mg-Li sheet in the thickness of 0.6mm and

0.17mm was obtained using 4H precision rolling mill

at ambient temperature...46 Fig. 3.8 The FLD test moulds, including semi-spherical punch,

die and blank holder, for the forming limit test...50 Fig. 3.9 The schematic diagram showed the main dimension of

the hemi-sphere punch drawing mould...51

(14)

Fig. 3.10 The Measurement of principal major and minor strains for the FLD test...51 Fig. 4.1 Optical microstructure of the as-cast LZ91 alloy...54 Fig. 4.2 α phase on the top of the photograph is almost free of

dislocations, while the β phase, on the bottom has a

high density of dislocations...54 Fig. 4.3 TEM analysis of as-cast specimen; oxide particles were

distributed in the α matrix, and were identified as ZnO

(round particles) and MgO (polygonal particles)...55 Fig. 4.4 Granular particles (a) and matrix (b) identified by EDS.

The particles contain Zn and O, and are ZnO...56 Fig. 4.5 Faceted particle (a) and matrix (b) identified by EDS.

The particles contain Mg and O, and are MgO...57 Fig. 4.6 Micro-structure of the as-rolled LZ91 alloy...59 Fig. 4.7 Optical micrograph of the as-rolled and solution-treated

LZ91 alloy...60 Fig. 4.8 TEM morphology of the rolled and solution-treated

LZ91 specimen...60 Fig. 4.9 Nano-beam diffraction pattern of as-cast LZ91

(15)

specimen. (a) TEM image, (b) diffraction pattern of the matrix, and (c) differaction pattern of the round ZnO

particle...61 Fig. 4.10 TEM images of solution-treated LZ91 alloy. Faceted

particles are present in the grain boundary area...62 Fig. 4.11 Nano-beam diffraction pattern of the polygonal

particle. (a) TEM image, (b) diffraction pattern of the polygonal MgO particle and (c) diffraction pattern of

the matrix...63 Fig. 4.12 The effects of temperature and soaking time on the

hardness of the LZ91 alloy during aging heat treating...67 Fig. 4.13 Optical micrographs of the LZ91 were aged at 50℃

for (a) 1 hour, (b) 3 hours, (c) 10 hours and (d) 100

hours...68 Fig. 4.13 Continued...69 Fig. 4.14 Optical micrographs of the LZ91 were aged at 100℃

for (a) 1 hour, (b) 3 hours, (c) 10 hours and (d) 100

hours...70 Fig. 4.14 Continued...71

(16)

Fig. 4.15 The effects of temperature and soaking time on the mechanical properties of the LZ91 alloy during aging

heat treatment...72 Fig. 4.16 XRD results of aging at 50℃ (a) and 100℃ (b),

showing an extra bump adjacent to the main peak of α (0002), indicated by arrows in the figures...73 Fig. 4.17 TEM image of the specimen aged at 100°C for 10h,

indicating a particle-like structure...74 Fig. 4.18 HRTEM image of the particle-like structure (a)

demonstrates the short-range structural modulation.

Fourier transformation of the image (b) presents many sticks that correspond to the modulation structure...75 Fig. 4.19 HR-TEM image of the specimen aged at 50°C for 100

hours showing the structure of the yarn-like particle...76 Fig. 4.20 Micro-hardness with respect to cold rolling reduction...78 Fig. 4.21 TEM observations of the cold rolled LZ91 alloy with a

reduction of 80%. (a)BFI, (b) DFI, (c) SADP, (d) Key diagram for SADP...79 Fig. 4.22 The results of hardness as a function of 30 minutes

(17)

annealing at various temperatures...81

Fig. 4.23 The nominal stress-strain curve of LZ61 in three different rolling directions...86

Fig. 4.24 The nominal stress-strain curve of LZ91 in three different rolling directions...87

Fig. 4.25 The nominal stress-strain curve of LZ101 in different rolling directions...88

Fig. 4.26 True stress-strain relationships of Mg-6Li-1Zn (Material A), Mg-9.5Li-1Zn (Material B) and Mg-12Li-1Zn (Material C) at various strain rates [20]...89

Fig. 4.27 True stress-strain relationships of LZ magnesium alloys at an initial strain rate of 1×10-3S-1...90

Fig. 4.28 The stress-strain curves of LZ and conventional AZ31B magnesium alloys...91

Fig. 4.29 The strain modes of stretching, plane strain and draw in the sheet forming processing [76]...99

Fig. 4.30 The forming limit test specimens of the LZ61 alloy...100

Fig. 4.31 The forming limit curve of LZ61...100

Fig. 4.32 The forming limit test specimens of the LZ91 alloy...101

(18)

Fig. 4.33 The forming limit test specimens of the LZ101 alloy...101 Fig. 4.34 The forming limit curve of LZ91...102 Fig. 4.35 The forming limit curve of LZ101...102 Fig. 4.36 Three maximum specimen sizes relative to forming

limit are shown...103 Fig. 4.37 The forming limit curves of LZ alloys at RT, and

AZ31 from RT to elevated temperature. The limit curves of LZ91 and LZ 101 were located at higher

position than their counterparts of AZ31 at 100℃...103

(19)

CHAPTER 1 OVERVIEW

1.1 Scope

Magnesium is the lightest metal that can be employed in structural applications when alloyed with other elements. Magnesium alloys have recently been adopted to make portable electronic devices because these alloys are light, show high electrical and thermal conductivities, and can be recycled. They have also been used in automotive and aerospace industries due to their attractive properties including low density, high specific strength and superior stiffness-to-weight ratio [1, 2].

Recently, environmental protection, especially by reduction of CO2

emissions, has become an urgent issue in the world. As one of attempts to reach the goal, considerable efforts are being made in applications of lightweight materials, such as Al, Ti and Mg alloys, to transportation systems [3]. For instance, research on Mg alloy focusing on mechanical properties has become very active in the last decade [4-6]. In light of hexagonal close packed (HCP) structure, Mg and its alloys have crucial drawback of poor formability, especially at room temperature, as compared with aluminum and its alloys. Since slip at room temperature is

(20)

limited to the basal plane, the processing and forming capabilities of magnesium alloys are generally fairly poor. As a result, aluminum-based alloys are employed for a considerable number of components; the widespread applications of Mg alloys to structural components are still limited at present.

It is well known that the addition of lithium to magnesium gives rise to highly workable, body-centered cubic alloys [7-9]. Alloying Mg with the lightest metal element, lithium, whose density is 0.534g/cm3, yields an Mg-Li alloy whose density is similar to those of plastics, and which has only half the density of aluminum alloys. Furthermore, the notorious HCP crystalline structure of general Mg alloys, which has inherently poor formability, can be altered to considerably enhance manufacturability [10-12]. Due to their ultra-low density, Mg-Li alloys are attracting much attention, not only in the automotive and aerospace industries applications, but also for use in electric appliances and consumer products.

As the amount of Li added to the Mg-Li alloy increases, the solid solution of Mg (α phase) still possesses HCP structure, but the crystal lattice axes ratio, c/a declines such that slippage between crystal planes become less difficult [8, 13]. Simultaneously, the BCC β phase of the Li

(21)

solid solution co-exists, further facilitating cold-working.

Mg-Li alloys were introduced in the sixties when NASA developed a series of Mg-Li-Al alloys centered on LA141A for aerospace applications [14]. LA141A is one of the Mg-Li alloys that has been successfully commercialized. However, the mechanical properties of LA141A alloy are not particularly favorable, as evidenced by its low tensile and yield strength as well as its inadequate corrosion ability. These characteristics offset the advantage of lightness, such that research and development of Mg-Li was not pursued. Currently, Mg alloys have numerous applications, with great potential in computer, communication and consumer products industry. Mg-Li alloys are only now regaining attention.

1.2 Problems and Issues

Haferkamp et al. [6] recently reported new magnesium-lithium alloy systems with low density, improved ductility and corrosion resistance.

Cold working and precipitation hardening [15, 16] have been suggested to improve on the poor mechanical characteristics of magnesium-lithium alloy. The Mg-Li phase diagram [17] indicates that when the Li content is

(22)

solid solution co-exists with the HCP α phase of the Mg solid solution.

Most presently standardized Mg-Li alloys basically have α or α+β phase(s). Adding third elements such as zinc, as in the material in this work, can manifest particular desirable properties. Due to inadequate research on these alloys, knowledge about them is poor. This includes the alloy’s unstable mechanical properties [18] and its inconclusive room temperature aging results [19].

The alloy design and mechanical properties of Mg-Li alloys have been investigated [20-27]. However few studies have been carried out to examine the evolution of microstructure of Mg-Li alloys during processing, though it is important to the development of a magnesium-lithium alloy with suitable mechanical characteristics.

Due to hexagonal closed-packed (HCP) crystal structures, magnesium alloys show low ductility at room temperature, and require thermal activation to increase their formability. However, it is well known that ductility of magnesium alloys can be improved with addition of lithium that helps developing the formation of body centered-cubic (BCC) crystal structures [7-9]. The BCC crystal structure gives rise to high formability at room temperature. The formability of different

(23)

magnesium-lithium alloy sheets, such as Mg-6%Li-1%Zn, Mg-9%Li-1

%Zn and Mg-10%Li-1%Zn, was investigated by conducting various experiments in the present study. Tensile tests and forming limit tests were first conducted to investigate the mechanical behavior and formability of magnesium-lithium alloys with various lithium contents.

1.3 Research Objective

Understanding micro-structural changes during processing is important to the development of a magnesium-lithium alloy with suitable mechanical characteristics. This work made TEM observations of the magnesium-lithium alloys, and this investigation explored the variations in the microstructures of a dual-phase Mg-9%Li-1%Zn alloy. The relationship between microstructures and mechanical behavior was analyzed to examine the strengthening mechanism. Furthermore, the difference in the mechanical behavior depending on the lithium contents of Mg-Li alloys was also investigated by means of tensile and forming limit tests. The formability results obtained in the present study might help to provide the fundamentals for the stamping die design of forming

(24)

initiated with two objectives. First, to examine the microstructures of Mg-9%Li-1%Zn alloy having a two-phase structure. Secondly, for Mg-6

%Li-1 % Zn, Mg-9 % Li-1 % Zn and Mg-10 % Li-1 % Zn alloys, quantitative measurements of their tensile properties and forming limits are to be performed to reach the conclusions concerning the formability of these alloys.

(25)

CHAPTER 2 LITERATURE SURVEY

2.1 Alloys Based on the Mg-Li System

Small numbers of ultra-light alloys have been developed over the years by alloying magnesium with Li, along with other elements such as Al, Zn, or Si. These alloys, however, have found only limited application and are not commercially available today.

Among those applications, two Mg-Li alloys, LA141 and LS141A [14], had received some commercial success in the 1960s. These were produced in the form of sheet, extrusions, and castings, mainly for use in aerospace and military applications. During that period, they were used in computer housings for the Saturn V project, circuit-module covers for the Gemini computer program, accelerometer housings for the Minuteman missile, and parts of the aiming device for the TOW missile launcher [14].

The reason that these alloys were widely used in these applications was due to their extreme light weight, good strength and very high stiffness that were needed in these parts. The properties of those alloys are shown in Tables 2.1-2.2 and Figs. 2.1-2.7 [19, 28, 29]. The room

(26)

temperature elastic modulus of LA141A is 42 GPa, which is almost as high as the value for conventional magnesium alloys, 45 GPa. However, the density of LA141A at room temperature is only 1.35 g/cm3, compared to around 1.80 g/cm3 for most conventional magnesium alloys. LS141A as an elastic modulus of 41 GPa and a density 1.33 g/cm3, which provides even greater bending stiffness than those of LA141A.

The lithium contents of both LA141A and LS141A range from 13 to 15%, while the aluminum content of LA141A is from 0.75 to 1.50%, and the silicon content of LS141A is from 0.5 to 0.8%. As it may be seen, the lithium contents of both alloys are greater than the approximately 11.5%

need to make them essentially all β-alloys and, therefore, quite formable at room temperature. However, some Mg-Li alloys such as LA91are produced with a mixed α+β microstructure. The intermetallic compounds found in the BCC LA141A alloy include FCC LiAl and FCC Li2MgAl, which is metastable phase.

As discussed in the section, Mg-Li alloys are amenable to age hardening, and LA141A was commonly used in the T7 temper (solution heat treated and stabilized) at 290℃ for 1 hour per mm of thickness, air cool and stabilized at 180℃ for 3 to 6 h. This stabilization treatment has

(27)

been found to eliminate the susceptibility of Mg-Li alloys to stress-corrosion cracking (SCC). All parts welded, however, must be stress relieved immediately after welding to prevent SCC.

Mg-Li alloys can be welded and machined much like other magnesium alloys. One problem with Mg-Li alloys, however, is that they are significantly more chemically active than other magnesium alloys.

Also, the corrosion resistance of Mg-Li alloys decreases with increasing aluminum contents due to microglvanic reactions. A fluoride anodizing treatment has been used as a paint base for Mg-Li alloys, but no coating has been found completely protects them in temperature-cycling and humidity tests.

(28)

Table 2.1 Properties of LA141A and LS141A at room temperature [19]

Property LA141A LS141A

Elastic modulus, GPa 42 41

Tensile strength, MPa 144 136

Tensile yield strength, MPa 123 110

Elongation in 50mm, % 23 23

Hardness, HRE 55-65 …

Density, g/cm3 1.35 1.33

Coefficient of linear thermal expansion, μ /m‧K

21.8 …

Specific heat, kJ/kg‧K 1449 …

Thermal conductivity, W/m‧K 80 …

Electrical resistivity, nΩ‧m 152 …

(29)

Table 2.2 The effect of temperature range on the coefficient of linear thermal expansion of LA141A [28]

Temperature range, ℃ Coefficient of linear thermal expansion, μm/m‧K

-130~+24 21.5 24~100 21.7 100~200 22.2

Fig. 2.1 Effect of temperature on the modulus of elasticity of LA141A [29]

(30)

Fig. 2.2 Effect of temperature on the properties of LA141A [29]

Fig. 2.3 Effect of temperature on the tensile elongation of LA141A [29]

(31)

Fig. 2.4 Effect of temperature on the specific heat of LA141A [29]

Fig. 2.5 Effect of temperature on the thermal conductivity of LA141A [29]

(32)

Fig. 2.6 Effect of temperature on the tensile strength of LS141A [29]

Fig. 2.7 Effect of temperature on the tensile elongation of LS141A [29]

(33)

2.2 Characterization of Mg-Li Alloys

As illustrated in Figure 2.8, binary Mg-Li phase diagram is of simple eutectic type with extensive solid solution regions HCP (α) and BCC (β), separated by the two-phase region (α+β) [17]. The α-MgLi assigned solid solution of magnesium with low lithium content (<5.5 wt.% Li) retains HCP magnesium structure with reduced lattice c-parameter which causes the activation of prismatic slip planes

{

1010

}

2110 in addition to basal

ones

{

0001

}

2110 , thus increasing considerably the ductility and preserving the moderate strength, wherein alloying as little as ~2 wt.% Li enhances markedly the tensile ductility and formability [10]. Large additions of lithium (>11.5 wt.% Li) transforms HCP lattice to BCC lithium-based phase β-MgLi which results in considerable strength decrease and essential ductility improvement at the same time. In addition, extremely high diffusivity of Li in cubic lithium-rich β-MgLi phase (~10-6cm2 s-1 at 420℃ [30]) is considered the main reason for pronounced creep of β-MgLi based alloys taking place even at room temperature [31].

Two-phase structure (α+β)-MgLi is an interesting material combination of α-phase with moderate strength and β-phase with excellent ductility.

(34)

stress-strain responses, as shown in Fig. 2.9 [22]. Dissolving of lithium in magnesium results in only small solid solution hardening effect without exhibiting any precipitation hardening capability due to the absence of Mg-Li intermetallic phases. Hence, the use of binary Mg-Li alloys in structural applications is largely frustrated with their low strength due to the lack of effective solid solution-, precipitation- and work hardening-mechanisms as well as poor creep resistance.

Ternary MgLiX alloys (usually X=Al, Zn, Ag) can be age-hardened by LiX type precipitates, as demonstrated by ternary MgLiAl system where formation of stable LiAl precipitates is preceded by the appearance of coherent MgLi2Al phase (Guinier-Preston type phase) with high hardening efficiency [7]. The hardening of β-MgLiAl takes place already at room temperature, nevertheless this effect is instable because MgLi2Al→AlLi transformation accelerates already at slightly enhanced temperatures (60-80℃) causing the overaging phenomena [32, 33], when, e.g., the ultimate tensile strength (UTS) of commercial Mg14Li1Al alloy (LA141A) at 90-150℃ decreased by ~50% from its room temperature (RT) value [26]. Several attempts were made to prevent rapid loss in strength of age-hardened MgLiX alloys above RT. Two processing

(35)

methods involving cold rolling +annealing + solution heating were proposed to fragment coarse dispersoids thus stabilizing the structure and inhibiting the recrystallization of LA141A up to elevated temperatures [26]. Since the growth of precipitates is governed by Li-diffusion, the overaging might be potentially retarded by the addition of a fourth element (e.g., Ag) which owing to its high solubility hampers the Li-migration, nevertheless this method has brought only limited effect [31]. The hardenability and creep resistance of β-MgLiAl alloys were significantly improved by the addition of ~0.1 wt.% Zr to form fine Al3Zr precipitates that are stable up to 200℃ due to low zirconium diffusivity [34]. In quaternary β-MgLiAlCu alloys the low soluble copper was considered to stimulate massive nucleation of fine precipitates (peak-aging), resulting in some room temperature hardening but maintaining excellent ductility at the same time [24].

The α-MgLiAl alloys can be hardened by the precipitation of stable AlLi and Mg17Al12 phases but such hardening is less effective than that attained in β-MgLiAl alloys when, moreover, lithium frustrates the micro-segregation of aluminum and precipitation of MgAl eutectics [6].

In two-phase (α+β)-MgLiAl alloys the sub-micron MgLi2Al particles and

(36)

coarse AlLi ones occurred in both α and β regions [35] while for high Al content these precipitates appeared predominantly in β regions at α/β interfaces in particular [31]. Figure 2.10 illustrates the effect of additional alloying of two-phase (α+β)-Mg8Li alloy on its tensile strength when the strengthening of (α+β)-Mg8LiAl alloys is attributable predominantly to the solution strengthening of α and β phases although the contribution of MgLi2Al precipitates cannot be excluded [36].

A serious drawback of Mg-Li alloys is their poor corrosion resistance caused by the high lithium reactivity and segregation of alloying elements thus creating localized cathodic centers. Increase in Li content drops dramatically the corrosion resistance of β-MgLi alloys while the additional alloying (e.g., Al, Zn, Sn) provides ambiguous effect as demonstrated by the minimum on the plot concentration vs corrosion rate (Fig. 2.11). In the case where the alloying element enters the solid solution (e.g., MgLiZn, MgLiSn), the corrosion resistance is usually improved but the precipitation of intermetallic products at higher concentrations leads to the worsening of the corrosion behavior (cathode effect) [31]. Likewise, the study of air oxidation of several Mg-Li and MgLiSi alloys has revealed that single-phase materials offer a superior

(37)

resistance to oxidation than do the two-phase materials [37]. The β-MgLiAl alloys suffer from severe aqueous corrosion and stress corrosion in chloride environments and this kind of corrosion can be markedly inhibited by the addition of n-butylamine and potassium chromate to the corrosive agent [38]. Alkaline impurities (Na, K) exhibit exceedingly detrimental effect on the grain boundary corrosion of Mg-Li alloys so that their concentration has to be maintained on the ppm level [39]. On the other hand, the corrosion stability of α-MgLi based alloys is significantly better than that of pure Mg wherein recently developed α-MgLi alloys slightly alloyed with Ca appear to be particularly promising in this regard [6].

(38)

Fig. 2.8 Binary magnesium-lithium phase diagram of Mg-9wt.%Li alloy, comprising α and β phases [17].

(39)

Fig. 2.9 Stress-strain response of MgLi alloys with different Li alloying (in wt.%) that correspond to single-phase structures α and β as well as two-phase one (α+β) [21]:

α-lines a, b, c; (α+β)-line d; β-line e.

(40)

Fig. 2.10 Tensile strength (UTS) of age-hardened (α+β)-MgLiAl and (α+β)-MgLiAlZn alloys demonstrating effect of additional Al and Zn alloying [31].

(41)

Fig. 2.11 Effect of additional alloying of Mg14Li alloy on its corrosion rate in an artificial tropical climate environment [7]

(42)

2.3 Deformation Textures and Slip in HCP Metals

When compared to crystal systems like face-centered or body-centered cubic, hexagonal close packed (hcp) metals exhibit a wider variety of deformation textures. Historically, hcp metals have been categorized in terms of c/a ratio and the observed rolling textures were similarly categorized. The distinct textures are actually due to the combined effects of c/a ratio and the fact that different hcp metals deform by different slip and/or twinning modes [40]. An example of how these aspects are intertwined is demonstrated by a discussion of the common

{ }

1012 twinning mode. For metals with c/a < 3 (e.g. beryllium,

titanium, zirconium, magnesium), the

{ }

1012 twin is activated by c-axis tension. During compression, grains are favorably oriented if their c-axis is perpendicular to the compression axis; and twinning reorients the c-axis of the twin nearly parallel to the compression axis. This is a major reason why basal or near-basal textures are common for cold-rolled hcp metals, as has been documented by Phillippe and collaborators [41-44] in an extensive study of texture evolution (during rolling) and mechanical behavior of hexagonal metals. Taking Zn as an exception to this rule, it is

{ }

(43)

compression along the c-axis (c/a> 3) and <c + a> slip is active. Hence, the common tendency for hcp metals to develop a basal compression texture is precluded in the case of Zn. Finally, it is often suggested that Mg will develop an ideal basal texture, since it has a nearly ideal c/a ratio of 1.624 [45, 46].

Although the slip mechanisms are the same in single and polycrystalline materials, their stress strain behaviour differs substantially.

Because neighbouring grains in a polycrystal constrain the deformation of each other, the deformation of the grains in a polycrystal are different to the response that would occur if each grain was tested as a single crystal.

Specifically, the displacements across grain boundaries must be matched so that the grains can deform in concert. Without cooperative displacements, cracks would develop at the grain boundaries. It can be shown that five independent slip systems are required to meet the boundary compatibility requirements in order to allow substantial plastic deformation in a polycrystalline aggregate. This is von Mises criterion.

Without these five independent slip systems, polycrystalline materials are brittle.

An independent slip system is one for which slip displacements on it

(44)

cannot be duplicated by a combination of displacements on other slip systems. The number of independent slip systems is usually less than the number of geometrical slip systems.

In HCP metals with a high c/a ratio, including Mg, slip occurs predominantly on the basal plane. This is because the Peierls stress is small because the burgers vector is small and the interplanar spacing is large. Prismatic slip does not occur in Mg, but does predominate in HCP metals with a small c/a ratio, such as Ti. Because slip in Mg only occurs in the basal plane and there are therefore only two independent slip systems, polycrystalline Mg is brittle.

(45)

Table 2.3 The slip systems of FCC, BCC and HCP structures [45]

Structure Slip Plane

Slip Direction

Number of non-parallel

planes

Slip directions

per plane

Number of geometrical

slip systems

Number of independent slip systems

FCC {111} <110> 4 3 12 5

{110} <111> 6 2 12 5

{112} <111> 12 1 12 5

BCC

{123} <111> 24 1 24 5

{0001} <11 2 0> 1 3 3 2

{1010} <11 2 0> 3 1 3 2

HCP

{1011} <11 2 0> 6 1 6 4

(46)

2.4 Corrosion Control of Mg-Li Alloys

The use of magnesium alloy is obstructed by insufficient corrosion resistance. Alloys of lithium, aluminum or calcium have influence on particular mechanisms and effects on corrosion. MgLi12at% has a higher corrosion resistance than that of magnesium. This can be enhanced by Al in atmosphere and by Ca in synthetic seawater [6]. Some ultra-light Mg-Li alloys have better corrosion resistance than that of commercial Mg ones. It is shown that active corrosion protection can be realized by a further lowered potential.

Li is chosen as an alloy addition owing to its attribute of not to react with OH- at any pH values. This permits the formation of a Mg(OH)2

out-layer which is stabilized by the Li increased pH value[47]. The selected test alloys include HCP magnesium, MgLi4at%, MgLi8at%, MgLi12at% and BCC MgLi40at%. According to ASTM D1141-52, these alloys are tested in synthetic seawater immersion consisting of the four main salts.

Mg is already dissolved after four days. The mass-time-course of the Mg-Li alloys shows that minor Li additions can lower the losses with

(47)

corrodes much faster and builds thick layers falling off at once (Fig.

2.12).

The results for HCP Mg-Li are backed by the conductance of MgLi12at% which is half the one of Mg after 200 h. The potential of MgLi12at% does not deviate intensely from that of Mg. Compared to AZ91D, MgLi12at% exhibits a smooth and constant course, which indicates a relatively high corrosion resistance. However, in discussing the mass-time-courses, the sticking products of corrosion on the plain and continuous attacked MgLi12at% have to be taken in account. The dense Li that induced Mg(OH)2 layer has to be stabilized by further additions.

In order to prevent a continuous formation and dissolution of the layer, these new additions should be adapted to the lattice of the Mg substrate and avoid dynamic alkalinization.

Al and Ca additions are selected for a future mechanical and chemical Mg(OH)2 layer stabilization. Mg-Al alloys are known to be sufficiently solid in aggressive media and highly resistant in a natural environment. Moreover, Al can form a solid solution with magnesium, or the so-called protecting layer component MgAl2O4, which is protected by its over-voltage. Another way to minimize the difference of the corrosion

(48)

potential is by this oxide forming element. However, high Al additions, e.g. 8 at%, support selective corrosion of the multi-phase structure. In neutral environments Al can be protected by the transfering Al2O3 on the Mg surface. Similarly, Ca is chosen because it forms a solid solution containing Ca and it does not react with OH-. Both additions are allowed to form the Mg(OH)2 based protective layer which makes the magnesium ternary alloys MgLiAl and MgLiCa a corrosion resistance system with increased ductility and creep resistance.

(49)

Fig. 2.12 The mass-time course of various Mg-Li alloys in the synthetic seawater [6].

(50)

2.5 Mg-Li Matrix Composites

Many process technologies including melt infiltration, reaction stirring, powder metallurgy technique and foil metallurgy have been applied to manufacture Mg-Li matrix composites. Except for reaction stirring where complete reaction between the added refractory powder and Mg-Li melt is requested, the management of the chemical reaction between the reinforcement and the matrix represents a crucial processing problem especially if the melt infiltration techniques are applied. Solid state techniques enable better control of interfacial reactions as a rule, but cause less intimate contact between components thus affecting the formation of effective interfacial bond.

Common reinforcements applied in metal matrix composites (e.g., alumina, silicon carbide, carbon, boron) are only slightly sensitive to the attack of pure magnesium, nevertheless thermodynamics and kinetics of these interfacial reactions is radically altered by the presence of lithium.

The activity concept should be used to assess correctly thermodynamic driving forces at interfaces in Mg-Li matrix composites wherein useful data about Li activity in Mg-Li and MgLiAl alloys provide the works [48].

(51)

reinforcements in particular would be considerably affected with MgLi matrix at relevant processing temperatures (<700℃) [49]. Briefly reviewed below are experimental observations concerning the chemical compatibility between Mg, Li and Mg-Li species and solid refractories (steel, titanium, carbon, boron, alumina, silicon carbide) applied, until now, as the reinforcements in Mg-Li matrix composites.

Both iron and titanium are relatively stable in molten magnesium and lithium when only negligible dissolving (0.1-0.01 at.%) in liquid metals takes place without formation of intermetallic phases [31].

Nevertheless, the wires of highly alloyed maraging steel (14Cr13Co5Mo) entered the reaction with molten Mg14Li1Al alloy (LA141A) to some extent, apparently under participation of alloying elements occurring in the steel wire [50].

Although magnesium boride could be formed by the reaction between the boron fiber and the magnesium vapor at 950 [℃ 51], no reaction between boron fibers and molten magnesium was found at 850℃

and thin MgB2 layer appeared at interfaces only after long-term heating of B/Mg composites at 400-500℃. The formation of lithium hexaboride Li2B6 during contact of elemental boron with molten lithium was reported

(52)

[31], nevertheless no affection of boron fibers with molten lithium was observed at the melting point (180℃) [52].

Silicon carbide whiskers SiCw (Tokawhisker) do not react neither with molten pure magnesium [53] and lithium [52], nor with molten Mg12Li and Mg8Li1Al alloys, even though there might exist some thermodynamic potential for the formation of magnesium silicide Mg2Si and lithium carbide Li2C2 [54, 55]. It is believed that the wetting of SiCw

with magnesium melt is promoted by the redox decomposition of thin surface silica film [56] and a similar effect might be anticipated if molten Li and/or Mg-Li alloy are the case. Polycrystalline SiC fibers (Nicalon) containing some amount of elemental Si and C readily absorbed lithium during contact with molten Mg12Li alloy and lithium vapor [54].

Likewise, the chemical vapor deposition (CVD) prepared SiC fibers (Sigma) were significantly affected with Mg12Li alloy and lithium vapors, predominantly along grain-boundaries, nevertheless thin surface Y2O3 layer provided good protection of the fibers [54].

Reaction between elemental carbon and magnesium does not take place because magnesium carbides MgC2 and Mg2C3 cannot be formed immediately from elements [57]. Consequently, carbon fibers are

(53)

considered stable during contact with magnesium; nevertheless some chemical change (nitrogen extraction?) and significant drop of the tensile strength of PAN-based carbon fibers occur during their long-term exposition to saturated magnesium vapors at 450-700℃ [58]. On the other hand, carbon species react readily with lithium above ~ 400℃ to form lithium carbide Li2C2 whereas below this temperature lithium intercalates LiCn (n>6) are formed via fast interplanar migration of Li ions between hexagonal planes (~10-6 cm2 S-1 at 390℃ [59]). Contact of carbon fibers with molten lithium [52] and Mg11Li alloy [54] caused their massive destruction apparently due to the fast formation of Li2C2 in the fiber bulk accompanied with huge volume change (∆V~100%) wherein the intercalation with lithium ions was believed to take place in the initial stage of the fiber attack [60, 61]. Surface CVD layer of pyrolytic carbon (PyC) essentially retards the kinetics of the attack of carbon fibers with Mg-Li melt, presumably due to the ordered onion-like PyC structure consisting of large blocks of graphene hexagons oriented parallel to the fiber axis [60] that are hardly cross-passable for Li+. [62].

Moreover, dense PyC deposit can plug the micro-pores on the fiber surface thus eliminating the network of interconnected micro-porosity in

(54)

carbon fibers as the short-circuit diffusion paths for Li penetration into fiber bulk.

Alumina products undergo displacement redox reaction with both magnesium and lithium to form simple oxides MgO and Li2O, respectively. Commercial δ- Al2O3 fibers (Saffil) were found to be susceptible to the attack with molten MgLi alloy when both Li2O and MgO were detected within entire fiber cross-section [63], whereas at the fiber/matrix interface Li2O was predominantly formed [31]. It was concluded that interfacial Li2O entered in the next step the reaction with alumina fiber via topotactic insertion of Li. into cation vacancies of δ- Al2O3 lattice [64], thus promoting the formation of strong interfacial bond in corresponding MgLi matrix composites [31] (Fig 2.13). Note that during the infiltration with pure Mg similar solid state reaction of MgO with δ- Al2O3 did not take place. Significant bulk attack of δ- Al2O3 fibers with MgLi melt was found only after long-term fiber/melt contact wherein structurally incoherent oxidic products (MgO, Li2O) were accumulated in the bulk of fibers to cause their embrittlement [65].

(55)

Fig. 2.13 Model of the formation of interfacial bond in δ-Al2O3/MgLi composites during melt infiltration process in which Li. ions are inserted topotactically into cation vacancies on the surface of δ-Al2O3 lattice [31]

(56)

CHAPTER 3 EXPERIMENTAL PROCEDURES

Figure 3.1 illustrates the overall experimental framework of this research in detail.

3.1 Material Preparation

The experiments were conducted using the cast type of magnesium alloys with different amounts of lithium, including Mg-6%Li-1%Zn, Mg-9%Li-1%Zn and Mg-10%Li-1%Zn alloy, namely LZ61, LZ91 and LZ101, respectively. The proceeding alloys were cast in order to produce a two-phase structure containing of α and β solid solution, which combined moderate strength and excellent ductility for forming processes.

The raw materials, magnesium and lithium with 3N grade, and zinc with 4N grade were melted to manufacture the LZ series alloy for this study.

A special melting procedure was used to limit the effect of lithium oxidation and losses on the alloy preparation. These Mg-Li alloys were melted in a high-frequency electric induction furnace with a vacuum capability, in an inert argon gas. Degassing of the crucible and its content was conducted under vacuum at 250℃ for 30 min in order to remove most of the absorbed oxygen and moisture. Heating in low carbon steel

(57)

crucible to a temperature between 710-750℃ depending upon the compositions of the Mg-Li alloys, magnesium and alloying element - zinc were melted to complete liquid phase in advance. After this procedure, the main alloying element, pure lithium, was added to the alloy melt by a successive addition method to obtain the nominal compositions. Still standing for 15 minutes, the melts were poured into a block or cylindrical mould in the chamber with a pressured argon gas atmosphere. Depending on the experimental requirement, there were two types of castings for decision: first it was a 37mm-thick flat slab; second that was a cylindrical billet which made to a diameter of 200mm (8”) and a length of 500mm, were cast in different steel moulds. The photography of LZ61, LZ91 and LZ101 alloys in as-cast condition was exhibited in Fig. 3.2.

The composition of these cast alloys were analyzed using an induction coupled plasma-atomic emission spectrometer (ICP-AES) for the lithium content, and a spark-optical emission spectrometry (Spark-OES) for the others elements. The major composition analyzed of various Mg-Li alloys was listed in table 3.1. Optical micrographs of these as–cast alloys, as shown in Fig.3.3, Fig 3.4 and Fig. 3.5, respectively, revealed that they were comprised of a mixture of α and β phases (bright

(58)

and dark regions, respectively).

Fig. 3.1 The overall experimental framework of this study

(59)

Table 3.1 Chemical analysis results of as cast magnesium-lithium alloys, in weight percent

Sample/element Li Zn Mg

LZ91 (T=37mm) 9.6 1.10 Remainder

LZ61 (φ=200mm) 5.8 0.46 Remainder

LZ91 (φ=200mm) 9.2 0.47 Remainder

LZ101 (φ=200mm) 10.2 0.49 Remainder

Fig. 3.2 The LZ61, LZ91 and LZ101 alloys, in the diameter of 200mm and as-cast condition, were successfully produced for this work.

(60)

Fig. 3.3 The optical microstructure of as-cast LZ61 alloy, comprising α and β phases (bright and dark regions, respectively).

Fig. 3.4 The optical microstructure of as-cast LZ91 alloy, comprising α and β phases (bright and dark regions, respectively).

(61)

The optical microstructure of as-cast LZ101 alloy, comprising α

Fig. 3.5

and β phases (bright and dark regions, respectively).

(62)

3.2 Thermo-mechanical Processing

Two processing methods were used, as shown in Fig. 3.1. In the first method initial 37mm-thick flat slab was thinned down to a final 2mm with a total reduction of around 94.6%; applying warm rolling, the roller was preheated to 160°C, about 5% reduction in thickness for each pass due to the inhomogeneous cast structure.

After casting process, the LZ91 alloy was studied with reference to three considerations - aging conditions, cold rolling and/plus subsequent annealing. Before aging treatment, LZ91 was maintained in silicon oil at 300°C for 30 minutes and quenched in water. Aging treatments were conducted in a furnace at 50 or 100°C. The cold rolling was investigated using specimens with four reduction ratios (20, 40, 60 and 80%) from the as-rolled 2mm thick plate to examine the effect of work hardening. These cold-rolled specimens were further annealed at 100, 150, 200 and 250°C for 30 minutes and air cooled.

The second processing method, for 200-mm cylindrical billet, was similar to the first except that an extrusion substituted for the warm rolling in this thermal fabrication study. Prior to the thermo-mechanical

(63)

heat treatments for uniform characterization at 300°C for around 12 hours.

The homogeneous billets were subsequently direct extruded to produce plates at 200°C, with an exactly thickness of 3 mm plates to change the shape and distribution of phases in these alloys, as shown in Fig.3.6. Fig.

3.7 illustrated that the 4H cold rolling was studied through use of samples with varied reduction ratios from the previous plates of extrusion and homogeneity. When the samples were applied approximately 60-85%

reduction (depending the alloy composition), the intermediate anneal was performed in an atmosphere of mixed Ar and H2 at 250°C for 3 hours, to produce a completely soft structure for future thinning processing. The cold rolled and annealed specimens with a final thickness of 0.6 or 0.17 mm were utilized for the microstructural analysis, mechanical property test and forming limit test.

(64)

Fig. 3.6 The Mg-Li plate of 3mm in thickness was produced by direct extrusion at the temperature of 200℃.

Fig. 3.7 The coiled Mg-Li sheet in the thickness of 0.6mm and 0.17mm was obtained using 4H precision rolling mill at ambient temperature.

(65)

3.3 Microstructural Characterization

The microstructural analysis of magnesium-lithium alloys was performed using an optical microscope (OM), a 400kV high-resolution transmission electron microscope (HR-TEM, JEM-4000FX) and a 200 kV field-emission TEM (FE-TEM, JEM-2010F). X-ray diffraction (XRD) and X-ray energy dispersive spectrometry (EDS) were utilized to identify any second phase presented in the structure. Samples for OM observations were etched using a reagent containing 6 g picric acid, 5 ml acetic acid, 10 ml distilled water and 100 ml ethyl alcohol for ~30S [66].

Foils for TEM observations were mechanically polished to a thickness about 0.1 mm, and then thinned to perforation by an ion milling method.

3.4 Mechanical Properties Test

Micro-hardness and tensile strength tests were conducted to evaluate mechanical behaviors. The sheet type tensile specimens were prepared according to the ASTM B-557 standard procedure with the gauge length and width were 25mm and 6mm, respectively. Uniaxial tension tests were carried out in the directions of 00, 450 and 900 to the rolling direction at

(66)

room temperature, and at a strain rate of 1×10-3S-1 using an Instron 4469 testing machine. The true stress-true strain data derived from the tensile tests were applied, to calculate n values, strain hardening exponent, for various magnesium-lithium alloys. The hardness of the aging and cold-rolled specimens was measured with a Micro-Vickers or a Rockwell scale tester. The hardness values were calculated by the average of 3 different random indentations.

3.5 Formability Evolution of Mg-Li Alloys

Circle grid analysis was used in this study to measure the strain distributions and thus to enable an analysis of the formability of the sheet metal by plotting the measured strains on the forming-limit diagram (FLD). Since Keeler and Backofen [67] introduced the concept of FLD in 1963, it has been a widely adopted criterion for fracture prediction in sheet-metal forming. In the present work, to determine the FLD of LZ91 and LZ101, stretching tests were performed for sheet specimens with different widths ranging from 10 mm to 90 mm in increments of 10 mm, using a semi-spherical punch of 50 mm in diameter. As the width of the

(67)

ranging from 10 mm to 70 mm in increments of 10 mm were also applied to the FLD tests. The FLD test mold and its geometric dimensions were shown in Fig. 3.8 and Fig. 3.9 respectively. The specimens were first electrochemically etched with circular grids that would be deformed into ellipses after being stretched. Before being stamped, the critical area of the sheet blank was imprinted with 2.5-mm diameter circles with 3-mm spacing between their centers. To imprint the circles the critical area of the sheet-blank was first cleaned using a special cleaner, then stencil with the correct grid pattern was placed in position on the part. Using the electrolyte as a conductor, the area covered by the stencil was marked with the grid pattern by an etching process.

The engineering strains measured in the major and minor axes of the ellipse are termed the ‘major strain’ and ‘minor strain,’ respectively, as indicated in Fig. 3.10. They are also the principal strains on the planes where the strains are measured. The major and minor strains measured in the location closest to the fracture for each specimen were recorded and were then plotted against one another with the major strain as the ordinate.

The curve fitted to the strain points was defined as the forming limit curve, also termed the ‘failure curve.’ Considering the safety factor for

(68)

the design purpose, a 10% off-set downward of the failure curve is adopted as the design curve.

Punch

Blank holder Die

Fig. 3.8 The FLD test moulds, including semi-spherical punch, die and blank holder, for the forming limit test.

(69)

Unit:mm

Fig. 3.9 The schematic diagram showed the main dimension of the hemi-sphere punch drawing mould.

Fig. 3.10 The Measurement of principal major and minor strains for the FLD test.

(70)

CHAPTER 4 RESULTS AND DISCUSSION

4.1 As-cast LZ91 Microstructures

The microstructure of the as-cast LZ91 under microscope included a β matrix plus a distributed α phase in lath form with a width of ~20µm and a length of ~100µm, as given in Fig. 4.1. The dark matrix in the SEM image is identified as the Li solid solution of low atomic number while the bright α phase is the Mg solid solution with a higher atomic number.

EDS demonstrated that the α solid solution dissolves more Zn than does β.

The former phase had a measured micro-hardness of Hv52, higher than that of the phase, which was Hv48. The volume fraction of each phase by image analyzer is about 27% α and 63% β, it is quite consistent with the value estimated from the binary magnesium-lithium phase diagram (Fig.

2.1).

The as-cast microstructure of LZ91 comprises a mixture of α and β phases (bright and dark regions, respectively) with some dispersed fine particles in the α phase, as presented in Fig. 4.2. The top α phase is almost free of dislocations, whereas the β phase on the bottom has a high density of dislocations. The α phase has the HCP structure, so it tends not to

(71)

deform plastically during the preparation of the TEM specimen. In contrast, plastic deformation during the preparation of the specimen increases the density of dislocations in the softer β phase, which has the BCC structure. Fig. 4.3 displays the two shapes of the ultrafine dispersed particles - spherical and faceted. All such particles are smaller than 40 nm, as shown in Fig. 4.3 The ultrafine particles were analyzed by FE-TEM equipped with EDS. The size of the electron beam was reduced to approximately 10 nm to ensure the accuracy of the analysis of the ultrafine particles. Figure 4.4 reveals that the rounded particles contain Zn and O, indicating that they should be ZnO. EDS analysis of the faceted particles indicated that the Mg and O elements were present, as presented in Fig. 4.5, suggesting that the faceted particles were the oxide, MgO.

The oxidation potentials of Mg and Li exceed that of Zn, so when oxides formed during melting, Li oxides should be observed in the as-cast ingot.

However, no particle was found to be an Li oxide in this investigation.

Therefore, the oxide detected in the as-cast ingot may have been present before or at melting process. The presence of dispersive nano-sized ZnO and MgO oxides could strengthen the LZ91 alloy.

(72)

Fig. 4.1 Optical microstructure of the as-cast LZ91 alloy

Fig. 4.2 α phase on the top of the photograph is almost free of dislocations, while the β phase, on the bottom has a high density of dislocations.

(73)

Fig. 4.3 TEM analysis of as-cast specimen; oxide particles were distributed in the α matrix, and were identified as ZnO (round particles) and MgO (polygonal particles).

(74)

Fig. 4.4 Granular particles (a) and matrix (b) identified by EDS.

The particles contain Zn and O, and are ZnO.

(75)

Fig. 4.5 Faceted particle (a) and matrix (b) identified by EDS.

The particles contain Mg and O, and are MgO.

(76)

4.2 As-rolled and Solution-treated LZ91 Microstructures

Figure 4.6 demonstrates that the as-rolled structure exhibits fibrous rolling texture, which is the typical plastically deformed structure. The states of the solution-treated 2mm as-rolled plate differ from the previous state of the plate that was not heat treated, mainly in the elimination of the fibrous rolling texture (Fig. 4.7). Unlike the as-cast, the α phase is aligned by post-rolling and refined in size down to a few microns. The β matrix now exhibits clear equi-axial structure with grains of size ~30µm.

Figure 4.8 displays the TEM microstructures of the rolled and solution-treated (as-quenched) specimen. The TEM micrograph presents the same oxide particles as were observed in the as-cast ingot, but aligned by heavy working. A nano-beam diffraction approach was then employed to elucidate the structure of the ultrafine particles. The nano-beam diffraction pattern demonstrates that ZnO has a Wurtzite structure (a=0.325 nm, c=0.521), as shown in Fig. 4.9. Due to the near lattice parameters of the same structure of ZnO and Mg, the contrast is low at the high magnification. The relationship between the orientations, displayed in Fig. 4.9, also reveals that the ZnO has a coherent structure

(77)

grain boundaries and displays the nearby faceted MgO. The nano-beam diffraction pattern indicates that the structure of MgO is not very coherent with the magnesium matrix, as presented in Fig. 4.11. Fig 4.11(b) and (c) were taken under the incident beam parallel to the direction of [001]MgO

and [1210]α , respectively.

Fig. 4.6 Micro-structure of the as-rolled LZ91 alloy

(78)

Fig. 4.7 Optical micrograph of the as-rolled and solution-treated LZ91 alloy

Fig. 4.8 TEM morphology of the rolled and solution-treated LZ91 specimen

(79)

Fig. 4.9 Nano-beam diffraction pattern of as-cast LZ91 specimen.

(a) TEM image, (b) diffraction pattern of the matrix, and (c) differaction pattern of the round ZnO particle.

(80)

Fig. 4.10 TEM images of solution-treated LZ91 alloy. Faceted particles are present in the grain boundary area.

(81)

Fig. 4.11 Nano-beam diffraction pattern of the polygonal particle.

(a) TEM image, (b) diffraction pattern of the polygonal MgO particle and (c) diffraction pattern of the matrix.

(82)

4.3 Effect of Aging Treatment on LZ91

The solution-treated specimens were then aged at 50 and 100°C.

Changes in hardness were plotted against aging time, as presented in Fig.

4.12. Hardness seems not to be influenced by the low temperature aging heat treatment in the early stage, and begins to increase at a treating time of about 3 hours. It took about 10 hours to reach peak hardness for treating at 100℃. However, when aged at 50°C, the corresponding time

was around 100 hours. As reported by Yamamoto et al. [16], peak hardness was also observed at room temperature and 55°C in an dual phase Mg-8.2wt.% Li-4.6wt.% Zn alloy, but it took less than 1 h to reach peak hardness at 55°C, and only softening occurred at temperatures of 75 and 110°C in their study [16]. The optical micrographs of LZ91 aged at 50 and 100°C were shown in Fig. 4.13 and Fig. 4.14, respectively. Prior to aging treatment, the specimens were solution treated at 300℃/30 minutes to eliminate the fibrous rolling texture. The microstructure of α phase was aligned by post-rolling and refined in size down to a few microns. The β matrix exhibited clear equi-axial structure with grains of size ~30-50µm. It appeared that the aging treatment did not affect the

(83)

Figure 4.15 demonstrates the change in the tensile properties during aging heat treatment at two different temperatures in the LZ91 alloy. It shows that aging treatment only a little effect on the variation of strength.

Basically, the dual phase LZ91 is not an age hardenable alloy; it must have been some other reason to cause the variation of hardness in the later stage of aging heat treating. X-ray diffraction was used to analyze the constituents and thus relate the hardness to the microstructure. All aged specimens were analyzed using XRD. Figures 4.16 (a) and 4.16 (b) show typical X-ray spectra obtained from LZ91, suggesting the presence of the two detectable phases - α and β. An extra bump (side band, indicated by arrows in the figures) adjacent to the main peak of α(002) is observed in the X-ray diffraction patterns of specimens aged at 50°C for 100 hours and 100°C for 10 hours. The appearance of the extra bump implies that spinodal decomposition may be responsible for hardening effect [68]. The aging curves, plotted in Fig. 4.12, also indicate that hardness and precipitation were maximal when the extra bump was present. However, the precipitate is an unstable phase because its diffracted peak disappeared in the XRD spectra after aging for a long period.

(84)

Figure 4.17 displays a typical TEM micrograph of the specimen aged at 100°C for 10 h, and it shows the particle-like structure. Although no clear precipitate was observed, the high-resolution TEM micrograph in Fig. 4.18 (a) demonstrated fine fringes in the particle-like structure. The structure is similar to that of short-range structural modulation, perhaps because of spinodal decomposition. Fourier transformation of the image (Fig. 4.18(b)) reveals the sticks that correspond to the modulation structure. The structure was further analyzed using nano-beam EDS and was found to vanish easily when electron beam radiation was applied. It was found not to be stable. Similar structure with finer fringes was also found in the specimen treated at 50°C for 100 hours, as shown in Fig.

4.19.

(85)

Fig. 4.12 The effects of temperature and soaking time on the hardness of the LZ91 alloy during aging heat treating.

(86)

(a)

100µm

(b)

100µm

Fig. 4.13 Optical micrographs of the LZ91 were aged at 50℃

for (a) 1 hour, (b) 3 hours, (c) 10 hours and (d) 100 hours.

參考文獻

相關文件

學博士,現為上海大學文學院教 授,兼任佛光山人間佛教研究院

Secondly then propose a Fuzzy ISM method to taking account the Fuzzy linguistic consideration to fit in with real complicated situation, and then compare difference of the order of

To tackle these problems, this study develops a three-stage approach (i.e., firstly create a correct CAD-oriented explosion graph and then find a graph-based assembly sequence

In order to analyze the tourists’ multi-phase recreation experience of railway tourism, this research discusses the dimensions, including customers’ involvement, perceived

To response the increasing competition between banks and it variation in operation, it is an significant issue to upgrade service quality and consolidate customer satisfaction,..

In the process control phase, by using Taguchi Method, the dynamic curve of production process and the characteristics of self-organizing map (SOM) to get the expected data

Keywords:Long-distance Steep Turn Pipe Impelling Construction Method, Analytic Hierarchy Process (AHP), technology transfer, Delphi Method, assessment standards... 誌

This dissertation utilizes the Kano Two-dimension Quality Model and the Importance-Performance Analysis to investigate the following studies: (1) Travel Motivation and