High Strain Rate Superplasticity of Mg Based Composites Fabricated by Friction Stir Processing

High Strain Rate Superplasticity of Mg Based Composites

Fabricated by Friction Stir Processing

C. J. Lee and J. C. Huang

Institute of Materials Science and Engineering, Center for Nanoscience and Nanotechnology, National Sun Yat-Sen University, Kaohisung, Taiwan 804, R. O. China

The nano-sized SiO2particles were added into the AZ61 Mg alloys via friction stir processing (FSP) to a volume fraction of 5–10%. After four FSP passes, the 10% composites had uniform dispersion of particles and grain size of 0.8mm. This composite exhibited high strain rate superplasticity, with a maximum ductility of 470% at1102_s1_{and 300C or 454% at310}1_s1_{and 400C while maintaining fine grains}

less than 2mmin size. [doi:10.2320/matertrans.47.2773]

(Received May 22, 2006; Accepted September 5, 2006; Published November 15, 2006)

Keywords: Friction stir processing, magnesium composite, high strain rate superplasticity

1. Introduction

Mg based alloys are the lightest structural materials, highly attractive in terms of weight and energy saving. However, the intrinsically poorer ductility of Mg alloys would limit their applications. A number of thermomechanical processing (TMP) methods, such as rolling, extrusion, equal channel angular pressing (ECAP) have demonstrated the feasibility to remarkably improve the ductility and formability.

Magnesium matrix composites, with the addition of ceramic reinforcements could further improve the strength, wear and creep resistance. But they are also inherited with the poor ductility, toughness and workability of Mg materials. High strain rate superplasticity (HSRSP), with a total

elon-gation >200% at a high strain rate greater than102_s1_{, in}

these composites may provide a chance to upgrade the workability and to form complicated shape for industrial applications.

The HSRSP behaviors observed in the Al based compo-sites usually occur at the partial melting temperatures with the help from a controlled amount of the liquid phase between the matrix and reinforcements to accommodate the stress concentration during deformation, as found in, for

example, the Al-Cu-Mg/Si3N4 composites.1) However, the

role of the liquid phase did not seem to be required in the Mg based composites due to the more rapid diffusion ability of

Mg atoms.2)The related HSRSP behaviors of the Mg based

composites3–6)_{are listed in Table 1. The reinforcement sizes}

are all in the micron range. These Mg based composites have all been subjected to secondary TMP to uniformly disperse the reinforcements and to refine the grain size.

Friction stir processing (FSP) is an extension of friction stir welding (FSW) originally developed by The Welding

Institute of UK in 1991.7) _{Mishra et al. first applied this}

process to achieve good superplasticity in the 7075 Al

alloys.8)_{After this, the HSRSP and low temperature}

super-plasticity (LTSP) performance in the related Al alloys9–15)

also could be achieved by FSP thermomechanical treatment. Moreover, Mishra showed the possibility to fabricate surface

composites in 5083 Al.16)Follow-up studies have extended

the FSP technique to fabricate many fine grained alloys,17)

composites,18)_{and intermetallic compounds.}19)_{The sizes of}

the inserted reinforcements have been varied from micron or submicron to the nano scales. FSP seems to be an effective

tool to disperse the nano fillers in the matrix.18)_{In this paper,}

we report the results of HSRSP in the AZ61 Mg based

composites, with nano-sized SiO2particles, prepared by FSP.

2. Experimental Methods

The AZ61A billets used in this study have a chemical composition in mass percent of Mg-6.02%Al-1.01%Zn-0.30%Mn. This alloy is a solution hardened alloy with minimal precipitation. The billet possessed nearly equiaxed

grains around 75mm (based on the linear line intercept

method from three cross-sectional planes). The billet was cut as rectangular samples 60 mm in width, 130 mm in length and

10 mm in thickness. The amorphous SiO2 nano particles

applied are nearly equiaxed in shape and have an average

diameter hdi 20nm and purity 99:9%, and density of

2.65 Mg/m3_.

The simplified FSP machine used in this experiment was a modified form of a horizontal-type milling machine, with a 5 HP spindle. The fixed pin tool was 6 mm in diameter and

6 mm in length. The shoulder diameter was 18 mm, and a 2

tilt angle of the fixed pin tool was adopted. The pitch distance was 1 mm. The advancing speed of the rotating pin was kept constant in this study to be 45 mm/min, with a fixed pin

rotation of the 800 rpm. To insert the nano SiO2particles, one

or two deep and shallow grooves6mm depth and 1.25 mm

wide were cut (termed as the 1D or 2D specimens), in which

the nano SiO2 particles to the desired amount were filled

before FSP. The volume fraction of the SiO2 nanoparticles

inserted in the 1D and 2D specimens within the stirred zone is estimated to be around 5 and 10 vol%, respectively. The groove(s) were aligned with the central line of the rotating pin. The detailed description of fabricating Mg based

composites has been presented in our previous paper.18)

The samples with SiO2 particles of one or two deep

grooves and four FSP passes (4P) were machined into tensile specimen, as shown in Fig. 1, with the gage length centered in the central region of the stirring zone. The tensile *_{Corresponding author, E-mail: [email protected]}

specimens have a gage length 3 mm and width 2 mm, and are

ground and polished to a thickness of 1:2mm. Constant

crosshead speed tensile tests were conducted using an universal testing machine equipped with a three-zone

furnace. The temperature was controlled with 2_{C. The}

AZ61 alloy processed by FSP for four passes without any

SiO2 filler was also conducted the same tensile tests for

comparison. The microstructure of FSP specimens was examined by scanning and transmission electron microscopy (SEM and TEM).

3. Results and Discussions

3.1 Microstructure of Mg based alloys and composites

According our previous results,18)_{the SiO}

2particles in the

groove(s) could be uniformly dispersed in the stirring zone after four FSP passes. In the stirred zone, the grains of Mg based composites are fine and equiaxed, as shown in Fig. 2. The average grain size of 1D4P and 2D4P samples was close

to 1.8mmand 0.8mm, respectively. The same FSP parameters

(800 rpm and 45 mm/min) were also applied to the AZ61

alloys without SiO2, and the grain size yielded was 7–8mm,

as shown in Fig. 3. It is apparent that the nano-sized particles in the groove(s) played a key role in restricting the rapid grain growth during FSP.

In terms of the critical microstructure parameter of

L¼ ðhdi=2Þð2=3VfÞ1=220)wherehdi is the average particle

diameter and Vf is the volume fraction, it is reasonable that

the 1D4P grain size is higher than the 2D4P one due to the lower nano filler volume fraction. Theoretically, if all of the

20 nm SiO2 nano particles are completely and uniformly

dispersed, the theoretically estimated grain size should vary

within 0.1–0.2mm. Nevertheless, the average SiO2cluster for

1D4P and 2D4P samples is 190 and 150 nm, respectively. Due to a certain level of local clustering, the average

Materials Process Particle size,

D/mm

Temp, T/_C

Strain rate, _

"

"/s1

Elongation,

% m value

_Ref.

Mg-A1/Mg2S/28p Rapidly solidification +_Extrusion 0.7 515 1101 370 0.5 3)

Mg-Zn/Mg2S/28p Rapidly solidification + 0.8 440 1101 290 0.5 3)

Extrusion

ZK60/SiC/17p PM + Extrusion 2 450 1101 _{360 0.5 4)}

Mg-5Zn/TiC/20p Vortex casting + 2–5 470 6:7102 _{340 0.33 5)}

Extrusion + rolling

Mg-5Al/AlN/15p Vortex casting + 0.72 450 5101 _{200 0.4 6)}

Extrusion + rolling

Fig. 1 The sampled tensile specimen from the FSP stir zone with the tensile loading direction perpendicular to the FSP travel direction.

(a) (b)

1 µm

resulting grain in the FSP stirred zone could only be refined to

0.5–2mm.

The grain size variation upon static annealing at 350_{C for}

the AZ61 alloy and composites are shown in Fig. 4. The grains of the 1D4P and 2D4P composites could still be

maintained within 1–3mmeven after 10 h static annealing. In

comparison, the grains of the FSP AZ61 alloy grew rapidly

from 7 to 44mmafter 5 h annealing. It is apparent that the

nano SiO2 did play an effective role in controlling the grain

growth during FSP and subsequent static annealing or superplastic deformation at elevated temperatures.

3.2 Superplasticity behavior

Systematic tensile tests were performed on the FSP 1D4P and 2D4P Mg based alloys and composites. Figure 5 shows the variation of tensile elongation as a function of strain rate

at 250–400_{C. It can be seen that the optimum strain rate for}

superplasticity performance increased with increasing load-ing temperature. For example, the optimum strain rate at

250C for the 2D4P samples with an elongation of 370% is in

the range of104–103s1; and the optimum strain rate at

300 and 400_{C with an elongation of 470 and 454% could be}

raised to1102_and3101_{, respectively, exceeding the}

criteria for HSRSP. In comparison, the highest superplastic elongations and the optimum strain rates of the 1D4P samples were all inferior to those of the 2D4P counterparts.

As for AZ61 alloys modified by FSP, it did not perform the same HSRSP behavior at the higher temperatures, such as

350 or 400_{C; nevertheless, the elongation could reach 235%}

at 300_{C and110}4_s1_{, as shown in Fig. 5. These results}

indicate that the current lower FSP travel speed of 45 mm/ min coupled with the medium pin rotation of 800 rpm would

result in relatively coarser grain sizes of 7–8mm, unfavorable

for HSRSP deformation. Meanwhile, the rapid grain growth

20 µm

Fig. 3 Optical micrograph of the AZ61 alloy processed by FSP for four passes.

0 8

0.1 1 10 100

Grain Size,

d

/

µ

m

Annealing Time,

t

/h

2D4P

2 4 6 10

Fig. 4 Variation of grain sizes of the FSP AZ61 alloy and composites at 350_{C during static annealing.}

10-5 10-4 10-3 10-2 10-1 100 0

100 200 300 400 500

(a)

Elongation (%)

Strain Rate, /s-1

400oC 350oC 300oC 250oC

10-5 10-4 10-3 10-2 10-1 100 0

100 200 300 400 500

(b)

Elongation (%)

Strain Rate, ε. /s-1

400oC 350oC 300oC 250oC

10-5 10-4 10-3 10-2 10-1 100 0

100 200 300 400 500

Elongation (%)

Strain Rate,

/s

400

C

350

C

300

C

250

C

(c)

ε.

ε

of the unreinforced AZ61 alloy at elevated temperatures to

30–50mm would further reduce the possibility for HSRSP.

Only at relatively lower temperature such as 300_{C or below,}

when rapid grain growth is unlikely to occur, superplasltic behavior can be retained.

The representative stress-strain curves of the FSP

compo-sites tested at 350_{C are shown in Fig. 6. The flow stresses of}

the 2D4P specimens at 350_{C did not exhibit strain}

harden-ing, maintaining relatively stable flow stress of 15 and

30 MPa at 1102 _{and 110}1_{. In contrast, the 1D4P}

samples showed slight strain hardening, especially in range

of1103–1102s1. The flow stress levels of the 2D4P

specimen are typically lower than those of the 1D4P one at the same rate, implying the more smooth operation of grain boundary sliding and thus the lower flow stress.

Figure 7 shows the 2D4P specimens deformed at 1

102_s1_and3101_s1_{. The 2D4P specimens deformed at}

3101_s1 _{and 350–400C show the neck-free gauge}

section characteristic of superplastic flow. But the specimens

show a certain level of local necking at 1102_s1 _and

400_{C. It might be a result of cavitation at high temperatures}

and lower strain rates with enough time and energy to make

vacancy diffusion for cavity nucleation.21)

The flow stresses at a true strain of 0.3 of the 2D4P

composites loaded at 250–400_{C as a function of strain rate}

are plotted in Fig. 8(a). The apparent m-value was estimated about 0.3–0.4 from the data in Fig. 8(a). With the consid-eration of the threshold stress, the true strain rate sensitivity

was calculated to be around 0.4–0.5, suggesting that the grain boundary sliding might be the dominant deformation

mech-anism. The apparent activation energyQa can be evaluated

according to the equation

Qa¼ R@ðln _""Þ

@ð1=TÞ

¼constant

; ð1Þ

where ""_, R andT are stain rate, gas constant and absolute

temperature, respectively. Taking ¼20MPa, Qa was

estimated from the slope of the plot of ln _"" against 1000/

RT, as shown in Fig. 8(b). The apparent activation energy

Qa over 250–400C is 106 kJ/mol, which is between the

activation energy for boundary self-diffusion of the Mg atoms (92 kJ/mol) and that for lattic self-diffusion of the Mg

atoms (135 kJ/mol).22)_{It is suggested that the}

accommoda-tion mechanism may be dislocaaccommoda-tion slip plus climb, the latter

controlled by grain boundary diffusion. Watanabeet al.2)_and

Mabuchiet al.3)_{have also reported the same deformation and}

accommodation mechanism in the Mg based composites made by other processing routes.

It is known that grain size strongly affects the optimum

superplastic strain rate by the relation below:2)

_ " "¼A

T b d

p

E

n

D; ð2Þ

0.0 0.3 0.6 0.9 1.2 1.5 1.8

0 10 20 30 40 50 60

(a)

True Stress,

σ

/MPa

True Strain

1x10-3 s-1 1x10-2 s-1 1x10-1 s-1

0.0 0.3 0.6 0.9 1.2 1.5 1.8

0 10 20 30 40 50 60

(b)

True Stress,

σ

/MPa

True Strain

1x10-3 s-1 1x10-2 s-1 1x10-1 s-1

Fig. 6 True stress and strain curves for (a) 1D4P and (b) 2D4P specimens at 350_C.

(a) 1x10-2 s-1

(b) 3x10-1 s-1

Undeformed

250oC, 135%

300oC, 470%

350oC, 331%

400oC, 125%

300oC, 100%

350oC, 391%

400oC, 454%

where A is a materials constant,the flow stress, E the elastic modulus, b the Burgers vector, D the diffusion coefficient, p

the grain size exponent, and n the stress exponent (¼1=m).

The 1D4P and 2D4P composites possessed the average grain

size around 2 and 0.8mm, respectively. The grain size

difference leads to the different optimal superplastic strain. Meanwhile, at the same strain rate, the 2D4P composites show the lower flow stresses and more smooth operation of grain boundary sliding, following the trend predicted by eq. (2).

3.3 Topography of deformed specimens

Figure 9 shows the surface topography of the 2D4P

specimen deformed at 350_{C and 110}1_s1 _{to a true}

strain of 1.37. It reveals the evidence of grain boundary sliding. The grain size can be seen to be maintained in the

range of 1–2mm. The inserted nano particles have played an

effective role in restraining grains growth not only during FSP but also during the subsequent static annealing and superplastic deformation at elevated temperatures.

The fracture surface of the 2D4P specimen deformed at

350_{C and 110}1_s1 _{to failure, showing the HSRSP}

elongation of 410%, is depicted in Fig. 10. The fracture surface of the HSRSP Al based composites often exhibit some long filaments after high strain rate deformation due to

the partial melting at grain or interface boundaries.23)Such a

phenomenon is not as apparent in the current Mg based

composites; but there are scratch-like marks on some exposed grain boundaries and occasional short grass-like appearance. The role of partial melting does not seem to be pronounced in the current Mg composites.

4. Conclusions

The four-pass FSP is demonstrated to fabricate an AZ61

Mg based composites with up to 10% of nano SiO2particles

measuring 20 nm. The nano particles can be reasonably uniformly dispersed, and restrict grain growth during FSP and subsequent static annealing or superplastic loading at

250–400_{C, limiting the grain size within 0.8–2.0mm.}

Satisfactory high strain rate superplasticity over 450% elongations can be achieved in the 2D4P composite as

loaded at 300–400_{C and 10}2_–101_s1_{. The strain rate}

sensitivity (0:5) and the activation energy (106kJ/mol),

coupled with the surface topography evidence, suggest that the dominant deformation mechanism is grain boundary sliding accommodated by dislocation slip plus climb, and the latter is controlled by grain boundary diffusion.

10-5 10-4 10-3 10-2 10-1 100 101 100

101 102

250oC 300oC 350oC 400oC

Flow Stress,

σ

/MPa

Strain Rate,

/s

(a)

0.17 0.18 0.19 0.20 0.21 0.22 0.23 -5.5

-5.0 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5

ln(Strain Rate)

,

ln

ε

.

/s

-1

1000(RT)

/mol J

Q

~ 106 kJ/mol

(b)

Fig. 8 The analyses on the 2D4P composites samples for (a) flow stress against strain rate cures and (b) apparent activation energy.

1 µm

Fig. 9 SEM micrograph showing the surface topography of the 2D4P specimen deformed at 350_{C and110}1_s1_to"¼1:37.

1 µm

Acknowledgement

The authors are gratefully acknowledge the sponsorship by National Science Council of Taiwan, ROC, under the project no. NSC 93-2216-E-110-021.

REFERENCES

1) M. Mabuchi, K. Higashi and T. G. Langdon: Acta Metal. Mater.42

(1994) 1739–1745.

2) H. Watanabe, T. Mukai, M. Mabuchi and K. Higashi: Acta Mater.49

(2001) 2027–2037.

3) M. Mabuchi and K. Higashi: Philo. Mag. A74(1996) 887–905. 4) T. G. Nieh and J. Wadsworth: Scripta Metal. Mater.32(1995) 1133–

1137.

5) S. W. Lim, T. Imai and Y. Nishida: Scripta Metal. Mater.32(1995) 1713–1717.

6) T. Imai, S. W. Lim, D. Jiang and Y. Nishida: Scripta Mater.36(1997) 611–615.

7) W. M. Thomas, E. D. Nicholas, J. C. Needham, M. G. Church, P. Templesmith and C. J. Dawes: Intl Patent No. PCT/GB92/02203. 8) R. S. Mishra, M. W. Mahoney, S. X. Mcfadden, N. A. Mara and

A. K. Mukherjee: Scripta Mater.42(2000) 163–168.

10) Z. Y. Ma, R. S. Mishra, M. W. Mahoney and R. Grimes: Mater. Sci. Eng. A351(2003) 148–153.

11) H. G. Salem, A. P. Reynolds and J. S. Lyons: Scripta Mater.46(2002) 337–342.

12) Z. Y. Ma, R. S. Mishra and M. W. Mahoney: Acta Mater.50(2002) 4419–4430.

13) Z. Y. Ma, R. S. Mishra and M. W. Mahoney: Scripta Mater.50(2004) 931–935.

14) Z. Y. Ma and R. S. Mishra: Scripta Mater.53(2005) 75–80. 15) I. Charit and R. S. Mishra: Acta Mater.53(2005) 4211–4223. 16) R. S. Mishra, Z. Y. Ma and I. Charit: Mater. Sci. Eng. A341(2003)

307–310.

17) C. I. Chang, C. J. Lee and J. C. Huang: Scripta Mater.51(2004) 509– 514.

18) C. J. Lee, J. C. Huang and P. J. Hsieh: Scripta Mater.54(2006) 1415– 1420.

19) C. H. Chuang, J. C. Huang and P. J. Hsieh: Scripta Mater.53(2005) 1455–1460.

20) A. J. Ardell: Metall. Mater. Trans. A16(1985) 2131–2165. 21) C. J. Lee and J. C. Huang: Acta Mater.52(2004) 3111–3122. 22) H. J. Frost and M. F. Ashby: Deformation Mechansim Maps (Pergamon

Press, Oxford, 1982) p. 44.