Hall­Petch Tensile Yield Stress and Grain Size Relation of Al­5Mg­0.5Mn Alloyin Friction-Stir-Processed and Post-Thermal-Exposed Conditions

Chun-Yi Lin+1, Truan-Sheng Lui+2, Li-Hui Chen and Fei-Yi Hung

Department of Materials Science and Engineering, National Cheng Kung University, Tainan 70101, Taiwan, ROC

Friction stir process (FSP) and post thermal exposure were carried out to a microstructure-thermostable cast Al­5Mg­0.5Mn alloy toexamine the grain size dependence of tensile yield stress by using Hall­Petch relation, ·y = ·0 + kyd¹1/2. The measurement of preferredcrystalline orientations and grain boundaries characteristic is used to explain the difference of the Hall­Petch parameters. FSP produces finegrain structure and, for the obtained average grain size from 3.7 to 12.8 µm, near-random crystalline orientations and certain proportion of rigidcoincident site lattice boundaries (CSLs) is detected. Two-step post thermal exposure leads to a quick development of rotated ©001ª typecrystalline orientations and the obtained coarse grains with average size from 229 to 508µm are also commonly surrounded by general highangle boundaries. The calculation of Taylor factor and CSLs proportion reveals that stronger texture and grain boundary hardening effects existin FSPed specimens than in thermal exposed specimens. These items are the major factors leading to larger ·0 and ky in FSPed specimens than inthermal exposed specimens. [doi:10.2320/matertrans.M2013334]

(Received August 28, 2013; Accepted October 29, 2013; Published January 25, 2014)

Keywords: friction stir process, thermal exposure, aluminum­magnesium alloy, Hall­Petch relation

1. Introduction

Friction stir welding1­3) (FSW) emerged as a solid-statejoining process produces defect-free joints with bettermechanical properties than fusion welding processes.Friction stir processing (FSP1,4)), which is essentially similarto FSW, can be applied to surface modification. Fine andequiaxed grain structure generated from combination offrictional heat and intense plastic deformation duringFSW/FSP. Previous studies5,6) on 5083 Al alloy, one ofthe widely investigated non-heat-treatable alloys, has shownthat the level of grain refinement can be controlled byprocess parameters and Hall­Petch equation can be used todescribe the grain-size dependence of tensile strength orhardness.

Hall­Petch equation7,8) correlates tensile yield stress (·y)and average grain size (d) as

·y ¼ ·0 þ kyd�1=2 ð1Þ

where ·0 is friction stress and ky is stress intensity coefficientassociated with the stress required to extend dislocationactivity into adjacent unyielded grains. Hall­Petch relationhas been investigated in 5083 Al alloy in several previousstudies. For the conditions of FSW/FSP,5,6) microhardnessdata showed that kH slope varies from 14 to 53Hv·µm¹1/2

with a variation of average grain size from 0.2 to 25 µm.These values can be converted to ky slopes as 46 to173MPa·µm¹1/2 by taking Hv = 3·y.9) On the other hand,the super-high rolled and consequently annealed specimenswith the grain size between 3.1 to 12 µm had a ky slope ofabout 188MPa·µm¹1/2.10) Above-mentioned reviews showthat different intense straining processes or annealingsituations can lead to a wide range of ky values in 5083 Alalloy. However, the influential factors on Hall­Petch pa-rameters in 5083 Al alloy are still not clearly clarified yet.According to several associated reports,4­6,11­13) FSP produc-

es quite fine grain structure through intense thermo-mechani-cal behavior and the subsequently post thermal exposurecauses an extreme grain growth phenomenon. Wide range ofgrain size is convenient for Hall­Petch investigation; for thisreason, a fully annealed and microstructure-thermostableAl­5Mg­0.5Mn cast alloy of compositions close to 5083wrought Al alloy was friction stir processed and post thermalexposed to analyze the governing factors of Hall­Petchrelation parameters, ·0 and ky as well.

2. Experimental Procedure

2.1 FSP, thermal exposure and tensile testNon-heat-treatable Al­5Mg­0.5Mn cast slab was com-

posed of 4.86% Mg, 0.58% Mn, 0.27% Fe, 0.07% Si, 0.06%Cr, 0.04% Cu (in mass%), and balancing Al. The cast slabwas machined to 4mm in thickness and friction-stir-processed by using a SKH-51 steel tool. The tool dimensionwas 15mm in shoulder diameter, 5.5mm in pin diameter, and3mm in pin length. The fixed parameters were 0.55mm/s inprocess speed, 1° in tool tilt angle and 70MPa in loadingpressure. Three tool rotational speeds, namely 450, 650 and850 rpm (labeled as 450F, 650F and 850F) were chosen toproduce three kinds of fine-grain specimens.

450F specimens were two-step thermally exposed to obtaindifferent levels of coarse-grained specimens. Strain-energy-release treatment took place at 300°C for 0, 1, and 6 h in thefirst step and subsequent thermal exposure was carried outat 500°C for 1 h in the second step. Which were labeledas TTEx, where x = 0, 1 and 6 means the holding hour ofstrain-energy-release treatment.

Following FSP and post thermal exposure, tensile testwas applied at room temperature with an initial strain rate,1.67 © 10¹3 s¹1. Flat tensile specimens in I-shape weremachined with gage longitudinal axis parallel to theprocessing direction. The gage length of tensile specimenswas 10mm and cross-section was 2mm in thickness and2.2mm in width, which was covered within the markedregion of stir zone (SZ) as shown in Fig. 1.

+1Graduate Student, National Cheng Kung University+2Corresponding author, E-mail: z7408020@email.ncku.edu.tw

Materials Transactions, Vol. 55, No. 2 (2014) pp. 357 to 362©2014 The Japan Institute of Metals and Materials EXPRESS REGULAR ARTICLE

2.2 Metallographic observationThe microstructural observation focused on the marked

region of SZ in the cross-sectional plane (perpendicular to thefriction-stir processing path). Electron back-scattered diffrac-tion (EBSD) analysis was used to identify the misorientationangle and grain size. The total scanning areas of as-FSPand as-TTEx specimens are approximate 8 © 104 and 1.2 ©107 µm2, respectively. The step size of scans was 0.4 µm, thebinning of Kikuchi pattern image was 8 © 8, and the furtherinformation was processed by OIM Data analysis software.According to the same viewpoint with previous studies,14,15)

only the conventional grains surrounded by high angle grainboundaries (HAGBs) were accounted for the grain sizecalculation. HAGBs were defined as misorientation angle>15° and grain size was calculated as the equivalent circlediameter of the grain area.

2.3 Texture analysisTexture data on the marked region was collected by X-ray

diffractometer and the orientation distribution functions(ODFs) were calculated from the incomplete pole figures{111}, {200} and {220}. The ODFs were calculated byBunge’s series method16) and represented in Euler’s space atconstant º2 sections. The texture component is labeled as(hkl)[uvw], where (hkl) is the crystallographic plane perpen-dicular to processing direction, and [uvw] is the crystallo-graphic direction parallel to normal direction.

In this study, Taylor orientation factor (M) was used tounderstand the textural effect on the Hall­Petch parameters.Since the tensile direction was parallel to the processingdirection, the stress pole of specific texture component(hkl)[uvw] was [hkl]. M value was calculated with Chin’smethod,17,18) which assumed tensile deformation as axisym-metric flow strain state and M value of each stress pole couldbe approached based on {111}h110i slip.

3. Experimental Results

3.1 Grain structureEBSD-OIM maps in Figs. 2(a)­2(c) show the grain

orientation and boundary characteristics in the markedregions of 450F, 650F and 850F. Equiaxed grains possessnear-random crystalline orientations and mainly neighborupon each other with high angle grain boundaries. Theaverage grain size is acquired as 3.7 µm at 450F, 7.8 µm at650F and 12.8 µm at 850F. The grain-to-grain misorientationangle distribution histograms (0 to 65 deg) in Fig. 2(d) revealthat all FSPed specimens own excess 90% HAGBs. Thestatistic data of coincident site lattice boundaries (CSLs)identification in Fig. 2(e) show that there are 10 to 13% ofCSLs within FSPed specimens and low ­ boundaries (i.e.,­ < 15) occupy about 2/3 in CSLs population.

For TTEx specimens, OIM graphs in Figs. 3(a)­3(c)indicate they possess coarse irregular grains and these grainsare typically surrounded by HAGBs. Average grain size isacquired as 508 µm at TTE0, 340 µm at TTE1 and 229 µm atTTE6. Grain boundary distribution histograms in Fig. 3(d)indicate that only less than 3% of low angle grain boundaries(LAGBs) exist in TTEx, which suggests that almost nosubgrain structure is identified within the coarse grains. CSLsstatistic histograms in Fig. 3(e) shows that there are about 8%CSLs in TTEx specimens and ­ distribution is more randomthan FSPed specimens.

3.2 Texture characteristicsThe results of texture analysis are represented with ODFs

in Figs. 4 and 5. The statistic data of texture componentsare synthesized in Tables 1 and 2, where “others” representsthose too weak components to be exactly identified and theyare reasonable to be considered as random texture. FromFig. 4 and Table 1, 450F, 650F and 850F specimens arecharacterized with weak texture signals. FSPed specimenspossess weak components f110gh001i (Goss orientation,labeled as G), ð045Þ½0�54� and ð302Þ½�203�. These specifictexture components contribute quite few and randomcomponents occupy more than 90% in volume fraction.According to Fig. 5, while 450F specimen is thermal exposedat 500°C, fine grains tend to grow into some particularcrystalline orientations. Comparison among all texturecomponents in Table 2 reveals that there are 50 to 75%rotated h001i type crystalline orientations in TTEx speci-mens. It suggests that grain coarsening under 500°C prefersto develop with {001} plane aligning to processing planewhether stain-energy-released treatment is carried out or not.

1mm

Fig. 1 Typical macrograph of the FSPed Al­Mg­Mn cast alloy and themarked region of SZ covers the gauge cross-section of tensile specimen.

Table 1 Volume fractions and associated Taylor factors of texture components in the as-FSP specimens.

450F 650F 850F

Component Vol% M Component Vol% M Component Vol% M

Goss 3.2% 3.67 Goss 2.1% 3.67 Goss 2.3% 3.67

ð045Þ½0�54� 1.2% 3.58 ð045Þ½0�54� 1.3% 3.58 ð045Þ½0�54� 1.5% 3.58

ð302Þ½�203� 2.8% 3.39 ð302Þ½�203� 1.2% 3.39 ð302Þ½�203� 1.0% 3.39

Others 93.8% 3.06 (230)[001] 1.2% 3.39 (230)[001] 1.1% 3.39

Others 94.2% 3.06 Others 94.1% 3.06

�M ¼ 3:10 �M ¼ 3:09 �M ¼ 3:09

C.-Y. Lin, T.-S. Lui, L.-H. Chen and F.-Y. Hung358

3.3 Tensile propertiesTable 3 shows the tensile properties of as-FSP and as-TTEx

specimens, in which deterioration of tensile strength andductility can be recognized in TTEx by comparing to FSP,especially in yield strength. As shown in Fig. 6, the yieldpoints are plotted as yield stress against the reciprocal ofsquare root of average grain size, and the linear relation isso-called Hall­Petch relation in eq. (1). The intercept of thestraight line on the stress axis is friction stress, ·0, and theline gradient is stress intensity coefficient, ky. The exper-imental data can be classified into two groups and each ofthem well fit on respective linear lines. The Hall­Petchparameters for group FSP in Fig. 6(a) are 128.7MPa in ·0and 128.2MPa·µm¹1/2 in ky. On the other hand, ·0 is111.1MPa and ky is 40.9MPa·µm¹1/2 for group TTEx in

Fig. 6(b). Both parameters in FSPed specimens are largerthan those in TTEx. The influential factors are discussed in thefollowing section.

4. Discussion

4.1 Parameter ·0Friction stress ·0, a grain-size-independent parameter,

is commonly interpreted as intragranular stress opposingthe passage of dislocations along the slip plane. The frictionstress is expressed as:

·0 ¼ M¸0 ð2Þwhere M is Taylor orientation factor, ¸0 is critical shear stresscorresponding to the tensile stress ·0 required to motivate

(e)

(b)

(a)

(c)

(d)

Fig. 2 Microstructure at the marked region in as-FSP specimens is shown as OIM graphs (a) 450F, (b) 650F, (c) 850F, and the distributionhistograms of grain boundary misorientation angle and CSLs are shown in (d) and (e).

Hall­Petch Tensile Yield Stress and Grain Size Relation of Al­5Mg­0.5Mn Alloy 359

dislocation slip in an isolated grain. The metallurgic factorson the friction stress are understood in terms of texturehardening,19) solute locking20) and precipitate hardening.21)

The statistic data of texture components in Table 1 showthat rotated h001i type components predominates in thespecimens from group TTEx but the specimens from group

(a) (d)

(e)

(b)

(c)

Fig. 3 Microstructure at the marked region in as-TTEx specimens is shown as OIM graphs (a) TTE0, (b) TTE1, (c) TTE6, and thedistribution histograms of grain boundary misorientation angle and CSLs are shown in (d) and (e).

Table 2 Volume fractions and associated Taylor factors of texture components in the as-TTEx specimens.

TTE0 TTE1 TTE6

Component Vol% M Component Vol% M Component Vol% M

ð001Þ½3�20� 36.3% 2.45 (100)[001] 35.2% 2.45 ð001Þ½3�20� 19.3% 2.45

ð001Þ½�2�30� 28.9% 2.45 ð314Þ½1�11� 17.0% 3.39 ð001Þ½�2�30� 14.7% 2.45

ð120Þ½�212� 11.3% 2.94 ð001Þ½1�30� 14.3% 2.45 ð001Þ½�1�10� 11.0% 2.45

ð001Þ½1�50� 5.65% 2.45 ð001Þ½�1�40� 9.5% 2.45 ð001Þ½�1�30� 8.0% 2.45

ð001Þ½�1�40� 3.55% 2.45 ð0�10Þ½301� 8.8% 2.45 ð001Þ½2�10� 6.0% 2.45

ð110Þ½1�11� 2.7% 3.67 ð143Þ½1�11� 1.0% 3.39 ð113Þ½1�52� 4.0% 2.89

ð015Þ½�100� 1.0% 2.40 ð110Þ½�1�11� 0.7% 3.67 ð113Þ½1�10� 2.2% 2.89

others 10.6% 3.06 others 13.1% 3.06 ð513Þ½�1�12� 2.0% 3.15

ð102Þ½�4�32� 1.6% 2.94

ð312Þ½�1�54� 1.5% 3.15

ð415Þ½�1�50� 1.0% 3.50

others 28.7% 3.06

rotated h001i components = 74.4% rotated h001i components = 67.8% rotated h001i components = 59.0%

�M ¼ 2:60 �M ¼ 2:71 �M ¼ 2:69

C.-Y. Lin, T.-S. Lui, L.-H. Chen and F.-Y. Hung360

FSP contain exceeding 90% “others” texture, i.e., randomcomponents. In Taylor’s calculation,22) M value of randomaggregates was obtained as 3.06 by averaging over all axialorientations. On the other hand, while the tensile stresspole lies at h001i, M value is 2.45. Average M values arecalculated in consideration of volume fraction of each texturecomponent. The value, �M, for FSPed specimens ranges from3.09 to 3.10, and for TTEx ranges from 2.60 to 2.71. Thevalue of �MFSP= �MTTEx is 1.16, which is quite similar to theratio of friction stress, (·0)FSP/(·0)TTEx = 1.158. This sug-gests that the difference of ·0 is mainly contributed fromtexture hardening effect. The value of critical shear stress, ¸0is similar between both groups which can be referred to theinsignificantly influential factors as precipitate hardening andsolute locking. Precipitate hardening is generally accepted tobe ignorable in non-heat treatable Al­Mg alloy. The workingtemperature of FSP and exposed temperature of TTEx arecommonly higher than 300°C. While the heating temperaturearising to 300°C, the solubility of major solute locking

element Mg exceeds 5mass%. Mg atoms are expected tocompletely dissolve into Al matrix in all experimentalconditions. The locking force from Mg solute atoms isinsignificantly different for all specimens.

4.2 Hall­Petch slope, kyky value in Hall­Petch relation expresses the intensity of

grain-size effect. Since grain boundaries provide priorpositions for stress concentration or relaxation during tensiledeformation, ky value is interpreted as the intensity oflocalized stress concentrations required for yielding prop-agation across grain boundaries. Wyrzykowski and Grabski’sresearch23) on aluminum alloy indicated CSLs possess a moreordered structure and lower boundary energy than generalhigh angle boundaries; for this reason, CSLs are capable ofsustaining more stress concentration. CSLs, especially in low­ boundaries, have higher boundary rigidity, provide moreretarded force for yielding propagation across and cause anincrease of ky value. The entirely distinct distribution profilesof CSLs between two groups are responsible for the differentky value. In which, ky = 128.2MPa·µm¹1/2 belonging togroup FSP is larger than ky = 40.9MPa·µm¹1/2 belonging togroup TTEx.

Besides HAGBs, which are accounted for grain sizecalculation in this study, LAGBs existing within the generalgrains also perform as obstacle barriers on the propagationroutes and provide an additional effect on ky.14,15) In thisstudy, more existence of LAGBs in FSPed specimens (4­10%) contributes more additional retarded force for yieldingdeformation than TTEx specimens (<3%). That is consideredas another factor which leads to higher ky value of FSPedspecimens.

(a)

(b)

(c)

Fig. 4 Texture characteristics of as-FSP specimens: (a) 450F, (b) 650F and(c) 850F.

(a)

(b)

(c)

Fig. 5 Texture characteristics of as-TTEx specimens: (a) TTE0, (b) TTE1and (c) TTE6.

Table 3 Tensile properties of as-FSP and as-TTEx specimens.

YS UTS UE TE

450F 198.6MPa 339.3MPa 24.4% 29.9%

650F 176.9MPa 335.2MPa 23.5% 28.5%

850F 163.0MPa 319.0MPa 26.4% 30.9%

TTE0 113.0MPa 290.0MPa 21.7% 25.2%

TTE1 113.4MPa 294.1MPa 19.8% 23.7%

TTE6 114.0MPa 292.2MPa 21.1% 25.5%

YS: yield strength, UTS: ultimate strength,UE: uniform elongation, TE: total elongation.

Hall­Petch Tensile Yield Stress and Grain Size Relation of Al­5Mg­0.5Mn Alloy 361

5. Conclusion

(1) Control of rotational speed and holding time of strain-energy-release heat treatment can be used to producedifferent levels of fine and coarse-grain structure. Byusing Hall­Petch relation, experimental specimens canbe divided into two groups, in which specimens fromgroup FSP possess higher values of ·0 and ky than thosefrom group TTEx.

(2) As-FSP specimens are characterized by weak Goss,ð045Þ½0�54� and ð302Þ½�203� texture components andTTEx specimens possess much stronger texture com-ponents with h001i plane aligning to processing plane.�M values for group FSP and group TTEx are 3.09­3.10 and 2.60­2.71, respectively. The relation ð·0ÞFSP=ð·0ÞTTEx ; �MFSP= �MTTEx suggests difference of ·0between two groups mainly depends on the averagevalue of Taylor’s orientation factor.

(3) As-FSP specimens contain more fraction of low ­

boundaries and larger total fraction of CSLs than TTEx

specimens. More ordered structure in CSLs owns bettercapability of stress concentration and it strengthens thegrain boundary rigidity. Besides special boundaries,more LAGBs within grain structure in specimens fromgroup FSP also contribute to higher ky value.

Acknowledgement

This paper has been supported by the Chinese NationalScience Council (Contract: NSC-102-2221-E-006-062), forwhich the authors are grateful.

REFERENCES

1) R. S. Mishra and Z. Y. Ma: Mater. Sci. Eng. R 50 (2005) 1­78.2) P. L. Threadgill, A. J. Leonard, H. R. Shercliff and P. J. Withers: Int.

Mater. Rev. 54 (2009) 49­93.3) P. Vilaça and W. Thomas: Adv. Struct. Mater. 8 (2012) 85­124.4) Z. Y. Ma: Metall. Mater. Trans. A 39 (2008) 642­658.5) Y. S. Sato, M. Urata, H. Kokawa and K. Ikeda: Mater. Sci. Eng. A 354

(2003) 298­305.6) T. Hirata, T. Oguri, H. Hagino, T. Tanaka, S. W. Chung, Y. Takigawa

and K. Higashi: Mater. Sci. Eng. A 456 (2007) 344­349.7) E. O. Hall: Proc. Phys. Soc. B 64 (1951) 747­753.8) N. J. Petch: J. Iron Steel Inst. 174 (1953) 25­28.9) M. F. Ashby and D. R. H. Jones: Engineering Materials 1: An

Introduction to their Properties and Applications, (Pergamon Press,Oxford, 1980) p. 105.

10) A. Niikura and Y. Bekki: Mater. Sci. Forum 331­337 (2000) 871­878.

11) K. A. A. Hassan, A. F. Norman, D. A. Price and P. B. Prangnell: ActaMater. 51 (2003) 1923­1936.

12) I. Charit and R. S. Mishra: Acta Mater. 53 (2005) 4211­4223.13) C. Y. Lin, T. S. Lui and L. H. Chen: Mater. Trans. 52 (2011) 1667­1673.14) H. Fujita and T. Tabata: Acta Metall. 21 (1973) 355­365.15) N. Hansen: Scr. Mater. 51 (2004) 801­806.16) H. J. Bunge: Texture Analysis in Materials Science, (Butterworth,

London, 1982) pp. 3­118.17) G. Y. Chin: The Inhomogeneity of Plastic Deformation, (R. E. Reed-

Hill, ASM, Ohio, 1973) pp. 83­112.18) C. Y. Chin and W. L. Mammel: Trans. Metall. Soc. AIME 239 (1967)

1400­1405.19) R. W. Armstrong: Yield, Flow and Fracture of Polycrystals, (T. N.

Baker, Applied Science Publishers, Barking, 1983) pp. 1­31.20) D. J. Lloyd and S. A. Court: Mater. Sci. Technol. 19 (2003) 1349­

1354.21) A. A. Mazilkin, B. B. Straumal, E. Rabkin, B. Baretzky, S. Enders,

S. G. Protasova, O. A. Kogtenkova and R. Z. Valiev: Acta Mater. 54(2006) 3933­3939.

22) G. I. Taylor: J. Inst. Met. 62 (1938) 307­324.23) J. W. Wyrzykowski and M. W. Grabski: Philos. Mag. A 53 (1986) 505­

520.

(a)

(b)

Fig. 6 The Hall­Petch grain size dependence on the tensile yield stress isclassified into (a) group FSP and (b) group TTEx.

C.-Y. Lin, T.-S. Lui, L.-H. Chen and F.-Y. Hung362