1

In-Situ Quantification of Solute Effects on Grain BoundaryMobility and Character in Aluminum Alloys During

Recrystallization

Mitra L. Taheri1, Anthony D. Rollett1, and Hasso Weiland2

1Department of Materials Science and Engineering, Carnegie Mellon University,Pittsburgh, PA 15232, USA

2Alcoa Technical Center, Alcoa Center, PA 15609, USA

Keywords: Recrystallization, grain boundary, coincidence site lattice, mobility

Abstract. Aluminum alloys exhibit recrystallization kinetics that vary strongly withcomposition. The conventional understanding is that certain alloying elements, e.g.chromium, retard grain boundary motion due to the formation of fine dispersions ofsecond phase particles, giving rise to particle drag of boundaries. There iscountervailing evidence, however, that suggests that solute drag provides a strongerinfluence on grain boundary mobility. This paper presents new evidence for apronounced effect of solute based on experiments in which individual boundariesmigrate under the driving pressure of stored energy from prior plastic strain. Assupported by the literature, boundaries exhibit a maximum mobility for a 38-39degree <111> misorientation in initial annealing experiments. Specifically, thismobility maximum is asymmetric with a sharp cutoff below 38-39 degrees but a moregradual decrease at misorientations beyond 40 degrees. The occurrence of other,smaller mobility peaks is discussed within the context of the sharpening of evolvingmaxima with discussed within the context of the sharpening of evolving maxima withincreased recrystallization. The presence of a minimum at 38-39 degrees is found atboth higher temperatures and higher solute concentrations. This transition from alocal mobility maximum to a minimum is discussed within the context of recenttheories solute drag activity.

Introduction

The effect of solutes on the migration of boundaries has long been a topic ofinvestigation due to its tremendous impact on its applications in materials science,both industrial and academic. All commercially significant alloys have substantiallevels of solute that strongly affect the migration mechanism. The source of this dragis the displacement between the position of the moving boundary and the position ofthe chemical potential defined by the solute distribution, assuming that there is asignificant segregation of solute to or away from the boundary. Unfortunately, verylittle experimental data has been collected for either low or high angle boundarymobility in alloys although some studies have been performed in nominally ‘pure’materials [6]. Even for these nominally pure metals, boundary migration is affectedby solute. The lack of measurement of the solute drag effect can be attributed to thedifficulty of defining the exact location and action of solute “drag” as the force ofattraction between the grain boundary and the alloying element(s) [1,5], to theoversimplification inherent in current solute drag theory, or most importantly, to thelimitations of current imaging capabilities.

There exists a limited amount of experimental investigation with respect to solutedrag, all of which has been performed within a Scanning Electron Microscope.

2

Evidence of both solute drag as well as differences in growth mechanisms of certainboundary types has been seen [2-4, 6].

Experimental

The mobility of individual grain boundaries during recrystallization was studied withan in-situ annealing procedure. A hot-stage was inserted into the XL-40 FEG SEM inorder to observe directly the mechanism of boundary movement, as shown in Fig. 1.During these heating experiments, electron backscatter diffraction patterns (EBSPs)were obtained to examine the orientation relationships of each grain at each stage ofdevelopment. This method is closely related to that of both Humphreys and Huang[3,4] and Molodov et al. [2], where both boundary velocity and orientation are studiedduring recrystallization. Single crystals were grown from compressed samples ofdiffering impurity concentrations (Table I) and subsequently cold rolled and scratchedbefore heating. Electron Backscatter Diffraction Patterns were acquired intermittentlyduring the in-situ annealing experiments. This process was repeated until the desiredgrain size was reached; this was commonly found around 200 µm.

Table I. Alloy Compositions (% solute)

Si Fe Cu Mn Mg Cr Ni Zn Ti Pb Be Zr

HPAl+Zr .002 .002 .004 .000 .002 .000 .002 .005 .000 .0001 .000 .031

CPAl+Zr .032 .014 .005 .000 .002 .000 .002 .005 .000 .000 .000 .032

Figure 1: Schematic representation of OIM setup with hot stage in XL-40 Chamber and annealingsequence representation

Results and Discussion

Annealing experiments with the two alloys, HPAl+Zr and CPAl+Zr, showed thatthere is a strong dependence of mobility and morphology on both boundary character(i.e. misorientation) and annealing temperature. In the HPAl+Zr alloy, a pronounced

3

38-39°<111> mobility maximum was found when annealed at 350°C; these highlymobile grain boundaries were characterized as near-sigma 7 type CSL boundaries,and exhibited pronounced faceting, growing parallel to the rolling direction with thesefaceted edges at the growing front of the grain. This phenomenon is shown in Fig. 2that shows micrographs at successive times for a sample of HPAl+Zr annealed at350°C. A maximum sharpening just below 40° is displayed in Fig. 3.

Figure 2: Faceted near-sigma 7 CSL type grains in an HPAl+Zr alloy annealed at 350°C taken atdifferent times to show the growth of recrystallized grains.

A sharp contrast in this trend was observed during annealing HPAl+Zr at 485°C. Themobility maximum at approximately 38° appeared to become a minimum; this shift intemperature also resulted in maxima at approximately 35° and 48° degrees,respectively. This is seen in Fig. 4a. This trend can be explained by the compensationtemperature (Tc) effect [6], i.e., that all boundaries in a given series have enthalpiesand entropies of mobility in a fixed ratio with the consequence that at some particular(compensation) temperature, all mobilities are equal. Variations in enthalpy amongboundary types in the same series may then lead to reversals in the rank order ofmobilities as the temperature is changed. Another result of the higher temperatureanneal was that the faceting of the growing grains was less pronounced, as shown inFig. 4b.

The absence of faceting mimics the behavior of the alloy with a higher solutecontent, CPAl+Zr, which is a commercial purity aluminum containing appreciableamounts of iron and silicon as well as the deliberate addition of zirconium. Theannealing sequence for CPAl+Zr alloy is shown below in Fig. 5. It should be notedthat the final annealing time (to obtain recrystallization) for this alloy at 485°C was106 minutes; this suggests that the additional Fe and Si exert appreciably greatersolute drag when compared to a final annealing time for the HPAl+Zr alloy at 485°Cof 20 minutes (to obtain comparable migration distances).

4

Figure 3: Diameter versus Misorientation Angle of individual recrystallized grains in a sample ofHPAl+Zr annealed at 350°C for 237 minutes.

(a) (b)

Figure 4: (a) Diameter versus Misorientation Angle for HPAl+Zr annealed at 485°C for 20 minutes,indicating shift of maximum from approximately 38° to a minimum yielding new maxima at 35°and

48°; (b) Reduction in amount and severity of faceting in highly mobile grains in HPAl+Zr annealed at485°C.

With respect to the most mobile grain boundaries, there is a maximum in migrationrate near the sigma 7 CSL boundary type (38.2°), with the largest grains possessingboundary misorientation angles of 38-39° and some at angles between 40-50°, asshown in Fig. 6a; this trend is similar to that seen in the HPAl+Zr alloy annealed at350°C. In contrast to this, the most mobile grain boundaries for this CP+Al alloyannealed at approximately 525°C had misorientation angles above 40°, with thelargest closer to 50°, as shown in Fig. 6b. The comparison of the two plots shows astrong compensation effect, similar to that found in the HPAl+Zr alloy. This shift inmaxima from low temperature anneals (presumably at T<Tc) to high temperatureanneals (T>Tc), is supported by the results of Gottstein et al. [6] in a study ofcurvature driven <111> boundaries in aluminum.

5

(a) (b)

Figure 5: (a) CPAl+Zr annealed at 485°C 25 minutes; (b) CPAl+Zr annealed at ~525°C for 20 minutes

(a)(b)

Figure 6: Diameter versus Misorientation Angle plots of the CPAL+Zr alloy annealed at (a) 485°C for25 and (b) 525°C for 20 minutes.

Summary

• The change in annealing temperature from 350°C to 485°C producesdrastically different trends in grain boundary mobility with respect to specificboundary types in an HPAl+Zr alloy. A maximum near the Σ7 CSL type isseen at 350°C whereas a minimum at this angle is seen at 485°C, with maximaat both 35° and 48°.

• A similar trend in character specific growth is found in a CPAl+Zr alloyannealed at 485°C, with a local maximum near the Σ7 CSL boundarymisorientation.

• A compensation effect is observed in both alloys when changing from low tohigh temperature anneals; the maximum at 38-39° is shifted due to this

6

change, suggesting low activation energy of the near Σ7 boundary growthprocess.

• Significant faceting is seen in the HPAl+Zr alloy annealed at 350°C; thesefaceted grain boundaries are of the Σ7 CSL boundary type. Faceting wasnegligible in both the CPAl+Zr alloy and the HPAl+Zr alloy when annealed at485°C.

• Slower recrystallization was noted in the CPAl+Zr sample compared to theHPAl+Zr; the recrystallization time was approximately 80 minutes longer.This suggests stronger solute drag corresponding to the higher levels of ironand silicon present in this material.

• The above observations suggest a connection between boundary character andmorphology with both temperature and solute content. At a given temperature,alloys with higher solute content exhibit mobility variations similar to thosewith lower solute content annealed at higher temperatures.

Acknowledgments

Support is gratefully acknowledged from the Alcoa Technical Center and by theMRSEC program of the National Science Foundation under Award Number DMR-0079996.

References

1. K. Lücke and H.P. Stüwe: Acta metall. Vol.19 (1971) p. 1087-1099; J. Cahn:

Acta metall. Vol. 10 (1962), p. 789-798.

2. D.A. Molodov et al.: Acta mater. Vol. 46 (1998), p. 553-562.

3. Y. Huang and F.J. Humphreys: in Recrystallization and Grain Growth, G.

Gottstein and D.A. Molodov, Eds, (Springer-Verlag, 2001) p. 409.

4. Y. Huang and F.J. Humphreys: Acta mater. Vol. 47 (1999), p. 2259-2271.

5. M.I. Mendelev and D.J. Srolovitz: Acta mater. Vol. 49 (2001), p. 589.

6. G. Gottstein and L.S. Shvindlerman: Grain Boundary Migration in Metals:

Thermodynamics, Kinetics, Applications (CRC Press, 1999).

7. R.A. Vandermeer and D. Juul Jensen: Metall. Mater. Trans A Vol.28 (1997),

p.749-754.

8. F.J. Humphreys and M. Hatherly: Recrystallization and Related Annealing

Phenomena (Pergamon Press, Oxford, 1995).

9. B. Hutchinson, S. Jonsson and L. Ryde: Scripta metall. 23 (1989), p. 671.

10. R.A. Vandermeer and D. Juul Jensen: Metall. Mater. Trans. A, Vol. 28

(1997), p.749-754.