Fe-ga薄片二次再结晶

Available online at

www.sciencedirect.com

Scripta Materialia 66(2012)

307–310

www.elsevier.com/locate/scriptamat

Single grain growth and large magnetostriction in secondarily recrystallized Fe–Gathin sheet with sharp Goss (011)[100]

orientation

Suok-Min Na ⇑and Alison B. Flatau

Department of Aerospace Engineering, University of Maryland, 3181Glenn L. Martin Hall, College Park, MD 20742, USA

Received 7October 2011; revised 7November 2011; accepted 15November 2011

Available online 22November 2011

Nearly single-grain-oriented Fe–Gathin sheets with a sharp Goss (011)[100]orientation have been developed through rolling and annealing processes. Annealing of rolled samples at 1200°C under a sulfur atmosphere produced abnormal Goss (011) grain growth, covering $98%of the sample surface. Compared with previous results from anneals under argon, the sulfur anneal resulted in the acceleration of abnormal (011) grain growth and a $79%increase in the observed magnetostriction, from 163to 292ppm, for the same anneal protocol.

Keywords:Fe–Gaalloy; Goss texture; Abnormal grain growth; Magnetostriction; Annealing

Magnetostrictive Fe–Gaalloys (Galfenol)have

promising attributes for application to sensors, actua-tors and energy harvesting as Clark et al. ﬁrstreported in 2000[1]. Single-crystal Galfenol has a body-centered cubic (bcc)crystal structure, and along the h 100i crystal orientations, it exhibits saturation magnetostriction of $400ppm in low applied magnetic ﬁeldsof $200Oe. It also has a mechanical strength of $500MPa, which is high relative to more costly rare earth magnetostric-tive materials such as Terfenol-D alloys which exhibit giant magnetostriction ($1600ppm) but are brittle and require much higher magnetic ﬁelds(>$1000Oe) for saturation [1–3]. The large magnetostriction and easy magnetization in single-crystal Galfenol alloys occur along the h 100i orientation. It is thus desirable to obtain the h 100i orientation in textured polycrystalline Galfenol, with the goal of providing enhanced mechan-ical properties and lower cost than single-crystal mate-rial, with similar magnetostrictive strain.

Two viable approaches have been employed to fabri-cate highly textured Fe–Gaalloys [4,5]. One is a directional solidiﬁcationgrowth process, and the other is thermome-chanical processing involving deformation via rolling and recrystallization through grain growth and orienta-tion mechanisms. Galfenol rods grown by the directional solidiﬁcationprocess have strong crystallite textures with h 100i preferred orientation aligned 14°oﬀfrom the rod

Corresponding

author. Tel.:+[1**********]; fax:+1301314

9001; e-mail:[email protected]

direction and a maximum magnetostriction (k k ) of 271ppm under compression [4]. In other works, Kellogg et al. reported that binary Fe 0.83Ga 0.17with a somewhat dispersed {001}h 100i texture along rolling direction (RD)exhibited magnetostrictive strain of $160ppm as a consequence of rolling and annealing at 1100°C for 4h [5]. Texture annealing of Fe 0.85Ga 0.15alloy with 1mol.%NbC at 1150–1300°C for 24h changed the texture from a strong a -ﬁbertexture h 110ik RD in as-rolled sheet to a preferred texture with h 100i orientation [6]. The authors did not report the magnetostriction values; however, an estimated of lower than 135ppm can be made based on their electron backscatter diﬀraction(EBSD)data and the nominal saturation value in a single crystal with the same composition. In our prior work, we have demon-strated the texture development of Goss {011}h 100i tex-ture through secondary recrystallization by using NbC particles as an inhibitor of normal grain growth (NGG)[7]. The dispersion of NbC particles resulted in promotion of the abnormal growth of {011}grains in a process that is similar to the inhibition of NGG in Fe–Sielectrical steel which occurs due to precipitation of second-phase parti-cles such as AlN and MnS [8,9]. The most common trig-ger for abnormal grain growth (AGG)is local pinning of low-energy non-Goss grain boundaries (GBs)and selec-tive depinning of high-energy Goss GBs as a result of pre-cipitation of ﬁnelydispersed particles [10]. This is because high diﬀusivityat high-energy Goss GBs leads to rapid coarsening of precipitates during annealing. The resulting large particles have a reduced pinning force, restricting

308S.-M. Na, A. B. Flatau /Scripta Materialia 66(2012)307–310

the movement of Goss grain boundaries. Thus, the Goss grain boundaries move more easily than other bound-aries, leading to AGG. Although AlN and MnS are eﬀec-tive particles for achieving AGG in Fe–Sialloys, they do not work well in Fe–Gaalloys. Dispersed NbC particles are an eﬀectivealternative, and they precipitate and coar-sen during annealing, leading to reduced pinning force and AGG in Fe–Gaalloys. Unfortunately, NbC particles have a melting point of 3610°C, which is signiﬁcantlyhigher than for AlN (2200°C) and MnS (1610°C), and in combination with having a larger particle size (1–5l m in diameter) this reduces the role of kinetics in the grain growth rate for abnormally grown Goss grains. This behavior limits the extent of abnormal growth of Goss grain in Fe–Gaunlike what occurs in Si steel. The AGG in Fe–Ga–NbCalloy annealed under argon is only dependent on the grain boundary energy diﬀerencebetween Goss and other grains, and ﬁnallystops when there is a balance between particle pinning force by Zener drag [11]and driving force by grain boundary en-ergy. Our previous EBSD results show a Goss area frac-tion of 60%of the scanned surface area, which is consistent with a restriction on the extent of Goss grain growth in the Fe–Ga–NbCsystem [7].

To overcome this limitation, we consider a global grain growth mechanism that takes into account surface energy diﬀerencesassociated with the misorientation an-gle between grains. Surface-energy-induced selective grain growth has been suggested by Ko ¨hlerand Kramer in Fe–Sialloy thin sheet annealed under a sulfur atmo-sphere using hydrogen sulﬁde(H) [12,13]. With the goal of approaching 2S) gas in hydrogen (Hstrictive 2the magneto-capability of single-crystal material, the desir-able texture needs to be more fully developed

sharper. In this paper, results on Goss grain growth a whole area of samples annealed under a atmosphere using H AGG behavior 2S gas are presented and with of argon-annealed samples Fe 0.81Ga 0.19plus 1mol.%NbC thin sheet. crystallographic orientation and strongly data on magnetostriction are shown as well.

Alloy material prepared by ETREMA Product consisted of melting elemental Ga with pure Fe and casting into ingot form through an induction technique. The ingot of Fe–19at.%Ga plus 1.0NbC ﬁneparticles (1–5l m in diameter) was enclosed a 321stainless steel can and sealed to prevent The 7.6mm thick ingot was hot rolled at 700°C to an 88.6%reduction. Subsequent warm rolling at 400provided a 34%reduction to a thickness of 0.54After a stress relief anneal at 600°C for 2h under cold rolling was undertaken to obtain a thickness 0.45mm (17%reduction). A series of subsequent ing was conducted on samples of two 12mm Â12mm Â0.45mm and 25.4mm Â25.4mm 0.45mm, at temperatures of 1200–1250°C under a atmosphere using 0.5%H 2S in argon for diﬀerentdurations to examine the inﬂuenceof sulfur on The gas ﬂowrate of 0–30sccm was precisely by a mass ﬂowcontroller (AalborgGFC) during ing. All samples were quenched in water when cooled. The saturation magnetostriction of each sample measured under a DC magnetic ﬁeldof 3500Gauss vided by Nd–Fe–Bpermanent magnets. The sample was rotated in this ﬁeldand the peak-to-peak saturation va-lue of (3/2)k S =k k Àk \for both front and back sur-faces of each sample was obtained using strain gauges with a gage area of 4.5mm Â7mm (basearea of 6.5mm Â14mm). The observed magnetostriction value in a specimen is determined by averaging the measured values at both front and back sides. EBSD patterns were captured and analyzed using OIM data collection soft-ware (TSLOIM Analysis 5) to obtain inverse pole ﬁgure(IPF)images, pole ﬁgure(PF),area fractions of each grain, fractions of each orientation and orientation dis-tribution function (ODF)plots.

Magnetostriction data and a curve ﬁtof the data for samples annealed at temperatures of 1200and 1250°C under either a sulfur or an argon atmosphere are plotted in Figure 1as a function of annealing time. The polyno-mial curves were determined using all data points for each dwell time.

The observed magnetostriction values for the sulfur-annealed samples increased from 28ppm in the as-rolled sample to maxima of 245and 247ppm, averaged on the front and back sides, at annealing temperatures of 1200and 1250°C, respectively. Most samples annealed at 1200°C showed an average variation of as much as 35ppm between the front and back surfaces of each sam-ple in samples with the dimensions of 12mm Â12mm. This was determined to be caused by the formation of somewhat diﬀerentgrain conﬁgurationson the front and back surfaces that were caused by the experimental set-up, even though the main orientation of both surfaces is exactly same. The processing tube geometry led to an unintended diﬀerencein the H 2S gas contact frequency

S.-M. Na, A. B. Flatau /Scripta Materialia 66(2012)307–310309

temperatures higher than 1200°C produced through-thickness grains over the entire sample surface and thus exhibited the same magnetostriction values at the both sides, with a peak value of 247ppm at 1250°C for 2h. The standard deviation of the magnetostriction values from the front to back of a sample were small, i.e. within the measurement error range of ±3ppm, except for the 3h data which varied by ±16ppm.

The observed magnetostriction values after annealing under H 2S were higher than the 163and 183ppm observed in samples annealed under a pure argon atmosphere at 1200and 1250°C, respectively, in our prior work [14]. The large increment in magnetostriction for annealing with the same time and temperature protocols is attributed to a change in texture caused by a change in the sample surface energy during the anneal due to the sulfur atmosphere. The grain conﬁgurationsand crystal orientations of some surfaces of annealed samples were analyzed through EBSD patterns captured by OIM and results on IPF images are given in Figure 2. Figure 2a shows the typical primary recrystallization with no microscopic texture, where (001), (011)and (111)grains each covered approx-imately one-third of the scanned area. The randomly ori-ented grain growth during primary recrystallization seems not to be aﬀectedby an annealing atmosphere of sulfur (usingH 2S) because the stored energy of deformed sheet is dominant for nucleation of the new strain-free grains. This initial microstructure, when subjected to subsequent annealing at longer times under the same atmosphere, leads to single-grain growth behavior. The sulfur-an-nealed specimen in Figure 2b shows the surface fully cov-ered by a (011)grain as if a single crystal, where area fraction of (011)grains in the scanned area (equalto the whole sample surface area) is 98.2%.Several small (001) and (111) grains were located in a (011)grain matrix like small islands, corresponding to 0.5%and 1.3%in the area fraction, respectively. On the other hand, both argon-an-

was grown until its grain walls became pinned at the boundaries of small matrix grains with diﬀerentorienta-tions due to the retardation of grain boundary motion due to pinning by NbC particles. Thus, even though NbC particles in Fe–Gasheet are eﬀectivefor promotion of AGG due to suppressing NGG, diﬀerencesin grain boundary energy alone were insuﬃcientto approach sin-gle-crystal-like grain growth throughout the sample. In contrast, sulfur annealing increases the rate of grain boundary migration by over 10-fold 1(from7.0l m min À1under argon to 70.1l m min À) as calculated from EBSD data. From these results, sulfur annealing using H accelerate the AGG of a single Goss 2S gas is believed to grain beyond what is achieved with the eﬀectsof NbC by intro-ducing a diﬀerencein surface energy between grains. In addition to single grain growth, crystal orientation of Goss grain along RD is of importance for maximizing magnetostrictive performance. The distribution of angu-lar deviation from an ideal Goss orientation with h 100i to within a tolerance of 0.25°was proﬁledfor the sulfur-annealed samples from the EBSD data. This allowed for assessment of the dependence of magnetostriction on the dispersion of Goss orientation around the ideal orienta-tion. It was observed that the deviation angles of a sam-ple annealed for 4h from h 100ik RD was less than 10°and the peak angle was 3.4°, accompanied with the max-imum magnetostriction. Even though the area covered by a Goss grain in the 3h sample is almost the same as from the 4h sample, the magnetostriction value is lower due to the broader deviation and peak angle of 6.1°. The magnetostriction values strongly depend on the extent of deviation angle and peak intensity. The much lower magnetostriction value of 163ppm, pro-duced by annealing for 3h under argon, was similarly aﬀectedby an even bigger deviation in Goss grain orien-tation, of as much as 21°.

The eﬀectof orientation on magnetostriction can be expressed as an area fraction of primary texture compo-nents such as g -ﬁberh 100ik RD, a -ﬁberh 110ik RD and h 111ik RD orientation (asquantiﬁedthrough calculation of EBSD patterns for the whole surface of each sample). The linear relationship between measured magnetostric-tion and each primary texture for all scanned sample sur-faces is shown in Figure 3, where the area fractions are calculated using the maximum deviation angle of 10°from each orientation parallel to the rolling direction (RD)in conventional Euler angle (u A linear regression for each texture 1, U , u was performed 2) notation. to determine which texture and microstructure components could be used to predict values corresponding to satura-tion magnetostriction. Regression results indicate that the g -ﬁbertexture, h 100ik RD, is the only signiﬁcantfactor examined [15]. Other a -ﬁberh 110ik RD and h 111ik RD orientations are not correlated with the mag-netostriction and instead are inversely proportional to it. The line ﬁtof h 100i texture on (011) grain yields satu-ration magnetostriction of 300ppm at 100area%.

Although the same time and temperature annealing protocols were carried out for the sample results pre-sented in Figure 1, diﬃcultiesin maintaining a constant H 2S atmosphere due to the gas ﬂowrate regulators em-ployed might explain the diﬀerentAGG patterns and the error bars associated with the magnetostriction val-

Materialia 66(2012)307–310

ues. After an update of the mass ﬂowcontrollers for the 0.5%H2S +Ar and Ar gas tanks, subsequent experiments were conducted with improved control of the gas ﬂowrate and content of H samples, with dimensions 2S to test this hypothesis. Slightly larger of 25.4mm Â25.4mm, were annealed at 1200°C under more precisely controlled H 2S ﬂowrates of 0, 4.5, 15and 30sccm for 4–6h to deter-mine if the scatter in the 1200°C annealed magnetostric-tion data could be attributed to poor control of the H The 4, 5and 6h anneal data had magneto-2S gas ﬂowrate. striction values of 247±44, 256±47and 269±22ppm, respectively. These results are inconclusive, although the variability in magnetostriction data is lowest for the 6h anneal, which is consistent with the error bars for the 4and 5h anneals being greater than for the 6h anneal. It is considered that the sulfur contamination is strongly correlated with the contact frequency to the sur-face, which in turn aﬀectsgrain conﬁgurationand texture change. Our previous results on the correlation between sulfur segregation and surface-energy-induced selective grain growth in (Fe0.813Ga 0.187) 99.5B showed that sulfur atoms 0.5doped with 50ppm sulfur segregated on the surface [16]. Observations indicated that {011}grains were dominant under low sulfur amounts, less than 0.5at.%,and {001}grains occurred under slight surface contamination with sulfur, at concentration levels of be-tween 0.5and 1.35at.%.After that, there was a threshold above which the selective growth of {001}grains changed to {111}grains, accompanied with a steep drop in magne-tostriction. The observed amount of segregated sulfur was analyzed by Auger electron spectroscopy (AES)depth proﬁles.Two other groups have observed a similar trend in the correlation between texture and sulfur contamina-tion in annealed 3%Si–Fethin sheets, with anneal under sulfur atmosphere using H 2S gas in one case [12]and with doping sulfur in bulk interior in the other study [17]. Kra-mer speciﬁcallyattributed surface-energy-induced selec-tive growth of (001) and (011) grains to control of the partial pressure of H 2S in hydrogen [13]. The maximum magnetostriction of 292ppm was reproducibly obtained at 4.5sccm. This occurs because the alignment of h 100i orientation is well matched with ideal Goss orientation and the deviation angle is less than the 3°determined from the (001) ODF ﬁgureand angular deviation plot. This is the ﬁrstreport of rolled Fe–Gathin sheet that exhibits this extent of sharpness in Goss texture. Ever since the inven-

tion of the principle production process for grain-oriented (GO)silicon steel by Goss in 1934[18], the importance of achieving texture sharpness has led to signiﬁcanteﬀorts.Even with recent work on silicon steels, the average devi-ation angle of h 100i orientation has only been improved to within 3°of the RD [19,20]. The ability to achieve such a high extent of sharpness in (011) grain-oriented Galfe-nol is attributed to an abnormal single grain growth. In conclusion, the eﬀectof sulfur anneal on abnormal grain growth, magnetostriction and resultant crystal ori-entation in polycrystalline (Fe0.81Ga 0.19) +1.0mol.%NbC alloy was examined through rolling and annealing processes. Goss texture {011}h 100i was formed in $98%of the annealed sheet at 1200°C for 4h that closely resembles the structure of a (011)[100]single crystal, while a number of small grains remained dispersed like small islands in the (011) grain matrix. The abnormal (011) grain growth was attributed to the combined eﬀectsof a dispersion of NbC ﬁneparticles and sulfur-induced surface energy eﬀectsduring annealing. The dispersive NbC in Galfenol sheet is eﬀectivefor suppressing normal grain growth, and sulfur anneal is believed to accelerate the abnormal (011) grain growth by providing the addi-tional driving force of surface energy needed to overcome the pinning forces that arise due to dispersed NbC parti-cles during AGG. The synergy of combining NbC parti-cles with H approach for 2S high-temperature anneals resulted in an making single-crystal-like Fe–Gathin sheet with magnetostriction of $292ppm.

This work was supported by ONR MURI

Grant No. N[1**********]0.

[1]A.E. Clark, J.B. Restorﬀ,M. Wun-Fogle, T.A. Lograsso, D.L. Schlagel, IEEE Trans. Magn. 36(2000)3238.

[2]A.E. Clark, K.B. Hathaway, M. Wun-Fogle, J.B. Restorﬀ,T.A. Lograsso, V.M. Keppens, G. Petculescu, R.A. Taylor, J. Appl. Phys. 93(2003)8621.

[3]S. Guruswamy, N. Srisukhumbowornchai, A.E. Clark, J.B. Restorﬀ,M. Wun-Fogle, Scripta Mater. 43(2000)239. [4]N. Srisukhumbowornchai, S. Guruswamy, J. Appl. Phys. 90(2001)5680.

[5]R.A. Kellogg, A.B. Flatau, A.E. Clark, M. Wun-Fogle, T.A. Lograsso, J. Appl. Phys. 93(2003)8495.

[6]N. Srisukhumbowornchai, S. Guruswamy, Metall. Mater. Trans. A 35A (2004)2963.

[7]S.M. Na, J.H. Yoo, A.B. Flatau, IEEE Trans. Magn. 45(2009)4132.

[8]A. Sakakura, J. Appl. Phys. 40(1969)1534.

[9]P. Lin, G. Palumbo, J. Harase, K.T. Aust, Acta Mater. 44(1996)4677.

[10]N. Rajmohan, J.A. Szpunar, Y. Hayakawa, Texture

Microstruct. 32(1999)153.

[11]S. Benum, E. Nes, Acta Mater. 45(1997)4593. [12]D. Ko ¨hler,J. Appl. Phys. 31(1960)408S. [13]J.J. Kramer, Metall. Trans. A 23(1992)1987.

[14]S.M. Na, A.B. Flatau, J. Appl. Phys. 103(2008)07D304. [15]E.M. Summers, R. Meloy, S.M. Na, J. Appl. Phys. 105

(2009)07A922.

[16]S.M. Na, A.B. Flatau, J. Appl. Phys. 101(2007)09N518. [17]N.H. Heo, K.H. Chai, J.G. Na, Acta Mater. 48(2000)2901. [18]N.P. Goss, US Patent 1965559, 1934.

[19]S. Taguchi, Trans. Iron Steel Inst. Jpn. 17(1977)604. [20]T. Kumano, T. Haratani, Y. Ushigami, ISIJ Int. 43

(2003)

736.