Damping Capacity of Fe­17mass%Mn High Damping Alloywith Variant Controlled Microstructure

Yoshimi Watanabe+, Yuusuke Suga, Hisashi Sato, Hideaki Tsukamoto and Yoichi Nishino

Nagoya Institute of Technology, Nagoya 466-8555, Japan

In our previous study, it is reported that damping capacity as well as hardness of an Fe­20mass%Mn alloy can be improved by the thermo-mechanical training featured by rolling deformation. In this study, the thermo-mechanical training of an Fe­17mass%Mn alloy is carried outwith bending mode, since vibration manner of the internal friction measurement refers to bending mode. An anisotropic damping capacity isobserved for samples subjected to bending mode training. Moreover, the trade-off between the damping capacity and hardness can be overcomeby thermo-mechanical training. To be concluded, the thermo-mechanical training is useful for enhancement of damping properties and hardnessof Fe­Mn alloys. [doi:10.2320/matertrans.ME201305]

(Received November 21, 2012; Accepted April 2, 2013; Published May 17, 2013)

Keywords: damping capacity, iron­manganese alloy, high damping alloy, anisotropy, variant, marteisitic transformation, hardness

1. Introduction

As one of actuator materials, shape memory alloys (SMAs)are well studied. A high damping capacity is considered asone of the important functional properties of SMAs. Forexample, it is well known that Ti­Ni and Cu­Al­Mn SMAscan exhibit high damping capacity.1­3) Those properties arerelated to a martensitic transformation.

In the meantime, it is well recognized that the increase ofdamping by traditional methods results in a reduction ofthe strength of mechanical properties.4) Figure 1 shows aninternal friction, Q¹1, vs. tensile strength map, where most ofmetallic materials are classified into three groups, the high-,intermediate- and low-damping materials.5) Therefore, thetrade-off relation between the damping capacity and strengthmust be overcome.

The Fe­Mn based alloys undergoing martensitic trans-formation from £-austenite (fcc) to ¾-martensite (hcp) areknown as low-cost high-damping alloys. It is reported that thedamping sources of Fe­Mn alloys were ¾-martensite variantboundaries, stacking faults in £-austenite and ¾-martensiteand £/¾ interphase boundaries.6) According to our previousstudies,7,8) it is found that the damping capacity of an Fe­20mass%Mn alloy, as well as hardness, can be improved bythe training treatments, where the training treatments in theSMA are known as a useful method to improve the shapememory effect (SME) in Fe­Mn alloys.9,10) Since the rollingdeformation was adopted as the thermo-mechanical training,7)

the deformation mode was different from that in vibration ofinternal friction measurement.

It has been well known that twelve ¾-martensite variantscan be produced according to twelve {111}©112ª shearsystems in martensitic transformation of £-austenite to ¾-martensite.11,12) However, the relative population of themartensite variants is strongly depended on the directionand sense of the external stress.13,14) If the bending mode isadopted for the thermo-mechanical training, it is expectedthat the favorable ¾-martensite variants for the dampingcapacity can be introduced since sound and noise are created

by bend mode vibrations in elastic amplitude. It is reportedthat an Fe­17mass%Mn alloy exhibits the highest dampingcapacity in Fe­Mn binary system.6) In this study, the effectsof the thermo-mechanical training featured by bending modeon microstructure, hardness and damping capacity in an Fe­17mass%Mn alloy have been studied.

2. Experimental Procedure

The chemical composition of the Fe­Mn alloy used in thisstudy is listed in Table 1. This alloy was prepared by meltingin a vacuum furnace. The rod shaped cast ingot of 30 kgwith 150mm diameter was homogenized at 1200°C for 24 h,and then hot forging was carried out at temperature rangebetween 1200 and 900°C. After the hot forging, the barshaped sample with 100mm © 20mm © 800mm was aircooled. The starting temperatures of martensitic transforma-tion (Ms) of Fe­Mn binary alloys are shown in Fig. 2.12,15) In

Mg-0.6%ZrMgCu-Al-Ni

Mn-Cu-Al

Fe-Cr-AlAl-Zn

TiNiFe-Co

12Cr-Steel

Ni

Fe

Grey CastIron (high C)

Grey CastIron

AZ81A

SUS304

S95C

Al Cast Alloy

BronzeBrass

Tensile Strength, σ / MPa102 103

10-1

10-2

10-3

Inte

rnal

fri

ctio

n, Q

-1

Fig. 1 Internal friction vs. tensile strength map.5)

+Corresponding author, E-mail: yoshimi@nitech.ac.jp

Materials Transactions, Vol. 54, No. 8 (2013) pp. 1288 to 1294Special Issue on New Functions and Properties of Engineering Materials Created by Designing and Processing II©2013 The Japan Institute of Metals and Materials

this figure, dotted line is the chemical composition used inthis study. As can be seen, Ms temperature for ¾-martensitictransformation of used Fe­17mass%Mn alloy is about150°C.

In this study, the effects of thermo-mechanical training ondamping capacity are investigated. Figure 3 shows thethermo-mechanical training program. First step of thetraining is heat treatment. The sample was austenitized at700°C for 2 h followed by water quenching. Second step isbending at room temperature (R. T.). The bending modetraining was adopted, since vibration manner in internalfriction measurement is bending, as described later. A coupleof deformation tests, bending and flatting deformation, iscarried out for one cycle of training, where dies for bendingare shown in Fig. 4(a) and die for flatting is metallic blockwith flat surface. To study the anisotropic damping capacity,two kinds of initial samples are used for thermo-mechanicaltraining. One is LD sample with dimensions of 1.5mm ©10.5mm © 81.0mm, and its longitudinal direction is per-pendicular with respect to the compression die. On the otherhand, WD sample has larger size (1.5mm © 81.0mm ©60.0mm), and its longitudinal direction is parallel with

respect to the compression die, as shown in Fig. 4(b). Thedeformation tests were carried out on an Instron type testingmachine at a cross-head speed of 1mm/min for bending

0 5 10 15 20 25 30

mass% Mn

900

800

700

600

500

400

300

200

100

0

γ

α

εTem

pera

ture

, T /

°C

Fig. 2 Ms temperature of martensitic transformation of Fe­Mn binaryalloys depending on Mn concentration.15,12)

flattingbending

R.T.

700°C

Deformation

2h 2h

Deformation

flattingbending

1 cycle

6 cycles

Fig. 3 Thermo-mechanical training program used in this study.

25mm

10mm

Specimen

Die

Compression Die

Press

φ 20mm

R=10mm

(a) (b)

81mm10.5mm

81mm

60mm

10.5mm

LD Sample

WD Sample

t=1.5mm

t=1.5mm100mm

WDLD

LDWD

23.6mm

10mm

11.8mm

10mm

r mm

1.5m

m

10mm

(c)

32m

m

r mm

Fig. 4 (a) Dies for bending used in the thermo-mechanical training.(b) Bending directions of thermo-mechanical training in LD and WDsamples. Vibration directions in internal friction measurement of LD andWD samples are parallel and perpendicular to the bending direction in thethermo-mechanical training, respectively.

Table 1 Chemical composition of used Fe­Mn damping alloy.

Element Fe Mn C Si Ni Cr T­N O

mass% Bal. 16.96 0.001 0.01 <0.01 <0.01 0.001 0.0057

Tungsten wire

(8mm x 4mm x 0.2mm)

1mm

Iron foil

Driving coil

Detector coil

Noise filter

A/D converter

Oscillator

Amplifier

Computer

22.4%

22.4%

80mm10mm

100%

Fig. 5 The setup for internal friction measurement.8)

10μm

Fig. 6 Typical microstructure of the sample without the training treatment.

Damping Capacity of Fe­17mass%Mn High Damping Alloy with Variant Controlled Microstructure 1289

deformation and 2mm/min for flatting deformation. Sincethe maximum stroke of the bending deformation is fixed to be10mm, radius of curvature of neutral axis is calculated to be18mm, as shown in Fig. 4(c). Then the surface strain of 0.04can be introduced by this deformation. Bending direction isnot changed by training cycle. The training cycling wasrepeated 2, 4 or 6 times for each sample. After the training,the sample with dimensions of 1.5mm © 10.5mm ©81.0mm was cut from WD sample, as shown in Fig. 4(b).As a consequence, vibration direction in internal frictionmeasurement of LD and WD samples is parallel andperpendicular with bending direction in the thermo-mechan-ical training, respectively.

In order to measure damping capacity, internal friction,Q¹1 was measured at R. T. in air atmosphere using a free-decay method of flexural resonant vibration with both freeends, as shown in Fig. 5.8) Since the £-austenite and ¾-martensite phases in Fe­17mass%Mn alloy show para-magnetism, a small piece of ferromagnetic Fe foils (8mm ©4mm © 0.2mm) were attached to both ends of the specimenin order to excite mechanical vibrations efficiently by anelectromagnetic force and to sensitively monitor the vibrationby an induced current. After a steady-state vibration, thedriving signal was turned off and the free-decay curve wasrecorded by a computer through a high-speed analogue­digital converter.8)

IPF

ε

21100001

1010

15μm

21100001

1010[001]

(a) (b)

(c)

WD

LD

LD

WD

Fig. 8 An inverse pole figure map of ¾-martensite (a), an inverse pole figure of ¾-martensite (b) and a (111) pole figure of £-austenite (c) inthe sample without training treatment. In inverse pole figure in (b), [001] presents the direction perpendicular to the LD and the WDdirections.

α

ε

γ 21100001

1010

101

111

001

101

111

00115μm 15μm

Phaseαεγ

0.0080.8240.168

Fraction

(a) (b)

Fig. 7 An inverse pole figure map (a) and a phase map measured (b) by EBSD for sample without training treatment.

Y. Watanabe, Y. Suga, H. Sato, H. Tsukamoto and Y. Nishino1290

The microstructure was observed using optical microscopeand electron backscattered diffraction (EBSD) after mechan-ically and electrically polishing. The electro-polishing wascarried out using Struers A2 electrolyte at 13°C and 32V for9 s in a Struers Electropol device. EBSD measurement wasperformed with step size of 0.3 µm. For the EBSD measure-ments, crystal orientation of ¾-martensite was analyzed withlattice ratio of c/a = 1.613.16) The hardness was measuredusing a Vickers hardness tester.

3. Results and Discussion

Figure 6 shows a typical microstructure obtained withoptical microscopy of the sample without the trainingtreatment. It is apparent that the sample without trainingcontains ¾-martensite phase, since Ms temperature for ¾-martensitic transformation of used Fe­17mass%Mn alloy isabout 150°C, as shown in Fig. 2. It is also found that the¾-martensite phase has some variants. This is because twelve¾-martensite variants can be produced according to twelve{111}£©112ª£ shear systems in £ to ¾ martensitic trans-

formation. Average grain size of prior £-austenite in thespecimen without training treatment is 20 µm.

EBSD results of the sample without training treatment areshown in Figs. 7 and 8. Figures 7(a) and 7(b) are an inversepole figure map and a phase map, respectively. From thephase map, it is seen that this microstructure consists of ¾-martensite and £-austenite, and ¡-martensite phase could notbe observed. Volume fractions of ¾-martensite and £-austenitephases are 82 and 17 vol%, respectively.

Since the damping sources of the Fe­Mn system arerelated to ¾-martensite phase,6) a detailed analysis on ¾-martensite phase is performed. Figures 8(a), 8(b) and 8(c)show an inverse pole figure map, an inverse pole figure ofonly ¾-martensite phase and (111) pole figure of £-austenitein the sample without training treatment, respectively. In theinverse pole figure in Fig. 8(b), [001] presents the directionperpendicular to the LD and the WD directions. As can beseen from Figs. 8(a) and 8(b), crystal orientation distributionof ¾-martensite in whole region of this sample is randomalthough transformation texture of ¾-martensite is observedin a prior £-austenite grain. This is because that there is no

10 μmLD

WD

(a)

10 μmLD

WD

(b)

10 μmLD

WD

(c)

10 μmLD

WD

(d)

10 μmLD

WD

(e)

10 μmLD

WD

(f)

Fig. 9 Microstructural change with thermo-mechanical training process. (a) to (c) are LD samples and (d) to (f ) are WD samples. Numberof cycling is 2 times for (a) and (d), 4 times for (b) and (e), and 6 times for (c) and (f ).

Damping Capacity of Fe­17mass%Mn High Damping Alloy with Variant Controlled Microstructure 1291

texture in crystal orientation of prior £-austenite grains in thisspecimen as shown in Fig. 8(c). This is in agreement with theresults obtained with optical microscopy. An average sizeof ¾-martensite measured from Fig. 8(a) is 14.4 µm in lengthand 2.7 µm in width.

Optical micrographs showing the change in ¾-martensitestructure by the thermo-mechanical training are shown inFig. 9. LD and WD directions in this figure are longitudinaland width directions in internal friction measurement,respectively, as defined in Fig. 4(b). In Fig. 9, (a) to (c) areLD samples and (d) to (f ) are WD samples, respectively.Number of cycling is 2 times for (a) and (d), 4 times for (b)and (e), and 6 times for (c) and (f ). Average grain sizes of

prior £-austenite in these treated specimens at 2 times, 4 timesand 6 times are 40, 39 and 39 µm, respectively. In case ofthe LD samples, lots of ¾-martensite are oriented with itslongitudinal direction parallel with LD direction. On the otherhand, ¾-martensite in the WD samples is oriented with itslongitudinal direction perpendicular with LD direction. Inthis way, the ¾-martensite phases have a preferred orientationin the samples with thermo-mechanical training.

Figures 10(a) and 10(b) are an inverse pole figure map anda phase map measured by EBSD, respectively, for the samplesubjected to 2 times of thermo-mechanical training. Themicrostructure consists of ¾-martensite, £-austenite and smallamount of ¡-martensite. Volume fractions of each phase are

α

ε

γ 21100001

1010

101

111

001

101

111

001

WD

LD

15μm

WD

LD

15μm

Phaseαεγ

0.0090.8460.145

Fraction

(a) (b)

Fig. 10 An inverse pole figure map (a) and a phase map measured (b) by EBSD for sample with thermo-mechanical training. Number ofcycling is 2 times.

IPF

ε

21100001

1010

15μm

21100001

1010[001]

WD

LD

(a) (b)

(c)

LD

WD

Fig. 11 An inverse pole figure map of ¾-martensite (a), an inverse pole figure of ¾-martensite (b) and a (111) pole figure of £-austenite (c)in the sample with thermo-mechanical training. In inverse pole figure in (b), [001] presents the direction perpendicular to the LD and theWD directions. Number of cycling is 2 times.

Y. Watanabe, Y. Suga, H. Sato, H. Tsukamoto and Y. Nishino1292

84.6, 14.5 and 0.9 vol%, respectively. Larger amount ¾-martensite phase is detected as compared with the samplewithout training. Crystal orientation textures and smallergrain sizes of the ¾-martensite phase are also observed.

Figures 11(a), 11(b) and 11(c) show an inverse pole figuremap, an inverse pole figure of only ¾-martensite and a(111) pole figure of £-austenite in the sample with thermo-mechanical treatment, respectively. In Fig. 11(b), [001] is thedirection perpendicular to the LD and the WD directions.As well as specimen without thermo-mechanical treatment,prior £-austenite grains have no texture. On the other hand,it is important to note that strong texture is observed for¾-martensite, as shown in Fig. 11(b). Moreover, an averagesize of ¾-martensite in this sample is 7.9 µm in length and0.9 µm in width. The size of ¾-martensite in this specimenwith thermo-mechanical treatment is smaller than that ofthe sample without treatment. Therefore, enhanced andanisotropic damping capacity will be expected for thethermo-mechanical trained samples.

Internal friction, Q¹1, in the trained samples as a functionof training cycle is shown in Fig. 12. The data of the samplewithout training is also shown in this figure. It is seen fromthis figure that enhanced damping capacity is found for thesamples with the thermo-mechanical training. It is importantto note that an anisotropic damping capacity is observed forthe thermo-mechanical trained samples. Namely, a strongtraining effect was found for LD samples, while a small effectfor WD samples. As the deformation of thermo-mechanicaltraining, the same deformation mode for vibration must bechosen if one would like to obtain a better damping capacityof Fe­Mn damping alloy. The increase in internal friction isdue to an increase in the number of the favorable dampingsources per unit volume, i.e., oriented ¾-martensite variantboundaries, oriented stacking faults and oriented £/¾interfaces. Therefore, the thermo-mechanical training bybending deformation, in which bending direction parallelwith vibration direction in internal friction measurement,increases the relative population of favorable ¾-martensitevariant boundaries, favorable stacking faults and favorable£/¾ interfaces for damping capacity.

Figure 13 shows the hardness of the samples with andwithout training treatment. It may have been noticed that

thermo-mechanical treatment increases the hardness of Fe­Mn alloy, and the hardness increases with increasing thenumber of training cycle. However, there was no significantdifference in hardness between LD sample and WD sample.This is because basically the same thermo-mechanicaltreatment was performed for LD and WD samples.

An anisotropic damping capacity is found for the thermo-mechanical training by bending mode. Usually vibrationdirection of noise souses is predetermined; our techniqueprovides the better damping material for industrial fields.Since it is reported that the 100% ¡-martensitic alloys such asFe­5mass%Mn and Fe­10mass%Mn alloys showed verylow damping capacities,12) existence of ¡-martensite must beavoided. The ¡-martensite in high manganese Fe­Mn alloyis induced by relatively large strain. Therefore, during thethermo-mechanical training, the surface strain by the bendingdeformation must be less than 0.04.

In this study, it is found that the thermo-mechanicaltreatment increases the hardness of Fe­Mn alloy, as well asdamping capacity. Therefore, the trade-off between thedamping capacity and hardness can be overcome by thetraining. It can be concluded that the training treatments arevery useful for enhancement of the damping property andhardness of Fe­Mn alloy.

4. Conclusion

In this study, the training effects on microstructure,hardness and internal friction in an Fe­17mass%Mn alloyhave been investigated. The thermo-mechanical trainingfeatured by bending mode is carried out, since vibrationmanner in internal friction measurement is also bendingmode. The results of the study are summarized as follows.

(1) The ¾-martensite phases have a preferred orientation inthe specimens with thermo-mechanical training with bendingdeformation.

(2) An anisotropic damping capacity is observed for thesamples subjected to bending mode thermo-mechanicaltraining.

(3) It is found that the trade-off between the dampingcapacity and hardness can be overcome by the thermo-mechanical training.

(4) The training treatments are useful for enhancement ofthe damping properties and hardness of Fe­Mn alloy.

1.4

1.2

0.8

0 2 4 6Number of Training, N

8

1.0

0.6

without Training

WD Sample

LD Sample

x 10-4

Inte

rnal

fri

ctio

n, Q

-1

Fig. 12 Internal friction, Q¹1, in the trained samples as a function oftraining cycle. The data of the sample without training is also shown inthis figure.

350

300

200

0 2 4 6Number of Training, N

8

250

150

without Training

WD Sample

LD Sample

Mic

ro V

icke

rs H

ardn

ess,

HV

Fig. 13 The hardness of the samples with and without training treatment.

Damping Capacity of Fe­17mass%Mn High Damping Alloy with Variant Controlled Microstructure 1293

Acknowledgments

The authors are grateful to Mr. Satoshi Tamaoka for hishelpful works in carrying out the experiments. The authorsalso would like to thank Mr. Kozo Ozaki of the DaidoSteel Co. Ltd., for providing the Fe­Mn alloy. The authorssincerely acknowledge the financial support of this studyby Ministry of Education, Culture, Sports, Science andTechnology, Regional Innovation Cluster Program.

REFERENCES

1) J. Van Humbeeck: J. Alloy. Compd. 355 (2003) 58­64.2) N. Igata, N. Urahashi, M. Sasaki and Y. Kogo: J. Alloy. Compd. 355

(2003) 85­89.3) Y. Sutou, T. Omori, N. Koeda, R. Kainuma and K. Ishida: Mater. Sci.

Eng. A 438­440 (2006) 743­746.4) R. Mulyukov, S. Mikhailov, R. Zaripova and D. Salimonenko: Mater.

Res. Bull. 31 (1996) 639­645.5) H. Mizubayashi, S. Murayama and H. Tanimoto: J. Alloy. Compd.

330­332 (2002) 389­392.6) S.-H. Baik, J.-C. Kim, D.-W. Han, T.-H. Kim, J.-H. Back and Y.-K.

Lee: Mater. Sci. Eng. A 438­440 (2006) 1101­1105.7) Y. Watanabe, H. Sato, Y. Nishino and I.-S. Kim: Mater. Sci. Eng. A 490

(2008) 138­145.8) Y. Watanabe, H. Sato, Y. Nishino and I.-S. Kim: Mater. Sci. Eng. A

521­522 (2009) 376­379.9) K. Tsuzaki, M. Ikegami, Y. Tomota and T. Maki: ISIJ Int. 30 (1990)

666­673.10) Y. Watanabe, Y. Mori and A. Sato: J. Mater. Sci. 28 (1993) 1509­1514.11) Z. Nishiyama, M. E. Fine, M. Meshii and C. M. Wayman: Martensitic

Transformation, (Academic Press, New York, 1978) p. 48.12) Y.-K. Lee, J.-H. Jun and C.-S. Choi: ISIJ Int. 37 (1997) 1023­1030.13) Y. Higo, F. Lecroisey and T. Mori: Acta Metal. 22 (1974) 313­323.14) M. Kato and T. Mori: Acta Metal. 25 (1977) 951­956.15) H. Schumann: Z. Metallkd. 58 (1967) 207­210.16) T. Gebhardt, D. Music, B. Hallstedt, M. Ekholm, I. A. Abrikosov,

L. Vitos and J. M. Schneider: J. Phys. Condens. Matter 22 (2010)295402.

Y. Watanabe, Y. Suga, H. Sato, H. Tsukamoto and Y. Nishino1294