Gn

Ga

b

a

ARRAA

KNENGG

1

mevhpsnmpifiagnmgmpna

0d

Materials Science and Engineering A 539 (2012) 324–329

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering A

journa l homepage: www.e lsev ier .com/ locate /msea

rain growth characteristics and its effect on deformation behavior inanocrystalline Ni

arima Sharmaa,∗, Jalaj Varshneyb, A.C. Bidayeb, J.K. Chakravarttya

Mechanical Metallurgy Division, Bhabha Atomic Research Centre, Mumbai 400 085, IndiaMaterial Processing Division, Bhabha Atomic Research Centre, Mumbai 400 085, India

r t i c l e i n f o

rticle history:eceived 25 September 2011eceived in revised form 10 January 2012ccepted 24 January 2012vailable online 1 February 2012

a b s t r a c t

The grain growth characteristics and its effect on deformation behavior of nanocrystalline(nc)-Ni havinga grain size of 60 nm have been studied. The kinetics of thermal grain growth was studied by DSC andresistivity experiments. Thermal instability leading to grain growth in nc-Ni was assessed by determiningthe activation energy required for grain growth in the temperature range of 100–400 ◦C. The activationenergy was found to be ∼100 kJ/mol below Curie temperature and ∼298 kJ/mol above Curie tempera-

eywords:anostructured materialslectron microscopyanoindentationrain growth

ture. The effect of grain size on nano-hardness and loading rate was investigated using nanoindentation

technique. The deformation parameters like strain rate sensitivity (m) and activation volume (∗V) for dif-

ferent grain sizes in nc-Ni were also evaluated. The interaction of dislocations-grain boundaries mediatedmechanism was appeared to be the rate controlling plastic deformation mechanism.

© 2012 Elsevier B.V. All rights reserved.

rain boundary

. Introduction

Materials with grain size less than 100 nm have been receivinguch attention because of their unparallel and exceptional prop-

rties. In comparison with conventional coarser grained materials,arious benefits offered by nanostructured materials include ultra-igh yield and fracture strengths, superior wear resistance, andossibly superplastic formability at low temperatures and/or hightrain rates. The high volume fraction of grain boundaries in theanocrystalline (nc) materials serves as effective obstacles for theotion of dislocations and therefore is expected to influence their

lastic flow characteristics. Thus, the strength and creep propertiesn nc-materials are believed to be directly related to the volumeraction of grain boundary present. However, nc-materials are notn thermal equilibrium due to the high amount of energy associ-ted with the grain boundaries, and they usually show tremendousrain growth with temperature [1–3]. Technological application ofanocrystalline materials requires stability of the nanocrystallineicrostructure at elevated temperatures which is often limited by

rain growth. Thus, it is very important to understand the ther-al stability of nc-materials from both technological and scientific

oint of view. However, it is seen that grain growth behavior ofanocrystalline materials has been studied in detail and that onlyfew investigations on grain growth kinetics for nanocrystalline

∗ Corresponding author. Tel.: +91 22 25590457; fax: +91 22 25505050.E-mail addresses: garimas@barc.gov.in, garimas@yahoo.com (G. Sharma).

921-5093/$ – see front matter © 2012 Elsevier B.V. All rights reserved.oi:10.1016/j.msea.2012.01.102

pure metals, oxides, compounds or composites have been reportedso far [4–6]. Recently, research interest has been focused to under-stand the plastic deformation behavior in nc-Ni [7–12]. However,the effect of transition from the nanocrystalline grains to ther-mal equilibrium grains on the mechanical property of nc-materialshas not been studied in detail, and therefore, a systematic studyis needed. Therefore the motivation of the present study is tounderstand the kinetics of grain growth in nc-Ni. In addition,nanoindentation technique has been employed to study the defor-mation behavior and deformation parameters such as strain ratesensitivity (m) and activation volume (V*) have been evaluated.

2. Experimental

nc-Ni foils of about 100 �m thickness were prepared by elec-trodeposition technique using a stainless steel substrate. Priorto deposition, the substrate surface was prepared by mechani-cal polishing on different grades of SiC abrasive paper and finallydiamond polished to a 7 �m finish. The stainless steel electrodeswere cleaned with solvent cleaner, acetone, and then chemicallycleaned with steelex and dilute HCl to remove dirt and oxide layerfrom its surface. The cleaned substrates were then passivated with50% HNO3 for 5 min at RT. Passivation helps in formation of non-adherent coating and thus could be removed easily by slight pulling

from one corner. The plating bath consisted of NiSO4, NiCl2 andNa3C6H5O7 with pH adjusted by the addition of dilute H2SO4. Thedeposited nc-Ni was removed from the substrate in the form of foil.The chemical composition of the Ni foil was determined by WDS

G. Sharma et al. / Materials Science and Engineering A 539 (2012) 324–329 325

(afmKceutaTuitotort

400300200

-0.1

0.0

0.1

0.2

0.3

Heat F

low

(m

W)

Temperature (oC)

Abnormal Grain Growth

3. Results and discussions

Fig. 1. Schematic diagram for four probe resistivity measurement system.

wavelength dispersive) analysis carried on Cameca SX 600 withbeam size of 1 �m. The nc-Ni was found to be 99.98% pure. Ni

oils were characterized using X-ray diffraction (XRD) and trans-ission electron microscopy (TEM). XRD was carried out using Cu

� radiation (1.5406 A) and Si was used as an external standard fororrecting the broadening associated with the instrument. Differ-ntial scanning calorimeter measurements (DSC) were carried outsing Mettler Toledo, DSC 822e with Au-AuPd thermopile. The elec-rical resistance was measured on thin nc-Ni foils in a four-terminalrrangement at constant temperature under Argon atmosphere.he constant current source along with Kithley nanovoltmeter wassed for the measurement of the voltage. The isothermal resistiv-

ty measurement was carried out by inserting the sample insidehe furnace heated to predetermined temperature. A standard cellf 1 ohm resistance was used with the current source for settinghe constant input current across the probing wires. Copper wires

f diameter 0.3 mm were spot welded as point contact on stripesistivity samples for 4 probe measurement. Standard eurothermemperature controller was used for controlling heating of the

10080604020

200

400

600

(311)(220)

(222)

(200)

Inte

nsity (

arb

. un

its)

Two Theta (degrees)

(111)

Fig. 2. XRD pattern of as received electro-deposited nanocrystalline-Ni.

Fig. 3. DSC curve showing grain growth in nc-Ni during constant heating rate at10 K/min.

furnace. A program was made in Lab View plateform with RS232 communication for the data recording in interval of 1 s. Also,samples of same size and thickness were used for resistivity mea-surements. The schematic representation of the resistivity setupused for the measurement is shown in Fig. 1.

nc-Ni was heat treated at different temperature under Argonatmosphere. Samples for TEM investigation were prepared by jetpolishing 3 mm discs using 20% perchloric acid and 80% methanolat 20 V and −45 ◦C. TEM observations were conducted with a JOEL2010 operated at 200 kV using conventional bright field and darkfield imaging techniques. The hardness experiments were car-ried out using nanoindentation technique with Berkovich diamondindenter (UNHT, CSM, Switzerland) to investigate the deformationbehaviour. The P (load) vs. h (penetration depth) measurementswere carried out at as a function of load. Strain rate changetest was performed by changing the loading rate from 2500 to25,000 �N/min, at a load of 4000 �N. Three experiments were con-ducted for each condition and average value of hardness was usedfor determining the m and V* parameters.

Fig. 2 shows the XRD pattern of the electrodeposited nc-Ni, withthe major peaks corresponding to (1 1 1), (2 0 0), (2 2 0) and (3 1 1)

Fig. 4. Resistivity plot showing grain growth as function of time during isother-mal heating of nc-Ni at different temperatures. Note R0 is the saturation resistivityobtained at the experimental temperature.

326 G. Sharma et al. / Materials Science and Engineering A 539 (2012) 324–329

F in ass 0 min

rfsg

Fs

ig. 5. Bright field TEM image of (a) as received nc-Ni; (b) Grain size distributionhowing isolated coarse grain; (d) grain growth after heat treatment at 370 ◦C for 3

eflections. The average grain size was estimated to be ∼60 nm

rom the diffraction line width based on Scherrer’s relation. Ithould be noted that this grain size was obtained without using anyrain refiner during electrodeposition experiments. The ratio of the

21201918171615

4.8

5.2

5.6

6.0

6.4

6.8

ln(d

gra

in d

ia.)

nm

/min

1/TTemperature

(10-4) K

Slope = -Q/nR

ig. 6. Plot showing the variation of grain size (d) with temperature (T) in nc-Ni;lope yield the activation energy required for grain growth.

received Ni; (c) grain growth after heat treatment at 200 ◦C for 40 min with inset.

major peaks intensity indicates random orientation of the grains.Fig. 3 shows the differential scanning calorimetry (DSC) curve fornc-Ni at a constant heating rate of 10 K/min. The results indicatea significant increase in grain growth above Curie temperature i.e.353 ◦C (shown by arrow in Fig. 3). As the grain boundaries in nc-materials are not in thermal equilibrium, annealing result in graingrowth in order to minimize their energy (associated with them)by attaining a stable grain structure. The kinetics of grain growthin bulk materials has been studied in detail by resistivity measure-ment [13,14]. In pure nc-metals with no impurities, resistivity isa good technique to estimate the grain growth, as the main con-tribution to resistivity comes from electron scattering from grainboundaries. However, not much literature is available for resistivitymeasurement in nc-materials [15–17]. For pure nc-Cu, it has beenreported that dislocation and vacancy annihilation have only neg-ligible effect on resistivity and that main contribution to resistivitycomes from grain growth [17]. This was due to the fact that dislo-cation and vacancy annihilation do not provide sufficient energy toinitiate grain growth in nc-materials. In contrast, the annihilationof grain-boundary area leads to an overall decrease of total energyand, consequently, drives the grain growth. Fig. 4 shows the varia-

tion of resistivity (R/R0; where R0 is the saturation resistivity) withtime during isothermal heating of nc-Ni. In this study, the resis-tivity was found to decrease with time indicative of the decreaseof the grain boundaries in pure nc-Ni. The grain growth profile at

G. Sharma et al. / Materials Science and Engineering A 539 (2012) 324–329 327

4.44.03.63.22.8

40

60

80

100

120

Submicron-sized grains

Ac

tiv

ati

on

Vo

lum

e (

V*)

Hardness (GPa)

nc-Ni

Nano-sized grains

20001500100050000

100

200

300

400

500

50mN10mN

100mN

200mN

300mN

400mN

Lo

ad

(m

N)

Displacement (nm)

500mNnc-Ni, RT

30020010000

1000

2000

3000

4000

5000

RT

HT-500HT-340

HT-200

Lo

ad

(μN

)

Displacement (nm)

HT-100

Heat Treated nc-Ni

1010.10.01

3.5

4.0

4.5

d=250 nm

d=100 nm

Ha

rdn

es

s (

GP

a)

Strain rate (s-1)

d=60 nm

nc-Ni

6004002000

50

75

100

125

Act.

Vo

l (V

*/b

3)

Grain Size (nm)

(a) (b)

(d)(c)

(e)

F ad; (bh e for nd

2ridorrb

ataganow

grains. Klement and da Silva [18] have also shown in their EBSDstudies on nc-Ni that first grown grains have a 〈3 1 1〉//ND orien-tation. However, further heating upto 90 min at 200 ◦C showed a

Table 1Effect of heat treatment on grain size of nc-Ni.

Heat treatment Avg. grain size (nm)

1 100 ◦C for 1 h 632 200 ◦C for 30 min Only selective grains grow upto ∼200 nm (max)

ig. 7. (a) Load displacement curve for as received nc-Ni as a function of applied loardness vs. strain rate plot for nc-Ni with different grain size; (d) activation volumifferent grain size.

00 ◦C in nc-Ni is shown in Fig. 4. A considerable decrease in theesistivity during the initial stages could be due to electron scatter-ng through high volume of grain boundaries. However, for longeruration resistivity does not change significantly and become morer less constant as there was no further grain growth and grainseached a thermal equilibrium state. At 350 ◦C, a sharp decrease inesistivity indicated that the fine-grained structure was consumedy large grains within 25 min (Fig. 4).

Based on the resistivity and DSC results, nc-Ni samples werennealed in the temperature range of 100–400 ◦C for variousime periods. Detailed TEM investigations were performed on thennealed samples so as to determine the grain size in nc-Ni. Table 1ives the grain size obtained as a function of annealing temperature

nd time. Fig. 5(a) shows the bright field TEM image of as preparedc-Ni with grain size distribution is shown in Fig. 5(b). The grain sizebtained by TEM investigations are found to be in good agreementith those obtained by XRD results in the present study. Fig. 5(c)

) load displacement curve for heat treated nc-Ni exhibiting different grain size; (c)c-Ni as a function of grain size; (e) activation volume as a function of hardness for

shows the selective grain growth of isolated grains at 200 ◦C for40 min with coarse grain is shown as inset. The main reason forselective growth could be due to the different orientations of the

3 200 ◦C for 90 min 954 300 ◦C for 30 min 2105 340 ◦C for 30 min 2406 370 ◦C for 25 min 500

3 and E

sloge

D

watcmvsgapwTini∼tdC

c[wrdttblnahmwpchNrim(piIgcngsghbiCtmf

28 G. Sharma et al. / Materials Science

light increase in overall grain size of the matrix in addition to iso-ated coarse grains. At 350 ◦C, the uniform coarse grain size wasbserved after 25 min as shown in Fig. 5(d). The kinetics of grainrowth with temperature in nc-materials can be described by thequation [12]:

n − Dn0 = K0t exp

(−Q

RT

)(1)

here D0 and D are the starting and current grain sizes, t is thennealing time, T the absolute temperature, R the gas constant, Qhe activation energy for grain growth, and n and K0 are materialonstants. The activation energy for the grain growth was deter-ined from the slope of plot of ln (d) vs. 1/T as shown in Fig. 6. The

alue of n (grain growth exponent coefficient) determined from thelope of plot ln (d − do) vs. ln (t) and was found to be close to ∼7. Therain growth behavior followed usual Arrhenius type relation withtransition in kinetics at Curie temperature [12]. At lower tem-

erature the value of activation energy was found to be 100 kJ/molhich increases to 298 kJ/mol above Curie temperature (353 ◦C).

he activation energy for grain growth below Curie temperatures very close to that for grain boundary diffusion in polycrystallineickel ∼109 kJ/mole [19]. However, at higher temperature it was

n close agreement with the activation energy for volume diffusion278 kJ/mol in polycrystalline nickel [20]. These results showed

hat the grain growth in nc-Ni was controlled by grain boundaryiffusion below Curie temperature and by volume diffusion aboveurie temperature.

Grain size is known to have a significant effect on the mechani-al behavior of materials, in particular, on the hardness or strength.2,3]. The dependence of yield stress on grain size in metals isell established in the conventional coarse-grained polycrystalline

ange, but for nc-materials it still requires focused studies. Nanoin-entation is a viable and most suitable technique for evaluatinghe deformation behavior of nc-materials. In the present study,he effect of grain size on the mechanical properties was studiedy nanoindentation technique. Fig. 7(a) shows the representative

oad, P, vs. indentation depth, h, curves obtained on as preparedc-Ni samples as a function of applied load. These P–h curves werenalyzed by employing the Oliver–Pharr method to extract theardness of the sample [21]. The hardness of nc-Ni did not varyuch with applied load and an average hardness value of 4.5 GPaas obtained for 60 nm as prepared nc-Ni. Fig. 7(b) shows the P–hlot for heat-treated nc-Ni samples for different grain sizes. The plotlearly shows the increase in the indentation depth with increase ineat treatment temperature. This indicate that the hardness of nc-i decrease with increase in grain size. The effect of strain/loading

ate on hardness for different grain size was determined by chang-ng the applied loading rate from 2500 to 25,000 �N/min, at a

aximum load of 4000 �N (Fig. 7(c)). The strain rate sensitivity,m = (∂ log H)/(∂ log ε)), was determined from the slope of theselots. The as-prepared nc-Ni yields a value of m ∼0.043, which is

n agreement with those reported for nc-Ni in the literature [10].t may be pertinent to mention that in fcc metals, the nano-sizerains are reportedly more sensitive to the rate of loading than theoarse grains with the former usually exhibiting an order of mag-itude higher m values. This has been attributed to the enhancedrain boundary mediated process in the nc-fcc metals [22]. Thetrain rate sensitivity was found to decrease with the increase inrain size and found to be ∼0.027 for the grain size of 250 nm. Theigh grain boundary density in nc-materials usually support grainoundary mediated process [22]. However, the value of m obtained

n present study was much lower and hence diffusion controlled

oble creep and grain boundary sliding mechanisms do not appearo be the deformation mechanism. Instead, the decrease in value of

with the decrease in grain boundary density is indicative of theact that dislocation-grain boundary interactions are responsible

ngineering A 539 (2012) 324–329

for high hardness and high strain rate sensitivity with smaller grain

size. The activation volume,∗V , was determined using the following

relation [23]:

∗V = 3

√3KT

(� ln

•ε

�H

)(2)

where K is the Boltzmann’s constant and T is the absolute tempera-

ture. The∗V obtained for nc-Ni as a function of grain size is shown in

Fig. 7(d). The activation volume obtained for as prepared Ni (60 nm)in the present study was found to be in good agreement with thosereported by Wang et al. [10] for nc-Ni with 30 nm grain size usingstress relaxation technique. Fig. 7(e) shows a clear trend for signifi-

cant decrease in∗V with the increase in hardness for different length

scale i.e. from submicron to nano-sized grains. For coarse grained

fcc metals,∗V usually lies in the range of 100–1000 b3 correspond to

cutting of forest of dislocation as the rate controlling mechanisms[21]. However, with the decrease in grain size to ultra fine or nanosize grains, high density of grain boundaries serve as the obstaclesto the motion of dislocation resulting in change in rate limiting pro-

cess [24]. In the present study, increase in∗V with increase in grain

size is indicative of the grain boundary mediated mechanism as therate limiting process.

4. Conclusions

Grain growth kinetics and influence of grain size on the defor-mation behavior have been studied for nc-Ni. On the basis of theresults obtained, following conclusions can be drawn:

1. The DSC results showed a significant increase in grain size aboveCurie temperature. Resistivity measurements at different tem-perature showed time for grain growth decrease with increasein temperature.

2. The grain growth behavior followed well-known Arrhenius typerelation with a transition in kinetics at Curie temperature. Theactivation energy calculated for nc-Ni showed that the graingrowth was controlled by grain boundary diffusion below Curietemperature and by volume diffusion above Curie temperature.

3. The nanohardness measurements confirmed an inverse relation-ship between hardness and grain size. Strain rate sensitivity (m)was found to decrease with increase in grain size.

4. The activation volume for nc-Ni was also found to be much lowerthan the activation volume for coarse-grained Ni. The activationvolume was found to increase with the increase in grain size.This observation suggested grain boundary mediated process asthe rate controlling mechanism.

Acknowledgments

The authors would like to thank Dr. A.K. Suri, Director, Mate-rial Group, BARC for the encouragement provided in carrying outthis work. The support provided by Prof. M. Sundararaman (Univ.of Hyderabad) and Dr. R.N. Singh Head, Fracture Mechanics Sec-tion is highly appreciated. Research fund provided by PRF-BRNS incarrying out this work is also greatly acknowledged.

References

[1] C. Detavernier, D. Deduytsche, R.L. Van Meirhaeghe, J. De Baerdemaeker, C.Dauwe, Appl. Phys. Lett. 82 (2003) 1863–1865.

[2] A.J. Haslam, D. Moldovan, S.R. Phillpot, D. Wolf, H. Gleiter, Comput. Mater. Sci.23 (2002) 15–32.

[3] G.D. Hibbard, K.T. Aust, U. Erb, J. Mater. Sci. 43 (2008) 6441–6452.[4] L.Z. Zhou, J.T. Guo, Scr. Mater. 40 (1999) 139.[5] C. Suryanarayana, Inter. Mater. Rev. 40 (1995) 41.

and E

[[[[[

[

[[[[

[[21] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564.

G. Sharma et al. / Materials Science

[6] K. Lu, Mater. Sci. Eng. R16 (1996) 161.[7] M.J.N.V. Prasad, S. Suwas, A.H. Chokshi, Mat. Sci. Eng. A 503 (2009) 86–91.[8] Y.M. Wang, A.V. Hamza, E. Ma, App. Phys. Lett. 86 (2005) 241917.[9] B. Xu, Z. Yue, Xi Chen, J. Phys. D: Appl. Phys. 43 (2010) 245401.10] C.I. Wang, M. Zhang, T.G. Nieh, J. Phys. D: Appl. Phys. 42 (2009) 115405.11] C.C. Koch, J. Phys. Conf. Ser. 144 (2009) 012081.12] M.A. Meyers, A. Mishra, D.J. Benson, Prog. Mater. Sci. 51 (2006) 427–556.

13] M.E. Day, M. Delfino, J.A. Fair, W. Tsai, Thin Solid Films 254 (1995) 285–290.14] A. Emre Yarimbiyik, Harry A. Schafft, Richard A. Allen, Mark D. Vaudin, Mona

E. Zaghloul, Microelectron. Rel. 49 (2009) 127–134.15] M.J. Aus, B. Szpunar, U. Erb, A.M. El-Sherik, G. Palumbo, K.T. Aust, J. Appl. Phys.

75 (7) (1994) 3632–3634.

[[[

ngineering A 539 (2012) 324–329 329

16] V. Stary, K. Sefchik, Vacuum 31 (1981) 345–349.17] V. Weihnacht, W. Bruckner, Thin Solid Films 418 (2002) 136–144.18] U. Klement, M. da Silva, J. Alloys Compd. 434–435 (2007) 714–717.19] I. Kaur, W. Gust, L. Kozma, Handbook of Grain and Interphase Boundary Diffu-

sion Data, Ziegler Press, Stuttgart, 1989.20] G. Neumann, V. Tolle, Phil. Mag. A 54 (1986) 619.

22] Robert J. Asaro, Subra Suresh, Acta Mater. 53 (2005) 3369.23] C.D. Gu, J.S. Lian, Q. Jiang, W.T. Zheng, J. Phys. D: Appl. Phys. 40 (2007) 7440.24] M. Dao, L. Lu, Q.A.J. Asaro, J.T.M. De Hosson, E. Ma, Acta Mater. 55 (2007)

4041.