Comparative Study of the Mechanical Properties of HMA Mixture: Field Vs Laboratory

Comparative study of the mechanical properties of nanostructured thin films onstretchable substratesS. Djaziri, P.-O. Renault, E. Le Bourhis, Ph. Goudeau, D. Faurie, G. Geandier, C. Mocuta, and D. Thiaudière

Citation: Journal of Applied Physics 116, 093504 (2014); doi: 10.1063/1.4894616 View online: http://dx.doi.org/10.1063/1.4894616 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/116/9?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Deformation and fracture behavior of composite structured Ti-Nb-Al-Co(-Ni) alloys Appl. Phys. Lett. 104, 071905 (2014); 10.1063/1.4865930 Microstructure and mechanical properties of Ti–B–C–N–Si nanocomposite films deposited by unbalancedmagnetron sputtering J. Vac. Sci. Technol. A 31, 061401 (2013); 10.1116/1.4815952 Microlattices as architected thin films: Analysis of mechanical properties and high strain elastic recovery APL Mat. 1, 022106 (2013); 10.1063/1.4818168 Mechanisms of reversible stretchability of thin metal films on elastomeric substrates Appl. Phys. Lett. 88, 204103 (2006); 10.1063/1.2201874 High ductility of a metal film adherent on a polymer substrate Appl. Phys. Lett. 87, 161910 (2005); 10.1063/1.2108110

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

Comparative study of the mechanical properties of nanostructured thin filmson stretchable substrates

S. Djaziri,1,2 P.-O. Renault,2 E. Le Bourhis,2 Ph. Goudeau,2,a) D. Faurie,3 G. Geandier,4

C. Mocuta,5 and D. Thiaudiere5

1Max-Planck-Institut f€ur Eisenforschung GmbH, Max-Planck-Straße 1, 40237 D€usseldorf, Germany2Institut P’ (UPR 3346 CNRS), Universit�e de Poitiers, ENSMA, Bd Pierre et Marie Curie,86962 Futuroscope Cedex, France3LSPM, (UPR 3407 CNRS), Universit�e Paris 13, Institut Galil�ee, 99 avenue Jean-Baptiste Cl�ement,93430 Villetaneuse, France4Institut Jean Lamour (UMR 3079 CNRS), Universit�e de Lorraine, Parc de Saurupt, CS 50840,54011 NANCY Cedex, France5Synchrotron SOLEIL, L’Orme des Merisiers, Saint-Aubin, BP 48, 91192 Gif-sur-Yvette Cedex, France

(Received 5 July 2014; accepted 22 August 2014; published online 4 September 2014)

Comparative studies of the mechanical behavior between copper, tungsten, and W/Cu

nanocomposite based on copper disperso€ıd thin films were performed under in-situ controlled

tensile equi-biaxial loadings using both synchrotron X-ray diffraction and digital image correlation

techniques. The films first deform elastically with the lattice strain equal to the true strain given by

digital image correlation measurements. The Cu single thin film intrinsic elastic limit of 0.27% is

determined below the apparent elastic limit of W and W/Cu nanocomposite thin films, 0.30% and

0.49%, respectively. This difference is found to be driven by the existence of as-deposited residual

stresses. Above the elastic limit on the lattice strain-true strain curves, we discriminate two differ-

ent behaviors presumably footprints of plasticity and fracture. The Cu thin film shows a large tran-

sition domain (0.60% true strain range) to a plateau with a smooth evolution of the curve which is

associated to peak broadening. In contrast, W and W/Cu nanocomposite thin films show a less

smooth and reduced transition domain (0.30% true strain range) to a plateau with no peak broaden-

ing. These observations indicate that copper thin film shows some ductility while tungsten/copper

nanocomposites thin films are brittle. Fracture resistance of W/Cu nanocomposite thin film is

improved thanks to the high compressive residual stress and the elimination of the metastable b-W

phase. VC 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4894616]

I. INTRODUCTION

The studies of the mechanical behavior of nanocomposite

thin films are of great interest not only because of their wide

technological applications but also because they offer the op-

portunity for studying the relation between the microstructural

length scale and the mechanical properties.1 In particular,

nanostructured metal thin films on polymer substrates are in

use in many technological applications, such as flexible elec-

tronics2 and polymer metallization.3 In aeronautics, metal/

polymer composites are very interesting thanks to their me-

chanical flexibility, light weight, and low thickness. This kind

of composites shows a larger strain at failure than freestanding

metal film.4,5 Moreover, the combination of different materi-

als in the form of nanolayered systems is known to improve

the mechanical properties of thin coatings.6 These systems

may show interesting electronic, optical, and magnetic proper-

ties.7–9 Combinations of the high thermal conductivity of cop-

per (Cu) and the low thermal expansion coefficient of

tungsten (W) makes W-Cu composites attractive for thermal

management applications.10 In addition to their electrical

properties, these metal composites present interesting me-

chanical strength resulting from the compromise between the

high strength of W and the ductility of Cu.11

The continuing trend of miniaturization in thin film

technology poses tremendous challenges for studying defor-

mation mechanisms at small scales. Indeed, when compared

with bulk counterparts the deformation mechanisms of nano-

structured materials are different.12,13 The size effects on the

plasticity of nanostructured materials have been extensively

studied over the past decade and although there is no funda-

mental understanding of deformation mechanisms, there is a

general agreement that the size effect manifests when film

thickness and/or grain size are smaller than 100 nm.14–16 At

this scale, deformation mechanisms based on grain bounda-

ries deformation processes come into play.12,17 Such tend-

ency has been observed for nanocrystalline films18–20 and

nanoscale multilayer films.21,22 Most of the previous studies

were conducted using conventional methods such as nanoin-

dentation and uniaxial tensile testing which do not reflect the

real thermo-mechanical loading that the thin-film undergoes

during preparation and in use. Mechanical testing under con-

trolled biaxial tension allows to overcome the limitations of

representative deformation modes. This has been achieved

thanks to the biaxial tensile device operating at room temper-

ature at DiffAbs beamline of the French synchrotron

a)Author to whom correspondence should be addressed. Electronic mail:

[email protected]. Tel.: þ33 549 496 726. Fax: þ33 549

496 692.

0021-8979/2014/116(9)/093504/8/$30.00 VC 2014 AIP Publishing LLC116, 093504-1

JOURNAL OF APPLIED PHYSICS 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

radiation facility SOLEIL.23 Combined strain measurements

approach has been applied using X-ray diffraction (XRD)

and digital image correlation (DIC) in order to accurately

characterize the early stages of the mechanical deformation

of nanostructured thin films.24 It has been shown that the de-

formation is transmitted unchanged through the metal film/

polyimide substrate interface even for two materials with

such a high mechanical contrast and where the interface mor-

phology plays an important role on the physical properties of

coatings.25,26 A previous study of gold thin film on Kapton

substrate has also confirmed the complete strain transfer

through the film-substrate interface where the elastic

responses of both film and substrate have been measured

using synchrotron XRD.27

This paper presents a study on the mechanical properties

of three types of thin films deposited on polyimide sub-

strates: Cu single thin film, W single thin film, and W/Cu

nanocomposite based on quasi-isotropic copper disperso€ıdthin film. These three systems have been selected for study-

ing the role of grain size and microstructure as well as cop-

per addition on the ductile/brittle behavior of thin films. The

samples were subjected to equi-biaxial stress tension and

their obtained elastic limit was compared to each other.

II. SAMPLE DESIGN

A. Metal thin films

Thin films were deposited by sputtering method in the

central area of the cruciform polyimide substrate, in a 20 mm

diameter disk. Three series of samples were fabricated: Cu,

W, and W/Cu nanocomposite thin films. Sputtering deposi-

tion was performed at room temperature with an argon ion-

gun sputtering beam at 1.2 keV; the chamber base pressure

was 7� 10�5 Pa, and the working pressure during film

growth was 10�2 Pa. The deposition rate was approximately

of 0.05 and 0.07 nm/s for W and Cu layers, respectively. The

thin films were also deposited on 200 and 650 mm-thick

(001)-oriented Si wafers in order to characterize their corre-

sponding microstructure in the as-deposited configuration.

1. Pure Cu thin film

Pure Cu thin film was deposited with a total thickness of

200 6 10 nm. This value represents ten times the estimated

grain size (see below) and thus ensures to be over the perco-

lation threshold associated with the deposition process.

Texture analysis revealed that Cu thin film grains present

mainly an isotropic h111i fiber texture component.

The as-deposited Cu thin films are subject to a slight

tensile residual stress. The stress magnitude is about

þ100 MPa according to X-ray diffraction stress analysis (see

Table I). The average grain size estimated through Scherrer

approximation is about 20 nm.28

2. Pure W thin film

Pure W thin films were fabricated with a total thickness

of 150 6 10 nm. This value has been chosen to avoid thin

film delamination related to elastic energy release. Indeed,

increasing the film thickness increases the elastic energy

value due to the high compressive residual stress state in the

film (see below). Thin film buckling is observed over 150 nm

film thickness.29 Phase analysis by XRD measurements

shows that W crystallites exhibit two different phases: the

equilibrium pure W phase, called a-W which has a body-

centered cubic—bcc—structure (space group Im-3 m) and a

second one called b-W having an A15 cubic structure (space

group Pm-3n) which is stabilized by a low impurity (O, C)

concentration.30 The volume fraction of the b-W phase has

been estimated to be about 10% by the method given in Ref.

28. This method relies on comparing the integrated inten-

sities (or peak areas) of a a-W peak and a b-W peak.

W-scans (W being the angle between the normal to the

sample surface and the normal to the diffracting planes)

were performed on a-W{110}, a-W{200}, and a-W{211}

diffraction peaks in order to analyze crystallographic texture

of the a phase of W thin films. Also, several h-scans were

carried out in order to determine the crystallographic texture

of the b-W phase in the thin films. Texture analysis show

that the a-W grains exhibit mainly two fiber texture compo-

nents: a-W h110i and a-W h111i, while the b-W grains are

found to be oriented along b-W h100i fiber texture axis.

From calculations using Scherrer approximation, the size of

the a-W grains is estimated to be about 7 nm.

The as-deposited W thin films are subjected to high com-

pressive residual stresses of about �3.0 6 0.2 GPa obtained

by X-ray diffraction stress analysis in a-W phase (Table I).

3. W/Cu nanocomposite based on copper disperso€ıdsthin films

In order to better control the microstructure of W, W, and

Cu have been sputtered alternatively on the polyimide sub-

strate. The effective thicknesses of W and Cu are 3.2 6 0.1 nm

and 0.6 6 0.1 nm, respectively. The thin film was fabricated

with 38 periods which leads to a total thickness of about

TABLE I. Microstructure of the studied thin films deposited on Kapton substrates: K, ti, and rri correspond to period, thickness, and residual stress, respec-

tively, associated to subscript i (W, Cu, or f). W or Cu refers, respectively, to tungsten or copper component and f to the film. Residual stresses in W or Cu

component were determined by XRD measurements.

Effective thicknesses (nm) Crystallographic textures Residual stresses (GPa)

Thin film K tW tCu tf a-W b-W Cu rrW rr

Cu

W … 152 … 150 {111}þ {110} {100} … �3.0 …

Cu … … 221 200 … … {111} … 0.1

W/Cu 3.8 3.2 0.6 150 {110} … … �3.0 …

093504-2 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

150 6 10 nm. Due to the introduction of Cu, W crystallites ex-

hibit only one phase: the equilibrium a-W phase and present

only the a-W h110i fiber texture component. The unique a-W

h110i texture component is obtained from a thickness thresh-

old of Cu determined at 0.6 nm.31 Energy-dispersive X-ray

spectroscopy was used in a scanning electron microscope in

order to determine the atomic concentrations of Cu and W in

the film. The volume fraction of the Cu phase has been esti-

mated to be about 20% of the global thin film volume. The

microstructural morphology of the nanocomposite thin film

was characterized by grazing incidence small angle X-ray

scattering (GISAXS). The analysis of the GISAXS patterns

shows that the thin film can be represented as an arrangement

of Cu nanoaggregates within a W matrix. The Cu clusters are

ellipsoidal with an average diameter close to 5 nm. The calcu-

lated grain size of a-W phase using Scherrer’s formula is also

about 5 nm. However, the diffracting signal from Cu nanoag-

gregates in W/Cu thin films is too weak to allow an accurate

analysis of copper in grain strains.

The W crystallites are subject to a high compressive re-

sidual stresses of about �3.0 6 0.2 GPa obtained by X-ray

diffraction stress analysis in a-W phase (Table I). This value

is equal to the one found for the a-W phase in a pure W thin

film. Then, the number of period has been fixed in order to

reach a total thickness slightly lower than the one of pure W

films (see previous paragraph).

B. Polyimide substrate

The sample structure consists of a cruciform polyimide

substrate coated by a metallic thin film. The substrate acts as

a mechanical support and provides a stronger constraint on

thin film cracking due to its stiffness.32

The metal thin film/compliant substrate structure poses

greater challenges for the development of flexible and

stretchable devices. In order to accurately determine the me-

chanical behavior of such structures in service, it is of utmost

importance to account for the realistic complex stress condi-

tions. To address these tasks, biaxial tests were carried out

using cruciform shape substrate loaded in biaxial device.

The substrate geometry has been optimized using finite

element modeling to obtain homogeneous stress condition on

a few square millimeters of the deposited area.33 Indeed, this

is crucial for the XRD measurements. The chosen substrate is

a KaptonVR

HN from DuPontTM with 125 lm thickness,

20-mm wide arms, and a 5-mm toe weld. The substrates were

ultrasonically cleaned with acetone before deposition.

III. MECHANICAL TEST AND STRAINMEASUREMENTS

The biaxial tensile experiments were performed using

the biaxial tensile device dedicated to DiffAbs beamline at

the French synchrotron radiation source SOLEIL. The exper-

imental setup is illustrated and described elsewhere.23,24 All

in-situ tensile tests were conducted using a step by step load-

ing procedure. The value of the applied load is given by the

load cell for each loading step; the corresponding applied

stress for a load of 100 N is 39 MPa based on the sample

geometry. XRD and DIC measurements were performed to

measure the lattice and true strains, respectively.

A. Lattice strain

The high intensity of the synchrotron radiation and the

use of a hybrid pixel detector (XPAD3.1) allowed the mea-

surement of the lattice strain into thin films in short counting

times with good accuracy.34 Sin2W measurements have been

performed for one azimuthal angle only (U¼ 0� being arbi-

trary fixed for one of the two cross line of the cruciform kap-

ton substrate), since the applied loadings were equi-biaxial.

Bragg peak profiles were obtained by azimuthal integration

of the partial Debye-Scherrer rings.35 Pearson VII function

and linear background have been used to fit the diffraction

diagram in order to extract the 2h peak position as well as

the peak broadening during deformation. The lattice strain

feghkl0W corresponding to {hkl} diffracting planes and a scat-

tering vector with (0,W) orientation has been obtained using

the following equation:

ef ghkl0W ¼

d0W � dð0Þ0W

dð0Þ0W

¼ lnd0W

dð0Þ0W

!¼ ln

sin h 0ð Þ0W

sin h0W

!

¼ e11 � e33ð Þ sin2Wþ e33; (1)

where dð0Þ0W is the reference lattice spacing, and hð0Þ0W the asso-

ciated reference diffraction angle which corresponds here to

the unloaded state. d0W and h0W are the lattice spacing and

the diffraction angle, respectively, for the loaded states. e11

and e33 correspond to in-plane and out-of-plane strains,

respectively.

The evolution of the peak width was used to address the

deformation mechanisms underlying the mechanical behav-

ior of the thin films. The biaxial stresses in thin films during

deformation were determined from the measured X-ray elas-

tic strains.36 In the current study, the monochromatic X-ray

beam energy was 8.8 keV. The focal point was at the center

of the 6-circle diffractometer (corresponding to the sample

position) with 0.3 lm in size (same value in both directions).

Experimental details can be found in Ref. 24.

B. True strain

The true strain of the samples was obtained by meas-

uring the displacement of speckle dots spray-painted on the

rear surface of the polyimide substrate, employing the opti-

cal camera of the testing device which looks at the rear face

of the sample. The size of the captured region of the sample

surface is 9� 6.3 mm2 and corresponds to the size of the uni-

formly strained zone as predicted by finite element calcula-

tions. DIC analysis was used to track the evolution of the

deformation of the polyimide substrate during the tensile

test.37 The DIC method is based on the conservation of the

gray level between two images

f ðxÞ ¼ g ½xþ �uðxÞ�; (2)

where g(x) and f(x) are gray level functions corresponding to

the reference and deformed image configurations at each

093504-3 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

pixel point of coordinate x. �u is the displacement field in the

measured direction.

The true strain is determined in terms of displacement

field, considering small strains and very small rotations

�eij ¼1

2

@�ui

@jþ @�uj

@i

� �; (3)

where (i,j)¼ (x,y), and �ui and �uj are displacement

components.

The results of preliminary experiments combining the

two techniques show that the two strains are measured with

accuracy better than 10�4 in the elastic domain for various

loading paths.24 It is worth noting that the strains reported

herein are very small and typically less than 1.8%.

IV. RESULTS

A. Pure Cu thin film

During the tensile test of pure Cu thin film, equi-biaxial

loadings ranging from 14 to 170 N have been applied to the

film/substrate composite. Only the {220}-planes of Cu have

been monitored from XRD measurements. These (hkl)

planes are selected to fit the following perquisites which are

reliable determination of strains: peak intensity and position

toward large diffraction angle, X-ray elastic constants close

to mechanical elastic constants. Figure 1(a) shows the com-

plete set of ln(1/sin h)� sin2w plots. The slope variation of

the sin2w plots reflects the variation of the stress states within

Cu crystallites during the tensile test. The total stress within

Cu thin film is tensile (positive slope) and increases progres-

sively during the test. As shown in a previous work,38 XRD

measurements have been compared to those obtained by DIC

in order to determine the elastic limit of the film which is

defined as the point from which lattice strain and true strain

differ by more than 0.02%.

The lattice strain has been plotted as a function of the

true strain in Figure 2(a) and is observed to increase monot-

onically. Three distinct deformation domains can be defined

for the pure Cu thin film and are labeled in Figure 2(a). The

linear behavior is elastic (domain I) since the lattice strain is

equal to the true strain (line of slope equal to one). Indeed,

the polyimide substrate behaves elastically up to a true strain

of 4%–5%. Then, the lattice strain continues to increase but

less rapidly than the true strain (domain II). This domain is

twice as larger as domain I in contrast to the two other films

(see Figure 2). Finally, the lattice strain saturates while the

true strain continues to increase (domain III). It is worth not-

ing that the evolution of the lattice strain is monotonous and

its derivative varies continuously which is regarded as typi-

cal for plastic deformation of the polycrystalline Cu. The

elastic limit of the film is quantified following the 0.02% off-

set yield strain. The resulting value is 0.27% which corre-

sponds to a total biaxial stress state of þ650 MPa.

B. Pure W thin film

Equi-biaxial loadings ranging from 14 to 170 N

have been carried out for the pure W thin film. Only the

{211}-planes of a-W have been studied to extract XRD

strains. Figure 1(b) shows the complete set of ln(1/sin

h)� sin2w plots. The slope variation of the sin2w plots

reflects the variation of the total stress within a-W crystalli-

tes during the tensile test. In contrast to Cu thin film observa-

tions, the total stress is first compressive (negative slope) and

its magnitude decreases when the applied load is increased.

Noteworthy, the slope is still negative for the largest applied

load at the end of the tensile test.

Figure 2(b) shows the lattice strain versus true strain for

the W single thin film. Initially, the response of the thin film

FIG. 1. ln (1/sin h) versus sin2W plots for (a) {220}-planes of Cu obtained

for the pure Cu thin film subjected to 30 consecutive equi-biaxial loadings

(from 14 N to 160 N), (b) {211}-planes of a–W obtained for the pure W thin

film subjected to 31 consecutive equi-biaxial loadings (from 14 N to 170 N),

and (c) {211}-planes of a–W obtained for the W/Cu nanocomposite thin

films subjected to 27 consecutive equi-biaxial loadings (from 10 N to

160 N); all the measurements are performed at sample azimuth angle U¼ 0�.The inset shows the total stress (i.e., residual and applied stresses) evolution

during the tensile test.

093504-4 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

(lattice strain) is linear elastic (domain I) as it is equal to the

polyimide substrate response (true strain). Then, the lattice

strain is no longer equal to the true strain but less than the

true strain (domain II). Within this domain, the lattice strain

continues to increase with a change on the slope of the curve

less smooth than for Cu film. Finally, the lattice strain satu-

rates while the true strain continues to increase (domain III).

It is worth noting that domain II is about half that observed

for Cu. Both this reduced excursion and the less smooth evo-

lution are attributed to brittle behavior. Moreover, the disper-

sion of data in the plateau of the domain III confirms this

result. The elastic limit of the film is quantified following the

0.02% offset yield strain. The resulting value is 0.30% which

corresponds to a total compressive biaxial stress state of

�1.7 6 0.4 GPa (applied stress of 1.3 GPa) in a-W crystalline

phase.

C. W/Cu nanocomposite based on copper disperso€ıdsthin films

During this test, equi-biaxial loadings ranging from 10

to 160 N have been applied to the film/substrate composite.

Only the {211}-planes of a-W have been studied by

XRD measurements. Figure 1(c) shows the complete set of

ln(1/sin h)� sin2w plots. Considering this case of a nano-

structured W thin film, the total stress within a-W crystallites

is first compressive as for the W single thin film and

decreases when the applied load is increased. However, the

stress becomes tensile from an applied load of 89 N approxi-

mately which is in contrast with the pure W thin film me-

chanical behavior shown above.

The lattice strain has been plotted as a function of the

true strain in Figure 2(c) where the three deformation

domains are labelled and defined as shown for the W single

thin film. The elastic limit of the film is found at 0.49% fol-

lowing the 0.02% offset yield strain criterion. The resulting

value corresponds to a total compressive stress of about

�0.4 GPa (applied stress of þ2.9 GPa) in a-W crystalline

phase.

V. DISCUSSION

The method used in the present study allows for a

straightforward and reliable determination of the elastic limit

of thin film composites thanks to combined measurements

by synchrotron XRD and DIC techniques. The studied sam-

ples show obviously three distinct deformation regimes as

observed in previous studies of metal thin films.38–41

The elastic domain was followed by energy dissipation

in the film volume as the film/substrate composite was fur-

ther strained until saturation (plateau stress). The obtained

elastic limit for each coating is determined in view of the

microstructure and it is clear that the combination of W and

Cu in the form of a nanostructured composite resulted in an

improved elastic limit. Changes in both coating elastic stress

and peak broadening during loading reveal different defor-

mation mechanisms which will be discussed in the light of

these two parameters.

It should be noted that all the samples were subjected to

the same applied strain conditions which were governed by

FIG. 2. Lattice strain as a function of true strain for (a) pure Cu thin film,

(b) pure W thin film, and (c) W/Cu nanocomposite thin films. The red straight

line has the slope equal to one, e being the x-ray strain and �e the true strain.

FIG. 3. True strain determined by DIC in the Kapton substrate vs. the

applied load for the different studied coatings (W, Cu, and W/Cu thin films).

093504-5 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

the mechanical properties of the substrate. As shown in

Figure 3, the obtained true strain in the Kapton substrate is

similar for the different coatings. This demonstrates the very

good repeatability of the measurements.

We investigated and compared the deformation modes

of each coating. Figure 4 shows the stress-strain curves of

each sample; two types of behavior are highlighted. Let us

notice that the stress values have been calculated from the

slopes of the curves in Figure 1 taking into account the bulk

elastic constants of copper and tungsten.42 The evolution of

stress in the Cu thin film is monotonous with a large transi-

tion domain II which is representative of a ductile material.

In the present case, the Cu thin film behaves elastically up to

an applied strain of about 0.27% in accordance with the line-

arity of the stress-strain behavior. In Figure 6, we plotted the

relative evolution of the diffraction peak width as a function

of the true strain. For the Cu thin film, the peak width

increases slightly within the elastic domain (domain I) which

may be due to the elastic anisotropy inducing elastic strain

heterogeneities.43 In transition domain II (between 0.27%

and 0.75% applied strain), both the stress and the peak width

increase indicating strain hardening effects. In domain III

(beyond 0.75% applied strain), the stress becomes constant

and the peak width increases less strongly with the applied

strain. The peak broadening is usually attributed to an

increased dislocation density and an inhomogeneous plastic

deformation. The present study concerns Cu thin film with a

grain size of about 20 nm. The investigation of the stress-

strain behavior reveals a reduced strain hardening and a

near-perfect plastic behavior of the Cu film. This is in good

agreement with the observations in Ref. 44 for nanocrystal-

line Cu of grain size between 50 and 80 nm. These results

may indicate that the dislocation activity is reduced and plas-

tic deformation is dominated by diffusional grain-boundary

sliding as it has been extensively reported for nanocrystalline

cubic metals with grain size below 40 nm.17,45–48

As shown in Sec. IV A, the evolution of elastic stress as

function of true strain is monotonous in a large transition do-

main (labelled II in Fig. 2). This may be interpreted as evi-

dence that the Cu thin film undergoes some ductility. Such

results have also been observed for similar composites of Cu

thin film-polyimide substrate.49,50

By contrast, the W thin film and the W/Cu nanocompo-

site thin film show a brittle behavior as revealed by the varia-

tion of the x-ray stress (Figure 4) and the peak width (Figure

5) versus the true strain. The stress-strain curves of these

films can also be divided into three deformation domains.

The first domain (I) corresponds to an elastic deformation as

indicated by the linear stress-strain relationship and a con-

stant peak width (up to 0.30% and 0.49% applied strain for

W thin film and W/Cu nanocomposite thin film, respec-

tively). By increasing the strain, the elastic stress increases

less rapidly in the second domain (II) and becomes constant

in the third domain (III). It is worth noting that, in contrast to

Cu, the peak width is almost constant during the entire ten-

sile test for the W thin film (it slightly decreases in domains

II and III for the W/Cu nanocomposite thin film).

Furthermore, the stress plateau in the domain III of these two

thin films shows a slight dispersion of data. The less smooth

variation of stress within the reduced domain II is attributed

to two competing processes: thin film cracking with an

increased crack density versus a continued elastic

FIG. 4. Variation of the x-ray stress as a function of the true strain for the

pure Cu thin film, the pure W thin film, and the W/Cu nanocomposite thin

films. The horizontal dotted line marks the zero stress level. Arrows indicate

the total stress (i.e., residual and applied stresses) for each film at yielding at

0.02% strain offset.

FIG. 6. ln (1/sin h) versus sin2W plots for {321}-planes of b–W obtained for

the pure W thin film subjected to equi-biaxial loadings (only three loadings

are presented); the measurements are performed at sample azimuth angle

U¼ 0�.

FIG. 5. Relative evolution of the peak width as a function of the true strain

for the Cu single thin film, the W single thin film, and the W/Cu nanocompo-

site thin films as a function of the true strain. Only diffraction data of {220}-

planes of Cu and {211}-planes of W are reported here. The horizontal dotted

line marks the reference level.

093504-6 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

deformation of the adherent parts of the film as the applied

strain is increased. The less rapid and smooth stress evolu-

tion in those thin films is a signature of crack nucleation and

propagation as observed in Ref. 41 for Ta thin films on polyi-

mide substrates.40 However, the set-up did not allow surface

imaging of the coating surface to confirm cracking develop-

ment during tensile loading. Moreover, no cracks have been

observed during and after sample unloading. The relaxation

of the polyimide substrate may induce the shrinkage of the

sample and closure of cracks upon full unloading as

observed in copper films on Poly Ethylene Terephthalate

(PET) substrate.51

It is interesting to note that the elastic domain of the

W/Cu nanocomposite thin film is larger than that of the pure

W thin film. This is related to microstructural differences

between the two thin films where the presence of the b-W

phase in the pure W thin film has the main impact on the me-

chanical properties of the film. This is demonstrated by a

reduced fracture resistance (1.3 GPa) compared to that of the

W/Cu nanocomposite (2.9 GPa) where W crystallites exhibit

only the stable a-W phase. Indeed, the elastic stress in the

b-W phase is initially compressive and becomes tensile

before the crack initiation observed in the a-W phase as

shown from the ln(1/sin h)� sin2w plots in Figure 6. This

tensile state in the b-W phase induces the initiation of cracks

in the W thin film at lower strain (0.30%) although the a-W

phase is under a high compressive stress at this stage of de-

formation. In the case of the W/Cu nanocomposite thin film,

the a-W phase exhibits a unique h110i fiber texture compo-

nent and more interestingly the b-W phase is suppressed.

The microstructure change induced by the addition of a

given amount of Cu within the W matrix explains the

enhanced fracture resistance obtained for the nanocomposite

thin film. The onset of cracking is determined at larger strain

(0.49%) where the a-W phase is under a low compressive

stress of about �0.4 GPa. Noteworthy, the compressive re-

sidual stress may delay the crack initiation in the coating

subjected to tensile loading conditions and the coating failure

occurs when the stress is in the tensile regime.52–54

Furthermore, the presence of the Cu nanocrystalline phase

appears to improve the fracture resistance of the thin film.

Indeed, the addition of second phase inclusions can influence

the crack propagation as already observed in Ni embedded

alumina nanocomposite thin films.55 The observed improve-

ment could be explained by crack bridging mechanism,56,57

according to which, the particles, when approached by a

crack, form bridging zones behind the crack front. This

bridging ligament exerts closure stresses, which reduce the

stress intensity at the crack tip. The propagation of cracks,

initiated in a brittle phase, would be stopped by particles that

debond partially from the brittle matrix, thus absorbing

energy and bridging the crack. This would lead to a limited

crack growth.58,59 The high fraction of atoms that is present

at grain boundaries may hinder the motion of dislocations

for W and Cu with grain size on the nanometer range.17 We

therefore assume that the energy dissipation in the film is

due to microcracks generation as it has been confirmed

unambiguously in a previous paper,41 plotting the yield sur-

face of the W/Cu film using different proportional loading

paths. Interestingly, thin film failure occurs by different

modes when non-equibiaxial loadings are applied.60

According to Figure 2(c), the second domain of deforma-

tion for the pure Cu thin film is approximately twice as large

as those of pure W and W/Cu thin films. This significant dif-

ference is caused by the difference in deformation mecha-

nisms. The investigation of the Cu film response revealed

some plasticity. This should be the main reason for the large

deformation of the Cu thin film before saturation of the elastic

strain. It is interesting to note in Figure 4 that the Cu thin film

is initially under a slight tensile biaxial stress, which tends

towards a near zero stress state, whereas the W and W/Cu thin

films are initially under a high compressive biaxial stress

(�3.0 6 0.4 GPa). Therefore, the obtained elastic limit for the

Cu film represents an intrinsic property of the film contrary to

the W and W/Cu thin films which behaviors depend strongly

on the initial residual stress prior to deformation.

VI. CONCLUSION

Controlled biaxial tensile testing with combined strain

measurements using both synchrotron X-ray diffraction and

digital image correlation has been used to test three metallic

thin films: Cu thin film, W thin film, and W/Cu nanocomposite

based on quasi-isotropic copper disperso€ıds thin film. From the

analysis of the stress-strain curves and the peak width, we find

that the thin films behave elastically up to 0.27%, 0.30%, and

0.49% for Cu, W, and W/Cu, respectively. Because the Cu

thin film is initially close to a zero stress state, the obtained

elastic limit corresponds to an intrinsic property. In contrast,

the elastic limits of W and W/Cu thin films are determined by

the high as-deposited residual stress present in these films.

Above the elastic limit, the observations suggest two different

behaviors, namely, plasticity for the Cu thin film showing

20 nm grain size and fracture for W and W/Cu thin films where

the a-W grain size is about 5 nm. Indeed, the Cu thin film

shows a large transition domain to a plateau with a smooth

evolution of the stress-strain curve which is associated to a

peak broadening whereas, W and W/Cu thin films show a less

smooth and smaller transition domain to a plateau with no

peak broadening. The elastic limit of the W/Cu thin film is

�60% higher than that of the W thin film. This enhanced frac-

ture resistance of the W/Cu thin film is due to the change of

the microstructure that plays an important role on this property.

The presence of the high compressive residual stress and the

elimination of the metastable b-W phase are considered as

main factors for the improved mechanical behavior of the W/

Cu thin film.

ACKNOWLEDGMENTS

We acknowledge Yannick Diot and Philippe Gu�erin

from the PPRIME institute for samples preparation and

SOLEIL for provision of synchrotron radiation facilities.

This work was partially funded by the French Government

program “Investissements d’Avenir” (LABEX INTERACTIFS,

reference ANR-11-LABX-0017-01).

1M. A. Meyers, A. Mishra, and D. J. Benson, Prog. Mater. Sci. 51, 427

(2006).

093504-7 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03

2S. P. Lacour, D. Chan, S. Wagner, T. Li, and Z. Suo, Appl. Phys. Lett. 88,

204103 (2006).3W. T. Li, R. B. Charters, B. Luther-Davies, and L. Mar, Appl. Surf. Sci.

233, 227 (2004).4T. Li, Z. Huang, Z. Suo, S. P. Lacour, and S. Wagner, Appl. Phys. Lett.

85, 3435 (2004).5J. Xiao, A. Carlson, Z. J. Liu, Y. Huang, H. Jiang, and J. A. Rogers, Appl.

Phys. Lett. 93, 013109 (2008).6P. A. Gruber, E. Arzt, and R. Spolenak, J. Mater. Res. 24, 1906 (2009).7X. Zhang, M. F. Hundley, A. Malinowski, A. Misra, H. Wang, and M.

Nastasi, J. Appl. Phys. 95, 3644 (2004).8A. Gurevich, Appl. Phys. Lett. 88, 012511 (2006).9Y. Wang, L. Y. Chen, S. M. Zhou, Y. X. Zheng, A. Hu, H. R. Zhai, R.

Naik, G. L. Dunifer, and G. W. Auner, J. Appl. Phys. 81, 5256 (1997).10A. Luedtke, Adv. Eng. Mater. 6, 142 (2004).11P. M. Geffroy, T. Chartier, and J. F. Silvain, Adv. Eng. Mater. 9, 547

(2007).12S. H. Oh, M. Legros, D. Kiener, P. Gruber, and G. Dehm, Acta Mater. 55,

5558 (2007).13G. Dehm, Prog. Mater. Sci. 54, 664 (2009).14A. Misra and H. Kung, Adv. Eng. Mater. 3, 217 (2001).15A. Misra, J. P. Hirth, and R. G. Hoagland, Acta Mater. 53, 4817 (2005).16J. R. Greer and J. T. M. De Hosson, Prog. Mater. Sci. 56, 654 (2011).17P. A. Gruber, J. B€ohm, F. Onuseit, A. Wanner, R. Spolenak, and E. Arzt,

Acta. Mater. 56, 2318 (2008).18Z. Budrovic, H. Van Swygenhoven, P. M. Derlet, S. Van Petegem, and B.

Schmitt, Science 304, 273 (2004).19X. Yu, J. Zhang, Y. Zhang, L. Wang, and Y. Zhao, J. Appl. Phys. 111,

113506 (2012).20G. Guisbiers, M.-S. Colla, M. Coulombier, J.-P. Raskin, and T. Pardoe,

J. Appl. Phys. 113, 024513 (2013).21S. Shao, H. M. Zbib, I. N. Mastorakos, and D. F. Bahr, J. Appl. Phys. 112,

044307 (2012).22J. J. Niu, J. Y. Zhang, G. Liu, P. Zhang, S. Y. Lei, G. J. Zhang, and J. Sun,

Acta. Mater. 60, 3677 (2012).23G. Geandier, D. Thiaudiere, R. N. Randriamazaoro, R. Chiron, S. Djaziri,

B. Lamongie, Y. Diot, E. Le Bourhis, P. O. Renault, P. Goudeau, A.

Bouaffad, O. Castelnau, D. Faurie, and F. Hild, Rev. Sci. Instrum. 81,

103903 (2010).24S. Djaziri, P.-O. Renault, F. Hild, E. Le Bourhis, Ph. Goudeau, D.

Thiaudiere, and D. Faurie, J. Appl. Cryst. 44, 1071 (2011).25G. Palasantzas and J. Th. M. De Hosson, J. Appl. Phys. 93, 320 (2003).26T. H€ochbauer, A. Misra, K. Hattar, and R. G. Hoagland, J. Appl. Phys. 98,

123516 (2005).27G. Geandier, P.-O. Renault, E. Le Bourhis, Ph. Goudeau, D. Faurie, C. Le

Bourlot, Ph. Dj�emia, O. Castelnau, and S. M. Ch�erif, Appl. Phys. Lett. 96,

041905 (2010).28P. Villain, P. Goudeau, J. Ligot, S. Benayoun, K. F. Badawi, and J.-J.

Hantzpergue, J. Vac. Sci. Technol., A 21, 967 (2003).29P. Goudeau, P. Villain, N. Tamura, and H. A. Padmore, Appl. Phys. Lett.

83, 51 (2003).30B. Girault, D. Eyidi, P. Goudeau, T. Sauvage, P. Guerin, E. Le Bourhis,

and P.-O. Renault, J. Appl. Phys. 113, 174310 (2013).31B. Girault, D. Eyidi, T. Chauveau, D. Babonneau, P. O. Renault, E. Le

Bourhis, and P. Goudeau, J. Appl. Phys. 109,014305 (2011).32N. Lu, X. Wang, Z. Suo, and J. Vlassak, Appl. Phys. Lett. 91, 221909

(2007).

33S. Djaziri, D. Thiaudiere, G. Geandier, P.-O. Renault, E. Le Bourhis, P.

Goudeau, O. Castelnau, and D. Faurie, Surf. Coat. Technol. 205, 1420

(2010).34K. Medjoubi, T. Bucaille, S. Hustache, J.-F. B�erar, N. Boudet, J.-C.

Clemens, P. Delpierre, and B. Dinkespiler, J. Synchrotron Radiat. 17, 486

(2010).35C. Le Bourlot, P. Landois, S. Djaziri, P. O. Renault, E. Le Bourhis, P.

Goudeau, M. Pinault, M. Mayne-L’Hermite, B. Bacroix, D. Faurie, O.

Castelnau, P. Launois, and S. Rouziere, J. Appl. Crystallogr. 45, 38

(2012).36D. Faurie, P. Djemia, E. Le Bourhis, P. O. Renault, Y. Roussign�e, S. M.

Ch�erif, R. Brenner, O. Castelnau, G. Patriarche, and Ph. Goudeau, Acta.

Mater. 58, 4998 (2010).37G. Besnard, F. Hild, and S. Roux, Exp. Mech. 46, 789 (2006).38S. Djaziri, D. Faurie, E. Le Bourhis, Ph. Goudeau, P.-O. Renault, C.

Mocuta, D. Thiaudiere, and F. Hild, Thin Solid Films 530, 30 (2013).39S. Van Petegem, L. Li, P. M. Anderson, and H. Van Swygenhoven, Thin

Solid Films 530, 20 (2013).40S. Frank, U. A. Handge, S. Olliges, and R. Spolenak, Acta. Mater. 57,

1442 (2009).41S. Djaziri, D. Faurie, P.-O. Renault, E. Le Bourhis, Ph. Goudeau, G.

Geandier, and D. Thiaudiere, Acta. Mater. 61, 5067 (2013).42V. Hauk, Structural and Residual Stress Analysis by Non Destructive

Methods: Evaluation, Application, Assessment (Elsevier, Amsterdam,

1997).43F. Siska, S. Forest, and P. Gumbsch, Comput. Mater. Sci. 39, 137

(2007).44Y. Champion, C. Langlois, S. Gu�erin-Mailly, P. Langlois, J. L.

Bonnentien, and M. J. H€ytch, Science 300, 310 (2003).45H. V. Swygenhoven, Science 296, 66 (2002).46J. Schiotz and K. W. Jacobsen, Science 301, 1357 (2003).47M. Chen, E. Ma, K. J. Hemker, H. Sheng, Y. Wang, and X. Cheng,

Science 300, 1275 (2003).48X. Z. Liao, Y. H. Zhao, S. G. Srinivasan, Y. T. Zhua, R. Z. Valiev, and D.

V. Gunderov, Appl. Phys. Lett. 84, 592 (2004).49M. Hommel and O. Kraft, Acta. Mater. 49, 3935 (2001).50D. S. Gianola, S. Van Petegem, M. Legros, S. Brandstetter, H. Van

Swygenhoven, and K. J. Hemker, Acta. Mater. 54, 2253 (2006).51O. Glushko, V. M. Marx, C. Kirchlechner, I. Zizak, and M. J. Cordill,

Thin Solid Films 552, 141 (2014).52F. Ahmed, K. Bayerlein, S. M. Rosiwal, M. G€oken, and K. Durst, Acta.

Mater. 59, 5422 (2011).53C. C. Aydıner, D. W. Brown, A. Misra, N. A. Mara, Y.-C. Wang, J. J.

Wall, and J. Almer, J. Appl. Phys. 102, 083514 (2007).54C. C. Aydıner, D. W. Brown, N. A. Mara, J. Almer, and A. Misra, Appl.

Phys. Lett. 94, 031906 (2009).55S. Neralla, D. Kumar, S. Yarmolenko, and J. Sankar, Composites: Part B

35, 157 (2004).56B. Budiansky, J. C. Amazigo, and A. G. Evans, J. Mech. Phys. Solids 36,

167 (1988).57L. S. Sigl, P. A. Mataga, B. J. Dalgleish, R. M. McMeeking, and A. G.

Evans, Acta Metall. 36, 945 (1988).58J. Y. Zhang, G. Liu, X. Zhang, G. J. Zhang, J. Sun, and E. Ma, Scr. Mater.

62, 333 (2010).59L. Liu and X. Chen, Acta. Mater. 56, 6214 (2008).60P.-O. Renault, E. Le Bourhis, Ph. Goudeau, D. Thiaudiere, and D. Faurie,

Thin Solid Films 549, 239 (2013).

093504-8 Djaziri et al. J. Appl. Phys. 116, 093504 (2014)

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:

194.254.166.120 On: Wed, 10 Sep 2014 08:54:03