Qin Y. Micromanufacturing Engineering and Technology

Подождите немного. Документ загружается.

In-situ Testing of Mechanical

Properties of Materials

Fredrik O

stlund, Karolina Rzepiejewska-Malyska, Laetitia Philippe,

Patrick Schwaller and Johann Michler

INTRODUCTION

The mechanical properties of small structures,

typically with dimensions within the range of a

few hundred microns down to sever al microns or

below, cannot simply be extrapolated from the

properties of bulk samples. This is due to two

effects. First, samples used for bulk mechanical

testing usually have dimensions which are much

larger than the micro-structural features, such as

grains or pa rticles. Second, mechanical behavior

is controlled by certain fundamental length scales.

For example, plasticity in met als involves the

motion of dislocations, which are hindered when

they try to pass between obstacles more closely

spaced than about 100 nm, and fracture in brittle

materials is initiated at flaws with a critical size of

several tens of micrometers [1]. The mechanical

properties of a material will fundamentally

change as the sample dimensions become smaller

than these fundamental length scales. It is there-

fore necessary to measure the mechanic al proper-

ties at a length scale comparable to the feature

sizes used in miniature or micro-electromechani-

cal devices. An example is the flow stress depen-

dence on pillar size in micro-compression experi-

ments shown by Uchic et al. [2], who performed

compression tests on Ni pillars ha ving different

diameters ranging from 40 mm down to 5 mm and

found that for the smallest samples the flow stress

is almost three times greater than for bulk Ni.

Accurate prediction of a material’s response

also requires an understanding of the fundamental

mechanisms of material deformation and fracture

in the micro- and nano-scale. It is therefore essen-

tial that small specimens, for instance for tensile

or compression testing, can be manufactured for

which appropriate handling and manipulation

techniques and equipment with appropriate load

and displacement resolution are available. Also it

has to be possible to observe the material under

load in-situ to study plastic-deformation and

crack-propagation mechanisms. The scanning

electron microscope (SEM) is an ideal platform

for in-situ mechanical testing, as it covers a mag-

nification range from 10 to 10

times, exhibits a

relatively large specimen chamber that can accom-

modate mechanical testing equipment and has

analytical capabilities for determining local chem-

ical composition and crystal structure.

In the following sections the reader will be

introduced to the peculiarities of integrating

materials testing within an SEM. The most popu-

lar techniques will be explored, i.e. miniaturized

tensile tests, compression tests and nano-indenta-

tion tests. Then the subject of image analysis

will be touched upon and finally a case study

CHAPTER

Figures 1, 2, 3 and Table 1, plus similarities in this

Chapter are re-produced with permission of MRS Jour-

nal of Materials Research.

331

combining the use of compressive- and tensile-

testing methods will be presented.

INTEGRATING A MECHANICAL

TESTING SET-UP IN AN SEM

At the time of writing there are no fully commer-

cialized products for in-situ materials testing on

the micrometer scale established on the market.

Integrating a mechanical testing set-up within an

SEM rais es some difficulties. In the following

some important details about this process will

be described.

It is of course necessary to use actuators and

load sensors that can deliver the required displa-

cements and measure the low loads encoun tered

at these scales. Additionally, a major limit is the

space available in an SEM chamber. This requires

that the actuators and sensors have small dimen-

sions. In particular, to achieve an optimal perfor-

mance of the SEM, the distance between the

objective lens of the SEM and the sample has to

be kept to a minimum. Typically, the optimal

working distance is between 2 and 5 mm.

The vacuum in the chamber also has to be

taken into account when designing a mechanical-

testing device. All elements of the device have to be

vacuum compatible, meaning that they do not

emit gas in vacuum. Special care should be taken

to avoid, for example, plastics and oil.

Note that some materials which normally do

not emit gas in vacuum can do so when exposed to

an electron beam. Teflon, for example, releases

fluorine when exposed to high energy electrons.

To facilitate the evacuation of the chamber, parts

that contain threads should have extra channels to

allow air to exit the components.

Magnetic materials should be avoided, since

their magnetic fields can affect the electron beam

of the SEM, resulting in a distorted image with bad

resolution. Similarly, electrical components must

have sufficient electrical shielding. All parts close

to the electron beam should also be conductive

and grounded in order to avoid electrical charging.

In particular, this applies to the tip used in com-

pression and indentation tests. Usually these tips

are made of diamond because of the high hardness

requirements. Pure diamond, however, is a very

poor electrical conductor. Therefore, boron-doped

diamond is a preferred choice. Even if these rules

are followed, the maximum resolution of the SEM

will generally not be achieved.

Note that not all experiments have to be per-

formed in-situ: cost and time have to be weighted

against gain of information. In addition, not all

materials can be test ed in-situ. The electron beam

might affect the sample by, for example, breaking

chemical bonds or the sampl e might become too

highly charged to be imaged. Charging problems

can often be solved by deposi ting a thin metal

layer on the samples before the testing. However,

this can potentially change the mechanical prop-

erties of the material.

TENSILE TESTS

Tensile testing is probably the most commonly

used standardized [3] method to determine the

mechanical properties of materials [4].Ina

typical tensile test a specimen having a small, uni-

form cross-section is strained until failure. The

deformation (increase in length) and applied load

are continuously measured during the whole

experiment. The applied engineering stress s is

defined by the load F divided by the initi al cross-

sectional area A

of the sample. The engineering

strain e is defined by the measured length change

Dl divided by the initial length L

s ¼

e ¼

ð1Þ=ð2Þ

A schematic representation of an engineering

stress/strain curve obtained this way is shown in

Figure 21-1.

From such a curve the following properties can

be inferred. The linear increase at low strain is an

elastic deformation that is described by Young’s

modulus E (the slope of the stress/strain curve).

The transition between elastic and plastic defor-

mation is defined by the 0.2% offset yield stress

p0.2

commonly used in materials technology. In

Figure 21-1 a straight line shifted by a strain of

0.002 is drawn parallel to the linear elastic part.

332 CHAPTER 21 In-situ Testing of Mechanical Properties of Materials

The intersection of this line with the stress/strain

curve is taken as the yield stress.

The tensile strength R

indicates the deforma-

tion where a necking of the sample starts. Finally,

the elongation at rupture A indicates the strain at

failure of the sample. The uniaxial deformation is

usually applied using a constant strain rate, i.e.

using a displacement control, in contrast to a load

control. This is because as a sample yields there

can be a decrease in the load. If that occurs with

load control the feedback system will try to re-

establish the load by increasing the strain. This

causes the sample to be rapidly deformed until

failure in an uncontrolled manner.

It is also worthwhile mentioning the difference

between engineering and true stress and strain

values. Engineering stress and strain are defined

by Eqns (1) and (2) as shown above. Both quan-

tities refer to the ini tial cross-section and initial

length. True stress s

true

and true strain e

true

values

take small changes of the deformation into

account starting from an instantaneous elastic or

plastic deformation. s

true

and e

true

can be calcu-

lated as follows:

true

¼ sð1 þ eÞ e

true

¼ lnð1 þ eÞð3Þ=ð4Þ

For samples in the millim eter range tensile

tests can be performed under an optical micro-

scope. If the samples are smaller, the higher r es-

olution of an SEM allows imaging. This is espe-

cially important if the images are used to extract

the strain.

From the image sequence, it is sometimes inter-

esting to observe the propagation of cracks

induced by the tensile strain and to correlate the

crack path or the crack type with other material

properties that can be inferred from SEM images.

Instead of imaging the sample using secondary

electrons, chemical mapping can be done using

energy-dispersive X-ray spectroscopy (EDX) and

the orientation of grains of the sample can be

recorded using electron backscatte r diffraction

(EBSD).

Due to the limited overall size of the test stage,

the sample dimensions also have to be small (typ-

ically only several millimeters length and of a few

100 micromete rs thickness). This means that sur-

face properties gain importance and rough surface

regions or a slightly different composition in the

surface region may influence the measured stress/

strain curves. The most critical issue, however, is

the correct fixing of the tensile test specimen.

Mechanical clamping may be unsuitable because

a deformation could already have been induced

prior to the actual test. A suit able solution is to use

sample holders having the ‘negative’ shape of the

specimen and to fit the specimen inside these hold-

ing forms: an example of this is shown in

Figure 21-2.

FIGURE 21-1 Engineering stress/strain curve for an NiCo

sample indicating the relevant quantities that can be

extracted from the curve (see text for the details).

FIGURE 21-2 A specimen for tensile testing positioned in

the sample holder of a tensile testing set-up.

CHAPTER 21 In-situ Testing of Mechanical Properties of Materials 333

Due t o the small dimensions of the tensile test

specimens, extensom eters or conve ntional strain

gauges cannot be clamped onto the samples.

However,theSEMimagescanbeusedtocalcu-

late strai n values. The quantitative evaluation of

the strain values from SEM images recorded dur-

ing the tensile tests, however, is difficul t. The

total elongation until fracture in the images is

often only a few pixels and specialized image

analysis routines allowing sub-pixel r esolution

have to be used to obtain reliable strain data:

this will be further discussed below. In contrast

to the displacement measurements, the load

measurements can be performed using conven-

tional load cells.

COMPRESSION TESTS

As an alternative and complement to tensile test-

ing, materials can be tested by compression tests.

This is also a method with very old roots and has

the advantage that the specimen does not have to

be attached at the ends of the testing set-up.

Compression tests are analogous to tensile tests:

a c ylindrical sample wi th uniform diameter is

compressed by applying an increasing force on

the ends until it fractures; typically starting

with a purely elastic region followed by plastic

deformation with strain hardening and finally

fracture. Throughout the compression the load

on the sample a nd the deformation are recorded.

Typically, these numbers are then converted

to engineering stress and strain and these are

plotted against each other, providing character-

ization of the material as described above for

tensile test s.

As for tensile tests, it is an advantage if the

actuators can be run with displacement control.

Although there is no necking phenomenon as for

tensile tests, crystalline samples can exhibit a load

drop at the onset of plasticity which can cause an

uncontrolled compression if load control is used.

In addition, the mechanical behavior of a material

is often dependent on the rate of compression,

making constant displacement more logical than

load control.

New load cells and piezo-actuators have

made it possible to perform compression tests

in the sub-micrometer scale [5] . Also vital for

this type of testing is the possibility to manufac-

ture rods or pillars using, for example, lithogra-

phy or focused ion beams (for an excellent

example of compression te sts on FIB-m achine d

metal pillars, cf. [2]).

Figure 21-3 shows an image series from a

compression test on a gallium arsenide pillar

of 4.2 mm diameter. This pillar was manufac-

tured by photolithography and reactive-ion

etching. The crystallographic direction of the

surface is (001). (a) and (d) show the pillar

before and after compression, respectively.

ferent (111) slip planes and (d) reveals cracking

along 110 planes.

The stress/st rain curve recorded during this

experiment is shown in Figure 21-4. The letters

FIGURE 21-3 Image sequence from the video captured during the compression of a gallium arsenide pillar. The total time

of the compression experiment was about 90 s and the frame rate of the video 0.58 frames/s. In (b) plastic deformation along

three different slip planes can be seen. (c) reveals some cracking of the pillar. (d) shows the pillar after compression. The

stress/strain curve for this particular experiment is shown in Figure 21-4.

334 CHAPTER 21 In-situ Testing of Mechanical Properties of Materials

a–d correspond to the images in Figure 21-3.

From this curve it is not possible to see the crack-

ing of the pillar. If the experiment had not been

performed in an SEM it would not have been

possible to deduce whether the cracks or the slip

occurred first. This is an example of the strength

of in-situ testing. In this particular study, the tran-

sition from brittle to ductile behavior of gallium

arsenide as a function of the dimensions was

investigated [7].

Micrometer-sized compression experiments do

not necessarily have to be performed within an

SEM, but doing so offers some advantages. First,

being able to see the sample and tip (compression

punch) facilitates the positioning of the sample

under the tip. A further benefit of this is that there

will not be any doubt as to whether the tip only

partially contacts the sample, which could be pos-

sible with an experiment without direct observa-

tion. Also, contaminants on the sample, which

potentially can affect the experiment, can be

detected immediately.

If the pillar and the tip are badly aligned the

pillar might buckle or bend, thus producing an

entirely different l oad/displ acement curve. Such

effects are easily detected in an in-situ experi-

ment so that experiments showing these effects

can be discarded. By analyzing the video with

image-analysis software (see below) it is possi-

ble to extract more information than just load

and displacement. For example, the lateral

expansion can be measured and from this the

value of Poisson’s ratio can be c alculated. Also,

during a typical compression experiment the

pillar sinks into the substrate. This effect,

called sink-in, can easily be measured. In fact,

because it is possible to m easure the top and

bottom of t he pillar, the strain extracted from a

video will be entirely free of instrum ental com-

pliance.

In principle, the imag e from an SEM only gives

two-dimensional information. This complicates

the approach of the tip to the sample. In order

to make the alignment of the tip over the pillar

possible it is necessary that the set-up be tilted to

the electron beam so that the sample substrate can

be seen (as seen in Figure 21-3). The focus of the

SEM can be used to give an estimation of the

height difference between the sample and the

tip. When the tip is very close to the substrate a

‘shadow’ of the tip can be seen. This shadow

arises from the constrained geometry, preventing

a fraction of the secondary electrons from reach-

ing the detector of the SEM. Most configurations

FIGURE 21-4 Stress/strain curve for a c ompre ssion expe rime nt of a galliu m a rsen ide pillar. T he le tter s refer to the

images i n Figure 21-3. Note that after the initial elastic part of the curve there is a drop in the measured force. This is

common for crystalline samples a nd is cau sed by the introduct ion and subsequent mu ltiplicatio n of dislocati ons within

the material [6].

CHAPTER 21 In-situ Testing of Mechanical Properties of Materials 335

allow for physically touching the substrate with

the tip at a position close to the pillar before the

experiment. After this the distance between the

sample and the tip is known and positioning

becomes easier.

INSTRUMENTED INDENTATION/

NANO-INDENTATION

Instrumented indentation or nano-indentation

is an excellent tool for the determination of

the hardness and Young’s modulus of thin

coatingsorsmallobjects.Inatypicalexperi-

ment a diamond indentation body is pre ssed

into the specimen. The term nano-indentation

is commonly used for indentation depths

ranging from a few nanometers to 100 nm.

The load P and the displacement into the

surface h are continuously measured during

loading and unloading: for an overview, cf.

[8]. An example of such a load/displacement

curve is shown in Fig. 21-5. Young’s modulus

Indent

can be calculated from the unloading

part of the curve, as the unloading is a purely

elastic recovery process. Howe ver, the pro-

cedure is not as straigh tforwar d as for tensile

and compression tests, therefore a method to

determine E

Indent

in more detail is described.

The full procedure is presented by Oliver and

Pharr in [9].

The first step of the Oliver–Pharr data analy-

sis procedure consists of fitting the unloading

part of the load/displacement data to a power/

law r elation derived from contact mechanics

theory:

P ¼ Bðh  h

ð5Þ

where P denotes the applied load, h the pene-

tration into the surface, h

the final displace-

ment after complete unloading (cf. Fig. 21-5)

and B and m are empirically determined fitting

parameters. From this equation, the unloading

stiffness S = dP/dh canbecalculatedbydiffer-

entiating the equation and evaluating it for

h = h

max

S ¼ Bmðh

max

 h

m1

ð6Þ

Using S, the so-called contact depth h

can be

calculated according to:

¼ h  eP=S ð7Þ

where e is a constant depending on the indenter

geometry. For three-sided pyramidal indenters

(Berkovich tips), e = 0.75. Note that the correction

for h

has to be used with some caution because it

is not valid in the case of material pile-up around

an indentation. Using h

,theprojectedcontact

area as a function of the penetration, A(h

)can

be calculated. With S and A, the so-called reduced

Young'smodulusE

can be determined:

ﬃﬃﬃﬃﬃﬃﬃﬃ

ð8Þ

b depends on the indenter geometry and is

equal to 1.034 for Berkovich pyramids. The

reduced Young’s modulus does not take the finite

stiffness of the tip into account. The Young’s

modulus of the sample – E

Indent

– can be extracted

from the relation:

ð1  n

Indent

ð1  n

ð9Þ

Here, n is the Poisson’s ratio of the test mate-

rial and E

(1141 GPa) and n

(0.07) are the

FIGURE 21-5 A typical load/displacement curve from an

indentation experiment.

336 CHAPTER 21 In-situ Testing of Mechanical Properties of Materials

Young’s modulus and the Poisson’sratio,

respectively, of the diamond indenter. It may

seem counter-intuitive that the Poisson’sratio

of the investigated material has to be known.

However, n is about 0.3 for most metals and

even an uncertainty of 0.1 produces an error

for E

Indent

of only about 5%. The scattering of

Indent

values from m easurements on the same

material with fixed measurement parameters is

within 10%.

The hardness of the material is defined by:

H ¼

Aðh

ð10Þ

The largest error source for the calculation of

hardness and Young’s modulus is the expression

used to calculate the area as a function of the

displacement. Only for an ideal (i.e. infinitely

sharp) Berkovich indenter is the relation A ¼

24:5

valid. In reality each tip will be blunted

and may have other defects. This can be corrected

to some extent by using an area function of the

form:

Aðh

Þ¼a

i¼1

ð11Þ

The parameters a

can be obtained by perform-

ing nano-indentation experiments on materi als

with a known value of Young’s modulus.

INDENTATION INSIDE THE SEM

Nano-indentation enables the measurement of

hardness and Young’s modulus of small samples

or thin coatings. However, it is not possible to

have an insight of the deformed zone of the sample

during the indentation process. As a consequence,

no information about possible crack formation or

pile-up (cf. Fig. 21-6) can be obtained. In Fig. 21-6,

an SiO

surface was indented by a cube-corner tip.

The information about cracking can be used to

evaluate the fracture toughness of the material.

Information about pile-up and sink-in can be used

in models of the flow, providing more accurate

data about the material. Another advantage of

in-situ indentations is the possibility to precisely

position the tip with respect to the sample surface.

This makes indentation experiments on small

structures or at specific locations possible. Finally,

there is no need to locate the indent after a test for

measuring the area of the residual imprint. For ex-

situ indentations this can be a very time consuming

task since the size of the imprints is only in the

order of nanometers. In the next paragraph issues

that have to be addressed for in-situ SEM devices

will be briefly described.

As for all in-situ m echanical tests, the SEM

environment induces several ‘design boundaries’

that do not have to be considered in a standard

indentation device working in air. It has to be

ensured that the set-up allows an unhindered

view of the indentation region. The indentation

tip axis normal to the sample surface has there-

fore to be inclined with respect to the beam axis

of the SEM. In addition the indentation tip

geometries which are normally used in indenta-

tion experiments (the four-sided Vickers pyra-

mid or t he three-sided Berkovich pyramid [10])

cannot be used because they have too large an

opening angle (140.6 degrees). For measure-

ments inside the SEM, sharper indenters such

as the c ube-corner (opening angle 84.6 degrees)

have to be used.

Figure 21-7 shows a custom-mad e instrumen-

ted indentatio n device working inside the SEM

FIGURE 21-6 In-situ indentation showing cracking and

pile-up.

CHAPTER 21 In-situ Testing of Mechanical Properties of Materials 337

[11]. The indentation head (4) is composed of a

parallel mechanism (flexure hinges) that holds the

diamond tip. The head is drive n by a stack piezo

(5) with 20 mm range and a built-in displacement

sensor (strain gauge). The indentation head and

its actuator are assembled on a coarse positioning

stage (6) (also a flexure mechanism) and driven by

a fine-pitch pr ecision screw, remotely controlled

with a cable connected to a knob installed on the

SEM chamber’s door. The stage is fixed on the

main body by a dovetail sliding bearing. An XY

slip-stick piezo stage (2) holds the load cell (1) and

the sample (3).

As an application example, Fig. 21-8 shows the

result of an indentation experiment of a thin

nano-composite coating. Whereas the load/dis-

placement curve does not show any irregularities,

the SEM images recorded simultaneously show

the formation of pile-up and the formation of a

crack upon unloadi ng. Interestingly, the crack

almost closes completely after full unloading

and it woul d therefore be very difficult to detect

it by inspection of the residual impression only.

This demonstrates nicely the added value that can

be gained by in-situ SEM indentation experi-

ments.

IMAGE ANALYSIS

Most modern SEMs allow the recording of a

video, which can be later used to extract the

FIGURE 21-8 In-situ SEM indentation of a thin nano-compositite coating.

FIGURE 21-7 The SEM-micro-indenter: 1 load cell; 2 XY-

positioning table; 3 sample holder; 4 tip holder and adapter;

5 piezo-actuator, displacement sensor; 6 sample coarse

positioning.

338 CHAPTER 21 In-situ Testing of Mechanical Properties of Materials

displacement of different features of the sample.

Rough measurements can be done simply by

directly measuring distances with some imaging

software. However, image-analysis tools can do

this with a far greater resolution. Also, through

image-analysis algorithms it is possible to track

features throughout a whole video automatically

in order to construct, for example, a time vs strain

relationship.

There are several algorithms available for

tracking motion in an image sequence. Here, it

will be assumed that an algorithm called c ross-

correlation is employed. For a description of this

algorithm, cf. [12]. This algorithm takes a small

image and a large image as inputs. T he output

will be a map showing how similar the small

image is to each position in the larger one. By

taking the peak value the position of the smaller

image in the larger image is found. This algo-

rithm works even if there is no exact copy of the

small image within the larger. By selecting a

small feature in the first frame of the video and

trying to find this part in the following frames

gives the displacement of this particular feature.

For the c ase of compressive or tensile tests, if

two features that are close to the ends of the

sample are tracked the strain can be calculated.

This strain will be entirely independent of instru-

ment compliance. As menti oned above, strain is

not the only property that can be measured from

the images. For example, by tracking several

parts on the samples, buckling and lateral

expansion of the samples can be measured.

The cross-correlation algorithm can be

improved, making it possible to track changes

with a sub-pixel resolution , provided that the

image is well focused and that the level of noise

is low. This is important, because the total size of

an image is often about 500 pixels in width. In this

case one single pixel corresponds to 0.2% strain

assuming that the sample covers the whole image.

This low resolution would normally be consid-

ered insufficient.

When working close to the resolution limit of

an SEM, the signal is usually very weak, requiring

a very long integration time for each image. In this

case, the acquisition time for a single frame can be

several seconds. It must therefore be taken into

account also that different parts of the image are

scanned at different times, usually starting from

the top and ending at the bottom.

CASE STUDY: A COMPARISON OF

IN-SITU MICRO-TENSION AND

MICRO-COMPRESSION FOR

STUDYING THE PLASTIC

PROPERTIES OF NANO-

CRYSTALLINE ELECTRODEPOSITED

NICKEL AT DIFFERENT LENGTH

SCALES

In order to evaluate the effects of grain size or

geometrical constraints on measured mechanical

properties, there is a need to understand the influ-

ence of the measurement technique, load distribu-

tion, strain rate, etc., on the measurement values

and to correlate the measurements with deforma-

tion mechanisms. Uniaxial in-situ methods are

interesting tools for this purpose. In the following

case study a comparison of in-situ and ex-situ

micro-compression with in-situ micro-tensile tests

is shown; all the tests being used to study the

mechanical properties of electrodeposited nano-

crystalline (nc) nickel. As the material is nano-

crystalline, it would be expected that the size of

the probed volume does not influence the mechan-

ical properties as long as it is at least of the order

of a cubic micrometer. Micro-tensile tests that

probe a volume of more than 2.10

show

reasonable agreement with results from micro-

compression tests that probe much smaller

volumes of down to a few mm

. In-situ uniaxial

solicitation in compression mode reveals several

advantages for studying stress/strain properties.

In the same way as for tensile meas urements, a

quantitative evaluation of the deformation during

compression through video frame records [13]

was found possible.

The material investigated was electrodeposited

nano-crystalline (nc) nickel. Figure 21-9 shows an

HRSEM picture of the nc Ni surface. The sizes of

the grains vary between 30 and 200 nm with an

average of 50 nm. The inset in Fig. 21-9 displays a

TEM cross-section of the film. The grains have a

CHAPTER 21 In-situ Testing of Mechanical Properties of Materials 339

predominantly columnar structure parallel to the

deposition direction.

The samples for tensile testing w ere manu-

factured by a LIGA process. The 10 mm pillars

for in-situ compression w ere also manufac-

tured by LIGA, whereas the 2 m m pillars for

ex-situ compressi on were produced by FIB

machining.

The in-situ tensile tests were carried out

at a constant strain rate of 0.2  10

3

1

Figure 21-10 shows the average true stress/strain

(s/e) curve (i.e. measurements made prior to neck-

ing) obtained from micro-tensile measurements.

The curve indicates weak strain hardening leading

to an increase of the required stress for further

deformation of the tested specimen. Figure

21-10(a) shows an SEM image of the dog-bone

central section extracted from the video frames

recorded at the beginning of a measurement and

Fig. 21.10(b) represents SEM video frame of the

micro-tensile bars just before fracture, revealing

necking in the central part. An inset on the left

part of the graph shows an optical picture of the

typical dog-bone used.

The in-situ compressive tests were carried out

using a flat diamond punch of 15 mm diameter.

The strain rate was 0.2  10

3

1

. Figure 21-11

displays true s/e curves obtai ned on 10 microme-

ter diameter pillars (aspect ratio 3:1) with in-situ

compressive tests. Figure 21-11(a) and (b) for the

10 micrometer diameter pillar tested represent

the video frames extracted at the beginning and

at the end of a typical in-situ compressive test,

FIGURE 21-10 Average tensile stress/strain curves obtained from four samples. The inset of the left part of the graph shows

an optical picture of the typical dog-bone specimen tested.

FIGURE 21-9 HRSEM image of the nc Ni surface; rough-

ness and grain size are easily identified on this picture. The

inset of the right of the picture shows a TEM bright-field

image of the specimen indicating the growth direction of

the electrodeposition.

340 CHAPTER 21 In-situ Testing of Mechanical Properties of Materials