Journal of Space Technology, Vol 1, No.

1, July 2012

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation
M. Zeeshan Siddiqui, Fawad Tariq, Nausheen Naz, M. Fahad Ahmed
Mostly, the Youngs modulus E is derived from the slope of
the linear part of the stress-strain curve during uniaxial tensile
testing. For getting accurate values of E, the strain data should
be reasonably accurate which is generally obtained from either
using foil strain gages or clip-on extensometers. However, the
use of strain gages and extensometers is sometimes limited
because of cost involved, and the size and shape of specimen
under test. Beside this traditional technique, several other
methods are documented in literature for accurate
determination of Youngs modulus (E) of materials. One is
ultrasonic pulse-echo technique [1]-[3]. In this method, an
ultrasonic beam is generated by a transducer which is attached
to the measured materials. By measuring the sound velocity in
the medium, the elastic properties can be determined. The
other is the resonance method [4], [5]. By measuring the
resonant frequency, elastic properties can be calculated
because resonant frequency is related to the structures
geometry as well as the elastic properties. Indentation methods
(micro and nano) are also widely adopted for determining
youngs modulus where value of E is determined directly from
indentation load and displacement data obtained during one
cycle of loading and unloading [6]-[11]. This method is very
famous because relatively small amounts of testing materials
are needed and there are no strict requirements for sample
shape. Some authors have also successfully used nondestructive Eddy Current technique (combined a/c and d/c
magnetic fields) to evaluate Youngs modulus of solids [12],
[13].
In the present paper, the direct measurement approach has
been employed wherein the strain measurement has been done
though Digital Image Correlation (DIC). From the outcome of
this work it is expected that this technique will prove very
useful for determination of Youngs modulus, particularly in
those cases where the use of strain gage or extensometer is not
applicable.

AbstractAccurate determination of Youngs Modulus of

materials is important for structural design, analyses, as well as
for quality control. The intricacies involved in such experiments
are generally related to the accurate measurement of strains in
the elastic region. For metals and composite laminates, the
problem is generally dealt with the use of extensometers or strain
gages. In case of fiber tows, such simple solution is often
restricted by the small specimen size and low bending rigidity of
the specimen. This paper presents an alternative approach for
the determination of in-plane Youngs modulus of these materials
by using Digital Image Correlation (DIC). Uniaxial tensile
experiments were performed with specimens made of aluminum
alloy, E-Glass/Polyester laminate and carbon fiber tows with
strain measurement done through an indigenously developed two
dimensional DIC algorithm in MatLab. The strain measurements
were validated for aluminum alloy and composite specimens by
using foil-type electric resistance strain gages. The elastic
modulus computed with these strain values were then compared
with literature data. The comparison shows good agreement with
the reference values. The proposed technique offers a quick, costeffective and widely adaptable alternative of traditional methods
for determination of elastic modulus of both homogeneous as well
as heterogeneous materials.
Index TermsComposites, Digital Image Correlation, Optical
Strain Measurement, Youngs Modulus

I. INTRODUCTION

HE knowledge of elastic constant of materials is

indispensable for structure design and quality control
purposes. Thus its determination is of utmost importance to all
engineering applications. The Youngs modulus or modulus of
elasticity E is the property which describes the deformation
behavior of a material or structure under loading. Materials
deform differently when loads are applied, and the relationship
between stress and strain is typically defined by the Youngs
modulus. The ability of any material to resist or transmit load
is important, and this property is often used to determine if a
particular material is suitable for a specific purpose.
The value of elastic constants can be experimentally
determined by different methods. Methods of determination
are further classified into Direct and Indirect methods. In
direct methods, these constants are determined from the
longitudinal and transverse deformations as well as by the
effects which depend directly on them; whereas, in the indirect
methods these are determined from the measurement of other
elastic constants like Shear or Bulk moduli. In static methods,
these constants are determined by the deformations produced
by the time dependent or slowly changing forces in contrast to
dynamic methods which often involve vibration studies [1].
For general engineering applications static methods are widely
utilized by engineers and scientist.

II. PROCEDURE FOR MEASUREMENT OF YOUNGS

MODULUS
For a uniaxial tensile test, the Youngs modulus (E) is defined
as the ratio between stress and strain during the elastic region.
For metallic specimens, it is given as:

E"

P
A!

(1)

Where:
P is the load measured during the elastic region through
a load cell;
A is the cross-sectional area of the specimen;
is the elastic strain;
32

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation

computed by (4)
For composite laminates, the quantity E is termed the chord
modulus (Fig. 1) and is calculated as:

Pu % Pl
E"
A ! # u % l $

&& #R

ij

C"

(2)

&& #R

ij % R $

Where:
E is the chord modulus as defined in fig. 1
Pu is the tensile load at upper strain limit;
Pl is the tensile load at lower strain limit;
u is the upper strain limit;
l is the lower strain limit;

MUL
f

(3)

Where:
MUL is the fiber mass per unit length
f is the fiber bulk density
The strain appearing in (1) and (2) is the quantity of interest
that has to be measured experimentally using a suitable strain
measurement device. The strain measurement, in this work,
has been done by extracting the surface strain values from
displacement data acquired through simple experimental setup
consisting of digital camera and numerical algorithm for
Digital Image Correlation devised in MatLab. For comparison,
the Youngs modulus has also been determined by measuring
strain from electric-resistance strain gages and extensometers.
III.

&& #T

ij % T $

(4)

Where Rij, Tij are intensity values at each pixel for the
reference and target images respectively while ,
are
corresponding mean intensities of the entire images. The
correlation surface near the optimum point, which is in the
vicinity of target location, is similar to Fig. 3. Due to the
limited resolution of most imaging systems, integer pixel
locations found with this coarse search is not sufficient and
sub-pixel level matching is required to get accuracies in
micro-strain range. Some type of image interpolation is
needed to achieve this objective. Figure 4 shows the
interpolation schemes which are normally used. In this work,
cubic spline interpolation was used due to the uniform
transitions from pixel to pixel gray values achieved with this
interpolation scheme.
Once the algorithm reaches in the vicinity of target point, the
sub-pixel accurate location of target point is calculated by
initiating a fine search. The fine search algorithm uses the
response surface of the correlation coefficient in the
neighborhood of target point to estimate its exact location with
sub-pixel accuracy [14].

The cross sectional area A of fiber tows is calculated with

the equation:

A"

% R $#Tij % T $

IV. EXPERIMENTAL SETUP

The experimental setup is similar to the one given in [14 and
consisted of a high resolution Nikon D90 digital camera which
records video streams at HD resolution (1280 X 720 pixels)
and Sony XR-550 high definition camcorder which records
video streams at full HD resolution (1920 X 1080 pixels). Use
of two cameras for strain measurement is not essential but was
done only to compare sensitivity of strain measurements from
different imaging sensors.
Specimen surfaces were spray painted full black (or white) to
form a uniform background. After the base color dried, a fine
mist of white (or black) was sprayed to form a random speckle
pattern as shown in Fig. 5. One camera was placed in front of
each of the two faces of the specimen and the distance was
adjusted between the camera and the specimen to properly
frame the area of interest. For enhancing the contrast,
specimens surface were illuminated by flexible fiber optic
dual halogen lamps (Meiji Techno, Japan). By properly
adjusting the orientation and distance of the two white light
sources, a relatively uniform light intensity was generated on
the specimen. The experimental setup with a Sony camcorder
is shown in Fig. 6.
The materials selected for this study were aluminum alloy
2024-T6, E-glass/polyester laminate and epoxy impregnated
specimens of carbon fiber tows. Tensile properties of the
materials used in this study are listed in Table 1. Rectangular
metallic strips were cut from the longitudinal direction of the
sheets and dog-bone shape tensile specimens were machined.

DIGITAL IMAGE CORRELATION

When an object undergoes deformation, each point on the

surface of the object moves in coherence with its
neighborhood. This coherence introduces a similarity between
the images of object surface before and after deformation.
Digital Image Correlation (DIC) works by matching this
similarity across images of the object surface taken before and
after deformation. The images for correlation matching can be
acquired either by using a simple tripod mounted digital
camera or through a microscope (optical, SEM or TEM).
Selection of image capturing device depends on the
dimensions of the specimen or strain levels to be measured.
As the object deforms, points in the initial or reference image
(left image in Fig. 2) move to new positions in the deformed
or target image (right image in Fig. 2). The correlation
algorithm goes through a set of selected points in the reference
image and finds their corresponding matches in the reference
image based on the optimum value of matching criterion.
Once the positions of all the points have been determined,
smoothing is applied to remove noise, and strain computations
may be carried out. The matching criteria for finding the best
match can be least squares, absolute integral error or
correlation coefficient. In this work, correlation coefficient is
33

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation

The dimensions of the metallic samples were in conformance

with ASTM standard E8-99.
Carbon fiber tow specimens were fabricated from 12k carbon
fiber tow (T700S, Torayca, Japan), resin (epoxy) impregnated
and cured in an autoclave at 150C for 6 h under vacuum.
Special care was given to avoid formation of resin lumps, uneven cross-section or filament breakage during specimen
fabrication. Especially designed end tabs were cast on resinimpregnated tows to prevent slippage or failure within the
grips and to facilitate proper transferring of load (Fig. 7). In
majority of specimens the volume percentage of fiber and
resin was kept approx. 60:40. Dimensions were set in
accordance with ASTM standard D 4018-99 [16].
E-Glass/polyester laminate were prepared by impregnating the
2 stacked plies of E-glass woven fabric (type: plain weave, 0
/ 90 direction, specific mass: 600 gm-2) in unsaturated
polyester thermoset resin using vacuum assisted resin transfer
molding (VARTM) process and cured at room temperature.
Thickness of the samples was about 1.0 mm and other
dimensions were set in accordance to ASTM standard D 303999 [19]. Picture of E-glass/polyester laminate specimen is
shown in Fig. 8.
All the tension tests were performed on 150 kN servo-electric
universal tensile testing machine (Tinius Olsen, UK) under
crosshead speed of about 20 mm/min at room temperature.
Axial alignment of the specimens in the wedge action grips
was done with the help of LCD grid. At least three specimens
of each material were tested.
For validation of measured axial strains electric resistancetype strain gages (Vishay Micro Measurements, USA) of
about 6.35 mm gage length and 350 were used. Strain gage
data was acquired by custom-made LabView 2010 software
interface using National Instrument Data Acquisition System.

The corresponding tensile chord modulus was calculated as

per ASTM standard D 4018-99 for a strain range of 0.1% to
0.6%. Fig. 11 shows the Stress-Strain curve for Eglass/polyester laminate. The tensile chord modulus was
computed by a similar technique as per ASTM standard
D3039-99. The initial part of the curve was excluded from the
chord modulus range to remove the effect of initial bending of
specimen. The average modulus along with observed standard
deviations for carbon fiber tows and E-glass/polyester
laminates are also given in Table II.
As shown in Table II, the measured values of Youngs
modulus for the materials under study are very close to
literature data whereas the standard deviation in DIC results is
on a higher side. This relatively high standard deviation,
although still within acceptable limits, could be due to number
of reasons. One possible reason is the low image quality
because the digital cameras used for this test were not
scientific grade and hence show more noise which could not
be completely removed by Gauss filter. Another possible
reason for this deviation is specimen alignment and bending
which were only ensured visually. A very small misalignment
between the image plane and specimen surface can result in
significant false strains due to the out of plane motion.
VI. CONCLUSION
Application of Digital Image Correlation technique has been
successfully demonstrated for measurement of Youngs
modulus of metallic as well as composite materials. For
aluminum alloy, which exhibits an approximately same elastic
range as the specified strain range for chord modulus
measurement in carbon fiber tow, the modulus values obtained
with this technique are in good agreement with strain gage
data. The modulus of carbon fiber tows was only calculated
with DIC and is very close to the supplier specifications.
Especially for carbon fiber tow, use of such a non-contact
strain measurement technique allows accurate measurement of
elastic constant without interfering with the loading condition.
A contact type extensometer, for example, would cause
significant bending of the carbon fiber tows and may result in
premature failure. The narrow specimen width also excludes
the possibility of mounting strain gages on the specimen,
further highlighting the need for a non-contact strain
measurement technique. Another benefit of using DIC for
strain measurement is the cost reduction due to avoiding
expensive gages. With Digital Image Correlation, it is possible
to take measurements at different locations within the gage
section. This makes it possible to check if there is non-uniform
deformation within the specimen.

V. EXPERIMENTS AND RESULTS

During the tensile tests, images of the specimen surfaces were
acquired at constant time intervals. These images were stored
in AVI format which were imported and processed with our
indigenously developed Optical Strain Measurement (OSM)
utility in MatLab environment. Instead of doing a full field
strain measurement as is normally done in commercial DIC
software only two markers were placed in the axial direction
for displacement read out. The displacement data was then
converted to engineering strain in post processing.
The inherent noise in the DIC results, which stems from the
noise in imaging sensor output, was removed with a Gaussian
filter of 10 pixel length and a standard deviation of 2. Further
averaging of consecutive frames was done to reduce the
effects of small vibrations and light intensity variations.
Fig. 9 shows the elastic part of Stress-Strain curve for an AA
2024-T6 specimen. Due to initial preload and toe effect, the
curve does not pass through the origin; accordingly the fitted
line was not constraint to pass through origin. The slope of
fitted solid line defines the Youngs modulus which is given in
Table II.
Fig. 10 shows the Stress-Strain curve for a carbon fiber tow.

ACKNOWLEDGMENT
We wish to thank Dr. Sajid Mirza (Senior Chief Manager) for
their valuable suggestions and guidance and Mr. Ahmed Bilal
(Chairman SUPARCO) for approval and provision of
facilities. We would further like to acknowledge the technical
assistance and meaningful discussion extended by Mr.
Tasleem Baig (Chief Manager) and Dr. Rasheed Ahmed
34

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation

Baloch (Deputy Chief Manager) throughout the experimental

work.

[2]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]
[17]
[18]
[19]

9-10

1.8

1.75

Poissons ratio

0.33

0.20-0.22

REFERENCES
[1]

Elongation at
fracture (%)

This value depends upon fiber/resin volume percentage and principal

loading direction.
b
This value is for 60% fiber volume percentage. The value of L depends
upon fiber/resin volume percentage and principal loading direction.

N. Parveen and G. V. S. Murthy, Determination of elastic modulus

in nickel alloy from ultrasonic measurements, Bull. Mater. Sci., vol.
34, no. 2, pp. 323326, April 2011
D. Ensminger, Ultrasonics Fundamental, Technology, Applications,
New York: 2nd ed. Marcel Dekker Inc., 1988 [3] P. McIntire, Ed.,
Nondestructive Testing Handbook: Ultrasonic Testing, American
Society for Nondestructive Testing, USA, 2nd ed., vol. 7, 1991
Q. M. Wang and L. Eric Cross, Determination of Youngs modulus
of the reduced layer of a piezoelectric RAINBOW actuator, J. Appl.
Phys., vol. 83, no. 10, pp. 5358-5363, 1998
L. kiesewetter, J.M. Zhang, D. Houdeau and A. Steckenborn,
Determination of Youngs Moduli of micromechanical thin films
using the resonance method, Sensor. Actuat. A-Phys., vol. 35, no. 2,
pp. 153-159, Dec. 1992
J. Gubicza, Determination of Youngs Modulus from depth sensing
Vickers Indentation Tests, Solid. State. Phenom., vol. 56, pp. 195200, 1997
S. Hong, T. P. Weihs, J. C. Bravman and W. D. Nix, Measuring
stiffnesses and residual stresses of silicon nitride films, J. Electron.
Mater., vol. 19, no. 9, pp. 903-909, 1990
G.M. Pharr, W.C. Oliver, and F.R. Brotzen, On the generality of the
relationship among contact stiffness, contact area, and elastic
modulus during indentation, J. Mater. Res. vol. 7, pp. 613-617,
March 1992
J. B. Pethica, R. Hutchings, and W. C. Oliver, Hardness
measurements at penetration depths as small as 20 nm, Philos. Mag.
A, vol. 48, no. 4, pp. 593-606, 1983
K. M. Mussert, W. P. Vellinga, A. Bakker and S. Van der Zwaag, A
nano-indentation study on the mechanical behaviour of the matrix
material in an AA6061 Al2O3 MMC, J. Mater. Sci., vol. 37, pp.
789-794, 2002
T. Rio, Gomez-Del, P. Poza and J. Rodrguez, Micromechanical
characterization of Al 8090/SiC composites by nanoindentation, J.
Mater. Sci., vol. 40, no. 6, pp. 1513-1515, 2005
T. Kwaaitaal and A. J. G. Schoofs, Determination of Young's
Modulus or Poisson's Ratio Using Eddy Currents, Exp. Mech., vol.
19, no. 12, pp. 450-455, March 1979
K. E. Fllstrm and M. Jonsson, A nondestructive method to
determine material properties in anisotropic plates, Polym. Compos.,
vol. 12, no. 5, pp. 293-305, Oct. 1991
M. Z. Siddiqui, F. Tariq and N. Naz, Application of a two step
digital image correlation algorithm in determining Poissons ratio of
metals and composites, presented at Materials and Structures
Symposium, 62nd International Aeronautical Congress, South Africa,
Oct. 2-5, 2011.
ASM
Aerospace
Specification
Metals
Inc.
Available:
[Link]
24t6
Standard Test Methods for Properties of Continuous Filament
Carbon and Graphite Fiber Tows, ASTM Standard D 4018, 1999
Torayca T700S Data sheet, Technical data sheet no. CFA-005,
America
Mechanical Properties of Carbon Fibre Composite Materials,
Available: [Link]
Standard Test method for Tensile Properties of Polymer Matrix
Composite Materials, ASTM Standard D 3039, 1999

TABLE II
MODULUS OF ELASTICITY (IN GPa)
Material
AA 2024-T6

Literature

Strain gages

DIC

72.5

73.40.6

73.52.5

2288

22.42

23.92

Carbon Fiber
230
Tow*
E-Glass fabric
22-25
composite**
All units are in GPa.
* strain range: 1000-6000
** strain range: 2300-6000

Fig. 1. The slope of dotted line is the chord modulus as defined in ASTM
standard D4018-99 [16]

TABLE I
TENSILE PROPERTIES OF MATERIALS USED IN THIS WORK
AA 2024- Carbon Fiber
E-Glass/Polyester
Property
T6 [16]
Tow [17]
Laminate [18]
Youngs Modulus
72.5
230
22-25
(GPa)
0.2% offset Yield
>345
-Strength (MPa)
Ultimate tensile
>427
4900
440
strength (MPa)

Fig. 2. Displacement of an image point after application of load

35

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation

&*)(

&*#

&*#(

&*+

&*+(

'

=>??3016->57=>3@@-2-356

&*+(
&*+
&*#(
&*#
&*)(
&*)
!)&
!'#

!)(

!"&
!""

!#&

!"%
!"$
!#(

,-./012343567-57<79/-:30.;

!"#

,-./012343567-57879/-:30.;

Fig. 3. Correlation coefficient as a function of X and Y displacements

F,7?3/?3.35616->5

A?1B7C10D3.7>@717'&:'&7-41E37?3E->5
"

%&&

%
"&&
$
&
'&

'&

'&
"

& &

'&

G-0-531?7-563?/>016->5
%&&

"&&

"&&

(&
& &

Fig. 2. Experimental setup for uniaxial tensile tests. The digital camera is
positioned in such a way that the area of interest remains in the frame.
Halogen lamps are visible in the foreground illuminating the specimen surface
with controlled light intensity suitable for correlation.

=DH-27./0-537-563?/>016->5

%&&

&
(&

&
(&

(&
& &

Fig. 4. Three dimensional representation of gray values in original and

interpolated images for different interpolation schemes
Fig. 7. Epoxy impregnated specimens of unidirectional carbon fiber tows

Fig. 5. Two types of speckle patterns with white and black background are
shown. As long as the speckle size remains the same, the choice of
background color does not affect the correlation

Fig. 8. E-glass/polyester laminate specimen with strain gage mounted for

measurement of elastic modulus

36

Determination of Youngs Modulus of Metallic and

Composite Materials by Digital Image Correlation

lass/polyester laminate specimen in the

Fig. 5 Stress-strain curve for an E-glass/polyester
strain range for chord modulus with strain measurement done by (a) DIC and
(b) strain gauge

Fig. 3. Stress-strain
strain curve for an Aluminum alloy 20242024-T6 specimen in the
elastic region with strain measurement done through (a) st
strain gage and (b)
DIC

Fig. 4. Stress strain curve for a Carbon fiber tow. Note the essentially linear
behavior up to failure. The strain range for chord modulus as per ASTM
standard D4018 is 0.1% to 0.6%.

37