Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F.

Petrou

EXPERIMENTAL ANALYSIS OF FIBER REINFORCED CEMENTITIOUS

COMPOSITES WITH INCREASED TOUGHNESS
1
2
3
Androula V. Georgiou , Stavroula J. Pantazopoulou and Michael F. Petrou
1

Department of Civil and Environmental Engineering

University of Cyprus
Nicosia, 1678, Cyprus
e-mail: ageorg44@ucy.ac.cy
Department of Civil and Environmental Engineering
University of Cyprus
Nicosia, 1678, Cyprus
e-mail: pantazopoulou.stavroula@ucy.ac.cy
Department of Civil and Environmental Engineering
University of Cyprus
Nicosia, 1678, Cyprus
e-mail: petrou@ucy.ac.cy

Keywords: Fibers, Toughness, Cementitious Composites.

ABSTRACT
Abstract. The emerging synthetic fiber reinforced cementitious composites present advantageous mechanical
properties as compared with normal concrete in terms of tensile strain capacity and ductility. In this paper new
types of deformable cementitious composites are explored, having an extensive strain capacity in tension in the
range of, or exceeding the value of 1%. A critical feature of the material for securing very large strain capacity
is the requirement that cracking stabilization be delayed, so that multiple rather than localized cracking may
develop for a large range of the plastic domain of the response. Preliminary experimental data from sample tests
on cementitious composites with unconventionally high tensile strain capacity and toughness, conducted at the
University of Cyprus are presented. Two types of fiber were used with the cementitious mixes, namely
Polypropylene (PP) and Polyvinyl Alcohol (PVA) fibers. Modulus of rupture tests were conducted to establish
the achieved material toughness. From the tests it was seen that cracking behavior marked by formation of a
large network of fine cracks yielded a superior overall material toughness. Aging counteracted ductility as the
bonds between cement hydrates and fibers strengthened with time, increasing the tendency for crack
localization, thereby causing an apparent reduction in specimen deformation capacities and an increase in
strength. Flexural strengths measured during the tests ranged from 6.5 to 12 MPa, with equivalent flexural
strength ratios based on nominal toughness values in the range of 11 to 52%; rotation capacity of concrete
exceeded 10% in all cases considered prior to the onset of strength deterioration. These preliminary findings
suggest that a whole new generation of Cement based materials is possible where tensile strength and strain
capacities are so high that they provide a totally different context for formulation of concrete mechanics and
consequent reinforcement detailing.

INTRODUCTION
The major weakness of concrete is its poor response in tension. Thus, the entire framework of structural
design is evolving around elimination or mitigation of the effects of tension. In older times the motto of concrete
design was to place reinforcement so that it crosses all possible crack paths; this is the intent behind the addition
of steel fibers the first type of fibers systematically studied as a means of mass reinforcement. Addition of
polypropylene (PP), carbon (CF), and polyvinyl-alcohol fibers (PVA) is a more recent innovation with
promising results as the fineness of these fibers gives a very large number of filaments per unit volume as
compared with steel fibers, and their elastic compliance has shown them more efficient in producing very large
strain capacities under tension. All these developments became possible only in the context of self consolidating
concrete where the viscosity increase caused by the addition of fibers may only partly compromise the

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

rheological properties of the mix during casting. Today the type and content of fibers used vary over a large
range of possibilities, and design may even be optimized in light of the performance requirements of the final
product. One type of Engineered Cementitious Composite (ECC) with reported strain capacities in direct tension
in excess of 5%, developed by Lepech and Li[1] reinforced with surface-treated PVA fibers is used as a point of
reference in the mix design considered in the present investigation. The research presented herein is attempting
to develop such a composite product with the use of the available materials in Cyprus. Two different types of
fibers are used in the study, with appropriate mix designs. The paper summarizes the test results along with an
analysis of the experimental findings.

2 DESCRIPTION OF THE EXPERIMENTAL PROGRAM

2.1 Scope of the Investigation
Cementitious materials are considered to be brittle due to the crack initiation that occurs internally during
hydration, even prior to the application of any external load; cracking concentrates particularly at the point of
contact between gel and aggregates due to the unilateral growth of binding, and is owing to the stress
concentrations due to different material stiffness. Fibers, if bonded properly, can bridge these cracks and transfer
the loads, delaying brittle failure by connecting and widening of the cracks. At the same time, through proper
mix design it is the intention to minimize stress concentrations eliminating coarse aggregate. A large content of
fly ash delays stiffening of the mix necessarily imparted by wetting of the large surface area of the fibers, also
assisted by the addition of superplasticizer.
Fiber reinforced cement composites generally comprise various kinds of fibers (Polyvinyl Alcohol,
Polyethylene, Polypropylene, Steel) in different amounts, separately or in combination. Each type of fiber
develops a different type and intensity of binding with the surrounding matrix and the final products properties
are greatly affected by this property, referred to hereon as specific bond. In the present investigation, two types
of fibers have been used, namely, PP and PVA fibers. The PP fiber had a density of 0.91gr/cm3, Youngs
modulus of 7-9GPa and a tensile strength of 400-500MPa (commercially known as MERAFLEX) with a
diameter of 20m and a length of 8mm. The PVA fibers were 8mm long and had a 39m diameter (commercial
name is Kurallon K-II) (specific surface of 2.5x10-8m2). The nominal tensile strength is 1600 MPa, the Youngs
Modulus is 40GPa and the density is 1300kg/m3.
The two kinds of fibers have completely different behaviors. The PP fibers develop no chemical bonding
with the surrounding concrete and a very low contact friction due to their hydrophobic surface, contrary to the
PVA fibers which are hydrophilic and display high chemical bonding to the cement hydrates (due to a very large
number of free hydroxyls on their surface). This is the reason why if used in the same matrix and amount of
fibers, the effect and contribution of the two fiber types will be different.
The mix design was according with published research on self-compacting ECC[2,3]. Tests were carried out
on cubes, under direct compression, and on prisms under three point loading under displacement control in order
to obtain toughness characteristics of the FRC mixes.
2.2 Matrix Characteristics
Extensive research has been carried out on optimizing the mix design for the cementitious matrix in light of
the requirements for the target self-consolidation and the fibers properties (volumetric content, surface area, and
surface binding). Selection was made from among a variety of proposed mix designs. Thus, for the PVA fibers
the mix design was ECC-M45[2]. This mix design was developed for large scale casting. The materials of the
original mix design consist of ordinary Portland cement (OPC, one part by weight of the total), fly ash (1.2 parts
by weight-p.b.w.), silica sand (0.8 p.b.w.), water (0.56 p.b.w.), high range water reducer (0.012 p.b.w.) and 2%
by volume of fibers.
Adjustments have been made on the materials and quantities according to availability of products in Cyprus.
Thus, the cement used was Portland Composite Cement En 197-1 Cem Ii / A-M (L-S) 42.5 R. This type of
Blended Cement is produced using pure calcite limestone and is more impermeable and dense as compared to
OPC, with a higher degree of workability and reduced plastic shrinkage. The compressive strength at 28 days is
equivalent to that of 42.5 R OPC and superior at 2 and 7 days.
Silica sand used in the ECC-M45 had a maximum grain size of 300m. This material is not available in
Cyprus and therefore a different kind of sand was used. The sand used was sieved to remove all particles greater
than 300m and washed with the 63m sieve to remove all smaller particles since they mostly comprised of high
quantities of clay that would induce instability in the final product. The fly ash was of type F as used in the
original design.
In the case of the PP-fiber reinforced mix, proportioning was adjusted to the one proposed for HTPP fibers [3]
due to the same surface binding property (i.e. hydrophobia) of the two fibers. The mix was modified with the use
of a small amount of sand, whereas none had been used in the reference mix.

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

Additional adjustments were made to the quantity of the high range water reducer (HRWR) used as it was
proven by experiments that for the specific HRWR available in this experimental investigation, approximately
twice the amount was required to achieve the required self-consolidating performance of the mix.
2.3 Mix design
The mix designs proposed in the reference studies[2,3] for the two types of fibers along with the one adopted
in the present investigation are listed in Table 1. Note that the surface properties of the fibers are considered in
the design; for example in the mix with the hydrophobic PP fibers a higher amount of fly ash was required.
Mix
Fly Ash Sand (fine
Cement
Designation
(Type F) silica sand)
M45[2]
testM45
PP[3]
testPP

1.0
1.0
1
1.0

1.2
1.2
2.8
2.8

0.8
0.8
1.12

Water
0.56
0.56
0.88
1.37

High Range
Fiber
Water Reducer
(Vol. %)
(HRWR)
0.012
0.02
0.024
0.02
0.016
0.02
0.024
0.02

Table 1 : Proposed and adopted mix designs (by weight unless stated otherwise)
2.3 Fabrication of the specimens
To prepare the PP1 mix sand was mixed with two thirds of the water for 1min and subsequently, cement, fly
ash and HRWR with the remaining one third of the water were added mixing continuously for 3min; lastly, the
fibers were added slowly and mixed for another 6min. This sequence presented some difficulties in that the
volume of the final composite was much less than the capability of the available mixing drum and from the start
it was noticed that the material added (only sand and water) could not be thoroughly mixed. The procedure had
to be stopped several times so that the contents of the drum would be manually mixed.
Prior to the addition of fibers the mix was clearly very flowable. After the fibers were introduced the mix
became evidently viscous with clumping of the fibers. Additional water was added in the mix. The additional
water made the mix more flowable, and reduced the stiffness that was introduced by the fibers. It was attempted
to compact the material when placed in the molds but it was obvious that the mix did not behave well to that.
Some compaction was obtained by tapping on the walls of the mold from the outside. Due to the poor
flowability, cavities of air were trapped on the side surfaces of the molds for the PP1 mix design, as became
obvious after demolding. The specimens were demolded more than 24hours later, (but the mix took several more
days to fully harden), and were placed in water for 28days.
In the second mix design (PVA-M45) a deferent mixing sequence was tested for comparison. The amount of
composite was too low for the specific drum dimensions so all the dry materials were firstly added together.
After mixing for 2-3min the full amount of water and half the superplasticizer were added and mixed thoroughly
until no clumps were visible. At this phase the mixing was stopped and the cement showed signs of very good
flowability. Then the fibers were slowly added with the rest of the superplasticizer and mixing was continued till
they were evenly distributed.
Prior to the addition of fibers the mix seemed to have self consolidating properties. As soon as the fibers
were added, even though in the relatively small percentage of 2% by volume, the mix became instantly less
flowable and the addition of extra water reducer (super-plasticizer) was necessary this time.

TESTS ON HARDENED CONCRETE

Usually when specimen size is very small from the stage of casting a marked bias emanates from the
tendency of the fibers to get layered in a two-dimensional plane rather that acting as actual mass reinforcement.
This was seen in the specimens of the first mix, and therefore in subsequent tests specimens were cut off from
bigger blocks to secure the three-dimensionality of dispersion. Only early samples of the PVA1 mix which were
cast directly in small molds (approximately 8mm equal to the length of the fiber) may be assumed to suffer from
a 2-D dispersion of fibers these are considered in this presentation in a qualitative light only.
Prior to testing the density of the mixes was measured. The specimens were weighted and according to their
volume the density is listed in the table:

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

material mix design

density (g/cm3)

PP1

PVA1

cement

3.14

1.00

1.00

Fly Ash

2.248

2.80

1.20

Sand

2.61

1.12

0.80

Water

1.37

0.56

Superplasticizer

1.05

0.02

0.02

Fibers (kg/0.015m3 of cement)

0.91 / 1.3

0.4

0.4

1797.82

1928.26

density (kg/m )

Table 2 : Density of hardened cementitious mortar

3.1 Compression
Direct compression tests were conducted on small concrete cubes (50mm) from the PP1 mix, that were
saw-cut from the bigger prism specimens. The compressive strength is calculated as the maximum load at failure
divided by the loaded area of the specimen. The load was applied at a constant rate of 0.4 MPa/s.

Figure 1. PP1B compaction tests-failure cracks

Under pure uniaxial compression the failure cracks generated were approximately parallel to the direction of
the applied load with some cracks forming at an angle. Friction at the end platens due to restrained lateral
expansion of concrete combined with the small aspect ratio of the specimen led to the development of a three
dimensional state of stress as evidenced by the inclined cracks near the ends. Note however that in no case was
there any spalling and deterioration of the test block, a result that is in sharp contrast with the reported behavior
of concrete cubes after crushing in the case of conventional concrete. Collapse in those cases would be marked
by the excessive lateral expansion beyond attainment of strength; this type of failure is avoided herein owing to
the confining action of the fibers.

3.2 Modulus of Rupture-Three point bending

Flexural strength of the concrete specimens was determined by the use of three-point loading on simple
prisms (having various thickness values, ranging from 8 to 20 mm, and a span length ranging from 80-130mm).
This testing procedure produces flexural strength values that tend to be higher than those obtained through fourpoint bending. The load was at first applied using force-control up to the peak value and then the procedure was
continued under displacement control sustaining an approximately constant rate. The specimens characteristics
are listed in Table 3:
BEAM DIMENSIONS (mm)
b
d
L (support)
days from cast

PVA1
71.37
8.07
103.50
4

PVA1A
55.68
8.19
80.25
28

PVA1B
53.20
8.75
80.25
28

Table 3 : Specimen size and age of testing

PP1A
50.50
18.70
130.15
50

PP1B
50.50
18.70
130.15
58

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

Figure 2. PP1A large deflection and fibers bridging crack action

The polypropylene specimens cracked upon attainment of peak strength but continued to be able to withstand
high values of load carrying capacity, very close to the peak, whereas additional cracking occurred at higher
deflection values. The fibers were visible bridging the crack opening without complete pullout and from their
stretched appearance it was evident that transfer of stress occurred throughout the crack height. The crack
progressively widened from the crack tip to the bottom edge of the specimen. At higher deflections the top side
of the specimen exhibited compression cracks but the material was still capable of sustaining high values of
stress. During displacement control displacement was increased in steps; for each increment, the specimen
reached a local peak force which immediately decreased by a margin of about 20%, possibly owing to
redistribution of internal stresses along the fibers.

Figure 3. PVA1 multiple cracking 3 point bending (4 days)

Specimens of all mix designs showed high values of deflection up to approximately 8mm. The PVA1 mix
specimen at 4 days exhibited a very ductile behavior with a high tensile strength. The basic difference from the
PP specimens is in the formation of multiple cracking with very small crack width extending over a wide process
zone rather than being concentrated under the load. The fluctuation of the load at the stations after a brief
displacement increment was much smaller as compared with the PP specimens.

Figure 4. PVA1a,b 3 point bending (28 days)

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

However, this characteristic disappeared when the same mix (PVA1) was tested at a later stage (28 days);
rather, a significant reduction of ductility with localization of deformation in a single crack was noted, followed
by an extreme rise of strength (almost double). The fibers bonding in the matrix increased with age altering the
deformation pattern of the individual fibers: from controlled pullout which led to distributed cracking in the early
stage, bond increased with age. At 28 days the adhesion was strong enough to render the fiber itself the most
compliant component of the mix whereby crack opening now occurred due to fiber elongation rather than
pullout, leading to brittle fracture of the fibers and severe localization of all deformation in a single crack. Even
in this case, however, the strain capacity attained by the composite was about 100 times more than the value
expected for conventional normal strength concrete whereas the tensile strength became much higher.
Note that the average fibre stress, ff , at a smeared axial strain of f, is equal to the macroscopic tensile
strength of the material after cracking, multiplied by the number Nf of fibres crossing a unit sectional area of the
final material (a value of 25Vf/(df2lf/4) has been proposed[4] in the past, where Vf the volumetric ratio of fibres
(here 2%)). Thus, for an average material tensile strength of 6 MPa, the multiplier would be 52.345, and the
corresponding fibre stress, 6x52.345=314 MPa. It has been shown[4] that ff = min {2ublf/df ; kff}ffu; where, ub is
the bond stress developing along the fibre surface (it changes with aging), kf is the axial stiffness of the fibre, and
ffu the fibre rupture strength. For example in the case of the PVA fibre, with an lf/df ratio of 8000/39, and kf=40
GPa, ffu<1600 MPa, the fibre stress at a strain of 1% corresponds to a fibre tensile stress of 400MPa, whereas the
corresponding fibre stress due to bonding would be, 205 MPa and 410 MPa for ub values of 0.5MPa and 1MPa,
respectively; although these values for bond are postulated, the range illustrates that in one case response would
be controlled from bond whereas in the other case it would be controlled from fibre axial deformation.

RESULTS

4.1 Compressive strength

In general, the compressive properties of ECC are compared to those of normal concrete, reportedly ranging
from 20 to 95 MPa, while the elastic modulus is lower (15-35 GPa) due to the lack of coarse aggregate in the
mix. The results for PP1 at 58 days are given in the following table:
dimensions

a (mm)

b (mm)

h (mm)

Pmax
(kN)

max (Mpa)

PP1C

48.2

48.87

52.2

81.08

34.42

PP1D

47.8

49

52.1

84.36

36.02

Table 4 : PP1 Compressive strength

4.2 Modulus of Rupture
The modulus of rupture is calculated according to the ASTM Standard C293[5] as follows:

3PL
2bd 2

(1)

Where:
R is the Modulus of rupture, MPa, P is the maximum applied load indicated by the testing machine, (in N), L is
the span length (in mm), and b/d is the average cross section aspect ratio (width/depth) of the specimen (in mm).
PVA1

PVA1A

PVA1B

PP1A

PP1B

Pmax (N)

200

378

352

698

632

R (MPa)

6.69

12.18

10.40

7.72

6.99

Table 5 : Modulus of rupture

The load-deflection curves obtained from the three-point loading tests are compared in Figure 3:

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

700
PP1B-58days
PVA1-4days

600

PP1B-50days
500

PVA1A-28days
PVA1B-28days

P (N)

400

300

200

100

0
0

10

(mm)

Figure 3. Comparative Force-Deflection diagrams

4.2 Flexural Toughness
Fracture Energy and Residual Strength Ratio are calculated according to ASTM Standard C1609[6] for
the three point bending tests. The peak strength and toughness is used to describe in a comparative sense the
flexural toughness of the material. Specimen toughness is measured up to a specified deflection by the absorbed
energy (A.E.). The values of AE depend directly on the size and shape as well as the loading arrangement. The
method used to fabricate the specimens also appears to have an influence (cut from the bulk as compared with
wet-layering in molds). Also the specimen aspect ratio plays a significant role due to the maximum crack widths
produced with a given mid-span deflection and the commensurate degree of pull out and extension of the fibers
to bridge the cracks.
From the load-deflection curves the toughness TD15 is obtained as the total area under the load-deflection
curve up to a net deflection of L/15 (a drift ratio of 13%). Both Polypropylene and PVA specimens (28 days) did
not reach up to that deflection level (the experiment was terminated to avoid separation of the fragments) so the
fracture energy was calculated up to the maximum available value. The PVA1 (4 days after casting) sustained
higher values of drift ratio but calculations are made to the same net deflection ratio for the sake of comparison.
PVA1

PVA1A

PVA1B

PP1A

PP1B

L/15

6.90

5.35

5.35

8.68

8.68

fracture energy (J=Nm)

1.08

0.36

0.30

3.28

3.54

Table 6 : Absorbed Energy

In order to compare the results from different specimen sizes the equivalent flexural strength ratio was
calculated using the first peak strength and toughness determined above. This ratio took into consideration the
specimens dimensions.

RTD,15

R(equivalent strength ratio)%=

150T15D
100%
fbd 2

(2)

PVA1

PVA1A

PVA1B

PP1A

PP1B

52

12

11

36

43

Table 6 : Residual Strength Ratio

Androula V. Georgiou, Stavroula J. Pantazopoulou and Michael F. Petrou

11 CONCLUSIONS
This paper presented results from a preliminary experimental investigation on sample tests conducted on
engineered composites with unconventionally high tensile strain capacity and toughness. Two types of fiber were
used with the cementitious mixes, namely Polypropylene (PP) and Polyvinyl Alcohol (PVA) fibers. Modulus of
rupture tests were conducted to establish the achieved material toughness. From the tests it was seen that best
crack control, marked by formation of a network of fine cracks yielding superior material toughness was
provided by the PVA fiber on relatively young concrete (4 days old specimens made with early high strength
cement). This response was eliminated, however, in identical specimens aged to 28 days, whose failure was
marked by the formation of a single crack under the point load. A similar type of failure mode but much better
toughness was obtained by the PP mixes. Flexural strengths measured during the tests ranged from 6.5 to 12
MPa, with respective toughness values in the range of 0.3-3.54; beam chord rotations were more than 10% in all
cases considered prior to the onset of strength deterioration. These preliminary findings suggest that a whole
new generation of Cement based materials in possible where tensile strength and strain capacities are so high that
they provide a totally different context for formulation of concrete mechanics and consequent reinforcement
detailing.

REFERENCES
[1] Yang, En-Hua, Li, V.C. (2010), Strain-hardening fiber cement optimization and component tailoring by
means of a micromechanical model J. Construction and Building Materials 24, pp. 130-139.
[2] Lepech, M.D., Li, V.C. (2008), Large-Scale Processing of Engineered Cementitious Composites ACI
Materials Journal V105, No. 4, pp. 358-366.
[3] Yang, En-Hua (2008), Design Added Functions in Engineered Cementitious Composites, PhD dissertation,
University of Michigan
[4] Manita, P., Pantazopoulou S.J. (2002) "Constitutive Model for the Stress-Strain Response of Fibre
Reinforced Concrete in Compression" Composite Materials in Concrete Construction: Proceedings of the
International Seminar Held at the University of Dundee, Scotland, UK on 5-6 September, 2002. Vol. 4.
Thomas Telford Services Limited
[5] Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Center-Point Loading),
ASTM Standard C293, 2010
[6] Standard Test Method for Flexural Performance of Fiber-Reinforced Concrete (Using Beam with ThirdPoint Loading), ASTM Standard C1609, 2010