Impact properties of geopolymer based extrudates incorporated

with y ash and PVA short ber

Zhang Yunsheng
a,
*
, Sun Wei
a
, Li Zongjin
b
, Zhou Xiangming
b
,
Eddie
b
, Chau Chungkong
b
a
School of Materials Science and Engineering, Southeast University, Nanjing 210096, PR China
b
Department of Civil Engineering, The Hong Kong University of Science and Technology, Clear water Bay, Kowloon, PR China
Received 6 July 2006; received in revised form 24 August 2006; accepted 25 August 2006
Available online 13 November 2006
Abstract
A PVA short ber reinforced y ash-geopolymer composites manufactured by extrusion technique (SFRGC) is developed in this
study. The eects of y ash content and ber volume fraction on the rheological and impact behaviors of SFRGC are systemically inves-
tigated. Freezethaw cycles and sulfuric acid attack tests are also employed to study the durability of SFRGC. The experimental results
show that for normally curing SFRGC, the addition of PVA ber increases greatly the ductility of SFRGC, especially in the case of high
volume fraction of ber, resulting in the change of impact failure mode from brittle pattern to ductile one. As a result, a great increase in
the impact toughness is seen in SFRGC with high ber content. Comparatively the addition of y ash exhibits obvious inuence on the
rhelogical and impact behaviors. The spherical shape of y ash can greatly improve the extrudability of fresh SFRGC pastes. SFRGC
without or with low percentage of y ash possesses very high impact strength and stiness. However, when too much y ash is incor-
porated, the impact resistance of SFRGC is reduced rapidly. For SFRGC undergone freezethaw cycles and 1 month of sulfuric acid
solution attack, only 5% or less of loss in impact strength is observed. The corresponding microstructure and failure mechanism are also
explored by Laser particle size analysis (LSA), X-ray diraction analysis (XRD), Scanning Electron Microscope (SEM), and Mercury
intrusion porosimetry (MIP) techniques.
2006 Elsevier Ltd. All rights reserved.
Keywords: Extrusion technique; Geopolymeric cement; Fly ash, PVA short ber; Rheological behavior; Impact properties; Durability; Microstructure
1. Introduction
Recent years have seen a great development in new type
of high performance inorganic cementitious materials
geopolymeric cement around the world. Geopolymeric
cement is one type of three dimensional CaO-free alumino-
silicate gel binders, which was rstly introduced into the
inorganic cementitious world by Davidovits in the later
1970s [1]. Geopolymeric cement can be synthesized by mix-
ing aluminosilicate reactive materials with less or no CaO
component (such as metakaolin, dehydrated clay) and
strongly alkaline solutions (such as NaOH or KOH), then
curing at room temperature. Under a strongly alkaline
solution, aluminosilicate reactive materials are rapidly dis-
solved into solution to form free SiO
4
and AlO
4
tetrahedral
units. With the development of reaction, mix water is grad-
ually split out and these SiO
4
and AlO
4
tetrahedral units
are linked alternatively to yield polymeric precursors
(SiO
4
AlO
4
, or SiO
4
AlO
4
SiO
4
, or SiO
4
AlO
4

SiO
4
SiO
4
) by sharing all oxygen atoms between two
tetrahedral units, and thereby forming monolithic like geo-
polymeric cement products [2].
Geopolymeric cement made with reasonable mix-design
and formulation can exhibit superior properties to Port-
land cement [36]: The production of geopolymeric cement
requires much lower calcining temperature (600800 C)
and emits 8090% less CO
2
than Portland cement.
0950-0618/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2006.08.006
*
Corresponding author.
E-mail address: zhangys279@163.com (Z. Yunsheng).
www.elsevier.com/locate/conbuildmat
Available online at www.sciencedirect.com
Construction and Building Materials 22 (2008) 370383
Construction
and Building
MATERIALS
Reasonable strength can be gained in a short period at room
temperature. In most cases, 70% of the nal compressive
strength is developed in the rst 12 h. Low permeability,
comparable to natural granite, is another property of geo-
polymeric cement. It is also reported that resistance to re
and acid attacks for geopolymeric cement are substantially
superior to those for Portland cement. Apart from the high
early strength, lowpermeability and good re and acid resis-
tance, geopolymeric cement also can attain higher uncon-
ned compressive strength and shrink much less than
Portland cement. Other documented properties include
good resistance to freezethaw cycles as well as excellent
solidication of heavy metal ions. These properties make
geopolymeric cement a strong candidate for substituting
Portland cement applied in the elds of civil, bridge, pave-
ment, hydraulic, underground and military engineering [7].
Extrusion is an important plastic forming process that
consists of pushing a viscous, dough-like plastic mixture
through a shaped die under high compressive and high
shear forces generated by extruder. The process is continu-
ous and simpler to use than other commonly used methods
(Casting, Slurry inltration, Spray suction, and Hatschek
process), making it most suitable for industrial mass pro-
duction. In addition, Extrusion technique enables intracta-
ble substances to be shaped into various products of
complicated cross-section with minimal water content
which is critical in developing high strength, low permeabil-
ity and excellent durability of products [825].
Fly ash is a by-product with pozzolanic reactivity from
coal-red power stations. Millions of tons of it are gener-
ated each year around the worlds. Only in china, the total
amount of y ash was accumulated up to 4 billion tons in
2002, which occupied about 40005000 hectares of areas,
and 200 million or more tons are still generated each year.
However, less than 30% of y ash is eectively used. In
order to enlarge the applications, the use of y ash as par-
tial replacement of metakaolin to synthesis geopolymeric
cement is systematically investigated in this study.
Fiber reinforcement has been employed in various
cementitious materials to improve exural, impact behav-
ior, toughness, and to shift failure mode. The use of short
ber is very preferable due to the simplicity and economical
nature in fabrication. Considering the quasi-brittle charac-
teristics of hardened geopolymeric cement paste, polyvinyl
alcohol (PVA) short ber is incorporated to improve the
ductility of hardened geopolymeric cement in this study.
In this paper, a PVA short ber reinforced y ash-geo-
polymer composites (SFRGC) manufactured by extrusion
technique is developed. The eects of y ash and ber on
the rheological and impact behaviors of SFRGC are sys-
temically investigated. In order to better understand the
impact behavior of the SFRGC with dierent y ash and
ber contents, laser particle size analysis (LSA), X-ray dif-
fraction analysis (XRD), Scanning Electron Microscope
(SEM), Mercury intrusion porosimetry (MIP) are employed
to explore the microstructure and failure mechanism. In
addition, the freezethaw, sulfuric acid attack tests are also
conducted to characterize the durability performance of
SFRGC.
2. Materials and methods
2.1. Materials
Metakaolin used in this study is obtained by calcining
pure kaolin at 700 C for 12 h. Grade I y ash, similar to
Class F y ash according to ASTM, is supplied by Qingz-
hou power plant, Hong Kong, PR China. Chemical grade
NaOH and sodium silicate solution with the molar ratio of
SiO
2
to Na
2
O of 3.2 are used as alkaline reagents. Two
types of silica sand (600300 lm and 15090 lm in diame-
ter) with the ratio of 0.6:0.4 and totally 32.5% by weight of
the binder system are used as aggregate. PVA ber is used
as reinforcement materials. The ber length is 6 mm with
an average diameter of 14 lm and a density of 1300 kg/
m
3
. The average tensile strength and elastic modulus of
the ber are 1500 MPa and 36 GPa. The chemical compo-
sitions and physical properties of raw materials are listed in
Table 1.
2.2. Mixture proportions and specimen preparation
2.2.1. Mixture proportions
The mixture proportions of various SFRGC used to
investigate the eect of y ash and ber on the rheological,
impact and durability behaviors are summarized in Table
2. Fb1 and Fb2 are used to study the eect of ber volume
fraction, while FA10 to FA50 are to investigate the eect of
y ash content. The mixture proportion of the pure
SFRGC without y ash and ber, i.e. Fb0, serves as a con-
trol sample and its mix proportion is given as follows:
molar SiO
2
molar Al
2
O
3
4:5;
molar Na
2
O
molar Al
2
O
3
0:8;
molar H
2
O
molar Na
2
O
6:4
The control sample is an optimum one, which is opti-
mized in the literature [26].
2.2.2. Specimen preparation
To uniformly disperse PVA bers, bers and metakao-
line powder are rst dry-mixed for 3 min in a Hobart
Table 1
Chemical compositions and physical properties of raw materials
Raw materials SiO
2
Al
2
O
3
Fe
2
O
3
CaO MgO TiO
2
MnO K
2
O P
2
O
5
SO
3
L.O.I Specic surface (m
2
/kg)
Metakaolin 62.97 26.91 2.62 0.60 1.24 0.01 0.18 0.74 4.44 350
Fly ash 60.70 24.72 6.90 0.70 1.13 1.50 2.35 400
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 371
planetary mixer at a low speed. Meanwhile, NaOH, sodium
silicate solution and water are mixed in a glass beaker and
cooled down to room temperature. The cooled alkaline
solution is then added into the metakaolin powders and
PVA ber mixture, and mixed for another 3 min at a high
speed. It should be noted that the premixing of NaOH and
sodium silicate solution, rather than the direct addition of
NaOH pellets into geopolymeric pastes, is the key measure
to ensure a long pot life to complete the extrusion process.
The nally mixed, dough-like fresh geopolymeric cement
paste is fed into a homemade ram rheometer to measure
the rheology or the pugmill chamber of a single-screw vac-
uum extruder (PVL100.3, KEMA) to produce SFRGC
thin sheet. Before making the nal thin sheet, the fresh
SFRGC pastes are further mixed, de-aired, and compacted
in the extruder, then the pastes are pushed through a thin
sheet die with a cross-section of 75 mm 6 mm. The fresh
extrudates are rst covered with a plastic sheet and, after
24 h moved to an isothermal environmental chamber and
cured at a standard condition of 20 C and 100% relative
humidity (R.H.) for 28 days. After curing, the hardened
SFRGC thin sheets are cut into 90 mm 75 mm 6 mm
thin plates and dried in normal laboratory conditions for
24 h before conducting test.
2.3. Methods
2.3.1. Rheology test
The fresh SFRGC pastes for extrusion exhibit higher
stiness and cohesion than the normal Portland cement
pastes, which makes common shear-driven rheometer inap-
propriate for measuring their rheology. In this study, a
homemade small-scale ram-type rheometer is used to sim-
ulate the extrusion process and measure the rheology of
various SFRGC pastes. The setup of the ram rheometer
is illustrated in Fig. 1. The key parameters of the ram rhe-
ometer are also listed in Table 3.
When performing the rheology test, the barrel of the
ram rheometer is rst lled with a certain amount of fresh
SFRGC pastes and the plunger is positioned to ensure the
plunger plate directly contact with the upper surface of the
paste in the barrel, then the piston is xed at the top grip of
a closed-loop servohydraulically controlled materials test
machine (Sintech 10/D MTS 810), subsequently applies
pressure to the SFRGC pastes in the barrel, forcing them
through a rigid die land. According to Benbows extrusion
rheological theory [27], at least two piston velocities and
two die land length-to-diameter ratios (L/D) are required
to characterize the rheological behavior of a paste. In this
study, ram extrusion testing will be conducted on Fb0,
Fb2 and FA30 with the piston velocity of 0.3 mm/s,
0.54 mm/s, 0.27 mm/s, 0.135 mm/s, 0.063 mm/s (the corre-
sponding extrusion velocities are 13.3 mm/s, 24.0 mm/s,
12.0 mm/s, 6.0 mm/s, 2.8 mm/s, respectively) and L/D of
0.83, 4.79, 8.75, 12.85, 16.67 to calibrate the extrudability
of the fresh SFRGC pastes. The large volume of the barrel
(80 in diameter 100 in length) makes it possible to drive
the above mentioned 5 piston velocities, starting with the
highest velocity, then decreasing velocity step to the smaller
velocity at the end. Each piston velocity will last for 60 s to
Table 2
Mixture proportions of various SFRGC
Batches Powder constituents of geopolymeric
binders
Sand volume fraction (%) Fiber volume fraction (%)
Metakaolin (%) Fly ash (%)
Fb0 100 0 32.5 0
Fb1 100 0 32.5 1
Fb2(or FA0) 100 0 32.5 2
FA10 90 10 32.5 2
FA30 70 30 32.5 2
FA50 50 50 32.5 2
Fig. 1. Set-up of a ram rheometer designed in this study.
Table 3
Ram rheometer dimensions
Barrel chamber
diameter,
D
0
(mm)
Circular die
land diameter,
D (mm)
Die length/die diameter, L/D
80.0 12.0 0.83 4.79 8.75 12.85 16.67
372 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
ensure that a stable state is observed and thus the accuracy
of data is assured. The length of each plateau is propor-
tional to its piston velocity. A data acquisition system is
connected to the ram rheometer, and the resulting piston
load-displacement is recorded. Typical load-displacement
curve obtained during ram test is displayed in Fig. 2. It
should be note that the rst plateau for load-displacement
curve is used to pre-pressure the SFRGC pastes at the
beginning of the ram test, which will not be suitable for
the calculation of rheological parameters. The extrusion
pressure of the SFRGC pastes are evaluated by dividing
the load required for ram test by the cross-sectional area
of the die land.
2.3.2. Impact test
The thin plate specimens with the dimensions of
75 mm 90 mm 6 mm are prepared for impact test on
an instrument impact tester (ITR-200, RADMANA).
When a test starts, the punch is raised to a specic height
and then pushed down by high-pressure gas. The average
velocity of the punch before and after impact is adjusted
to approximately 3 m/s. The peak resistant load, fracture
energy, and load gradient of the specimen can be measured
and recorded. These parameters provide useful indices for
impact strength, toughness, and stiness of the SFRGC
specimen.
2.3.3. Freezing and thawing tests
With respect to the freezethaw test, one freezethaw
cycle involves the alternate lowering the temperature of
specimen from 4.4 to 20 C, and raising temperature
from 20 to 4.4 C. One cycle lasts 9h including 4 h of
freezing, 2.5 h of thawing, a certain amount of time for
deceasing and raising temperature, and time for draining
and lling water in the chamber used for depositing speci-
mens. Fig. 3 shows the typical temperature change for one
cycle. All the steps are automatically controlled by com-
puter. For each batch, at least three samples are tested.
The thin plates will be subjected to 20 freezethaw cycles.
2.3.4. Sulfuric acid attack test
The resistance to the acid solution attack in this study is
tested by immersion of SFRGC specimens in a solution of
sulfuric acid of pH=1. The acid solution is refreshed every
one week to ensure a relatively constant pH value. The
impact test is conducted after 1 month of exposure. The
surface images are also taken to study the eect of acid
attack.
2.3.5. Microstructure analysis
Laser particle size analysis (LSA), X-ray diraction
analysis (XRD), Scanning Electron Microscope (SEM)
and Mercury intrusion porosimetry (MIP) are employed
to examine the microstructure characteristics and impact
failure mechanism of the SFRGC.
2.3.5.1. LSA. To measure the particle size of raw materials,
about 2 g of powder materials are added to a 100 ml beaker
fully lled with water. About 0.5 g of sodium diphosphate
(Na
4
P
2
O
7
), serving as dispersant, is added into the beaker
to increase the dispersability of the powder materials. After
a few minutes of vibration, the dispersed sample is poured
into the chamber of the LSA particle Analyzer (Coulter
LS2300) for testing. The testing duration is limited to
10 min to avoid the possible hydration of the sample parti-
cles. The data are recorded and analyzed automatically
with Zirconia model by computer.
2.3.5.2. XRD. For the XRD test, samples from dierent
mixtures are nely crushed and collected at 28 days, and
then these crushed fractions are immersed for about 3 days
in an acetone to stop the further hydration process and
removed free water, subsequently oven-dried at 50 C for
12 h to eliminate the absorbed water. After that, these geo-
polymeric fractions are further ground into very ne parti-
cles, and sieved to remove most of the quartz sand and
ber. Therefore the diraction lines representing quartz
tuned out to be vary quite a bit from sample to sample.
This is an artifact that does not inuence interpretation
of the results of the present study and therefore it will be
ignored.
-20 0 20 40 60 80
0
500
1000
1500
2000
2500
V
0
=13.3mm/s
V
1
=24.0mm/s
V
2
=12.0mm/s
V
3
=6.0mm/s
V
4
=2.8mm/s
L
o
a
d

(
N
)
Distance travelled by ram (mm)
Fig. 2. Typical load vs. stroke displacement curve.
0 1 2 3 4 5 6 7 8 9 10
-25
-20
-15
-10
-5
0
5
10
15
20
25
T
e
m
p
e
r
a
t
u
r
e

(

C
)
Time (hour)
Fig. 3. Typical temperature change for one freezethaw cycle.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 373
X-ray powder diraction is recorded on a Philips
PW1830 spectrometer using CuKa radiation with a scan-
ning rate of 1 per min from 10 to 80(2h). The wavelength
is selected as 15.40562 nm (Cu).
2.3.5.3. SEM. A JEOL-6300 SEM microscope is used to
examine the microstructure of the fractured composites at
the accelerating voltage of 15 kV. These samples that are
used to conduct SEM observation are rstly dried at
60 C until the constant weight is reached. Then, these sam-
ples are bonded on the sample holders with conducting
glue, followed by 30-min sputter coating of carbon. After
that, the morphologies of the nal products in samples
are observed at microscopic level using SEM.
2.3.5.4. MIP. An Autopore IV 9500 MIP is used for to
investigate the porosity and pore size distribution of dier-
ent SFRGC. The samples with size of approximate
10 mm 10 mm 10 mm are obtained by cutting up the
harden SFRGC pastes at an age of 28 days. These samples
are rst oven-dried for 24 h at 50 C to remove the physi-
cally absorbed water in capillary, and then placed into
the glass penetrometer with the volume of 5 cc. After that,
low pressure test is performed, followed by high pressure
test (the maximum mercury press reaches 30,000 psi). The
intrusion and evacuation of mercury is automatically con-
trolled by computer with Autopore IV 9500 program. The
cumulative intrusion and dierential pore volume vs. pore
diameter are also measured and recorded with an increase
of the pressure. The diameters of the pores intruded by the
mercury are calculated using Washburn equation, with a
mercury contact angle of 130.
3. Results and discussions
3.1. Rheological test program and results
Rheological behavior of fresh paste is very important to
successfully extrude thin sheet products. When the rheolog-
ical behavior of SFRGC paste is not properly designed,
defects such as lamination, surface fracture or edge tear
will form in the composites, thus reducing its nial mechan-
ical and durable properties. It is well known that y ash
can improve the uidity of Portland cement paste. How-
ever, it is not still understood that the eect of y ash on
the rheological behaviors of the new type of geopolymeric
cement paste. In this study, rheological behavior of the
fresh SFRGC pastes without and with y ash is measured
to characterize the eect of y ash. The eect of ber on the
rheological behavior is also investigated by varying the
ber content from 0% to 2%. The ram extrusion mecha-
nism can be well illustrated by the BenbowBridgwater
model based on the assumption of plastic deformation in
the die entry and plug ow in the die land [27]. According
to this model, the total extrusion pressure drop P
e
, through
circular die with a die entry angle of 90 can be simply pre-
sented by the following equation:
P
e
P
1
P
2
2r
0
aV lnD
0
=D 4s
0
bV L=D 1
where P
e
is the overall extrusion pressure of piston ap-
plied by MTS machine, while P
1
and P
2
are pressure
drops at the die entry and the die land, respectively; D
0
and D are the barrel and the die land diameters, respec-
tively; L is the die land length; V is the extrudate velocity
in the die land; r
0
is the initial bulk yield stress of fresh
SFRGC paste, i.e., the least stress to push SFRGC paste
from the barrel into the die land; a is a factor character-
izing the eect of velocity in the die entry, which is similar
to a viscosity coecient g; both r
0
and a are regarded as
the material constants independent of die geometries and
extrusion velocities and only associated with the move-
ment of the paste as it ows from the barrel into the entry
of the die land. s
0
is the initial wall shear yield stress of
the fresh SFRGC paste near the die land wall, i.e., the
least shear stress to overcome the wall shear stress exerted
by the wall of the die land; b is a factor characterizing the
eect of velocity on die land extrusion pressure drop;
Both s
0
and b are also regarded as the material constants
and only associated with the resistance to ow in the die
land.
The extrusion pressure vs. L/D curves of the three
SFRGC pastes (Fb0, Fb2 and FA30) at dierent extru-
sion velocities are measured and displayed in Fig. 4(ac).
The four rheological parameters (r
0
, a, s
0
, b) of each mix-
tures are then calculated according to Eq. (1) and listed in
Table 4. For comparison, bulk yield stress r = r
0
+ aV,
Wall shear stress s = s
0
+ bV at an extrusion velocity of
24 mm/s and extrusion pressure P
e
at 24 mm/s and L/
D = 16.67 are also calculated and given in Table 4.
Bashed on the calculated the values of r
0
, a, s
0
, b, the
extrusion pressure can be obtained by Eq. (1) and shown
in Fig. 58.
3.1.1. Eect of ber
It can be seen in Table 4 that the initial yield stressr
0
is
increased from 12.579 kPa to 67.094 kPa after 2% PVA
ber is incorporated into the fresh SFRGC paste, indicat-
ing that the addition of ber cause the starting ow of
SFRGC paste become more dicult. However, once the
SFRGC paste with ber moves, the bulk yield stress a
exhibits a rapid drop, as shown in Table 4 and Fig. 5. This
may be explained by the facts that the twisting and enlace-
ment eects of ber drives SFRGC pastes to smoothly ow
into the rigid die land.
A careful look at the wall shear stress s at dierent
extrusion velocities (Fig. 6) demonstrates that the addition
of ber considerably increases the friction between the
SFRGC paste and die land wall, leading to an increase of
ow resistance. It is especially true in the case of high extru-
sion velocity. For a larger extrusion velocity, a higher
extrusion pressure is required to push out the SFRGC
paste from die opening (Fig. 7), while for smaller L/D, a
relative small pressure is enough (Fig. 8).
374 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
3.1.2. Eect of y ash
When comparing Fb2 to FA30, it can be found that the
addition of y ash (FA30) signicantly reduces the initial
yield stress (r
0
, s
0
). The bulk yield stress a and wall shear
stress s are also much lower than those of the SFRGC
paste without y ash (Fb2), as shown in Table 4 and Figs.
5 and 6. The SFRGC paste with y ash can be easily
extruded out of die opening with a relative small pressure
(Figs. 7 and 8). This may be attributed to the spherical
shape of y ash and low pozzolanic reactivity.
0 3 6 9 12 15 18
0
300
600
900
1200
1500
1800 V=24.0mm/s
V=12.0mm/s
V=6.0mm/s
V=2.8mm/s
E
x
t
r
u
s
i
o
n

p
r
e
s
s
u
r
e

(
k
P
a
)
Die land lenth/ die diameter (L/D)
0 3 6 9 12 15 18
0
500
1000
1500
2000
2500
3000
V=24.0mm/s
V=12.0mm/s
V=6.0mm/s
V=2.8mm/s
E
x
t
r
u
s
i
o
n

p
r
e
s
s
u
r
e

(
k
P
a
)
Die land length/ die diameter (L/D)
0 3 6 9 12 15 18
0
100
200
300
400
500
600
E
x
t
r
u
s
i
o
n

p
r
e
s
s
u
r
e

(
k
P
a
)
Die land length/ die diameter (L/D)
V=24.0mm/s
V=12.0mm/s
V=6.0mm/s
V=2.8mm/s
a b
c
Fig. 4. Total extrusion pressure vs. length/diameter (L/D) at various extrudate velocities.
Table 4
Rheological parameters of various SFRFGC pastes
Batches r
0
(kPa) s
0
(kPa) a (kPa s mm
1
) b (kPa s mm
1
) r at v = 24 mm/s
(kPa)
s at v = 24 mm/s
(kPa)
Extrusion Pressure at
v = 24 mm/s, L/D = 16.67 (kPa)
Fb0 12.579 9.221 17.534 0.145 433.395 12.699 2491.022
Fb2 67.094 8.480 4.4717 0.978 174.415 31.971 2793.228
FA30 16.282 0.647 0.858 0.139 36.874 3.979 405.154
0 5 10 15 20 25
0
100
200
300
400
500
B
u
l
k
y
i
e
l
d

s
t
r
e
s
s

(
k
P
a
)

Extrudate velocity (mm/s)
Fb0
Fb2
FA30
Fig. 5. Bulk yield stress of various SFRFGC at dierent extrudate
velocities.
0 5 10 15 20 25
0
5
10
15
20
25
30
35
D
i
e

w
a
l
l

s
h
e
a
r

s
t
r
e
s
s

(
k
P
a
)
Extrudate velocity (mm/s)
Fb0
Fb2
FA30
Fig. 6. Wall shear stress of various SFRFGC at dierent extrudate
velocities.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 375
3.2. Impact test program and results
Fig. 9(a) and (b) shows the impact response curves of
various SFRGC with dierent y ash and ber contents.
To compare the eect of ber and y ash content on the
impact behavior, Fb0 without ber and y ash addition
is served as a control sample. Fb1 to Fb2 are used to inves-
tigate the eect of the volume fraction of PVA ber. FA10
to FA50 are designed to study the eect of y ash content.
The mixture proportions of various SFRGC are listed in
Table 1. The impact strength, stiness and toughness for
Fb0 to FA50 are also shown in Table 5.
As can be seen in Fig. 9(a), Fb0 has highest impact
strength (449.7 N), while the impact displacement is very
small, only around 0.84 mm. And after peak load, Fb0
specimen cannot sustain any loading, resulting in a sharp
drop of its impact curve. This indicates that SFRGC with-
out ber will be broken without any warning when the
peak load is reached. This may be very dangerous for prac-
tical construction. Fig. 10(a) displays the broken photo of
Fb0 specimen after impact test. Fb0 specimen is separated
into 5 small fractions during impact test. The fracture pat-
tern is typically of a brittle failure.
When 1% PVA ber is incorporated (Fb1), the impact
response curve takes on a new look. Unlike Fb0, the
impact curve of Fb1 does not directly fall down to 0 after
peak load, rather than only drops to about 300 N and sus-
tains the load for a long displacement. This indicates the
addition of PVA ber prevent the cracks from rapid prop-
agating, thus an increase in ductility. However, 1% ber is
so little that they can only bear a part of the whole impact
load. Fig. 10(b) shows the photo of Fb1 after impact test.
The geopolymer plate with 1% ber is broken into three
small pieces after subjecting to impact test. This suggests
that the geopolymer plate with 1% ber shows some char-
acteristics of ductile materials.
When 2% PVA ber is incorporated (Fb2), the impact
curve shows an obvious strain hardening response after
peak load. Fb2 sample can sustain the peak load
(429.6 N) up to around 2.5 mm of displacement. After that
the bearing capacity falls slowly up to 7.5 mm of displace-
ment. Fig. 10(c) illustrates the broken photo of Fb2 after
impact test. No crack is found on the surface of SFRGC
with 2% of PVA ber. Only a hole caused by the high speed
0 3 6 9 12 15 18
0
500
1000
1500
2000
2500
3000
T
o
t
a
l

e
x
t
r
u
s
i
o
n

p
r
e
s
s
u
r
e

P
e

(
k
P
a
)
Die land length/ die diameter (L/D)
Fb0
Fb2
FA30
Fig. 7. Extrusion pressure at various extrudate velocities for a xed
L/D = 16.67.
0 5 10 15 20 25
0
500
1000
1500
2000
2500
3000
T
o
t
a
l

e
x
t
r
u
s
i
o
n

p
r
e
s
s
u
r
e

P
e

(
k
P
a
)
Extrudate velocity (mm/s)
Fb0
Fb2
FA30
Fig. 8. Extrusion pressure at various L/D for a xed extrudate velocities
V = 24 mm/s.
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d
(
N
)
Displacement(mm)
Fb0
Fb1
Fb2
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d
(
N
)
Displacement(mm)
FA0
FA10
FA30
FA50
a b
Fig. 9. Impact curves of various SFRFGC: (a) with dierent ber contents and (b) with dierent y ash contents.
Table 5
Impact resistance of various SFRGC
Batches Impact
strength (N)
Impact stiness
(N/mm)
Impact
toughness (mJ)
Fb0 449.7 1150 209
Fb1 426.0 873 1177
Fb2 (or FA0) 429.6 749 1833
FA10 443.3 1007 2108
FA30 290.0 665 1587
FA50 268.8 456 1307
376 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
impact of punch is seen in the center of sample after the
impact test. This demonstrates that 2% ber can ensure
the integrality of SFRGC near the hole in the case of a
large deection caused by high speed impact of the punch.
By comparing the impact test results in Table 5, it can be
seen that the addition of ber show a decreasing trend for
the impact strength (peak load), however the reduction is
very little, only 5.3% for Fb1 and 4.47% for Fb2. In con-
trast, a great inuence of ber content is clearly seen on
impact stiness (gradient in linear section for impact
curves) and impact toughness (area between impact curves
and the x-axis), especially for the latter. For example, the
impact toughness of Fb1 and Fb2 is 4.63 and 7.77 times
higher than that of Fb0.
Fly ash is vastly produced as by-product from the ther-
mal power plants across the world. Due to good pozzolanic
reactivity, y ash is widely used to partially replacement of
Portland cement in the led of civil engineering. Consider-
ing that y ash can, on the one hand, improve the extrud-
ability of geopolymeirc pastes; on the other hand, contains
an amount of reactive Al
2
O
3
and SiO
2
, which are similar to
metakaolin, a part of metakaolin is replaced by y ash in
this study. Fig. 9(b) shows the impact response curves for
SFRFGC with dierent percentages of y ash at the same
ber content (FA10, FA30 and FA50). The impact test
results are also shown in Table 5. As can be seen in
Fig. 9(b) and Table 5, the impact resistance is a function
of the percentages of y ash. When 10% by weight of y
ash is added (FA10), not only are the impact strength
and stiness enhanced, but also is the toughness (absorbed
energy) increased from 1833 mJ to 2108 mJ. When too
much y ash are incorporated into geopolymeric mixture
(FA30 and FA50), the impact resistance of SFRFGC is
reduced rapidly, and the reducing percentage is increased
with an increase in the percentages of y ash. It is especially
true for 50% y ash addition. For instance, when compar-
ing with FA0 the impact strength, stiness and toughness
of FA50 are reduced by 37.4%, 39.1% and 28.7%,
respectively.
As mentioned above, 10% y ash addition enhances the
impact resistance of SFRFGC. The enhancing mechanism
of y ash is investigated by LSA, SEM, XRD and MIP
techniques. Fig. 11 shows the particle size distributions of
various types of raw materials. It can be seen in Fig. 11 that
coarse sand has the largest particle size amongst the four
raw materials, then ne sand. Metakaolin and y ash have
the smallest particle size, which ll in voids among sands,
resulting in a compacted system. Although the particle size
of y ash is very closed to that of metakaolin particle, the
spherical shape of y ash greatly improve the extrudability
of geopolymeric mixture, as a result a very highly quality
extrudates are produced. In addition, the particles of y
ash are also served as the nucleation sites of geopolymer-
ization reaction, which promotes the reaction rate and
the formation of geopolymeric products. The phenomenon
can also be conrmed by comparing the amorphous areas
of XRD spectra for pure geopolymeric paste without y
ash (Fb2) and geopolymeric paste with 10% y ash
(FA10), as shown in Fig. 12.
As can be seen from Fig. 12, both pure geopolymeric
paste and y ash based geopolymeric paste have a large dif-
fuse halo peak at about 2040 (2h
max
CuKa). This means
that geopolymers with or without y ash are mainly X-ray
amorphous materials consisting of randomly developed Si
Al polyhedra with a lack of periodically repeating SiAl
atomic ordering. Comparing the area under X-ray dirac-
togram at 2040, it is found that y ash based geopoly-
meric paste shows larger area than pure geopolymeric
paste without y ash, which indicates y ash based goe-
polymeric paste contains more amorphous products than
pure geopolymer. In addition, several sharp characteristic
peaks (7.09 A

, 4.23 A

, 3.33 A

, 1.81 A

, 1.54 A

,1.37 A

) are
also seen from Fig. 12. According to the XRD-pattern,
these peaks are identied as quartz and kaolin. With
respect to X-ray diractogram of metakaolin, kaolin and
quartzes are induced by metakaolin, and in the process of
geopolymerization, the kaolin and quartzes do not take
part in reaction.
Fig. 10. Fracture photos of various SFRFGC after impact test.
0.1 1 10 100 1000
Particle Diameter(m)
metakaoline
fly ash
fine sand
coarse sand
Fig. 11. Particle size distributions of metakaolin, y ash, ne and coarse
sand.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 377
Figs. 13(a) and (b) and 14(a) and (b) show the SEM
images for FA0 and FA10, respectively. Some small pores
and cracks can be easily found in FA0, while the micro-
structure of FA10 with 10% y ash is much denser. If
too much y ash is added into geopolymeric paste, the
microstructure of hardened geopolymeric paste (FA50) will
become relatively loose (Fig. 15(a) and (b)). This can be
attributed to the fact that no enough geopolymeric prod-
ucts are formed in the interspaces amongst sands due to
too less high reactive metakaolin.
The inuence of y ash content on porosity and pore
size distribution of the hardened SFRFGC can also obvi-
ously be seen in Figs. 16 and 17. It can be seen from
Fig. 16 that the addition of 10% y ash greatly reduced
the total porosity. This is deemed to be responsible for
the denser and compacted microstructure of geopolymeric
paste, which is consistent with the observation of SEM
images. In contrast, when 50% y ash is incorporated into
geopolymeric mixture, the hardened matrix shows a very
high the porosity (Fig. 16), thus a formation of some large
pore and cracks, which is also be conrmed by observing
Fig. 15.
Fig. 17 gives the cumulative intrusion and its distribu-
tion of various SFRFGC with dierent percentages of y
ash. By comparing with the mercury intrusion curves,
cumulative pore volume with pore diameter smaller than
495 nm accounts for 60% of the total porosity for Fb2 sam-
ple without y ash replacement, 90% for FA10, and 55%
for FA50. This indicates that the addition of 10% of y
ash obviously reduces the pore size and improves pore
structure in hardened SFRFGC. However, too much y
1. Geopolymer with 10% fly ash
2. Pure geopolymer without fly ash
3. Metakaolin
10 20 30 40 50 60 70 80
7.09 1.37 1.54
1.81
4.24
3.33
1
2
3
2 ()
Fig. 12. X-ray diractograms of Metakaolin, pure geopolymeric paste,
and y ash based geopolymeric paste.
Fig. 14. SEM images of geopolymeric paste with 10% y ash (FA30).
Fig. 13. SEM images of geopolymeric paste without y ash (FA0).
378 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
ash addition will increase the percentage of large size pore,
leading to the formation of poor structure. The results are
good agreement with above SEM observations.
3.3. Durability test program and results
3.3.1. Freezethaw cycles
The impact response of various SFRGC after subjecting
to 20 freezethaw cycles is shown in Fig. 18(ad). For com-
parison, the impact characteristic parameters (impact
strength, stiness and toughness) of various SFRFGC
before and after freezethaw cycles are calculated and
listed in Table 6.
It can be from Table 6 and Fig. 18 that no obvious
change in the impact strength is seen for the SFRFGC
without y ash after 20 freezethaw cycles, as compared
to the same composite before freezethaw cycles. However,
the SFRFGC with y ash addition show a quite dierent
manner than the composite without y ash (Fig. 18(a
d)). The impact resistance of the composites with y ash
is not reduced, but an increase is found. This is especially
true for the composite with high percentage of y ash
(FA50). The observation is not consistent with what to
expect: the impact behavior of hardened cementitious paste
will generally be reduced after freezethaw cycles. The rea-
son for this may be explained by the following facts:
(1) The SFRFGC is so dense and compacted that water
can not be penetrate into inside the plates, resulting in
almost no freezethaw deterioration. (2) Geopolyeric prod-
ucts will further formed during the long duration of freeze
thaw test (about 10 days).
Based on the test results and analysis, the SFRFGC pos-
sessES excellent resistance to freezethaw cycles. To char-
acterize the freezethaw resistant properties of SFRFGC,
more number of freezethaw cycles will be required.
3.3.2. Sulfuric acid solution attack
Unlike conventional Portland cement, geopolymeric
cement do not produce lime (CaO) during the hydration
process, thus theoretically being not dissolved in acidic
solutions. However, the practical facts still require to be
investigated by experiments. This present work is devoted
to study the impact behaviors of various SFRGC before
and after sulfuric acid solution (pH=1 H
2
SO
4
) attack for
1 month. The mechanism of H
2
SO
4
attack on SFRGC is
explored by using XPS and IR techniques.
Fig. 19(a) and (b) shows the photos of the SFRGC
without y ash and with 50% y ash before and after sul-
furic acid solution attack. As can be seen from the pho-
tos, the surface layer of the SFRGC becomes white and
relatively hard but brittle after H
2
SO
4
attack compared
Fig. 15. SEM images of geopolymeric paste with 50% y ash (FA50).
FA0 FA10 FA50
0
10
20
30
40
34.21%
20.72%
25.59%
P
o
r
o
s
i
t
y

(
%
)
Types of fly ash based geopolymer
Fig. 16. Porosity of dierent geopolymeric pastes.
10 100 1000 10000 100000
0
20
40
60
80
100
(495,45)
(495,40)
(495,10)
C
u
m
m
l
a
t
i
v
e

I
n
t
r
u
s
i
o
n

(
%
)
Pore Diameter (nm)
FA0
FA10
FA50
Fig. 17. Mercury intrusion curves of dierent geopolymeric pastes.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 379
to those without attack. However, the weight loss of the
same batch after attack is so little that it may be ignored.
The phenomenon is possibly caused by the facts that
some of Na
+
ions in the surface layer of SFRGC are
exchanged by acid proton from H
2
SO
4
solution due to
stronger electrophilic eect of H
+
than Na
+
, resulting in
only remaining the acid insoluble polymeric SiOAl
framework. The replacement of Na
+
by H
+
can be con-
rmed by an obvious chemical shifts in Si2p XPS and
Al2p XPS spectra for the surface layer of the SFRGC
after sulfuric acid attack, as shown in Fig. 20(ad). The
corroded layer can be considered as a partially dealkalized
SiAl residue, which is little white and relatively hard, and
has potential to inhibit further corrosion by acting as bar-
rier to the acid ions.
The impact response behavior of various SFRFGC is
shown in Fig. 21(ad) after H
2
SO
4
solution attack. For
comparison, the impact response curves of the correspond-
0 5 10 15 20 25 30
0
200
400
600
800
L
o
a
d

(
N
)
Displacement (mm)
Before F-T cycles
After F-T cycles
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d

(
N
)
Displacement (mm)
Before F-T cycles
After F-T cycles
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d

(
N
)
Displacement (mm)
Before F-T cycles
After F-T cycles
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d
(
N
)
Displacement (mm)
Before F-T cycles
After F-T cycles
a
c d
b
Fig. 18. Impact curves s of various SFRFGC after 20 freezethaw cycles.
Table 6
Impact resistance of SFRGC after 20 freezethaw cycles
Batches Impact strength (N) Impact stiness (N/mm) Impact toughness (mJ)
Before cycles After cycles Before cycles After cycles Before cycles After cycles
FA0 429.6 442.1 749 1198 1833 1822
FA10 443.3 578.3 1007 786 2108 2216
FA30 290.0 345.0 665 658 1587 1939
FA50 268.8 332.4 456 600 1307 1828
Fig. 19. Photos of the SFRFGC before and after sulfuric acid solution attack.
380 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
108 106 104 102 100 98
Before acid attack
After acid attack
Before acid attack
After acid attack
Binding energy (ev)
78 76 74 72 70
Binding energy (ev)
78 76 74 72 70
Before acid attack
After acid attack
Before acid attack
After acid attack
Binding energy (ev)
108 106 104 102 100 98
Binding energy (ev)
a b
c d
Fig. 20. XPS spectra of the SFRFGC before and after sulfuric acid solution attack.
0 5 10 15 20
0
100
200
300
400
500
600
L
o
a
d
(
N
)
Displacement(mm)
Before acid attack
After acid attack
0 5 10 15 20
0
100
200
300
400
500
L
o
a
d
(
N
)
Displacement(mm)
Before acid attack
After acid attack
0 5 10 15 20
0
100
200
300
400
L
o
a
d
(
N
)
Displacement(mm)
Before acid attack
After acid attack
0 5 10 15 20
0
50
100
150
200
250
300
350
L
o
a
d
(
N
)
Displacement(mm)
Before acid attack
After acid attack
a b
c d
Fig. 21. Impact curves s of various SFRFGC before and after sulfuric acid solution attack.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 381
ing SFRFGC before attack are also displayed in gure.
The impact characteristic parameters (impact strength,
stiness and toughness) are listed in Table 7.
It can be seen from Fig. 21(ad) and Table 7 that the
impact strength (4.56%) and stiness (11.75%) of FA0 that
undergoes H
2
SO
4
solution attack is slightly lower than the
corresponding SFRFGC that does not go through H
2
SO
4
solution attack. However, the impact toughness has a great
increase, around 35.30%. This improvement in toughness is
assigned to the fact that there exists a platform with long
displacement and high load resistance in impact response
curve of FA0 after acid attack, as shown in Fig. 21(a).
The eects of y ash on acid resistance of SFRFGC are
shown in Fig. 21(b) and (c). As can be seen from these g-
ures, the impact resistance of SFRFGC with various per-
centage of y ash (FA10, FA30 and FA50,) is not
reduced after acid attack, but a considerable improvement,
particularly for FA30 and FA50, which is quite dierent
from that of FA0.
When observing the broken photos, no obvious dier-
ences in impact fracture pattern are found for various
SFRFGC before and after acid attack.
In order to investigate the mechanism of acid attack on
SFRFGC, IR technique is used to identify the changes in
microstructure, geopolymeric products of SFRFGC before
and after acid attack, as shown in Fig. 22(ad).
Taking into account that the IR spectra of FA0 to
FA50 are conducted by the IR instrument and under
the same conditions, it is seen that there is almost no dif-
ference in IR pattern before and after attack for FA0
and FA10 (Fig. 22(a) and (b)). This indicates that acid
attack does not destroy the bond environment and struc-
ture of the two mixtures, thus no too much change in
impact behaviors. In contrast, a visible reduction at
around 800 cm
1
band can be observed for FA30 and
FA50 after acid attack, as shown in Fig. 22(c and d).
Peaks at 800 cm
1
band are assigned to the bending
vibration of 6-fold coordination AlOH. According to
Davidovits [1], the decrease might be attributed to the
fact that some 6-fold coordination Al are converted into
4-fold coordination Al that are one of the constructing
element of geopolymeric products, that is, acid attack
maybe promotes the formation of amorphous geopoly-
meric products.
4. Conclusion
(1) The study demonstrates that extrusion is an ecient
forming process to produce high performance short
ber reinforced y ash-geopolymer composites
(SFRGC).
Table 7
Impact resistance of SFRGC after acid solution attack
Batches Impact strength (N) Impact stiness (N/mm) Impact toughness (mJ)
Before attack After attack Before attack After attack Before attack After attack
FA0 429.6 410 749 661 1833 2480
FA10 443.3 505 1007 734 2108 2582
FA30 290.0 372 665 545 1587 2412
FA50 268.8 326 456 496 1307 2739
1200 1000 800 600 400
Before acid attack
After acid attack
Wavenumber(cm
-1
)
1200 1000 800 600 400
After acid attack
Before acid attack
Wavenumber(cm
-1
)
1200 1000 800 600 400
After acid attack
Before acid attack
Wavenumber(cm
-1
)
1200 1000 800 600 400
Before acid attack
After acid attack
Wavenumber(cm
-1
)
a
c
b
d
Fig. 22. IR spectra of various SFRFGC before and after sulfuric acid solution attack.
382 Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383
(2) The spherical shape of y ash can greatly improve the
extrudability of geopolymeric mixture, resulting in a
denser microstructure of the SFRGC with low per-
centage of y ash. However, when too much y ash
is incorporated, the improvement of the microstruc-
ture caused by y ash cannot compensate the
decrease of geopolymeric products due to the low
pozzolanic reactivity of y ash in case of high per-
centage of y ash. As a result, the microstructure of
SFRGC becomes relatively loose.
(3) The addition of high volume fraction PVA ber
changes the impact failure mode of SFRGC from a
brittle pattern to ductile one, resulting in a great
increase in impact toughness.
(4) SFRGC without and with low percentage of y ash
possess very high impact strength and stiness. How-
ever, when too much y ash is incorporated, the
impact resistance is reduced obviously.
(5) SFRGC has excellent resistance to freezethaw cycles
and strong acid attack. After 20 freezethaw cycles or
1 month of pH = 1 sulfuric acid solution attack, no
obvious decrease in impact strength and stiness is
observed for various SFRGC. In some case, some
increase is even observed.
Acknowledgements
Authors gratefully acknowledge the nancial support
from the Jiangsu Province Natural Science project of No.
BK2006555; the project of No. A1420060186; the Su Tong
Bridge project of No. STDQ-04HT20SY-010; Opening re-
search project from Shenzhen Laboratory of Civil Engi-
neering Durability, Shenzhen University; Opening and
owing research project funded by Nanjing Hydraulic Re-
search Institute (No. Yk90508) and the doctorial founda-
tion (No. 9212002242) from southeast university.
References
[1] Davidovits J. Geopolymers and geopolymeric new materials. J
Thermal Anal 1989;35(2):42941.
[2] Davidovits J, Davidovits M. Geopolymer: ultrahihg-temperature
tooling material for the manufacture of advanced composites. In:
Adsit R, Gordaninejad F, editors. 36th Annual SAMPE symposium
and exhibition, Covina, CA, USA, vol. 36 pt 2; 1991. p. 193949.
[3] Davidovits J. Geopolymer cement to minimize carbon-dioxde green-
house-warming. In: Moukwa M, Sarkar SL, Luke K, Grutzeck MW,
editors. Cement-based materials: present, future, and environmental
aspects. Ceramic transactions, vol. 37. Westerville, America: Amer-
ican Ceramic Society; 1993. p. 16582.
[4] Davidovits J. Properties of geopolymer cement. In: S

kvara F, editor.
Proceedings 1st international conference on Alkaline cements and
concretes. Kiev Ukraine: Scientic Research Institute on Binders and
Materials, Kiev state technical university; 1994. p. 13149.
[5] Lyon RE, Foden A, Balaguru PN, Davidovits M, Davidovits J.
Fire-resistant aluminosilicate composites. J Fire Mater 1997;21(2):
6773.
[6] Davidovits J. High alkali cements for 21st century concretes. In:
Mehta PK, editor. Concrete technology, past, present, and future.
Detroit, SP-144: American Concrete Institute; 1994. p. 38397.
[7] Davidovits J. Geopolymer chemistry and properties. In: Davidovits J,
Orlinsl J, editors. Proceedings of the rst european conference on soft
mineralogy, vol. 1. Compiegne, France: The Geopolymer Institute;
1988. p. 2548.
[8] Shao Y, Marikunte S, Shah SP. Extruded ber-reinforced composites.
Concr Int 1995;17(4):4852.
[9] Shao Y, Marikunte S, Shah SP. High performance ber-cement
composite by extrusion processing. In: Chong KP, editors. Proceed-
ings of the 4th materials engineering conference: materials for the new
millennium. vol. 2, Washington, DC; 1996. p. 25160.
[10] Li Zongjin, Mu Bin, Chui Stanley NC. Systematic study of properties
of extrudates with incorporated metakaolin or silica fume. ACI Mater
J 1996;96(5):5749.
[11] Li Zongjin, Mu Bin, Chui Stanley NC. Static and dynamic behavior
of extruded sheets with short bers. J Mater Civil Eng
2001;13(4):24854.
[12] Shao Yixin, Shah SP. Mechanical properties of PVA ber reinforced
cement composites fabricated by extrusion processing. ACI Mater J
1994;94(6):55564.
[13] Corina A, Marikunte S, Shah SP. Extruded ber reinforced cement
pressure pipe. Adv Cement Based Mater 1998;8(2):4755.
[14] Li Zongjin, Mu Bin. Application of extrusion for manufacture of
short ber reinforced cementitious composite. J Mater Civil Eng,
ASCE 1998;10:24.
[15] Stang H, Pederson C. HPFRCC-extruded pipes. In: Chong KP,
editor. Proceedings of the 4th materials engineering conference:
materials for the new millennium. vol. 2, Washington, DC; 1996. p.
26170.
[16] Stevens MJ. Extruder principles and operation. London: Elsevier
Applied Science; 1985.
[17] Stang H, Li VC. Extrusion of ECC-material. In: Reinhardt HW,
Naaman AE, editors. Proceedings of the 3th international Workshop
on HPFRCC3. 1999. p. 20312.
[18] Shao Yinxin, Qin Jun, Shah SP. Microstructure of extruded cement-
bonded berboard. Cement Concr Res 2001;31(8):115361.
[19] Li Zongjin, Mu Bin. Rheological properties of cement-based extru-
dates with bers. J Am Ceram Soc 2001;84(10):234350.
[20] Shah SP, Peled A, DeFord D. Extrusion technology for the
production of ber-cement composites. In: Moslemi AA, editor.
Proceeding of inorganic-bonded wood and ber composite materials,
vol. 6. Idaho, USA: Forest Products Society; 1998. p. 26177.
[21] Peled A, Shah SP. Processing eects in cementitious composites:
extrusion and casting. J Mater Civil Eng 2003;15(2):1929.
[22] Akkaya Y, Peled A, Shah SP. Parameters related to ber length
and processing in cement composites. Mater Struct 2000;33(232):
51524.
[23] Igarashi S, Bentur A, Mindess S. The eect of processing on the bond
and interfaces in steel ber reinforced cement composites. Cement
Concr Comp 1996;18(2):31322.
[24] Peled A, Akkaya Y, Shah SP. Eect of ber length in extruded and
cast cement composites. In: Peled A, Shah SP, Banthia N, editors.
ACI SP-190 on high performance concrete ber reinforced thin
products. Detroit: American Concrete Institute; 2000. p. 120.
[25] Peled A, Cyr MF, Shah SP. High content of y ash (class F)
in extruded cementitious composites. ACI Mater J 2000;97(5):
50917.
[26] Yunsheng Z. Research on structure formation mechanism and
properties of high performance geopolymer concrete. Ph.D. Thesis,
School of Materials science and Engineering, Southeast University,
Nanjing, PR China; 2004.
[27] Benbow J, Bridgwater J. Paste ow and extrusion. Oxford, UK:
Clarendon; 1993. p. 153.
Z. Yunsheng et al. / Construction and Building Materials 22 (2008) 370383 383