Professional Documents
Culture Documents
DOI 10.1617/s11527-014-0322-7
ORIGINAL ARTICLE
Received: 10 February 2014 / Accepted: 23 April 2014 / Published online: 23 January 2015
Ó RILEM 2015
Abstract The effect of strain rate on the compres- loading. It is found that for the quasi-static loading and
sive behaviours of geopolymer concrete and mortar is low strain rate loading, cracks propagate along
reported. Split Hopkinson pressure bar was adopted interface transition zone (ITZ) and matrix of geopoly-
for the high strain rate testings. The dynamic increase mer concrete specimens whereas cracks occur at both
factors for compressive strength (DIFfc ) and critical the aggregates and ITZ under high strain rate loading.
strain (DIFec ) were measured and compared with
Concrete Comite Euro-international du Beton (CEB) Keywords Geopolymer concrete Strain rate effect
recommendations. The results show that alkaline SHPB Compressive strength
activators have significant influence on the quasi-
static compressive strength of geopolymer concrete.
With high strain rate loading, the DIFfc of geopolymer
concrete and mortar mixes increase with respect to 1 Introduction
increasing strain rates and in agreement with CEB
recommendations. In addition, the coarse aggregates Geopolymers, as ceramic materials, are produced by
in geopolymer concrete mixes play important role in alkali activation of aluminosilicate raw materials (fly
the increase of compressive strength. However, CEB ash or metakaolin), which are transformed into a
recommendations underestimate the DIFec of critical reaction product by polymerization in a high pH
strain for geopolymer concrete in the high strain rate environment and hydrothermal conditions at relatively
low temperatures (up to 120 °C) [1, 2]. As fly ash is an
by-product from power stations, the use of fly ash to
K. N. Feng Z. Pan F. Collins Y. Bai
make concrete can significantly reduce greenhouse
W. H. Duan (&)
Department of Civil Engineering, Monash University, emissions due to the production of cement [3].
Clayton, VIC 3800, Australia Therefore, geopolymer concrete has received increas-
e-mail: wenhui.duan@monash.edu ing interest in research community.
There are many experimental work conducted on
D. Ruan
Faculty of Engineering & Industrial Sciences, Swinburne quasi-static mechanical properties of geopolymer
University of Technology, Hawthorn, VIC, Australia based materials, including Young’s modulus, Pois-
son’s ratio, compressive strength, splitting tensile
C. M. Wang
strength, and fracture energy [4–9]. It is found that
Engineering Science Programme and Department of Civil
and Environmental Engineering, National University of these mechanical properties could be affected by a
Singapore, Kent Ridge, Singapore 119260, Singapore number of mixing parameters [4]. Specifically, the
672 Materials and Structures (2015) 48:671–681
compressive strength of geopolymer concrete can be effect of low strain rates, i.e. 10-5, 5 9 10-5, 10-4
increased by using a high ratio of sodium silicate-to- and 5 9 10-4 s-1, on Young’s modulus, Poisson’s
sodium hydroxide liquid, a high curing temperature in ratio and compressive strength of geopolymer mortar.
the range of 30–90 °C, or a long curing time. It is found that these properties of geopolymer mortar
Moreover, it is found that the constitutive relation increase with respect to the strain rate. Li and Xu [18]
for conventional ordinary Portland cement (OPC) explored the effect of high strain rate on the basalt
concrete can be applied to geopolymer concrete with fibre reinforce geopolymer concrete. The dependency
minor modifications [10]. It is worth mentioning that on the strain rate increases approximately linearly
fly ash-based geopolymer concrete is suitable for from the strain rate of 30–100 s-1.
structural applications and the current standards (the Compared with the limited research on the strain
Australian Standard AS3600 and the American Con- rate effect of geopolymer concrete, OPC concrete has
crete Institute Building Code ACI318-02) can be used been extensively studied. The strain rate effect can be
for the design of geopolymer concrete structures [7]. assessed by the dynamic increase factor (DIF, ratio of
In comparison with OPC-based materials, geopoly- dynamic strength/strain to static strength/strain). Con-
meric materials generally exhibit superior performance crete Comite Euro-international du Beton (CEB)
when they are exposed to extreme and aggressive recommendations are commonly used to predict the
conditions. The geopolymer exposed to the action of DIF of the compressive strength of OPC concrete
high temperatures (up to 800 °C) increased strength under the high strain rate loading [19]. The empirical
(due to a further geopolymerization) whilst the OPC equations of DIFfc . (for compressive strength) and
paste fully deteriorated due to the rehydration of the DIFec (for critical compressive strain) by CEB recom-
dissociated Ca(OH)2 following cooling [11, 12]. At mendations are given by
elevated temperatures, geopolymer concretes have less
potential to suffer from spalling in comparison with DIFfc ¼ rd =rs ¼ ðe_d =e_s Þ1:026as e_d 30 s1
ð1Þ
OPC concrete [13]. The durability of geopolymeric ¼ rd =rs ¼ cðe_d =e_s Þ1=3 e_d [ 30 s1
materials was investigated by Garcı́a-Lodeiro et al. [14],
who found that the potential for, and severity of, alkali- DIFec ¼ ecd =ecs ¼ ðe_d =e_s Þ0:02 ð2Þ
aggregate reaction (ASR) in geopolymeric materials is
lower than in OPC equivalents. The reason for this is where e_d is the strain rate (s ); e_s = 30 9 10 s-1;
-1 -6
attributed that low calcium fly ash based geopolymer as ¼ ð5 þ 9fcm =fcm0 Þ1 ; logc ¼ 6:156as 2; fcm is the
matrix is lack of readily available calcium which acts as concrete compressive strength; fcm0 is the reference
an instigator for ASR. strength of 10 MPa.
While most papers are focused on the quasi-static The current paper reports an experimental study on
mechanical properties, only a few studies have been the effect of strain rate on the compressive behaviours
conducted on the dynamic properties of geopolymer of geopolymer concrete and mortar. For comparison
concrete. Pan et al. [15] tested the damping of purposes, OPC concrete was tested as well. The
geopolymer pastes by using a cantilever beam. It is geopolymer concrete was prepared by using three
shown that overall damping capacities of geopolymer alkaline activators, i.e. NaOH/Na2SiO3, KOH/K2SiO3
pastes have comparable performance with OPC and KOH/Na2SiO3. A Shimadzu AG–X 300 kN
pastes. The damping mechanisms in geopolymer testing machine was adopted to measure the compres-
concrete are the combination of (a) moisture move- sive behaviours at quasi-static strain rate (10-6 s-1)
ment in the geopolymer matrix (b) sliding friction and low strain rates (10-5 to 10-4 s-1), while split
within geopolymer gels, and (c) movement between Hopkinson pressure bar (SHPB) techniques were used
fly ash particles and the matrix. In addition, strain rate at high strain rates (from 20 to 370 s-1). DIFfc and
effect on materials, induced by high-amplitude short- DIFec of geopolymer mixes were calculated and
duration loads, was identified as an important dynamic compared with the values predicted by CEB recom-
property in the design and analysis of structures to mendations. The mechanisms for the increase in the
resist dynamic loads from conventional weapons compressive strength of geopolymer concrete speci-
explosions, accidental explosions and high-speed mens were discussed with regards to the different
impact [16]. Khandelwal et al. [17] investigated the strain rates.
Materials and Structures (2015) 48:671–681 673
carried out for 4 min. The mixture was poured into condition was assumed as 10-6 s-1 in this investigation.
moulds (depending on the test for which they were In addition, two strain rates, i.e. 10-5 and 10-4 s-1 were
intended) in three equal layers. Each layer was selected to test the effect of low strain rate. A Shimadzu
vibrated for 15–30 s on a vibration table. AG–X 300 kN testing machine was used for the quasi-
static loading and low strain rate loading. Three speci-
2.2.3 Curing conditions mens were tested under each strain rate loading.
The OPC concrete specimens were cured under poly- 2.4 Setup of SHPB compressive tests
ethylene sheets for 24 h in a laboratory environment.
They were then removed from the moulds and trans- The experimental setup of SHPB consists of two long
ferred to a tank of saturated limewater at 23 ± 2 °C as slender steel bars with a short specimen sandwiched in
the moist-curing regime to satisfy ASTM C192 between (Fig. 1). When strike bar impacts the incident
requirements [20]. Specimens were cured for 28 days. bar, a stress wave was generated and travelled through
The geopolymer specimens in moulds were sealed the bar towards the specimen. When the wave reached
using plastic sheets to prevent moisture loss during the specimen, part of the stress wave reflected back
curing. The covered moulds were placed in a pre- down the incident bar and the remainder transmitted
heated oven at 60 °C for 24 h. The specimens were through the specimen into the transmitter bar. Remain-
then demoulded. ing energy in the transmitter bar was absorbed by the
After curing, both OPC and geopolymer specimens momentum bar.
were stored in a controlled environment kept at As shown in Fig. 1a, b, two SHPB testing systems
relative humidity 50 ± 3 % and temperature were adopted for mortar and concrete specimens,
23 ± 2 °C. This environment meets the International respectively. SHPB tests on mortar specimens
Organization Standardization requirements as a stan- (12.5 mm diameter 9 12.5 mm height) were tested by
dard atmosphere for conditioning and testing of a 14.5 mm (diameter of incident bar and transmitted bar)
materials known to be sensitive to variations in SHPB testing system. Because the ratio of maximum
temperature or relative humidity. The specimens were aggregate size to specimen diameter should be smaller
subjected to various tests after 28 days. In order to than 0.33 [24], a 50 mm SHPB testing system was
obtain smooth and parallel ends, the specimens were adopted for the concrete specimens (45 mm diameter and
polished with a rock-grinding machine before testing. 45 mm height). Specimen arrangement in the compres-
sion test is shown in Fig. 1c. Both end surfaces of
2.3 Quasi-static and low strain rate compressive specimens were carefully grounded in order to assure the
tests parallelism of the end surfaces. In order to minimise the
contact surface friction, grease was applied evenly on the
Density was calculated using the mass and dimensions interfaces between specimens and bars [25]. Strain rate
of cylinders (45 mm diameter and 90 mm height for loading of 50–370 and 300–1,500 s-1 were generated for
concrete specimens and 22.5 mm diameter and 45 mm concrete and mortar specimens, respectively.
height for mortar specimens) in accordance with
ASTM C1688 [21]. For each specimen, two cylinders 2.5 Data processing of SHPB tests
were measured for obtaining the average density of
each mix, and a third specimen was introduced if the Two sets of data were acquired at 1 ls interval,
first two densities differed by more than 5 %. corresponding to a sampling rate of 106 Hz, by oscillo-
The quasi-static and low strain rate tests were scope connected with strain gauges (SG1 and SG2 in
conducted in accordance with ASTM C39-94 [22], Fig. 1a, b) on the incident bar and transmitter bar,
which recommends a quasi-static loading rate of respectively. Stress correction factor and strain correction
0.14 MPa to 0.34 MPa per second. For typical concrete factor were calibrated to eliminate the effect of minor
(the modulus of elasticity is between 20 and 40 GPa), test misalignment of the strain gauges and the effect of
strain rates should be between 3 9 10-6 and adhesives used to bond the strain gauges to the bar
17 9 10-6 s-1 when considering the linear part of the surface. In addition, dispersion corrections were con-
stress–strain curve [23]. Therefore, a quasi-static loading ducted for the test data of 14.5 mm SHPB and 50 mm
Materials and Structures (2015) 48:671–681 675
(b)
(c)
SHPB systems. It was shown that there was no significant eT(t) are the reflected tensile wave and the transmitted
dispersion effect of the stress waves within bars [26]. compressive wave, respectively.
1D wave propagation is the key assumption in the One typical raw data set of geopolymer concrete
data processing, which assumes that the stress of a specimen (NaGC with strain rate of 150 s-1) is shown
long bar (homogeneous, isotropic and uniform in cross in Fig. 2a. It consists of incident wave, reflected wave
section over the entire length) is in a linear elastic and transmitted wave. The stress history and the strain
state. For the compressive test, strain rate, engineering history of specimens can be calculated by using Eqs.
stress and engineering strain of the specimens for (4–5). The combination of the stress history and strain
1-wave analysis are given by history results in the stress–strain relationship as
shown in Fig. 2b. DIFfc can be calculated by the ratio
2c0B
_ ¼
eðtÞ eR ðtÞ ð3Þ of dynamic strength to quasi-static strength while
Hs0
DIFec can be calculated by the ratio of critical dynamic
strain to quasi-static strain. Critical compressive strain
AB EB
rS ðtÞ ¼ eT ðtÞ: ð4Þ is defined as the stain where the stress reaches the
AS0
peak. Its corresponding strain rate was calculated by
Zt the average strain rate, which is the strain over the time
2c0B duration of the reflected wave.
eS ðt Þ ¼ eR ðt Þ ð5Þ
Hs0 Derivation of Eq. (4) from 1D stress wave propaga-
0
tion theory requires the force equilibrium condition, i.e.
where c0B is the elastic wave speed of the bar, Hs0 is F1 = F2. F1 and F2 are forces at the two ends of the
the length of the specimen; AB and AS0 are the cross specimen and can be calculated according to 1-wave
section area of the bar and the specimen, respectively; analysis and 2-wave analysis respectively [26]. As an
EB is the Young’s modulus of the bar; eR(t) and example, the force equilibrium condition for
676 Materials and Structures (2015) 48:671–681
(b)
(a) (b)
(c) (d)
(e)
Fig. 5 Strain-rate dependence of the compressive strength: a OPC; b NaGC; c KGC; d Na/KGC; e NaGM
inertia and lateral confinement effect. This statement 20 s-1. In addition, NaGM has a lower DIFfc when
can be further confirmed by the comparison between compared with CEB recommendations. These obser-
the experimental measurements and CEB recommen- vations indicate that the crack propagation through the
dations as shown in Fig. 5b–d. The comparison shows aggregates played important role in the increases of
that the DIFfc for geopolymer concrete are fitted well compressive strength in the high strain rate range.
with CEB recommendations. This tendency is inde-
pendent of the type of activator. 3.4 Effect of strain rate on critical strain
In Fig. 5e, the DIFfc of NaGM shows that high
strain rate sensitivity starts from 300 s-1, by contrast, The critical strain, which is considered as a key
high strain rate sensitivity for NaGC starts from parameter in the characterization of material
Materials and Structures (2015) 48:671–681 679
Fig. 6 Effect of strain rate on critical strain of concrete a critical strain determined by SHPB tests and b DIFec and CEB
recommendations
behaviour subjected to high strain rate loading, is contrast, the specimen was pulverized into small
presented in Fig. 6a. It can be seen that the critical pieces with the size of about 10 mm when subjected to
strains of both OPC concrete specimens and geopoly- high strain rate loading as shown within the red circle
mer concrete specimens are in a similar range and rise of Fig. 7c. It is clearly shown that the crack goes
with increasing strain rate, which is in agreement with through both the aggregates and the matrix.
previous work [34, 35]. The increase in the critical As mentioned above, no significant difference
strain could be attributed to the significant amount of could be observed from the failure modes of OPC
cracks formed in the concrete under high strain rate and geopolymer concrete mixes subjected to different
loading [34]. loading rates. Fracture process of OPC concrete mixes
The effect of strain rate on the DIFec for the could be applied to that of geopolymer concrete mixes.
concrete and mortar is shown in Fig. 6b. When Most cracks would propagate in ITZ and geopolymer
compared with CEB recommendations, it is shown matrix under quasi-static loading and low strain rate
that CEB recommendations underestimate the DIFec loading due to their relatively low strength when
for both OPC concrete and geopolymer concrete, compared with the aggregates. The final failure
especially in the range of high strain rate. Similar happens due to the propagation and perforation of
conclusions for OPC concrete were obtained by Wang one or more main cracks. However, at high strain rate,
et al. [34]. lots of tiny cracks occur at the aggregates and ITZ
simultaneously. Therefore, more energy need be
3.5 Effect of strain rate on failure pattern consumed in the process of the generation and
propagation of cracks [36]. Consequently, higher
All concrete mixes exhibited similar failure patterns DIFfc of the specimens are expected, which is in
subjected to different strain rate loadings. As an agreement with the observation in Fig. 7.
example, Fig. 7 shows the failure modes of NaGC
subjected to quasi-static loading (10-6 s-1), lower
strain rate loading (10-4 s-1) and high strain rate 4 Conclusions
loading (320 s-1).
In Fig. 7a, b, the specimens are failed as fragmen- Mechanical behaviors of geopolymer concrete and
tation after losing the strength. However, they still mortar as well as OPC concrete were investigated
keep their integrity. A further observation of fracture experimentally using both Shimadzu AG–X 300 kN
surfaces shows that most of cracks are propagated testing machine and the SHPB technique, in the strain
along ITZ between the matrix and aggregates for both rates range from 10-6 to 370 s-1. Five concrete and
quasi-static loading and lower strain rate loading. In mortar mixes were studied including one OPC
680 Materials and Structures (2015) 48:671–681
Failure modes
Cleavage of aggregate
10mm 10mm 3mm
Fig. 7 Failure modes and fracture surfaces of NaGC mixes subjected to: a quasi-static loading, 10-6 s-1; b lower strain rate loading,
10-4 s-1; and c high strain rate loading, 320 s-1
concrete mix (OPC), three geopolymer concrete mixes compressive strength of geopolymer concrete
(NaGC, KGC, and Na/KGC) and one geopolymer under high strain loading.
mortar (NaGM). The following conclusions can be
obtained: Acknowledgments The authors gratefully acknowledge the
financial support from the Australia Research Council in
(a) Geopolymer concrete and OPC concrete have conducting this study.
similar densities whereas geopolymer mortar
has a lower density due to the absence of coarse
aggregates. Alkaline activators have significant References
influence on the quasi-static compressive
strength of geopolymer concrete, which may 1. Davidovits J (1991) Geopolymers: inorganic polymeric new
be attributed to the modification of microstruc- materials. J Therm Anal 37(8):1633–1656
2. Rahier H, Mele B, Biesemans M, Wastiels J, Wu X (1996)
tures of geopolymer concrete. Low-temperature synthesized aluminosilicate glasses.
(b) Under the low strain rate loading (from 10-5 to J Mater Sci 31(1):71–79
10-4 s-1), the DIFfc of geopolymer concrete 3. Duxson P, Fernandez Jimenez A, Provis JL, Lukey GC,
Palomo A (2007) Geopolymer technology: the current state
and mortar mixes a slightly higher than 1.0 with
of the art. J Mater Sci 42(9):2917–2933
the increasing strain rates, which may attribute 4. Hardjito D, Wallah SE, Sumajouw DMJ, Rangan BV (2004)
to the viscous effect of free water in the On the development of fly ash-based geopolymer concrete.
geopolymer matrix. ACI Mater J 101(6):467–472
5. Sarker PK (2009) Analysis of geopolymer concrete col-
(c) With high strain rate loading (from 20 to
umns. Mater Struct 42(6):715–724
370 s-1), the DIFfc of geopolymer concrete 6. Hardjito D, Wallah SE, Sumajouw DMJ, Rangan BV (2005)
mixes increase with the increasing strain rates. Fly ash-based geopolymer concrete. Aus J Struct Eng
In addition, CEB recommendations underesti- 6(1):77–86
7. Sumajouw DMJ, Hardjito D, Wallah SE, Rangan BV (2007)
mate the DIFec of geopolymer concrete.
Fly ash-based geopolymer concrete: study of slender rein-
(d) For the quasi-static loading and low strain rate forced columns. J Mater Sci 42(9):3124–3130
loading, cracks propagate along ITZ and matrix 8. Rangan BV (2009) Chapter 11 in Geopolymers: structures,
of geopolymer concrete specimens whereas processing, properties, and applications. In: Provis J, De-
venter JV (eds) Engineering properties of geopolymer
cracks occur at both the aggregates and ITZ
concrete. Woodhead Publishing Limited, London
under high strain rate loading. This is believed 9. Pan Z, Sanjayan JG, Rangan BV (2011) Fracture properties
to be one of the mechanisms for the increase in of geopolymer paste and concrete. Mag Concr Res 63(10):9
Materials and Structures (2015) 48:671–681 681
10. Sarker PK (2009) Analysis of geopolymer concrete col- 25. Lu YB, Li QM (2011) About the dynamic uniaxial tensile
umns. Mater Struct 42(6):715–724 strength of concrete-like materials. Int J Impact Eng 38(4):
11. Pan Z, Sanjayan JG, Collins F (2014) Effect of transient 171–180
creep on compressive strength of geopolymer concrete for 26. Gama BA, Lopatnikov SL, Gillespie JW (2004) Hopkinson
elevated temperature exposure. Cem Concr Res 56:182–189 bar experimental technique: a critical review. Appl Mech
12. Pan Z, Sanjayan JG (2012) Factors influencing softening Rev 57(4):223–250
temperature and hot-strength of geopolymers. Cem Concr 27. Duxson P, Mallicoat SW, Lukey GC, Kriven WM, Van
Compos 34(2):261–264 Deventer JSJ (2007) The effect of alkali and Si/Al ratio on
13. Pan Z, Sanjayan JG, Kong DLY (2012) Effect of aggregate the development of mechanical properties of metakaolin-
size on spalling of geopolymer and Portland cement con- based geopolymers. Colloids Surf A 292(1):8–20
cretes subjected to elevated temperatures. Constr Build 28. Fernandez-Jimenez AM, Palomo A, Lopez-Hombrados C
Mater 36:365–372 (2006) Engineering properties of alkali-activated fly ash
14. Garcı́a-Lodeiro I, Palomo A, Fernández-Jiménez A (2007) concrete. ACI Mater J 103(2):106–112
Alkali–aggregate reaction in activated fly ash systems. Cem 29. Hardjito D, Rangan BV (2005) Development and properties
Concr Res 37(2):175–183 of low-calcium fly ash-based geopolymer concrete. Curtin
15. Pan Z, Feng KN, Gong K, Zou B, Korayem AH, Sanjayan J University of Technology, Perth
et al (2013) Damping and microstructure of fly ash-based 30. Ranjith PG, Jasinge D, Song JY, Choi SK (2008) A study of
geopolymers. J Mater Sci 48(8):3128–3137 the effect of displacement rate and moisture content on the
16. Ross CA, Tedesco JW, Kuennen ST (1995) Effects of mechanical properties of concrete: use of acoustic emission.
strain-rate on concrete strength. ACI Mater J 92(1):37–47 Mech Mater 40(6):453–469
17. Khandelwal M, Ranjith PG, Pan Z, Sanjayan JG (2011) 31. Rossi P (1997) Strain rate effects in concrete structures: the
Effect of strain rate on strength properties of low-calcium LCPC experience. Mater Struct 30(1):54–62
fly-ash-based geopolymer mortar under dry condition. Arab 32. Rossi P, Toutlemonde E (1996) Effect of loading rate on the
J Geosci 6(7):2383–2389 tensile behaviour of concrete: description of the physical
18. Li W, Xu J (2009) Mechanical properties of basalt fiber mechanisms. Mater Struct 29(186):116–118
reinforced geopolymeric concrete under impact loading. 33. Li QM, Meng H (2003) About the dynamic strength
Mater Sci Eng A 505(1–2):178–186 enhancement of concrete-like materials in a split Hopkinson
19. CEB Comite Euro-International Du Beton (1993) CEB-FIP pressure bar test. Int J Solids Struct 40(2):343–360
model code 1990: design code. Telford 34. Wang SS, Zhang MH, Quek ST (2012) Mechanical
20. ASTM (2007) C192. Standard practice for making and behavior of fiber-reinforced high-strength concrete sub-
curing concrete test specimens in the laboratory jected to high strain-rate compressive loading. Constr Build
21. ASTM (2011) C1688. Standard test method for density and Mater 31:1–11
void content of freshly mixed pervious concrete 35. Bischoff PH, Perry SH (1985) Compressive strain rate
22. ASTM (2007) C39-94. Standard specification for ready- effects of concrete. MRS Online Proceedings Library 64:
mixed concrete 425–450
23. Malvar LJ, Crawford JE (1998) Dynamic increase factors 36. Ren ZG, Chen M, Lu ZA, Xu WG (2012) Dynamic
for concrete. DTIC Document mechanical property of hybrid fiber reinforced concrete
24. ASTM (2007) C31/C31M. Standard practice for making (HFRC). J Wuhan Univ Technol Mater Sci Ed 27(4):783–
and curing concrete test specimens in the field 788