geopolymer and portland cement concretes in simulated Þre · 2016-12-05 · geopolymer concrete...

Magazine of Concrete Research, 2011, 63(3), 163–173

doi: 10.1680/macr.9.00110

Paper 900110

Received 25/06/2009; last revised 08/03/2010; accepted 06/04/2010

Published online ahead of print 10/01/2011

Thomas Telford Ltd & 2011

Magazine of Concrete ResearchVolume 63 Issue 3

Geopolymer and Portland cement concretesin simulated fireZhao and Sanjayan

Geopolymer and Portlandcement concretes in simulatedfireR. ZhaoDepartment of Civil Engineering, Monash University, Clayton, VictoriaAustralia

J. G. SanjayanFaculty of Engineering and Industrial Sciences, Swinburne University ofTechnology, Hawthorn, Victoria, Australia,

High-strength Portland cement concrete has a high risk of spalling in fire. Geopolymer, an environmentally friendly

alternative to Portland cement, is purported to possess superior fire-resistant properties. However, the spalling

behaviour of geopolymer concrete in fire is unreported. In this paper, geopolymer and Portland cement concretes of

strengths from 40 to 100 MPa were exposed to rapid temperature rises, simulating fire exposures. Two simulated fire

tests, namely rapid surface temperature rise exposure test and standard curve fire test, were conducted. In both

types of test, no spalling was found in geopolymer concretes, whereas the companion Portland cement concrete

exhibited spalling. This can be attributed to different pore structures of the two concretes. The sorptivity test found

that geopolymer concrete had a significantly higher sorption, therefore more connected pores, than Portland cement

concrete when compared at the same strength level. Hence, it is suggested that the water vapour can escape from

the geopolymer matrix quicker than in Portland cement concrete, resulting in lower internal pore pressure. The paper

concludes that, when compared at the same strength level, the geopolymer concrete possesses higher spalling

resistance in a fire than Portland cement concrete due to its increased porosity.

Introduction

Spalling of concrete

Concrete can spall when exposed to fire, leading to disintegration

of concrete structure in an accidental fire. Sometimes the spalling

is explosive. Explosive spalling is characterised by large or small

pieces of concrete being violently expelled from the surface

(Phan, 1996). The pieces may be as small as 100 mm or as large

as 300 mm in length and 15–20 mm deep in the concrete

structure elements. This type of spalling occurs during the early

part of a fire, usually within the first 30 min or so of a standard

furnace test. Various researches have been reported on the

spalling behaviour of Portland cement concrete and blended

Portland cement concrete. It is believed that high-strength con-

crete is more vulnerable to spalling when exposed to fire than

normal-strength concrete (Ali et al., 2001, 2004; Phan, 1996 ).

Spalling mechanisms

There are three main theories commonly used to explain the

cause of spalling.

(a) Moisture clog spalling: this was first proposed by Shorter and

Harmathy (1961), who hypothesised that spalling was caused

by the steam pressure build-up in the pores of concrete in

fire. During heating, the heat flow will increase the

temperature of the pore water in the concrete. When the pore

water reaches a sufficiently high temperature, it will begin to

vaporise, resulting in the increase of pore pressure. The

vapour will migrate along the temperature gradient, and

either escape from the concrete or move in the material until

it reaches a lower-temperature area and condenses. As this

process continues, pore water will build up in the cooler

region and form a saturated layer. The saturated layer will

impede the pore water from further migration. If vaporised

water cannot escape fast enough, the internal pore pressure in

the material will keep rising until it exceeds the material’s

tensile strength and causes spalling. This theory was later

adopted by Consolazio et al. (1998) and Kalifa et al. (2001).

(b) Bazant (1997) hypothesised that spalling results from

restrained thermal dilation close to the heated surface, which

leads to compressive stresses parallel to the heated surface,

further leading to brittle fractures of concrete. Similar

alternative theories include: that developed by Ulm et al.

(1999), the chemoplastic softening model; Stabler and Baker

(2000), the coupled thermomechanical damage model; and

Nechnech et al. (2002), the elastoplastic damage model.

(c) Thermal incompatibility between the aggregates and the

cement paste (Phan, 1996) may also cause spalling,

particularly in concrete with siliceous aggregates.

It has also been concluded by many researchers (Bazant and

Thonguthai, 1979; Harada and Terai, 1997; Khoury, 2000; Phan

et al., 2001) that concrete spalling is caused by the combination

163

of pore pressures and differential thermal stresses. In these

reports, it is believed that spalling is largely due to the moisture

presence and rapid heating rate. The rapid heating rate causes

thermal gradients and build-up of pore pressures. Pore pressure

and differential thermal stress can act solely or in combination

depending on the heating rate, concrete size, concrete structure,

moisture content and so on. Spalling may occur when the

combination of these pressures exceeds the tensile strength of the

concrete.

Spalling test methods

Spalling is difficult to predict and characterise. Spalling predic-

tion during heating has been largely an imprecise empirical

exercise. Large-scale tests are commonly used to assess the

spalling propensity of a concrete (Sanjayan and Stocks, 1993).

The specimens are usually tested in a fire furnace which is set to

follow a standard fire or hydrocarbon fire curve. This test method

is the standard method to predict risk of spalling. However, it is

costly and time-consuming.

Han et al. (2005, 2009) proposed a small-scale test using a gas

fire furnace. The standard size (100 mm diameter 3 200 mm

high) concrete cylinders were placed in the gas fire furnace

before testing. The cylinders were heated in the furnace by a

preset temperature plotted against time curve. This test method is

more economic and convenient than a large-scale test; however, it

does not consider the effect of size. Phan and Carino (2002) also

used the small-scale cylinder specimens to evaluate the concrete

behaviour exposed to the elevated temperatures. They concluded

that spalling tendency increased as the water/cementitious materi-

al ratio decreased. This tendency is consistent with the notion

that the tendency of explosive spalling is related to the resistance

to water vapour transport during heating. The present study used

this type of test for spalling assessment.

Hertz and Sorensen (2005) proposed a small-scale test by

exposing confined concrete cylinders to a pre-heated electric

muffle furnace. They developed a steel mantle to confine the

concrete cylinder. The confined concrete cylinder was exposed to

the pre-heated furnace chamber to achieve a rapid increase in

temperature. The end of the confined cylinder could be consid-

ered as a part of a fire-exposed surface of a concrete slab or wall.

A modified form of this test is also used in the present study.

Geopolymer cements

Ordinary Portland cement (OPC) is the main ingredient used in

the production of concrete – the most widely used construction

material in the world. In the past, concrete was simply a compo-

site of Portland cement paste with aggregates; however, modern-

day concrete incorporates other cementitious materials which act

as partial replacements of Portland cement. Fly ash is often used

in concrete as a supplementary cementitious material. Using fly

ash blended cement in concrete brings environmental benefits by

reducing resource, energy consumption and carbon dioxide emis-

sions.

Davidovits (1991) introduced the word ‘geopolymer’ to describe

an alternative cementitious material which has ceramic-like

properties. As opposed to OPC, the manufacture of fly-ash-based

geopolymer does not consume high levels of energy, as fly ash is

already an industrial by-product. This geopolymer technology has

the potential to reduce emissions by 80% (Davidovits, 1991)

because high-temperature calcining is not required. Geopolymer

can be produced by combining a pozzolanic compound or

aluminosilicate source material with highly alkaline solutions

(Davidovits and Davidovics, 1991). Fly ash, which is available

abundantly worldwide from coal-burning operations, is an excel-

lent aluminosilicate source material. In Australia, fly ash is

currently underutilised; according to figures taken from the year

2000, 12 million tonnes per annum were produced but only 10%

were effectively utilised in cementitious applications (Heindrich,

2002).

Fly ash was activated by the alkali to form an inorganic alumino-

silicates polymer which has a similar structure to the zeolite

minerals (Davidovits, 1991; Davidovits and Davidovics, 1991). It

hardens like organic resin, but is stable up to 1000,12008C. It

has been reported that geopolymer material had ceramic-like

properties with high strength and fire resistance (Davidovits,

1991; Davidovits and Davidovics, 1991). It has also been noted

by many researchers that geopolymer is a porous material.

Duxson et al. (2007) reported that a highly distributed pore

network existed in the geopolymer gel. Sindhunata et al. (2006)

reported that high-temperature curing increased the geopolymer-

isation extent and rate and increased mesopore volume.

Geopolymer generally requires high temperature to develop con-

siderable high strength. In the present paper, a curing temperature

of 808C is used. Therefore, it may not be practical to use for on-

site construction. However, in precast construction, 808C is used

for precast concrete slabs, beams, columns and pipes to eliminate

the lag between the time the on-site concrete is placed and the

time at which it can carry loads. Geopolymer can be prepared in

the precast construction plant.

It has also been noted by many researchers that geopolymer is a

porous material. Duxson et al. (2007) reported that a

highly distributed pore network existed in the geopolymer gel.

Sindhunata et al. (2006) reported that high-temperature curing

increased the extent and rate of geopolymerisation and increased

mesopore volume.

Research studies on concrete in fire include two main areas:

(a) residue strength performance of the concrete subjected to the

elevated temperature

(b) spalling behaviour of the concrete.

There are a few reports on the residual strength performance of

geopolymers at elevated temperature. Kong et al. (2007) noted

164


Geopolymer and Portland cementconcretes in simulated fireZhao and Sanjayan

that the residual strength of fly-ash-based geopolymer paste

increased by 6% after exposure to 8008C, whereas the strength of

metakaolin-based geopolymer paste reduced 34% after exposure

to 8008C. Kong et al. (2007) concluded that during heating,

the high permeability of fly-ash-based geopolymer provides the

escape routes for moisture in the matrix, thereby decreasing the

damage to the matrix. Sintering of fly ash geopolymer increases

the strength at 8008C. Similar strength increase was also reported

by Pan et al. (2009) on fly-ash-based geopolymer mortars.

However, there has been no study reported on the spalling

behaviour of the geopolymer material subjected to rapid tempera-

ture rise.

The aim of the present paper is to study the spalling behaviour of

geopolymer concrete subjected to rapid temperature rise simulat-

ing a fire. Small-scale test methods including a surface exposure

test by using an electric furnace and a gas fire furnace test were

conducted. Several Portland cement concrete cylinders with the

same strengths were tested to compare the spalling behaviour.

Further testing was carried out to explore the reasons for the

different behaviours in fire of Portland cement and geopolymer

concretes.

Spalling test

Materials

Ordinary Portland cement conforming to the requirements of

Australian standard AS3972 was used as the binder material of

Portland cement concrete. The fly ash was sourced from Pozzo-

lanic Gladstone and it was a low-calcium fly ash (class F). The

chemical composition of the fly ash was determined by X-ray

fluorescence (XRF) and is presented in Table 1.

Alkaline activators used for making the geopolymers consisted of

alkali silicate and hydroxide solutions. The alkali silicate was

Grade D sodium silicate solution with a specific gravity of 1.53

and modulus ratio (Ms) equal to 2 (where Ms ¼ SiO2/Na2O,

Na2O ¼ 14.7% and SiO2 ¼ 29.4%). The hydroxide solution was

prepared to 8 M and 12 M concentrations using a mixture of

distilled water and a commercial grade of pellets with 90% purity,

supplied by PQ Australia.

The basalt coarse aggregate had a maximum size of 14 mm and

was sourced from Readymix, Mount Shamrock quarry, Victoria,

Australia. The fine aggregate consisted of Lynhurst sand with a

fineness modulus of 2.19. The superplasticiser used in the high-

strength Portland cement concrete was Glenium 27 (high-range

water reducer, provided by BASF).

Specimen preparation

Portland cement concrete

The sand and coarse aggregate was dry mixed in a 70 litre pan

mixer for 2 min. The cement and water were added and mixed

for 2 min. After resting for a further 2 min, the concrete was

remixed for another 2 min before sampling and testing. Three

Portland cement samples were prepared for each test result. The

samples were cured for 28 days in saturated lime water kept at

238C prior to testing.

Geopolymer concrete

Sodium hydroxide pellets were mixed with distilled water to

prepare an alkaline solution one day in advance of the day of the

mixture preparation. On the day of the mixture preparation, the

sand and coarse aggregate were initially blended with fly ash and

dry mixed first in the 70 litre pan mixer for 1 min. The alkaline

solution prepared the day before was then introduced to the

mixture, and the wet mixing continued for another 4 min. Then

the mixture was cast into standard 150 mm diameter and 300 mm

high cylindrical moulds. The fresh concrete was compacted by a

vibration table to release any residual air bubbles. The concrete

in the cylinder moulds was placed in an oven at 808C for curing.

Four batches of samples were cured in the oven for 3 h, 8 h, 48 h

and 96 h respectively to reach the different target strengths. Then

the samples were placed in a constant-temperature room (238C,

50% humidity) for 28 days before testing. Six geopolymer

samples were prepared for each test result.

The mixture proportions and curing regime are presented in Table

2. The aggregates weights shown in Table 2 are in the saturated

surface dry condition. Compressive strength was tested for the

concrete cylinders by using an AMSLER compressive strength

test machine at a loading rate of 20 MPa/min. Both Portland

cement and geopolymer concrete cylinders were tested at 28 days

after curing for compressive strength.

Surface exposure test

Based on Hertz’s method (Hertz and Sorensen, 2005), the present

study has developed a similar method to test spalling of concrete

(Figures 1 and 2). A circular hole was created on the top of the

furnace chamber and it was initially covered by a ceramic board

during the pre-heating. The furnace was preheated to a tempera-

ture of 10008C and allowed to stabilise at this temperature for

30 min. Then the top cover board was removed and one end of

the 150 mm diameter cylinder was exposed through the 100 mm

Chemical constituent: % Portland cement Fly ash

SiO2 19.90 48.8

Al2O3 4.70 27.0

CaO 63.93 6.2

Fe2O3 3.38 10.2

K2O 0.446 0.9

MgO 1.30 1.4

Na2O 0.17 0.2

SO3 2.54 0.2

LOI 2.97 1.7

Table 1. Composition of fly ash and Portland cement as

determined by XRF

165



diameter hole, which transferred the heat from the furnace

chamber. This resulted in a temperature of about 8508C at the

surface of the cylinder in approximately 2 min. The temperature

plotted against time curve of the cylinder surface increased more

rapidly than the ISO standard fire curve (ISO, 1975) and was

close to the hydrocarbon fire curve (BSI, 2002) (Figure 3).

Therefore, this test can be used to simulate a standard fire or

more severe fire conditions than a standard fire. It should be noted

here that the rate of temperature rise in the first 30–40 min is the

most critical for concrete spalling (Sanjayan and Stocks, 1993);

O40* O60* O80* O110* G40y G60y G80y G100y

Water:cement 0.58 0.46 0.34 0.3 — — — —

Sand:aggregate:cement 2.4:3.8:1 1.9:3.1:1 1:2.1:1 1.4:2.9:1 — — — —

Water: kg/m3 190.00 180.00 190 143 — — — —

Cement: kg/m3 327.58 391.33 561 430 — — — —

Silica fume: kg/m3 — — — 47 — — — —

Alkaline liquid/fly ash — — — — 0.45 0.45 0.45 0.35

Fly ash: kg/m3 — — — — 381 381 381 409

NaOH solution (8M): kg/m3 — — — — 49 49 49 —

NaOH solution (12M): kg/m3 — — — — — — — 41

NaSiO4 solution (Grade D): kg/m3 — — — — 122 122 122 102

Fine aggregate: kg/m3 780.70 765.33 572.7 602 554 554 554 554

Coarse aggregate: kg/m3 1241.70 1218.33 1193 1280 1294 1294 1294 1294

Superplasticiser: l/m3 — — — 10 — — — 10

Curing condition 238C, saturated lime water 808C oven 908C oven

Curing time 28 days 3 h 6 h 48 h 96 h

* O40, O60, O80, O110 represent the ordinary Portland cement concrete specimens with a target compressive strength of 40 MPa, 60 MPa,80 MPa, 110 MPa respectively.y G40, G60, G80, G100 represent the geopolymer concrete specimens with a target compressive strength of 40 MPa, 60 MPa, 80 MPa, 100 MParespectively.

Table 2. Concrete mix proportions (kg/m3) of the concretes

Finance chamber

Thermocouple100 mm

Ceramic board cover

Cyl

inde

r

150 mm

Stee

l mou

ld

Neo

pren

e

8 m

m20

0 m

m

Figure 1. Testing rig for the surface exposure test method

Figure 2. Test furnace and specimen confined by steel mould

166



unlike other materials such as steel, performance is determined

by the level of absolute temperature and length of exposure.

The cylinder size was 150 mm in diameter and 300 mm high.

Before the test, the cylinder was confined by a 25 mm thick steel

cylindrical mould tightened by high-strength bolts. A layer of

neoprene was placed in between the cylinder and the mould to

compensate for the irregularities of the concrete cylinder surface.

A thermocouple was placed on the surface of the cylinder that

was subjected to temperature exposure. The temperature rate of

change on the surface of the cylinder was recorded by a

thermocouple. The test was conducted for a duration of 1 h.

The moisture content of the specimen was measured by a

TRAMEX electronic moisture meter before the test. Also,

identical specimens were dried and weighed to determine the

moisture content of the specimens at the time of test (Table 3).

After the test, the specimens exposed to simulated fire test also

were weighed and the degree of spalling was observed.

Gas fire furnace test

A hydrocarbon gas fire furnace can generate rapid temperature

rise during a short time (BSI, 2002). The gas fire furnace test

using small cylinders has been used previously to assess con-

cretes for spalling (Han et al., 2005, 2009; Phan and Carino,

2002). In the test used in this study, concrete cylinders 100 mm

in diameter and 200 mm high were placed in the furnace (Figure

4), which was heated by a preset ISO standard fire temperature

plotted against time curve (ISO, 1975). In contrast with the

surface exposure test (described in the previous section), the

cylinders were not confined. Therefore, the confining effect of

surrounding cool concrete in a large specimen was not simulated

in these tests. However, this test was used to investigate the

spalling behaviour of geopolymer concrete in standard fire curve

exposure and also to confirm the effects of the surface exposure

test method by comparing the same samples.

Results and discussion

Surface exposure test results

No spalling was observed on the surface of any of the geopoly-

mer concrete cylinders. The heat exposure 100 mm diameter

circular area on the cylinder surface turned a brown colour. This

may have been caused by the oxide of the iron from the fly ash.

1200

1000

800

600

400

200

0

Tem

pera

ture

: °C

0 5 10 15 20 25 30 35 40 45 50 55 60Time: min

Hydrocarbon fire curve

Surface exposuretest curve

Standard fire curve

Figure 3. Temperature plotted against time curves in the spalling

test

O40 O60 O80 O110 G40 G60 G80 G100

Compressive strength: MPa 42 63 82 110 37 62 78 98

Moisture content: % 5 5.3 5.5 5.5 5.2 5.1 5.2 5.3

Spalling specimen number* 1 3 3 3 0 0 0 0

No spalling specimen number* 2 0 0 0 6 6 6 6

Average spalling depth: mm 0.1 0.3 0.7 2.5 0 0 0 0

Spalling area percentage: %y 1.6 30 83 100 0 0 0 0

* Spalling and no spalling specimen number means the numbers of concrete specimens that showed spalling or no spalling after the test.y Spalling area percentage represents the percentage of spalling area in the total heated area.

Table 3. Compressive strength and spalling occurrence of the

test samples

Figure 4. Concrete cylinders placed in the gas fire furnace before

testing

167



The outer unheated 50 mm layer remained intact and acted as a

restraint to the heat exposure area (Figure 5).

High-strength Portland cement concrete samples (O80, O110)

showed signs of spalling (Figure 5). Significant spalling occurred

on the O110 Portland cement concrete cylinders. Minor spalling

occurred on O60 specimens. The spalling area or depth of O60

was much smaller compared with the high-strength Portland

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 5. Geopolymer and Portland cement concrete cylinders

after surface exposure test: (a) G40; (b) O40; (c) G60; (d) O60; (e)

G80; (f) O80;(g) G100; (h) O110

168



cement concretes. One cylinder spalled out of three tested in

40 MPa specimens; however, the degree of spalling was very

small (Figure 5).

The spalling percentages are presented in Table 3. The numbers

represent the percentage of spalled area of the total exposed area.

O80 and O110 specimens had 83% and 100% spalling percen-

tages, which are the maximum spalling patterns in the test. Other

Portland cement cylinders with lower compressive strengths

exhibited significantly lower spalling percentages.

Cracking or popping sounds were noticed during the test, which

were understood to be accompanying the incidences of spalling.

The times of the sounds were recorded. Higher-strength concretes

(O80, O110) were accompanied by louder noise and much more

severe spalling than low-strength concretes. The recording of the

times of the sounds indicated that spalling occurred in the initial

2–3 min of the rapid heating. No sound was recorded after the

first 5 min (Figure 6).

Gas fire furnace test results

The gas fire temperature rate was set to follow the ISO standard

fire (ISO, 1975). Figure 7 shows the test results of geopolymer

and Portland cement concrete cylinders with strength levels

varying from 40 to 110 MPa subjected to the standard fire. No

spalling occurred on any of the geopolymer concrete specimens.

High-strength Portland cement concrete specimens exhibited

severe spalling. Normal- and low-strength Portland cement con-

crete specimens exhibited minor spalling. These results are

consistent with the surface exposure test results reported above.

Sorptivity testThe spalling test results demonstrate that geopolymer concrete

has a better resistance to spalling than Portland cement concrete.

In order to explore further the reason for this phenomenon,

sorptivity tests on the geopolymer and Portland cement concretes

were carried out.

Sorptivity represents the material’s ability to absorb and transmit

water through the matrix by capillary suction. Compared with

permeability, which is used to measure the flow of water under

pressure in a saturated porous material, sorptivity is a more

suitable parameter for evaluating the pore connectivity and

capillary network, which is a major factor influencing water

transmission in the concrete when subjected to fire (Consolazio et

al., 1998; Kalifa et al., 2001; Shorter and Harmathy, 1961). Thus,

the sorptivity test was conducted to compare the pore structure

characteristics of both types of concrete (geopolymer and Port-

land cement concretes).

Test method

Following 28 days of curing, three concrete cylinders for the

water sorptivity test were prepared for each batch. The sorptivity

tests were conducted according to the test method specified by

ASTM C1585 (ASTM, 2004). According to this test method, the

test specimens were first dried until constant weight at 238C in a

desiccator before testing. The test specimen was exposed to the

water at one end by placing it in a pan (Figure 8). The water in

the pan was maintained at about 5 mm above the base of the

specimen. The lower surface on the sides of the specimen was

coated with paraffin to achieve unidirectional flow. At certain

times (0, 5, 10, 20, 30, 60, 180, 360, 1440 min), the weight of the

specimen was measured. The volume of water absorbed was

calculated.

The sorptivity coefficient (Collins and Sanjayan, 2008; Gonen

and Yazicioglu, 2007; Igarashi et al., 2005; Khan, 2003; Olivia et

al., 2008) was obtained by the following equation

Q

A¼ k

ffiffi

tp

1:

where Q is the volume of water absorbed (mm3); A is the cross-

sectional area of specimen that was in contact with water (mm2);

t is the time (min); and k is the sorptivity coefficient of the

specimen (mm/min1=2) – this is the sorptivity measured per mm2

of wetted area per min1=2. To determine the sorptivity coefficient,

Q/A was plotted against the square root of time (ffiffi

tp

), then k was

calculated from the slope of the linear relation between (Q/A) andffiffi

tp

.

Results and discussion

As shown in Figure 9, the sorptivity coefficients (k) of geopoly-

mer concrete specimens were significantly higher than for the

Portland cement concrete specimens. When compared at the same

strength level, the k value of G40 is about twice that of the k

value of O40. On the higher strength levels, the difference of k

value was increased. The k value of G90 is almost three times

greater than the k value of O110.

Portland cement concrete from the O60 to O110 specimens

showed a higher compressive strength the lower the sorptivity

coefficient pattern (Figure 10). This is consistent with past

reports (Gonen and Yazicioglu, 2007; Igarashi et al., 2005;

O110

O80

O60

O40

Con

cret

e m

ixtu

re t

ype

Spalling testing time: min0 1 2 3 4 5

Loud crack

Pop

Figure 6. Sound recording of spalling time of Portland cement

concretes in the surface exposure test

169



(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 7. Geopolymer and Portland cement concrete cylinders

after the gas fire furnace test: (a) G40; (b) O40; (c) G60; (d) O60;

(e) G80; (f) O80; (g) G100; (h) O110

170



Khan, 2003). The addition of silica fume in the O110 specimen

acts as a fine filler to fill the pores in the paste and interfaces

between aggregate and paste (Gonen and Yazicioglu, 2007;

Igarashi et al., 2005). Therefore, the O110 concrete processes a

densified structure resulting in a significantly low sorptivity

coefficient. O40, however, showed a similar value of sorptivity

coefficient compared with O60, instead of a higher value. Olivia

et al. (2008) proposed that variation of aggregate content,

grading and binder content can influence the sorptivity coeffi-

cient of concrete. It was observed (Olivia et al., 2008) that

increasing the aggregate/binder ratio can result in a decrease of

the water absorption. Therefore, it is suggested that the high

water/binder ratio was not the major factor to influence O40’s

sorption. It was mainly influenced by its high aggregate/cement

ratio, which was 6.2 compared with O60 with an aggregate/

cement ratio of 5.

The sorptivity coefficient of geopolymer concrete decreased with

increasing strength from G40 to G80 (Figure 10). This pattern is

also consistent with the Portland cement concrete. However,

G100 showed a significantly high sorptivity coefficient. This

phenomenon can be explained by the continuous pore structure

development of geopolymer gel during the extended oven-curing

regime, resulting in a more highly developed porous structure

compared with that produced by the normal length of oven curing

(Sindhunata et al., 2006).

A high sorptivity coefficient indicates the existence of a highly

connected porous structure of the material. When geopolymer

concrete is subjected to fire, according to the moisture clog

theory (Shorter and Harmathy, 1961), it is suggested that the

highly porous structure will be beneficial to decrease the thermal

gradient because the high volume of pore water in the structure

can absorb heat and distribute the heat flow in the matrix. A

highly porous structure can also accelerate the water flow in the

concrete skeleton, thereby slowing the temperature rise in the

concrete. It is also suggested that the highly connected porous

structure will slow the pressure build-up by releasing the water

vapour from the concrete.

ConclusionsThis paper has compared the spalling behaviour of geopolymer

and Portland cement concretes by using the surface exposure test

and standard gas furnace fire test. No spalling occurred on any of

the geopolymer concrete specimens, while spalling was observed

on some of the companion Portland cement concrete specimens.

The high-strength Portland cement concrete cylinders (O110,

O80) displayed severe spalling. Normal-strength Portland cement

concrete cylinders (O60, O40) exhibited minor spalling. The test

results on both Portland cement and geopolymer specimens by

using the surface exposure test and standard gas fire test are

consistent. These results showed that the geopolymer concrete

had a better spalling resistance to rapidly rising temperature

exposure than Portland cement concrete.

The sorptivity test demonstrated that the geopolymer concrete

specimen’s structure is more porous than the Portland cement

concrete specimens. The highly porous structure of geopolymer

concrete facilitates the release of the internal steam pressure

during heating. Hence, less tensile stress is imposed in the

Concrete

Paraffin

Water

Figure 8. Sorptivity test set-up

20 40 60 80 100 120Compressive strength: MPa

Geopolymer

OPC

12

10

8

6

4

2

0

Sorp

tivity

coe

ffic

ient

,:

10m

m/m

ink

��

30·

5

Figure 9. Sorptivity coefficient plotted against strength of

concrete

3·5

3·0

2·5

2·0

1·5

1·0

0·5

0

Sorp

tion:

mm

T 0·5 0·5: min

0 50 100 150 200 250 300 350

G100G80

G60G40

O60O40O80

O110

Figure 10. Sorptivity plots for Portland cement and geopolymer

concrete

171



geopolymer concrete than for Portland cement concrete during

heating, thereby reducing the geopolymer’s risk of spalling. This

comparison was carried out at the same strength levels of the two

concretes.

It can be concluded that, at the same strength level, geopolymer

concrete has a significant advantage over Portland cement con-

crete when exposed to fire.

AcknowledgementsThe authors gratefully acknowledge the financial support from

the Australian Research Council Discovery Grant No.

DP0664309 for this research work. Laboratory assistance from

Jeff Doddrell and Mr Long are also acknowledged.

REFERENCES

Ali F, O’Connor D and Abu-Tair A (2001) Explosive spalling of

high-strength concrete columns in fire. Magazine of Concrete

Research 53(3): 197–204.

Ali F, Nadjai A and Talamona D (2004) Assessment of the

susceptibility of normal and high strength concrete for

explosive spalling. Journal of Applied Fire Science 13(1):

79–88.

ASTM (American Society for Testing and Materials) (2004)

ASTM C1585: Standard test method for measurement of rate

of absorption of water by hydraulic-cement concretes. ASTM

International, West Conshohocken, Pennsylvania.

Bazant ZP (1997) Analysis of pore pressure, thermal stresses and

fracture in rapidly heated concrete. Proceedings of an

International Workshop on Fire Performance of High-strength

Concrete. US Department of Commerce, Washington DC,

NIST Publication 919, p. 155.

Bazant ZP and Thonguthai W (1979) Pore pressure in heated

concrete walls: theoretical prediction. Magazine of Concrete

Research 31(107): 67–76.

BSI (British Standards Institution) (2002) BS EN 1991–1–2:

Eurocode 1: Actions on structures: Part 1-2: General actions.

Actions on structures exposed to fire. BSI, London.

Collins FG and Sanjayan JG (2008) Unsaturated capillary flow

within alkali activated slag concrete. Journal of Materials in

Civil Engineering 20(9): 565–570.

Consolazio GR, McVay MC and Rish JW III (1998) Measurement

and prediction of pore pressures in saturated cement mortar

subjected to radiant heating. ACI Materials Journal 95(5):

525–536.

Davidovits J (1991) Geopolymers – inorganic polymeric new

materials. Journal of Thermal Analysis 37(8): 1633–1656.

Davidovits J and Davidovics M (1991) Geopolymer. Ultra-high

temperature tooling material for the manufacture of advanced

composites. Proceedings of an International SAMPE

Symposium and Exhibition, California, USA, vol. 36, Pt 2,

pp. 1939–1949.

Duxson P, Fernandez-Jimenez A, Provis JL, Lukey GC, Palomo A

and van Deventer JSJ (2007) Geopolymer technology: the

current state of the art. Journal of Materials Science 42(9):

2917–2933.

Gonen T and Yazicioglu S (2007) The influence of compaction

pores on sorptivity and carbonation of concrete. Construction

and Building Materials 21(5): 1040–1045.

Han CG, Hwang YS, Yang SH and Gowripalan N (2005)

Performance of spalling resistance of high performance

concrete with polypropylene fiber contents and lateral

confinement. Cement and Concrete Research 35(9): 1747–

1753.

Han CG, Han MC and Heo YS (2009) Improvement of residual

compressive strength and spalling resistance of high-strength

RC columns subjected to fire. Construction and Building

Materials 23(1): 107–116.

Harada K and Terai T (1997) Heat and Mass Transfer in the Walls

Subjected to Fire. National Institute of Standards and

Technology, Gaithersburg, Maryland, NISTIR 6030, p. 423.

Heindrich C (2002) Ash utilisation – an Australian perspective.

Proceedings of Geopolymers 2002, Melbourne, Australia 1

(CD-ROM).

Hertz KD and Sorensen LS (2005) Test method for spalling of fire

exposed concrete. Fire Safety Journal 40(5): 466–476.

Igarashi S-I, Watanabe A and Kawamura M (2005) Evaluation of

capillary pore size characteristics in high-strength concrete at

early ages. Cement and Concrete Research 35(3): 513–519.

ISO (International Organization for Standardization) (1975) ISO

834: Fire-resistance tests – elements of building construction.

ISO, Geneva.

Kalifa P, Chene G and Galle C (2001) High-temperature behaviour

of HPC with polypropylene fibres – from spalling to

microstructure. Cement and Concrete Research 31(10): 1487–

1499.

Khan MI (2003) Permeation of high performance concrete.

Journal of Materials in Civil Engineering 15(1): 84–92.

Khoury GA (2000) Effect of fire on concrete and concrete

structures. Progress in Structural Engineering and Materials

2(4): 429–447.

Kong DLY, Sanjayan JG and Sagoe-Crentsil K (2007) Comparative

performance of geopolymers made with metakaolin and fly

ash after exposure to elevated temperatures. Cement and

Concrete Research 37(12): 1583–1589.

Nechnech W, Meftah F and Reynouard JM (2002) An

elasto-plastic damage model for plain concrete subjected to

high temperatures. Engineering Structures 24(5):

597–611.

Olivia M, Nikraz H and Sarker P (2008) Improvements in the

strength and water penetrability of low calcium fly ash based

geopolymer concrete. Proceedings of the 3rd ACF

International Conference – ACF/VCA, Ho Chi Minh City,

Vietnam, 384.

Pan Z, Sanjayan JG and Rangan BV (2009) An investigation of

the mechanisms for strength gain or loss of geopolymer

mortar after exposure to elevated temperature. Journal of

Materials Science 44(7): 1873–1880.

Phan LT (1996) Fire Performance of High-Strength Concrete: A

172



Report of the State-of-the-Art. National Institute of Standards

and Technology, Maryland, USA, p. 5.

Phan LT and Carino NJ (2002) Effects of test conditions and

mixture proportions on behavior of high-strength concrete

exposed to high temperatures. ACI Materials Journal 99(1):

54–66.

Phan LT, Lawson L and Davis F (2001) Effects of elevated

temperature exposure on heating characteristics, spalling, and

residual properties of high performance concrete. Materials

and Structures 34(2): 83–91.

Sanjayan G and Stocks LJ (1993) Spalling of high-strength silica

fume concrete in fire. ACI Materials Journal 90(2): 170–173.

Shorter GW and Harmathy TZ (1961) Discussion of the fire

resistance of prestressed concrete beams. Proceedings of the

Institution of Civil Engineers 20(2): 313–315.

Sindhunata X, van Deventer JSJ, Lukey GC and Xu H (2006)

Effect of curing temperature and silicate concentration on fly-

ash-based geopolymerization. Industrial and Engineering

Chemistry Research 45(10): 3559–3568.

Stabler J and Baker G (2000) Fractional step methods for thermo-

mechanical damage analyses at transient elevated

temperatures. Journal of Numerical Methods in Engineering

48(5): 761–785.

Ulm FJ, Coussy O and Bazant ZP (1999) The ‘chunnel’ fire. I:

chemoplastic softening in rapidly heated concrete. Journal of

Engineering Mechanics 125(3): 272–281.

WHAT DO YOU THINK?

To discuss this paper, please submit up to 500 words to

the editor at www.editorialmanager.com/macr by 1

September 2011. Your contribution will be forwarded to

the author(s) for a reply and, if considered appropriate

by the editorial panel, will be published as a discussion

in a future issue of the journal.

173



geopolymer and portland cement concretes in simulated Þre · 2016-12-05 · geopolymer concrete...

Documents