The properties of fly ash based geopolymer mortars made with dune sand

S. Chuah, W.H. Duan, Z. Pan, E. Hunter, A.H. Korayem, X.L. Zhao, F.

Collins, J.G. Sanjayan

PII: S0264-1275(15)30926-6
DOI: doi: 10.1016/j.matdes.2015.12.070
Reference: JMADE 1094

To appear in:

Received date: 27 July 2015

Revised date: 5 December 2015
Accepted date: 14 December 2015

Please cite this article as: S. Chuah, W.H. Duan, Z. Pan, E. Hunter, A.H. Korayem, X.L.
Zhao, F. Collins, J.G. Sanjayan, The properties of y ash based geopolymer mortars made
with dune sand, (2015), doi: 10.1016/j.matdes.2015.12.070

This is a PDF le of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its nal form. Please note that during the production process
errors may be discovered which could aect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

The Properties of Fly Ash Based Geopolymer Mortars Made with Dune Sand

S. Chuaha, W.H. Duana, Z. Panb*, E. Huntera, A. H. Korayemc, X.L. Zhaoa, F. Collinsd and J.
G. Sanjayane
a

T
Department of Civil Engineering, Monash University, Clayton, Victoria, Australia, 3800
b
Institute for Infrastructure Engineering, University of Western Sydney, Kingswood, Australia, 2747

IP
c
Department of Civil Engineering, Iran University of Science and Technology, Tehran, Iran, 1684613114
d

R
Institute for Frontier Materials, Deakin University, Geelong Waurn Ponds, Australia, 3216
e
Faculty of Engineering & Industrial Sciences, Swinburne University of Technology, Hawthorn, Victoria,

SC
Australia
Keywords: Microstructure; Compressive Strength; Elastic Moduli; Aggregate; Alkali
Activated Cement

NU
ABSTRACT
MA
This paper reports the properties of fly ash based geopolymer mortars made with dune sand.

The geopolymer mortars of different cation type, namely sodium based (Na), potassium
D

based (K) and a mixed Na/K, were prepared with Dune Sand (DS) and River Sand (RS). The
TE

corresponding geopolymer pastes were also prepared. A series of tests including compressive

strength, modulus of elasticity, splitting tensile strength, microanalysis (using scanning

P

electron microscopy), porosity (using mercury intrusion porosimetry), sorptivity and air void
CE

(using section analysis method) were carried out. The results showed a strong correlation
AC

between strength and porosity of geopolymeric materials. The addition of DS had influences

on the chemical compositions and physical properties of geopolymers mortars. These

influences were dependent on the type of cation. Based on the results of mechanical

properties, DS can be utilised as the fine aggregate for the production of geopolymer based

construction material.

*
Corresponding author. E-mail address: Z.Pan@uws.edu.au

Address: Institute for Infrastructure Engineering, University of Western Sydney, Building Z, Kingswood
Campus, Second Avenue, Kingswood. NSW 2747. Australia

Tel.: +61 2 47360088

1
 ACCEPTED MANUSCRIPT

INTRODUCTION

Geopolymer has garnered immense attention since emerging as the new generation of

the environmentally friendly alternative to Ordinary Portland Cement (OPC). Geopolymer is

T
an amorphous to semi-crystalline three-dimensional framework structure by the

IP
geopolymerization of silicon, aluminium and oxygen with the presence of a highly

R
concentrated alkali as the reaction catalyst [1]. As one of common raw materials, fly ashes

SC
consist largely of alumina and silica that could be dissolved from the solid particles in

hydrothermal conditions and a high alkali environment. Alkali-activated fly ash derives its

NU
strength from the polymerisation of aluminosilicate gels instead of calcium silicate hydrate
MA
gels found in OPC [2].

Owing to the difference in binder forming the composite matrix, geopolymer offer
D

superior durability properties than OPC in terms of enhanced fire resistance [3], acid
TE

resistance [4] and immune to alkali-aggregate reaction [5]. The excellent durability of

geopolymer in aggressive environments to resist ingress of aggressive agents arise from the
P
CE

geopolymers low permeability [6]. These characteristics have vast applications ranging from

the fabrication of geopolymeric structural members [7, 8], durable well cement for carbon
AC

sequestration [9], ability to immobilize toxic [10] or radioactive [11] contaminants by waste

encapsulation and improve the impact resistance [12].

Compared with OPC, the manufacture of geopolymer is desirable to convert waste

material into useful construction material while leaving a smaller carbon footprint. For

making concrete in an environmentally friendly way, designers may consider not only

replacing OPC with green binders (e.g. geopolymer) but also using aggregates in a green way

because aggregates generally account for 60 to 80 percent of the volume of the concrete. In

2
 ACCEPTED MANUSCRIPT

concrete, aggregates which may be coarse or fine, are critical ingredients to enhance the

mechanical properties, volume stability and inexpensive materials. For centuries, river sand

(RS) is being used as fine aggregate. Presently, RS is dredged from river channels or

T
floodplains. This extraction process can be detrimental to the marine ecosystem as a

IP
consequence of river bank erosion, shifts in river courses, suspended solids contamination

R
and floods. This has resulted in the search for a suitable alternative. The dune sand is

SC
abundant in many countries such as Algeria, Tunisia and Oman. In Australia, dunes were

estimated to cover 40% of the surface area [13]. In these countries, a great availability of

NU
dune sand provides a practical solution for using it as the fine aggregate in concrete. In recent
MA
years, a number of studies have been conducted to investigate the use of dune sand (DS) in

concrete [14]. Those who were the forerunners of this field examined the use of DS as fine
D

aggregate discovered that the poor gradation of DS due to the presence of very fine particles
TE

(VFP) causes a decline in the compressive strength when DS completely replaced normal

sand as the fine aggregate in concrete [15]. However, if the proportion of VFP is optimized
P

by adjusting the ratio of sand to cement, the strength of DS concrete could reach a higher
CE

level as compared to concrete made with normal sand [16]. The major reason for this
AC

enhancement is pozzolanic effect [17]. Si4+ was found to be dissolved from the dune sand

particles. This cation reacts with the anhydrous or hydrated components of cement in the

presence of water. As a result, the mechanical strength was enhanced by the new formed

mineral phases in the cement matrix [17].

Besides DS, Tasong et al. [18] have found that other siliceous rocks such as basalt,

quartzite, and silica sand could also chemically interact with the cement pore solution. In a

alkali-activated fly ash system, Lee and Deventer [19] provided the evidence that the

3
 ACCEPTED MANUSCRIPT

interaction (between RS particles and pore solution) was further enhanced due to the highly

alkaline activating solutions used in such a system.

Not only influenced by the alkaline nature of pore solution, the chemical stability of

T
siliceous aggregates is also influenced by their fineness. In general, dune sand has a much

IP
higher fineness as compared to normal sand. In dune sand, typically 25% of grains (by weight)

R
have a diameter smaller than 150 m. On the other hand, this fraction is less than 6% in

SC
normal sand used in concrete, as specied in ASTM C33. In a geopolymer system, there

NU
might be chemical interactions taking place on the surface of these very ne particles. At the

highly alkaline activating solutions, the extent of such reactions might be significant. As a
MA
result, the physical and chemical performances of geopolymer systems would be affected [20,

21]. Therefore, the very fine dune sand particles may play a different role in matrix formation
D

in a geopolymer system, in comparison with those in an OPC system.

TE

Although the use of DS in OPC based materials has been studied to a certain degree,
P

the use of DS in geopolymer binder received little attention and is not yet researched despite
CE

the vast resources of desert sand in the Australian outback. Therefore, this approach

regarding the feasibility of geopolymer and DS to gradually phase out traditional concrete is
AC

certainly worth exploring. The research will present the results of the conventional

mechanical tests (compressive strength, indirect tensile strength and elastic modulus)

performed on the DS geopolymer pastes and mortars made with different alkali activators.

The effects of DS on strength will be discussed in terms of pore structure of geopolymer

matrix and microstructural characteristics of the interfaces between DS particles and the

geopolymer matrix.

4
 ACCEPTED MANUSCRIPT

EXPERIMENTAL PROGRAM

Characterisation of materials

All geopolymer samples were prepared with low-calcium Class F fly ash as per

T
ASTM C618-12, which was supplied by Cement Australia. Table 1 presents the chemical

IP
composition of the fly ash, as determined by X-ray Fluorescence (XRF) analysis.

R
SC
Table 1 Chemical composition of materials
Materials Al2O3 SiO2 CaO Fe2O3 K2O MgO Na2O SO3 LOI
Fly ash (%) 30.60 48.40 2.70 12.10 0.30 1.30 0.20 0.30 1.70

NU
DS (%) 2.00 94.8 0.23 0.66 0.34 0.11 0.06 - 0.83
RS (%) 1.05 96.7 0.08 0.56 0.12 - 0.12 - 0.29

Both dune sand and river sand were used as fine aggregates. The chemical
MA
compositions of these sands were analysed by XRF as shown in Table 1. The DS and RS

exhibit similar chemical compositions but different gradation and grain size. Figure 1 shows
D

the sieve analysis performed on DS and RS to characterise the particle size distribution (PSD)
TE

for sand aggregates. DS has a steep curve and lies within the narrow range of particle size,
P

inferring that DS is poorly graded. In contrast, RS is well graded compared to DS as it covers

CE

a wider range of sand sizes. The PSD curve suggests DS is finer than RS which corresponds

to the DSs lower Coefficient of uniformity (Cu) value of 2.13 than RSs 2.83.
AC

100%
Cumulative % Passing

80%

60%
Desert
Sand
40%
River
Sand
20%

0%
0.01 0.1 1 10
Sieve Size (mm)
Fig. 1 Particle size distribution curve
5
 ACCEPTED MANUSCRIPT

The alkaline activators required to synthesise geopolymers include potassium silicate,

sodium silicate, potassium hydroxide and sodium hydroxide. The potassium and sodium

silicate solutions were supplied by PQ Australia, under the commercial brand name of Kasil

T
2040 (26.7% SiO2, 13.3% K2O and 60.0% H2O) and Grade D (29.4% SiO2, 14.7% Na2O and

IP
55.9% H2O).

R
The pellets of Potassium Hydroxide (KOH) and Sodium Hydroxide (NaOH) were

SC
purchased from Redox Chemicals and PQ Australia respectively. Both solid potassium and

NU
sodium hydroxide were dissolved using distilled water to provide highly alkaline solutions.

The detailed mixes design of geopolymer pastes is given in Table 2. The compositions of
MA
mortar are the same as those of the corresponding pastes except that sand (two times by

weight of fly ash) was added.

D

Table 2 Compositions and compressive strength of geopolymer pastes

TE

Materials Mass (kg/m3)

Na Na/K K
P

Fly ash 1050 1050 1050

Sodium hydroxide solution 135 175 -
CE

Sodium silicate solution 338 - -

Potassium hydroxide solution - - 262.5
Potassium silicate solution - 350 262.5
Component
AC

M2O/Al2O3 0.42 0.42 0.70

H2O/Al2O3 5.24 5.93 6.3
SiO2/Al2O3 3.23 3.21 3.80
Compressive strength 77 55 27
Note: M = Na, K or Na/K

Mixing and casting

The dry ingredients were mixed together briefly before the alkaline solution was

added. The whole mixing procedure took place in a constant-speed mixer, in accordance with

ASTM C192-07. Mixes were cast into hollow PVC tubes of 50 mm length and 25 mm

diameter. The tops were sealed with consumer cling-wrap to avoid moisture evaporation and
6
 ACCEPTED MANUSCRIPT

the samples were cured at 60C for 24 hours [7]. The PVC tube moulds were removed to

uncover the sufficiently hardened mortar samples, which were ground down to have smooth

and parallel top and bottom edges.

T
IP
Mechanical properties testing

R
At least three samples of each mix underwent uniaxial compression testing after 28

SC
days in accordance with ASTM C39. The Shimadzhu 300 kN Compression Test Machine

was used loaded at a strain-rate of 0.02 mm/minute until compression failure. The strain

NU
gauges were attached to the sample to record the deformation of the sample as it was loaded
MA
under compression. The deformation was recorded to plot the stress versus strain curve. The

modulus of elasticity was determined as the secant modulus obtained from the gradient of the

plot at a stress level equal to 40% of the compressive strength of the sample. Splitting tensile
D
TE

tests were performed on four samples of each mix. The samples were positioned on their

sides, so that the long axis was perpendicular to the load direction. The tests were carried out
P

in accordance with ASTM C496.

CE

SEM and EDS analysis

AC

A JEOL 7001F FEG high resolution scanning electron microscope (SEM) was used to

investigate the morphology and microstructures of the geopolymer mortars. The elemental

composition of samples was observed using EDAX (energy-dispersive X-ray) spectrometry.

Sample preparation is very important for accurate SEM analysis of cementitious materials.

The grinding and polishing process may alter the pore structure. To stabilize the

microstructure of the specimens, they were impregnated in an EPO-TEK 301 epoxy resin.

Then the specimens were placed in a vacuum chamber operated at a pressure of 30 mbar.

7
 ACCEPTED MANUSCRIPT

Following the impregnation, the specimens were carefully grinded, lapped, and polished. The

finished samples were coated with gold-palladium.

MIP testing

T
IP
The porous structure of samples was characterized using mercury intrusion

R
porosimetry (MIP). For this test, small cube samples of approximately 5 x 5 x 5 mm were cut

SC
from the larger samples. The sample of this size almost fills the penetrometer, which requires

less mercury to fill the penetrometer. The use of large single sample promotes the accuracy of

NU
the entire analysis process due to 1) reduce mercury compressibility effects and 2) increase
MA
accuracy of density measurements. To remove moisture, the samples were dried at ambient

temperature in a vacuum oven. MIP tests were performed by using a Micromeritics 9420

Autopore III porosimeter.

D
TE

Sorptivity testing
P

Three samples of each mix underwent the sorptivity test, conforming to ASTM C1585.
CE

The entire surface for each sample was covered with three layers of tightly applied

commercial Duct-tape to make these sides impermeable. The bottom face of each sample
AC

was placed on bolts submerged in a shallow bath of water. The surface of the water was 1

mm above the base of each sample. At fixed points in time the weight gain attributed to the

absorption of water was measured. The coefficient of sorptivity, k was calculated by taking

the capillary activity into account and moderating it with regard to the influence of geometric

properties.

8
 ACCEPTED MANUSCRIPT

Air voids measurements

The modified point-count method was employed to characterize the air voids. The

implementation of procedures is in agreement with ASTM C457. Two samples with polished

T
surfaces from each mixes were analysed. The microscope used in this study was a

IP
stereoscopic microscope (ProSciTech). The machine was operated at a magnification of 50.

R
The total number of points and the number of points crossing air voids (for each surface)

SC
were measured. The air content was calculated by the relative amount of points crossing air

NU
voids to the total number of points. The area of approximate 35 35 mm on each polished

section was scanned.

MA
RESULTS AND DISCUSSION
D

Strength and pore structure of geopolymer pastes

TE

As the strength of cementitious composites is essentially dependent on its binder, the

P

strength of geopolymer pastes is studied and presented in Table 2. From Table 2, it was found
CE

that strength of geopolymer pastes was influenced by fabrication parameters. The effects of

these parameters (Si/Al ratio, type of alkali activators, alkali concentration and H2O/Al2O3
AC

ratio) on compressive strength have also been identified in the literatures and studied in some

depth [22-26]. All these parameters are varied through different mixes in the current research.

However, each particular parameter contributes to the strength at a different degree. Among

all these parameters, the H2O/Al2O3 ratio shows a strong correlation to strength in the current

research. A high H2O/Al2O3 ratio is associated with a geopolymer paste of low strength.

Water plays a key role in the synthesis of geopolymers [27]. On the one hand, water is the

medium for the dissolution of aluminosilicate precursors and speciation. Moreover, water

9
 ACCEPTED MANUSCRIPT

helps with various ion transfers and polycondensation of Al and Si monomeric and

oligomeric species. As water is not a reactive medium in geopolymerization, it remains

entrapped in the matrix after geopolymerization is completed.

T
The pore structure of geopolymer pastes is presented in Fig. 2. The figure shows a

IP
sharp increase in the cumulative pore volume curve at pore sizes of 30 nm, 110 nm and 1100

R
nm in samples Na, Na/K and K respectively. These pore sizes are referred to a threshold

SC
diameter above which the amount of mercury intrusion is negligible, and immediately below

NU
which commences the significant portion of intrusion.
MA
D
P TE

Fig. 2 Cumulative mercury intrusion against Fig. 3 Differential intrusion curves of

CE

pore diameter of geopolymer pastes geopolymer pastes

AC

Differential intrusion curves in Fig. 3 provide extended information on the pore

structure of geopolymer pastes. For all geopolymer pastes, pore size distributions were

mainly characterised by a class of pore access located in the vicinity of 10 nm and above.

Pores in 10 nm is believed to represent the main porous structure, which is the cluster (s) of

aluminosilicate particles that constitute the Si-O-Al- network of geopolymer matrix [28].

Different from the Na-based geopolymer paste, the addition of potassium in activator

leads to two larger pore accesses around 35 and 80 nm, with a lower peak density. This

suggests that the extent of geopolymerization in Na/K geopolymer pastes is lower than that in
10
 ACCEPTED MANUSCRIPT

Na geopolymer pastes. The produced geopolymer gels are not enough to fill the gaps among

unreacted fly ash particles, leaving relatively large voids within the geopolymer matrix.

Compared to Na and Na/K geopolymer pastes, the K geopolymer paste has a higher

T
H2O/Al2O3 ratio. After geopolymerization, the excess water remains entrapped in the pores to

IP
generate larger pores and greater pore volume. As a result, an accessible large pore of around

R
1000 nm is clearly observed in K geopolymer pastes.

SC
Besides the pore size distribution, porosity is a more important factor that influences

NU
the mechanical properties of cementitious materials. A semi-empirical relationship between

strength and porosity has been reported in the literature[29]. In the current study, the porosity
MA
is calculated based on the pore whose diameter is greater or equal to 50 nm, as such pore is

believed to have influence on the strength of porous materials [30]. On plotting (Fig. 4) the
D

compressive strength of geopolymers against their respective porosity of large pore, a strong
TE

correlation between strength and porosity was observed, suggesting that the strength of
P

geopolymer pastes is intrinsically dependent on their pore structure.

CE
AC

Fig. 4 Compressive strength against porosity Fig. 5 Compressive strength of geopolymer

of geopolymer pastes pastes and mortars

11
 ACCEPTED MANUSCRIPT

Properties of geopolymer mortars

The H2O/Al2O3 ratio shows a pronounced effect not only on strength (discussed above)

but also on water absorption capacity. Fig. 6 shows that a high H2O/Al2O3 ratio is associated

T
with high water absorption capacity of geopolymer mortars. The reason for this correlation

IP
can be understood by a further examination of the MIP results presented in Fig.2 that

R
indicates a high H2O/Al2O3 ratio results in a large threshold diameter. The threshold diameter

SC
is defined as the minimum diameter of pores that form a continuous network through the

NU
specimen. Thus, in a specimen with large threshold diameter, an increase in sorptivity may be

due to the rapid filling water in the pores through the large interconnected pores.
MA
D
P TE
CE

Fig.6 H2O/Al2O3 ratios against sorptivity of Fig.7 Sorptivity against compressive strength
DS and RS mortars of DS and RS mortars
AC

In general, a high absorptivity coefficient is associated with a large pore volume in the

porous materials. This correlation has been extensively reported in cementitious materials

[31]. Therefore, it is not surprising that a correlation exists between the strength and

absorptivity coefficient for geopolymer mortar (Fig. 7). This is an indication that the strength

of geopolymer mortars is also reliant on their pore structure.

A comparison of strength between pastes and mortars is presented in Fig.5. The

examination of this figure shows that the addition of sand has little influence on strength. In

12
 ACCEPTED MANUSCRIPT

terms of the relationship between the strength of geopolymer paste and mortar, it appears that

a stronger paste is always associated with a stronger mortar. This is in agreement with the

previous research [32]. This tendency suggests that the strength of geopolymeric composites

T
is mainly dependent on the pore structure of the matrix rather than the interface between

IP
inclusions and the matrix. Unlike OPC where one narrow region (known as interfacial

R
transition zone located) around the aggregate particles has on average fewer cement particles

SC
but more water, no such area exists in geopolymer mortar. This will be further discussed in

the next section.

NU
Morphology and EDS analysis of geopolymer mortars
MA
The SEM images taken on the polished RS and DS mortars are shown in Figs. 8.

Images (a) and (b) that were taken at the same magnification clearly showed the difference in
D

aggregate sizes between NS and DS. The microstructure of the interface between a river sand
TE

particle and geopolymer matrix is shown in Fig 8 (c). The area to the left is the sand particle
P

and the area to the right is the geopolymer matrix. The figure shows that sand particles are
CE

tightly bonded with the geopolymer gels. The interface between DS particles and the matrix

is shown in Fig. 8 (d). This is similar to the morphology presented in Fig. 8 (c). For both
AC

figures, the crystalline products were not observed in the interfacial area, as those exist

between sand particles and the cement matrix [33]. From the previous work [33], the

interfacial area contains excess calcium hydroxide and ettringite compared to the bulk paste.

The increased concentration of the crystalline products was most significant in the first 15-20

m closet to the aggregates. Compared to the bulk paste, this zone has the higher porosity

due to i) the participation of excess crystalline products and the presence of fewer cement

grains in the fresh state (due to wall effects of aggregates). Unlike the cement system, no such

13
 ACCEPTED MANUSCRIPT

a zone in geopolymers was observed. The geopolymer gel close to the sand surfaces

morphologically resembles the bulk gel. There is no apparent variation in gel density between

the interface and the bulk gel.

T
(a)

IP
Sand grain

R
Sand grain

SC
Sand grain

Air void
Air void

NU
100 m
MA
Fly Ash

Fly Ash
D

Sand grain
Sand grain
TE

Fig. 8 SEM image of geopolymer (a) mortars with RS (b) mortars with DS (c) interface
P

between RS and geopolymer matrix (d) interface between DS and geopolymer matrix
CE

Compositions of the samples were investigated by SEM-EDS analyses and results are

summarized in Table 3. Each data is averaged four results that were separately obtained at
AC

different locations. The locations within interfacial zone were chosen within 5 m distance

away from sand particles while the locations in bulk matrix were chosen at least 10 m away

from sand particles. All the locations of analyses were intentionally selected to avoid

unreactive ash particles within the matrix. From Table 3, it was observed that Si

concentration is increasing continuously from the bulk gel to the sand surfaces, in agreement

with other observations on geopolymeric systems [19]. Lee and van Deventer [19] proposed

that the formation of the Si-rich interfacial layer is a result of heterogeneous reactions that

14
 ACCEPTED MANUSCRIPT

accelerated dissolution of Si from the sand particles. Then the soluble Si ions precipitate in

the area around the sand particles.

Regardless of the sand type, the EDS result shows a strong correlation between the

T
cation type and Si/Al ratio. The K geopolymer mortar has highest Si/Al ratio, followed by the

IP
Na/K and Na geopolymer mortar. This could be explained by considering the role played by

R
the alkaline metals being incorporated to the system during geopolymerization. When Si4+

SC
atoms are replaced by Al3+, the alkaline metals compensate the electrical load of the

NU
microstructure. Compared to sodium, potassium is more readily associated with silicate and

aluminosilicate anions via cation-ion pairing, due to a lower charge density [34]. As a result,
MA
the amount of silicate in geopolymers increases with the K/Na ratio in the activating solution.

Table 3: Si/Al ratio of geopolymer mortars

D

Mortar Activator Si/Al

Interface Matrix
TE

DS K 2.13 0.25 1.907 0.11

Na/K 1.927 0.19 1.627 0.09
Na 1.713 0.2 1.591 0.18
P

RS K 2.03 0.13 1.814 0.12

Na/K 1.963 0.21 1.731 0.2
CE

Na 1.689 0.17 1.581 0.19

It is suggested that siliceous rocks are not chemically inert in geopolymer systems
AC

[35]. The structural breakdown and dissolution of silicate from sand particles have been

observed in the highly alkaline activating solutions [19]. Therefore, high Si concentration

observed in the current investigation at the interface could also be a result of dissolved Si

away from the sand particles. A detailed examination of Si/Al results shows that sand type

has influence on the ratio. This influence is affected by the cation type. In K and Na

geopolymer mortars, the average Si/Al ratio of DS mortars (at interface and matrix) is higher

than that of RS mortars. As discussed above, the fineness of DS is higher than that of RS. It is

proposed that the presence of larger surface area in DS may promote chemical interactions
15
 ACCEPTED MANUSCRIPT

between aggregate and solution, thereby accelerating the dissolution of Si from sand particles.

However, this tendency is not followed by Na/K geopolymer mortars. For these mortars, DS

specimens have lower Si/Al ratio as compared to RS specimens. The difference between the

T
Na, K and Na/K geopolymer mortars suggests that the alkali corrosion process is influenced

IP
not only by the physical properties (e.g., fineness) but also the chemical properties (e.g.,

R
mineral nature of sand). A considerable future research should be conducted to understand

SC
this issue.

NU
Effects of air voids on strength difference between DS and RS mortars

The sand type affects not only the chemical compositions but also strength of
MA
geopolymer mortars. A comparison of DS and RS mortars in Fig. 5 shows that the replacing

RS by DS decreases the strength in Na and Na/K geopolymer mortars. On the other hand, in
D

K geopolymer mortars, replacing RS by DS has little influence on strength. A comparison of

TE

Fig.6 and Fig.7 show that the relationship between strength and capillary pores cannot be
P

applied here to explain the strength differences due to sand type. For example, the sorptivity
CE

of Na and Na/K DS mortars is lower than that of Na and Na/K RS mortars. However, the

strength of Na and Na/K DS mortars is also lower than that of Na and Na/K RS mortars.
AC

a) b)

Fig. 9: Microscope images of mortars made with (a) river sand and (b) dune sand

16
 ACCEPTED MANUSCRIPT

It is noted that sorptivity test results are mainly attributed to the level of the

continuous pores. The strength level of the cementitious composites is dependent on the total

volume of pores including not only continuous pores but also closed pores [36]. The

T
absorptivity coefficient thus cannot be used to describe the relationship between the strength

IP
and pore structure concisely.

R
Air voids, which are closed pore, are produced in mixing, placing and handling. By

SC
occupying the space between unreacted grains, creating a more porous matrix (Fig. 9), and

NU
reducing the density of the specimens, air voids will reduce the strength of cementitious

materials. The results of the measured air voids content are presented in Table 4.
MA
Table 4: Air voids content calculated from images presented in Fig. 9
Activator K (%) Na/K (%) Na (%)
RS 2.39 2.16 2.80
D

DS 2.58 3.40 3.67

TE

It can be seen that the air void content of DS geopolymer mortars is higher than that

of RS geopolymer mortars. A similar tendency is also observed in DS OPC concrete [16].

P

Due to its higher surface area, DS particles have high amount of adsorbed layer water,
CE

leading to less free layer water in DS mixes compared to RS mixes. Free layer water helps
AC

facilitate the separation of solid particles from each other and thus contributes to workability

[37]. The DS mortar was less workable compared to the RS mortar due to the lack of free

layer water. In a less workable mix, more air voids were introduced as a result of placing,

compacting and handling. Besides the above processes, the majority of air voids are

introduced in the mixing. These voids originate from the air initially trapped among the dry

constituents, dissolved air in the water, and the air introduced as a result of the kneading and

folding action of mixing and trapped in the mix. However, these voids may be lost in placing

and handling. The retention of air voids is dependent on the surface tension at aqueous
17
 ACCEPTED MANUSCRIPT

interface which is significantly affected by the ions presented in the solution [38]. This may

explain the variation of air content in the mixes prepared by using different alkali cation.

However, the formation, retention and dispersion of air voids involve physical, chemical,

T
atmospheric and biological processes which are not well understood. The further research is

IP
required to understand this issue.

R
When DS replaced RS in the mortars, the relationship between the change in strength

SC
and that of air content is presented in Fig.10. Regarding air content, the growth in Na/K

NU
mortars is the fastest while the change in K mortars is the slowest. Regarding strength, the

reduction in Na/K mortars is the largest while the reduction in K mortars is the smallest. This
MA
correlation indicates that the reduction is due to higher amount of air voids trapped in the DS

mortar.
D

It was revealed from the results (Fig. 10) that the volume of air voids is the parameter
TE

affecting the strength of geopolymer mortars. On the other hand, the results (Table 4) show
P

that there is no correlation between the volume of air voids and the strength level of
CE

geopolymer mortars. This could be attributed to the fact that the air void (regarded as closed

pore) is only a small fraction of the total volume of pores (including both closed pores and
AC

continuous pores) in cementitious composites.

18
 ACCEPTED MANUSCRIPT

T
R IP
Fig. 10: Plot of changes in strength versus Fig. 11: Comparison between the measured
changes in air voids content when DS splitting tensile strength of geopolymer mortars

SC
replaces RS. and the predicted splitting tensile strength of
OPC concrete according to ACI 318-08

NU
Tensile strength

Fig. 11 presents the mean tensile strengths plotted against the square root of
MA
compressive strength. For a point of comparison the following empirical relationship from

ACI 318-08, fct=0.56(fc)1/2, was presented below the experimental mean tensile strengths.
D

Previous results also showed that the experimentally determined values of tensile strength
TE

remains above the required values predicted by ACI equation [39]. This means that for the
P

same compressive strength, the tensile splitting strength of geopolymeric materials satisfies
CE

the design equation given by the ACI standard. This tendency is independent on the type of

sand.
AC

Part of the reason for the ACI 318-08 curve to be lower than the experimental values

is that the code errs on the side of safety for designers. Another hypothesis for this is the good

bonding between the aggregates and the aluminosilicate binder. The dense and uniform

interface (presented in Figure 8 (c) and (d)) enhances the tensile strength of geopolymer

mortars

19
 ACCEPTED MANUSCRIPT

Modulus of elasticity

For OPC based materials, the modulus of elasticity is generally linked to its

compressive strength. This is also true in geopolymeric materials, as demonstrated in Fig. 12.

T
This linkage implies that the same main parameters influence both the compressive sterngth

IP
and the modulus of elasticity at the same time. According to Wongpa et al. [40], the modulus

R
of elasticity for geopolymer concrete may be taken as: E 5.39 f c 16.09 . The predictions

SC
by Wongpas equation were higher than the measured modulus of elasticity for all the

NU
samples. However, it is not reasonable to compare directly the modulus of elasticity present

in this study to the predictions, as the modulus of elasticity in their study was obtained with
MA
concrete. In general, density of concrete is higher by 30% as compared to that of mortar. The

low density is associated with low value of which is an factor of the type EC fC n -the
1

D
TE

general emperical relationship linking the modulus of elasticity to the compressive strength.

Therefore, for the same compressive strength, it is not surprising that the modulus of
P

elasticity of mortars is lower than that of concrete.

CE

When the compressive strengths of mortars are lower than 39 MPa, the rate of change
AC

for the modulus of elasticity with respect to compressive strength is similar to that presented

in the previous study [40]. However, once the strength is above 39 MPa, this rate change in

the current study and previous study are 5.39 and 3.51, respectively. This suggests that a

separate model for the modulus of elasticity should be developed for geopolymeric materials

with different strength level. For low strength geopolymer mortars, the weakest link exists in

the matrix which has a very porous structure. For high strength geopolymer mortars, the

structure of matrix becomes less porous and the sand particle could be the weakest link of

20
 ACCEPTED MANUSCRIPT

mortars. Due to this difference in the microstructure, it is seen that the empirical equation in

low strength mortar, could not work any more in high strength mortar. In the current study,

the equation predicting the modulus of elasticity from the compressive strength (higher than

T
39 MPa) is shown in the following: E 3.51 f c 7.33 . However, it is noted that further

IP
research is required to propose a general relationship between the modulus of elasticity and

R
the compressive strength, as the number of specimens tested in this study is limited.

SC
NU
MA
D
P TE
CE

Fig. 12 Relationship between modulus of elasticity and compressive strength of geopolymer

mortars
AC

CONCLUSION

Based on the results obtained in the current study, the following conclusions are drawn:

1. Test results show that H2O/Al2O3 ratio is the most dominant factor affecting

compressive strength of geopolymeric materials. This is due to that the higher

H2O/Al2O3 ratios lead to higher porosity.

2. The higher H2O/Al2O3 ratios result in increased pore volumes and threshold diameter

of pores, and thus higher water absorption ability.

21
 ACCEPTED MANUSCRIPT

3. Regardless of sand type, the compressive strength of geopolymer mortars shows a

strong correlation with its water sorptivity. This implies that the strength of mortars is

essentially dependent on the porosity of their corresponding geopolymer matrix.

T
4. The EDS results show the formation of a Si rich interfacial layer between the sand

IP
particle and the matrix, suggesting the chemical interactions between siliceous

R
aggregates and the highly alkaline solutions. During the interactions, Si could be

SC
dissolved from the aggregates initially and then precipitate at the interfacial layer after

a period of time.

NU
5. The cation type has an influence on Si/Al ratio of geopolymer mortars. The K
MA
geopolymer mortar has highest Si/Al ratio, followed by the Na/K and Na geopolymer

mortar. This is independent on the type of sand.

D

6. Compared with RS, the addition of DS increases Si/Al ratio in K and Na geopolymer
TE

mortars. This is proposed that the presence of larger surface area of dune sand may

promote chemical interactions between aggregate and solution, thereby accelerating

P

the dissolution of Si from the sand particles. However, this tendency is not followed
CE

by Na/K geopolymer mortars, suggesting that the alkali corrosion process is

AC

influenced not only by the physical properties (e.g., fineness) but also the chemical

properties (e.g., mineral nature) of sand particles.

7. The addition of DS increased the air void trapped in geopolymer mortars. The

retention of air voids is affected by the cation ions in the solution.

8. When RS is replaced by DS, the degree of decrease in strength is correlated to the

degree of increase in air void content. This suggests that the decrease in strength is

due to the increased air void in geopolymer DS mortars.

22
 ACCEPTED MANUSCRIPT

9. Regardless of sand type, the splitting tensile strength of geopolymer mortars was

shown to be higher than the predictions by the ACI code.

10. The square root of compressive strength linearly affects the modulus of elasticity of

T
geopolymer mortars. The slope of this relation decreases with strength level.

R IP
ACKNOWLEDGMENTS

SC
The authors are grateful for the financial support provided by the Australian Research

Council to conduct this study.

NU
MA
REFERNCES

1. Bagheri, A. and A. Nazari, Compressive strength of high strength class C fly ash-
based geopolymers with reactive granulated blast furnace slag aggregates designed
by Taguchi method. Materials and Design, 2014. 54: p. 483-490.
D

2. Jambunathan, N., et al., The role of alumina on performance of alkali-activated slag

TE

paste exposed to 50 C. Cement and Concrete Research, 2013. 54: p. 143-150.

3. Sarker, P.K., S. Kelly, and Z. Yao, Effect of fire exposure on cracking, spalling and
residual strength of fly ash geopolymer concrete. Materials and Design, 2014. 63: p.
584-592.
P

4. Bakharev, T., Resistance of geopolymer materials to acid attack. Cement and

CE

Concrete Research, 2005. 35(4): p. 658-670.

5. Garcia-Lodeiro, I., A. Palomo, and A. Fernandez-Jimenez, Alkali-aggregate reaction
in activated fly ash systems. Cement and Concrete Research, 2007. 37(2): p. 175-183.
AC

6. Zhang, Z., X. Yao, and H. Zhu, Potential application of geopolymers as protection

coatings for marine concrete I. Basic properties. Applied Clay Science, 2010. 49(1-2):
p. 1-6.
7. Pan, Z., et al., Damping and microstructure of fly ash-based geopolymers. Journal of
Materials Science, 2013. 48(8): p. 3128-3137.
8. Deb, P.S., P. Nath, and P.K. Sarker, The effects of ground granulated blast-furnace
slag blending with fly ash and activator content on the workability and strength
properties of geopolymer concrete cured at ambient temperature. Materials and
Design, 2014. 62: p. 32-39.
9. Nasvi, M.C.M., P.G. Ranjith, and J. Sanjayan, The permeability of geopolymer at
down-hole stress conditions: Application for carbon dioxide sequestration wells.
Applied Energy, 2013. 102: p. 1391-1398.
10. Van Jaarsveld, J.G.S., J.S.J. Van Deventer, and L. Lorenzen, Potential use of
geopolymeric materials to immobilize toxic metals: Part I. Theory and applications.
Minerals Engineering, 1997. 10(7): p. 659-669.

23
 ACCEPTED MANUSCRIPT

11. Shi, C. and A. Fernandez-Jimenez, Stabilization/solidification of hazardous and

radioactive wastes with alkali-activated cements. Journal of Hazardous Materials,
2006. 137(3): p. 1656-1663.
12. Feng, K.N., et al., Effect of strain rate on splitting tensile strength of geopolymer
concrete. Magazine of Concrete Research, 2014. 66(16): p. 825-835.
13. Laity, J.J., Deserts and desert environments. 2009: Wiley-Blackwell.

T
14. Wang, W.-H., et al., Behavior of concrete-filled steel tubular stub columns and beams
using dune sand as part of fine aggregate. Construction and Building Materials, 2014.

IP
51: p. 352-363.
15. Al-Harthy, A.S., et al., The properties of concrete made with fine dune sand.

R
Construction and Building Materials, 2007. 21(8): p. 1803-8.
16. Luo, F.J., et al., Effect of very fine particles on workability and strength of concrete

SC
made with dune sand. Construction and Building Materials, 2013. 47(0): p. 131-137.
17. Guettala, S. and B. Mezghiche, Compressive strength and hydration with age of
cement pastes containing dune sand powder. Construction and Building Materials,

NU
2011. 25(3): p. 1263-1269.
18. Tasong, W.A., J.C. Cripps, and C.J. Lynsdale, Aggregate-cement chemical
interactions. Cement and Concrete Research, 1998. 28(7): p. 1037-1048.
MA
19. Lee, W.K.W. and J.S.J. van Deventer, Chemical interactions between siliceous
aggregates and low-Ca alkali-activated cements. Cement and Concrete Research,
2007. 37(6): p. 844-855.
20. Vickers, L., W.D.A. Rickard, and A. Van Riessen, Strategies to control the high
D

temperature shrinkage of fly ash based geopolymers. Thermochimica Acta, 2014. 580:
p. 20-27.
TE

21. Chi, M. and R. Huang, Binding mechanism and properties of alkali-activated fly
ash/slag mortars. Construction and Building Materials, 2013. 40: p. 291-298.
22. Duxson, P., et al., Understanding the relationship between geopolymer composition,
P

microstructure and mechanical properties. Colloids and Surfaces A: Physicochemical

and Engineering Aspects, 2005. 269(1-3): p. 47-58.
CE

23. Duxson, P., et al., The effect of alkali and Si/Al ratio on the development of
mechanical properties of metakaolin-based geopolymers. Colloids and Surfaces A
(Physicochemical and Engineering Aspects), 2007. 292(1): p. 8-20.
AC

24. Hardjito, D., et al., On the development of fly ash-based geopolymer concrete. ACI
Materials Journal, 2004. 101(6): p. 467-472.
25. Okada, K., et al., Water retention properties of porous geopolymers for use in cooling
applications. Journal of the European Ceramic Society, 2009. 29(10): p. 1917-23.
26. Dombrowski, K., A. Buchwald, and M. Weil, The influence of calcium content on the
structure and thermal performance of fly ash based geopolymers. Journal of Materials
Science, 2007. 42(9): p. 3033-3043.
27. Lizcano, M., et al., Effects of water content and chemical composition on structural
properties of alkaline activated metakaolin-based geopolymers. Journal of the
American Ceramic Society, 2012. 95(7): p. 2169-2177.
28. Sindhunata, et al., Effect of curing temperature and silicate concentration on fly-ash-
based geopolymerization. Industrial and Engineering Chemistry Research, 2006.
45(10): p. 3559-3568.

24
 ACCEPTED MANUSCRIPT

29. Taylor, H., Discussion of the paper Microstructure and strength of hydrated cements
by RF Feldman and JJ Beaudoin. Cement and Concrete Research, 1977. 7(4): p. 465-
468.
30. Aligizaki, K.K., Pore structure of cement-based materials: testing, interpretation and
requirements. 2005: CRC Press.
31. Sahmaran, M. and V.C. Li, Influence of microcracking on water absorption and

T
sorptivity of ECC. Materials and Structures/Materiaux et Constructions, 2009. 42(5):
p. 593-603.

IP
32. Lee, W.K.W. and J.S.J. Van Deventer, The interface between natural siliceous
aggregates and geopolymers. Cement and Concrete Research, 2004. 34(2): p. 195-

R
206.
33. Scrivener, K.L., A.K. Crumbie, and P. Laugesen, The interfacial transition zone (ITZ)

SC
between cement paste and aggregate in concrete. Interface Science, 2004. 12(4): p.
411-421.
34. Kinrade, S.D. and D.L. Pole, Effect of alkali-metal cations on the chemistry of

NU
aqueous silicate solutions. Inorganic Chemistry, 1992. 31(22): p. 4558-4563.
35. Brough, A.R. and A. Atkinson, Sodium silicate-based, alkali-activated slag mortars -
Part I. Strength, hydration and microstructure. Cement and Concrete Research, 2002.
MA
32(6): p. 865-879.
36. Kumar, R. and B. Bhattacharjee, Porosity, pore size distribution and in situ strength
of concrete. Cement and Concrete Research, 2003. 33(1): p. 155-164.
37. Wang, A., C. Zhang, and W. Sun, Fly ash effects: I. The morphological effect of fly
D

ash. Cement and Concrete Research, 2003. 33(12): p. 2023-2029.

38. Jungwirth, P. and D.J. Tobias, Specific ion effects at the air/water interface. Chemical
TE

Reviews, 2006. 106(4): p. 1259-1281.

39. Pan, Z., J.G. Sanjayan, and B.V. Rangan, Fracture properties of geopolymer paste
and concrete. Magazine of Concrete Research, 2011. 63(10): p. 763-771.
P

40. Wongpa, J., et al., Compressive strength, modulus of elasticity, and water
permeability of inorganic polymer concrete. Materials & Design, 2010. 31(10): p.
CE

4748-4754.
AC

25
 ACCEPTED MANUSCRIPT

50
45
40
Elastic modulus (GPa)

35

T
High strength line
30

IP
y = 3.51 - 7.33
25

R
20

SC
15
10 High strength

NU
Low strength line
5 Low strength
y = 5.52 - 19.77
0
MA
4 5 6 7 8 9 10
fc (MPa1/2)
D

Graphical abstract
P TE
CE
AC

26
 ACCEPTED MANUSCRIPT

Highlights

Mechanical properties of dune sand and river sand geopolymer mortar are comparable.

The geopolymer gel around dune sand surfaces morphologically resembles the bulk gel.

T
IP
Empirical formula for elastic modulus changes as strength of mortars is increased.

R
The cation type affects properties of geopolymer mortars.

SC
The strength of geopolymeric materials is governed by the H2O/Al2O3 ratio.

NU
MA
D
P TE
CE
AC

27