Construction and Building Materials 52 (2014) 473–481

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Synthesis of geopolymer from large amounts of treated palm oil fuel ash:Application of the Taguchi method in investigating the main parametersaffecting compressive strength

0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.11.039

⇑ Corresponding author. Tel.: +60 4 5996208; fax: +60 4 5941009.E-mail addresses: eng.mustafajuma@gmail.com (M.J.A. Mijarsh), cemamj@

eng.usm.my (M.A. Megat Johari), zainal@eng.usm.my (Z.A. Ahmad).

M.J.A. Mijarsh a,b, M.A. Megat Johari a,⇑, Z.A. Ahmad c

a School of Civil Engineering, Universiti Sains Malaysia, Malaysiab Civil Engineering Department, Faculty of Engineering, Al-Merghab University, Al-Khums, Libyac School of Materials and Mineral Resources Engineering, Universiti Sains Malaysia, Malaysia

h i g h l i g h t s

� Geopolymer was synthesized from high content of treated palm oil fuel ash.� TPOFA-based geopolymer mortar gained compressive strength of 47 MPa at 7 days.� The optimized TPOFA-geopolymer mortar mixture was analysed using XRD and FTIR.� The main binding phases consist of N–A–S–H and C–S–H gels.

a r t i c l e i n f o

Article history:Received 23 August 2013Received in revised form 4 November 2013Accepted 12 November 2013Available online 12 December 2013

Keywords:GeopolymerPalm oil fuel ashCompressive strengthTaguchi method

a b s t r a c t

The aim of this study was to synthesize geopolymers using a large amount of treated palm oil fuel ash(TPOFA). The efficiency of the TPOFA (as the source material) in producing geopolymer products wasenhanced via six factors which were optimized using the Taguchi method L25. The six factors weredivided into two different components: (i) additive materials i.e. Ca(OH)2, silica fume (SF), Al2(OH)3,and (ii) alkaline activators; i.e. NaOH concentration (moles), Na-silicate: NaOH ratio, and alkali-activa-tor:solid-material ratio. Each of these factors was examined on five levels in order to obtain the optimummixture. A total of 25 mixtures were prepared in accordance to the L25 array proposed by the method.The performance of the specimens was evaluated by compressive strength tests. The results show thatthe optimum mixture consisted of 65 wt.% TPOFA and 35 wt.% additive materials which achieved a com-pressive strength of 47.27 ± 5.0 MPa after 7 days of curing. The properties of the optimized mixture werefurther analyzed via X-ray diffractography (XRD) and Fourier transform infrared spectroscopy (FTIR)analyses. The results show that the main binding phases consist of aluminosilicate type gel ‘‘N–A–S–H’’ (Na2O–Al2O3–SiO2–H2O) and calcium silicate hydrate (C–S–H) gels, formed simultaneously, withinthe TPOFA-based geopolymer mortar.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

A geopolymer is also known as an alkali-activated aluminosili-cate material. A geopolymer binder has a three-dimensional amor-phous structure of (N–A–S–H) type gel (Na2O–Al2O3–SiO2–H2O).The gel is formed by either the polymerization of individual[SiO4]4� and [AlO4]5� species as building blocks in the systemwhich are the main reaction product of alkali-activated aluminosil-icate (geopolymer) materials derived from low-calcium content[1,2] or a calcium alumina silicate hydrate C–(A)–S–H type gel

(CaO–Al2O3–SiO2–H2O) which is the main binding phase in the sys-tem of alkali-activated aluminosilicate calcium-rich source mate-rial [3,4]. Therefore, geopolymer synthesis is dependent on theuse of material rich in aluminum–silicate glass activated with alka-line solutions [5]. The most commonly used material is calcinedkaolin or metakaolin (MK) [2,6]. However, industrial wastes suchas fly ash and ground granulated blast furnace slag (GGBFS) arealso used due to their high silica and alumina content with littleand high CaO content, respectively [1,3]. In the major palm oil pro-ducing countries, palm oil fuel ash (POFA) is another abundantlyavailable waste material which can be used for geopolymer syn-theses. POFA [7] has high SiO2 (61.33) content, low Al2O3 (7.018)content, and high P2O5 (4.55) content compared to class C and classF fly ash, GGBFS, and MK or calcined kaolin [1–4,6]. It also has

Table 1Chemical compositions using XRF technique and physical properties of TPOFA.

Chemical Component (%)

SiO2 61.33Al2O3 7.018Fe2O3 5.11CaO 8.20MgO 4.69P2O5 4.55K2O 6.50SO3 0.27TiO2 0.25MnO 0.097Na2O 0.123C 1

Physical propertiesSpecific surface area (m2/g) 1.775Loss on ignition (%) 2.53Median particle size d50 (lm) 2.06

474 M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481

higher Fe2O3 (5.11) content compared to GGBFS, and MK or cal-cined kolin [2,3,6], and higher MgO (4.69) and CaO (8.20) contentcompared to class F fly ash and MK [1,2]. Theoretically, thereshould be a direct correlation between the geopolymerization re-sults and the total silica content of the source materials. Increasingthe amount of silica causes the number of Si–O–Si bonds which arestronger than Si–O–Al and Al–O–Al bonds [8], to increase; this im-plies that the strength of geopolymers would also increase with theSi/Al ratio because the density of the Si–O–Si bonds increases withthe increase in Si/Al ratio [9,10]. Hence, the high silica content ofthe POFA reflects its feasibility as source material for geopolymersynthesis.

Untreated POFA has been used in combination with pulverizedfuel ash (PFA) at a ratio of 30:70, and activated with an alkalineactivator to produce geopolymer with a compressive strength of30 MPa after 28 days of curing [11]. This rather limited strengthpotential of the resulting geopolymer could be attributed to thehigh content of unburned carbon in POFA, as unburned carbon ofbase material has been found to affect compressive strength ofgeopolymer [12]. Therefore, POFA has to be treated to reduce thecarbon content in order to promote the geopolymerization process.The potential of POFA, especially treated POFA (TPOFA) [7], as asource material for geopolymer synthesis has not been adequatelyexplored as in the case of class C and class F fly ashes, GGBFS, andMK or calcined kaolin [1–4,6]. Nonetheless, this is changing asgreat interest in POFA has sparked among researchers, especiallythose from palm oil producing countries.

To the authors’ best knowledge, no geopolymerization of TPOFAhas ever been reported, possibly due to its efficiency as it containsa high amount of silica, but a low amount of alumina, as well ascalcium which need to be initially enhanced. Recent researchworks have shown that the inclusion of specific amounts of cal-cium hydroxide (Ca(OH)2) [13], aluminum hydroxide (Al(OH)3)[14], and silica fume (SF) [15] can enhance the geopolymerizationprocesses. In all cases, they were done individually in differentaluminosilicate source materials. A significant amount of thesematerials showed considerable impact on the geopolymerizationprocesses. Therefore, it is important to introduce TPOFAs contain-ing different amounts of these materials in order to investigatethe enhancement of TPOFA for efficient geopolymer synthesis.

Another crucial parameter that affects the TPOFA geopolymersynthesis is alkali activator types (combination and concentration).Several works have been carried out on geopolymers, yet no clearexplanation has been provided on the effects of various factors forupsetting source materials, such as how the strength increasedwith high NaOH concentrations [16,17]. However, some otherresearchers [18,19] found that high concentrations of NaOH havea negative effect on strength. Moreover, previous research has con-cluded that the use of Na2SiO3/NaOH in a weight ratio of 1.0 givesstrength of up to 70 MPa [20]. The highest compressive strengthvalue of 71 MPa was observed at 2.5 Na2SiO3/NaOH weight ratio[21].

Investigating all of these parameters (Ca(OH)2 wt.%, SF wt.%,Al(OH)3 wt.%, NaOH concentration (mole), Na2SiO3/NaOH (weightratio), and alkali-activator/solid materials (weight ratio)) in a sin-gle work may not be possible. However, with a suitable designmethod, one may consider some of the factors affecting these prop-erties. The Taguchi experimental design is one of the most famousmethods used to design the parameters of specific problems. Theapplication of the Taguchi method in geopolymers by pastresearchers for similar purposes [22–24], but with different sourcematerials, has been successful.

This study is an investigation on the synthesis and characteriza-tion of geopolymers produced using TPOFA as the source material.Six design factors were examined at five levels by the Taguchimethod to obtain the optimum mixture. The six factors are

Ca(OH)2 wt.%, SF wt.%, Al(OH)3 wt.%, NaOH concentration (mole),Na2SiO3/NaOH (weight ratio), and alkali-activator/solid materials(weight ratio). A total of 25 experiments were conducted accordingto L25 array proposed by the Taguchi method. The efficiency of theTPOFA was enhanced by investigating the effects of the abovemen-tioned parameters on the mechanical properties of the product inorder to obtain the optimum mixture. This mixture was thenexamined using X-ray diffraction (XRD) and Fourier transforminfrared spectroscopy (FTIR).

2. Materials and methods

2.1. Materials

Four basic raw materials were used to produce the geopolymer mortar. RawPOFA was collected from a nearby palm oil mill in Nibong Tebal, Penang, Malaysia.The incompletely combusted fibers and kernel shells were separated using a300 lm sieve. The POFA was then ground in a ball mill to obtain particle sizes ofabout 10 lm. To remove the unburned carbon, POFA was heated at 500 �C for 1 hin a gas furnace and then the POFA was subjected to second stage grinding to formthe TPOFA. The chemical and physical properties of the TPOFA in this study are pro-vided in Table 1. Based on ASTM C618 [25], the treated POFA could be classified as aclass F mineral admixture. The same approach has recently been used and was re-ported to be effective in increasing the efficiency of TPOFA for use in high-strengthgreen concrete [7] and engineered cementitious composites [26]. The TPOFA is themajor source material in this research. The alkaline activators used were solutionscombining NaOH and Na2SiO3. The analytical grade NaOH was in pellet form with98% purity and the commercial Na2SiO3 was in liquid form with a specific gravityof 1.53 g/cm3 at 20 �C and a silica modulus (Ms, where Ms = SiO2/Na2O) of 2;14.7% Na2O, 29.4% SiO2, and 55.9% H2O were combined. The additives used wereCa(OH)2, Al(OH)3, and SF. The surface area of Ca(OH)2, Al(OH)3 and SF were0.6214 m2/g, 0.3298 m2/g, and 0.1364 m2/g, respectively. Clean river sand passinga 1.18 mm sieve and retained on a 150 lm sieve, with a fineness modulus of 2.8and specific gravity of 2.65, was used.

2.2. Design of the mixtures

In this study, the Taguchi method was used to design the mixtures and to obtainthe optimum mixture design by considering the effects of the parameters on theirmechanical properties. Six primary factors were examined followed by a statisticalstudy of the Ca(OH)2 wt.% (designated as A); the SF wt.% ‘‘B’’; the Al(OH)3 wt.% ‘‘C’’;the NaOH concentration (moles) ‘‘D’’; the Na2SiO3 to NaOH weight ratio ‘‘E’’; and theliquid alkaline to solid material weight ratio ‘‘F’’. Each factor was examined at thefive levels as described in Table 2. The level of each factor and the values of thetested factors were chosen based on previous researches [14,27–29]. The amountof water added to each mixture was calculated based on the percentage of geopoly-mer paste at 5% by weight ratio [30]. All of the geopolymer mortars were made withsand to solid material weight ratios of 1.5 which is similar to that used in previousresearches [31,32]. This is the optimum ratio for the binder (solid material) andsand; anything beyond this ratio will cause the compressive strength to be dramat-ically reduced [33]. The design suggested by the Taguchi method for six factors atfive levels is the L25 array, as shown in Table 3, while Table 4 shows the valuesand trial mixture proportions used in the series of 25 mixtures. The compressive

Table 2The introduced levels for each factor in Taguchi experimental design.

Factor Unit Level1

Level2

Level3

Level4

Level5

A: Ca(OH)2 wt.% 15 17.5 20 22.5 25B: Silica fume wt.% 2.5 3.75 5 6.25 7.5C: Al(OH)3 wt.% 5 6.25 7.5 8.75 10D: NaOH concentration Mole 10 12 14 16 24E: Na2SiO3-to-NaOH wt.

ratio0.67 1 1.5 2 2.5

F: Alkali-activator-to-solid material

wt.ratio

0.46 0.47 0.48 0.49 0.50

Table 3Taguchi method of orthogonal arrays [L25 (5 * 6)] of the experimental design.

Trial Factor A Factor B Factor C Factor D Factor E Factor F

T1 2 1 2 3 4 5T2 1 4 4 4 4 4T3 3 5 2 4 1 3T4 1 1 1 1 1 1T5 3 4 1 3 5 2T6 4 3 1 4 2 5T7 2 2 3 4 5 1T8 3 2 4 1 3 5T9 2 3 4 5 1 2T10 3 1 3 5 2 4T11 2 5 1 2 3 4T12 4 4 2 5 3 1T13 5 3 2 1 5 4T14 1 5 5 5 5 5T15 5 5 4 3 2 1T16 5 1 5 4 3 2T17 1 3 3 3 3 3T18 5 2 1 5 4 3T19 1 2 2 2 2 2T20 4 1 4 2 5 3T21 4 5 3 1 4 2T22 5 4 3 2 1 5T23 2 4 5 1 2 3T24 3 3 5 2 4 1T25 4 2 5 3 1 4

Table 4Mix proportions of geopolymer used for Taguchi optimization.

Mix POFA Ca(OH)2 Silicafume

Al(OH)3 Sand Na2SiO3 NaOH Addedwater

(g) (g) (g) (g) (g) (g) (g) (g)

T1 500 120 20 40 1030 240 120 100T2 490 100 40 60 1040 230 120 100T3 470 140 50 40 1080 130 200 100T4 540 100 20 30 1050 150 220 100T5 480 140 40 30 1040 250 100 100T6 480 160 40 40 1060 170 170 100T7 500 120 30 50 1050 240 100 100T8 470 140 30 60 1050 200 140 100T9 490 120 40 60 1070 140 200 100T10 490 140 20 50 1050 170 170 100T11 490 120 50 30 1050 210 140 100T12 470 160 40 40 1070 190 130 100T13 450 180 40 40 1060 230 90 100T14 470 100 50 70 1050 240 100 100T15 430 180 60 60 1100 150 150 100T16 450 180 20 70 1080 190 120 100T17 500 100 30 50 1040 210 140 100T18 470 180 30 40 1060 220 110 100T19 520 100 30 40 1040 180 180 100T20 470 160 20 60 1060 230 90 100T21 450 160 50 50 1080 210 100 100T22 440 180 50 50 1090 130 190 100T23 470 120 40 70 1070 170 170 100T24 460 140 40 70 1070 210 110 100T25 460 160 30 70 1080 130 200 100

Table 5Changes of compressive strength of trial mixes.

Trial mix Combination Compressive strength (MPa)

Response 1 Response 2 Response 3

1 day 3 days 7 days

T1 A2B1C2D3E4F5 20.65 21.98 23.07T2 A1B4C4D4E4F4 18.6 19.8 20.855T3 A3B5C2D4E1F3 15.67 16.69 18.42T4 A1B1V1D1E1F1 18.59 19.79 22.79T5 A3B4C1D3E5F2 30.39 31.31 33.63T6 A4B3C1D4E2F5 27.74 29.33 30.52T7 A2B2C3D4E5F1 30.65 32.71 33.454T8 A3B2C4D1E3F5 27.94 30.19 31.19T9 A2B3C4D5E1F2 24.66 25.1 26.25T10 A3B1C3D5E2F4 17.81 18.97 20.185T11 A2B5C1D2E3F4 18.69 19.9 20.185T12 A4B4C2D5E3F1 32.73 33.72 35.855T13 A5B3C2D1E5F4 37.77 39.97 42.15T14 A1B5C5D5E5F5 18.91 19.8 20.915T15 A5B5C4D3E2F1 27.71 29.2 31.94T16 A5B1C5D4E3F2 36.95 38.26 39.77T17 A1B3C3D3E3F3 26.02 28.27 29.55T18 A5B2C1D5E4F3 30.65 31.98 34.91T19 A1B2C2D2E2F2 19.09 20.85 23.43T20 A4B1C4D2E5F3 30.21 33.01 35.02T21 A4B5C3D1E4F2 31.08 33.57 35.63T22 A5B4C3D2E1F5 24.38 26.29 28.06T23 A2B4C5D1E2F3 26.65 27.77 30.55T24 A3B3C5D2E4F1 41.82 42.62 44.74T25 A4B2C5D3E1F4 27.26 28.7 31.445

M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481 475

strength values of the trial mixtures results were evaluated by calculating a re-sponse index for each factor based on signal-to-noise ratio (S/N) principles [34],i.e. a higher S/N provides a ‘better’ response index.

2.3. Geopolymer synthesis

The synthesis of geopolymer mortar from TPOFA was carried out using mixturesof solid materials and an alkaline activator. The solid materials were TPOFA, thesource material, which was substituted with various ratios of Ca(OH) 2, SF, andAl(OH)3. The alkaline activator was made from Na2SiO3 and NaOH solutions. TheNaOH pellets were poured into the beaker with distilled water to obtain concentra-tions of 10, 12, 14, 16, and 24 M. The solution was stirred until the NaOH pellets dis-solved and the solution became clear. During this process, a significant amount ofheat can be released. To ensure that the heat did not interfere with the geopolymerreaction, the solution was covered and sealed for at least 3 h which allowed thesolution to cool down to ambient temperature (28 ± 2 �C, 70% RH). The Na2SiO3

solution was used without further preparation and poured into the beaker withthe aforementioned NaOH solution and mixed for 2 min. Water was added to thesolid material and alkaline activator prior to the slow addition of sand, as seen inTable 4. Immediately after mixing, the geopolymer mortar was cast into 50 � 50� 50 mm steel molds in two layers so that it becomes compact, as described inthe ASTMC109M [35]. This was followed by an additional vibration of 15 s usinga vibrating table. The specimens were wrapped with cling film to avoid moistureevaporation, left in the laboratory for 1 h and then cured at 75 �C for 48 h to activatethe geopolymerization [36,37]. The specimens were left in the laboratory to cooldown. After that, the specimens were demoulded and kept under ambient temper-ature (28 ± 2 �C, 70% RH) for 1, 3, and 7 days of curing time.

2.4. Specimen analysis

The specimens were tested for compressive strength according to ASTM C109M[35]. The test was performed on three specimens from each mixture. The optimummixture was also analyzed by X-ray diffraction (XRD) and Fourier transforms

infrared spectroscopy (FTIR). XRD was used to determine the phases present inthe specimens and FTIR was used to identify the different types of chemical bondspresent in the materials on a molecular level. Analysis was carried out using an XRDmachine (Bruker D8 Advance) with Cu Ka radiation (1.5406 Å) from 10 to 90 of 2h�.Diffraction patterns were then analyzed using Expert HighScorePlus software. ForFTIR, the equipment used was Perkin–Elmer Spectrum One. Approximately 5 mgof powdered specimen was mixed with 95 mg of potassium bromide (KBr) andground using an agate mortar and pestle. The powder was pressed for 2 min under2.758 MPa using hydraulic pressing to prepare the translucent pellets which werethen inserted into the infrared spectrometer. The infrared spectra were recordedin the range of 400–4000 cm�1 wavelength.

0

10

20

30

40

1 day

3 days

7 days

A1A2

A3A4

A5

Com

pres

sive

Str

engt

h (M

Pa)

Curin

g ag

e (da

y)

Calcium Hydroxide (wt.%)

(a)

0

10

20

30

40

1 day

3 days

7 days

B1B2

B3B4

B5

Com

pres

sive

Str

engt

h (M

Pa)

Curin

g ag

e (da

y)

Silica Fume (wt.%)

0

10

20

30

40

1 day

3 days

7 days

C1C2

C3C4C5C

ompr

essi

ve S

tren

gth

(MP

a)

Curin

g ag

e (da

y)

Aluminum Hydroxide (Wt.%)

(c)

0

5

10

15

20

25

30

35

1 day

3 days

7 days

D1D2

D3D4

D5

Com

pres

sive

Str

engt

h (M

Pa)

Curin

g ag

e (da

y)

NaOH Concentration (Mole)

(d)

0

5

10

15

20

25

30

35

1 day

3 days

7 days

E1E2

E3E4

E5

Com

pres

sive

Str

engt

h (M

Pa)

Curin

g ag

e (da

y)

Na2SiO

3-to-NaOH (wt.ratio)

(e)

0

10

20

30

40

1 day

3 days

7 days

F1F2

F3F4

F5

Com

pres

sive

Str

engt

h (M

Pa)

Curin

g ag

e (da

y)

Alkali-Activator-to-solid material (wt.ratio)

(f)

(b)

Fig. 1. Effect of (a) Ca(OH)2 wt.%, (b) SF wt.%, (c) Al(OH)3 wt.%, (d) NaOH concentration (in terms of molar), (e) Na2SiO3-to-NaOH weight ratio, (f) alkali-activator-to-solidmaterial weight ratio on each response of compressive strength at different curing ages.

476 M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481

3. Results and discussion

Table 5 shows the results of the compressive strength testsconducted on the 25 trial mixtures prepared as suggested by theTaguchi method. Based on these results, the evaluation of compres-sive strength (response index) for each factor was calculated by

averaging the strengths at 1, 3, 7 days for each trial mixturecontaining a particular factor. For example, the response indexfor factor A1 was tested on trial mixtures marked with factor A1,as shown in Table 5 (i.e. T2, T4, T14, T17, and T19). Therefore,the response index for factor A1 at 1 day after curing was the aver-age of the response index for trial mixtures of T2, T4, T14, T17, and

Table 6Optimization of the factors combination of TPOFA-based geopolymer mortar mixtureat different curing ages.

Optimum mixture Combination Compressive strength (MPa)

1 day 3 days 7 days

1 A3B3C5D1E5 F1 42.64 45.55 47.27

M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481 477

T19. Similarly, the calculations were done for factors A1, A2, A3, A4,and A5 at 1, 3, and 7 days, respectively. The results have been plot-ted in Fig. 1a. The plot indicates that the Factor A5 gave the highestresponse index; therefore, it is the optimum value for factor A.Optimum values for factors B (Fig. 1b), C (Fig. 1c), D (Fig. 1d), E(Fig. 1e) and F (Fig. 1f) were calculated in the same way. A detaileddiscussion is given in the following subsections.

3.1. Effect of additive materials

3.1.1. Effect of Ca(OH)2 content (Factor A)The response index for Factor A is shown in Fig. 1a. The re-

sponse index increased with the increase in Factor A amount. Thereis an obvious positive influence on the response index when theFactor A is above A3. However, in the actual experiment, specimensproduced using A3 onwards exhibited very rapid setting (less than3 min) and poor workability due to higher dissolution of Factor A.Therefore, replacing TPOFA with Factor A above A3 in geopolymersynthesis is not useful.

Factor A plays a significant role in improving the strength of thehardened geopolymer mortar due to the pozzolanic reaction. Posi-tive ions such as Ca2+ need to be present in the framework cavitiesto balance the negative charge of the aluminate group. However, itis still not clear why Ca(OH)2 plays such a significant role in thestrength of alkali-activated binders [13,15]. In particular, the addi-tion of highly alkaline activating solutions into TPOFA containingCa(OH)2 has the net effect of lowering the pH of the activatingsolution. This is due to the removal of the OH� ions which signifi-cantly affects the rate of further dissolution/precipitation processesto form C–S–H and geopolymeric gel at low alkalinity [38]. Hence,the unreacted Ca(OH)2 is slowly recarbonated into calcite overtime, while free OH� hydroxyl ions that could be involved in thecarbonation of Ca(OH)2 into CaCO3 move in the other direction.This carbonation principle contributes to the growth of geopoly-mer strength. It has been noted that the strength of geopolymersthat contain reasonable quantities of calcite had improved withtime on account of the interaction between calcite and geopoly-meric phase. In addition, it also functions as a physical filler/mi-cro-aggregate [39]. Other researchers have found that loss ofstrength occurs when Ca(OH)2 is above 10% at 14 days of curing;however, the source material of the geopolymer was tungstenmine waste mud, not POFA or FA. This relates to the fact that theC–S–H and geopolymeric reactions compete against each otherfor soluble silicates, resulting in a binder composed of two poorphases that leads to loss of strength [27].

3.1.2. Effect of SF content (Factor B)The response index for Factor B is shown in Fig. 1b. The re-

sponse index increased with the increase of Factor B up to B3and started to reduce once it was above B3. It was expected thatadding Factor B to the TPOFA-based geopolymer mortar mixtureswould increase the response index. However, this increase in re-sponse index is not more than the efficacy of TPOFA containing61.33% SiO2 that would chemically react with Factor A to formC–S–H or with Factor C and alkaline activation to form (N–A–S–H) gels. Nevertheless, in actual experiments, the workability ofspecimens fabricated with amounts above B3 started to reduce

due to the finer sized particles of SF (Factor B) which hinders theprocess of geopolymerization [28]. Therefore, adding Factor B upto B3 crammed the voids existing between the sand particles andcontributed to the improvement of porosity, producing densermicrostructure [15]. The decrease in response index with additionsbeyond B3 is also linked to the amount of unreacted material in thespecimens [40].

3.1.3. Effect of Al(OH)3 content (Factor C)The response index for Factor C is shown in Fig. 1c. The response

index increased with the increase of Factor C up to C5. It was alsofound in the actual experiments that the specimens containingFactor C close to C5 produced higher workability. The results indi-cated that Factor C played an important role in gel structure devel-opment in the early stages of the reaction during the initialstrength development of the geopolymer by promoting the geo-polymeric reaction and strengthening the matrix. These resultsgenerally agree with the experimental results obtained in an ear-lier study [14].

3.2. Effects of alkaline activator combinations and concentrations

3.2.1. Effects of sodium hydroxide concentration (Factor D)The response index for Factor D is shown in Fig. 1d. The re-

sponse index reduced with the increase of Factor D up to D5. Inthe actual experiment, specimens fabricated with more Factor Dhad reduced workability and higher porosity. This result is attrib-uted to the high number of OH� groups that inhibit the geopoly-merization and C–S–H reaction which produces the gel binderand increase the total porosity to directly influence the responseindex. In addition, Na+ changed the balance requirements withinthe structure by satisfying the sodium content, thus leading tohigher strengths. In contrast, with increased sodium content, so-dium carbonate starts to form. This carbonate compound willcompete with the geopolymerization process, producing poorermechanical properties, thereby reducing the response index[18,19]. However, previous researchers found that, when using dif-ferent source materials, the increase of NaOH concentrations eitherincreased [16,17] or reduced [18,19] the strength of the specimens.

The typical geopolymer composition is generally expressed asnM2O�Al2O3�xSiO2�yH2O, where M is an alkali metal [32]. The mostwidely adopted alkaline activators are MOH-type caustic alkalisand R2O�(n)SiO2 type silicates which are used individually or incombination, as exhibited in previous researches [16–19]. Here,M denotes the alkaline activator, which is generally NaOH, Na2-

SiO3, or Ca(OH)2 containing alkaline metal ions such as Na+ andalkaline earth metal such as Ca2+ serving as an accelerator of thereaction speed by activating Al and Si through a reaction withthe binder. In previous researches, the alkaline Na2SiO3 was usedwithout specifying its Ms (silica modulus) SiO2/Na2O ratio in theirsystem [17,19], as shown in Eq. (1); however, the modulus wasspecified in other system [16]. The SiO2/M2O ratio in an alkalinesilicate solution affects the degree of polymerization of the dis-solved species [41]. This presents some difficulties when compar-ing results and findings in previous researches. The SiO2/M2Omolar ratio (M = Na+) of the activating solution (NaOH and Na2-

SiO3) is the most critical factor for the synthesis of geopolymers[42]. Previous researches used NaOH without Na2SiO3 or/andCa(OH)2 in their systems [18,43,44]. It is important to understandthe roles of Ca(OH)2, NaOH, Na2SiO3 in optimizing the properties ofuser-friendly geopolymer systems.

There should be strong correlation between the total solid SiO2/Na2O ratio and water content H2O/Na2O ratio in alkaline activatorand they must be taken into account [32]. In this research, the totalwater content H2O/Na2O resulted from the contribution of watercontent from the alkaline activator (NaOH and Na2SiO3) as well

10 20 30 40 50 60

G Q

Q

H J

Q Q C Q Q

G H J Q

G G P P P P G

P

P A A A

Q Q Q Q

k Cr Q Q Q Q Q

Q

Q

C

Q H J

G A G

Cr k Cr (a)

(b)

(c)

A-Albite, C-Calcite, Cr-Cristobalite, G-Gibbsite, H-Hillebrandite, J-Jadeite, K-[ K3Al2(PO4)3], P- Portlandite, Q-Quartaz Low.

k

Fig. 2. X-ray diffractograms of (a) TPOFA, (b) TPOFA substituted with optimumamount of (Ca(OH)2, Al(OH)3 and SF) and (c) TPOFA-based geopolymer mortar.

478 M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481

as free water from the decomposition of (Al(OH)3, Ca(OH)2, andNaOH) as shown in Eqs. (2)-(4) and the amount of added water.The effects of alkaline activator to solid weight ratio is explainedin the following subsection.

Na2O � nSiO2 ðaqÞ ! Na2Oþ nSiO2 þH2O ð1Þ

2AlðOHÞ3 ! Al2O3 þ 3H2O ð2Þ

CaðOHÞ2 ! CaOþH2O ð3Þ

2NaOH! Na2OþH2O ð4Þ

Nevertheless, the formation of gel is dependent on the concen-tration of alkaline OH� group to initiate the dissolution of amor-phous phases (e.g. aluminosilicate source) at the initial stage ofthe geopolymerization reaction [45]. The dissolution process inalkaline medium with the inclusion of Al(OH)3 and Ca(OH)2 formedmore OH� than NaOH alone. The OH� attacks the covalent bond(oxygen) of source materials, implying that Si–O–Si, Al–O–Al andSi–O–Al bonds are broken, yielding Al–OH and Si–OH groups [46].

As discussed in the previous section, the influence of Ca(OH)2 asan activator increases the response index while NaOH reduces it.This observation could be correlated to previous work which con-cluded that when NaOH concentration is 10 M or above, dissolu-tion of Ca(OH)2 is very difficult due to the presence of hydroxide[43,44]. This means that there will not be enough for the formationof C–S–H gel. Instead, sodium-based aluminosilicate is formed,attracting OH� to its structure, lowering the total amount ofhydroxide groups and allowing the formation of C–S–H gel as asecondary reaction product [13]. Previous work [22] indicates thatNaOH concentrations may depend on the aluminosilicate source;this has been verified by the findings of the present work. There-fore, one may conclude that for geopolymers with TPOFA as thealumino silicate source, NaOH with a concentration of 10 M bestincreases the strength.

3.2.2. Effects of Na2SiO3 to NaOH weight ratio (Factor E)The response index for Factor E is shown in Fig. 1e. The response

index increased with the increase of Factor E up to E5. Nonetheless,in the actual experiment, the workability of specimens fabricatedwith amounts above E1 started to decrease due to higher viscosityof Factor E. The increased response index contributes to higher[SiO4]4� concentrations which in turn increases the reaction rate(a higher concentration of reactants results in a higher reactionrate).

The analysis of the response index which increased with the in-crease in amorphous silica content testified to the insufficient reac-tion of the amorphous silica introduced in the form of Factor B‘‘SF’’, as indicated in (Fig. 1b). This phenomenon can be comparedto that introduced in the form of Factor E ‘‘soluble sodium silicate(water glass)’’, as indicated in (Fig. 1e). This difference was attrib-uted to the rate of influence on the response index by the forma-tion of C–S–H and geopolymer (N–A–S–H) gels with an increasein the amorphous silica content. In general, the trends observedwere in agreement with those from a previous investigation [29]which showed that Factor E at E5 leads to the highest compressivestrength. The addition of Factor E into the activating solutions alsoenhances the polymerization process of the ionic species present inthe system, thus leading to higher response index. On the otherhand, the decreased silica concentration in these solutions leadsto a less-polymerized distribution of silicon species, thereby reduc-ing the compressive strength.

3.2.3. Effects of alkaline activator to solid weight ratioThe response index for Factor F is shown in Fig. 1f. The response

index decreased with the increase in Factor F up to F5. The actual

experimental study produced a similar observation. The workabil-ity started to increase with the increase in Factor F; however, evenat very low amounts of Factor F, the mixture still had moderateworkability. This may have occurred because increasing the activa-tor content in the mixture led to an increase in water content. Geo-polymer reactions are dependent mainly on polymerization andcondensation reactions. The reasons for the increase in water con-tent are less obvious, although it is clearly necessary to providesufficient water to facilitate mixing and to provide a mechanismfor ionic transport. The effect of excess water may be to dilutethe reaction or leach the more soluble components and transportthem away from the reaction zone [47].

3.3. Optimum level of key components of treated POFA-basedgeopolymer mortar

Finally, the statistical analysis of the Taguchi method can deter-mine the ideal combination of the optimum levels of all factorswhile simultaneously meeting the highest response index for TPO-FA-geopolymer mortar mixture. The compressive strength obtainedfrom the various trial mixtures (Table 5) ranged from 15.67 to44.74 MPa at 1, 3, and 7 days. Optimum levels achieved for theeffective performance of each response are shown in Table 6. How-ever, the results can be verified by fabricating new mixtures andproducing nine specimens according to the optimum levels of allfactors and testing them at 1, 3, and 7 days. The compressivestrength obtained ranged between 42.64 and 47.27 ± 5.0 MPa at 1,3, and 7 days. This is higher compared to that of the 25th mixtures,as shown in Table 5. Thus, the Taguchi method can be effectivelyapplied to optimize the compressive strength of TPOFA-geopoly-mer mortar. The detailed analysis for this mixture is given in thenext section.

4. Characterization of the TPOFA-based geopolymer mortar mix

4.1. Mineralogical analysis

Fig. 2 shows the XRD diffractograms for TPOFA, TPOFAcontaining the optimum amount of Factors A, B, & C, and the finaloptimized TPOFA geopolymer mortar. The phases found in rawTPOFA are similar to the XRD pattern observed by previousresearchers [7,26,48]. It contains quartz (SiO2), cristobalite (SiO2)and potassium aluminum phosphate (K3Al2(PO4)3), as shown in

M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481 479

Fig. 2a. However, for the TPOFA containing the optimum amountsof Ca(OH)2, Al(OH)3 and SF, the XRD pattern shows the presence ofvarious new compounds, as shown in Fig. 2b. The main constitu-ents are quartz low (SiO2) (ICSD No. 98-007-9362), portlandite(Ca(OH)2) (ICSD No. 98-003-5165), gibbsite (ICSD No. 98-011-2963), and albite Na(AlSi3O8) (ICSD No. 98-005-2543). This clearlyindicates that various active reactions occurred between TPOFAand the additives.

XRD diffractograms of the TPOFA-based geopolymer mortar, asshown in Fig. 2c, indicate the formation of different phases dur-ing geopolymer synthesis. The peaks correspond to hillebrandite(ICSD No. 98-003-7574), quartz low (ICSD No. 98-005-6888),jadeite (ICSD No. 98-011-0201), calcite (ICSD No. 98-000-5340),and gibbsite (ICSD No. 98-011-2963). The presence of the twosignificant phases, i.e. jadeite and hillebrandite indicate the for-mation of high quality geopolymer binders [49]. Jadeite (NaAlSi2-

O6) belongs to the family of aluminosilicates and its crystalstructure contains Si which are tetrahedrally coordinated in sin-gle chains, with Al and Na in octahedral coordination [50]. Hille-brandite, Ca2Si03(OH)2, is a natural member of the CaO–SiO2–H2Oternary system which includes numerous natural and syntheticcalcium silicate hydrate (C–S–H) phases that result in animprovement in strength. Hillebrandite is composed of a three-dimensional network of Ca–O polyhedra that accommodate wol-lastonite-type Si–O tetrahedral chains which resemble thosereported for many calcium silicate hydrate (C–S–H) phases[49,51]. The calcite (CaCO3) phase resulting from the carbonationof Ca(OH)2 or C–S–H with carbon dioxide to form calcium car-bonate which presents itself between 20 and 30 (2h) [52]. Therewas a small peak near 20� (2h), which was attributed to gibbsite,Al(OH)3. It should be noted that preparation was not performedin an inert atmosphere. The presence of gibbsite likely resultedfrom unreacted material or the carbonation of excess Al(OH)4�

species. In the presence of excess sodium, this reaction can alsolead to the formation of sodium carbonate, though none wasdetected by the XRD.

4.2. Chemical bond development analysis using FTIR

Fig. 3 exhibits the FTIR analysis for the TPOFA (source materialwithout additives, Fig. 3a), TPOFA (substituted with the optimumamount of Ca(OH)2, SF, and Al(OH)3 (Fig. 3b), and alkali-activatedTPOFA (with additives, Fig. 3c) specimens. The three different

Wavenumber (cm-1)

5001000150020002500300035004000

1112

875

669

1015-990 451

3467

1650

1420

950

3589 34

06

795

693

and

713

(a)

(c)

(b)

Fig. 3. FTIR spectra of (a) TPOFA, (b) TPOFA with additive (substituted withoptimum amount of Ca(OH)2, Al(OH)3 and SF) and (c) TPOFA-based geopolymermortar.

spectra can be easily distinguished. TPOFA exhibiting a broad bandaround �1103 cm�1 is associated with the asymmetric stretchingof Si–O–Si bonds of SiO2, along with a vibration mode at�474 cm�1 attributed to the symmetric stretching of Si–O–Sibonds of SiO4 [53]. A band at 693 and 713 cm�1 is associated withthe asymmetric vibration mode of the quartz double band [54],that at 795 cm�1 is attributed to the bending vibration mode ofthe Si–O–Si bonds [55], that at 875 cm�1 is associated with theasymmetric vibration mode of the Al–O–H [56], that at�1445 cm�1 is associated with the asymmetric vibration mode ofthe O–C–O bonds in carbonates, and that at 3600–2200 cm�1 isattributed to the stretching vibration mode of the (–OH, HOH)[57]. The FTIR spectrum of the TPOFA with additive material spec-imen showed certain differences when compared to the spectrumof the starting TPOFA (source materials). The band at 3600–2200 cm�1 disappeared and the new band at the �3640 cm�1 re-gion is attributed to the stretching vibration mode of the Ca(OH)2

[58]. A band indicated at source material TPOFA around�1103 cm�1 which shifted to higher frequency at 1112 cm�1 wassimilarly associated with the asymmetric stretching of Si–O–Sibonds of silicon dioxide tetrahedra due to an increase in silicacount derived from SF [53]. A new vibration band was also indi-cated at the �950 cm�1 region and was associated with the asym-metric stretching of Al–O–Si bonds due to increased number ofoctahedral aluminum atoms derived from Al(OH)3 [55].

The differences between the FTIR absorption frequencies for thesource materials TPOFA (with and without additive) and geopoly-mer products could provide evidence of effective geopolymeriza-tion in this study. Bands indicated at source material TPOFA(with and without additive) around 3640 cm�1, �950 cm�1, and875 cm�1 disappear and the main characteristics bands of thevibration spectra signaled at regions of �1103–1112 cm�1 at TPO-FA (with and without additive) shifted to lower frequencies atwave number region 990–1015 cm�1, indicating stretching vibra-tion of Si–O–Al or Si–O–Si chains. This was correlated to the de-crease in the number of Si–O–Al bonds (N–A–S–H gel) and theincrease in the number of Si–O–Si bonds (C–A–S–H gel) due tomore cross-linked (N–A–S–H) and (C–S–H) binder gels whichcaused the compressive strength to increase [8–10]. Theoretically,there should be a direct correlation between the mechanicalstrength and silica content because increasing the amount of silicaincreases the number of the Si–O–Si bonds which are stronger thanSi–O–Al and Al–O–Al bonds. In agreement with alkali-activatedmaterials only based on slag or fly ash, this vibration band is usu-ally identified between 950–1100 cm�1, and is typically associatedwith the binding gels (C–(A)–S–H for slag and N–A–S–H for fly ash)[11]. On the other hand, TPOFA (with and without additive) vibra-tions bands at 795 cm�1 disappeared and double bands (693 and713 cm�1) shifted to higher frequency at 779 cm�1 and 796 cm�1

which is associated with the asymmetric vibration mode of thequartz double band [54], and vibrations band at 474 cm�1 shiftedto lower frequency �451 cm�1 region due to the bending of Si–O–Si and O–Si–O bonds of SiO4 [53]. New vibrations band at�669 cm�1 is attributed to the bending vibration mode of the O–Si–O bending [54]. The common atmospheric carbonation of thegeopolymer was revealed by the band centered at 1500–1400 cm�1, which indicated the appearance of O–C–O stretchingvibrations. Another interesting finding related to geopolymeriza-tion includes the significant bands located at approximately3467 cm�1 and 1650 cm�1 for the O–H, H–O–H stretches and H–O–H bending, respectively [57].

The observation from both XRD and FTIR is well-correlated andexplains the formation of the gel binder (C–S–H and N–A–S–H) inthe geopolymer. This gel binder is responsible for the high com-pressive strength which has been proven by the experimentalworks.

480 M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481

5. Conclusion

This study has shown that TPOFA is suitable for the productionof normal strength geopolymer mortar. The Taguchi method L25indicated that the efficiency of TPOFA can be effectively enhancedwith substitution of additive materials (20 wt.% Ca(OH)2, 5 wt.% SF,and 10 wt.% Al(OH)3) and alkaline activator (10 M NaOH, Na2SiO3/NaOH = 2.5, and alkaline activator/solid = 0.47). The optimumamount of TPOFA used to synthesize the optimum geopolymermortar is 65 wt.%, producing 47.27 ± 5.0 MPa compressive strengthat 7 days of curing. The high compressive strength is due to the for-mation of gel binder (C–S–H and N–A–S–H) in the geopolymer, asproven by XRD and FTIR analyses.

Acknowledgements

The authors gratefully acknowledge the Universiti Sains Malay-sia for providing the financial support through the Research Uni-versity (1001/PAWAM/814103) Grant Scheme for undertakingthe research work. Special thanks are due to United Palm Oil Indus-tries for providing the palm oil fuel ash.

References

[1] Temuujin J, van Riessen A, MacKenzie KJD. Preparation and characterisation offly ash based geopolymer mortars. Constr Build Mater 2010;24(10):1906–10.

[2] Rovnaník P. Effect of curing temperature on the development of hard structureof metakaolin-based geopolymer. Constr Build Mater 2010;24(7):1176–83.

[3] Aydın S, Baradan B. Mechanical and microstructural properties of heat curedalkali-activated slag mortars. Mater Des 2012;35:374–83.

[4] Guo X, Shi H, Dick WA. Compressive strength and microstructuralcharacteristics of class C fly ash geopolymer. Cement Concr Compos2010;32(2):142–7.

[5] Van Jaarsveld JGS, Van Deventer JSJ, Lorenzen L. The potential use ofgeopolymeric materials to immobilise toxic metals: Part I. Theory andapplications. Miner Eng 1997;10(7):659–69.

[6] Liew YM, Kamarudin H, Mustafa Al Bakri AM, Luqman M, Khairul Nizar I,Ruzaidi CM, et al. Processing and characterization of calcined kaolin cementpowder. Constr Build Mater 2012;30:794–802.

[7] Megat Johari MA, Zeyad AM, Muhamad Bunnori N, Ariffin KS. Engineering andtransport properties of high-strength green concrete containing high volumeof ultrafine palm oil fuel ash. Constr Build Mater 2012;30:281–8.

[8] de Jong BHWS, Brown Jr GE. Polymerization of silicate and aluminatetetrahedra in glasses, melts, and aqueous solutions—I. Electronic structure ofH6Si2O7, H6AlSiO71�, and H6Al2O72�. Geochim Cosmochim Acta1980;44(3):491–511.

[9] Duxson P, Provis JL, Lukey GC, Mallicoat SW, Kriven WM, van Deventer JSJ.Understanding the relationship between geopolymer composition,microstructure and mechanical properties. Colloids Surf A 2005;269(1–3):47–58.

[10] He J, Jie Y, Zhang J, Yu Y, Zhang G. Synthesis and characterization of red mudand rice husk ash-based geopolymer composites. Cement Concr Compos2013;37:108–18.

[11] Ariffin M, Bhutta M, Hussin M, Mohd Tahir M, Aziah N. Sulfuric acid resistanceof blended ash geopolymer concrete. Constr Build Mater 2013;43:80–6.

[12] Lee S, Seo M-D, Kim Y-J, Park H-H, Kim T-N, Hwang Y, et al. Unburned carbonremoval effect on compressive strength development in a honeycombbriquette ash-based geopolymer. Int J Miner Process 2010;97(1):20–5.

[13] Rashad AM. Alkali-activated metakaolin: a short guide for Civil Engineer – anoverview. Constr Build Mater 2013;41:751–65.

[14] Rattanasak U, Chindaprasirt P, Suwanvitaya P. Development of high volumerice husk ash alumino silicate composites. Int J Miner Metall Mater2010;17(5):654–9.

[15] Rashad AM. A comprehensive overview about the influence of differentadditives on the properties of alkali-activated slag – a guide for Civil Engineer.Constr Build Mater 2013;47:29–55.

[16] Nazari A, Riahi S, Bagheri A. Designing water resistant lightweightgeopolymers produced from waste materials. Mater. Des. 2012;35:296–302.

[17] He J, Zhang J, Yu Y, Zhang G. The strength and microstructure of twogeopolymers derived from metakaolin and red mud-fly ash admixture: acomparative study. Constr Build Mater 2012;30:80–91.

[18] Somna K, Jaturapitakkul C, Kajitvichyanukul P, Chindaprasirt P. NaOH-activated ground fly ash geopolymer cured at ambient temperature. Fuel2011;90(6):2118–24.

[19] Rajamma R, Labrincha JA, Ferreira VM. Alkali activation of biomass fly ash–metakaolin blends. Fuel 2012;98:265–71.

[20] Rattanasak U, Chindaprasirt P. Influence of NaOH solution on the synthesis offly ash geopolymer. Miner Eng 2009;22(12):1073–8.

[21] Liew YM, Kamarudin H, Bakri AMMA, Binhussain M, Luqman M, Nizar IK, et al.Influence of solids-to-liquid and activator ratios on calcined kaolin cementpowder. Phys Procedia 2011;22:312–7.

[22] Riahi S, Nazari A, Zaarei D, Khalaj G, Bohlooli H, Kaykha MM. Compressivestrength of ash-based geopolymers at early ages designed by Taguchi method.Mater Des 2012;37:443–9.

[23] Peyronnard O, Benzaazoua M. Alternative by-product based binders forcemented mine backfill: recipes optimisation using Taguchi method. MinerEng 2012;29:28–38.

[24] Nazari A, Khanmohammadi H, Amini M, Hajiallahyari H, Rahimi A. Productiongeopolymers by Portland cement: designing the main parameters’ effects oncompressive strength by Taguchi method. Mater Des 2012;41:43–9.

[25] ASTM C618-12a. Standard Specification for Coal Fly Ash and Raw or CalcinedNatural Pozzolan for Use in Concrete. ASTM International, WestConshohocken, PA; 2012.

[26] Altwair NM, Megat Johari MA, Saiyid Hashim SF. Flexural performance of greenengineered cementitious composites containing high volume of palm oil fuelash. Constr Build Mater 2012;37:518–25.

[27] Pacheco-Torgal F, Castro-Gomes JP, Jalali S. Investigations on mix design oftungsten mine waste geopolymeric binder. Constr Build Mater2008;22(9):1939–49.

[28] Thokchom S, Dutta D, Ghosh S. Effect of incorporating silica fume in fly ashgeopolymers. World Acad Sci, Eng Technol 2011;60 243–7.

[29] Pacheco-Torgal F, Moura D, Ding Y, Jalali S. Composition, strength andworkability of alkali-activated metakaolin based mortars. Constr Build Mater2011;25(9):3732–45.

[30] Chindaprasirt P, Chareerat T, Sirivivatnanon V. Workability and strength ofcoarse high calcium fly ash geopolymer. Cement Concr Compos2007;29(3):224–9.

[31] Winnefeld F, Leemann A, Lucuk M, Svoboda P, Neuroth M. Assessment of phaseformation in alkali activated low and high calcium fly ashes in buildingmaterials. Constr Build Mater 2010;24(6):1086–93.

[32] Ryu GS, Lee YB, Koh KT, Chung YS. The mechanical properties of fly ash-basedgeopolymer concrete with alkaline activators. Constr Build Mater2013;47:409–18.

[33] Thakur RN, Ghosh S. Effect of mix composition on compressive strength andmicrostructure of fly ash based geopolymer composites. ARPN J Eng Appl Sci2006;4(4):68–74.

[34] Ross PJ. Taguchi techniques for quality engineering. New York: McGraw-Hill;1988.

[35] ASTM C109/C109M-12. Standard test method for compressive strength ofhydraulic cement mortars (Using 2-in [50 mm] cube specimen. Annual book ofASTM standards; 2012.

[36] Sathonsaowaphak A, Chindaprasirt P, Pimraksa K. Workability and strength oflignite bottom ash geopolymer mortar. J Hazard Mater 2009;168(1):44–50.

[37] Sata V, Sathonsaowaphak A, Chindaprasirt P. Resistance of lignite bottom ashgeopolymer mortar to sulfate and sulfuric acid attack. Cement Concr Compos2012;34(5):700–8.

[38] Yip CK, Lukey GC, Provis JL, van Deventer JSJ. Effect of calcium silicate sourceson geopolymerisation. Cem Concr Res 2008;38(4):554–64.

[39] Yip CK, Provis JL, Lukey GC, van Deventer JSJ. Carbonate mineral addition tometakaolin-based geopolymers. Cement Concr Compos 2008;30(10):979–85.

[40] Bernal SA, Rodríguez ED, de Gutiérrez RM, Gordillo M, Provis JL. Mechanicaland thermal characterisation of geopolymers based on silicate-activatedmetakaolin/slag blends. J Mater Sci 2011;46(16):5477–86.

[41] Swaddle TW. Silicate complexes of aluminum (III) in aqueous systems. CoordChem Rev 2001;219:665–86.

[42] Lee WKW, van Deventer JSJ. The effects of inorganic salt contamination on thestrength and durability of geopolymers. Colloids Surf A 2002;211(2–3):115–26.

[43] Alonso S, Palomo A. Calorimetric study of alkaline activation of calciumhydroxide–metakaolin solid mixtures. Cem Concr Res 2001;31(1):25–30.

[44] Alonso S, Palomo A. Alkaline activation of metakaolin and calcium hydroxidemixtures: influence of temperature, activator concentration and solids ratio.Mater Lett 2001;47(1–2):55–62.

[45] Krizan D, Zivanovic B. Effects of dosage and modulus of water glass on earlyhydration of alkali–slag cements. Cem Concr Res 2002;32(8):1181–8.

[46] Tashima M, Soriano L, Borrachero M, MonzÓ J, Payá J. Effect of curing time onthe microstructure and mechanical strength development of alkali activatedbinders based on Vitreous Calcium Aluminosilicate (VCAS). Bull Mater Sci2013;36(2).

[47] Barbosa VFF, MacKenzie KJD, Thaumaturgo C. Synthesis and characterisationof materials based on inorganic polymers of alumina and silica: sodiumpolysialate polymers. Int J Inorg Mater 2000;2(4):309–17.

[48] Chandara C, Azizli KAM, Ahmad ZA, Hashim SFS, Sakai E. Analysis ofmineralogical component of palm oil fuel ash with or without unburnedcarbon. Adv Mater Res 2011;173:7–11.

[49] Kumar S, Kumar R, Mehrotra S. Influence of granulated blast furnace slag onthe reaction, structure and properties of fly ash based geopolymer. J Mater Sci2010;45(3):607–15.

[50] Kupwade-Patil K, Allouche E. Effect of alkali silica reaction (ASR) ingeopolymer concrete. World of Coal Ash (WOCA) Conference – May 9–12,2011, in Denver CO, USA.

[51] Zhang YJ, Li S, Wang YC, Xu DL. Microstructural and strength evolutions ofgeopolymer composite reinforced by resin exposed to elevated temperature. JNon-Cryst Solids 2012;358(3):620–4.

M.J.A. Mijarsh et al. / Construction and Building Materials 52 (2014) 473–481 481

[52] Villa C, Pecina ET, Torres R, Gómez L. Geopolymer synthesis using alkalineactivation of natural zeolite. Constr Build Mater 2010;24(11):2084–90.

[53] Bernal S, Rodríguez E, Mejia de Gutiérrez R, Provis J, Delvasto S. Activation ofmetakaolin/slag blends using alkaline solutions based on chemicallymodified silica fume and rice husk ash. Waste Biomass Valorization2012;3(1):99–108.

[54] Criado M, Palomo A, Fernández-Jiménez A. Alkali activation of fly ashes. Part 1:Effect of curing conditions on the carbonation of the reaction products. Fuel2005;84(16):2048–54.

[55] Bakharev T. Durability of geopolymer materials in sodium and magnesiumsulfate solutions. Cem Concr Res 2005;35(6):1233–46.

[56] Yunsheng Z, Wei S, Zongjin L. Preparation and microstructure of K-PSDSgeopolymeric binder. Colloids Surf A 2007;302(1–3):473–82.

[57] Li Q, Xu H, Li F, Li P, Shen L, Zhai J. Synthesis of geopolymer composites fromblends of CFBC fly and bottom ashes. Fuel 2012;97:366–72.

[58] Ylmén R, Jäglid U, Steenari B-M, Panas I. Early hydration and setting ofPortland cement monitored by IR, SEM and Vicat techniques. Cem Concr Res2009;39(5):433–9.