Construction and Building Materials 75 (2015) 189–199

Contents lists available at ScienceDirect

Construction and Building Materials

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

New synthesis method for the production of coal fly ash-based foamedgeopolymers

http://dx.doi.org/10.1016/j.conbuildmat.2014.07.0410950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +27 021 9593304; fax: +27 021 9593878.E-mail address: lpetrik@uwc.ac.za (L.F. Petrik).

Nuran Böke, Grant D. Birch, Sammy M. Nyale, Leslie F. Petrik ⇑Environmental and Nano Science Research Group, Department of Chemistry, University of the Western Cape, 7535 Cape Town, South Africa

h i g h l i g h t s

� Foamed geopolymer was synthesised using coal fly ash, NaOH, and NaOCl at 90 �C.� NaOCl foaming agent did not cause early foaming of the geopolymer precursor slurry.� Varying the NaOH/FA ratio controls geopolymerisation and foaming.� Geopolymers were obtained with porosities in the range 35–62%.

a r t i c l e i n f o

Article history:Received 11 December 2013Received in revised form 13 June 2014Accepted 16 July 2014

Keywords:Coal fly ashNaOHNaOClFoamed geopolymerGeopolymerisationPorosity

a b s t r a c t

Foamed geopolymers were synthesised using a South African Class F coal fly ash (FA), sodium hydroxide(NaOH), and the novel foaming agent sodium hypochlorite (NaOCl) at slightly elevated temperatures(90 �C). The synthetic method has the advantage of maintaining control over foaming; the mixed and pre-pared cementitious slurry containing NaOCl is stable for at least 1 h at room temperature, thus, avoidingpremature foaming of the slurry before moulding. In this study, the effect of NaOH/FA on the formation offoamed geopolymers was investigated at predetermined NaOCl/FA ratios (by weight). The NaOH/FA ratiowas changed over the range of 0.16–0.24 at a constant NaOCl/FA composition of 0.50. Increasing NaOH/FA improved geopolymerisation up to a ratio of 0.20 and porosity up to a ratio of 0.22, after which geo-polymerisation and porosity declined. These results were substantiated through XRD, SEM, FTIR, Ramananalyses, mercury intrusion porosimetry, nitrogen adsorption/desorption, and helium density measure-ments. The formulations developed in this study may be used to fabricate materials suitable for applica-tion in construction as light weight fireproof insulation, internal walls or ceiling tiles.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Coal-burning electric power plants generate large amounts offly ash (FA) as a by-product of the coal combustion process. InSouth Africa, approximately 36.7 million tons of coal FA was pro-duced in 2009, of which only around 5.7% was recycled, with theremaining FA is deposited in ash dumps [1]. The disposal of FAcould have a negative impact on the receiving environment if thepH of the solution in contact with the ashes is not appropriatelymonitored [2]. Thus, significant research efforts have beenexpended to find novel applications for FA use other than only asan additive in cement and concrete production or in the synthesisof zeolites [3,4].

Geopolymers are a new class of materials that can be synthes-ised using metakaolin, natural minerals, and waste materials suchas FA, slag, and red mud [5–7]. Among these feedstocks, the use ofFA ash is desirable as it involves the conversion of a copious wastematerial into a useful product. In general, geopolymers are three-dimensional, amorphous-to-semi-crystalline aluminosilicatematerials [8]. They are composed of successive SiO4 and AlO4

tetrahedra connected through an oxygen-bridged bonding frame-work, in which positive ions (Na+, K+, Li+, or Ca2+) balance the neg-ative charge of Al3+ in IV-fold coordination [9]. The dissolution ofaluminosilicate materials from the surface of FA particles in ahighly alkaline environment via hydrolysis is generally acceptedas the first step in geopolymerisation [5], followed by oligomeri-sation, polymerisation, and finally, condensation of dissolvedspecies. The newly formed geopolymeric gel binds unreacted or half-reacted FA particles together [10]. The (hardened) geopolymersprepared in this way have apparent densities of >1 g/cm3 [11].

190 N. Böke et al. / Construction and Building Materials 75 (2015) 189–199

The foamed geopolymers are lightweight materials (bulk den-sity <1 g/cm3). They can be synthesised at high temperature or atlow temperature processing conditions, the latter of which isapplied at ambient or slightly-elevated temperatures and thusmore cost effective and suitable for larger-scale production suchas in the construction industry [12]. The low temperature process-ing comprises of mechanical pre-foaming, chemical foaming(hydrogen peroxide and aluminium powder as foaming agents)and silica fume in situ foaming techniques [12–15]. In an alkalineenvironment, hydrogen peroxide decomposes into water and oxy-gen, and aluminium and silicon elements in the fumed silica initi-ate the release of hydrogen gas, leading to a porous structure.Foaming is rapid when using the above-mentioned foaming agents,but making it difficult, if not impossible, to control reactions in lar-ger-scale production. Conversely, the foaming agent perboraterequires heat to begin foaming, allowing the foaming process tobe controlled via the agency of heat [16]. In the present study,FA-based foamed geopolymers were synthesised using sodiumhypochlorite (NaOCl) as the foaming agent, allowing the controlof foaming via heat in this case as well [17]. The gases released dur-ing this process, although not investigated in this study, are pre-sumably both oxygen produced by the dissociation of NaOCl and,to some extent, carbon dioxide produced by the reaction of NaOClwith the carbon present in FA according to the following reactions[18]

NaOCl! NaClþ 1=2 O2

2NaOClþ C! 2NaClþ CO2

The FA-based foamed geopolymers prepared in this way can beused in the construction industry to make light weight fireproofinsulation, internal walls or ceiling tiles [17].

The foaming agent/FA and alkali activator/FA compositions areimportant parameters in the synthesis of foamed geopolymerssince they control both the extent of foaming and geopolymerisa-tion. The aim of this study was to investigate the effect of NaOH/FAon the microstructure and textural properties of foamed geopoly-mers fabricated using a predetermined amount of NaOCl as thefoaming agent at a constant Si/Al composition.

)

120

2. Materials and methods

2.1. Materials

The coal fly ash (DuraPozz) was supplied by Ash Resources (Pty Ltd) from theLethabo power station in South Africa. The FA had a mean particle size of 25 lm;typically, 90% of the product’s particles passed through a 45 lm sieve accordingto the supplier. NaOCl (Kimex, South Africa, 12%, density 1.173 g/L) and sodiumhydroxide (Merck, 98%) were the other reagents used during experiments.

2.2. Synthesis of foamed geopolymers

A series of samples were prepared in two sets. In the first set, NaOCl/FA was var-ied (0.40, 0.50 and 0.60) while keeping NaOH/FA constant (0.20) (by weight). Theresults of nitrogen adsorption and desorption tests done on these samples revealedthat the sample with NaOCl/FA = 0.50 had the highest specific and mesoporous sur-face area compared to the other samples. Thus, a second set of experiments were

Table 1Initial compositional ratios of the foamed geopolymer samples (by weight).

Sample no NaOCl/FA = 0.50

NaOH/FA H2O/FA

1 0.16 0.442 0.18 0.443 0.20 0.444 0.22 0.445 0.24 0.44

done by varying NaOH/FA (from 0.16 to 0.24 with 0.02 increments) while keepingNaOCl/FA constant (0.50) (wt) (Table 1). Those samples were named according toonly their initial NaOH/FA ratios in the later sections of the paper.

During synthesis, firstly, the NaOCl solution was added to FA and stirred for10 min using a flat blade impeller to obtain a homogeneous slurry. Thereafter thetemperature of the slurry was kept below 40 �C whilst adding small portions ofNaOH pellets until all the predetermined amount of NaOH dissolved. The cement-itous slurry was stirred for an additional half hour. 150 g portions of the mainslurry, having entrapped gas, such as air and probably carbon dioxide, were pouredinto individual polypropylene containers with lids, each container was sealed witha plastic bag [19]. The purpose of sealing was to ensure minimal loss of water ormoisture during the subsequent hydrothermal treatment. The sealing also limitedthe absorption of atmospheric water vapour and carbon dioxide by the NaOH solu-tion in the slurry. Temperature was gradually increased during four days of hydro-thermal treatment to prevent instant expansion of the geopolymer paste and tomaximize dissolution and polycondensation reactions. The sealed samples wereplaced in an oven and aged at 30 �C for a day. The aged samples were heated forthree days: the oven temperature was held at 60 �C during the first day and 90 �Cduring the second and third day (Fig. 1). After completion of hydrothermal treat-ment, each foamed geopolymer sample was unwrapped, cured further for a dayat 30 �C and afterwards kept in its sealed container. Fig. 2 shows the FA used inthe synthesis and the foamed geopolymer with NaOH/FA = 0.20 produced from thisFA.

2.3. Characterisation

Elemental composition of FA was determined by XRF spectroscopy on a PANal-yticalAxios Wavelength Dispersive spectrometer. The loss on ignition (LOI) test wasdone at 950 �C. The XRD patterns of the fly ash and the powder geopolymers sam-ples were recorded on a Bruker powder diffractometer (D8 Advance) equipped witha theta-theta goniometer setting which includes a Cu target X-ray tube (CuKa1 lineat k = 1.5406 Å) and PSD Vantec-1 detector. Data evaluation was done using the EVAsoftware from Bruker based on the International Centre for Diffraction Data (ICDD)PDF database 1998. Small samples of the geopolymer blocks were carbon coatedand their surfaces investigated by SEM/EDS analysis. Micrographs were recordedusing a Zeiss Auriga field-emission scanning electron microscope. The elementalanalysis was done using an Oxford Aztec EDS attached to the SEM. Attenuated totalreflectance Fourier transform infrared (ATR-FTIR) absorption spectra were recordedin the range 4000–450 cm�1 using a Perkin–Elmer-Spectrum100 FTIR spectrometer,equipped with the Universal ATR top plate and diamond crystal. Spectra wererecorded at a spectral resolution of 4 cm�1 and a scan speed of 0.2 cm/s and datawere normalised according to the Spectrum software from Perkin–Elmer. Ramanspectra were acquired using a Jobin–Yvon LabRAM HR Raman spectrometer withthe 514.5 nm line from a Lexel argon ion laser as excitation source.

Before mercury intrusion porosimetry (MIP) analysis, the geopolymer mono-liths (1 ⁄ 1 ⁄ 2 cm3) were dried in an oven at 95 �C for 24 h. The Auto pore II 9220Micromeritics mercury intrusion porosimeter used during measurements can workat low (to 200 kPa) and high pressures (from 414 MPa to 200 kPa), respectively. TheWashburn equation relates the pressure measured during the test to the poreradius as follows by assuming that all the pores are cylindrical and thus pore open-ings are circular in cross section, which makes it possible to derive the modelequation

P ¼ �2c cos ðhÞ=R

where P (psi) is the applied pressure, c is the surface tension of mercury (485 dyne/cm), h is the advancing/receding contact angle (130�) and R is the capillary radius (Å)[20,21]. Nitrogen adsorption and desorption analyses of the foamed geopolymerswere performed using an automatic Quantachrome gas sorption analyzer (instru-ment model ASiQCOV102-2). Small pieces of geopolymer, about 0.6 ⁄ 0.6 ⁄ 0.6 cmin size, were degassed for 14 h under a nitrogen atmosphere at a temperature of100 �C to remove air from the pores. The samples were then cooled under a nitrogen

0

30

60

90

0 1 2 3 4

Tem

pera

ture

(°C

Day

Fig. 1. Curing regime of foamed geopolymers.

Fig. 2. SEM micrographs of FA (a) and the foamed geopolymer with NaOH/FA = 0.20 (b).

N. Böke et al. / Construction and Building Materials 75 (2015) 189–199 191

atmosphere before analyses. Specific surface area of the samples were calculated bythe software depending on the Brunauer–Emmett–Teller method (BET) [22]. Themesoporosity was calculated with the Barrett–Joyner–Halenda (BJH) method usingthe data obtained from the adsorption branch of the nitrogen isotherm [23]. Theskeletal densities of geopolymer monoliths (1 ⁄ 1 ⁄ 3 cm3) were determined usinga helium-based pycnometer (Model 1330, Micromeritics, Norcross, GA). After insert-ing the geopolymer block, the sample chamber was purged with helium 50 timesbefore analysis to ensure removal of atmospheric gases. A total of 10 measurementswere acquired for each sample. For every geopolymer, three blocks were preparedfor the He pycnometry measurements. Average results were used for plotting thegraphs.

Compressive strength tests (CS) were done on 10 ⁄ 10 ⁄ 10 cm3 cubes using aContest compression machine (max force 2000 kN). All samples waited for 28 daysbefore testing. An average of five CS measurements for each sample was used fordata evaluation.

Duplicate leaching tests were done using 1 cm3 geopolymer cubes at 1/10 solid/water ratio for 15 min and 24 h. Upon completion of the shaking, solutions wereseparated by filtration of the leachate from the blocks. The leachates were analysedby a Varian axial (710-ES) inductively coupled plasma-optical emission Spectrom-eters (ICP-OES) and a Thermo Scientific Dionex ICS 1600 ion chromatograph (IC)with a IonPac AS22 column and AG22 Guard column.

FA

0.24

0.22

0.20

0.18

0.16

M M

Q

M M

H Q

M MSS

H

M

3. Results

3.1. X-ray fluorescence (XRF)

XRF analysis is a reliable technique to resolve the elementalcomposition (major, minor, and trace elements) of FA materials.The major oxides in the South African Lethabo FA are SiO2, Al2O3,Fe2O3, and CaO, followed by Mg, Ti, and K (Table 2), which are rep-resentative of the typical composition of a South African FAobtained from combustion of the regional bituminous coal [24].The present FA ash can be defined as Class F according to the Amer-ican Society for Testing and Materials (ASTM) Specification No.C618-84, since the total sum of its SiO2, Al2O3, and Fe2O3 content(88.2%) is above the limit given by the standard (70%) [25]. TheCaO content of FA is lower than 10%, fulfilling another requirementof the standard to define it as Class F [26]. The CaO found in FA isreported to exist both as free lime and bonded to the glassy phaseof FA [27]. The low CaO percentage of the current FA (4.5%) indi-cates low pozzolanity of the material. The presence of alkali earthelements in the form of their oxides (CaO, MgO, and K2O) but anabsence of Na2O indicates that all the Na during geopolymerisationoriginates from the added NaOH. The presence of iron leads to

Table 2Chemical composition of Lethabo fly ash.

SiO2 Al2O3 Fe2O3 CaO MgO TiO2 K2O Na2O LOI other total

54.4 30.6 3.2 4.5 1.1 1.6 0.8 0.0 0.8 0.6 97.5

coloration of the FA. The loss of ignition (LOI) shows the unburnedcarbon content of the FA.

3.2. X-ray diffraction (XRD)

During XRD analysis, the geopolymer samples were scanned inthe region 2h = 10–65�, where the important crystalline peaks typ-ical of mineral phases found in South African FA such as quartz,mullite and hematite, were noted. Fig. 3 shows X-ray diffracto-grams of the FA and geopolymers prepared by the activation ofthe same FA using NaOH/FA = 0.16–0.24. Although the geopoly-mers were prepared using an increased amount of alkali, a compar-ison of XRD patterns confirmed the presence of similar mineralphases as those found in the FA. The X-ray diffractogram of FAexhibited a broad hump in the region 2h = 18–34�, showing thata significant proportion of the material is amorphous. Some peaksattributable to typical crystalline mineral phases such as quartz(SiO2, PDF No. 01-083-0539) and mullite (Al6Si2O13, PDF No. 00-015-0776) were also present.

In the geopolymer diffractograms, the amorphous hump of eachgeopolymer shifted to a higher 2theta value in the region 2h = 18–38� with increasing alkali activation. This can be ascribed to theformation of a geopolymer gel created during dissolution of the

10 15 20 25 30 35 40 45 50 55 60 65 2θ (degree)

Fig. 3. XRD patterns of FA and the foamed geopolymers with respect to increasingNaOH/FA in the range 0.16 and 0.24.

192 N. Böke et al. / Construction and Building Materials 75 (2015) 189–199

FA glassy phase [28]. It is notable that the quartz and mullite peakswere still present but with reduced intensity, after NaOH activa-tion. Quartz and mullite are known to exhibit a lower reactivitytoward alkali media compared to the glassy phase covering FA par-ticles [4]. Thus, the decrease in peak intensities for quartz andmullite in the geopolymer XRD patterns can be ascribed to limitedtransformation of these phases to the geopolymer, as well as adilution effect [29]. The geopolymers exhibited new crystallinephases attributable to halite (NaCl, PDF No. 00-005-0628) andsodalite (Na4Al3Si3O12Cl, PDF No. 00-037-0476), in addition tomullite and quartz. Halite formation originates from reactionscaused by the added NaOCl. Zeolitic products such as sodalite(Si/Al = 1) are frequently formed in FA-based geopolymers viatwo possible routes after aluminosilicate oligomer formation: first,via polymerisation of oligomers followed by gel formation, andfinally, partial conversion of the gel to zeolites; second, via oligo-mer condensation, nucleation, and subsequent crystallisation ofzeolites [30].

3.3. Scanning electron microscopy (SEM)

SEM analyses elucidated the microstructural changes thatoccurred upon FA activation using NaOH. Energy dispersive spec-troscopy (EDS) combined with SEM was used to demonstrate theelemental composition of the geopolymers. Fig. 4 shows themicrostructural changes that occurred with increasing geopoly-mer NaOH/FA ratios. The surfaces of all geopolymers appearedheterogeneous, which is typical of FA-based geopolymers. Asshown in Fig. 4a, the geopolymer synthesised with NaOH/FA = 0.16 was composed mainly of whole FA particles, the sur-faces of which were covered by a layer composed of speciesformed by dissolution of Si, Al and some impurities from the sur-face [31]. The FA particles were apparently loosely compacted byboth these precipitates and the small amount of geopolymeric gelformed due to the low amount of NaOH used as activator(Fig. 4a). On the other hand, the geopolymer with NaOH/FA = 0.18 consisted of partly-reacted FA particles that were wellembedded in the formed geopolymeric gel (Fig. 4b). Geopolymerswith NaOH/FA = 0.20–0.24 exhibited extensive geopolymerisa-tion, which served to bind the whole structure together(Fig. 4c–e); however, the existence of some partly-reacted FAparticles indicated that the geopolymerisation was not complete,which is common for geopolymers based on complex wastematerials such as FA and red mud [5,7]. The reason for thismay be that as the condensation reaction proceeds, the newlyformed aluminosilicate gel covers the surface of FA particles, pre-venting the release of Si and Al from the surface and thus inhib-iting further reaction [32]. Fig. 4d shows a close-up view of thesample prepared with NaOH/FA = 0.20, displaying prism-shapedcrystals and abundant gel covering all the FA particles. EDS anal-ysis of the gel on the point shown in Fig. 4d indicated that O, Na,Al, and Si were the main elements constituting the geopolymergel, in addition to minor impurities such as Fe, K, and Ca originat-ing from FA, and Cl from the foaming agent NaOCl (Fig. 4g).

When the geopolymerisation is not complete, excess Na+ accu-mulated on the geopolymer surface in the form of a salt [15]. Thetabular/rod shaped particles shown in Fig. 4c were determined toconsist mainly of sodium according to EDS. However, the analysiscould not determine the mineral phase of the alkali salts that formthose particles nor many types of crystallites observed on the geo-polymer pore walls (not shown here). On the other hand, it waspossible to determine the nature of diamond-shaped crystals usingEDS (Fig. 4a, Fig. 5a and b). Those crystals were most likely solubleNaCl or halite crystals as EDS indicated that Na and Cl were themain elements and the ratio of Na to Cl was very close to one. Thisassumption was confirmed by disappearance of these crystals from

the geopolymer as could be seen SEM micrographs after waterleaching (Fig. 5c). The presence of anions such as Cl� in the alkaliactivation solution and consequent crystallization in the geopoly-mers may affect the formation kinetics and nature of the geopoly-meric gel [33]. Small amounts of impurities and heavy metals fromFA may also have an effect on the geopolymerisation [34].

3.4. Infrared spectroscopy (FTIR)

FTIR spectral analysis can be used to identify specific functionalgroups present within the Si-containing glassy or crystallinephases of FA that are formed during coal combustion, and alsowithin the geopolymer obtained by alkali activation of the FA.Analyses of FA and geopolymers via FTIR have been widely inves-tigated and well documented in the literature [35–38]. The FTIRwavenumber (cm�1) assignment of Si species in the FA andselected geopolymers are summarised in Table 3 and plotted inFig. 6. The spectrum of FA exhibited a broad (not distinct) bandin the range 1200–900 cm�1 attributed to the heterogeneous struc-ture of FA (Fig. 6a) [37]. The constituents forming FA, that is, theglassy surface layer (SiO2) and crystalline matrix components suchas mullite and quartz, have peaks in that region [38]. Thus, theoverlapping peaks gave rise to the broad FA band, the centre ofwhich was located at approximately1040 cm�1. This band can beassigned to the asymmetric stretching vibration of Si–O–T (T = Sior Al) bonds of SiO4 or AlO4 tetrahedra [39]. The FA sample alsoshowed a distinct band at 550 cm�1, which can be attributed tothe bending of Si–O–Al bonds, indicating octahedrally coordinatedaluminium in mullite [35,38].

The centre of the broad FA band (1040 cm�1) shifted to lowerwavenumbers (1050–890 cm�1) in the geopolymer spectra,becoming intense and narrow (Fig. 6a and b). This shift indicatesthe Al incorporation into the Si–O–Si structure and an increase inthe number of non-bridging oxygens (NBOs) in the silicon tetrahe-dra [37,40]. Above NaOH/FA > 0.18, no further shift was observed(Fig. 6b). The peaks at 880 cm�1 could be assigned to Si–O stretch-ing and –OH bending (Si–OH) [36]. The peaks at 730 cm�1 and665 cm�1 could be assigned to symmetric stretching of Si–O–Siand Al–O–Si [38] and the Si–O–Si bending vibration [41], respec-tively. The Si–O–Al band (550 cm�1) of FA decreased in intensityin the geopolymer spectra with an increasing amount of NaOH, aresult that may be due to limited quartz/mullite conversion or adilution effect, as observed in XRD analysis. The FTIR spectra con-firmed that the ash had been converted into a geopolymer struc-turally under the various applied conditions.

3.5. Raman

Raman spectroscopy can be used to distinguish between vari-ous Si glassy phases. In general, bands between 1200 and800 cm�1 are associated with Si–O stretching vibrations oftetrahedral silicate units [42]. Four major Raman bands at1100–1050 cm�1, 1000–950 cm�1, 900 cm�1, and 850 cm�1, corre-sponding to the disilicate (Si2O5), metasilicate (Si2O6), pyrosilicate(Si2O7), and orthosilicate (SiO4) compositions, were generallyattributed to symmetric stretching vibrations of silicate tetrahedra,respectively, with one (Q3), two (Q2), three (Q1), and four (Q0) NBOs[42]. The bands in the region of 700–400 cm�1 are associated withinter-tetrahedral Si–O–Si linkages [42]. The structures of Si tetra-hedra and the positions of NBOs are re-plotted in Fig. 7 based onliterature values [42].

Peak identifications from Raman analyses for FA and geopoly-mers with NaOH/FA = 0.18 and 0.20 are compiled in Table 4. TheRaman spectra seen in Fig. 8 were plotted using background-sub-tracted data. Four groups of bands were visible in the low fre-quency region of the Raman spectrum of FA: a relatively small

(a) (b)

(d)(c)

tabular par�cles

(g)

(e) (f)

crystal

gel

Fig. 4. SEM micrographs of the foamed geopolymers with NaOH/FA = 0.16–0.24 (a–e) and geopolymeric gel (f) and EDS (g) of the sample with NaOH/FA = 0.20.

N. Böke et al. / Construction and Building Materials 75 (2015) 189–199 193

peak at 430 cm�1 was assigned to four (or greater)-memberedrings present in vitreous silica, a sharp band around 460 cm�1

associated with a-quartz, and a small peak at 495 cm�1 (D1) andstrong peak at 606 cm�1 (D2) assigned to four- and three-mem-bered siloxane rings, respectively [42–47]. Clear identification ofthe mullite phase was difficult due to the overlap of some mullitepeaks with vitreous silica peaks. The bands at 860 and 950 cm�1

were assigned to tetrahedral silica units with four and two NBOs,respectively [42]. The sharp peak at 1025 cm�1 can be attributedto tetrahedral silica units with one NBO [48,49]. The peaks at1350 and 1600 cm�1 were assigned to the disordered (D) andgraphite (G) carbon bands that exist in the FA from the coal com-bustion process [48,50].

All peaks seen in the Raman spectrum of FA drastically changedwith alkali activation; bands at 430 and 460 cm�1 disappeared,most likely due to alkali activation, with the emergence of twosharp peaks at 590 and 662 cm�1on the spectrum of the geopoly-mer with NaOH/FA = 0.18. The peak at 590 cm�1 probably wasnot derived from the vitreous silica band at 606 cm�1, but was anew band formed by alkali addition [42]. This band had a similarfrequency to that of the silica band, implying that it is the resultof similar vibrations detected in both silica and the other silicaglasses [42]. The band located at 662 cm�1 can be assigned to Si–O–Si stretching [51]. The peaks at 590 and 662 cm�1 lost theirsharpness on the spectrum of the geopolymer with NaOH/FA = 0.20 most likely due to further alkali activation of the FA.

(a)

(b)

(c)

holes

Fig. 5. Crystals observed on the surface of the foamed geopolymers with NaOH/FA0.16 (a) 0.20 (b) and the surface of the geopolymer with NaOH/FA = 0.20 afterleaching in water for 24 h (c).

Table 3FTIR bands of the FA and foamed geopolymers.

cm�1 Assignment References

1200–900 broadband

Asymmetric stretching (Si–O–Si and Al–O–Si)

[37,39]

960 Asymmetric stretching (Si–O–Si and Al–O–Si)

[36]

880 Si–O stretching and OH bending (Si–OH) [36]730 Symmetric stretching (Si–O–Si and Al–

O–Si)[33]

665 660–670 cm�1 (Si–O–Si bendingvibration)

[41]

550 Si–O–Al [35,38]

194 N. Böke et al. / Construction and Building Materials 75 (2015) 189–199

Assignment of the small peak at 406 cm�1 on the spectrum of thegeopolymer with NaOH/FA = 0.20 and the broad bands between662 and 1280 cm�1 on the spectra of both geopolymers was notpossible, since these bands have not yet been identified for thegeopolymers in the literature to date. The wide bands at 1383and 1520 cm�1 were most likely the reduced-intensity D and Gbands of carbon.

3.6. Textural properties (MIP, nitrogen adsorption/desorption, heliumpycnometry)

The porous properties of the foamed geopolymer monolithswere determined by mercury intrusion porosimetry (MIP) andnitrogen adsorption and desorption analyses. Pores were definedaccording to the IUPAC classification as micropores (0–2 nm), mes-opores (2–50 nm), and macropores (>50 nm) [52]. The MIP tech-nique is based on the Washburn equation, which assumescylindrical pore geometry, and is able to measure a very widerange of pore sizes (>0.5 lm) [21,53,54]. The porosimeter used inthis study was able to measure pores in the capillary diameterrange of 0.003–100 lm (3–100,000 nm).

The large pores induced by foaming (>18 lm) were above thedetection limit of the MIP method (Figs. 2b and 9a). In Fig. 9a,the proportion of large macropores (8–18 lm) decreased asNaOH/FA increased, indicating the enhancing effect of the alkaliactivator on pore formation. Smaller pores were not clearly obser-vable in Fig. 9a, but could be easily distinguished from each otherin the differential pore volume graphs. Fig. 9b–f shows that meso-porous volume increased regularly up to the geopolymer withNaOH/FA = 0.22, followed by a decrease for the geopolymer withNaOH/FA = 0.24. The mesopores below10 nm (Fig. 9b and f),12.4 nm (Fig. 9c), 23.6 nm (Fig. 9d), and 6.7 nm (Fig. 9e), respec-tively, could not be detected most likely due to their destructionunder mercury pressure. Despite these limitations, it was possibleto derive total porosity from the MIP analysis using the bulk andskeletal density values. Table 5 shows that the foamed geopoly-mers had considerable porosity in the range 35–62%.

The mesopores (up to 0.05 lm or 50 nm) and macropores(between 0.05 lm and mostly up to 0.2 lm or 50–200 nm)observed in Fig. 9d–f may be formed during several stages of geo-polymerisation as follows. Briefly, the decomposition of FA startson the glassy surface of the FA [4]. NaOH is the demineralisingagent, and attacks Si- and Al-bearing mineral phases via hydrolysis,leading to depolymerisation. The bonds forming the glass structure(Si–O–Si and Si–O–Al) break down, resulting in the formation ofSi(OH)4 and Al(OH)4

� oligomers under highly alkaline conditions[55]. Polycondensation of the hydrated Si and Al tetrahedra occurthrough olation and oxolation processes [56], which leads to theformation of a supersaturated solution or gel consisting of a collec-tion of the aluminosilicate (Si–O–Al) particles causing inherentmacroporosity at the early geopolymerisation stages [16]. Improv-ing geopolymerisation conditions leads to the formation of longeraluminosilicate chains that fill the gel macropores, thus creatingmesoporosity inside the gel and causing reduction of macroporousvolume [31].

The nitrogen adsorption/desorption analysis performed on thefoamed geopolymers indicated that micropores (<2 nm) occupieda relatively small volume in their structure. The nitrogen adsorp-tion/desorption isotherms and the corresponding interpolated sur-face-area plots of the foamed geopolymers are shown in Fig. 10.The shape of an isotherm is the result of the porous structure ofthe material investigated [57]. According to IUPAC classifications,the isotherms in Fig. 10 corresponded to Type IV isotherms withhysteresis loops, which are indicative of capillary condensation inthe mesopores [52]. The pore volumes of foamed geopolymers(Fig. 10) showed notable changes; the sample pore volumes

5006508009501100Wave number (cm-1)

FA

(a)

0.20

0.18

0.16

0.24

0.22

9009309609901020

Wave number (cm-1)

968

966

960

960

960

0.16

0.18

0.20

0.22

0.24

(b)

Fig. 6. FTIR vibrational spectra of FA and the foamed geopolymers (a) band shift with increasing NaOH/FA (b).

Fig. 7. Silicate structural units. Bridging and non-bridging oxygen in Si tetrahedron(a) the four major polarised high frequency bands generally ascribed to symmetricstretching vibrations of tetrahedral silica units with one, two, three and four non-bridging oxygens (b) [42].

N. Böke et al. / Construction and Building Materials 75 (2015) 189–199 195

increased with an increase in NaOH/FA, showing an improved con-nectivity of the pore structure. These results are in accordance withthose of the results obtained by the MIP. Type H3 hysteresis loopsseen in Fig. 10 are usually due to aggregated flat particles formingslit-shaped pores [58]. The formation of aggregates among geo-polymeric gels and half-reacted FA particles was demonstrated inSEM micrographs (Fig. 4a–d). Some structural parameters deter-mined from nitrogen adsorption and desorption analyses of the

foamed geopolymers are compiled in Table 5. With an increasingNaOH/FA, the BET specific surface area, pore volume, and mesopor-ous surface area increased. up to the sample prepared with NaOH/FA = 0.22; showing the beneficial effect of the NaOH content on theformation of porosity in the geopolymers as observed with the MIPtests. But the mesoporous area of the geopolymer with NaOH/FA = 0.24 decreased sharply compared to the sample with NaOH/FA = 0.22, indicating that geopolymerisation was very fast in thiscase and thus, resulted in reduced geopolymeric gel formation(Table 5). The interpolated surface area plots of the foamed geo-polymers exhibited that the pore radii fell in three regions:r < 10 nm, 10 < r < 20 and 20 < r < 30 nm, respectively, indicatingwell-connected pore structure (Fig. 10b, d, f, h and j).

Since mercury intrudes most of the pores it gives correct porevolume data [59]. But, there is criticism of the reliability of poresize distribution obtained from the mercury porosimetry methodin hydrated cementitious solids, such as, ‘‘closed pores in thesetypes of systems are not able to contact with mercury’’ and ‘‘thepores are actually not cylindrical, so application of Washburnequation to obtained mercury intrusion data leads to improperpore size distribution results’’ or ‘‘since mercury porosimetry mea-sures the mouth of the pore in the case of ink-bottle shaped poresthe measured distribution will shift to smaller pore sizes’’[60,61,59]. As for BJH, it calculates pore size distributions fromexperimental isotherms using the Kelvin model of pore filling[62]. Moreover, it applies only to the mesopore and small macro-pore size range [62]. In a study, the BJH method was found to giveoverly sharp pore size distributions that are systematically shifted(by about 1 nm) to lower pore sizes compared to the pore sizesobtained from Monte-Carlo simulation that mimics the experi-mental processes that produce Vycor and controlled-pore glasses[63].

Several other methods were reported to be suitable to representpore structure of porous aluminosiliate materials such as quantita-tive transmission electron microscopy (TEM) and small angle neu-tron/X-ray scattering (SANS and SAXS) [31]. The MIP and BJHcannot give quantitative information on pore size distribution ifthe pore network is different from the one assumed by the model[31]. But the MIP and BJH methods are still useful to investigateeffect of varying synthesis and reaction conditions on the pore sizedistributions of the geopolymers, since the systematic error will bethe same in all cases and the relative difference between samples isthe main interest [31].

Table 4Raman bands of the FA and foamed geopolymers.

Compound Bands at wavenumber (cm�1) Assignment Reference

410Fly ash Vitreous silica 430 >five -membered rings (w1) [42–44]

Alpha quartz 460 (strong) [45]495 (D1) Planar four membered rings (D1) [44,46]

Meta silicate (Si2O6) (Q2) 606 (D2) Planar three membered rings (D2) [44,46,47](Qo) Ortho silicate (SiO4

4�) 860 Tetrahedral silica units with four NBOs [42](Q2) Meta silicate (Si2O6) 950 Tetrahedral silica units with two NBOs [42](Q3) 1025 Tetrahedral silica units with one NBO [48,49]

1350 (D band) Carbon [45,50]1600 (G band) Carbon [45,50]

Geopolymer 590 Silica or silica glass [42]662 Si–O–Si [51]

200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700wavenumber (cm-1)

662

1025

1600

430460

606

860

950

0.18

1280590

1350

FA

495 1383 1520

1155

0.20

Fig. 8. Raman spectra of FA and selected geopolymers with NaOH/FA = 0.18–0.20.

0 0.020.040.060.080.1

0 5 10 15 20

Incr

emen

tal v

olum

e (m

L/g)

Pore size (μm)

0.180.16

0.200.220.24

(a)

0

5E-05

0.0001

0.01 0.1 1 10 100dV/D

(mL/

g.Å

)

Pore size (μm)(b)

0

5E-05

0.0001

0.01 0.1 1 10 100

dV/D

(mL/

g.Å

)

Pore size (μm)(c)

0

0.0003

0.0006

0.01 0.1 1 10 100

dV/D

(mL/

g.Å

)

Pore size (μm)(d)

0

0.0003

0.0006

0.001 0.01 0.1 1 10 100

dV/D

(mL/

g.Å

)

Pore size (μm)(e)

0

0.0003

0.0006

0.01 0.1 1 10 100

dV/D

(mL/

g.Å

)

Pore size (μm)(f)

Fig. 9. The mercury intrusion porosimetry (MIP) curves of the foamed geopolymers with NaOH/FA = 0.16–0.24 plotted using incremental volume (a) and differentialvolume(b–f) versus pore size.

196 N. Böke et al. / Construction and Building Materials 75 (2015) 189–199

Skeletal density can be determined by measuring the volumeof a solid with known mass by a gas displacement techniqueusing gas pycnometry. The volume of gas displaced by the solidduring helium pycnometric measurements includes both the vol-ume of the closed pores and the solid, while the bulk densityvalue is comprised of open and closed pores plus the volumeof the solid. Helium pycnometry provides reliable skeletal

density values since helium is a small molecule (diameter2.57 Å), having access to even the smallest pores [64]. As shownin Table 5, the skeletal densities of the foamed geopolymersexhibited minor increase with increasing NaOH/FA up to a valueof 0.22, indicating that the expected increase in the skelataldensity with increasing NaOH/FA was neutralized by theincreasing foaming.

Table 5The textural properties of the foamed geopolymers with varying NaOH/FA determined by nitrogen adsorption/desorption analysis, MIP and He pycnometry.

NaOH/FA SBET VBJH, ads SBJH, ads RBJH, ads Porosity (MIP)a Skeletal density STDEV of skeletal densityb

(m2/g) (cc/g) (m2/g) (nm) (%) (g/cc) (g/cc)

0.16 7.6 0.022 5.4 2.1 34.5 2.30 0.0040.18 9.6 0.022 6.7 2.1 36.5 2.33 0.0030.20 17.0 0.071 10.2 1.9 55.0 2.40 0.0030.22 31.4 0.078 14.6 1.7 62.1 2.41 0.0050.24 22.3 0.046 6.7 2.0 46.5 2.34 0.005

a Porosity was calculated using the equation of porosity = 1 � (dbulk/dskeletal) equation where the bulk and skeletal density values determined from MIP analyses.b STDEV is acronym for the standard deviation of three skeletal density measurements.

0

20

40

60

0 0.5 1

Pore

vol

ume

(cc/

g)

p/p0(a)

0

2

4

6

0 10 20 30

Surf

ace

area

(m

²/g)

Pore radius (nm)(b)

0

20

40

60

0 0.5 1

Pore

vol

ume

(cc/

g)

p/p0(c)

0

2

4

6

0 10 20 30

Surf

ace

area

(m

²/g)

Pore radius (nm)(d)

0

15

30

45

60

0 0.5 1

Pore

vol

ume

(cc/

g)

p/p0(e)

0

2

4

6

0 10 20 30 40

iSur

face

are

a(m

²/g)

Pore radius (nm)(f)

0

15

30

45

60

0 0.5 1

Pore

vol

ume

(cc/

g)

p/p0(g)

0

2

4

6

0 10 20 30 40

Surf

ace

area

(m²/

g)

Pore radius (nm)(h)

0

15

30

45

60

0 0.5 1

Pore

vol

ume

(cc/

g)

p/p0

(i)

0

2

4

6

0 10 20 30 40

Surf

ace

area

(m²/

g)

Pore radius (nm)(j)

Fig. 10. The nitrogen adsorption desorption isotherms of the foamed geopolymers with NaOH/FA = 0.16 (a) 0.18 (c) 0.20 (e) 0.22 (g) 0.24 (i) and the correspondinginterpolated surface area plots (b d f h j).

N. Böke et al. / Construction and Building Materials 75 (2015) 189–199 197

3.7. Compression strength (CS)

Representative CS range was covered by selecting the sampleswith NaOH/FA 0.18 and 0.20, which had porosity in between thelowest (NaOH/FA = 0.16) and most porous (NaOH/FA = 0.22) sam-ples. Average bulk density of the geopolymers with NaOH/

FA = 0.18–0.20 was 0.76 (±3%) and 0.82 (±10%) g/cm3, respectively.Results of CS tests showed that these samples had average CS val-ues of 3.3 (±1.5%) and 3.1 (±18%) MPa, respectively. The samplewith NaOH/FA = 0.20 was synthesised using higher NaOH/FA, andis therefore expected to be stronger than the sample with NaOH/FA = 0.18, which is the case with the hardened geopolymers. The

198 N. Böke et al. / Construction and Building Materials 75 (2015) 189–199

lower CS value for the sample with NaOH/FA = 0.20 is most likelyrelated to its more porous structure (Table 5); however, the differ-ence between the two CS values is low due to the fact that the sam-ple with NaOH/FA = 0.20 probably had more geopolymeric gel thanthe sample with NaOH/FA = 0.18 as it had more mesoporous areawhich rendered it resistant to compression (Table 5). It is not fea-sible to expect higher CS values for samples with NaOH/FA > 20;the sample with NaOH/FA = 0.22 may have had a higher mesopor-ous area (14.6 m2/g) compared to the sample with NaOH/FA = 0.20(10.2 m2/g), but greater porosity (62%). The sample with NaOH/FA = 0.24 had less porosity, but overall less mesoporosity com-pared to the sample with NaOH/FA = 0.20 (Table 6). The strengthvalues (3.33 and 3.1 MPa, respectively) determined for the sampleswith NaOH/FA = 0.18 and 0.20 were close to that of an autoclavedaerated concrete produced from a mix of coal bottom ash, Portlandcement and other components such as quartz sand, limestone, andwater [65].

Several synthetic parameters are effective toward enhancinggeopolymer strength. The amorphicity of the geopolymer, i.e., theamount of geopolymeric gel, is known to increase the CS of thematerial [66]. The initial mixture solid/liquid (S/L) ratio, CaO fromthe FA, or CaO addition to the initial mixture and activator/FA areother parameters that affect compressive strength of the final geo-polymer. In the synthesis of an FA-based hardened geopolymer, thecompressive strength of the material has been shown to increasealmost exponentially until S/L = 2.05 kg/L, after which it sharplydecreases [67]. Addition of CaO and Ca(OH)2 (1–3%) is reportedto improve CS of FA-based geopolymers prepared at 20 �C, proba-bly due to precipitation of calcium silicate hydrate or calcium sili-cate aluminate hydrate phases [68]. The activator/FA ratio is alsoknown to increase the strength until a definite value, after whichCS decreases [18]. The addition of chloride salts such as KCl andCaCl2 has been reported to decrease the CS in FA/kaolin-basedhardened geopolymers synthesized using alkaline silicate solutionsof relatively low concentration [69].

3.8. Leaching studies

Leaching tests were performed according to the German stan-dard leaching (DIN-S4) test method with minor modification byusing block foamed geopolymers with NaOH/FA = 0.20 [2]. Leach-ing tests were also performed over 15 min to obtain comparativedata. The results of XRD analysis done on the leached, dried andground geopolymer samples indicated that halite peaks almost dis-appeared from the diffractograms after 24 h of leaching comparedto the un-leached sample (not shown in this paper). Leachingresults were evaluated by ICP-OES and ICS measurements. Table 6shows average leached Si, Al, Na, and Cl content in the leachatesfrom the geopolymer specimens prepared with NaOH/FA = 0.20.Si and Al are framework elements and are expected to remainfirmly attached to the amorphous gel and sodalite unless polymer-isation was incomplete and soluble monomeric or dimeric precur-sor species still remained unincorporated.

Release of Si and Al was measured as 305.5 and 1.52 mg/kg,respectively, after 24 h of shaking; this indicated that the

Table 6Leaching in water from the foamed geopolymer with NaOH/FA = 0.20.

Time of leaching RSD(%)a

15 min 24 h 15 min 24 h

pH 10.34 10.83 0.07 0.07Si (mg/kg) 17.70 305.5 2.6 11.4Al (mg/kg) 0.58 1.52 6.6 33.5Na (%) 3.2 6.5 3.3 2.3Cl (%) 3.8 5.8 11.1 4.5

a Percentage of relative standard deviation.

leachability of Si species was 200 times higher than that of Al,probably due to dissolution of Si, which did not participate in thegeopolymerisation and may be present as uncondensed Si mono-mers formed during hydrolysis. The results of Na and Cl leachingafter 24 h of shaking were determined to be 6.5 and 5.8% (wt),respectively. On the other hand, excess Na and Cl included intothe structure of the foamed geopolymer from NaOCl during syn-thesis was calculated to be 4.9 and 6.7% (wt) of the geopolymer,respectively. Thus, it is possible to estimate how much Na and Clwas leached out or trapped in the structure of the foamed geopoly-mer after leaching. The leached Na was 1.6% (wt) of geopolymerwhen excess Na from the NaOCl was excluded. This value maybe the sum of soluble Na species unused during geopolymerisation,released after 24 h of shaking. The Cl content of the geopolymerwith NaOH/FA = 0.20 was determined to be 0.9% (wt) after 24 hof shaking, showing that the foamed geopolymers have a smallcapacity to trap Cl in the structure.

4. Conclusions

Foamed geopolymers were prepared using coal fly ash as thealuminosilicate source and NaOH as the alkali activator. The foam-ing agent used, NaOCl, has the advantage of providing controlledfoaming. The prepared geopolymer precursor slurry remained sta-ble for 1 h before foaming was initiated by means of heating. Theeffect of NaOH/FA ratios on the formation of foamed geopolymerswas investigated in the range 0.16–0.24 at a predetermined NaOCl/FA ratio of 0.50 (wt). The properties of the product geopolymers,determined using different methods of characterisation (XRD,SEM/EDS, FTIR, Raman analyses), showed various degrees of FAconversion to geopolymers depending on the applied formulations.MIP and nitrogen adsorption/desorption analyses revealed the por-ous structure of the geopolymers. Although the geopolymer withNaOH/FA = 0.22 had the highest mesoporous surface area and thesmallest mesopore radius (14.6 m2/g and 1.7 nm, respectively),the geopolymer with NaOH/FA = 0.20 (the mesoporous surfacearea 10.2 m2/g and the mesopore radius 1.9 nm, respectively)was determined to be the optimum product because this productneeded less alkali activator to provide almost the same character-istics as the geopolymer with NaOH/FA = 0.22 had. The novel foam-ing agent, NaOCl, used in this study provided good foaming of thegeopolymers: the foamed geopolymer with NaOH/FA = 0.20 had aporosity of 55% with a compression strength of 3.1 ± 18% MPa.But porosity may increase and strength may reduce as halite crys-tals originating from NaOCl dissolve. Therefore, the materialsshould not be exposed to direct water flows. Further research isrequired to investigate this topic. The formulations developed inthis study may be used to fabricate materials suitable for applica-tion in construction as light weight fireproof insulation, internalwalls or ceiling tiles.

Acknowledgement

This work was financially supported by the THRIP programmeof the DTI in South Africa.

References

[1] Musyoka NM, Petrik LF, Balfour G, Wilson MG. Synthesis of hydroxy sodalitefrom coal fly ash using waste industrial brine solution. J Environ Sci Health ATox Hazard Subst Environ Eng 2011;46:1699–707.

[2] Wilson WM, Fatoba OO, Petrik LF, Vadapalli VRK. Leaching characteristics ofselected South African fly ashes: effect of pH on the release of major and tracespecies. J Environ Sci Health A Tox Hazard Subst Environ Eng 2009;44:206–20.

[3] Papadakis VG. Effect of fly ash on Portland cement systems: Part I. Low-calcium fly ash. Cem Conc Res 1999;29(11):1727–36.

[4] Inada M, Eguchi Y, Enomoto E, Hojo J. Synthesis of zeolite from coal fly asheswith different silica–alumina composition. Fuel 2005;84:299–304.

N. Böke et al. / Construction and Building Materials 75 (2015) 189–199 199

[5] Duxon P, Fernandez-Jimenez A, Provis JL, Lukey GC, Palomo A, Van DeventerJSJ. Geopolymer technology: the current state of the art. J Mater Sci2007;42:2917–33.

[6] Slag Li C, Sun H, Li L. A review: the comparison between alkali-activated slag(Si+Ca) and metakaolin (Si+Al) cements. Cem Concr Res 2010;40:1341–9.

[7] 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.

[8] Davidovits J. Geopolymers and geopolymeric materials. J Therm Anal1998;35:429–41.

[9] Xu H, Van Deventer JSJ. Effect of source materials on geopolymerization. IndEng Chem Res 2003;42(8):1698–706.

[10] Dimas D, Giannopoulou I, Panias D. Polymerization in sodium silicatesolutions: a fundamental process in geopolymerization technology. J MaterSci 2009;44:3719–30.

[11] Andini S, Colangelo F, Grieco T, Montagnaro F, Santoro L. Coal fly ash as rawmaterial for the manufacture of geopolymer-based products. Waste Manage2008;28:416–23.

[12] Zhang Z, Provis JL, Reid A, Wang H. Geopolymer foam concrete: an emergingmaterial for sustainable construction. Const Build Mater 2014;56:113–27.

[13] Vaou V, Panias D. Thermal insulating foamy geopolymers from perlite. MinerEng 2010;23:1146–51.

[14] Rickard WDA, Vickers L, Van Riessen A. Performance of fibre reinforced, lowdensity metakaolin geopolymers under simulated fire conditions. App Clay Sci2013;73:71–7.

[15] Delair S, Prud’homme E, Peyratout C, Smith A, Michaud P, Eloy L, et al.Durability of inorganic foam in solution: the role of alkali elements in thegeopolymer network. Corros Sci 2012;59:213–21.

[16] Davidovits J. Geopolymer chemistry and applications. Saint-Quentin: InstituteGeopolymere; 2008.

[17] Birch GD. Cellular cementitious composition. US Patent 8167994 B2; May 1,2012.

[18] Powell Fabrication & Manufacturing, Inc. Sodium Hypochlorite BleachDecomposition & Sodium Chlorate Formation Calculator. <http://www.powellfab.com/technical_information/tools/sodiumhypo_decomp.aspx>.

[19] Mulvey RF, Crepeau CE. Thermal insulation composite of cellular cementitiousmaterial. US Patent 4351867, Sep 28, 1982.

[20] Leon CAL. New perspectives in mercury porosimetry. Adv Coll Inter Sci1998;76–77:341–72.

[21] Curran JA, Clyne TW. Porosity in plasma electrolytic oxide coatings. Acta Mater2006;54:1985–93.

[22] Brunauer S, Emmett PH, Teller E. Adsorption of gases in multi molecular layers.J Am Chem Soc 1938;60(2):309–19.

[23] Barrett EP, Joyner LG, Halenda PP. The determination of pore volume and areadistributions in porous substances. I. Computations from nitrogen isotherms. JAm Chem Soc 1951;73(1):373–80.

[24] Babajide O, Musyoka N, Petrik L, Ameer F. Novel zeolite Na-X synthesized fromfly ash as a heterogeneous catalyst in biodiesel production. Catal Today2012;190:54–60.

[25] Musyoka NM, Petrik LP, Fatoba OO, Hums E. Synthesis of zeolites from coal flyash using mine waters. Miner Eng 2013;53:9–15.

[26] Muriithi GN, Gitari WM, Petrik LF, Ndungu PG. Carbonation of brine impactedfractionated coal fly ash: implications for CO2 sequestration. J Environ Manage2011;92:655–64.

[27] Enders M. The CaO distribution to mineral phases in a high calcium fly ashfrom eastern Germany. Cem Concr Res 1996;26(2):243–51.

[28] Ahmari S, Ren X, Toufigh V, Zhang L. Production of geopolymeric binder fromblended waste concrete powder and fly ash. Constr Build Mater2012;35:718–29.

[29] Heah CY, Kamarudin H, Al Bakri AMM, Bnhussain M, Luqman M, Nizar K, et al.Study on solids-to-liquid and alkaline activator ratios on kaolin-basedgeopolymers. Constr Build Mater 2012;35:912–22.

[30] Van Deventer JSJ, Provis JL, Duxon P, Lukey GC. Reaction mechanisms in thegeopolymeric conversion of inorganic waste to useful products. J Hazard Mater2007;A139:506–13.

[31] Sindhunata, Van deventer JSJ, Lukey G, Xu H. Effect of curing temperature andsilicate concentration on fly-ash-based geopolymerization. Ind Eng Chem Res2006;45:3559–68.

[32] Murayama N, Yamamoto H, Shibata J. Mechanism of zeolite synthesis fromcoal fly ash by alkali hydrothermal reaction. Int J Miner Process 2002;64:1–17.

[33] Lee WKW, Van Deventer JSJ. Effects of anions on the formation ofaluminosilicate gel in geopolymers. Ind Eng Chem Res 2002;41:4550–8.

[34] Van Jaarsveld JGS, Van Deventer JSJ. Effect of the alkali metal activator on theproperties of fly ash-based geopolymers. Ind Eng Chem Res 1999;38:3932–41.

[35] Van Jaarsveld JGS, Van Deventer JSJ, Lukey GC. The characterisation of sourcematerials in fly ash-based geopolymers. Mater Lett 2003;57:1272–80.

[36] Lee WKW, Van Deventer JSJ. Use of infrared spectroscopy to studygeopolymerization of heterogeneous amorphous aluminosilicates. Langmuir2003;19:8726–34.

[37] Rees CA, Provis JL, Lukey GC, Van Deventer JSJ. Attenuated total reflectancefourier transform infrared analysis of fly ash geopolymer gel aging. Langmuir2007;23:8170–9.

[38] Fernandez-Jimenez A, Palomo A. Mid-infrared spectroscopic studies of alkali-activated fly ash structure. Microporous Mesoporous Mater 2005;86:207–14.

[39] Criado M, Fernandez-Jimenez A, Palomo A. Alkali activation of fly ash: effect ofthe SiO2/Na2O ratio Part I: FTIR study. Microporous Mesoporous Mater2007;106:180–91.

[40] Skvara F, Kopecky L, Smilauer V, Bittnar Z. Material and structuralcharacterization of alkali activated low-calcium brown coal fly ash. J HazardMater 2009;168:711–20.

[41] Hidalgo A, Petit S, Domingo C, Alonso C, Andrade C. Microstructuralcharacterization of leaching effects in cement pastes due to neutralisation oftheir alkaline nature Part I: Portland cement pastes. Cem Concr Res2007;37:63–70.

[42] McMillan P. Structural studies of silicate glasses and melts-applications andlimitations of Raman spectroscopy. Am Mineral 1984;69:622–44.

[43] Aguilar H, Srra J, Gonzalez P, Leon B. Structural study of sol–gel silicate glassesby IR and Raman spectroscopies. J Non-Cryst Solids 2009;355:475–80.

[44] Sykes D, Kubicki JD. Four-membered rings in silica and aluminosilicate glasses.Am Mineral 1996;81:265–72.

[45] Potgieter-Vermaak SS, Potgieter JH, Belleil M, DeWeerdt F, Van Grieken R. Theapplication of Raman spectrometry to the investigation of cement Part II: amicro-Raman study of OPC, slag and fly ash. Cem Concr Res 2006;36:663–70.

[46] Barba D, Martin F, Ross GG. Evidence of localized amorphous silicon clusteringfrom Raman depth-probing of silicon nanocrystals in fused silica. Nanotechnol2008;19:115707 (5pp).

[47] Kubicki JD, Sykes D. Molecular orbital calculations of vibrations in three-membered aluminosilicate rings. Phys Chem Minerals 1993;19:381–91.

[48] Dauce R, Keding R, Sangleboeuf JC. On the relations between ISE and structurein some RE(Mg)SiAlO(N) glasses. J Mater Sci 2008;43:7239–46.

[49] Ziemath EC. Raman and infrared investigations of glass and glass-ceramicswith composition 2Na2O.1CaO.3SiO2. J Mater Res 1994;9(1):216–25.

[50] Böke N, Godongwana ZG, Petrik LF. Synthesis of an ordered mesoporouscarbon with graphitic characteristics and its application for dye adsorption. JPorous Mater 2013;20:1153–61.

[51] Dai W, Zhu M, Hou YD, Wang H, Yan H, Shao M, et al. Preparation andcharacterization of Ba2TiSi2O8 ferroelectric films produced by sol–gel method.Mater Lett 2004;58:2927–31.

[52] Sing KSW, Everett DH, Haul RAW, Moscou L, Pierotti RA, Rouquerol J, et al.Reporting physisorption data for gas/solid systems with special reference tothe determination of surface area and porosity. Pure App Chem1985;57(4):603–19.

[53] Webb PA. An Introduction to the physical characterization of materials bymercury intrusion porosimetry with emphasis on reduction and presentationof experimental data. <www.micromeritics.com/pdf/app_articles/mercury_paper.pdf>.

[54] Lubda D, Lindner W, Quaglia M, du Fresne von Hohenesche C, Unger KK.Comprehensive pore structure characterization of silica monoliths withcontrolled mesopore size and macropore size by nitrogen sorption, mercuryporosimetry, transmission electron microscopy and inverse size exclusionchromatography. J Chromatogr A 2005;1083:14–22.

[55] Hench LL, West JK. The sol–gel process. Chem Rev 1990;90:33–72.[56] Livage J. Sol–gel synthesis of heterogeneous catalysts from aqueous solutions.

Catal Today 1998;41:3–19.[57] Leofanti G, Padovan M, Tozzola G, Venturelli B. Surface area and pore texture

of catalysts. Catal Today 1998;41:207–19.[58] Miranda-Garcia N, Suarez S, Sanchez B, Coronado JM, Malato S, Maldonado I.

Photocatalytic degradation of emerging contaminants in municipalwastewater treatment plant effluents using immobilized TiO2 in a solar pilotplant. Appl Catal B: Environ 2011;103:294–301.

[59] Curran JA, Clyne TW. Porosity in plasma electrolytic oxide coatings. Acta Mater2006;54:1985–93.

[60] Diamond S. Mercury porosimetry an inappropriate method for themeasurement of pore size distributions in cement-based materials. CemConcr Res 2000;30:1517–25.

[61] Leofanti G, Padovan M, Tozzola G, Venturelli B. Surface area and pore textureof catalysts. Catal Today 1998;41:207–9.

[62] Gas adsorption theory-Micromeritics. <http://www.micromeritics.com/repository/files/gas_adsorption_theory_poster.pdf>.

[63] Gelb LD, Gubbins KE. Pore size distributions in porous glasses: a computersimulation study. Langmuir 1999;5:305–8.

[64] Nakayama A, Suzuki K, Enoki T, Ishii C, Kaneko K, Endo M. Anomalous helium-gas-induced spin-lattice relaxation and the evidence for ultra micropores inmicroporous carbon. Solid State Commun 1995;93(4):323–6.

[65] Kurama H, Topçu _IB, Karakurt C. Properties of the autoclaved aerated concreteproduced from coal bottom ash. J Mater Process Technol 2009;209:767–73.

[66] Duxon 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: Physicochem EngAspects 2005;269:47–58.

[67] Panias D, Giannopoulou IP, Perraki T. Effect of synthesis parameters on themechanical properties of fly ash-based geopolymers. Colloids Surf A:Physicochem Eng Aspects 2007;301:246–54.

[68] Temuujin J, Van Riessen A, Williams R. Influence of calcium compounds on themechanical properties of fly ash geopolymer pastes. J Hazard Mater2009;167:82–8.

[69] Lee WKW, Van Deventer JSJ. The effects of inorganic salt contamination on thestrength and durability of geopolymers. Colloids Surf A Physicochem EngAspects 2002;211:115–26.