Geopolymerization of glass- and silicate-containing heated clay

Construction and Building Materials 159 (2018) 598–609

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Geopolymerization of glass- and silicate-containing heated clay Kenza El Hafid, Mohamed Hajjaji ⇑ Laboratoire de Physico-chimie des Matériaux et Environnement, Unité Associée au CNRST (URAC20), Faculté des Sciences Semlalia, Université Cadi Ayyad, Bd. Prince My Abdellah, B.P. 2390, 40001 Marrakech, Morocco

h i g h l i g h t s
Na-Chabazite and gels formed in the cured alkali activated amended heated clay.
The neoformation processes involved metakaolinite and glass or Na-silicate derivatives.
The properties of the geopolymers and the operating factors were related by using polynomials.
The suitable factors for bricks manufacturing from the cured materials were evaluated.

a r t i c l e

i n f o

Article history: Received 16 July 2017 Received in revised form 5 November 2017 Accepted 7 November 2017

Keywords: Heated clay Waste glass Geopolymer Microstructure Mechanical/physical properties Response surface methodology

a b s t r a c t The microstructures of cured alkali-activated glass-modified heated clay and sodium silicate-containing heated clay were investigated by using X-ray diffraction, Fourier transform infrared spectroscopy, thermal analysis and scanning electron microscope. Moreover, the effects of the SiO2/Na2O ratio, ageing time and curing temperature and their mutual interactions on bending strength and water absorption of the cured alkali-activated materials were assessed by using the response surface methodology (RSM). It was found that the cured materials were composed of gels, neoformed crystalline phases (Na-chabazite, sodium carbonate) together with starting constituents (illite, quartz, metakaolinite, glass). The zeolite/ gel essentially neoformed from metakaolinite derivatives. Sodium carbonate was the product of samples carbonation. A part of the gel formed from waste glass and sodium silicate derivatives respectively. The RSM results showed that the weights of the effects of the experimental factors on the measured properties of the materials obtained were well predicted by using polynomial models, and the SiO2/Na2O ratio was the most influencing factor. The effects of the factors studied were discussed in relation to the microstructure characterization. Also, the results showed that cured alkali activated glass-amended heated clay could be suitable for bricks manufacturing. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction In concentrated alkaline solutions, siliceous materials (e.g., fly ash) and aluminosilicates (e.g., metakaolinite) could lead to the formation of three-dimensional structures of inorganic polymers (geopolymers), considered as promising eco-friendly cementitious materials [e.g., 1]. With the use of sodium hydroxide as activator and aluminosilicates as source materials, the frameworks of geopolymers should be composed of linked tetrahedral units of aluminate and silicate. OH ions provided by the base used plays a key role in the depolymerization of the structure of the aluminosilicate, whilst Na+ ions contributes to the neoformation of zeolites and to the offset of the charge deficit [2–4]. ⇑ Corresponding author. E-mail address: [email protected] (M. Hajjaji). https://doi.org/10.1016/j.conbuildmat.2017.11.018 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

Geopolymerization could be also realized with sodium silicate solution. In this case, the degree of polymerization of the dissolved species was proven to be dependent on the ratio of the amount of SiO2 to that of Na2O. The use of excessive silica may result in a premature solidification of the geopolymer-forming paste, and consequently it leads to the decrease of the rate of the geopolymerization reaction [3]. Referring to some authors [4], higher geopolymerization rates were obtained with mixtures of NaOH and Na2SiO3 solutions. In the use of such a mixture, sodium hydroxide was presumably involved in the dissolution of the source materials, while Na2SiO3 acted as binder [5]. The geopolymerization processes depend on, among others, the chemistry of the activator, the source material, the curing temperature and ageing time [6–9]. Dehydroxylated clay minerals, such as

599

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609 Table 1 Chemical compositions (wt.%) of the clay and waste glass used.

Clay Waste glass *

SiO2

Na2O

CaO

MgO

Al2O3

K2O

Fe2O3

TiO2

*

51.4 68.7

1.0 20.9

3.1 4.4

3.3 3.7

21.7 1.4

3.5 0.4

4.0 0.1

0.6 –

10.7 0.2

L.O.I

Loss on ignition at 1000 °C.

a I

I Q

I

F

I,F

F I,H IH Q

F

Ch,K,I,F

K,Ch

Ch

Q

Q

Q

Q

Q

Q

Q

HC D D D

D

D

RC 10

20

30

40

50

60

°2

70

b

Fig. 1. X-ray diffraction patterns of the raw (RC) and heated (HC) clay (a), and the waste glass used (b). I: illite (PDF# 43-0685); Ch: chlorite (PDF# 73-2376); K: kaolinite (PDF# 83-0971); Q: quartz (PDF# 05-0490); F: K-feldspar (PDF# 76-0831); D: dolomite (PDF# 84-1208); H: hematite (PDF# 85-0987).

Fig. 2. Spatial distribution of the planned experiments listed in Table 2.

metakaolinite, are convenient source materials for geopolymer synthesis. Metakaolinite and NaOH and/or sodium silicate were extensively used for geopolymers production [e.g., 10]. The use of

sodium silicate resulted in good mechanical/physical properties of the geopolymer. But, because of its adverse effect on the environment [11], an intense research activity has been developed in order to find out alternative efficient activators. In strong alkaline solutions, glass could be the subject of dissolution, and gives rise to various siliceous forms [12], which could be used as an alternative substitute of sodium silicate in the elaboration of geopolymers [13]. As soda-lime glass, wastes of glass windows could be used in conjunction with aluminosilicates for geopolymers preparation. In fact, recently waste glass together with fly ash [13] or metakaolinite [14] was used for geopolymers synthesis. However, to the best of our knowledge no attention was paid to the geopolymerization of alkali activated glass-modified raw clays or heated clays. Moreover, very little attention has been paid to the evaluation of the simultaneous effects of SiO2/Na2O molar ratio, ageing time and curing temperature on the mechanical/physical properties of raw clay-based geopolymers.

600

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

Table 2 Planned experimental conditions (experimental design matrix) and measured values of the studied properties of the cured alkali-activated samples (Y1, Y2: Bending strength; Y3, Y4: Water absorption). Experiment number

X1

X2

X3

SiO2/Na2O

t (days)

T (°C)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0 1 0.5 0.5 1 0.5 0.5 0.5 0.5 0 0.5 0.5 0 0 0

0 0 0.866 0.2887 0 0.866 0.2887 0.866 0.2887 0.5774 0.866 0.2887 0.5774 0 0

0 0 0 0.8165 0 0 0.8165 0 0.8165 0.8165 0 0.8165 0.8165 0 0

5 7 6 6 3 4 4 6 6 5 4 4 5 5 5

14.5 14.5 26.2 18.4 14.5 2.8 10.6 2.8 10.6 22.3 26.2 18.4 6.7 14.5 14.5

57.5 57.5 57.5 84 57.5 57.5 31 57.5 31 31 57.5 84 84 57.5 57.5

Glass-modified heated clay

Na-silicate-containing heated clay

Y1 (MPa)

Y3 (%)

1.55 0.26 1.12 5.21 2.87 3.47 17.65 1.18 0.46 6.52 3.01 1.13 9.49 1.42 0.9

18.2 26.19 21.16 21.81 13.7 15.56 15.01 21.02 22.96 20.14 13.63 16.52 20.51 19.05 18.37

Y2 (MPa)

Y4 (%)

0.87 0.29 0.85 1.18 1.96 1.41 0.65 0.64 0.16 1.66 1.83 2.48 2.09 1.15 0.93

20 25.6 12.2 19.5 18.8 16.4 16.9 18.5 19.9 18.7 10.8 15.3 18.9 19.4 19.9

This study focused on the geopolymerization processes of alkali-activated waste glass -modified heated clay and sodium silicate-containing heated clay. For this objective, the microstructure of the geopolymers prepared was investigated. Moreover, the mechanical strength and water absorption of the materials prepared were measured, and the effects of the above factors on these properties were evaluated by using the response surface methodology. In this study, the potential use of the cured glass-modified heated clay for bricks preparation was assessed.

2. Materials The basic raw material was from the clay pit of Safi (Morocco). It was composed of illite, kaolinite and chlorite (33, 25 and 10 wt% respectively), and some ancillary minerals (quartz (19 wt%), dolomite (6 wt%), hematite (4 wt%) and K-feldspar (3 wt%)). Considering the chemical composition of this clay (Table 1), the mass ratio of silica to alumina of aluminosilicates was estimated to be 1.5. This ratio supported the above assemblage of clay minerals. Based on the X-ray diffraction pattern of the raw clay (RC) given in Fig. 1a, the oxides of iron and calcium (Table 1) derived essentially from hematite and dolomite respectively. To be used as a source material for geopolymer preparation, the raw clay was heated at 700 °C for 2 h. As a result of heating, kaolinite and chlorite disappeared, whereas illite persisted (Fig. 1a). The particle size distributions D10, D50 and D90 of the heated clay were determined to be: 30, 270 and 700 mm respectively.

Fig. 3. Typical X-ray diffractograms of cured alkali activated glass-modified heated clay (G-a, G-b) and alkali activated sodium silicate-containing heated clay (S-a, S-b). Conditions of preparation: G-a and S-a: ratio = 5, t = 6.7 days, T = 84 °C; S-a and S-b: ratio = 4, t = 10.6 days, T = 31 °C. C: Na-chabazite (PDF# 84-0698); S: sodium carbonate (PDF# 86-0298); I: illite (PDF# 43-0685); Q: quartz (PDF# 05-0490).

The waste glass used was supplied by a local shop of glass windows. The vitreous feature of this material was revealed by the broad X-ray reflexion of the diffractogram shown in Fig. 1b. Refer-

Table 3 ANOVA results and fitting coefficients (R2) related to the adopted polynomial models. Sources of variation Glass-modified heated clay

Y1

Y3

Sodium silicate-containing heated clay

Y2

Y4

a ***

Sum of squares

Degrees of freedom

Mean squares

F-ratio

a

Signification ***

R2

Regression Residual Total Regression Residual Total

271.0913 37.813 308.9044 179.1205 0.6574 179.778

9 7 16 9 7 16

30.1213 5.4019

444.3974

0.0279

19.9023 0.0939

211.9051

<0.01***

0.996

Regression Residual Total Regression Residual Total

6.3631 0.1738 6.5369 167.4364 8.4499 175.8863

9 7 16 9 7 16

0.707 0.0248

28.4756

0.0234***

0.973

18.604 1.2071

352.4828

0.0330***

0.952

Probability (p) of obtaining a ratio of mean squares greater than F (0 < p < 1). Statistically significant at the level >99.9% (p < .001).

0.878

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

601

G-b G-a S-b S-a

4000

3800

3600

3400

2974

3465 3435

3628

HC

3200

3000

2800

2600

2400

2200

2000

-1

Wavenumber (cm )

G-b G-a S-b S-a

1600

1400

731 663

1000

800

688 642 556 486 461 438

870 800 1200

849 773

1800

1016 2000

1483 1446 1410 1396

1634

HC

600

400

-1

Wavenumber (cm ) Fig. 4. Infrared spectra of the heated clay (HC) and samples of cured alkali activated glass-modified heated clay (G-a, G-b) and sodium silicate-containing heated clay (S-a, S-b). Conditions of preparation: G-a and S-a: ratio = 5, t = 6.7 days, T = 84 °C; S-a and S-b: ratio = 4, t = 10.6 days, T = 31 °C.

ring to the chemical composition of the glass given in Table 1, this material is a typical soda-lime glass composed of about 69% formers, 29% modifiers and 1% intermediates. The glass pieces were crushed and milled with a PM100 Retsch planetary ball mill operating at 500 rpm for 5 min. The grain size of the particles of the milled glass used was <50 mm.

3. Experimental procedures and techniques Pastes composed of the heated clay, glass (3–20 wt%) and NaOH solution (5–14 M) on one hand, and the heated clay, sodium silicate (8–42 wt%) and NaOH solution (3 and 4 M) on the other hand were manually homogenized for 15 min. In both studied cases, the molar ratio of SiO2 to Na2O ranged from 3 to 7. In relation to glass addition, weighted samples of the milled glass (particle size <50 mm) were added to NaOH solutions and stirred with an ordinary magnetic agitator for 4 h, enough time for total dissolution. Paste bars (1 cm  2 cm  5 cm) were shaped and aged at 25 < T < 90 °C in open air atmosphere. The ageing time varied between 1 and 29 days. Phases identification was carried out by using X-ray diffraction (XRD), thermal analysis and Fourier transform infrared spectroscopy (FT-IR). The XRD analysis was realized with a Philips X’Pert MPD diffractometer operating with a copper anode

Fig. 5. SEM micrograph showing needles of Na-chabazite formed in cured glassmodified heated clay. The EDS spectrum corresponds to the latter phase. Preparation conditions: ratio = 5, ageing time: 6.7 days, curing temperature: 84 °C.

(kKa = 1.5418 Å), step size (°2h): 0.013 and scan step time (s): 1. For thermal analysis, a Setaram Setsys 24 apparatus was used. The operating conditions were as follows: atmosphere: air, heating rate: 10 °C/min, reference material and crucible: alumina, sample masse: 78 mg. Fourier transform infrared spectra were recorded with a Perkin Elmer spectrophotometer, operating in the range of 4000–400 cm1. The analysis was performed on discs of blends composed of 1 mg of powdered sample and 99 mg of KBr. The deconvolution of the IR bands was performed with the PeakFit v4.12 software (Peak type: Gaussian shape). The best fit was evaluated based on the values of the correlation coefficient (R2), the standard error (SE) and F-statistic. The microstructure of the cured samples was examined with a JEOL JMS 5500 scanning electron microscope equipped with an EDAX Falcon spectrophotometer. For this goal, freshly fractured pieces were carbon-coated The flexural strength (r) of the cured bars was calculated with the relation:

r¼

3FL 2

2bh

ð1Þ

(F: load at failure (N) determined with an Instron 3369 apparatus with the three-point loading method; L: distance between the supports; b: width of the sample; h: thickness of the samples). For water absorption (WA) determination, cured samples were placed on a water immersed sponge and periodically weighted. The amount of absorbed water at equilibrium was determined from the saturation plateau of the kinetic curves.

602

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

Table 4 Elemental compositions of the zeolite and gel formed in glass- and silicate-containing heated clay samples. Element

Weight%

Atomic%

Net Intensity

Error%

K ratio

Z

R

A

F

Zeolite CK OK Na K Mg K Al K Si K Cl K KK Ca K Fe K

3.65 36.47 11.03 0.92 9.18 17.20 0.35 3.82 4.83 12.57

6.75 50.58 10.65 0.84 7.55 13.59 0.22 2.17 2.67 4.99

27.43 937.29 433.19 57.31 742.51 1498.45 29.21 301.45 327.52 425.67

13.74 8.65 8.63 12.69 6.32 5.40 18.35 4.82 4.37 3.14

0.0065 0.1084 0.0359 0.0038 0.0487 0.1030 0.0026 0.0326 0.0427 0.1096

1.1234 1.0772 0.9803 0.9975 0.9609 0.9824 0.9160 0.9120 0.9288 0.8291

0.9378 0.9602 0.9873 0.9952 1.0027 1.0097 1.0285 1.0394 1.0444 1.0664

0.1581 0.2761 0.3308 0.4099 0.5468 0.6067 0.7900 0.9045 0.9215 0.9971

1.0000 1.0000 1.0039 1.0071 1.0089 1.0054 1.0213 1.0356 1.0333 1.0547

Gel CK OK Na K Al K Si K KK Ca K Fe K

7.07 44.09 39.54 2.15 6.26 0.30 0.35 0.24

10.93 51.15 31.92 1.48 4.14 0.14 0.16 0.08

86.58 2346.26 2700.51 184.79 658.78 30.44 30.25 11.16

11.10 6.69 6.63 8.24 5.91 17.64 17.02 56.57

0.0154 0.2073 0.1726 0.0094 0.0351 0.0026 0.0031 0.0022

1.0897 1.0429 0.9473 0.9277 0.9480 0.8785 0.8945 0.7970

0.9655 0.9862 1.0111 1.0250 1.0313 1.0576 1.0619 1.0798

0.2004 0.4507 0.4600 0.4684 0.5896 0.9394 0.9651 1.0126

1.0000 1.0000 1.0017 1.0046 1.0035 1.0231 1.0289 1.1463

ratio, ageing time (t) and curing temperature (T)) were determined by using RSM and by adopting a quadratic model (Eq. (2)). In this purpose, it should be mentioned that RSM studies can be carried out by using a primary order equation or a quadratic model. Based on the literature data [15–18], the quadratic model is considered as a useful model because it allows to assess the weight of the effect of each factor as well as the weights of the effects of the interactions between the factors on the studied property (response). In addition, it enables to determine the optimum values of the property.

Y ¼ bo þ

k k k X X X bi X i þ bii X 2i þ bij X i X j i¼1

i¼1

ð2Þ

16i6j

Y is the studied property. bo, bi, bii and bij are constant. bi represents the weight of the experimental factor ‘‘i” on the property studied (Y). bij expresses the effect of the interaction between ‘‘i” and ‘‘j” factors. bii can be regarded as a curve shape parameter. k is the number of the studied factors (k = 3 in the present study). Xi is the coded variable related to the real value (vi) of the experimental factor ‘‘i” according to the relation:

Xi ¼

v i v oi b Dv i

ð3Þ

v oi is the real value at the centre of the experimental domain. Dvi is the step of variation of the real value, and b is the major coded limit value. Considering the investigated domains of the studied factors: 3  SiO2/Na2O  7; 1  t  28 days; 25  T  90 °C, the values at the centres and the variation steps were: v oSiO2 =Na2 O ¼ 5; v 0t ¼ 14:5 days; v oT ¼ 57:5  C, and D(SiO2/Na2O) = 2, Dt = 13.5 days, DT = 32.5 °C. The Doehlert design, which is considered as a practical and highly efficient second-order experimental matrix [17], was adopted for the experimental design. The required

Fig. 6. Typical SEM micrograph of the cured alkali activated sodium silicatecontaining heated clay. The EDS spectrum represents the chemical analysis of the ‘‘a” area. Conditions of preparation: ratio = 5, ageing time: 2.8 days, curing temperature: 57.5 °C.

4. Experimental design and modelling The mathematical relations between the studied properties of the cured samples and the experimental factors (SiO2/Na2O molar

2

number (N) of experiments (N ¼ k þ k þ 1Þ was 13. The representative points of these experiments are spatially distributed as shown in Fig. 2. It is worthy to note that, in order to validate the experimental error, the run at the centre was duplicated. The measured properties of the cured samples, which prepared in conformity with the conditions of the planned experiments, are given in Table 2. These data were used for the calculation of the model constants (bo, bi, bij and bii) by using the method of least-square regression [19].

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

603

Fig. 7. Particles of sodium carbonate formed in cured alkali activated sodium silicate-containing heated clay. Conditions of preparation: ratio = 5, ageing time: 14.5 days, curing temperature: 57.5 °C.

The analysis of variance (ANOVA) was used as a mathematical tool to validate the fittingness of the above model [20]. The ANOVA results showed that the fisher variance ratio (F-ratio)  1 and the statistical significance was very high 99.9% (Table 3). Moreover, the correlation coefficients (R2) exceeded 0.8 (Table 3). In this purpose, it should be noted that the model is considered acceptable as R2 > 0.8 [21]. Based on the obtained statistical data, the polynomial models (given below) well described the variations of the mechanical strength and water absorption of the materials prepared against the change of SiO2/Na2O ratio, ageing duration and curing temperature. 5. Results and discussion 5.1. Microstructural characterization of the cured blends The X-ray diffraction analysis of the cured alkali activated glass- and sodium silicate-containing heated clay samples indicated the presence of neoformed phases identified to Na-chabazite and sodium carbonate together with starting minerals (illite and quartz) (Fig. 3). The amount of illite, characterized by the intensity of the basal distance at around 10 Å, was sensitive to the change of the experimental factors. Unreacted metakaolinite and residual glass, distinguished by the well resolved IR frequencies 560–556 cm1 and 469–461 cm1, were encountered in the cured samples (Fig. 4). As far as the IR analyses are concerned, the frequency 3628 cm1 was linked to the stretching vibration of OH of illite [8], which was the only

heat-resistant clay mineral (Fig. 1a). The bands at about 3465 and 3435 cm1 were attributed to the stretching vibrations and the OH bonds of water. The unexpected frequency 2974 cm1 was probably due to an organic contaminant. The bending deformation of the bonds of water manifested at about 1634 cm1. The bands, which occurred in the range of 1485–1415 cm1 and at 870 cm1, were assigned to the bonds of CO2 3 of sodium carbonate. The frequencies 800, 773 and 688 cm1 were linked to quartz. Bands assigned to zeolite occurred at about 731, 663 and 438 cm1. Details regarding the IR bands of the zeolite/gel are reported hereafter. Na-chabazite occurred in cured glass-modified heated clay samples as dense interlocked needles with an average length of 5 mm (Fig. 5). Considering the EDS data given in Table 4, and the number of oxygen anions (24O2) per 1/3 of unit cell of chabazite [22], the chemical formula of the neoformed chabazite was determined to be (Na4.6Ca1.26K0.94Mg0.36) (Al3.24Fe2.14) Si5.82O24xH2O (x = 12). Short and scarce acicular particles of chabazite together with abundant zones of piles of matter were found in cured sodium silicate-containing samples (Fig. 6). These zones corresponded presumably to a gel. Based on the EDS data given in Table 4, and assuming that the cationic charges were balanced by appropriate amount of oxygen anions, the calculated chemical formula of the neoformed gel was to be Na0.85Al0.4 Si0.11O2.5zH2O, and the ratio Si/Al approximated 3. Thus, its chemical structure seemed to be mainly composed of poly(sialate-disiloxo) units [2]. The microscopic examinations showed that some cured samples enclosed particles of sodium carbonate looking like interconnected ropes

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

a

0.0

DTG (%/min)

HeatFlow (mW)

30

20

-0.2

0

-2

mass variation (%)

604

-4 -0.4 -6 -0.6 -8

10 -0.8

-1.0

-10

-12

0 200

400

600

800

1000

b DTG (%/min)

HeatFlow (mW)

30

20

0.0

0

-0.2

-2

-0.4

-4

-0.6

-6

-0.8

10

-1.0

-1.2

0 200

400

600

800

-8

Mass variation (%)

T (°C)

-10

-12

1000

T (°C) Fig. 8. Thermal curves of cured alkali activated glass-modified heated clay (a) and sodium silicate-containing heated clay (b). Conditions of preparation: ratio = 5, ageing time: 6.7 days, curing temperature: 84 °C.

(Fig. 7). Particles of sodium carbonate also manifested as lumps placed at the adjacent environment of the ‘‘pseudo-ropes” (Fig. 7). Considering these observations, and based on the fact that the intense IR band of Na-carbonate was splitted into two bands: 1483 and 1446 cm1 (Fig. 4), sodium carbonate was probably present in well and poorly crystallized states. The similarities in terms of the mineral compositions of the cured glass- and sodium silicate-containing blends were supported by thermal analysis (Fig. 8). Adsorbed water (about 3.3 wt%) was lost at about 140 °C, and zeolite/gel was the subject of dehydroxylation at 570–645 °C. The amount of the released water was estimated to be 2.2–2.7 wt%. Sodium carbonate decomposed at 740–750 °C and nepheline neoformed in the range of 775–800 °C [8]. The endotherm at 245 °C (Fig. 8a) was associated to the loss of zeolitic water. It was manifested particularly on the thermal curve of zeolite-rich samples, i.e., cured glass-modified heated clay samples.

To obtain a better understanding of the geopolymerization process, the hidden IR bands of the spectra of cured samples were identified (Figs. 9 and 10), and their assignments are reported in Table 5. The main set of the new bands of the spectrum of the cured glass-containing heated clay was observed in the range of 1230–870 cm1. These bands confirmed the formation of zeolite/ gel (1079–1075, 979–971, 725–726, 661 and 461–463 cm1), and showed the presence of polymer species (1181 cm1) and various structural units of silicon: linear Q1 (987 cm1), monomer and dimer (941 cm1) species. The presence of the above identified gel, simply represented by N-A-S-H (Si/Al = 3:1) was supported by the frequencies 1119, 1018 and 463 cm1 very close to those reported in [27]. In the presence of Na-silicate, the new bands were associated especially to gel/zeolite (1075 and 979 cm1). In the alkaline solutions used, glass depolymerised and various ‘‘Si” units formed in accordance to the following transformations [12]:

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

605

and silicate species combined to form aluminosilicate oligomer, which yielded to amorphous (gel) and crystalline (zeolite) aluminosilicate structures. 5.2. Mechanical/physical properties of the cured materials 5.2.1. Mechanical strength The results of the response surface methodology showed that the bending strengths of the cured alkali-activated heated clay containing glass (Y1) and Na-silicate (Y2) depended on the coded variables X1 (SiO2/Na2O ratio), X2 (ageing time) and X3 (curing temperature) as follow:

Y1 ¼ 1:33  2:81X1  1:39X2  1:80X3 þ 0:23X21 þ 1:07X22 þ 7:81X23 þ 0:23X1 X2 þ 12:96X1 X3  2:38X2 X3

ð5Þ

Y2 ¼ 1:01  0:86X1 þ 0:28X2 þ 0:67X3 þ 0:11X21 þ 0:19X22 þ 0:46X23  0:12X1 X2  0:45X1 X3  1:04X2 X3

Fig. 9. Deconvoluted bands of the IR spectrum of cured glass-modified heated clay (ratio = 5, ageing time: 6.7 days, curing temperature: 84 °C), and curve fitting parameters.

BSi AO ASi B+ (Na+, OH-)(aq) !BSi AOH þ BSiAO Naþ BSiAO Naþ þ H2 O ! BSiAOH þ ðNaþ ; OH ÞðaqÞ ð4Þ o

In such a way, monomer (Si(OH)4, Q ), dimmer ((HO)3SiAOASi(OH)3, Q1), linear and cyclic trimmer (Q2) and polymer occurred. On the other hand, the structure of metakaolinite was partially depolymerised, and consequently Al(OH) 4 and [SiOx (OH)4x]x/[Si2O1+x(OH)6x]x formed [3]. The aluminate

ð6Þ

The comparison of the absolute values of the coefficients (bj) of the linear terms showed that the weights of the effects of the experimental factors decreased in the order: SiO2/Na2O > T > t. The significant effect of the SiO2/Na2O ratio seemed to be the consequence of the marked influence of Na+ ions and silica on the geopolymerization process [2]. In fact, Na+ ions played a catalytic role for the formation of geopolymers. The effect of silica on the geopolymerization process depended on its content. The use of high amounts of silica impeded the formation of geopolymer due to the solidification of the paste before the reaction completion [3]. In view of these results, strengthening of the geopolymer materials necessitated concentrated alkaline solution but optimum amount of silica. Considering the algebraic values of the coefficient of X2 and X3 (Eq. (5)), the increase of the curing temperature or ageing time had a negative impact on the mechanical resistance of glass amended heated clay samples. In contrast, it had a positive effect on the mechanical strength of the cured samples of silicate-containing heated clay. The differences regarding the response towards the increase of temperature or ageing time could be linked to the observed differences in the nature of the gel. In addition, we believed that due to the rise of temperature, the access of OH and Na+ ions to the aluminosilicate particles was limited probably because of the formation of a semi-solid boundary layer consisting of glass-derivative species. The negative impact due to the increase of the curing time on the mechanical strength could be linked to the depletion of the gluing gel due to its involvement in the development of the particles of chabazite. On the other hand, the positive effect of the rise of the curing temperature or ageing time on the mechanical strength of sodium silicate-containing samples could be associated to their positive effect on the gel development. Based of the values of the coefficients bij of Eq. (6), the simultaneous increase of a couple of factors, among the three studied ones, had an adverse effect on the mechanical strength of cured alkali activated silicate-containing samples. This was not the case for cured samples of glass-modified heated clay. Considering the high and positive value of the coefficient of X1 X3 (Eq. (5)), the simultaneous increase of the SiO2/Na2O ratio and curing temperature had a marked positive effect on the mechanical strength of the glassmodified heated clay samples. However, an antagonistic effect manifested between curing temperature and ageing time. It was believed that due to the simultaneous increase of the latter factors, samples were the subject of a loss of water and subsequently microcracks of shrinkage occurred. Thereby, the mechanical strength slightly declined.

606

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

Fig. 10. Deconvoluted bands of the IR spectrum of cured sodium silicate-containing heated clay (ratio = 5, ageing time: 6.7 days, curing temperature: 84 °C), and curve fitting parameters.

5.2.2. Water absorption The amounts of absorbed water by cured glass-modified heated clay (Y3) and silicate-containing heated clay (Y4) varied as a function of the coded variables according to the following equations:

Y3 ¼ 18:51 þ 6:40X1  0:42X2 þ 0:15X3 þ þ

1:46X23

1:44X21



þ 1:19X1 X2  2:05X1 X3  2:76X2 X3

1:36X22 ð7Þ

Y4 ¼ 19:76 þ 3:03X1  2:73X2  0:37X3 þ 2:41X21  7:84X22  0:98X23  0:41X1 X2 þ 0:91X1 X3  4:88X2 X3

ð8Þ

It was derived that the weights of the effects of the factors studied evolved in the same manner as before (SiO2/Na2O > T > t). The increase of water absorption with increasing the SiO2/Na2O ratio seemed to be mainly due to the abundance of open porosity.

607

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609 Table 5 Positions and assignments of the IR Bands of the spectra of the heated clay and glass used, and cured alkali activated materials. a

HC

Glass

3637 3454

1663 1634

3464 3418

1638 1620

1526

1401 1385

1137

b

c

3624 3524 3462 3418 3390 3297 1675 1638

3625 3551 3445

1531 1468 1436 1409 1393 1384 1238 1181 1141 1119 1079

1531 1463 1442

Glass-HC

Silicate-HC

3305 1668 1638

1397

1075

1036 981

798 778

1018 987 971 941 898 865 798 778

870 797 777

726 693 661

725 693 661

597 560 502 487

593 556 517 485

463

461

429

433

979

773 722 692 642 598 556 516 483 444 469 420 a b c

Assignment

References

OH (Illite) OH (Zeolite) OH (Water/Glass) OH (Glass) Glass Zeolite/Gel Water Water/Glass Water MK CO2 3 (Na-carbonate) CO2 3 (Na-carbonate) MK MK Glass Glass derivatives Polymer MK Gel(Si/Al = 3:1) Zeolite/Gel Glass Glass Linear Q1 Zeolite/Gel Monomer and dimer SiO small anions SiAOH (gel)/CO32 (Na-carbonate) SiAO (quartz) SiAO (quartz) Glass OASiAO (MK/Zeolite/gel) SiAO(quartz)/Zeolite/gel(AlO4(condensed) and SiO4(rings) OASiAO (Zeolite) K-feldspar MK MK SiAOAAl (Silicates) SiAOASi (Silicates) MK Gel/zeolite/glass Glass Silicates

[23] [24] [8]

[25]

[8] [26]

[27] [8]

[26] [8] [26] [8,28] [8] [8] [25] [8,28] [25] [8]

Heated clay. Glass-modified heated clay. Sodium silicate-containing heated clay.

Indeed, the geopolymerization process became limited as the amount of silica increased and the quantity of Na+ species diminished. In such a condition, the amount of the geopolymer seemed to be too low to seal off the open pores. It should be noted that the quantitative presence of pores due to the increase of the SiO2/Na2O ratio was likely responsible for the weakening predicted by Eqs. (5) and (6). For both mixtures studied, the increase of the ageing time resulted in a reduction of water absorption. This fact could be due to the reduction of the size and/or the amount of pores because of the development of the geopolymer. The effect of the change of the curing temperature on water absorption was less important. Water absorption by cured glass-heated clay blends increased with increasing temperature. However, it diminished for sodium silicate-containing samples. These results could be explained on the basis of pores sealing. The gel formed in cured sodium silicatecontaining samples played a key role in the latter process. Based on the values of the coefficients of XiXj, expressing the weights of the effect of the interaction between the factors studied, cured materials with low water absorption ability could be prepared by a simultaneous increase of ageing time and curing temperature.

5.3. Bricks suitability of cured alkali-activated glass-modified heated clay Cured samples of glass-modified heated clay had good mechanical strength (up to 18 MPa) compared to the cured samples of sodium silicate-containing heated clay (up to 2.5 MPa). Thus, we were interested in evaluating the potential use of the former blend in bricks manufacturing. This evaluation was realized by using the response surface methodology and following the optimization method described elsewhere [16]. For this objective, the technological requirements for bricks (bending strength 10 MPa, water absorption 16%) were considered, and used for the plot of the desirability functions (Fig. 11), needed for calculation purposes. Based on the sketches of Fig. 12, especially the white domains, which defined the suitable experimental factors for bricks manufacturing, bricks made of cured glass-modified heated clay could be successfully fabricated by adopting the following conditions: 3 < SiO2/Na2O < 4.5, 1 < ageing time < 29 days, 25 < curing temperature < 43 °C. The theoretical values of the bending strength and water absorption at the optimum operating conditions (SiO2/Na2O = 3; t = 15 days; T = 37 °C), were found to be 14.5 MPa

608

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

Fig. 11. Desirability functions related to the bending strength (a) and water absorption (b) of bricks.

and 13.9%. These values were very close to the experimental ones: 14.3 ± 0.3 MPa and 13.6 ± 0.4%. This result supported the validity of the above optimization study. 6. Conclusion The results of this study enabled us to conclude that: i) Cured alkali-activated glass- and sodium silicate-containing heated clay were the subject of formation of gels, zeolite (Na-chabazite) and sodium carbonate. ii) The neoformation processes involved metakaolinite derivatives, carbon dioxide of air and glass originated species. The reactivity of illite, metakaolinite and glass were sensitive to the SiO2/Na2O ratio, ageing time and curing temperature. iii) The changes of the bending strength and water absorption of the cured materials prepared against the above operating factors were well described with polynomial models. iv) The weights of the effects of the factors on the measured properties followed the order: SiO2/Na2O > T > t. The use of significant amounts of soda and optimum quantities of silica could improve the technical properties of the cured materials. v) In cured alkali activated glass-containing samples, chabazite coalesced as needles at the expense of the gels, and contributed to the improvement of the bending strength (up to 18 MPa). vi) Cured glass-modified heated clay could be convenient for bricks

Fig. 12. Desirability domains for bricks manufacturing. (a) Curing temperature: 37 °C, (b) ageing time: 15 days, (c) SiO2/Na2O molar ratio: 3.

preparation. For this purpose, the SiO2/Na2O ratio, ageing time and curing temperature should be in the ranges of 3–4.5, 1–29 days and 25–43 °C.

K.E. Hafid, M. Hajjaji / Construction and Building Materials 159 (2018) 598–609

References [1] L.K. Turner, F.G. Collins, Carbon dioxide equivalent (CO2-e) emissions: a comparison between geopolymer and OPC cement concrete, Constr. Build. Mater. 43 (2013) 125–130. [2] K. Komnitsas, D. Zaharaki, Geopolymerisation: a review and prospects for the minerals industry, Miner. Eng. 20 (2007) 1261–1277. [3] C. Li, H. Sun, L. Li, A review: the comparison between alkali-activated slag (Si +Ca) and metakaolin (Si+Al) cements, Cem. Concr. Res. 40 (2010) 1341–1349. [4] A. Palomo, M.W. Grutzeck, M.T. Blanco, Alkali-activated fly ashes a cement for the future, Cem. Concr. Res. 29 (1999) 1323–1329. [5] Z. Yahya, M.M. Al Bakri Abdullah, K. Hussin, K.N. Ismail, R. Abd Razak, A.V. Sandu, Effect of solids-to-liquids, Na2SiO3-to-NaOH and curing temperature on the palm oil boiler ash (Si + Ca) geopolymerisation system, Materials 8 (2015) 2227–2242. [6] P. Rovnanik, Effect of curing temperature on the development of hard structure of metakaolin-based geopolymer, Constr. Build. Mater. 24 (2010) 1176–1183. [7] F. Pacheco-Torgal, J. Castro-Gomes, S. Jalali, Alkali-activated binders: a review. Part 2. About materials and binders manufacture, Constr. Build. Mater. 22 (2008) 1315–1322. [8] K. El Hafid, M. Hajjaji, Effects of the experimental factors on the microstructure and the properties of cured alkali-activated heated clay, Appl. Clay Sci. 116 (2015) 202–210. [9] A.D. Hounsi, G.L. Lecomte-Nana, G. Djétéli, P. Blanchart, Kaolin-based geopolymers: Effect of mechanical activation and curing process, Constr. Build. Mater. 42 (2013) 105–113. [10] L.Y. Gomez-Zamorano, E. Vega-Cordero, L. Struble, Composite geopolymers of metakaolin and geothermal nanosilica waste, Constr. Build. Mater. 115 (2016) 269–276. [11] K.A. Komnitsas, Potential of geopolymer technology towards green buildings and sustainable cities, Proc. Eng. 21 (2015) 1023–1032. [12] P.W.J.G. Wijnen, T.P.M. Beelen, J.W. de Haan, C.P.J. Rummens, L.J.M. van de Ven, R.A. van Santen, Silica gel dissolution in aqueous alkali metal hydroxides studied by 29Si-NMR, J. Non-Cryst. Solids 109 (1989) 85–94. [13] M. Torres-Carrasco, F. Puertas, Waste glass in the geopolymer preparation. Mechanical and microstructural characterisation, J. Clean. Prod. 90 (2015) 397–408. [14] M.R. El-Naggar, M.I. El-Dessouky, Re-use of waste glass in improving properties of metakaolin-based geopolymers: mechanical and microstructure examinations, Constr. Build. Mater. 132 (2017) 543–555. [15] M. Nechar, M.F. Molina, J.M. Bosque-Sendra, Application of Doehlert optimization and factorial designs in developing and validating a solid-phase

[16]

[17]

[18]

[19] [20]

[21]

[22] [23] [24] [25]

[26]

[27]

[28]

609

spectrophotometric determination of trace levels of cadmium, Anal. Chim. Acta 382 (1999) 117–130. M. Zougagh, P.C. Rudner, A. Garcia de Torres, J.M.C. Pavón, Application of Doehlert matrix and factorial designs in the optimization of experimental variables associated with the on-line preconcentration and determination of zinc by flow injection inductively coupled plasma atomic emission spectrometry, J. Anal. At. Spectrom. 15 (2000) 1589–1594. S.L.C. Ferreira, W.N.L. dos Santos, C.M. Quintella, B.B. Neto, J.M. Bosque-Sendra, Doehlert matrix: a chemometric tool for analytical chemistry-review, Talanta 63 (2004) 1061–1067. M.A. Bezerra, R.E. Santelli, E.P. Oliveira, L.S. Villar, L.A. Escaleira, Response surface methodology (RSM) as a tool for optimization in analytical chemistry, Talanta 76 (2008) 965–977. D. Mathieu, R. Phan Tan Luu, Software NEMROD, Université d’Aix, Marseille III France, 1980. D. Mathieu, R. Phan Tan Luu, Approche méthodologique des surfaces de réponse, in: J.J. Droesbeke, J. Fine, G. Saporta (Eds.), Plans d’expériences: Applications à l’entreprise, Edition Technip, Paris, 1997, pp. 270–309. T. Lundstedt, E. Seifert, L. Abramo, B. Thelin, Å. Nyström, J. Pettersen, R. Bergman, Experimental design and optimization, Chemometrics Intelligent Lab. Syst. 42 (1998) 3–40. E. Passaglia, The crystal chemistry of chabazites, Am. Mineral. 55 (1970) 1278– 1301. W. Vedder, Correlations between infrared spectrum and chemical composition of mica, Am. Mineral. 49 (1964) 736–768. R. Szostak, Handbook of molecular sieves, Published by Van Nostrand Reinhold, NY (USA), 1992. M. Criado, A. Fernández-Jiménez, A. Palomo, Alkali activation of fly ashes. Part 1: effect of curing conditions on the carbonation of the reaction products, Fuel 84 (2005) 1048–2054. J.L. Bass, G.L. Turner, Anion distributions in sodium silicate solutions. Characterization by 29Si NMR and infrared spectroscopies, and vapor phase osmometry, J. Phys. Chem. B 101 (50) (1997) 10638–10644. Q. Wan, F. Rao, S. Song, R.E. García, R.M. Estrella, C.L. Patino, Y. Zhang, Geopolymerization reaction, microstructure and simulation of metakaolinbased geopolymers at extended Si/Al ratios, Cement Concr. Compos. 79 (2017) 45–52. A. Palomo, M.T. Blanco-Varela, M.L. Granizo, F. Puertas, T. Vazquez, M.W. Grutzeck, Chemical stability of cementitious materials based on metakaolin, Cem. Concr. Res. 29 (1999) 997–1004.