Construction and Building Materials 139 (2017) 342–354

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Waste glass as a precursor in alkaline activation: Chemical process and

hydration products
M. Torres-Carrasco ⇑, F. Puertas
Eduardo Torroja Institute for Construction Sciences (IETcc-CSIC), Spain

h i g h l i g h t s

 Composite binders of urban and industrial waste glass were studied.

 The main reaction product obtained is an alkali-activated material (AAM).
 Curing conditions have a significant effect on mechanical properties.
 Waste glasses as precursor favours the minimization of the highly polluting effects.

a r t i c l e i n f o a b s t r a c t

Article history: This paper discusses the results of alkali-activating waste glass in pursuit of a new type of alkali activated
Received 17 November 2016 materials (AAMs). Waste glass pastes were prepared with different alkaline activators (NaOH, NaOH/
Received in revised form 14 February 2017 Na2CO3, sodium silicate hydrate and KOH) with different concentrations and cured for 20 h at 85 °C
Accepted 15 February 2017
and at 99% or 6.5% relative humidity to assess the effect of these variables on paste mechanical strength
Available online 24 February 2017
and microstructure. Compressive strength performance was better at the higher alkaline concentrations
and when the pastes were cured at low relative humidity (6.5%), which induced low porosity (MIP) and
Keywords:
high specific surface (BET). Irrespective of the type of activator or curing process, the main reaction pro-
Waste glass
Alkali-activated materials
duct was a Si-high, Al-low gel. As a raw material for alkaline cement manufacture, waste glass delivered
Mechanical behaviour compact, high-strength cementitious skeletons.
Microstructure Ó 2017 Elsevier Ltd. All rights reserved.
Curing conditions

1. Introduction studies might then be conducted on their use as compositional cor-

rectors or supplements for more conventional geopolymers. Glass
The alkaline activation of aluminosilicates is a well-known might be one such type of precursor [17]. Compositionally speak-
alternative to portland cement [1–4]. Aluminosilicates may be of ing, glass is comparable to its crystalline source materials with
natural (primarily clay) origin [5–8] or derived from industrial one exception: due to the cooling to which the molten material
by-products, such as glassy blast furnace slag [8–12] or fly ash is subjected, it has a disorderly structure which might, a priori,
[13–15]. When these precursors are dissolved in alkaline solutions be compatible with use in alkaline cement production.
they condense and polymerise, generating so-called alkali- The amount of urban and industrial waste has increased the
activated materials (AAMs) or geopolymers, depending on the world over in the last few decades [17]. Glass is an inert material
chemical composition of the starting material used (rich or poor under common environmental conditions that can be recycled in
in CaO or Al2O3) and the final composition of the main reaction many ways without altering its chemical properties. Waste glass
product. These materials, a new type of inorganic polymers with collection and management are, then, increasingly common ele-
cementitious properties, are regarded as good candidates for appli- ments of environmental policy in the developed world. According
cations including refractory panels [15] and the encapsulation of to EU data, Europeans recycle over eight million tonnes of glass
radioactive materials [16]. containers yearly. In Spain over 700,000 tonnes of glass, or 59 glass
The possible use of other aluminosilicate materials as a solution containers per capita, were deposited in specific street-side bins
to possible metakaolin or fly ash shortages is an area of particular and recycled in 2015 [18], for a recycle rate of approximately
interest. If their suitability for alkaline activation is confirmed, 70%. Waste glass must meet a series of requirements for reuse in
the manufacture of other glass articles, however. The tendency is
⇑ Corresponding author. to collect and sort urban and industrial waste glass by type. Even
E-mail address: mtorres@ietcc.csic.es (M. Torres-Carrasco). so, the mix of glass types, chemical compositions and particle sizes

http://dx.doi.org/10.1016/j.conbuildmat.2017.02.071
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 343

involved renders its reuse by conventional technological processes crushed and ultimately ground to a particle size of <45 lm and
highly complex. As a result, from 10 to 30% of waste glass is not specific surface area of 1.75 cm2/g [33–35].
recyclable for these purposes and alternative valorisation path- The chemical composition given in Table 1 confirms that it con-
ways must be sought. sisted primarily in SiO2, Na2O and CaO.
One possible outlet for such waste is the construction industry. The waste glass was activated with different solutions (Table 2).
Many studies were conducted In the nineteen sixties on the possi- Panreac analytical grade (98% pure) NaOH and (99.8%) Na2CO3
bility of using crushed waste glass as concrete aggregate [19–21]. were used; Merck analytical grade (99% pure) KOH was used;
Over the last 10 years, due to the high costs of eliminating waste and a Merck commercial waterglass (27 wt% SiO2; 8 wt% Na2O;
glass and in the wake of European and US legislation on the envi- 65 wt% H2O) solution with a SiO2/Na2O ratio of 0.86 was used.
ronmental impact of the use of natural aggregate, such research Moreover, two alternative solutions to commercial waterglass
has been renewed [22–28]. were used:
Papers [29–31] have also been published on the use of waste
glass as a raw material to produce cement, given that its main  A NaOH/Na2CO3 solution with the preceding components plus
chemical component is SiO2. One of the concerns addressed in 25 g of dissolved glass as an alternative to commercial water-
those papers is how glass alkalis might affect cement clinker min- glass [33,35]. This is a solution resulting from 6 h at 85 °C of
erals and how much alkali might remain in the clinker or affect the chemical treatment to obtain an alternative to commercial
walls of the kiln. Both depend on the amount of waste glass used as sodium silicate.
a raw material: small replacement ratios can yield good results  A likewise NaOH solution with the preceding component plus
[30]. (15 g per mL) waste glass as an alternative to commercial
Recent studies [30,32] have explored the pozzolanic properties waterglass [34].
of waste glass and its use as a cement replacement in concrete
In both cases, to prepare the alternatives solutions to a commer-
manufacture. A number of benefits would ensue from the reuse
cial sodium silicate, the procedure was as follows: the waste glass
of waste glass in cement and concrete production: environmental
with a particle size <45 mm is brought into contact with the alka-
conservation, longer landfill life, significant energy savings and
line solution (NaOH/Na2CO3 or NaOH) during a stirring time of
lower greenhouse gas emissions.
6 h and a temperature of 85 °C. After this time, the solution was fil-
The feasibility of using vitreous urban and industrial waste as a
tered to be used as an alternative solution in the activation of the
source of silica to wholly or partially replace sodium silicate
waste glass.
(waterglass) in the alkaline activation of slag and fly ash has like-
wise been researched [33–38]. The process entails the partial dis-
2.2. Sample preparation and test methods
solution of the waste glass in highly alkaline media (NaOH/
Na2CO3 and NaOH) to obtain the type of silicon-rich solution that
Paste specimens measuring 1  1  6 cm were prepared from
has been shown to be usable as an alternative to commercial
the waste glass and aforementioned activators at a liquid/solid
sodium silicate activators. These waste glass-based alternative
ratio of 0.4. The conditions for activating and preparing the waste
solutions are less energy-intensive and more eco-friendly than
glass are given in Table 2.
waterglass, which is economically and environmentally problem-
Two curing methods were used.
atic to manufacture, for it calls for heating SiO2 and Na2CO3 to tem-
peratures of over 1000 °C. For every kilogramme of sodium silicate
 Method 1: sealed mould curing [48]. Used in prior research, this
produced, an estimated 1.5 kg of CO2 is emitted into the atmo-
procedure ensures that humidity is held constant to prevent
sphere [39].
quick setting [49–51] and carbonation [52] while the material
Synthesised glasses has been used as a precursor in geopolymer
sets and hardens. The trial consisted of placing the moulds with
preparation [40–43]. The CaO content in synthesised glass and the
the fresh paste in water-containing sealed individual plastic
Si/Al ratios of the starting material are variables that must be borne
bags to prevent evaporation during initial oven curing for 20 h
in mind for their possible effect on the mechanical properties and
at 85 °C and 99% relative humidity. RH was logged continuously
microstructure of the gels forming during alkaline activation [40].
with an EL-USB-2-LCD temperature ( 35 °C–80 °C) and relative
The initial proportions of oxides used to generate the vitreous
humidity (0%–100%) sensor.
phase must be carefully controlled to ensure that the chemical
 Method 2: the pastes were placed in sealed individual plastic
composition of the aluminosilicate precursor optimises alkaline
bags with no water. The specimens were cured at 85 °C for
activation and the production of high performance cements [40].
20 h, during which the relative humidity, held at approximately
Very few studies [44–47] have been conducted on the alkaline
6.5%, was monitored constantly with the EL-USB-2-LCD data
activation of urban and industrial waste glass, however, i.e., on
logger described above.
the suitability of such waste as a prime material (or precursor)
for activation. This study aimed to determine how intrinsic vari-
The prismatic specimens were tested for compressive strength
ables such as the nature and concentration of the alkaline activator
after 24 h of curing. After the strength tests, the pastes were frozen
and curing conditions affect glass activation and the microstruc-
in acetone/ethanol to detain the hydration/activation reactions
ture of the reaction products. That involved assessing the
prior to Hg intrusion porosimetry (MIP), BET, XRD, FTIR, SEM and
mechanical strength and microstructure of the mixes prepared
BSEM/EDX exploration on the following instruments.
with alkali-activated waste glass.
 Mechanical strength tests were conducted to failure on an Ibert-
est Autotest 200/10 hydraulic press as specified in European
2. Experimental procedures
standard EN 196-1 at a rate of 2 400 N/s ± 200 N/s. Compressive
strengths was determined on 1 day specimens.
2.1. Materials
 Total porosity and pore size distribution were found with Hg
intrusion porosimetry (MIP) on a Micromeretics Autopore IV
Soda-lime-silica glass sourced from an industrial and urban
9500 analyser able to exert pressure of up to 32,000 Psi (equiv-
waste management plant at Ajalvir, Madrid, Spain, was used in this
alent to pore sizes down to 0.0067 mm).
study (Fig. 1). Waste glass used was a mix of different colours glass,
344 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

Glass Recycling in Spain 2015

724 000 t of recycled glass containers
Recycling rate 67,4%
Each Spanish recycled 15.5 kg of glass

Experimental treatment of waste glass from Ajalvir plant (Spain)

Waste glass (mixed)
Grinding the waste
A B glass in a ball mill

Sieve
C D

Waste glass <45 µm

Fig. 1. Treatment of waste glass of different colours (from a recycle plant at Ajalvir, Madrid, Spain) to obtain a particle size of <45 lm for subsequent alkaline activation: A)
clear waste glass; B) green waste glass; C) brown waste glass; D) mixed waste glass (used in the study).

Table 1
Chemical composition of waste glass (wt%) (XRF).

% wt CaO SiO2 Al2O3 MgO Fe2O3 Na2O K2O LoI*

Waste glass 11.75 70.71 2.05 1.17 0.52 11.71 1.08 0.83
*
LoI: loss on ignition.

Table 2
Paste preparation: activation conditions.

Sample name Activator type L/S pH % Na2O SiO2/Na2O Curing process

Temperature (°C) Humidity (%)
N/C NaOH/Na2CO3 0.4 13.37 5.0 – 85 99.0
N/C NaOH/Na2CO3 0.4 85 6.5
*
N/C + 25g NaOH/Na2CO3-glass 0.4 13.40 5.0 0.86 85 99.0
*
N/C + 25g NaOH/Na2CO3-glass 0.4 85 6.5
WG 0.86 Waterglass 0.4 13.80 5.0 0.86 85 99.0
WG 0.86 Waterglass 0.4 85 6.5
N10 NaOH 0.4 14.10 9.75 – 85 99.0
N10 NaOH 0.4 85 6.5
**
N10 + 15g NaOH-glass 0.4 13.60 9.75 0.11 85 99.0
**
N10 + 15g NaOH-glass 0.4 85 6.5
N10 + 15Wg NaOH-waterglass 0.4 13.80 8.41 0.19 85 99.0
N10 + 15Wg NaOH-waterglass 0.4 85 6.5
***
KOH KOH-10M 0.4 14.94 – – 85 99.0
***
KOH KOH-10M 0.4 85 6.5
*
Solution prepared by dissolving 25 g of waste glass in 100 mL of NaOH/Na2CO3 [33].
**
Solution prepared by dissolving 15g of waste glass in 100 mL of 10M NaOH [34].
***
10M KOH solution.

 BET specific surface area measurements, pore size distribution  The FTIR spectra were obtained from 64 scans of
and nanoporosity were found from nitrogen adsorption iso- compressed KBr pellets containing 1.0 mg of sample in
therms generated by a Micromeritics ASAP 2010 analyser. 300 mg of KBr run on an ATIMATTSON Genesis Series
 The XRD patterns for the samples were recorded on a Bruker FTIR-TM spectrometer operating in a range of 4000–
AXS D8 Advance diffractometer fitted with a Lynxeye super 400 cm 1.
speed XR detector and a 2.2-kW Cu anode. No monochromator  Carbon-coated samples were examined under a JOEL JSM 5400
was used. The scanning range, from (2h) 5° to 50°, was covered scanning electron microscope fitted with a solid-state backscat-
in a 24 h period. The instrument was set at 40 kW and 30 mA tered electron detector (SEM) and a LINK-ISIS energy dispersive
and the sample was not rotated during scanning. (BSEM/EDX) analyser.
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 345

3. Results 3.1.1. Hg intrusion porosimetry

Total porosity was determined in the systems exhibiting the
3.1. Mechanical strength and porosity (MIP and BET) best mechanical strength, i.e., the specimens cured at low (6.5%)
relative humidity. The total porosity and pore size distribution (in
The 1-day compressive strength values are depicted in Fig. 2A. the 100 mm–0.1 mm range) for the pastes are shown in Fig. 3, where
The activated waste glass exhibited strength ranging from the porosity values are given in percentage of the total sample vol-
17 MPa to 88 MPa, depending on the activator and curing ume. The waste glass pastes activated with the NaOH/Na2CO3-glass
conditions. The highest performance was observed in the exhibited the highest total porosity (29.72%). The presence of sili-
specimens cured at 6.5% relative humidity, for which the values con from the glass waste did not favour 1-day mechanical strength
were 50% to 75% greater than in the specimens cured at high development. In contrast, the commercial sodium silicate (water-
humidity for nearly all the activators (Fig. 2B). The exception was glass) reduced porosity, thereby raising mechanical strength.
the specimens activated with a 10M KOH solution, in which no The waste glass pastes activated with NaOH-glass (N10 + 15g)
strength differences were observed between the two curing showed the lowest total porosity, at around 1%. That finding was
conditions. consistent with the strength values of up to 88 MPa obtained for
The nature of the activator played a significant role in paste this system (Fig. 3B).
strength development. Of the solutions studied, the lowest perfor-
mance was delivered by NaOH/Na2CO3 (pH = 13.3), with values on 3.1.2. BET and nanoporosity
the order of 29 MPa (Fig. 2A), which declined to 17.1 MPa when The differences in mechanical strength found in waste glass
25 g of waste glass was added to that solution as an alternative pastes activated with different activators may also be explained
to commercial sodium silicate, yielding a pH of 13.4. by the specific surface of the samples, and indirectly by the micro-
A possible explanation for such low mechanical strength may porosity of the reaction products. The specific surface was
lie in the pH values (around 13.4) (Table 2). As pH governs alumi- observed to vary between the samples with the highest and lowest
nosilicate dissolution, which favours nucleation or condensation compressive strength (N10 + 15g and N/C + 25g, respectively). In
and polymerisation, it plays a crucial role in activating these mate- low (6.5%) relative humidity curing, the sample activated with
rials. Nonetheless, when a commercial sodium silicate (waterglass) the NaOH-glass solution had a specific surface 12% higher than
with a SiO2/Na2O modulus of 0.86 and pH of 13.8 was used, the the pastes hydrated with NaOH/Na2CO3-glass and 26% higher than
strength measured was on the order of 54 MPa. As the liquid/solid those activated with KOH (Fig. 4A). Fig. 4 B–D) show that pores
ratio (L/S = 0.4) was identical in all the systems, however, when the under 1 nm, associated with gel formation, were present in the
commercial sodium silicate solution was used, the paste had a three aforementioned systems.
lower consistency which may have partially explained the higher Fig. 4A shows how curing in different RH conditions affected the
mechanical strength observed [53]. Moreover, the reactivity of Si specific surface and pore size of the pastes activated with a given
from a commercial sodium silicate is better than the dissolve sili- solution. At 6.5% relative humidity the specific surface of samples
con from waste glass. N10 + 15g was 21% higher than in the respective sample cured at
When the glass pastes were activated with a high OH- concen- 99% relative humidity, whilst in the N/C + 25g material the differ-
tration (such as characteristically used to activate fly ash), mechan- ence was 11%. In contrast, the waste glass pastes activated with
ical strength rose. The best results were observed for the NaOH- 10M KOH exhibited very similar specific surface and pore size val-
glass waste solution (N10 + 15g), which delivered strength of ues (Fig. 4D) irrespective of the relative humidity at which they
88 MPa. Torres-Carrasco and Puertas [34] observed the same were cured, which was consistent with the similar mechanical
strength performance in fly ash activated with this alternative to strength found for the respective specimens. Here also, the smaller
commercial sodium silicate, prepared by dissolving the waste glass ionic radius of Na+ than K+ would explain the lesser mobility of the
in the presence of heat [35]. potassium ion and hence less intense waste glass dissolution.
Lastly, with the 10M KOH solution as an activator, the waste
glass attained strength of around 50 MPa, a value lower than found 3.2. Characterisation of reaction products
for the Na+ ion-containing media. As the K+ cation dissolves less of
the Ca and Mg present in the starting material, less reaction pro- In light of the strength and porosity findings that revealed that
duct precipitates [54,55]. the best results were obtained with low relative humidity curing,

Fig. 2. A) Compressive strength for waste glass activated with different activating solutions and cured at high (99%) or low (6.5%) relative humidity; B) percentage difference
between high RH (99%) and low RH (6.5%) curing conditions.
346 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

Fig. 3. Glass waste pastes activated with different solutions and cured at 6.5% relative humidity: A) pore size distribution vs total porosity; B) compressive strength vs total
porosity.

Fig. 4. Waste glass paste cured at 6.5% or 99% relative humidity: specific surface for A) N10 + 15g, N/C + 25g and KOH; pore size distribution for B) N10 + 15g, C) N/C + 25g and
D) KOH.

the main reaction products in the pastes prepared under those con- unhydrated soda-lime waste glass was amorphous with low range
ditions were characterised mineralogically and microstructurally. structural order. The respective XRD trace exhibited no reflections
attributable to any identifiable crystallised compound, but only an
3.2.1. XRD and FTIR amorphous hump (Fig. 5).
The 1 day diffractograms for the waste glass and the pastes pre- The XRD findings showed that the reaction products in the
pared with the activators studied are reproduced in Fig. 5. The pastes hydrated with each activator retained an amorphous struc-
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 347

ture with scantly any crystalline phases. The amorphous hump was at around 960 cm 1, characteristic of the formation of a CAAASAH
shifted slightly toward higher (30°–32°) 2-theta values, however, a gel (with a higher calcium content) [10].
finding attributed to a higher Na2O content in the main reaction Whilst carbonates were present in all the systems, the content
products formed [44,56]. was particularly high in the waste glass activated with KOH. The
The FTIR spectra for the waste glass pastes prepared with the intense wide band observed at around 1424 cm 1 is characteristic
activators used in this study are shown in Fig. 5B). The narrow sig- of K2CO3. In that vein, paste carbonation or weathering during
nal at 460–480 cm 1 was generated by the bending vibrations in preparation and analysis is difficult to prevent.
the SiAOASi bonds [57]. The likewise narrow signal at 775–
800 cm 1 was attributed to the symmetrical stretching vibrations
generated by the OASiAO bonds [57]. An intense narrow band at 3.2.2. SEM and BSEM/EDX
around 1050 cm 1 was related to the asymmetrical stretching The SEM micrographs of the waste glass pastes activated with
vibrations induced by bridging oxygen. Two small shoulders were the various alkaline activators and cured for 20 h at 85 °C and
observed on this signal, at 940 cm 1 (from non-bridging oxygen) 6.5% relative humidity are shown in Fig. 6. The waste glass pastes
and 1120 cm 1 (asymmetric vibrations from bridging oxygen in activated with NaOH/Na2CO3 and waterglass solutions (N/C, N/C
SiAOASi bonds) [57]. The weak signal at 1460 cm 1 was due to + 25g and 0.86 WG) exhibited non-uniform morphology. The unre-
the carbonate groups present in the glass [57,58]. acted (anhydrous) waste glass particles and the regions where the
As Fig. 5B shows, the main band (TAOAT: Si or Al) for all the glass had dissolved to form the main reaction products can be
activated glass pastes was located at wavenumbers (990– clearly distinguished.
1020 cm 1) slightly lower than observed on the spectrum for the On the other hand, the morphology of the pastes activated with
anhydrous waste glass (around 1050 cm 1). That finding may be a high alkaline concentration (N10, N10 + 15g and N10 + 15Wg)
regarded as evidence of the de-polymerisation of the silica network differed, mainly due to the use of higher concentrations. These
in the waste glass [45,59]. Moreover, the former signal is normally pastes contained less unreacted glass and the formation of hydra-
associated with the precipitation of a silica gel that incorporates tion products on and from the waste glass was more visible, with a
cations Ca2+ and Al3+ into its structure, for the replacement of a larger size that can form agglomerates of the main reaction prod-
Si4+ ion with a Al3+ ion shifts the signal to lower frequencies. The uct. No secondary zeolite-like phases were observed in these low
presence of system modifiers such as Na+ ions also lowers the band relative humidity systems, even when the activator provided extra
frequency [45]. This pattern is observed irrespective of the activa- silica (N10 + 15g and N10 + 15Wg). Although some authors have
tor used [41,60]. In this case, the shift was nonetheless less signif- reported that the presence of silicon sourced from the activator
icant than observed when blast furnace slag is activated with the induces greater zeolite crystallisation [61–63], in this study, the
same alkaline solutions [33]. There, the band generated by the presence of silicon in the activator was offset by the absence of alu-
SiAOASi or AlAOAAl bonds in the main reaction product is located minium, which deterred zeolite formation.

Fig. 5. Unhydrated waste glass and activated waste glass pastes: A) XRD patterns; B) FTIR spectra.
348 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

Fig. 6. SEM micrograph of 1-day alkali-activated pastes cured at 85 °C and 6.5% relative humidity.

The same pattern was observed in the pastes cured at 99% rel- strength and low porosity of the WG system. When more highly
ative humidity (Fig. 7). The microstructures of the pastes activated concentrated activators were used (Fig. 9), gel morphology was
with NaOH-glass (N10 + 15g) and NaOH-Waterglass (N10 + 15Wg) more compact and less unreacted waste glass was observed. The
exhibited more cracks than the materials cured at 6.5% RH. Mor- calcium content was found in all samples, although it was still
phologies typical of zeolites were not identified in these pastes insufficient to form a CASAHA, CAAASAHA or (N,C)A
either, likewise as a result of the low aluminium content [64]. AASAHAlike gel [10].
SEM analysis confirmed an open, scantly compact morphology, Lastly, the waste glass paste activated with KOH had a very
which would explain the lower strength and higher porosity of compact morphology with very few microcracks. Further to the
these materials. analysis in Fig. 10, the aluminium, calcium and sodium were dis-
The BSEM/EDX findings showed that the reaction products con- tributed across both the starting material and the products (gels)
sisted essentially in a silicon-high Ca- and Al-low gel, irrespective formed. FTIR analysis had identified potassium carbonate-like min-
of the type of activator (Figs. 8 and 9). The Si/Al ratios in the gel eralogical species in these pastes.
formed were consequently very high (Table 3).
The BSEM micrographs in Fig. 8 were obtained from the waste 4. Discussion
glass pastes activated with NaOH/Na2CO3, NaOH/Na2CO3-glass
and waterglass solutions, i.e., the solutions with a concentration This study assessed the possibility of using recycle plant waste
of 5% of Na2O. The lighter areas in each region constitute unreacted glass as the staring material (precursor) in the preparation of
(anhydrous) waste glass and the darker areas the reaction alkali-activated materials (AAMs). Such application would be envi-
product, a silicon-rich gel. When the activator used was waterglass ronmentally beneficial, for it would provide an outlet for waste
(WG 0.86), the gel formed was much more compact and glass that cannot be reused in glass manufacture.
uniform than the gels formed with NaOH/Na2CO3 (N/C) and The findings showed that the materials resulting from the alka-
NaOH/Na2CO3-glass (N/C + 25g), confirming the high mechanical line activation of waste glass exhibit high mechanical strength and
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 349

Fig. 7. SEM micrograph of 1-day N10 + 15g and N10 + 15Wg pastes cured at 99% relative humidity.

Table 3
EDX determination of atomic ratios.

Ratio Number of analyses Anhydrous waste glass N/C N/C + 25g WG 0.86 N10 N10 + 15g N10 + 15Wg KOH
Si/Al 20 29.81 24.12 24.74 25.01 24.52 25.31 31.26 22.77
Ca/Si 20 0.25 0.25 0.20 0.22 0.22 0.24 0.25 0.23

that curing conditions (high and low relative humidity) and the The chemical composition of the AAMs produced and analysed
nature and concentration of the activator impacted mechanical is shown on the CaO-Al2O3-SiO2 ternary diagrams (Figs. 8 and 9). In
strength and microstructure. the activated waste glass, the Si/Al molar ratio for the main reac-
XRD, FTIR and SEM-BSEM--EDX exploration confirmed that the tion product was over 20 in all systems due to the low aluminium
reaction product formed in the alkaline activation of waste glass is content in the starting material (around 2.05%) (Table 3). The cal-
a silicon-high gel. Further to those findings, gel composition does cium content (Ca/Si ranging from 0.20 to 0.25), in turn, was lower
not appear to be affected by curing conditions or activator nature than in the materials obtained in blast furnace slag activation. Sev-
or concentration. On the contrary, microstructural characteristics eral authors have shown the latter to have a Ca/Si molar ratio of
such as moisture-induced cracking intensity do affect gel porosity 0.8–1.2, depending on the activator [33,68]. Due more than likely
and consequently strength. to the chemical composition of the waste glass used as a precursor
This main reaction product was a silicon-high gel identified by (Si-high and Al- and Ca-low), the alkali-activated products were
FTIR (Fig. 5B), in which, with the uptake into its structure of Al3+ scantly affected by the activating conditions.
and Ca2+, the band observed for the paste shifted to lower The microstructure developing in the pastes did appear to be
wavenumbers than the signal for the unhydrated waste glass, from impacted by the activating conditions (Fig. 7). Low (6.5%) humidity
around 1050 cm 1 to 990–1020 cm 1. As a rule, a high purity silica thermal curing yielded a much more cohesive microstructure with
gel generates a SiAOASi vibration band at around 1050– fewer macropores and microcracks and higher mechanical
1100 cm 1. That shift would therefore infer and be explained by strength than high (99%) relative humidity curing. The BET findings
some minor replacement of Si4+ with Al3+ in the main reaction pro- appeared to indicate that the gel forming at low RH had greater
duct, along with the presence of a network modifier such as Na microporosity and specific surface than the gel in the high RH
[45]. The possibility of the co-existence of this Si-high gel with pastes. Both these favourable characteristics would also help
other types of gels (CASAH, CAAASAH, NAAASAH or (C,N)A explain the higher mechanical performance in the former.
AASAH) was slight given the paucity of aluminium and calcium The findings illustrated in Fig. 2 infer that the curing conditions
in the system and the absence of a band at around 960 cm 1 on (relative humidity) had a likewise significant impact on strength
the infrared spectra that would denote the presence of CAAASAH development in the waste glass pastes activated by all the solu-
gel [10]. tions used except KOH. Whilst a number of studies have analysed
Geopolymers are normally non-crystalline materials containing changes in alkali-activated materials properties under different
silicon, aluminium and sodium in the main gel. After the activation curing temperatures [7,69–72], very few have been conducted on
of this waste glass, however, with an aluminium-low chemical the effect of relative humidity on curing [73,74]. One possible
composition (Table 1), the gels generated are likewise Al-low explanation for the higher mechanical strength when the pastes
(Table 3). Many studies have shown that the presence of alu- were heat-cured at a lower relative humidity is the significant role
minium in these systems, expressed as the Si/Al ratio, is directly played by the water (adsorbed, absorbed or both) present in the
related to mechanical strength. Although metakaolin-based this materials and its subsequent evaporation in the generation
geopolymers are characterised by an Si/Al ratio of around 1.8–2.0 of a product with fewer microcracks or micropores [73]. During
[65,66], values of about 16 raise mechanical strength more effec- the dissolution of (very aluminium-low) waste glass in a highly
tively. Mechanical strength is optimal in fly ash systems when Si/ alkaline medium, the water available in the aqueous phase to form
Al  5 [67], whereas in alkali-activated slag systems the optimum Si(OH)4 and Al(OH)4 monomers is partially released during subse-
ranges from 2 to 4 [33,68]. As noted earlier, the low Al content quent polymerisation [73]. The reaction between the glass and the
explains the absence of zeolite-like secondary species in the pastes OH groups always hydrolyses oxygen bridges, partially destroying
prepared in this study, where species such as sodalite hydrate the network [35,75]. In alkaline media, glass may be said to be
(Na4Al3Si3O12OH) or Na-chabazite (NaAlSi2
3H2O) were lacking, depolymerised, resulting in total destruction of its network and
for they are generally found at Si/Al ratios of 1 or 2, respectively. gradual dissolution [35,75]. Low RH leads to the formation of more
350 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

Fig. 8. Chemical composition of waste glass paste hydration products after activation with: A) NaOH/Na2CO3 (N/C); B) NaOH/Na2CO3-glass (N/C + 25g); C) waterglass
(0.86 WG).

cohesive AAMs with a smaller pore size, as the Hg porosimetry and specific surface and adsorb much of the water in the medium [45].
BET findings showed (Figs. 3 and 4). The N10 + 15g sample exhib- In dryer environments, much more compact systems are formed,
ited the lowest total porosity and a higher specific surface with as water is eliminated from the pores, inducing the formation of
fewer <1 nm pores in the gel. Silicon-high gels tend to have a high a less porous, more cohesive gel.
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 351

Fig. 9. Chemical composition of waste glass paste hydration products after activation with: A) NaOH (N10); B) NaOH-glass (N10 + 15g); C) NaOH-waterglass (N10 + 15Wg).

In connection with activator nature and concentration, the solu- favoured the breakage of the SiAOASi bonds in the glass, prompt-
tions that induced the formation of pastes with higher mechanical ing more intense dissolution and generating reaction products that
strength were the ones with a high OH ion content. The OH ions afforded the system suitable mechanical stability [75]. Moreover,
352 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

Fig. 10. KOH-activated waste glass paste: A) BSEM map; B) chemical composition of hydration products; C) SEM micrograph.

the silicon-containing solutions (0.86 WG) delivered higher turn induced lower total porosity and enhanced pore size refine-
strength materials than the NaOH/Na2CO3 solution because silicon ment, raising 1-day mechanical strength.
favoured glass activation. The results obtained after the activation of waste glass with dif-
When the KOH solution was used as the waste glass activator, ferent alkaline activators and under different curing conditions
strength varied very little between the specimens cured at high provided us information at short study ages (1 day) in pastes, thus
and low relative humidity. From the compositional standpoint, leaving a door open in the investigation of these AAMs at more
however, all the systems generated the same main reaction advanced ages and in materials intended for an application in con-
product (silicon-high gel), irrespective of the nature and concen- struction, such as in concrete. Moreover, a need has been identified
tration of the activator (Figs. 8–10). The cation effect on precursor for future studies geared to determining the chemical and physical
dissolution was likewise significant. Cation K+ dissolves less of the stability of pastes generated by the alkaline activation of waste
Ca and Mg present in glass and, having a lower charge density than glass in aggressive media, for Si-high gels are known to exhibit
cation Na+, induces a smaller proportion of secondary products dimensional or chemical instability or both.
[54]. This study revealed differences between sodium- and
potassium-based activators in terms of strength development 5. Conclusions
and microstructure [44]. One of the determining factors in reaction
kinetics is cation size. As Na+ is smaller, it forms zeolites more The main conclusions drawn from the present findings are
readily [54]. listed below.
The solution pH, however, was observed to affect the formation
of the reaction product, with higher values generating a more - Waste glass is apt as a precursor for alkaline cements, for it
cohesive, higher strength gel. The use of solutions with a high alka- develops materials with suitable mechanical strength.
line content (N10, N10 + 15g and N10 + 15Wg) delivered a much - The main reaction product obtained in waste glass activation is
more compact microstructure (more compact hydration products) an alkali-activated material (AAM). The composition of this
(Fig. 6), as confirmed by the significant decline in total porosity in silicon-high, Al- and Ca-low gel is unaffected by activation con-
the systems with highest mechanical performance. Moreover, the ditions or the nature or concentration of the alkaline activator.
more alkaline solutions dissolved the waste glass more effectively, - Curing conditions, and more specifically relative humidity, have
raising system reactivity and favouring the formation of more pri- a significant effect on strength development in alkali-activated
mary reaction product. glass pastes. When cured for 20 h at 85 ± 2 °C and 6.5% relative
The BET study of these systems revealed a correlation between humidity, the pastes developed strength 50–75% greater than
the specific surface of the systems and the RH at which they were when cured for the same time and at the same temperature
cured. At low (6.5%) RH, paste specific surface was higher. That in but at 99% relative humidity.
 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354 353

- High alkaline activator concentrations induce the formation of [26] E.A. Byars, B. Morales-Hernandez, H.Y. Zhu, Waste glass as concrete aggregate
and pozzolan, Concrete 38 (2004) 41–44.
less porous, higher strength pastes with more compact
[27] I.B. Topcu, M. Canbaz, Properties of concrete containing waste glass, Cem.
microstructures. Concr. Compos. 34 (2004) 267–274.
- The absence of zeolite-like secondary reaction products would [28] C.H. Chen, R. Huang, J.K. Wu, C.C. Yang, Waste E-glass Particles used in
be explained by the low Al content in the precursor waste glass. cementitious mixtures, Cem. Concr. Res. 36 (2006) 449–456.
[29] G. Chen, H. Lee, K.L. Young, P.L. Yue, A. Wong, T. Tao, K.K. Choi, Glass recycling
in cement production–an innovative approach, Waste Manage. 22 (2002) 747–
753.
[30] C. Shi, K. Zheng, A review on the use of waste glasses in the production of
Acknowledgements cement and concrete, Resour. Conserv. Recycl. 52 (2007) 234–247.
[31] Z. Xie, Y. Xi, Use of recycled glass as a raw material in the manufacture of
This research was funded by the Ministry of Economy and Com- Portland cement, Mater. Struct. 35 (2002) 510–515.
[32] A. Khmiri, M. Chaabouni, B. Samet, Chemical behaviour of ground waste glass
petitiveness under projects BIA2010-15516 and BIA2013-47876- when used as partial cement replacement in mortars, Constr. Build. Mater. 44
C2-1-P and the Spanish National Research Council under project (2013) 74–80.
PIE201460E06. The authors wish to thank Eduardo Torroja Insti- [33] F. Puertas, M. Torres-Carrasco, Use of glass waste as an activator in the
preparation of alkali-activated slag. Mechanical strength and paste
tute for Construction Science technicians Alfredo Gil, Francisco characterisation, Cem. Concr. Res. 57 (2014) 95–104.
Morales and Patricia Rivilla for their assistance with the laboratory [34] M. Torres-Carrasco, F. Puertas, Waste glass in the geopolymer preparation.
trials. Mechanical and microstructural characterisation, J. Cleaner. Prod. 90 (2015)
397–408.
[35] M. Torres-Carrasco, J.G. Palomo, F. Puertas, Sodium silicate solutions from
References dissolution of glass wastes: statistical analysis, Mater. Construcción. 64 (2014).
[36] M. Torres-Carrasco, M. Tognonvi, A. Tagnit-Hamou, F. Puertas, Durability of
alkali-activated slag concretes prepared using waste glass as alternative
[1] A.O. Purdon, The action of alkalis on blast-furnace slag, J. Soc. Chem. Ind.-Trans.
activator, ACI 112 (2015) 791–800.
Commun. 59 (1940) 191–202.
[37] F. Puertas, M. Torres-Carrasco, M.M. Alonso, Reuse of urban and industrial
[2] V. Glukhovsky, G. Rostovskaja, G. Rumyna, High strength slag-alkaline
waste glass as novel activator for alkali-activated slag cement pastes: a case
cements, in: 7th Int. Congr. Chem. Cem, Paris, 1980, pp. 164–168.
study, in: Handb. Alkali-Activated Cem. Mortars Concr., 2014, pp. 75–110.
[3] V. Glukhovsky, Y. Zaitsev, V. Pakhomow, Slag-alkaline cements and concretes
[38] H.K. Tchakouté, C.H. Rüscher, S. Kong, E. Kamseu, C. Leonelli, Geopolymer
structures, properties, technological and economic aspects of the use, Silic.
binders from metakaolin using sodium waterglass from waste glass and rice
Indus. 48 (1983) 197–200.
husk ash as alternative activators: a comparative study, Constr. Build. Mater.
[4] J. Davidovits, Geopolymer Chemistry and Applications, Institud Géopolymère,
114 (2016) 276–289.
San Quintin, France, 2008.
[39] L.K. Turner, F.G. Collins, Carbon dioxide equivalent (CO2) emissions: a
[5] S.A. Bernal, E.D. Rodríguez, R. Mejia de Gutiérrez, J.L. Provis, S. Delvasto,
comparison between geopolymer and OPC cement concrete, Constr. Build.
Activation of metakaolin/slag blends using alkaline solutions based on
Mater. 43 (2013) 125–130.
chemically modified silica fume and rice husk ash, Waste Biomass
[40] I. Garcia-Lodeiro, E. Aparicio-Rebollo, A. Fernández-Jimenez, A. Palomo, Effect
Valorization 3 (2011) 99–108.
of calcium on the alkaline activation of aluminosilicate glass, Ceram. Int. 42
[6] A. Palomo, M.T. Blanco-Varela, M.L. Granizo, F. Puertas, T. Vazquez, M.W.
(2016) 7697–7707.
Grutzeck, Chemical stability of cementitious materials based on metakaolin,
[41] I. Garcia-Lodeiro, A. Fernández-Jimenez, P. Pena, A. Palomo, Alkaline activation
Cem. Concr. Res. 29 (1999) 997–1004.
of synthetic aluminosilicate glass, Ceram. Int. 40 (2014) 5547–5558.
[7] S. Alonso, A. Palomo, Alkaline activation of metakaolin and calcium hydroxide
[42] B. Walkley, R. San Nicolas, M.-A. Sani, G.J. Rees, J.V. Hanna, J.S.J. van Deventer, J.
mixtures: influence of temperature, activator concentration and solids ratio,
L. Provis, Phase evolution of C-(N)-A-S-H/N-A-S-H gel blends investigated via
Mater. Lett. 47 (2001) 55–62.
alkali-activation of synthetic calcium aluminosilicate precursors, Cem. Concr.
[8] E. Robayo, R.M. de Gutierrez, M. Gordillo, Natural pozzolan-and granulated
Res. 89 (2016) 120–135.
blast furnace slag-based binary geopolymers, Mater. Constr. 66 (2016) e077.
[43] B. Walkley, R. San Nicolas, M.-A. Sani, J.D. Gehman, J.S.J. van Deventer, J.L.
[9] F. Puertas, Cementos de escorias activadas alcalinamente: Situación actual y
Provis, Phase evolution of Na2O–Al2O3 –SiO2 –H2O gels in synthetic
perspectivas de futuro, Mater. Construcción. 45 (1995) 53–64.
aluminosilicate binders, Dalt. Trans. 45 (2016) 5521–5535.
[10] F. Puertas, M. Palacios, H. Manzano, J.S. Dolado, A. Rico, J. Rodríguez, A model
[44] M. Cyr, R. Idir, T. Poinot, Properties of inorganic polymer (geopolymer) mortars
for the C-A-S-H gel formed in alkali-activated slag cements, J. Eur. Ceram. Soc.
made of glass cullet, J. Mater. Sci. 47 (2011) 2782–2797.
31 (2011) 2043–2056.
[45] R. Redden, N. Neithalath, Microstructure, strength, and moisture stability of
[11] F. Puertas, Escorias de alto horno: composición y comportamiento hidráulico,
alkali activated glass powder-based binders, Cem. Concr. Compos. 45 (2014)
Mater. Construcción. 43 (1993).
46–56.
[12] S.A. Bernal, R. San Nicolas, R.J. Myers, R. Mejía de Gutiérrez, F. Puertas, J.S.J. van
[46] U. Avila-López, J.M. Almanza-Robles, J.I. Escalante-García, Investigation of
Deventer, J.L. Provis, MgO content of slag controls phase evolution and
novel waste glass and limestone binders using statistical methods, Constr.
structural changes induced by accelerated carbonation in alkali-activated
Build. Mater. 82 (2015) 296–303.
binders, Cem. Concr. Res. 57 (2014) 33–43.
[47] R. Martinez-Lopez, J. Ivan Escalante-Garcia, Alkali activated composite binders
[13] C. Shi, P. Krivenko, D. Roy, Alkali-Activated Cements and Concretes, Taylor and
of waste silica soda lime glass and blast furnace slag: strength as a function of
Francis, London and New York, 2006.
the composition, Constr. Build. Mater. 119 (2016) 119–129.
[14] A. Palomo, P. Krivenko, E. Kavalerova, O. Maltseva, A review on alkaline
[48] G. Kovalchuk, A. Jiménez-Fernández, A. Palomo, Alkali-activated fly ash: effect
activation: new analytical perspectives, Mater. Construcción. 64 (2014).
of thermal curing conditions on mechanical and microstructural development-
[15] A. Cornejo, C. Leiva, L.F. Vilches, Properties of fly ash and metakaolín based
Part II, Fuel 86 (2007) 315–322.
geopolymer panels under fire resistance tests, Mater. Construcción. 65 (2015)
[49] P.V. Krivenko, G. Kovalchuk, Fly ash based zeolite cements. Innovations and
e059.
developments in concrete materials and construction, in: Challenges Concr.
[16] C. Shi, A. Fernández-Jiménez, Stabilization/solidification of hazardous and
Constr., Dundee, 2002, pp. 123–132.
radioactive wastes with alkali-activated cements, J. Hazard. Mater. 137 (2006)
[50] P.V. Krivenko, G. Kovalchuk, G. Yu, Heat-resistant fly ash based geocements, in:
1656–1663.
Geopolymer 2002, Melbourne, Australia, 2002.
[17] I. Escalante-García, Overview of the potential of urban waste glass as a
[51] G. Kovalchuk, G. Yu, Heat resistant gas concrete based on alkaline
cementitious material in alternative chemically activated binders, in: 14th Int.
aluminosilicate binder, Ukrainian, 2002.
Congr. Chem. Cem., Beijing, 2015, p. 25.
[52] M. Criado, A. Palomo, A. Fernández-Jiménez, Alkali activation of fly ashes. Part
[18] Ecovidrio, Ecovidrio, http://www.ecovidrio.es/, 2015.
1: effect of curing conditions on the carbonation of the reaction products, Fuel
[19] R.G. Pike, D. Hubbard, E.S. Newman, Binary silicate glasses in the study of
84 (2005) 2048–2054.
alkali-aggregate reaction, High Res. Board Bull. 275 (1960) 39–44.
[53] K. Yang, J. Song, Workability loss and compressive strength development of
[20] A. Scmidt, W.H.F. Saia, Alkali-aggregate reaction tests on glass used for
cementless mortars activated by combination of sodium silicate and sodium
exposed aggregate wall panel work, ACI Mater. J. 60 (1963) 1235–1236.
hydroxide, J. Mater. Civ. Eng. Civ. Eng. 21 (2009) 119–127.
[21] C.D. Johnston, Waste glass as coarse aggregate for concrete, J. Test. Eval. 2
[54] A. Fernández-Jiménez, A. Palomo, M. Criado, Alkali activated fly ash binders. A
(1974) 344–350.
comparative study between sodium and potassium activators, Mater.
[22] C. Meyer, S. Baxter, W. Jin, Alkali-aggregate reaction in concrete with waste
Construcción. 56 (2006) 51–65.
glass as aggregate, in: Proc. 4th Mater. Eng. Conf. Mater. New Millenn., Reston,
[55] H. Xu, J.S.J. Van Deventer, The effect of alkali metals on the formation of
1996, pp. 1388–1397.
geopolymeric gels from alkali-feldspars, Colloids Surf., A 216 (2003) 27–44.
[23] C. Meyer, S. Baxter, Use of recycled glass and fly ash for precast concrete, J.
[56] R.T. Hemmings, E.E. Berry, On the glass in coal fly ashes: recent advances,
Mater. Civ. Eng. Civ. Eng. 11 (1999) 89–90.
Mater. Res. Soc. 113 (1987).
[24] C. Pollery, S.M. Cramer, R.V. De la Cruz, Potential for using waste glass in
[57] S.M. Abo-Naf, F.H. El Batal, M.A. Azooz, Characterization of some glasses in the
portland cement concrete, J. Mater. Civ. Eng. Civ. Eng. 10 (1998) 210–219.
system SiO2, Na2O
RO by infrared spectroscopy, Mater. Chem. Phys. 77 (2002)
[25] Z.P. Bazant, G. Zi, C. Meyer, Fracture mechanics of ASR in concretes with waste
846–852.
glass particles of different sizes, J. Eng. Mech. 126 (2000) 226–232.
354 M. Torres-Carrasco, F. Puertas / Construction and Building Materials 139 (2017) 342–354

[58] M.A. Villegas, J.M. Fernández-Navarro, Preparación y caracterización de vidrios [67] A. Palomo, M.W. Grutzeck, M.T. Blanco, Alkali-activated fly ashes A cement for
del sistema CaO-SiO2 por el procedimiento sol-gel, Bol. La Soc. Cerámica Y the future, Cem. Concr. Res. 29 (1999) 1323–1329.
Vidr. 27 (1988) 349–357. [68] F. Puertas, A. Fernández-Jiménez, M. Blanco-Varela, Pore solution in alkali-
[59] I. García-Lodeiro, A. Fernández-Jiménez, M.T. Blanco, A. Palomo, FTIR study of activated slag cement pastes. relation to the composition and structure of
the sol-gel synthesis of cementitious gels: C-S-H and N-A-S-H, J. Sol-Gel. Sci. calcium silicate hydrate, Cem. Concr. Res. 34 (2004) 139–148.
Technol. 45 (2008) 63–72. [69] E. Altan, S.T. Erdoğan, Alkali activation of a slag at ambient and elevated
[60] C.I. Merzbacher, W.B. White, The structure of alkaline earth aluminosilicate temperatures, Cem. Concr. Compos. 34 (2012) 131–139.
glasses as determined by vibrational spectroscopy, J. Non Cryst. Solids. 130 [70] G.F. Huseien, J. Mirza, M. Ismail, M.W. Hussin, Influence of different curing
(1991) 18–34. temperatures and alkali activators on properties of GBFS geopolymer mortars
[61] M. Criado, A. Fernández-Jiménez, A. Palomo, Alkali activation of fly ash: effect containing fly ash and palm-oil fuel ash, Constr. Build. Mater. 125 (2016)
of the SiO2/Na2O ratio, Microporous Mesoporous Mater. 106 (2007) 180–191. 1229–1240.
[62] M. Criado, A. Fernández-Jiménez, A. Palomo, I. Sobrados, J. Sanz, Effect of the [71] P. Nath, P.K. Sarker, Fracture properties of GGBFS-blended fly ash geopolymer
SiO2/Na2O ratio on the alkali activation of fly ash. Part II: 29Si MAS-NMR concrete cured in ambient temperature, Mater. Struct. 50 (2017) 32–44.
Survey, Microporous Mesoporous Mater. 109 (2008) 525–534. [72] A.A. Aliabdo, A.E.M. Abd Elmoaty, H.A. Salem, Effect of cement addition,
[63] J.L. Provis, G.C. Luckey, J.S.J. Van Deventer, Do Geopolymers actually contain solution resting time and curing characteristics on fly ash based geopolymer
nanocrystalline zeolites? A reexamination of existing results, Chem. Mater. 17 concrete performance, Constr. Build. Mater. 123 (2016) 581–593.
(2005) 3075–3085. [73] D.S. Perera, O. Uchida, E.R. Vance, K.S. Finnie, Influence of curing schedule on
[64] D.W. Breck, Zeolite Molecular Sieves: Structure, chemistry and Uses, New the integrity of geopolymers, J. Mater. Sci. 42 (2007) 3099–3106.
York, USA, 1973. [74] A. Tawfik, F.A. El-raoof, H. Katsuki, K.J.D. Mackenzie, S. Komarneni, K-Based
[65] P. Duxson, J.L. Provis, G.C. Luckey, S.W. Mallicoat, W.M. Kriven, J.S.J. Van Geopolymer from metakaolin: roles of K / Al ratio and water or steam curing at
Deventer, Understanding the relationship between geopolymer composition, different temperatures, Mater. Constr. 66 (2016) e081.
microstructure and mechanical properties, Colloids Surf., A 269 (2005) 47–58. [75] J.M. Fernández-Navarro, El vidrio, Editorial CSIC, Madrid, 2003.
[66] P. Duxson, S.W. Mallicoat, G.C. Lukey, W.M. Kriven, J.S.J. van Deventer, The
effect of alkali and Si/Al ratio on the development of mechanical properties of
metakaolin based geopolymers, Colloids Surf., A 292 (2007) 8–20.