Pressure-Induced Geopolymerization in Alkali-Activated Fly Ash

The present study investigated geopolymerization in alkali-activated fly ash under elevated pressure conditions. The fly ash was activated using either sodium hydroxide or a combination of sodium silicate solution and sodium hydroxide, and was cured at 120 ◦C at a pressure of 0.22 MPa for the first 24 h. The pressure-induced evolution of the binder gel in the alkali-activated fly ash was investigated by employing synchrotron X-ray diffraction and solid-state 29Si and 27Al MAS NMR spectroscopy. The results showed that the reactivity of the raw fly ash and the growth of the zeolite crystals were significantly enhanced in the samples activated with sodium hydroxide. In contrast, the effects of the elevated pressure conditions were found to be less apparent in the samples activated with the sodium silicate solution. These results may have important implications for the binder design of geopolymers, since the crystallization of geopolymers relates highly to its long-term properties and functionality.


Introduction
Geopolymers are a cementitious material belonging to a group of alkali-activated binders, which can be synthesized by alkaline activation of aluminosilicate-rich precursors [1,2].Coupled with the global CO 2 emissions associated with the production of Portland cement, this binder system is viewed as a potential sustainable cement [3][4][5].Geopolymers are known to exhibit excellent durability performance [6,7], providing a potential application to various fields in place of ordinary Portland cement [8].
Fully-reacted geopolymeric gels are totally amorphous [9].However, they may present a nanostructural analogue to zeolite, possibly due to the similarities between the chemical compositions of both materials [10,11].The nanostructure of geopolymers consists of alkali-aluminosilicate gel, often abbreviated as N-A-S-H (where N denotes sodium, which is the most commonly employed alkali), where both Si and Al are in tetrahedral coordination [2].A recent study conducted by Brant et al. [12] revealed that Q 4 (4Al), Q 4 (3Al), Q 4 (2Al), and Q 4 (1Al) Si units are predominantly present in sodium aluminosilicate gels, which are charge-balanced by three to four coordinated Na + ions and H 2 O molecules, or six coordinated Al atoms.This study also identified that the presence of Al 3+ in sites with lower symmetry provided charge-balancing capacity [12].
The durability and other physicochemical properties of geopolymers are attributed to the nanostructure of the aluminosilicate gel.Binders derived from alkali-activated fly ash present excellent resistance against a number of chemical degradation phenomena, which are known to be devastating in the case of Portland cement [13][14][15].The aluminosilicate gel formed in geopolymers possesses negatively-charged surfaces and a high specific surface area, which are ideal for the adsorption of cationic pollutants such as heavy metals (Pb 2+ ) [16,17] and radionuclides (Cs + , Sr 2+ ) [8,18,19].
The nucleation of zeolites with the ageing of geopolymers can lead to reduced mechanical strength and durability, as observed by Silva and Crenstil [20].Thus, it is crucial to assess the factors that lead to the transformation of amorphous geopolymeric gel into crystalline zeolites to ensure the long term properties of geopolymers.The effect of elevated pressure during the manufacture of geopolymers and enhanced zeolite yield has been extensively investigated in previous studies.The strength of fly ash-based geopolymers varies according to the curing conditions and precursor/activator composition.While the strength of those cured at an elevated temperature and ambient pressure can be as high as 40 MPa at 28 days [21,22], and have the potential to develop further at later ages [23], the geopolymer specimens, in which a significant amount of aluminosilicate gel is transformed into crystalline zeolite at an elevated pressure, show limited strength (i.e., as low as 2 MPa at 28 days [24]).While the strength decline in geopolymers via crystallization is a generally acknowledged deterioration phenomenon of geopolymers, its nanostructural characterization remains limited.
Zeolite formation in alkali-activated materials can be affected by a number of factors, i.e., the Ca content in the precursors can lead to precipitation of Ca-chabazite [24].Faujasite and sodalite are the two zeolite crystals commonly observed in fly ash-based binders, in which inert minerals such as quartz and mullite can be consumed as reactants for zeolite formation in the presence of elevated pressure [16].Hydrothermal treatment is a conventional method for artificially synthesizing zeolites from silica and alumina source materials [25].Under suitable pressure and humidity conditions, pure phase zeolites can be synthesized with high yield.Curing geopolymers under elevated pressure conditions at an early age can be a robust way to study the factors controlling this transformation, and to predict their long-term properties.Hence, geopolymers were cured under elevated pressure conditions in this study to explore the effects of different activating solutions.

Materials
The chemical composition of the fly ash (supplied by Hadong thermal power plant in South Korea) used in this study is shown in Table 1.The alkali-activated fly ash paste samples were synthesized using 9 M NaOH and a sodium silicate solution (SiO 2 = 29 wt%, Na 2 O = 10 wt%, H 2 O = 61 wt%).Two series of samples were prepared to investigate the effect of the activators as follows: H-series activated with NaOH only, and S-series activated with the sodium silicate solution with a SiO 2 /Na 2 O of 1.0.The Na 2 O dosage of both series was 7.5 g per 100 g of fly ash.The H 2 O/fly ash was fixed at 0.338 in both series to isolate the effect of water content.Note that all the tested samples were produced from the same batch of fly ash.

Sample Preparation
The solids and liquids were mechanically stirred for five minutes at room temperature.The fresh samples were poured into vial molds (25 mm diameter and 50 mm height).The molds were placed into a chamber, and the samples were cured at 120 • C at a pressure of 0.22 MPa for the first 24 h.The reference samples were cured at 120 • C at atmospheric pressure.The samples were cured at room temperature after the first 24 h of curing.Note that a pressure of 0.22 MPa was reached at this temperature, which was deemed effective in terms of the crystallization in the geopolymer [24,26].The experimental parameters including mixture proportion and curing conditions employed in this study were adapted from the relevant literature [8,16].This level of pressure is known to onset the crystallization of geopolymer gels [24,26], and can be experienced in repository conditions.

Experimental Methods
The pressure-induced geopolymerization in alkali-activated fly ash was explored using synchrotron X-ray diffraction (XRD) and solid-state magic angle-spinning nuclear magnetic resonance (MAS NMR) spectroscopy at 24 h, 7and 28 days.The samples were ground and passed through a sieve of 64 µm for analysis.The synchrotron XRD was conducted at beamline 5 A at the Pohang Accelerator Laboratory, South Korea.The conditions for the data collection were: an angular range of 2-20 • , a scan rate of 0.01 • /s and a wavelength of 0.69265 Å.The conditions for the solid-state 27 Al and 29 Si MAS NMR spectra collected using 600 MHz solid-state NMR (14.1 T, unity INOVA, Agilent Technologies: Santa Clara 95050, CA, USA) at the KBSI Western Seoul Center were as follows: - 29 Si MAS NMR spectra: data acquisition at a transmitter frequency of 119.14 MHz, a pulse length of 30 • (2.2 µs), a spinning rate of 10 kHz, and a relaxation delay of 22 s, using a 5 mm HX-CPMAS probe and a 5 mm zirconia rotor.The chemical shifts were referenced to TMS (0 ppm). - 27 Al MAS NMR spectra: data acquisition at a transmitter frequency of 156.32 MHz, a pulse length of 30 • (1.8 µs), a spinning rate of 22 kHz, and a relaxation delay of 2 s, using a 2.5 mm HX-CPMAS probe and a 2.5 mm low Al zirconia rotor.The chemical shifts were referenced to aqueous AlCl 3 (0 ppm).
The 27 Al MAS NMR spectra of the H-series samples are shown in Figure 2a.The asymmetry in the obtained spectra was caused by the resonance of the quadrupolar nuclei, hence, the line width of the spectra was largely affected by the quadrupole moment of the nucleus, and the size of the molecule [28].Although this technique was able to resolve the local environment of Al species, accurate descriptions of quadrupolar coupling effects in these materials are currently unavailable in the literature.The spectrum of the raw fly ash showed resonance centered at 52 and 1 ppm due to the presence of tetrahedral and octahedral Al sites, respectively, where the tetrahedral site corresponds to the Al in the tetrahedral aluminosilicate glass network [21] and the octahedral site of the mullite and mullite-like glass [14,29].Upon alkaline activation, the spectra of the H-series samples both with and without pressure showed a significant reduction in the intensity of the resonance in the octahedral site, indicating the consumption of the glassy phase.The resonance at 58 ppm was observed in these samples, which is due to the presence of tetrahedral Al in the N-A-S-H gel [21], showing a shift from 54 ppm from the resonance of the raw fly ash.The spectrum of the sample with pressure, however, showed much narrower width at this resonance, indicating that the tetrahedral Al environment constructed herein was more clearly defined and less amorphous (i.e., crystallization).In addition, the intensity at the octahedral site was found to be lower in the sample where pressure was applied during the curing process, which implies that the reactivity was relatively higher in this sample.showing a shift from 54 ppm from the resonance of the raw fly ash.The spectrum of the sample with pressure, however, showed much narrower width at this resonance, indicating that the tetrahedral Al environment constructed herein was more clearly defined and less amorphous (i.e., crystallization).In addition, the intensity at the octahedral site was found to be lower in the sample where pressure was applied during the curing process, which implies that the reactivity was relatively higher in this sample.The 29 Si MAS NMR spectra of the H-series samples are shown in Figure 2b.The Si environment in both the raw and alkali-activated fly ash consists of an aluminosilicate framework, resonating at −87, −92, −96, −103, and −108 ppm due to the presence of Q 4 (4Al), Q 4 (3Al), Q 4 (2Al), Q 4 (1Al), and Q 4 (0Al), respectively [21,30,31].The resonance of Q 4 (0Al) due to the Si-O-Si linkage in the crystalline (i.e., quartz) or glassy phase [32,33] was relatively lower in the sample with pressure.This implies that the reactivity of the fly ash vastly improved in elevated pressure conditions.The sample without pressure mostly displayed the highest intensity at the Q 4 (4Al) site and an increased resonance at the The 29 Si MAS NMR spectra of the H-series samples are shown in Figure 2b.The Si environment in both the raw and alkali-activated fly ash consists of an aluminosilicate framework, resonating at −87, −92, −96, −103, and −108 ppm due to the presence of Q 4 (4Al), Q 4 (3Al), Q 4 (2Al), Q 4 (1Al), and Q 4 (0Al), respectively [21,30,31].The resonance of Q 4 (0Al) due to the Si-O-Si linkage in the crystalline (i.e., quartz) or glassy phase [32,33] was relatively lower in the sample with pressure.This implies that the reactivity of the fly ash vastly improved in elevated pressure conditions.The sample without pressure mostly displayed the highest intensity at the Q 4 (4Al) site and an increased resonance at the Q 4 (3Al) site at 28 days.In contrast, the sample with pressure showed the highest intensity at the Q 4 (2Al) and Q 4 (1Al) sites throughout all ages, implying that the elevated pressure was effective for the formation of Si-rich reaction products even at an early age.The resonance at these two sites was much narrower due to the higher degree of crystallization in the sample with pressure.
Sustainability 2018, 10, x FOR PEER REVIEW 5 of 10 Q 4 (3Al) site at 28 days.In contrast, the sample with pressure showed the highest intensity at the Q 4 (2Al) and Q 4 (1Al) sites throughout all ages, implying that the elevated pressure was effective for the formation of Si-rich reaction products even at an early age.The resonance at these two sites was much narrower due to the higher degree of crystallization in the sample with pressure.
Figure 2. (a) 27 Al and (b) 29 Si MAS NMR spectra of H-series samples.black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Sodium Silicate-Series Samples
The synchrotron XRD patterns of the S-series samples are shown in Figure 3.The XRD patterns of the samples with and without pressure showed the peaks corresponding to the presence of crystalline phases contained in the raw fly ash, i.e., quartz, mullite and hematite.The absence of peaks attributed to zeolite minerals indicates that the growth of zeolite crystals (or zeolite with a sufficient  27 Al and (b) 29 Si MAS NMR spectra of H-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Sodium Silicate-Series Samples
The synchrotron XRD patterns of the S-series samples are shown in Figure 3.The XRD patterns of the samples with and without pressure showed the peaks corresponding to the presence of crystalline phases contained in the raw fly ash, i.e., quartz, mullite and hematite.The absence of peaks attributed to zeolite minerals indicates that the growth of zeolite crystals (or zeolite with a sufficient degree of crystallinity) was significantly hindered by the dissolved silica, even with the elevated pressure.The effect of pressure on the reactivity of the fly ash was seen to be much less in comparison with the H-series; the intensity of the octahedral site in the 27 Al MAS NMR spectrum of the sample with pressure was only slightly reduced (Figure 4a); and the resonance at the Q 4 (0Al) in the 29 Si MAS NMR spectra was similar to that of the sample without pressure.pressure.The effect of pressure on the reactivity of the fly ash was seen to be much less in comparison with the H-series; the intensity of the octahedral site in the 27 Al MAS NMR spectrum of the sample with pressure was only slightly reduced (Figure 4a); and the resonance at the Q 4 (0Al) in the 29 Si NMR spectra was similar to that of the sample without pressure.   Al and (b) 29 Si MAS NMR spectra of S-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Discussion
Previous studies reported that NaOH-activated metakaolin and slag generally show a higher tendency of crystallization over time in comparison with those activated with sodium silicate [34,35].A similar result was observed in this study, in which the H-series samples exhibited a higher extent of crystallization in elevated pressure conditions.An ageing phenomenon of geopolymers is generally accepted to accompany crystallization [36] in a similar manner to that of hydrated calcium aluminate cement, in which hydrates (mainly calcium aluminate octa-and deca-hydrate) lose structural water, increasing the pore volume [37].The difference between the ageing phenomena of the two binder systems is the presence of water; neither geopolymer gels nor zeolite crystals contain  27 Al and (b) 29 Si MAS NMR spectra of S-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Discussion
Previous studies reported that NaOH-activated metakaolin and slag generally a higher tendency of crystallization over time in comparison with those activated with sodium silicate [34,35].A similar result was observed in this study, in which the H-series samples exhibited a higher extent of crystallization in elevated pressure conditions.An ageing phenomenon of geopolymers is generally accepted to accompany crystallization [36] in a similar manner to that of hydrated calcium aluminate cement, which deca-hydrate) lose increasing the pore volume [37].The difference between the ageing phenomena of the two binder systems is the presence of water; neither geopolymer gels nor zeolite crystals contain as much water as calcium aluminate cement hydrates, hence the ageing effect of geopolymers is not associated with the release of structural water from the binder gel, but instead is related to the density of the tetrahedral per unit volume which is significantly reduced by crystallization from amorphous gel into zeolite [36].Such phenomena were validated in the case of the alkali-activated metakaolin, which experienced a significant loss in compressive strength after 28 days of ageing at 95 • C [36].Moreover, the strength loss was approximately in an inverse relationship with the content of zeolite P [36].
To date, only a few studies have looked at the ageing effect of alkali-activated fly ash (i.e., [36,38]).For instance, fly ash activated with sodium silicate solutions with varying compositions (7% SiO 2 and 7-15% Na 2 O by the mass of fly ash) showed a higher content of faujasite in a system with lower Na 2 O, and a higher content of zeolite P in a system with higher Na 2 O [36].Considering that zeolite A [39] and faujasite [40] succeed to zeolite P, the presence of these zeolite crystals in the H-series sample with elevated pressure may correlate to its likely durability performance.Specifically, employing NaOH as an activator for fly ash could be plausible for use as a structural element where long-term durability is mandatory, and where mechanical properties are of less concern (i.e., molecular sieve [41] or radioactive waste immobilization [24,26]).On the other hand, the zeolite phases formed in these samples may not be considered to have followed a general phenomenon, since the synthesis of zeolites is significantly influenced by temperature In particular, the formation of zeolite Na-P1 or sodalite is reportedly preferred at the temperature applied in this study (120 • C) [43].

Conclusions
The present study explored pressure-induced geopolymerization in alkali-activated fly ash using synchrotron XRD and solid-state MAS NMR spectroscopy.Two series of samples were prepared, one activated with NaOH (H-series) and another activated with a sodium silicate solution (S-series).An initial curing stage at an elevated 24 h.The synchrotron XRD results showed that significant growth of zeolites (zeolite A, zeolite Na-P1, chabazite, sodalite) in the H-series samples was induced by pressure, while the S-series samples remained amorphous in identical conditions.The NMR spectra of the samples suggest that the reactivity of the fly ash, or geopolymerization in the H-series samples was enhanced by pressure to a greater extent than that of the S-series samples.
The obtained results showed that the elevated pressure was effective in enhancing the reactivity of the fly ash and growth of the zeolite crystals in the H-series samples.The geopolymer gel in the sample with pressure was relatively rich in Si.In contrast, the effects of elevated pressure were much less clear in the S-series samples, showing that the dissolved silica that was supplied by the activator inhibited the crystallization of the binder gel.The crystallization of the binder gel can be viewed in two different aspects.First, the ageing of geopolymer gels tends to induce microstructural changes that may affect their mechanical properties [44], while zeolite formation plays an important role as an ion adsorbent [24,26].
Further research is necessary to reveal the crystallization of fly ash-based geopolymers and their mechanical properties.The crystallization of this binder system was found to be less likely in comparison with that of metakaolin.The results obtained in this study suggest that alkali-activated fly ash can undergo significant crystallization when activated with NaOH under elevated pressure.In addition, the phase transformation of zeolite formed in this binder system may be an important topic of future studies, since it is highly related to the ion sorption kinetics and capacity.The influence of zeolite crystallization on the mechanical properties of these materials could be an important factor, which should be explored in forthcoming studies.

Figure 2 .
Figure 2. (a)27 Al and (b)29 Si MAS NMR spectra of H-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Figure 4 .
Figure 4. (a)27 Al and (b)29 Si MAS NMR spectra of S-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Figure 4 .
Figure 4. (a)27 Al and (b)29 Si MAS NMR spectra of S-series samples.The black solid and blue dotted lines indicate the spectra of the reference sample and the sample with pressure, respectively.

Funding:
This study was supported by the National Research Foundation (NRF) of the Korean government (Ministry of Science & ICT) [Grant Number: 2017R1A5A1014883] through the Smart Submerged Floating Tunnel System Research Center.

Table 1 .
Chemical composition of fly ash obtained using XRF.