3.1. Characterization of deactivated ZSM-5 and fly ash
3.1.1. Phase identification and quantification
The crystalline solid phase was determined by mineralogical analysis; and the XRD patterns of deactivated ZSM-5 and fly ash are illustrated in Fig. 2. The existence of a broad diffuse peak between 15-30° in the fly ash pattern shows that the amorphous phase, calcined silica and alumina are the main components with some crystalline phases, such as mullite and quartz. Note that chemical analysis (Table 2) shows that deactivated ZSM-5 contains a small amount of alumina (Al2O3); hence, a broad diffuse peak between 20 and 30° suggests that the amorphous phase was mainly derived from silica and small amount of alumina. Moreover, the predominant crystalline phases of deactivated ZSM-5 are quartz, ZSM-5 and cristobalite.
The phase quantitative analysis of deactivated ZSM-5 and fly ash was carried out by using X'pert HighScore Plus software. The amorphous content and absolute content of the crystalline phases were calculated by an external standard method. The quantitative analysis of deactivated ZSM-5 and fly ash are listed in Table 6 and Table 7. The calculation results show that the fly ash contains 10.43% mullite, 4.16% quartz and 85.41% amorphous. Additionally, deactivated ZSM-5 consists of 59.63% ZSM-5, 5.12% cristobalite, 5.36% quartz and 29.89% amorphous.
Table 6
Phase composition of deactivated ZSM-5.
Phase (Deactivated ZSM-5)
|
Phase composition (g/100 g)
|
ZSM-5 (ICSD:201183)
|
59.63
|
Cristobalite SiO2 (ICSD:047221)
|
5.12
|
Quartz SiO2 (ICSD:100341)
|
5.36
|
Amorphous
|
29.89
|
Total
|
100
|
Table 7
Phase composition of fly ash.
Phase (Fly ash)
|
Phase composition (g/100 g)
|
Quartz SiO2 (ICSD:89278)
|
4.16
|
Mullite Al1.69Si1.22O4.85 (ICSD:43297)
|
10.43
|
Amorphous
|
85.41
|
Total
|
100
|
3.1.2. FTIR
Fig. 3 presents the FTIR pattern of the raw materials. Absorption bands could be observed at 3439 and 3436 cm-1, which are associated with the stretching vibrations of O-H groups. The bands located near 1629 cm-1 and 1630 cm-1 are assigned to the existence of interlayer adsorbed water, which may be caused by the leftover in the samples or the higher content of water in the raw materials, which is depicted as the molecular water bending vibrations. The absorption bands at approximately 1428 and 1429 cm-1 are ascribed to antisymmetric stretching vibrations of CO32-. The bands near 1093 and 1110 cm-1 correspond to the antisymmetric stretching vibration of T-O-T (T = Al and Si). The bands located near 556 and 548 cm-1 are attributed to the stretching vibration of T-O. The absorption bands near 462 and 464 cm-1 are derived from the asymmetric tensile vibration of vibration of Si-O. In the deactivated ZSM-5 pattern, the shoulder peak at 796 cm-1 is related to the traces of cristobalite, which are the crystalline phases from the deactivated ZSM-5. The absorption band at approximately 876 cm-1 is ascribed to the stretching vibration of Si-O.
3.2. XRD
The XRD patterns of the geopolymer pastes are shown in Fig. 4. It is clearly observed that the sharp peaks in M1.4T25, M1.4T50, M1.4T75, M2.0T25, M2.0T50 and M2.0T75 did not significantly change. A broad hump at 18-30° 2θ was present in all geopolymer pastes, which was the characteristic reflection of amorphous geopolymers. The superposition of amorphous or semicrystalline phases may be responsible for the formation of a broad amorphous hump: (1) the insufficiency reaction and dissolution of the amorphous phase that ranges from 18-30° 2θ, which is derived from fly ash and deactivated ZSM-5, may occur in the geopolymerization reaction[46]. (2) The 2θ of the newly formed amorphous geopolymer generally ranges from 18 to 30°[47]. This finding was consistent with the current view that geopolymers are amorphous to semicrystalline aluminosilicate materials[47-50]. Moreover, the existence of crystalline minerals of mullite, quartz and cristobalite still existed in the XRD results of geopolymer pastes, which were derived from fly ash and deactivated ZSM-5, respectively, indicating that crystalline phases essentially did not participate or were only partially involved in the geopolymerization reaction but filled the geopolymer samples in the form of inactive fillers. In addition, crystalline components such as ZSM-5, which were confirmed in the XRD pattern of deactivated ZSM-5 (Fig. 2), were not detected in the final geopolymer products. One possible reason was that the concentration or content of crystalline minerals ranges from extremely low to not detectable. Another possible explanation was that the desilication of deactivated ZSM-5 under strong alkaline conditions causes the dissolution and collapse of the zeolite framework, which eliminates the crystallinity of ZSM-5[51]. The peaks of crystalline zeolite phases of gismondine were observed in geopolymer pastes. In a geopolymer cementitious matrix, geopolymer-like gels, through nucleation and growth, were considered to be a dominant means of forming gismondine[52].
3.3. Compressive strength
Fig. 5 shows the effect of curing temperature and modulus of water glass on the compressive strength of the geopolymer. Notably, varying the modulus of water glass caused diverse mechanical performance; thus, this modulus is an important variable influencing the performance of geopolymers. For the two groups of geopolymers, a modulus of water glass of 2.0 and 1.4 was utilized to fabricate geopolymers. When the polymers were thermally cured at 25 ℃; the compressive strengths were 2.5 and 3.2 MPa, respectively, when they were thermally cured at 50 ℃, the compressive strengths were 4.4 and 5.2 MPa, respectively. This finding indicates that the water glass modulus did not significantly affect the compressive strength of the geopolymer when cured at 25 ℃ and 50 ℃. Although the low-modulus solutions produced more reaction products, the geopolymers did not form hardened structures in a short period due to the retarded setting of geopolymer slurry. However, the compressive strength of the geopolymer reached 6.3 MPa with a water glass modulus of 2.0. The compressive strength was 8.5 MPa with a water glass modulus of 1.4 and increased nearly one-third when cured thermally at 75 ℃. The results show that the compressive strength of the geopolymer increases with a decrease in water glass modulus at three different temperatures. In general, low-modulus water glass solutions lead to extensive dissolution of raw materials and promote the activation of silica, alumina and amorphous materials so that silicon and aluminium leach more in the water glass solution, increasing the extent of the geopolymerization reaction and resulting in the generation of larger quantities of reaction products[30, 53, 54]. Hence, low-modulus alkali solutions enhanced the compressive strength of the final geopolymer pastes.
Based on these data (Fig. 5), the curing temperature is more significant than the modulus of water glass in dictating the compressive strength. Moreover, the different compressive strengths caused by various curing temperatures can be confirmed, that is, the compressive strength increased with an increase in curing temperature. It is commonly agreed that the curing temperature has a significant impact on the properties of geopolymers[29, 30, 53, 55, 56]. In this study, increasing the curing temperature yields a geopolymer with higher strength, irrespective of the water glass modulus. Note that the compressive strength that is thermally cured at 25 ℃ is significantly lower than that of the geopolymer at 50 and 75 ℃. The main reasons for this finding are presented as follows: (a) at low temperatures, the geopolymerization rate is low with gradual development of the compressive strength. (b) As confirmed by the XRD finding, a large quantity of crystalline phases and an unreactive amorphous phase derived from raw materials are present in the geopolymer, which may affect the rate of geopolymerization. In summary, increasing the heat curing temperature from 25 ℃ to 75 ℃ imparts a higher degree of geopolymerization reaction and facilitates the production of aluminosilicate gels, which is the main binding phase in particle bonding that accelerates the development of compressive strength[30, 57]. Moreover, a higher curing temperature also facilitates the formation of a hard structure in a short time.
3.4. Bulk density
Fig. 6 shows the effect of curing temperature and modulus of water glass on the bulk density of the geopolymer. The viscosities of the two geopolymer pastes are shown in Table 8. The bulk densities of samples M1.4T25, M1.4T50, M1.4T75, M2.0T25, M2.0T50 and M2.0T75 were 910, 859, 703, 866, 712 and 627 kg/m3. Notably, increases in both curing temperature and water glass modulus reduced the bulk density of the geopolymer. The following critical reasons explain the influence of water glass modulus on the bulk density of geopolymers: (1) the higher viscosity of low-modulus water glass was pronounced in hindering the foaming effect of deactivated ZSM-5; and (2) as previously discussed, more abundant aluminosilicate gels were produced by a lower modulus water glass, which converted the free water to bond water and hence increased the viscosity of geopolymers (Table 8)[58], causing a detrimental effect on the bulk density. During the geopolymerization reaction, strong alkaline solutions dissolved the structure of deactivated ZSM-5, and therefore, the air was released from the zeolite cavity. However, higher viscosity limits the foaming process; hence, the lower modulus of water glass yields geopolymer samples with higher bulk density at a given temperature. In addition, it was determined that the bulk density significantly decreased with an increase in curing temperature, indicating that a more porous hardened structure was formed. An elevated curing temperature promotes the thermal expansion of deactivated ZSM-5 and accelerates the release rate of air from the zeolite cavity. Therefore, the higher the curing temperature is, the lower the bulk density, irrespective of the water glass modulus. In this study, however, there was variation in the relationship between density and compressive strength, which contradicts a finding of previous studies: geopolymers with a higher bulk density exhibit a higher compressive strength. This trend is mainly attributed to the delayed coagulation and hardening of geopolymers at low temperatures; thus, compressive strength is negatively correlated with density. In general, it is preferable to set the curing temperature above 50 ℃.
Table 8
Viscosity of the two geopolymer pastes.
Sample
|
M1.4
|
M2.0
|
Viscosity (mPa·s)
|
10521
|
9038
|
3.5. FTIR
The FTIR spectra of geopolymer pastes with various curing temperatures and water glass moduli are shown in Fig. 7. The bands located near 3450 cm-1 and 1640 cm-1 are assigned to the existence of interlayer adsorbed water, which may be caused by the leftover in the samples or the higher content of water in the raw materials, which is depicted as molecular water bending vibrations and symmetric stretching of O-H bands. The band of all geopolymer spectra at approximately 2356 cm−1 is related to the formation of HCO3−, indicating that CO2 reacted with a large amount of sodium hydroxide. The bands detected at approximately 1430 cm-1 and 877 cm-1 are related to the different vibration modes of n3 carbonate [CO3]2- and n2 carbonate [CO3]2-, respectively[59]. Furthermore, the traces of quartz were identified in spectra located near 790 cm-1 which derived from incompletely reacted fly ash particles and are related to the symmetrical stretching vibration of Si-O (Q1) bonds[60]. The band near 580 cm-1 are associated with vibration of Si-O bands. The absorption bands near 460 cm-1 can be assigned to the vibration of Si-O-Si bands.
The broad humps between 800 and 1300 cm-1 were associated with the superposition of reaction products and precursors. As reported in previous studies[61-65], this band was regarded as the T-O asymmetric stretching band, which is widely employed to investigate the amorphous gel structure in geopolymers. Due to the low amount of Ca2+ contained in the raw material, in low-calcium environments, the broad and asymmetrical stretching band of T–O, where T is Si or Al between 1030 and 1043 cm−1, is typical of tridimensional frameworks of N-A-S-H-type gels from geopolymers. The investigated reaction products should be notably consistent with the SEM–EDS analysis.
To analyse the main band in question at varying water glass moduli and curing temperatures, the position shifts of the T-O band were identified, as shown in Fig. 8. The T-O band moves from lower to higher wavenumbers with low water glass moduli. On the other hand, the curing temperature also significantly changed the T-O band position. The main reason for the decrease in wavenumber with an increase in modulus is that the binding energy of Al-O is lower than that of Si-O bands, which causes Al to more readily dissolve from solid precursor structures[62]. Therefore, the substitution of Si by Al in the reaction products led to the T-O band position located at lower wavenumbers. Furthermore, higher curing temperatures and lower water glass moduli significantly affected the dissolution of the solid precursor, desilication process of deactivated ZSM-5 and polymerization of the main reaction product N-A-S-H gels. In particular, the quantity of available silica (especially monomers) significantly changes with different moduli[61, 65]. This change is reflected in the shift of the main T-O band in the infrared spectrum to higher wavenumbers, suggesting that a larger amount of Si was introduced to the gel network and that the high Si content in cementitious systems enables a Si-rich structure of the reaction products. The Si is obtained from two main sources: (a) desilication process of deactivated ZSM-5 and (2) Si-O groups separated in raw fly ash. This finding further illustrates the higher degree of polymerization of the reaction products.
3.6. TG/DTA analysis
The thermogravimetric and differential thermogravimetric curves of the geopolymer samples are presented in Fig. 9. The DTA curve of all geopolymer samples changed mainly in the range of room temperature to 200 ℃ and tended to be stable at temperatures above 560 ℃. The mass loss from room temperature to 80 ℃ on the TG curves was ascribed to physically bound water. The main mass loss fell within the range of 80 to 160 ℃ after decomposition of the reaction product in all geopolymer samples. As described in the EDS analysis in Section 3.7, the main reaction products N-A-S-H in all geopolymer binders were associated with mass loss in this range. Additionally, this range was consistent with N-A-S-H gels in cementitious systems[62, 63]. The negligible decline curves observed between 160 and 500 ℃ were relevant to the polymerized/condensed water evaporated from Si-OH and Al-OH groups[66]. Moreover, the obvious DTA peak range of 500-560 ℃ is caused by the oxidation and combustion of unburned coal in the residual fly ash particles[67]. The weak peak at 800 ℃ was attributed to the carbonated phase.
It is assumed that the main binding phase in geopolymer samples comprised N-A-S-H gels and that the quantity can be calculated by the mass loss percentage of the thermogravimetric (TG) curve (Fig. 9) in the range of 80-160 ℃. As a result, the mass loss from 80-160 ℃ of all geopolymer samples is given in Fig. 10. Notably, the curing temperature and water glass modulus have a significant effect on the formation of gel products. Samples with a lower water glass modulus have obviously higher mass loss at the same curing temperature. On the other hand, the curing temperature was an important factor that determined the quantity of gel product formation. The trends in the mass loss of gel products showed agreement with the compressive strength of geopolymers.
3.7. SEM/EDS analysis
The microstructure of geopolymer samples at different curing temperatures and water glass moduli and selected representative EDS points are given in Fig. 11 and Fig. 12.
Fig. 11 shows the microstructure of representative samples at low magnification. Notably, the curing temperature and water glass modulus have a substantial influence on the pore structure of the geopolymer samples. The relationship between pore structure and curing temperature and modulus will be discussed in detail in Section 3.8. Fig. 12 shows the microstructure of all samples at high magnification. It was evidently observed that microcracks existed in the porous and nonhomogeneous microstructure. The cracks may be generated in two ways: (I) shrinkage of geopolymers due to evaporation of water during curing and (II) cracks induced by loading in the compressive strength test[47]. These cracks may have a negative impact on the mechanical properties of the samples.
Fig. 12(a) and (b) show micrographs of samples M2.0T25 and M1.4T25. It was clearly seen that the morphology of the paste was loose and had less density. Fig. 12(c), (d), (e) and (f) display a uniform and denser microstructure, indicating a relatively higher degree of geopolymerization with an increase of curing temperature.
The EDS spectra were employed to analyse the chemical compositions of the reaction products in geopolymers. The EDS points within the binder region were located far from the unreacted particles[68]. This was clearly shown by all EDS data, which mainly consists of Si, Al, Na and O, with trace Fe, Ti and Ca. Considering the typical chemical bonds observed in the FTIR analysis, the main reaction product was the N-A-S-H gel. The structure of N-A-S-H gel is often considered associated with "zeolites" and is closely related to the formation of crystalline zeolites in the system[69, 70]. In addition, N-A-S-H gel is a key product in geopolymers, which are widely employed in the synthesis of zeolites. Impurities such as Fe, Ti and Ca in the raw materials, which have a certain impact on the geopolymerization process, were shown in the EDS spectrum[48].
In this study, the effects of curing temperature and water glass modulus on the N-A-S-H gel were reflected by its Si/Al ratio. With the same curing temperature, the sample with a low water glass modulus has a higher Si/Al ratio due to abundant Si induced by a decrease in water glass modulus, suggesting that low-modulus water glass solutions lead to extensive dissolution of raw materials. Additionally, Si was also incorporated into the structure with an increase in curing temperature. The Si/Al ratio increases with an increase in curing temperature for a given water glass modulus, indicating that a higher curing temperature promotes the dissolution of Si species from raw materials. The results showed that a low water glass modulus and high curing temperature yielded a higher degree of polymerization of the reaction products, which showed agreement with the FTIR analysis (Section 3.5).
3.8. MIP
Single pore micro, meso- and macropore structures exhibit excellent properties. However, single pore structure materials also have intrinsic deficiencies. For example, certain adsorbent materials that have a single pore channel always possess a limited adsorption capacity due to the small internal pores, so the large molecules formed in the pore channel cannot quickly escape. By comparison, the porous structure retains the adsorption and exchange properties of the original pore structure while creating new pore structures that can increase the pore volume, surface area, porosity, and active adsorption sites of the system and thus enable it to have a wide range of adsorption and exchange capabilities. This section focuses on the pore structure analysis of porous zeolite precursors.
The pore size range of 0.002 to 0.1 μm was very fine and hardly visible in the foamed geopolymer systems. The nitrogen adsorption method is therefore not suitable for the determination of the pore structure of foamed geopolymer systems. Mercury intrusion porosimetry (MIP) in this experiment has been widely applied to determine pore size distributions larger than 0.1 μm. However, when applying this technique in the analysis of geopolymer pore structures, the following issues should be considered. The pressure exerted by intruding mercury will destroy the microstructure of the reaction products and produce high pore volumes. In addition, MIP does not measure the total pore volume but measures the percolated pore volume as MIP assumes that all pores are cylindrical. Hence, the technique was a comparative measurement tool that was employed in the comparison of various geopolymer systems.
Table 9 lists the porosity of the geopolymer samples. The porosities of samples M1.4T25, M1.4T50, M1.4T75, M2.0T25, M2.0T50 and M2.0T75 were 37.71%, 38.2%, 44.84%, 42.53%, 47.01% and 49.68%, respectively. Notably, an increase in water glass modulus and curing temperature causes an increase in the porosity of the samples. Many studies have reported on the relationship between the porosity and the compressive strength of cementitious materials[71-73]. In this study, however, the results contradict the previous research finding that geopolymer with low porosity yields higher compressive strength. Although geopolymer samples with low porosity exhibit a denser pore structure, as mentioned in Section 3.3, longer curing times are required to develop compressive strength due to delayed coagulation and hardening of geopolymers at low temperatures.
The pore volume and pore size distribution of the geopolymer determined by mercury intrusion porosimetry are given in Fig. 13. For the given 1.4 and 2.0 moduli of the water glass series, the geopolymer with a higher curing temperature displays a higher cumulative intrusion, indicating the formation of a porous structure. On the other hand, the cumulative intrusion also increases with an increase in water glass modulus from 1.4 to 2.0. Additionally, as shown in Fig. 13(b), with an increase in curing temperature and water glass modulus, the pore structure becomes porous, which is expressed as a shift in the critical pore size towards larger values on the pore size distribution curve.
As shown in Fig. 13(c), with an increase in water glass modulus and curing temperature, the ratio of pores (<5 μm) decreased, the ratio of pores (5-30 μm) increased, and the ratio of pores (>30 μm) decreased. It can be concluded that a higher water glass modulus and curing temperature yield a more uniform distribution of pores and that the redistribution of pore sizes between 5 µm and 30 µm may explain the change in compressive strength. During the geopolymerization reaction, the desilication of deactivated ZSM-5 induced more mesopores on the surface of zeolite, which enabled the alkali solution to enter the microporous inlet. As the alkali solution continued to damage the micropores and dissolve the framework of zeolite (as determined by the XRD result), the air was released from the collapse of the zeolite cavity. Furthermore, the thermal expansion of deactivated ZSM-5 is improved by the higher curing temperature. These are the two main reasons for the change in pore structure caused by the incorporation of deactivated ZSM-5. In addition, the different moduli of water glass and curing temperature have a significant effect on the porous structure of geopolymers and can regulate and optimize the pore structure.
Table 9
Porosity of geopolymer.
Sample NO.
|
M1.4T25
|
M2.0T25
|
M1.4T50
|
M2.0T50
|
M1.4T75
|
M2.0T75
|
Porosity (%)
|
37.71
|
42.53
|
38.2
|
47.01
|
44.84
|
49.68
|