Effect of H2O Activity on Zeolite Formation

In an effort to understand the effects of H2O activity on zeolite formation, we have synthesized LTA zeolite using a combination of freezing processes and varying drying temperatures. Sodium aluminate and sodium silicate were used to form LTA zeolite, according to the IZA (International Zeolite Association) protocol. The synthesis steps were modified by adding the precursor frozen process by a rapid liquid nitrogen (−196 °C) treatment or slow conventional freezer treatment (−20 °C). The samples were subsequently sonicated and then dried at 80 °C or 40 °C. X-ray diffraction (XRD) and scanning electron microscopy (SEM) were performed on the samples immediately after the drying process as well as after 2 weeks and 1 month of aging the solid products. The results indicated that LTA zeolite does not form. The silica-alumina precursor after both freezing processes and after being dried at 80 °C showed the presence of sodalite displaying stable behavior over time. Both sets of samples dried at 40 °C and did not show the presence of zeolite immediately after the drying process. However, after 2 weeks, the liquid nitrogen–frozen precursor was characterized by the presence of EMT whereas zeolites never formed in the −20 °C samples. These results suggest that freezing processes differently control the H2O activity during the drying and aging processes in the solid state. Thus, although the precursor chemical composition is the same, the type of zeolite formed is different.

Numerous studies have also been performed to investigate the evolution of this type of zeolite into more thermodynamically stable hydroxisodalite [22][23][24]. In our previous papers, two processes for LTA synthesis were analysed concluding that rapid zeolite crystallization by sonication treatment ensures rapid transformation into more stable sodalite, whereas the slower double-step mechanism of geopolymer transformation into the crystalline phase by a conventional hydrothermal process is responsible for a very slow transformation of LTA into sodalite [25,26].
Zeolites are typically synthesized by a hydrothermal process [27] and, indisputably, the water plays a key role in this process. Water molecules have, in fact, an important structural directing action in zeolite formation [28]. The action of H 2 O molecules in precursor materials and metastable products can also be influenced by their position in restricted geometries (nanopores). Literature data have documented that H 2 O confined in mesopores and nanopores shows properties different from those of common bulk water [29][30][31]. 2 of 16 Generally, it is assumed that, under ambient conditions, ice occurs in the stable hexagonal form (ice I h ) or in the metastable cubic form (ice I c ). However, in the last few years, ice crystallization in nanopores in the form of intercalated cubic and hexagonal layers has been documented [32][33][34]. Jażdżewska and co-workers [35] demonstrated that ice confined in nanopores having pore diameters of 5.9 nm has cubic sequences interlaced with hexagonal sequences, producing a stacking-disordered ice (ice I sd ). According to the authors [35], the unique behavior in the nanoscale environment is due to a competition between the H 2 O-wall and H 2 O-H 2 O intermolecular interactions as well as an unspecified size effect.
Many other research data indicate that the properties of H 2 O in pores <2 nm are different from the properties of H 2 O in pores with a larger size, and it is common to consider that the freezing temperature of H 2 O in pores decreases with a decline in pore size [36]. For nanopores with a pore size of 2 nm, water can be supercooled to −90 • C [37,38], and Jannes et al. [39] showed that homogeneous nucleation of structural H 2 O molecules takes places at~−40 • C inside the microporous structure of zeolites. Bordoninskii and Olov [36] described a multiphase transition of ice structures in natural zeolite (natural powder zeolite composed of 90% clinoptilolite) with pore sizes from 0.2 to 2 nm over a range of temperature from −150 to −100 • C. Some simulation studies carried out using zeolite A indicated that the H 2 O nanoclusters in this type of zeolite are too small to crystallize, thus displaying amorphous ice behavior [29]. The formation of low-density amorphous ice in nano-structural zeolite-template carbon was also documented by Kyakuno et al. [40]. Water (liquid water) "freezing" in zeolite (powder zeolite) is influenced by the type of freezing process used. The vitrification determined by liquid nitrogen is based on a rapid temperature lowering process that does not favor the nucleation of crystal water. Due to this, the technique is largely used to preserve biomaterials without ice or with a very limited amount of ice formation [41][42][43]. The slow cooling due to conventional freezer treatments, instead, determines the crystallization of larger water clusters [44]. In the last few years, the application of ultrasound to liquid freezing has also attracted great interest, and freezing by sonication has been used in many applications such as medical science or food engineering [45][46][47][48]. Literature data have shown that an ultrasound can accelerate mass transfer determining rapid droplet cooling [49]. Sonication causes cavitation with formation of bubbles inside the liquid. These bubbles can affect mass transfer and heat transfer mainly at the liquid surface, thereby facilitating liquid cooling and freezing. Some studies have analysed the variation of ultrasound propagation speed based on the physical properties of a medium that they go through (e.g., ice) [50,51]. This mechanism has also been investigated for crystal ice nucleation in water [52][53][54]. However, according to our knowledge, a study on the effects of an ultrasound on frozen hydrogel-colloidal precursors of zeolite has not been performed.
We performed a preliminary study on the effects of a rapid and a slow freezing treatment (by liquid-nitrogen and conventional freezer, respectively) of zeolite precursor combined with subsequent low (−40 • C) or high temperature (80 • C) drying. The potential effect of brief sonication on frozen samples before the drying processes was also investigated. The study aims to provide insights into the fundamental role of water and its form during zeolite crystallization with the hope of providing new insights into understanding formation mechanisms.

Samples Preparation
An NaOH (Sigma Aldrich-Europe, Darmstadt, Germany, 98+% NaOH) solution was prepared by gently mixing 0.723 g of NaOH in 80 mL of distilled water until it is dissolved. Half of this alkaline solution was then used to dissolve 8.3 g of sodium aluminate (Sigma Aldrich-Europe, Darmstadt, Germany, Al (Al 2 O 3 ):50-56% Na (Na 2 O):40-45%) (Solution A). In addition, 15.48 g of sodium silicate solution (Sigma Aldrich) were dissolved in the second half (Solution B). Finally, solution A was quickly added to solution B under vigorous stirring. Solutions A and B were prepared according to the IZA protocol and the method proposed by Thompson and Huber [1,55]. However, in this study, the next steps to form LTA zeolite were modified as follows: (i) the resulting suspension (combination of Solution A and B) was divided in two parts, (ii) one part was frozen by liquid nitrogen (~−196 • C) (LN) for five minutes (rapid freezing treatment) and the other by conventional freezer treatment (~−20 • C) (CF) for 2 h (slow freezing treatment), (iii) both frozen samples were sonicated using an ultrasonic water bath (240 W, 35 kHz) for 5 min. US temperature was 10 • C, thus melting the samples, iv) both sonicated LN and CF samples were split into two parts and then dried in open vessels at 40 • C and 80 • C for 24 h, respectively. The experiments were also performed by washing LN with ethanol before drying at 40 • C and 80 • C for 24 h (LNE). Finally, additional tests were carried out without any freezing treatment. In detail, the suspension resulting from mixing of Solution A and Solution B was divided into two portions. One was centrifuged and dried at 40 • C and 80 • C for 24 h (ZC), separately. The other portion was directly dried at 40 • C or 80 • C without being centrifuged (Z). The scheme of the experiments is illustrated in Figure 1. the IZA protocol and the method proposed by Thompson and Huber [1,55]. However, in this study, the next steps to form LTA zeolite were modified as follows: (i) the resulting suspension (combination of Solution A and B) was divided in two parts, (ii) one part was frozen by liquid nitrogen (~ −196 °C) (LN) for five minutes (rapid freezing treatment) and the other by conventional freezer treatment (~ −20 °C) (CF) for 2 h (slow freezing treatment), (iii) both frozen samples were sonicated using an ultrasonic water bath (240 W, 35 kHz) for 5 min. US temperature was 10 °C, thus melting the samples, iv) both sonicated LN and CF samples were split into two parts and then dried in open vessels at 40 °C and 80 °C for 24 h, respectively. The experiments were also performed by washing LN with ethanol before drying at 40 °C and 80 °C for 24 h (LNE). Finally, additional tests were carried out without any freezing treatment. In detail, the suspension resulting from mixing of Solution A and Solution B was divided into two portions. One was centrifuged and dried at 40 °C and 80 °C for 24 h (ZC), separately. The other portion was directly dried at 40 °C or 80 °C without being centrifuged (Z). The scheme of the experiments is illustrated in Figure 1.

Sample Characterization
The mineralogical composition of all the synthetic products was determined by X-ray powder diffraction (XRD) using a Rigaku Rint 2200 powder diffractometer (Tokyo, Japan) with Cu Kα graphite monochromatized radiation (30 kV-40 mA). XRD profiles were collected in θ-θ geometry over the angular range 3 • -70 • 2θ, step size 0.02 • , and scan-step time of 3 s. The synthetic products were analysed immediately after the drying process (1 day) as well as 2 weeks and 1 month after. The samples were stored at room temperature (about 20 • C). Table 1 summarizes the codes for different samples, according to the processes used for their synthesis. Morphological observation of products was performed by scanning electron microscopy (SEM) using a Zeiss Supra 40 microscope (Oberkochen, Germany) equipped with energy dispersive X-ray analysis (EDX)  Figure 2 shows X-ray diffraction patterns of the LN 1d and CF 1d samples immediately after drying at 40 and 80 • C. The samples dried at 40 • C (CF 1d-40 and LN 1d-40 ) are characterized by the presence of aluminium silicate and sodium silicate in addition to amorphous material (Figure 2a). The XRD peaks for the samples dried at higher temperatures (CF 1d-80 and LN 1d-80 ) were mainly indexed to sodalite with minor aluminium silicate and sodium silicate as well as amorphous material (Figure 2b).  (LN2w-40) is, instead, characterized by the presence of EMT zeolite, which increased after 1 month (LN1m-40) (Figure 4b and Figure S1). silicate. The XRD pattern of LN dried at 40 • C and analysed after 2 weeks of aging (LN 2w-40 ) is, instead, characterized by the presence of EMT zeolite, which increased after 1 month (LN 1m-40 ) (Figure 4b and Figure S1).     The typical morphology of this newly formed zeolite is displayed in Figure 5a,b. The presence of other zeolites (e.g., sodalite) was confirmed by SEM (Figure 5c). The typical morphology of this newly formed zeolite is displayed in Figure 5a,b. The presence of other zeolites (e.g., sodalite) was confirmed by SEM (Figure 5c).   Figure 8). Finally, Figure 9 shows XRD profiles of ZC and Z samples   (Figures 7b and 8). Finally, Figure 9 shows XRD profiles of ZC and Z samples at low drying temperature. The results demonstrate the synthesis of EMT in Z samples after 2 weeks and 1 month of aging. EMT is not present immediately after drying Z samples at 40 • C (Figure 9a). LTA is the primary newly formed zeolite in centrifuged samples (ZC) (Figure 9b). At higher temperature (80 • C), sodalite and LTA with sharp XRD peaks characterize Z and ZC products, respectively (Figure 10a,b).

Results
Materials 2020, 13, 4780 8 of 16 at low drying temperature. The results demonstrate the synthesis of EMT in Z samples after 2 weeks and 1 month of aging. EMT is not present immediately after drying Z samples at 40 °C (Figure 9a). LTA is the primary newly formed zeolite in centrifuged samples (ZC) (Figure 9b). At higher temperature (80 °C), sodalite and LTA with sharp XRD peaks characterize Z and ZC products, respectively (Figure 10a,b). The right-most column in Table 1 summarizes the mineralogical composition of each synthetic product.      Likely, the higher drying temperature makes the water molecules available, which are still confined in colloidal precursor after ethanol treatment. The data for untreated samples (no freezing or centrifugation) confirm the absence of zeolite immediately after drying at 40 °C. The progressive transformation of aluminosilicate and amorphous material into EMT and LTA takes place after 2 weeks and 1 month (Figure 9a). The results after centrifugation indicate the crystallization of LTA zeolite takes place at 40 °C, and LTA is stable over time (2 weeks and 1 month) (Figure 9b). At 80 °C, sodalite formed in Z and is unchanged after solid-state aging (Figure 10a). These data are in accord with the results obtained at the same temperature for the LN sample. The results after centrifugation (ZC) reveal the crystallization of LTA zeolite at 80 °C as well as its presence after 2 weeks and 1 month (Figure 10b). Our hypothesis is that centrifugation has an influence on water available for the zeolite synthesis, conditioning LTA zeolite synthesis. Finally, no conclusion can be drawn on the effect of sonication.  The right-most column in Table 1 summarizes the mineralogical composition of each synthetic product.

Discussion
These preliminary results reveal the complex behavior of aluminosilicate precursor of LTA zeolite when subjected to combined freezing and drying processes. The data show that LTA zeolite does not form when the aluminosilicate precursor is rapidly frozen by liquid nitrogen and dried at 40 or 80 • C nor when it is dried at the same temperatures after conventional freezer treatment. These results suggest that the type of freezing treatment affects the availability of water in the precursor during the drying process (also based on the drying temperature), thereby influencing the type of zeolite formed and its stability after aging in the solid state. In detail, XRD profiles for both CF and LN immediately after drying at 40 • C are characterized by the presence of peaks indexed to aluminium silicate and sodium silicate. This is due to rapid precipitation in the supersaturated solution. Short time and lower heating temperature (40 • C) do not facilitate the synthesis of LTA zeolite (Figure 2a). This is also confirmed by the large presence of amorphous material. The data for the process at 80 • C show sodalite formation in both CF and LN samples (CF 1d-80 and LN 1d-80 ) (Figure 2b), and LTA zeolite does not form. SEM images are displayed in Figure 3a,b. After 2 weeks of aging in the solid state at room temperature, LTA zeolite did not form in CF samples initially dried at 40 • C (Figure 4a) whereas LN 2w-40 is characterized by the presence of newly formed EMT zeolite detectable after 1 month in LN 1m-40 (Figure 4b). This is well detectable in Figure 5a,b showing the typical EMT morphology. Both CF and LN samples that initially dried at higher temperature (80 • C) are characterized by the presence of sodalite (Figure 6a,b) after 2 weeks and 1 month of aging in the solid state.
We do not assume that the nature and the aggregation state of the aluminosilicate phases precipitated during the freezing process condition the behavior of the different samples. Our speculative hypothesis to explain the different behavior between CF and LN samples dried at a low temperature (40 • C), which is based on the assumption that rapid freezing by liquid nitrogen (−196 • C for few minutes) does not favor the nucleation of crystalline ice in the aluminosilicate precursor matrix, but, instead, leads to the formation of amorphous ice [29,40]. The presence of amorphous ice leads to a higher H 2 O activity in the matrix, which controls the transformation of aluminosilicate and amorphous material into EMT zeolite in 2 weeks (LN 2w-40 ). Single crystals of sodalite also formed ( Figure 5c). We propose that the process can be explained by considering partial hydrolysis at room temperature in the presence of enhanced H 2 O availability [56]. Conversely, the conventional freezer treatment is a slow process (2 h) occurring at −20 • C. This allows the formation of crystalline clusters [57] into the large amorphous material, thus making water molecules unavailable for zeolite crystallization at 40 • C and after aging in the solid state (CF 2w-40 and CF 1m-40 ). The presence of sodalite in both CF and LN samples dried at 80 • C (Figures 2b and 6b) indicates that different processes take place, making water molecules available to form this more stable mineral. The presence of this newly formed phase after 2 weeks and 1 month of solid-state aging are in accordance with literature data [58,59]. Our hypothesis on the role of the H 2 O form (crystalline cluster or amorphous ice) in type and stability of zeolite formed is indirectly confirmed by additional experiments performed by washing the LN precursor with ethanol (before drying) or by repeating the synthesis without freezing. In the first case, the results indicate that, at 40 • C, no zeolite formed immediately after drying as well as after 2 weeks and 1 month of solid-state aging (LNE 1d-40 , LNE 2w-40 , and LNE 1m-40 ) (Figure 7a). This is due to the partial replacement of water molecules by ethanol [60], confirming the importance of water availability in zeolite formation. After drying at 80 • C, sodalite formed (Figure 7b) despite the morphology of this newly formed mineral not being well organized, and the crystals are a few hundred nm in size ( Figure 8).
Likely, the higher drying temperature makes the water molecules available, which are still confined in colloidal precursor after ethanol treatment. The data for untreated samples (no freezing or centrifugation) confirm the absence of zeolite immediately after drying at 40 • C. The progressive transformation of aluminosilicate and amorphous material into EMT and LTA takes place after 2 weeks and 1 month (Figure 9a). The results after centrifugation indicate the crystallization of LTA zeolite takes place at 40 • C, and LTA is stable over time (2 weeks and 1 month) (Figure 9b).
At 80 • C, sodalite formed in Z and is unchanged after solid-state aging (Figure 10a). These data are in accord with the results obtained at the same temperature for the LN sample. The results after centrifugation (ZC) reveal the crystallization of LTA zeolite at 80 • C as well as its presence after 2 weeks and 1 month (Figure 10b). Our hypothesis is that centrifugation has an influence on water available for the zeolite synthesis, conditioning LTA zeolite synthesis. Finally, no conclusion can be drawn on the effect of sonication.

Conclusions
In this study, the effects of freezing on zeolite synthesis from an LTA precursor was analysed with the aim to evaluate the role of H 2 O and its form as amorphous or crystalline ice in determining zeolite type and long-term stability.
Based on the results, our speculative hypothesis is that the formation of amorphous ice rather than crystalline ice in the aluminosilicate precursor controls the synthesis of different zeolites both immediately and after solid-state. The presence of amorphous ice in the precursor by a rapid liquid nitrogen freezing process (−196 • C) raises the possibility of the hydrolysis mechanism in the samples dried at 40 • C and zeolite evolution to a more stable form at room temperature. The formation of ice crystalline clusters by slow conventional freezer treatment at −20 • C, instead, is responsible for low water molecules availability in the samples, making no zeolite crystallization over time. In all the samples, the drying process of the precursor at 80 • C makes available water molecules due to a melting process of ice, forming sodalite with a stable behavior up to 1 month.
The results also display that the action of short time sonication after the freezing treatment is negligible.
In our opinion, these types of experiments make an important contribution to understanding crystallization mechanisms of minerals forming in extreme environmental conditions. However, further investigation is needed.