*Article* **Thermophysical Properties of Kaolin–Zeolite Blends up to 1100** ◦**C**

**Ján Ondruška <sup>1</sup> , Tomáš Húlan <sup>1</sup> , Ivana Sunitrová 1 , Štefan Csáki <sup>2</sup> , Grzegorz Łagód 3 , Alena Struhárová <sup>4</sup> and Anton Trník 1,5,\***


**Abstract:** In this study, the thermophysical properties such as the thermal expansion, thermal diffusivity and conductivity, and specific heat capacity of ceramic samples made from kaolin and natural zeolite are investigated up to 1100 ◦C. The samples were prepared from Sedlec kaolin (Czech Republic) and natural zeolite (Nižný Hrabovec, Slovakia). Kaolin was partially replaced with a natural zeolite in the amounts of 10, 20, 30, 40, and 50 mass%. The measurements were performed on cylindrical samples using thermogravimetric analysis, a horizontal pushrod dilatometer, and laser flash apparatus. The results show that zeolite in the samples decreases the values of all studied properties (except thermal expansion), which is positive for bulk density, porosity, thermal diffusivity, and conductivity. It has a negative effect for thermal expansion because shrinkage increases with the zeolite content. Therefore, the optimal amount of zeolite in the sample (according to the studied properties) is 30 mass%.

**Keywords:** kaolin; zeolite; kaolinite; clinoptilolite; thermal expansion; thermal diffusivity; thermal conductivity; specific heat capacity

#### **1. Introduction**

Ceramic production is known as one of the oldest sectors of human activity. Commonly used materials in the production of traditional ceramics are kaolin, illitic clays, feldspars, quartz, and Al2O3. Traditional ceramics are usually used in the building industry, such as bricks and tiles. The partial substitution of raw materials with waste or new materials can improve the properties of ceramic products and also reduce the cost of their production [1]. Nowadays, many published studies deal with partial substitution of traditional input raw materials for production of ceramics by waste materials such as fly or bottom ash [2–12], waste glass [13–16], waste calcite [17], etc. Húlan et al. [7] determined that a higher Young's modulus was reached after sintering with a lower amount of fly ash. The Young's modulus and the flexural strength decreased linearly with the amount of fly ash. Hasan et al. [14] observed an increase in the compressive strength and a decrease in water absorption of the samples with the addition of waste glass. They also found that the partial replacement of natural clay in a brick with waste soda-lime glass made the brick production sustainable and eco-friendly. Kováˇc et al. [17] showed that a high content of waste calcite may double the energy consumption during the creation of anorthite at a

**Citation:** Ondruška, J.; Húlan, T.; Sunitrová, I.; Csáki, Š.; Łagód, G.; Struhárová, A.; Trník, A. Thermophysical Properties of Kaolin–Zeolite Blends up to 1100 ◦C. *Crystals* **2021**, *11*, 165. https:// doi.org/10.3390/cryst11020165

Academic Editor: Magdalena Król Received: 4 January 2021 Accepted: 3 February 2021 Published: 7 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

temperature of 950 ◦C and also that the waste calcite had a slight positive effect on the final contraction of samples.

In this study, kaolin and natural zeolite were used for the preparation of blends. Kaolin is the most commonly used material in the ceramic and paper industries, cosmetics, medicine, etc. It mainly consists of mineral kaolinite (Si2Al2O5(OH)4) with about 80 mass%, depending on its origin and type [18]. Kaolinite belongs to the group of phyllosilicates and it crystalizes in the triclinic crystal system. Kaolinite has a 1:1 sheet structure composed of tetrahedral [Si2O5] <sup>2</sup><sup>−</sup> sheets and octahedral [Al2(OH)4] 2+ sheets. It has a pseudo-hexagonal symmetry with the lattice parameters *a* = 0.515 nm, *b* = 0.895 nm, *c* = 0.740 nm, *α* = 91.68◦ , *β* = 104.87◦ , and *γ* = 89.9◦ [19–21]. The distance between layers is 0.72 nm [22]. During heating up to 1100 ◦C, three important thermal reactions take place in kaolinite. The first is the dehydration, which occurs in the temperature interval from 35 to 250 ◦C, where physically bound water is removed from the surface of crystals and pores [23]. The second reaction is the dehydroxylation of kaolinite [19,21,24–26], where the chemically bound water escapes from its structure and kaolinite is transformed into a new phase, metakaolinite. It may be described by the following equation:

$$\rm Al\_2O\_3 \cdot 2SiO\_2 \cdot 2H\_2O \rightarrow Al\_2O\_3 \cdot 2SiO\_2 + 2H\_2O\_{(g)}.\tag{1}$$

This reaction takes place in the temperature interval from 450 to 700 ◦C. Metakaolinite (Al2Si2O7) has a similar structure to kaolinite, but the lattice parameter *c* is changed to 0.685 nm (it is smaller by about 0.055 nm). The structure of metakaolinite does not contain OH<sup>−</sup> ions, and the distance between the layers is shorter than in kaolinite. In addition, metakaolinite is more defective and less stable than kaolinite [24,25,27]. The third and last process begins above 925 ◦C and is connected with the transformation of metakaolinite to an Al-Si spinel, γ–Al2O3, and amorphous SiO2, according to the following equation [28,29]:

$$\text{2(Al}\_2\text{O}\_3\cdot2\text{SiO}\_2\text{)} \to \text{2Al}\_2\text{O}\_3\cdot3\text{SiO}\_2 + \text{SiO}\_{\text{2(amorphous)}}.\tag{2}$$

Moreover, above the temperature of 700 ◦C, solid-state sintering occurs, which is a high-temperature technological process that transforms individual ceramic particles into a compact polycrystalline body. For traditional, kaolin-based ceramics (earthenware, stoneware or pottery), solid-state sintering occurs when the powder compact is densified entirely in the solid state [24,25,30,31].

Zeolites are microporous, hydrated crystalline aluminosilicates which are porous and widely used due to their structure and absorption properties. Many studies deal with natural or synthetic zeolites [32–41]. Usually, zeolites contain alkaline metals or metals of alkaline earth and frequently (e.g., in the case of clinoptilolite) crystallize in a monoclinic crystal system [32]. The three-dimensional structure of zeolites consists of tetrahedral silicate and aluminum, which are interconnected by oxygen atoms. The charge of their structure is negative and this charge can balance between monovalent and divalent cations [33,34]. Zeolites are porous and they are widely used due to their absorption properties (for example, in agriculture, ecology, the rubber industry, the building industry, households, and medicine [35–38]). Several important processes occur in zeolites during heating. Above the temperature of 100 ◦C, physically bound water escapes from the crystal surface and pores (capillaries) [39]. During heating up to 900 ◦C, the infrared spectral features attributed to the Si (Al)-O stretching and bending vibration modes do not show significant differences from the features for unheated (raw) zeolite. These spectral results are consistent with the fact that the three-dimensionally rigid crystal structure of zeolite is more stable than the layer structure of phyllosilicates [40]. Above the temperature of 1000 ◦C, the structure of zeolite is definitely destroyed and an amorphous phase is formed [21,33,41].

Our previous study [42] concerned the thermal expansion and Young's modulus of samples made from kaolin and zeolite. The samples were not studied in-situ; instead, they were preheated at different temperatures from room temperature up to 1100 ◦C, and

selected properties were measured after cooling at room temperature. The next study [43] focused on comparing the thermal expansion of the kaolin–zeolite and illite–zeolite samples. Our previous paper [39] aimed at conducting a thermogravimetric analysis and differential scanning calorimetry of a kaolin–zeolite sample. The aim of this paper was to estimate the influence of natural zeolite in the kaolin–zeolite samples on their thermophysical properties, such as the thermal expansion, thermal diffusivity and conductivity, and specific heat capacity, during heating up to 1100 ◦C and to determine the possible usage of natural zeolite in ceramic materials.

#### **2. Materials and Methods**

Samples were prepared from Sedlec kaolin (Czech Republic) and natural zeolite (Nižný Hrabovec, Slovakia). The major mineral in Sedlec kaolin is kaolinite (77.8%). In addition, there are impurities such as mica clay (17.4%) and quartz (1.5%) [44]. Natural zeolite mainly contains mineral clinoptilolite (58.2%) from the group of heulandite and has impurities such as cristobalite (12.2%), illite with mica and feldspar (albite) (9.6%), quartz (0.7%), and also amorphous phase (19.3%) [45]. The chemical compositions of the Sedlec kaolin and natural zeolite are given in Table 1.

**Table 1.** The chemical compositions of the Sedlec kaolin and natural zeolite (in mass%).


The samples were prepared as follows. Kaolin was partially replaced with natural zeolite in the amounts of 10, 20, 30, 40, and 50 mass%. Pure kaolin and zeolite samples were also prepared. The studied samples were labeled as KZ10, KZ20, KZ30, KZ40, and KZ50, according to the natural zeolite content, whereas the pure kaolin and zeolite samples were labeled as SLA and ZEO (see Table 2). The kaolin pellets were crushed and milled to pass a 100-µm sieve. Zeolite was used as a powder, passing a 50-µm sieve. After these procedures, the powders were mixed with deionized water to obtain a plastic mass. Cylindrical samples with a diameter of 14 mm were extruded from this mass. Then, the samples were dried in open air until an equilibrium of moisture was reached (from 1.4 to 2.9 mass% of the physically bound water). The dry samples were cut to the lengths needed for the analyses.

**Table 2.** The compositions of the samples made from Sedlec kaolin and natural zeolite (in mass%).


Differential thermal analysis (DTA) and thermogravimetry (TG) of the compact samples (∅14 × 16 mm) with a mass of 3.5 g were performed by means of a Derivatograph 1000 analyzer (MOM Budapest, Budapest, Hungary) [46], in which a pressed alumina reference sample with similar dimensions to the studied sample was used. Thermodilatometry (TDA) was carried out using a horizontal pushrod alumina dilatometer [47] on samples with dimensions of ∅14 × 35 mm. All measurements were performed in the temperature interval from 30 to 1100 ◦C in static air atmosphere at a heating rate of 5 ◦C·min−<sup>1</sup> .

Differential scanning calorimetry (DSC) was performed on a Netzsch DSC 404 F3 Pegasus apparatus (NETZSCH Holding, Selb, Germany) in dynamic argon atmosphere with a flow rate of 40 ml/min. Al2O<sup>3</sup> crucibles with a lid and powder samples with mass of ~30 mg were used. The temperature increased linearly with the heating rate of 5 ◦C/min and in the temperature interval from 30 up to 1100 ◦C. In order to determine the enthalpy of reactions, a baseline of the tangential type was selected.

The bulk density was calculated from thermogravimetric and thermodilatometric results to obtain its actual values during firing, according to the following equation:

$$
\rho = \rho\_0 \frac{\Delta m / m\_0}{\left(1 + \Delta l / l\_0\right)^3} \,, \tag{3}
$$

where *ρ*<sup>0</sup> is the bulk density of green samples at room temperature.

The open porosity was calculated with the help of the experimentally determined bulk density and matrix density. The bulk density was obtained from the volume and mass of the cylindrical samples. The matrix density was measured by means of helium pycnometry (Pycnomatic ATC, Thermo Fisher Scientific, Waltham, MA USA).

Measurements of the thermal diffusivity (*a*), thermal conductivity (*λ*), and specific heat capacity (*cp*) of samples were performed by means of the flash method using a Netzsch LFA 427 LaserFlash apparatus (NETZSCH Holding, Selb, Germany) in the temperature interval from 30 to 1000 ◦C at a heating rate of 5 ◦C/min and in nitrogen atmosphere with a flow rate of 100 mL/min. The dimensions of samples were ∅12.5 × 2.5 mm. The samples were covered by graphite on both sides before measurements. Measuring the heat capacity of a sample required an additional measurement of a reference material with a known heat capacity and density. The basis of this method is in the application of a laser pulse with the same parameters as the measured and reference samples. In this way, the same amount of heat is provided to both samples. Next, the limit of adiabatic temperature is calculated for both measurements. The specific heat capacity of measured sample is calculated according to the formula:

$$\mathcal{L}\_p = \frac{m\_R \ c\_{pR} \Delta T\_{\infty \text{R}}}{m \ \Delta T\_{\infty}} \, ^\circ \tag{4}$$

where *m<sup>R</sup>* is the mass of the reference sample (in software, the mass of a sample is calculated from its bulk density and dimensions), *cpR* is the heat capacity of the reference sample, ∆*T*∞*<sup>R</sup>* is the adiabatic temperature of the reference sample after the amount of thermal energy is received, *m* is the mass of the studied sample, and ∆*T*<sup>∞</sup> is the adiabatic temperature of the studied sample after the same amount of thermal energy as for reference sample is received.

Microstructure observations were carried out by means of a scanning electron microscope (FEI QuantaTM FX200, Thermo Fisher Scientific, Waltham, MA, USA) in low vacuum mode (100 Pa) with an accelerating voltage of 10 kV on the compact polished raw samples, and the samples were heated at 1100 ◦C.

#### **3. Results and Discussion**

#### *3.1. Differential Thermal Analysis*

The DTA results of the studied samples are shown in Figure 1. Three significant peaks are visible. The first peak (up to 300 ◦C) corresponds to the process of dehydration, where the liberation of physically bound water from pores and surface of crystals occurs [23]. This peak is endothermic and its magnitude increases with the amount of zeolite. The second peak (from 450 to 700 ◦C) is also endothermic and belongs to the dehydroxylation of kaolinite [19,48]. During this reaction, the chemically bound water is evaporated, and this causes the structure of kaolinite to transform into metakaolinite. The last peak (from 940 to 1030 ◦C) is exothermic and has been interpreted as the result of the formation of an Al-Si spinel or of spinel and/or mullite by [28] and by many papers extensively discussed in [31]. The magnitudes of last two peaks decrease with the amount of zeolite in the samples. In natural zeolite, only one reaction is observed and it is endothermic (up to 450 ◦C). This

peak corresponds to dehydration, where the physically bound water evaporates from pores and surface of crystals [21].

**Figure 1.** Differential thermal analysis (DTA) of the studied samples.

#### *3.2. Differential Scanning Calorimetry*

The DSC results of the kaolin–zeolite samples are shown in Figure 2. Three peaks can be observed, as it was in the DTA results (see Figure 1). The first endothermic peak (from 30 to 200 ◦C) represents the liberation of physically bound water [23]. The second endothermic peak (from 400 to 600 ◦C) corresponds to the dehydroxylation of kaolinite [19,48]. The third peak (from 950 to 1000 ◦C), which is exothermic, is related to the crystallization of high-temperature phases, as indicated before [28,31]. In zeolite and natural clinoptilolite, only two peaks are observed in the curve. The first is the endothermic peak corresponding with dehydration. This means that physically bound water evaporates from the pores and surface of crystals. The second peak is also endothermic, and this reaction begins above 850 ◦C, when the structure of clinoptilolite is definitely destroyed and an amorphous phase is formed [33,41].

The influence of zeolite in the samples is visible on the enthalpy of reactions (dehydroxylation of kaolinite and Al-Si spinel formation) occurring in the kaolin–zeolite samples

during thermal treatment (see Figure 3). The enthalpies were determined from the DSC results (peaks) (see Figure 2). A tangential baseline, which most realistically represents the progress of reactions, was used. The first studied reaction was the dehydroxylation of kaolinite, which is associated with the evaporation of chemically bound water [19,48]. The enthalpy of this reaction decreases linearly with the amount of zeolite in the samples from 248.9 (sample SLA) to 128.1 J/g (sample KZ50). The second reaction was the transformation of metakaolinite into an Al-Si spinel [28]. The values of enthalpy also decrease linearly with the zeolite content, from 58.8 (sample SLA) to 34.4 J/g (sample KZ50).

**Figure 3.** The enthalpy of the studied samples during the dehydroxylation of kaolinite (grey) and crystallization of spinel (black).

As can be seen from the enthalpy results for both reactions, the enthalpies decrease along with the amount of kaolinite. Therefore, it can be concluded that zeolite has no significant influence on both reactions in the studied samples.

#### *3.3. Thermogravimetric Analysis*

− The relative mass changes of the kaolin–zeolite samples are shown in Figure 4. Two significant mass losses occur. The first loss is in the temperature interval from 30 to 250 ◦C and corresponds to the process of dehydration, during which physically bound water evaporates [23]. The process of dehydration is the least significant for sample SLA (1.56%) and increases with the amount of zeolite (3.98% for sample KZ50). The second mass loss, in the temperature interval from 450 to 700 ◦C, corresponds to the dehydroxylation of kaolinite [19,48]. The mass loss for sample SLA is 11.35% and it decreases with the amount of zeolite (6.12% for sample KZ50). The mass loss of the zeolite sample decreases continuously up to 800 ◦C and reaches 11.27%. Then, the mass loss remains almost constant. During the dehydroxylation, the mass loss is only 0.98%. This is because the dehydroxylation does not occur in zeolite. The reason is that zeolite does not have OH<sup>−</sup> ions in the matrix structure. The DTA, DSC, and TG results of kaolin and natural zeolite are very similar to the results presented in [21,26,49]. In these studies, the mass losses of the kaolinite subgroup and clinoptilolite are in good agreement with our results. For kaolinite, it is in the interval from 11.8% to 13.31% (we obtained 11.35%, but the kaolinite content in kaolin was only 77.8%), and for clinoptilolite, it is 9.54% at 1000 ◦C (we achieved 11.27%). The obtained difference can be explained by a different content of water in the prepared sample from kaolin or natural zeolite. Moreover, the studied zeolite was not pure clinoptilolite as it also contained the amount of cristobalite, illite with mica, and feldspar.

**Figure 4.** Relative mass change of the studied samples.

The results of the relative mass change of the kaolin–zeolite samples at temperatures of 800, 900, and 1000 ◦C are plotted in Figure 5. It is visible that the mass loss of the samples increases with temperature and decreases with the amount of zeolite. At a temperature of 1000 ◦C, it ranges from 14.3% to 12.6%. The trend of decrease is identical for all three selected temperatures and it is almost linear up to 40 mass% of zeolite. Then, the mass loss for samples KZ40 and KZ50 is comparable (the difference is only 0.04%).

**Figure 5.** Relative mass change of the studied samples at 800, 900, and 1000 ◦C.

#### *3.4. Thermodilatometric Analysis*

The relative thermal expansion of the kaolin–zeolite samples is shown in Figure 6. In the temperature interval from 30 to 250 ◦C, the process of dehydration occurs, where physically bound water escapes from the pores and surface of crystals. The length of all samples, except the zeolite sample, increases linearly and reaches about 0.07%. The length of the zeolite sample decreases and its shrinkage is about 0.14%. The expansion of samples decreases with the amount of zeolite. The next process occurring in the studied samples starts at about 500 ◦C, when the chemically bound water escapes and the structure of kaolinite is transformed into metakaolinite [19,48]. The structure of metakaolinite is

similar to the structure of kaolinite. The lattice parameters *a* and *b* remain the same, but the parameter *c* is changed to 0.685 nm. The distance between layers is shorter (about 0.055 nm) than in kaolinite. In addition, it is more defective and less stable than kaolinite [19]. Therefore, the shrinkage of samples in the temperature interval from 500 to 700 ◦C is observed. Nevertheless, after the dehydroxylation is finished, the contraction continues up to 950 ◦C due to the sintering process [30]. As a result of the reactions above ~950 ◦C, there is a rapid contraction of the produced bodies. In the temperature interval from 700 to 1100 ◦C, the sintering process occurs as well. Therefore, this shrinkage is caused by both processes. The total shrinkage of the kaolin–zeolite products increases with the amount of zeolite from 2.59% for sample SLA to 3.91% for sample KZ50. Similar behavior of thermal expansion of kaolin samples was also obtained in [50,51]. The shrinkage of kaolin samples reached 2.5% and 3.2% at 1100 ◦C, respectively, which is in good agreement with our results. −

**Figure 6.** Relative thermal expansion of the studied samples.

Different results were obtained for the zeolite sample because the structure of zeolite does not contain free OH<sup>−</sup> ions (zeolite is just a hydrate). The liberation of the physically bound water from the zeolite sample proceeds until up to 850 ◦C (the shrinkage reaches 1.82%). Above the temperature of 850 ◦C, very intensive sintering and the formation of a glassy phase occur. The total shrinkage of the zeolite sample at 1100 ◦C reaches 14.59%. Dell'Agli et al. [52] also studied the thermal expansion and mass loss of different types of clinoptilolite, but only up to 700 ◦C. They found out that the final shrinkage was in the interval from 0.6% to 3%. The shrinkage of our zeolite sample reached 1.2% at 700 ◦C. Nevertheless, the trend of the measured curves is similar.

The linear shrinkage of the kaolin–zeolite samples at different temperatures is shown in Figure 7. The results show that the linear shrinkage at temperatures of 800 and 900 ◦C with the increasing amount of zeolite is almost constant (the differences are very small, about 0.15%). The curves at the temperatures of 1000 and 1100 ◦C exhibit a different trend. The shrinkage increases with the amount of zeolite, but the samples with 10, 20, and 30 mass% of zeolite reach almost the same values of shrinkage. Above 30 mass% of zeolite, the linear shrinkage increases. Those results show that the most intensive changes occur when the Al-Si spinel crystallization is finished and the sintering process starts.

#### *3.5. Bulk Density*

− − − − − − The results of the bulk density of the kaolin–zeolite samples are shown in Figure 8. The bulk density decreases with the amount of zeolite and also with the temperature increase to 650 ◦C. The bulk density for the green SLA sample is 1464 kg·m−<sup>3</sup> , and for the green KZ50 sample, it amounts to 1340 kg·m−<sup>3</sup> . The first significant decrease is up to 250 ◦C, which is caused by the liberation of physically bound water [23]. Then, the bulk density decreases almost linearly until the dehydroxylation of kaolinite starts (at about 500 ◦C) [19,48]. The structural changes (transformation of kaolinite into metakaolinite) cause the decrease in the bulk density. This decrease occurs due to an intensive mass loss (11.35% for sample SLA) (see Figure 4) and contraction (0.68% for sample SLA) (see Figure 6). During dehydroxylation, the bulk density decreases from 135 kg·m−<sup>3</sup> for sample SLA to 50 kg·m−<sup>3</sup> for sample KZ50. After dehydroxylation, the bulk density slightly increases for all studied samples up to 950 ◦C. Then, a sharp increase occurs due to the transformation of metakaolinite into the spinel phase [28]. Above 1000 ◦C, the bulk density increases only slightly. At 1100 ◦C, the values of the bulk density are lower than for the green samples. The differences are from 11 kg·m−<sup>3</sup> for sample KZ50 up to 106 kg·m−<sup>3</sup> for sample SLA, which means that the difference decreases with the zeolite content. − − − − − −

**Figure 8.** Bulk density of the studied samples.

The bulk density of the zeolite sample decreases up to 550 ◦C, then a slight increase occurs, and above 900 ◦C, a sharp increase is visible. Finally, the bulk density reaches 1760 kg·m−<sup>3</sup> at 1100 ◦C, which is about 42% higher than for the green sample. −

The bulk density of kaolin–zeolite samples at different temperatures is shown in Figure 9. The results show that the bulk density at temperatures of 800 and 900 ◦C decreases with the amount of zeolite. The differences between the SLA and KZ50 samples are about 40 kg·m−<sup>3</sup> at 800 ◦C and about 60 kg·m−<sup>3</sup> at 900 ◦C. The values of the bulk density at the temperatures of 1000 and 1100 ◦C also decrease, almost linearly in this case, except for sample KZ10. Those results show that the bulk density decreases with the zeolite content by about 6%. − −

**Figure 9.** Bulk density of the studied samples at 800, 900, 1000, and 1100 ◦C.

#### *3.6. Thermal Diffusivity*

− − − − − − − − The results of the thermal diffusivity of the kaolin–zeolite samples in the temperature interval from 30 to 1000 ◦C are plotted in Figure 10. The thermal diffusivity of the green samples decreases with the amount of zeolite from 0.57 mm<sup>2</sup> ·s −1 for sample SLA to 0.37 mm<sup>2</sup> ·s −1 for sample KZ50. The green zeolite sample has a thermal diffusivity of only 0.21 mm<sup>2</sup> ·s −1 . The thermal diffusivity of all studied samples decreases with temperature, except for the zeolite sample. The thermal diffusivity of the zeolite sample is almost the same in the whole temperature range (0.21 mm<sup>2</sup> ·s −1 ). The decrease in thermal diffusivity is caused by the escape of physically and chemically bound water in the samples (dehydration and dehydroxylation) [19,23,48]. The differences in thermal diffusivity between samples also decrease with temperature. They are 0.20 mm<sup>2</sup> ·s −1 for the green samples at room temperature and 0.07 mm<sup>2</sup> ·s −1 for the samples measured at 1000 ◦C. A similar decreasing trend of the thermal diffusivity of clays was obtained in [17,49]. The values were from 0.17 to 0.27 mm<sup>2</sup> ·s −1 . A more interesting study was published by Antal et al. [49], where the thermal diffusivity of textured kaolin samples was assessed. It is visible that the thermal diffusivity of kaolin samples depends significantly on the direction of crystal orientation. The results showed that thermal diffusivity is in the interval from 0.2 to 0.75 mm<sup>2</sup> ·s −1 measured at room temperature and from 0.15 to 0.4 mm<sup>2</sup> ·s <sup>−</sup><sup>1</sup> at the temperature of 1100 ◦C.

−

**Figure 10.** Thermal diffusivity of the studied samples.

− The thermal diffusivity of the kaolin–zeolite samples at different temperatures is shown in Figure 11. The thermal diffusivity decreases with the amount of zeolite up to 40 mass%. Sample KZ50 has almost the same value of thermal diffusivity as sample KZ40. All three curves (the results at 800, 900, and 1000 ◦C) exhibit the same trend. At a temperature of 1000 ◦C, samples KZ40 and KZ50 have values of thermal diffusivity similar to the zeolite sample. The difference is only 0.01 mm<sup>2</sup> ·s −1 .

−

**Figure 11.** Thermal diffusivity of the studied samples at 800, 900, and 1000 ◦C.

#### *3.7. Specific Heat Capacity*

− − − − − − − − − − − − The results of the specific heat capacity of the kaolin–zeolite samples during thermal treatment up to 1000 ◦C are shown in Figure 12. The heat capacity of the green samples is in the interval from 1.02 to 1.42 kJ·kg−<sup>1</sup> ·K−<sup>1</sup> . The values of the heat capacity for all samples decrease when the liberation of physically bound water is finished. The heat capacity of water is 4.2 kJ·kg−<sup>1</sup> ·K −1 (much higher than for solid materials); therefore, its decrease was expected in the temperature interval up to 300 ◦C. Then, the specific heat capacity of samples slightly increases up to 800 ◦C. The values are in the interval from 0.88 to 1.45 kJ·kg−<sup>1</sup> ·K−<sup>1</sup> . The dehydroxylation of kaolinite also influences the heat capacity of samples. The OH<sup>−</sup> groups are removed from the structure of kaolinite, and therefore,

the heat capacity does not increase, as it is excepted from the Debye model. Above 800 ◦C, a sharp increase in the heat capacity occurs due to crystallization of spinel [28] and a normal characterization of the heat capacity at higher temperatures. At a temperature of 1000 ◦C, the specific heat capacity of the samples ranges from 1.25 up to 2.24 kJ·kg−<sup>1</sup> ·K −1 , which is about 50% higher than it was for the green samples. Michot et al. [53] studied the heat capacity and thermal conductivity of kaolinite/metakaolinite in the temperature interval from room temperature up to 1000 ◦C. Their results showed that the heat capacity of metakaolinite above the temperature of 700 ◦C is about 1.2 kJ·kg−<sup>1</sup> ·K −1 , which is lower than our values. This can be explained by a different content of kaolinite in our kaolin sample and the application of a method which is not accurate for measurement of heat capacity. − − − − −

**Figure 12.** Specific heat capacity of the studied samples.

∙ − ∙ − The specific heat capacity of the kaolin–zeolite samples at temperatures of 800, 900, and 1000 ◦C is presented in Figure 13. The heat capacities of samples KZ30, KZ40, and KZ50 at 800 ◦C are almost the same. At temperatures of 900 and 1000 ◦C, the trend of the heat capacity is similar. Zeolite in the samples causes a decrease iin the specific heat capacity. The difference between samples SLA and KZ50 at 1000 ◦C is 0.99 kJ·kg−<sup>1</sup> ·K−<sup>1</sup> , which is about 80%. − −

**Figure 13.** Specific heat capacity of the studied samples at 800, 900, and 1000 ◦C.

∙ − ∙ −

∙ − ∙ −

#### *3.8. Thermal Conductivity*

The results of the specific heat capacity of the zeolite–kaolin samples during thermal treatment up to 1000 ◦C are shown in Figure 14. The thermal conductivity of the green samples (at 30 ◦C) is the highest for sample SLA (1.03 W·m−<sup>1</sup> ·K −1 ) and the lowest for the zeolite sample (0.37 W·m−<sup>1</sup> ·K −1 ). It decreases with the amount of zeolite. Then, in the temperature interval from 30 to 300 ◦C, the thermal conductivity of all studied samples decreases due to the liberation of physically bound water from the pores and crystal surfaces (water has a higher thermal conductivity than ceramic materials) [23]. Significant changes are observed in the temperature interval from 400 to 700 ◦C. This decrease is caused by dehydroxylation of kaolinite [19,48] and it is smaller with higher zeolite content in the samples. At 700 ◦C, all samples reach the lowest values of thermal conductivity, which range from 0.52 (sample SLA) to 0.25 W·m−<sup>1</sup> ·K−<sup>1</sup> (the zeolite sample). Above the temperature of 800 ◦C, a sharp increase in the thermal conductivity occurs due to the sintering process and crystallization of spinel [28]. The final values at 1000 ◦C are from 0.86 W·m−<sup>1</sup> ·K−<sup>1</sup> for sample SLA to 0.35 W·m−<sup>1</sup> ·K−<sup>1</sup> for sample KZ50. The results obtained by Michot et al. [53] showed a thermal conductivity of the green kaolin body of up to 0.5 W·m−<sup>1</sup> ·K−<sup>1</sup> , which is in good agreement with the value of 0.52 W·m−<sup>1</sup> ·K−<sup>1</sup> obtained in this study. On the other hand, our values of thermal conductivity were not measured but calculated from density, heat capacity, and thermal diffusivity. The surface of samples was coated by graphite before the analysis, and above 800 ◦C, graphite could burn out, although the analysis was performed in nitrogen atmosphere. Therefore, greater uncertainties are expected. − − − − − − − − − − − − − −

**Figure 14.** Thermal conductivity of the studied samples.

− − − − − − The thermal conductivity of the kaolin–zeolite samples at temperatures of 800, 900, and 1000 ◦C is shown in Figure 15. At a temperature of 800 ◦C, thermal conductivity has a decreasing trend up to 40 mass% of zeolite. Nevertheless, samples KZ40 and KZ50 have similar values of thermal conductivity (the difference is only 0.003 W·m−<sup>1</sup> ·K−<sup>1</sup> ). At temperatures of 900 and 1000 ◦C, thermal conductivity decreases with the amount of zeolite and this decrease is almost linear. The difference between samples SLA and KZ50 at 1000 ◦C amounts to 0.45 W·m−<sup>1</sup> ·K−<sup>1</sup> , which is about 53%. Typically, values of the thermal conductivity of ceramic materials are in the interval from 0.28 to 1.2 W·m−<sup>1</sup> ·K−<sup>1</sup> [53]. Therefore, the obtained results of the thermal conductivity are in good agreement with the literature.

**Figure 15.** Thermal conductivity of the studied samples at 800, 900, and 1000 ◦C.

#### *3.9. The Open Porosity and Scanning Electron Microscopy*

The results of bulk density were also confirmed with the results of the porosity of kaolin–zeolite samples at temperatures of 800, 900, 1000, and 1100 ◦C (see Figure 16). The porosity of the studied samples decreases and the bulk density increases with temperature. The porosity at 800 and 900 ◦C decreases when the sample contains 10 mass% of zeolite. Then, the porosity increases and reaches about 50% for samples KZ30, KZ40, and KZ50. The porosity of sample KZ10 at temperatures of 800, 900, and 1000 ◦C is almost the same. At a temperature of 1100 ◦C, porosity decreases from 48.5% (sample SLA) down to 43.0% (sample KZ50). The decrease in the porosity with temperature is caused by the sintering process in the samples [30]. This process causes a decrease in the sample volumes and vanishing of pores. The results show that the porosity decreases with the amount of zeolite in the samples.

**Figure 16.** Porosity of the studied samples at 800, 900, 1000, and 1100 ◦C.

An SEM picture of the green kaolin sample is presented in Figure 17A. Grains of kaolinite with a different size and also agglomerates are visible. Kaolinite crystals have a plate-like shape and they are ordered randomly. The samples contain also many small pores. This is confirmed with high values of the porosity (black or dark points). A picture

of the kaolin sample heated at 1100 ◦C is shown in Figure 17B. The porosity decreased and the kaolin sample stayed more compact. The structure remains in the same condition up to 1100 ◦C. Moreover, the pseudoamorphous shape of a grain of initial agglomerates is still the same.

**Figure 17.** SEM pictures. (**A**) green Sedlec kaolin; (**B**) Sedlec kaolin fired at 1100 ◦C; (**C**) green zeolite; (**D**) zeolite fired at 1100 ◦C.

An SEM picture of the green zeolite sample is presented in Figure 17C. Grains as well as agglomerates (clusters of grains) with different sizes and shapes can be seen. A picture of the zeolite sample heated at 1100 ◦C is shown in Figure 17D, where large pores (black or dark points) are mainly visible. The initial shape of grains is not preserved. The samples are more compact (higher density (see Figure 8) and smaller porosity), and the melting of grains is visible.

#### **4. Conclusions**

The thermophysical properties (thermal expansion, thermal diffusivity and conductivity, and specific heat capacity) of kaolin–zeolite samples were investigated during a heating stage of the firing. The samples were prepared from Sedlec kaolin (Czech Republic) and natural zeolite (Nižný Hrabovec, Slovakia). Kaolin was partially replaced with a natural zeolite in the amounts of 10, 20, 30, 40, and 50 mass%. Pure kaolin and zeolite samples were also prepared. From the obtained values for the thermophysical properties (thermal expansion, thermal diffusivity and conductivity, and specific heat capacity) of Sedlec kaolin and blends of that kaolin with natural zeolite, we can conclude that increasing zeolite content of the blends decreases the values of all studied properties (except thermal expansion), which is positive for bulk density, porosity, thermal diffusivity, and conductivity. Zeolite has a negative effect on thermal expansion due to an increased shrinkage. Therefore, the optimal amount of zeolite in the blends (for the studied properties) is 30 mass%.

**Author Contributions:** Conceptualization, A.T.; methodology, T.H., J.O., Š.C. and A.T.; investigation, I.S., T.H., Š.C., J.O., G.Ł., A.S. and A.T.; writing—original draft preparation, A.T.; writing—review and editing, J.O., T.H., Š.C., I.S., G.Ł., A.S. and A.T.; visualization, A.T.; supervision, A.T.; funding acquisition, T.H. and J.O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Ministry of Education of Slovak Republic, grant numbers VEGA 1/0810/19 and VEGA 1/0425/19, by RVO: 11000.

**Acknowledgments:** Authors wish to thank the company ZEOCEM, a.s. (Slovakia), for the supply of natural zeolite.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


## *Article* **Alkaline Activation of Kaolin Group Minerals**

## **Oliwia Biel, Piotr Ro ˙zek \* , Paulina Florek , Włodzimierz Mozgawa and Magdalena Król**

Faculty of Materials Science and Ceramic, AGH University of Science and Technology, 30 Mickiewicza Av., 30-059 Krakow, Poland; oliwia.biel@gmail.com (O.B.); paulina@agh.edu.pl (P.F.); mozgawa@agh.edu.pl (W.M.); mkrol@agh.edu.pl (M.K.)

**\*** Correspondence: prozek@agh.edu.pl

Received: 2 March 2020; Accepted: 31 March 2020; Published: 2 April 2020

**Abstract:** Zeolites can be obtained in the process of the alkali-activation of aluminosilicate precursors. Such zeolite–geopolymer hybrid bulk materials merge the advantageous properties of both zeolites and geopolymers. In the present study, the effect of the type and concentration of an activator on the structure and properties of alkali-activated metakaolin, and metahalloysite was assessed. These two different kaolinite clays were obtained by the calcination of kaolin and halloysite, and then activated with sodium hydroxide and water glass. The phase compositions were assessed by X-ray diffraction, the microstructure was observed via scanning electron microscope, and the structural studies were conducted on the basis of the infrared spectra. The structure and properties of the obtained alkali-activated materials depend on both the type of a precursor and the type of an activator. The formation of zeolite phases was observed when the activation was carried out with sodium hydroxide alone, or with a small addition of water glass, regardless of the starting material used. The higher proportion of silicon in the activator solution does not give crystalline phases, but only an amorphous phase. Geopolymers based on metahalloysite have better compressive strength as the result of the better reactivity of metahalloysite compared to metakaolin.

**Keywords:** zeolite; alkali-activation; geopolymer; metakaolin; metahalloysite

#### **1. Introduction**

Zeolites are crystalline aluminosilicates (having the general chemical formula Me2/nO·Al2O3·xSiO2·yH2O [1], with the possibility of [SiO4] substitution by another elements, e.g., [PO4]) whose characteristic feature is the microporous, regular structure with a constant channel and pore size. This feature is an advantage of zeolites over other known sorbents, as it enables the selective adsorption of ions and molecules. Zeolites are one of the products of alkali-activated aluminosilicate precursors. The activation is conducted by treating the material with an alkaline solution, and keeping for an appropriate time at an elevated temperature. This method of obtaining zeolites is well known, and has been used since many years; however, an improvement of its parameters is still the subject of research [2]. Depending on the used parameters, geopolymer gel and zeolites can be found among the reaction products. That type of composite is known as a zeolite–geopolymer hybrid bulk material.

Zeolite–geopolymer hybrid bulk materials are composites that combine the advantages of zeolites as a dispersed phase and geopolymer as a matrix. Their structure contains the meso-porosity of the geopolymer and the micro-porosity of the zeolite [3]. The range of porosity in those materials is very wide, and reaches from 5 Å up to 2 µm [4]. It was proven [3] that there is a relationship between the properties of initial kaolinite, its meta phase and the final product: zeolite–geopolymer hybrid bulk material, which gives opportunity to adjust the contents of desirable phases, and in consequence, the range of porosity.

There are known various aluminosilicate precursors, such as synthetic ones (coal fly ash, ash from waste incineration, blast furnace slag, etc.) or natural clays, for instance, kaolinite and

halloysite. Using those natural minerals for the synthesis of zeolite and zeolite–geopolymer hybrid bulk materials is a topical issue [3,5,6]. Despite the fact that both kaolinite (Al2Si2O5(OH)4) and halloysite (Al2Si2O5(OH)4·2H2O) have similar compositions [7], their reactivity is different. The source of the differences is the presence of interlayer water in halloysite's structure, which results in its greater porosity, chemical activity and specific surface area [8].

In this work, the results of the structural studies of different composites, obtained using the alkali-activation method and two different kaolinite clays as starting materials, were presented. It is well known that kaolinite and halloysite change into hydroxysodalite by the treatment with sodium hydroxide. It is also known that sodium zeolite type A can be synthesized from both raw materials [9–11]. However, the methodology for preparing composites varies in all available works. Therefore, the aim of this work was to compare the impact of the raw material structure on the course of the synthesis process. As activators, NaOH and water glass were applied. The specific objectives were to compare the properties of materials obtained on the basis of metakaolin and metahalloysite, their structural characterization, and the assessment of the impact of the activator type on the properties and structure of the resulting zeolite–geopolymer hybrid bulk materials. Such composites would possess the synergistic benefits of both zeolites and geopolymers, and could be used as monolithic sorbents, self-supported zeolite sieves or membranes, rather than construction materials.

#### **2. Materials and Methods**

#### *2.1. Chemicals*

Clays from two different Polish deposits were used in this study: kaolin from the Maria III (KSM Surmin-Kaolin S.A., Nowogrodziec, Poland) and halloysite from the Dunin (Kopalnia Haloizytu DUNINO Sp. z o. o., Krotoszyce, Poland). Sodium hydroxide (analytically pure NaOH as microbeads) and water glass (the content of Na2O 11.4 wt.% and SiO<sup>2</sup> 27.6 wt.%) were used as the activator.

#### *2.2. Pre-Treatment of Natural Clays*

The homogenized samples of kaolin and halloysite were thermally treated at 700 ◦C for 2 h (the temperature was chosen on the basis of the previous work [12]), in order to obtain more reactive materials, metakaolin and metahalloysite.

#### *2.3. Alkali-Activation*

The alkali-activated paste samples were synthesized using 4 mL of the activator per 5 g of metakaolin. The compositions of the specimens are summarized in Table 1. The solid and solution were mechanically stirred for several minutes at room temperature. The fresh paste was poured into a silicone mold (20 mm × 20 mm × 20 mm cubic samples) and activated at 80 ◦C for 24 h. After this time, the samples were demolded and left to mature for another 27 days. With each mixture at least eight samples were prepared.

**Table 1.** The mixture proportions of geopolymer pastes and their molar ratios of silica, alumina, water and sodium oxide.


#### *2.4. Characterization*

The solid-state characterization techniques, such as X-Ray Fluorescence (XRF), X-Ray Diffraction (XRD), Infrared Spectroscopy (FT-IR) and Scanning Electron Microscopy (SEM), were carried out on both raw materials, and on the as-synthesized materials.

The chemical compositions of the starting materials were determined using X-ray fluorescence. The spectrum was detected using the wavelength dispersive X-ray fluorescence spectrometer (WD-XRF) Axios mAX 4 kW, PANalytical equipped with Rh source. The PANalytical standardless analysis package Omnian was used for the quantitative analysis of the spectra.

The resulting materials were analyzed in the terms of the phase composition by means of Philips X-ray powder diffraction X'P-ert system (CuKα radiation). The measurements were carried out in the 2θ angle range of 5–90◦ for 2 h, with a step of 0.007. Phases were identified with the use of an X'Pert HighScore Plus application, and the International Centre for Diffraction Data.

The existence of zeolite frameworks was also confirmed by the analysis of the spectra in the mid infrared (4000–400 cm−<sup>1</sup> ) that were measured on Bruker VERTEX 70v vacuum FT-IR spectrometer using the standard KBr pellets methods. They were collected in after 64 scans at 4 cm−<sup>1</sup> resolution.

The microstructure of the resulting samples was observed using a scanning electron microscope FEI Nova NanoSEM 200. The samples were sprayed with graphite.

The bulk density was calculated by dividing the mass of the sample by its volume. The compressive strength of the samples, measured by using the ZwickRoell device machine that works on the principle of a hydraulic press, is defined as the ratio of the sample breaking force to the area on which the force acts. Eight similar samples were tested, and the average of the eight measurements was taken.

#### **3. Results and Discussion**

#### *3.1. Characterization of Raw Materials*

Both kaolinite and halloysite are considered to be the polytypical forms of aluminum monophyllosilicate with 1:1 packets [13]. They differ in the way the layer packages are arranged in the crystal structure [14]. The kaolinite crystals have a lamellar habit. The shifting of two layers relative to each other, tetrahedral (Si–O) and octahedral (Al–OH), creates a tubular halloysite structure. The characteristic feature of halloysite is also the presence of water molecules in the inter-package space. The presence of water causes an increase in the distance between the layer packs, where for halloysite it is 10.1 Å, while for kaolinite, this is 7.15 Å. Both minerals dehydroxylate at 600–800 ◦C [15]. As the products, more active amorphous metakaolin and metahalloysite, are obtained.

The chemical compositions of the obtained intermediates, determined on the basis of X-ray fluorescence spectrometry, are summarized in Table 2. By comparing both compositions, it can be stated that the metakaolin sample has a higher silica content, whereas halloysite has higher amounts of iron oxide. In addition, the molar ratio of the main components (Si/Al) in both cases was about 2. The content of CaO in both materials is below 1%, so the resulting products can be called geopolymers (due to the limitation of this name to low-calcium alkali-activated materials).

**Table 2.** The chemical compositions of the metakaolin and metahalloysite.


Figure 1 shows the XRD-patterns of both clays (**K** = kaolinite and **H** = halloysite) before and after the calcination. The XRD analysis of kaolin (Figure 1a) showed that it consisted of kaolinite and quartz. After the thermal treatment, only the peaks from quartz, which is not decomposed, are visible. The

distribution of kaolinite to the amorphous metakaolin.

appearance of an amorphous "halo" in the 20–30◦ range on the 2-theta scale indicates the distribution of kaolinite to the amorphous metakaolin. also gave the material a brown color. Attention should also be paid to the quartz content lower than in the case of kaolin, which may result in better reactivity in an alkaline environment, and thus achieving better strength parameters for samples obtained under the analogous conditions.

such that the high iron content in the chemical composition (Table 2) confirms its presence; and it

*Crystals* **2020**, *10*, x FOR PEER REVIEW 4 of 11

The appearance of an amorphous "halo" in the 20–30° range on the 2-theta scale indicates the

In turn, the analysis of the phase composition of halloysite (Figure 1b) showed that it included, in addition to halloysite, quartz and hematite. For both halloysite and metahalloysite, the reflections are wider, compared to the samples of kaolin and metakaolin, respectively, which indicates a greater

**Figure 1.** The X-ray diffractometry (XRD) patterns of initial clays and clays after the calcination at 700 °C: (**a**) kaolinite; (**b**) halloysite. **Figure 1.** The X-ray diffractometry (XRD) patterns of initial clays and clays after the calcination at 700 ◦C: (**a**) kaolinite; (**b**) halloysite.

Figure 2 presents the mid-infrared absorption spectra of kaolin and halloysite (marked as **K** and **H**, respectively) and their calcination products (**MK** and **MH**). The strong similarity between the spectra of the samples of kaolin and halloysite can be observed due to the structural similarity of both materials. The bands in the range of 1100–400 cm–1 are associated with the aluminosilicate structure of the material, and come from the vibrations of the Si–O–Si(Al) bridges. The band at 913 cm–1 is characteristic of the kaolinite structure, and comes from Al–OH stretching vibrations in the octahedral layer. The band at 695 cm–1, in turn, comes from the vibrations of the Al–O bond for aluminum in the octahedral position. In the spectra of all samples, a characteristic doublet of bands In turn, the analysis of the phase composition of halloysite (Figure 1b) showed that it included, in addition to halloysite, quartz and hematite. For both halloysite and metahalloysite, the reflections are wider, compared to the samples of kaolin and metakaolin, respectively, which indicates a greater defect or the smaller size of crystallites. In addition, the XRD pattern shows reflections from hematite, such that the high iron content in the chemical composition (Table 2) confirms its presence; and it also gave the material a brown color. Attention should also be paid to the quartz content lower than in the case of kaolin, which may result in better reactivity in an alkaline environment, and thus achieving better strength parameters for samples obtained under the analogous conditions.

at about 800 and 780 cm–1 can also be seen, indicating the presence of quartz. The spectra of the materials after the calcination are clearly changed. The bands derived from the water and the OH groups disappear, which is caused by the dehydroxylation process. Another difference is the increase in FWHM bands derived from Si–O and Al–O bond vibrations. For the samples of the starting materials, FWHM is small, which means that they had a crystal structure, while in the spectrum of the materials after the heat treatment, they are characterized by a larger FWHM, which indicates a reduced degree of structure ordering. The characteristic band, which indicated the presence of aluminum in the octahedral position (band at 913 cm–1) also disappeared, which indicates that a spatial amorphous aluminosilicate structure was obtained. Figure 2 presents the mid-infrared absorption spectra of kaolin and halloysite (marked as **K** and **H**, respectively) and their calcination products (**MK** and **MH**). The strong similarity between the spectra of the samples of kaolin and halloysite can be observed due to the structural similarity of both materials. The bands in the range of 1100–400 cm−<sup>1</sup> are associated with the aluminosilicate structure of the material, and come from the vibrations of the Si–O–Si(Al) bridges. The band at 913 cm−<sup>1</sup> is characteristic of the kaolinite structure, and comes from Al–OH stretching vibrations in the octahedral layer. The band at 695 cm−<sup>1</sup> , in turn, comes from the vibrations of the Al–O bond for aluminum in the octahedral position. In the spectra of all samples, a characteristic doublet of bands at about 800 and 780 cm−<sup>1</sup> can also be seen, indicating the presence of quartz.

The spectra of the materials after the calcination are clearly changed. The bands derived from the water and the OH groups disappear, which is caused by the dehydroxylation process. Another difference is the increase in FWHM bands derived from Si–O and Al–O bond vibrations. For the samples of the starting materials, FWHM is small, which means that they had a crystal structure, while in the spectrum of the materials after the heat treatment, they are characterized by a larger FWHM, which indicates a reduced degree of structure ordering. The characteristic band, which indicated the presence of aluminum in the octahedral position (band at 913 cm−<sup>1</sup> ) also disappeared, which indicates that a spatial amorphous aluminosilicate structure was obtained.

**Figure 2.** The mid-infrared (MIR) spectra of initial clays (**K** = kaolin (left) and **H** = halloysite (right)) and clays after the calcination at 700 °C (**MK** = metakaolin (left) and **MH** = metahalloysite (right)). **Figure 2.** The mid-infrared (MIR) spectra of initial clays (**K** = kaolin (**left**) and **H** = halloysite (**right**)) and clays after the calcination at 700 ◦C (**MK** = metakaolin (**left**) and **MH** = metahalloysite (**right**)).

#### *3.2. Characterization of Alkali-Activated Clays 3.2. Characterization of Alkali-Activated Clays*

identification was difficult [16].

typical for the N-(A)-S-H phase [19].

Zeolite–geopolymer hybrid bulk materials can be obtained using a thermal activated kaolinitic clay. Due to the chemical composition of metakaolin (Si/Al = 1), the most expected crystalline phase is zeolite A [12], although the appearance of sodalite is not excluded [16]. Zeolite–geopolymer hybrid bulk materials can be obtained using a thermal activated kaolinitic clay. Due to the chemical composition of metakaolin (Si/Al = 1), the most expected crystalline phase is zeolite A [12], although the appearance of sodalite is not excluded [16].

Figure 3 shows the XRD patterns of the materials obtained as the result of an alkaline activation of both metakaolin (Figure 3a) and metahalloysite (Figure 3b). The analysis of the phase composition showed the presence of quartz (in all samples) and hematite (only in the series based on halloysite). Both mentioned phases were present in the starting material (Figure 1), and did not degrade in the discussed process. Figure 3 shows the XRD patterns of the materials obtained as the result of an alkaline activation of both metakaolin (Figure 3a) and metahalloysite (Figure 3b). The analysis of the phase composition showed the presence of quartz (in all samples) and hematite (only in the series based on halloysite). Both mentioned phases were present in the starting material (Figure 1), and did not degrade in the discussed process. *Crystals* **2020**, *10*, x FOR PEER REVIEW 6 of 11

**Figure 3.** The XRD patterns of alkali-activated: (**a**) metakaolin; (**b**) metahalloysite. **Figure 3.** The XRD patterns of alkali-activated: (**a**) metakaolin; (**b**) metahalloysite.

The presence of zeolite phase was confirmed by the SEM observations. The selected results are shown in Figures 4 and 5. Especially in the samples containing zeolite A (**MK2**; Figure 4a), the cubic crystallites surrounded and joined with a layer of amorphous phase were visible. Zeolite X obtained in the analogous conditions had less developed morphology (**MH2**, Figure 5a), hence its As expected, zeolite A was formed as the result of activation with the highest sodium solution content (**MK1** and **MH1** samples). Interestingly, in the case of metakaolin (Figure 3a), the activation with the solution with a slightly lower Na/Si ratio (comparing **MK1** and **MK2**) gave the greater amounts of zeolite A. The probable reason was the increased proportion of silicon in the solution in the first

comparing the microstructure of **MK5** (Figure 4b) and **MH5** (Figure 5b) samples, it could be easily seen that, while in the case of the sample based on metakaolin (**MK5**), the low ordered amorphous phase was visible, the sample based on metahalloysite (**MH5**) was characterized by a microstructure

(**a**) (**b**)

**Figure 4.** The microstructure of metakaolin-based composites: (**a**) **MK2**; (**b**) **MK5**.

stages of the geopolymerization process [17,18], which promoted the condensation of tetrahedra into the double-6-ring units, characteristic for LTA structures. In the case of halloysite (**MH2**), zeolite X was formed under the same conditions. The literature data indicate [1] that this zeolite can be obtained using a longer synthesis time, or in reaction systems with a higher silicon content. It can therefore be assumed, that either the metahalloysite dissolved faster than the metakaolin, or the halloysite contained more silicon in the active phase (despite the similar silicon content in both raw materials (Table 1), metakaolin contained inert quartz (Figure 1)). In the systems with higher silicon content (**MK3**–**MK5** and **MH3**–**MH5**), zeolite phases were not formed. **Figure 3.** The XRD patterns of alkali-activated: (**a**) metakaolin; (**b**) metahalloysite. The presence of zeolite phase was confirmed by the SEM observations. The selected results are shown in Figures 4 and 5. Especially in the samples containing zeolite A (**MK2**; Figure 4a), the cubic crystallites surrounded and joined with a layer of amorphous phase were visible. Zeolite X obtained in the analogous conditions had less developed morphology (**MH2**, Figure 5a), hence its identification was difficult [16]. The higher reactivity of metahalloysite compared to metakaolin was demonstrated by the

(**a**) (**b**)

*Crystals* **2020**, *10*, x FOR PEER REVIEW 6 of 11

The presence of zeolite phase was confirmed by the SEM observations. The selected results are shown in Figures 4 and 5. Especially in the samples containing zeolite A (**MK2**; Figure 4a), the cubic crystallites surrounded and joined with a layer of amorphous phase were visible. Zeolite X obtained in the analogous conditions had less developed morphology (**MH2**, Figure 5a), hence its identification was difficult [16]. analysis of samples obtained with the higher proportion of water glass in the activating solution. By comparing the microstructure of **MK5** (Figure 4b) and **MH5** (Figure 5b) samples, it could be easily seen that, while in the case of the sample based on metakaolin (**MK5**), the low ordered amorphous phase was visible, the sample based on metahalloysite (**MH5**) was characterized by a microstructure typical for the N-(A)-S-H phase [19].

**Figure 4.** The microstructure of metakaolin-based composites: (**a**) **MK2**; (**b**) **MK5**. **Figure 4.** The microstructure of metakaolin-based composites: (**a**) **MK2**; (**b**) **MK5**.

**Figure 5.** The microstructure of metahalloysite-based composites: (**a**) **MH2**; (**b**) **MH5**. **Figure 5.** The microstructure of metahalloysite-based composites: (**a**) **MH2**; (**b**) **MH5**.

In the MIR spectra (Figure 6) the bands due to the characteristic vibrations of bonds observed in both types of oxygen bridges, Si–O–Si and Si–O–Al, were assigned. These bridges constitute basic structural units, forming tetrahedral geopolymer chains. It was found that the slag composition influences the presence of the bands connected with the phases formed during the hydration in the MIR spectra. Additionally, significant effect of amorphous phases share on the spectra shape was established. Based on IR spectra, it was also possible to determine the influence of the activator type The higher reactivity of metahalloysite compared to metakaolin was demonstrated by the analysis of samples obtained with the higher proportion of water glass in the activating solution. By comparing the microstructure of **MK5** (Figure 4b) and **MH5** (Figure 5b) samples, it could be easily seen that, while in the case of the sample based on metakaolin (**MK5**), the low ordered amorphous phase was visible, the sample based on metahalloysite (**MH5**) was characterized by a microstructure typical for the N-(A)-S-H phase [19].

on the products formed. The geopolymerization of alkali-activated metakaolin/metahalloysite (**MK**/**MH**) was indicated by the bands corresponding to vibrations of Si–O–Si(Al) at 1200–950 cm–1 (*νas*) and at 750–650 cm–1 (*νs*). Their presence in the IR spectra is due to the aluminosilicate character of the structure. These In the MIR spectra (Figure 6) the bands due to the characteristic vibrations of bonds observed in both types of oxygen bridges, Si–O–Si and Si–O–Al, were assigned. These bridges constitute basic structural units, forming tetrahedral geopolymer chains. It was found that the slag composition influences the presence of the bands connected with the phases formed during the hydration in the

observation agrees with our previous conclusions [18].

with the lowest sodium content).

First, the slight shift of the most intense band located at about 1000 cm–1 with the increasing amount of water glass can be observed. On the one hand, the reason may be the increasing share of silica in the structure. On the other hand, the degree of structure polymerization can be increased, which would explain the systematic increase in the compressive strength of the **MK1**–**MK5** samples, and which agrees well with the microscopic observations. The analogous conclusions can be drawn by interpreting the spectra of the series of samples based on metahalloysite (**MH1**–**MH5**). This

The significant differences in the IR spectra of the **MK1** and **MK2** samples are visible in the socalled pseudolattice range, 800–550 cm–1, in which the vibrations of the over-tetrahedral structural units of the zeolite framework can be visible. Especially, characteristic is the band at 557 cm–1, which is connected with the vibrations realized in the zeolite A structure. In turn, the spectrum of the sample

The carbonate group bands were identified at about 1450 and 880 cm–1, as they may be the result of the carbonation of N-(A)-S-H or other sodium compounds, or an unreacted alkali-activator. The interesting observation is the fact that the presence of zeolites in the geopolymer structure inhibits the carbonation process. In this case, sodium was probably bound in the zeolite structure, and there was no free OH groups that could be involved in the carbonation. On the other hand, the intensity of this band decreased as the sodium content in the reaction systems decreased (hence the disappearance of this band in the spectra of the MH5 sample, obtained by activation with the solution

**MH2** shows two bands characteristic for the faujasite-type structure (zeolite X).

MIR spectra. Additionally, significant effect of amorphous phases share on the spectra shape was established. Based on IR spectra, it was also possible to determine the influence of the activator type on the products formed. *Crystals* **2020**, *10*, x FOR PEER REVIEW 8 of 11

**Figure 6.** The MIR spectra of alkali-activated: (**a**) metakaolin; (**b**) metahalloysite. **Figure 6.** The MIR spectra of alkali-activated: (**a**) metakaolin; (**b**) metahalloysite.

The bulk density and compressive strength of the geopolymer samples after 28 days of maturation were determined. The results are shown in Figure 7. Generally, the strength achieved by geopolymers based on clay minerals is relatively low compared to materials based on fly ash or blast furnace waste [20]. In the case of this work, these values did not exceed 12 MPa. The highest compressive strength values were achieved for the samples **MK5**/**MH5**. This is in line with the The geopolymerization of alkali-activated metakaolin/metahalloysite (**MK**/**MH**) was indicated by the bands corresponding to vibrations of Si–O–Si(Al) at 1200–950 cm−<sup>1</sup> (ν*as*) and at 750–650 cm−<sup>1</sup> (ν*s*). Their presence in the IR spectra is due to the aluminosilicate character of the structure. These bands do not differ significantly from each other, however, the fact that the envelope in this range is a superposition of several bands should be kept in mind.

literature data [21], which indicates that the higher the silicon content in the activator, the higher the strength parameters. It is related to the content of the amorphous phase, that is, the higher the glassy content, the more visually homogenous gel structure (Figures 4 and 5), which have possibly contributed to its high strength after the activation. What is surprising is the high strength of the samples in which zeolite phases were found. Two factors related to the zeolite presence in a geopolymer matrix affect the compressive strength, where one is the amount of zeolites, and the second is the zeolite crystallite size. Their growth causes the First, the slight shift of the most intense band located at about 1000 cm−<sup>1</sup> with the increasing amount of water glass can be observed. On the one hand, the reason may be the increasing share of silica in the structure. On the other hand, the degree of structure polymerization can be increased, which would explain the systematic increase in the compressive strength of the **MK1**–**MK5** samples, and which agrees well with the microscopic observations. The analogous conclusions can be drawn by interpreting the spectra of the series of samples based on metahalloysite (**MH1**–**MH5**). This observation agrees with our previous conclusions [18].

decrease in the compressive strength of the composite [2]. The relatively high strength of the obtained samples with zeolites suggest that these factors are at the levels which can be borne by a geopolymer matrix without causing its weakening. When comparing the two raw materials used, it can be stated that the samples based on metahalloysite were characterized by a higher apparent density, and thus a higher compressive The significant differences in the IR spectra of the **MK1** and **MK2** samples are visible in the so-called pseudolattice range, 800–550 cm−<sup>1</sup> , in which the vibrations of the over-tetrahedral structural units of the zeolite framework can be visible. Especially, characteristic is the band at 557 cm−<sup>1</sup> , which is connected with the vibrations realized in the zeolite A structure. In turn, the spectrum of the sample **MH2** shows two bands characteristic for the faujasite-type structure (zeolite X).

strength. The obtained better strength values may have resulted from the abovementioned absence of quartz in the phase composition of this raw material, as well as from the greater reactivity of metahalloysite compared to metakaolin, which could have affected the formation of phases with the higher structural density. This was confirmed by both the microscopic observations (Figures 4 and 5), and also in the analysis of the infrared spectra (the higher positions of the band originating from Si–O–Si(Al) bridges with similar raw material compositions; (Figure 6)). A significant difference in the iron content of the compared raw materials (Table 2) can be controversial. The role of iron in the geopolymerization process is widely discussed [22–24]. It is generally accepted that the contribution of iron oxide inhibited the geopolymer formation, and it is The carbonate group bands were identified at about 1450 and 880 cm−<sup>1</sup> , as they may be the result of the carbonation of N-(A)-S-H or other sodium compounds, or an unreacted alkali-activator. The interesting observation is the fact that the presence of zeolites in the geopolymer structure inhibits the carbonation process. In this case, sodium was probably bound in the zeolite structure, and there was no free OH groups that could be involved in the carbonation. On the other hand, the intensity of this band decreased as the sodium content in the reaction systems decreased (hence the disappearance of this band in the spectra of the MH5 sample, obtained by activation with the solution with the lowest sodium content).

necessary to control the content of Fe2O3 to enhance the physical characteristics of geopolymer paste [22]. The distribution of iron in geopolymers made with iron-rich precursors has been also investigated [23,24]. The authors agree that iron (Fe3+) occupies primarily octahedral positions. However, information can be found that the coordinated Fe3+ replaced Al3+ in the aluminosilicate The bulk density and compressive strength of the geopolymer samples after 28 days of maturation were determined. The results are shown in Figure 7. Generally, the strength achieved by geopolymers based on clay minerals is relatively low compared to materials based on fly ash or blast furnace waste [20]. In the case of this work, these values did not exceed 12 MPa. The highest compressive

structure of the geopolymer [24]. On the other hand, some observation suggests the replacement of

strength values were achieved for the samples **MK5**/**MH5**. This is in line with the literature data [21], which indicates that the higher the silicon content in the activator, the higher the strength parameters. It is related to the content of the amorphous phase, that is, the higher the glassy content, the more visually homogenous gel structure (Figures 4 and 5), which have possibly contributed to its high strength after the activation. in fly ash has no negative effect on the strength development of geopolymer [26]. Based on the results described in this paper, it is impossible to clearly assess the role of iron in the formation of the geopolymer structure. Although **MH** contains much higher amounts of iron (Table 2), it is mainly in the form of hematite (Figure 1), and in this form it can also be found in the material after activation (Figure 3). In addition, composites based on **MH** have a higher compressive strength, which is due to the higher degree of structure polymerization observed both in FT-IR and SEM analysis.

exhibited better heat resistance than Portland cement concrete [25]. Also, a high iron oxide (31.1 wt.%)

*Crystals* **2020**, *10*, x FOR PEER REVIEW 9 of 11

It was observed that the optimal composition of geopolymer based on fly ash and red mud (Fe2O3 content = 33.9 wt.%) was at a weight ratio of 50/50 that gave higher compressive strength than

**Figure 7.** The physical and mechanical properties of the obtained samples: (**a**) bulk density; (**b**) compressive strength. **Figure 7.** The physical and mechanical properties of the obtained samples: (**a**) bulk density; (**b**) compressive strength.

**4. Conclusions**  The aim of this study was to determine the effect of the type and concentration of an activator on the structure and properties of the alkali-activated metakaolin and metahalloysite. Therefore, the first stage of the work was based on the obtaining of the metakaolin and metahalloysite phases as a result of the calcination of the starting materials, kaolin and halloysite, respectively. The obtaining of What is surprising is the high strength of the samples in which zeolite phases were found. Two factors related to the zeolite presence in a geopolymer matrix affect the compressive strength, where one is the amount of zeolites, and the second is the zeolite crystallite size. Their growth causes the decrease in the compressive strength of the composite [2]. The relatively high strength of the obtained samples with zeolites suggest that these factors are at the levels which can be borne by a geopolymer matrix without causing its weakening.

the reactive phases with a similar structure was confirmed by the analysis of the XRD and IR results. The chemical and phase composition analysis confirmed that both of the starting materials are mainly in the amorphous aluminosilicate phase. The next stage of the work was an attempt to alkali-activate the previously obtained metakaolin and metahalloysite. A number of activation solutions based on sodium hydroxide and sodium silicate were used. The structure and properties of the resulting geopolymer binders depend on both the type of starting material used, and the type of activator. The activation carried out with sodium hydroxide alone, or with a small addition of water glass, causes the formation of zeolite phases, regardless of When comparing the two raw materials used, it can be stated that the samples based on metahalloysite were characterized by a higher apparent density, and thus a higher compressive strength. The obtained better strength values may have resulted from the abovementioned absence of quartz in the phase composition of this raw material, as well as from the greater reactivity of metahalloysite compared to metakaolin, which could have affected the formation of phases with the higher structural density. This was confirmed by both the microscopic observations (Figures 4 and 5), and also in the analysis of the infrared spectra (the higher positions of the band originating from Si–O–Si(Al) bridges with similar raw material compositions; (Figure 6)).

the starting material used. The alkaline activation with a solution with a higher proportion of silicon did not give crystalline phases, but only an amorphous N-(A)-S-H phase. The bulk density and compressive strength were determined. The obtained strength values were in the range of 0.7–7.1 MPa for metakaolin, and in the range of 1.1–11.3 MPa for metahalloysite. The geopolymers based on metahalloysite had better strength results, which was probably due to their higher bulk density. A higher density, and thus better strength parameters, were probably the result of the better reactivity of metahalloysite compared to metakaolin. Zeolites are known for their ability to immobilize heavy metal ions. Although the resulting materials did not obtain the strength parameters valuable for construction, subsequent tests may prove to be attractive materials for neutralizing hazardous waste and water treatment. A significant difference in the iron content of the compared raw materials (Table 2) can be controversial. The role of iron in the geopolymerization process is widely discussed [22–24]. It is generally accepted that the contribution of iron oxide inhibited the geopolymer formation, and it is necessary to control the content of Fe2O<sup>3</sup> to enhance the physical characteristics of geopolymer paste [22]. The distribution of iron in geopolymers made with iron-rich precursors has been also investigated [23,24]. The authors agree that iron (Fe3+) occupies primarily octahedral positions. However, information can be found that the coordinated Fe3<sup>+</sup> replaced Al3<sup>+</sup> in the aluminosilicate structure of the geopolymer [24]. On the other hand, some observation suggests the replacement of Al3<sup>+</sup> by Fe3<sup>+</sup> in octahedral sites most likely within the kaolinite phase, which indicate that iron is not necessarily deleterious to geopolymer formation, as has sometimes been suggested [23].

**Author Contributions:** Conceptualization, M.K.; sample preparation, O.B.; measurement, O.B and P.R.; data curation, O.B. and M.K; writing—original draft preparation, M.K., P.R. and P.F.; writing—review and editing, P.R. and P.F.; project administration, M.K..; funding acquisition, W.M. All authors have read and agreed to the published version of the manuscript. It was observed that the optimal composition of geopolymer based on fly ash and red mud (Fe2O<sup>3</sup> content = 33.9 wt.%) was at a weight ratio of 50/50 that gave higher compressive strength than at any other ratio [24]. Geopolymers based on fly ash with a high content of iron oxide (48.8 wt.%) exhibited better heat resistance than Portland cement concrete [25]. Also, a high iron oxide (31.1 wt.%) in fly ash has no negative effect on the strength development of geopolymer [26]. Based on the results described in this paper, it is impossible to clearly assess the role of iron in the formation of the geopolymer structure. Although **MH** contains much higher amounts of iron (Table 2), it is mainly in the form of hematite (Figure 1), and in this form it can also be found in the material after activation (Figure 3). In

addition, composites based on **MH** have a higher compressive strength, which is due to the higher degree of structure polymerization observed both in FT-IR and SEM analysis.

#### **4. Conclusions**

The aim of this study was to determine the effect of the type and concentration of an activator on the structure and properties of the alkali-activated metakaolin and metahalloysite. Therefore, the first stage of the work was based on the obtaining of the metakaolin and metahalloysite phases as a result of the calcination of the starting materials, kaolin and halloysite, respectively. The obtaining of the reactive phases with a similar structure was confirmed by the analysis of the XRD and IR results. The chemical and phase composition analysis confirmed that both of the starting materials are mainly in the amorphous aluminosilicate phase.

The next stage of the work was an attempt to alkali-activate the previously obtained metakaolin and metahalloysite. A number of activation solutions based on sodium hydroxide and sodium silicate were used. The structure and properties of the resulting geopolymer binders depend on both the type of starting material used, and the type of activator. The activation carried out with sodium hydroxide alone, or with a small addition of water glass, causes the formation of zeolite phases, regardless of the starting material used. The alkaline activation with a solution with a higher proportion of silicon did not give crystalline phases, but only an amorphous N-(A)-S-H phase.

The bulk density and compressive strength were determined. The obtained strength values were in the range of 0.7–7.1 MPa for metakaolin, and in the range of 1.1–11.3 MPa for metahalloysite. The geopolymers based on metahalloysite had better strength results, which was probably due to their higher bulk density. A higher density, and thus better strength parameters, were probably the result of the better reactivity of metahalloysite compared to metakaolin.

Zeolites are known for their ability to immobilize heavy metal ions. Although the resulting materials did not obtain the strength parameters valuable for construction, subsequent tests may prove to be attractive materials for neutralizing hazardous waste and water treatment.

**Author Contributions:** Conceptualization, M.K.; sample preparation, O.B.; measurement, O.B. and P.R.; data curation, O.B. and M.K; writing—original draft preparation, M.K., P.R. and P.F.; writing—review and editing, P.R. and P.F.; project administration, M.K.; funding acquisition, W.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was financially supported by The National Science Centre Poland under grant no. 2018/31/B/ST8/03109.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Communication*
