Study of the Crystallographic Distortion Mechanism during the Annealing of Kaolinite

Abstract

The distortion process of kaolinite annealed from 25 °C to 550 °C for different holding times can be characterized using a thermogravimetric/differential scanning calorimeter (TG/DSC) for thermal analysis, X-ray diffraction (XRD) for establishing the crystal structure, the Fourier transform infrared spectrum (FTIR) for identifying the functional groups, and a scanning electron microscope (SEM) for establishing the microstructure. Dehydroxylation is the main reaction during annealing from 25 °C to 550 °C and leads to kaolinite crystal distortion. A stable crystal structure during distortion was obtained by optimizing the bulk phase with quantum chemistry. Then, the crystal structure was studied by using ab initio multiple scattering calculations for X-ray absorption of the fine structure (XAFS). The results of X-ray absorption near the edge structures (XANES) determined that peak shifts and intensity phases slightly increased. The crystal structure distortion of kaolinite during annealing can be explained by the experimental and simulation results. This work provides theoretical support for identifying kaolinite with different degrees of distortion and has the potential for further developments in coal gangue separation.

1. Introduction

With the rapid development of artificial intelligence, compared with traditional coal gangue separation technology, including manual separation, gravity coal preparation, flotation, and other methods, intelligent photoelectric separation has the advantages of nondestructive testing and intelligent remote control; it creates no sewage, and can be adapted to an underground environment, and therefore has been widely researched [1,2]. To remove gangue from raw coal is conducive to improving the efficiency of coal utilization. There are numerous different separation methods, such as flotation, gravity separation, etc.; however, for downhole sorting, photoelectric separation has previously not been possible. In recent years, intelligent photoelectric separation has been widely developed and applied, based on X-ray detection technology and image detection technology. Recognition depends on the different attenuations of X-rays through the coal and gangue, but a deep recognition principle has not yet been found [3,4,5]. A variety of clay minerals are widely mixed in coal seams and are important in the composition of gangue. Among them, kaolinite is used in a large variety of technological applications; it is necessary to separate it from the coal gangue mixture to improve the purity of the coal and facilitate subsequent processing and application. Kaolinite is a form of water-containing aluminosilicate. Its theoretical structure is Al₂Si₂O₅(OH)₄; octahedral aluminosilicate with a structure of 1:1 is formed by connecting the octahedral aluminosilicate with the tetrahedral aluminosilicate and stacking along the c axis [6]. In this study, kaolinite is the main object of study.

The existing coal gangue identification equipment is based on the use of the level of gray value to represent the percentage of X-ray attenuation after passing through coal gangue [7,8]. Coal and gangue are distinguished by the difference in gray value. However, it is difficult to distinguish gangue when the gray value difference is not obvious. In order to improve the identification accuracy of coal gangue using intelligent photoelectric separation, based on X-rays for kaolinite and coal, the physical and chemical characteristics of coal and kaolinite have attracted much attention due to their attenuation contribution to X-rays.

The structure of kaolinite has been studied since the 1930s [9]; experimental, theoretical, and combined strategies have been developed to analyze the structure of kaolinite, and several crystal models of kaolinite have been proposed, inspired by changes in the crystal structure of kaolinite caused by heating [10,11,12,13,14,15]. When kaolinite is annealed at 550 °C, as the heating time increases, the hydroxyl of kaolinite is gradually removed and the crystal is transformed into amorphous kaolin [16,17]. In this work, annealing is used to change the crystal structure of kaolinite to obtain kaolinites with different degrees of distortion.

XAFS (X-ray absorption fine-structure spectroscopy) techniques, such as X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS), is a widely-used material characterization technique that is able to obtain information regarding the oxidation state of the atoms, the average coordination number, interatomic distances, and degrees of disorder from the atomic environment to the absorbing atom [18,19,20].

In this study, kaolinite with different crystallinities was obtained for experimentation by using the characteristics of the dehydroxylation of kaolinite at 550 °C and the complete disappearance of the crystal structure [17], identified by thermogravimetric analysis, scanning electron microscopy, Fourier-transform infrared spectrometry, and X-ray diffraction [21,22]. At the same time, material studio software was used to simulate the theoretical kaolinite crystal and the partially deformed kaolinite [23]. The kaolinite crystal structure data was based on mp-541152 Al₂Si₂H₄O₉ reported by the Materials Project [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40], and the theoretical XRD was compared with the actual implementation results. The EXAFS signal changes of kaolinite crystal and partially deformed kaolinite crystal were simulated by multiple scattering computation, based on ab initio and ATHENA software [41,42,43]; multiple scattering computation based on ab initio is an implementation of real-space multiple scattering theory (RSMS) and is capable of calculating a wide range of X-ray and electron spectra, while ATHENA is a program capable of supporting common XAS data-processing chores. These results will help to provide new options for separating coal (which is amorphous) and gangue (which is mostly crystalline) from coal gangue.

2. Materials and Methods

2.1. Materials

In this work, the standard kaolin clay sample Raw-Ka was purchased from Aladdin (Shanghai Aladdin Biochemical Technology Co., Ltd., Shanghai, China) of pharmaceutical grade. The sample has high purity, fine particle size, and high kaolinite content (>95%). The sample was dried at 90 °C for 24 h before the experiment. Anhydrous alcohol (at analytically pure grade) was purchased from Aladdin, Shanghai, China. The chemical composition and particle size analysis (PSD) results are listed in

Table A1. XRF analysis of the Raw-Ka sample.

Oxide	Content (%)	Standard Error
SiO2	54.82	0.25
Al2O3	43.16	0.25
TiO2	0.654	0.033
Fe2O3	0.48	0.024
Na2O	0.225	0.011
MgO	0.164	0.008
P2O5	0.162	0.008
K2O	0.153	0.008
SO3	0.0566	0.0028
ZnO	0.0415	0.0021
CaO	0.0305	0.0015
V2O5	0.0147	0.0008
Cl	0.0138	0.0009
Cr2O3	0.0058	0.0005
CeO2	0.0056	0.002
Nd2O3	0.003	0.0011
Re2O7	0.0025	0.0012
Ga2O3	0.0022	0.0003
MnO	0.0019	0.0004
Sc2O3	0.0017	0.0004
ZrO2	0.0015	0.0003

and Figure A1, respectively.

2.2. Preparation

The evolution of the kaolinite crystal structure distortion was achieved through thermal treatment. Firstly, in order to find the appropriate temperature for studying the distortion of kaolinite under heat treatment, Raw-Ka was heat-treated in a laboratory in a programmable muffle furnace (tube furnace) in the air for 2 h at different temperatures (400, 500, 550, 600, 700, and 800 °C). Raw-Ka was also heated at the same rate from 25 °C to 550 °C, with different holding times (10, 20, 30, 60, 90, and 120 min, respectively), labeled as Ka-10 min, Ka-20 min, Ka-30 min, Ka-60 min, Ka-90 min, and Ka-120 min, respectively.

2.3. Characterization

The composition of Raw-Ka was investigated using an X-ray fluorescence spectrometer (XRF, Thermo Fisher, Waltham, MA, USA). The PSD was carried out with a SALD-7101 ultraviolet laser nanoparticle-size analyzer (Shimadzu, Kyoto, Japan). The TG/DSC of Raw-Ka was performed using an SDTA851e thermogravimetric (Maitler Toledo, Greifensee, Switzerland), in a temperature range from 25 to 800 °C at 10 °C/min, in the air. The powder XRD data were recorded using a Smartlab SE (Rigaku, Tokyo, Japan, using CuKα radiation). The FTIR absorption spectra were recorded on a NICOLET 380 (Thermo Fisher, Waltham, MA, USA) from 400 to 400 cm⁻¹, using 32 scans with a resolution of 4 cm⁻¹ on a KBr sampling sheet. The SEMs were recorded with a Zeiss Sigma300 (Zeiss, Oberkochen, Germany).

2.4. Simulation of Crystal Structure Distortion

Firstly, an ideal kaolinite crystal structure was constructed, a representative kaolinite crystal structure was formed according to the Materials Project mp-541152 Al₂Si₂H₄O₉ (denoted as Sim-Ka). The Materials Project mp-541152 corresponds to a crystal structure file of kaolinite [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40]. The kaolinite lattice constant values are a = 5.213 Å, b = 7.479 Å, c = 9.052 Å, α = 91.787°, β = 89.728°, γ = 104.975°, in the space group P1. With the disordering of the kaolinite, in the process of annealing, hydroxyl groups were removed from the alumina tetrahedron in order of O33H15→O32H14→O31H13→O27H9. After constructing the kaolinite cell model, geometric optimization was carried out. All calculations were based on DFT and were carried out using the CASTEP module [44] in the Material Studio software. Generalized gradient approximation (GGA) with a Perdew–Burke–Ernzerhof (PBE) scheme and a 3 × 2 × 2 Monkhorst-Pack K-point grid was used to achieve geometric optimization [45]. The geometric optimization was set according to the following convergence accuracy: The plane-wave energy cut-off that was used in the calculations was 450.0 eV. The total self-consistent field (SCF) energy change was 1 × 10⁻⁶ eV/atom, while the energy between optimization steps was 1 × 10⁻⁵ eV/atom. The atomic orbitals involved in the calculation were H¹S¹, O 2s²2p⁴, Al 3s²3p¹, and Si 3s²3p², respectively.

Notably, each hydroxyl removal was accompanied by applying geometric optimization. The CSTEP outputs were denoted as Phase 1, Phase 2, Phase 3, and Phase 4.

2.5. XAS Calculation

The absorption coefficient μ, which gives the probability that the X-ray will be absorbed according to the Beer–Lambert Law [46], is mainly involved when discussing X-ray absorption:

$$I=I_0{e}^{-\mu t}$$

(1)

where I₀ is the X-ray intensity incident on a sample, t is the sample thickness, and I is the intensity transmitted through the sample.

At most X-ray energies, the absorption coefficient μ is a function of energy, sample density ρ, the atomic number Z, atomic mass A, and the X-ray energy, E, which is roughly that reported in [47].

$$\mu \approx \frac{\rho Z^4}{A{E}^3}$$

(2)

When the incident X-ray has an energy equal to that of the binding energy of a core-level electron, there is a sharp rise in absorption, with an absorption edge corresponding to the core level to the continuum. The energy dependence of the X-ray absorption coefficient μ(E) in transmission mode is that reported in [48].

$$\mu \left(E\right)=\ln \frac{I}{I_0}$$

(3)

The EXAFS fine-structure function χ(E) as a change to μ(E) and μ₀(E), as reported in [48]:

$$\chi \left(E\right)=\frac{\mu \left(E\right)-{\mu }_0\left(E\right)}{\Delta {\mu }_0\left(E\right)}$$

(4)

where μ(E) is the measured absorption coefficient and μ₀(E) is a smooth background function, representing the absorption of an isolated atom.

EXAFS is best understood in terms of the wave behavior of the photoelectron that is created in the absorption process. It is common to convert the X-ray energy to k [24], the wave number of the photoelectron, which is defined as:

$$k=\sqrt{\frac{2m(E-E_0)}{{\mathit{ħ}}^2}}$$

(5)

where E₀ is the absorption edge energy, m is the electron mass, and ħ is Planck’s constant. The theoretical expression of EXAFS [49] can be obtained by introducing k to the equation:

$$x\left(k\right)={\displaystyle \sum }_j\frac{N_j{f}_j\left(k\right)e^{-2k^2{\sigma }_j^2}}{k{R}_j^2}\sin \left[2k{R}_j+{\delta }_j\left(k\right)\right]$$

(6)

Under the conditions of known scattering amplitude $f\left(k\right)$ and phase displacement $\delta \left(k\right)$ via numerical analysis, the number of neighboring atoms is N, the distance to the neighboring atom is R, and the disorder in the neighboring distance is ${\sigma }^2$.

The theoretical structure and simulation analysis of the intermediate state of dehydroxylation and the structure change of kaolinite are carried out using Materials Studio and ATHENA. The EXAFS change caused by the increase of kaolinite disorder has been studied elsewhere [41,42,43].

The CASTEP module in the Materials Studio software was used to optimize the structure. The kaolinite model of mp-541152 was used to optimize the crystal model by using the PBE exchange-correlation function in the generalized gradient approximation (GGA) [21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,42].

The XAFS at the Si4 sites was calculated by multiple scattering computations based on ab initio, with Si as the central absorption atoms. The coordination number, coordination distance, and species information of the absorption atoms were given. The first coordination shells of Si4 at the K edge were calculated and analyzed.

The self-consistent field (SCF) and the full multiple scattering (fms) were calculated for the unit cell. A Hedin–Lundqvist-type exchange-correlation potential was used.

3. Results and Discussion

3.1. TG/DSC Measurements

The distortion of kaolinite is highly correlated with heat treatment. The results obtained by the thermogravimetric/differential scanning calorimeter (TG/DSC) for Raw-Ka was presented in Figure 1. The sample has a strong internal thermal effect in the range of 450 °C to 600 °C. The endothermic peak at 500 °C corresponds to the dehydroxylation of kaolinite, which is consistent with the descriptions found in previous studies. In addition, 17.53% of weight loss occurred throughout the heating process, which corresponds to the loss of the hydroxyl weight of kaolinite during heat treatment [16].

3.2. FTIR

Dehydroxylation is the main change of kaolinite during annealing; it plays a pivotal role in the distortion of kaolinite and the corresponding structural characteristics of the samples. The FTIR absorption spectra of Raw-Ka, Ka-10 min, Ka-30 min, and Ka-120 min is shown in Figure 2a.

Details of FTIR spectra is shown if Figure 2b. The prominent peaks of the stretching vibration phase of the Al-OH groups appeared at 3695, 3654, and 3620 cm⁻¹ in the FTIR spectra, respectively, corresponding to the inner surface hydroxyls, the out-of-plane stretching vibrations, and the inner hydroxyl between the tetrahedral and octahedral. Several bands within 1100 cm⁻¹ to 1000 cm⁻¹ corresponded to Si-O in the in-plane stretching modes. For inner Al-Al-OH bonds, a peak was observed at 916 cm⁻¹. The weak bends at 792, 755, and 698 cm⁻¹ were assigned for Si-O bending. In the lower-frequency region, the Si-O-Al, Si-O-Si, and Si-O bending peaks appeared at 538, 470, and 428 cm⁻¹, respectively [50,51]. With the increase in heat-treatment time, the reflection intensity has obviously been reduced in Ka-10 min, Ka-20 min, and Ka-120 min. In terms of annealing holding for 120 min, almost all the peaks of Al-OH groups were absent, while the FTIR spectrum degraded into a relatively flat line, which consisted of three smooth peaks at 1064, 809, and 462 cm⁻¹. The peak at 1064 cm⁻¹ was for tetrahedral sheet vibration; the peak at 809 cm⁻¹ was for free silica or quartz; a new peak appeared at 451 cm⁻¹ for structural distortion [11,22,52]. The results show the dehydroxylation phenomenon, which leads to the distortion of the kaolinite.

3.3. Microstructural Characterization

Figure 3 shows images of the original kaolinite (Raw-Ka) and annealed kaolinite samples taken with a Zeiss Sigma300 (Zeiss, Oberkochen, Germany). Pictures show several random particles of the four samples, scanned by a scanning electron microscope, with the typical agglomeration features of clay aggregates and clear platelet morphology. Aggregates of hexagonal spoon kaolinite slabs can be clearly observed, dominated by face-to-face and edge-to-edge contacts.

For Raw-Ka, Ka-10 min, the nano-scale internal structure of the deposition layer can be identified. The Ka microphotograph (Figure 3a) shows that the samples have an obvious layered structure and that these layers are stacked approximately directly. The Ka-10 min photograph (Figure 3b) shows that the kaolinite at 550 °C for 10 min still has an obviously layered structure, but it also shows that this reflects a certain degree of deformation. As the holding time increased from 30 min to 120 min, the layered structure of kaolinite from Figure 3c,d fractured and deformed due to dehydroxylation and crystal structure collapse, which was consistent with the results obtained by XRD.

3.4. Crystal Structure

The crystal structures of Mp-541152 Al₂Si₂H₄O₉ [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40] as Sim-Ka, Phase 1, Phase 2, Phase 3, and Phase 4 are shown in Figure 4. Among them, Phase 1, Phase 2, Phase 3, and Phase 4 are stable structures that may exist in the annealing distortion process of geometrically optimized kaolinite and are shown in Figure 4b–e. With the dehydroxylation and distortion of kaolinite, the Al-O octahedral structure degrades into a relatively unstable Al-O tetrahedral structure. At the same time, the distortion of the Al-O groups leads to a series of movement and rotation transformations of Si-O groups, but the Si-O tetrahedron structure remains intact, on the whole. With the transformation of the Al-O octahedron into an Al-O tetrahedron, the structure of the whole aged rock tends to be unstable. With the continuous annealing process, kaolinite eventually loses its crystal structure and becomes amorphous metakaolin. In this study, based on the results of TG/DSC and FTIR XRD patterns, combined with the results of other studies, we used software to simulate the possible change process of crystal structure in the distortion process of kaolinite [11,17]. The distortion of kaolinite in the process of Sim-Ka → Phase 4 change was mainly studied.

3.5. Powder XRD

The distortion of kaolinite due to the elimination of structure hydroxyl groups at 550–800 °C [17] corresponds with the obvious changes in the XRD reflections between Ka-550 °C and Ka-800 °C (Figure A2). The results of Figure A2. show that 550 °C is suitable for preparing kaolinite with different degrees of distortion.

To study the relationship between the structural distortion of kaolinite and X-ray absorption, kaolinite was annealed at 550 °C for different holding times. Figure 5 shows the XRD patterns of the sample at Raw-Ka, Ka-10 min, Ka-20 min, Ka-30 min, Ka-60 min, and Ka-120 min. With the annealing process, the diffraction peak of the sample gradually weakened and finally became flat, indicating the collapsing behavior of the crystal texture, especially of the (001) reflection at 2θ = 12° and (002) reflection at 2θ = 25°. These peaks are gradually decreasing with the increase in time, which corresponds to the poorly ordered kaolinite. The presence of poorly ordered kaolinite implies that annealing this kaolin will transform it into highly reactive metakaolin. Combined with previous studies, it can be concluded that the disappearance of the diffraction peak of kaolinite corresponds to the dehydroxylation of kaolinite into amorphous kaolin during heat treatment [53].

After a heating duration of 120 min, the peak of (001) diffraction cannot be distinguished from the background intensity, which indicates that the crystal structure of kaolinite is completely distorted and that most of the kaolinite in the sample is transformed into amorphous metakaolin [53]. It was proved that with the increase in holding time, the kaolinite crystal degenerated and finally lost its crystal structure, evolving into amorphous metakaolin.

Figure 6a shows the XRD simulation patterns of Sim-Ka, Phase 1, Phase 2, Phase 3, and Phase 4. Material Studio was used to simulate the XRD pattern and the XRD pattern with the removal of partial hydroxyl and the changes of atomic position that denoted Phase 1, Phase 2, Phase 3, and Phase 4. With the removal of the partial hydroxyl and the change of Si atom position, it can be observed that the diffraction peak changes significantly. The θ_m moves from 12.25° to 11.79°, and the peak intensity in the range of 2θ = 20–30° is no longer consistent with the intensity shown in PDF 78–2110, while certain peaks with a smaller intensity have appeared. The details are shown in Figure 6b. The spectrum shows an amorphous structure with crystalline peaks of kaolinite, suggesting incomplete dehydroxylation compared with the spectrum shown in Figure 5 and in previous studies [11,17].

3.6. XAFS Calculation

To compare the different structures of kaolinite, the spectrums of the simulated kaolinite structure are calculated by multiple scattering computations, based on ab initio [19,54,55,56]. The multiple scattering calculation is based on the simulated samples including Sim-Ka, Phase 1, Phase 2, Phase 3, and Phase 4, which correspond to the incomplete dehydroxylation of kaolinite.

Figure 7 presents the simulated XAFS spectroscopy for the samples at different reaction stages including Si K-edge XANES absorption, the EXAFS functions k² χ(k), and the radial structural function (RSF) of samples with the calculated theoretical multiple scattering spectra.

The peak features shown in Figure 7a correspond to the simulated XANES spectra at the Si K-edge of Sim-Ka and samples at different reaction stages. The calculated main characteristic pick of Si of the Sim-Ka K-edge (absorption peak A) is E = 1847.5 eV, due to the permitted transition of Si 1 s electrons to an antibonding 3p-like state, and is related to tetrahedral coordination, which is one of the apparent characteristics of SiO₂ and is consistent with the report [12,20].

For the normalized X-ray absorption spectra of the Si K-edge, the position of the characteristic peak of Si changed from 1846.7 eV to 1845.1 eV.

In addition, in all cases, the post-edge resonances that are present are similar to those reported in other articles for kaolinite. The resonances at 1860.7 eV (absorption peak B) are the result of multiple scattering processes within the (SiO₄)-tetrahedron. Slight differences are observed between Sim-Ka and Phase 1, Phase 2, and Phase 3, while mutation occurred during Phase 1 to Phase 2, and during Phase 3 to Phase 4, which corresponds to the distortion of the kaolinite crystal structure, while the structural change of (SiO₄)-tetrahedron resulted in the change in X-ray absorption.

The X-ray absorption spectrum was converted into wave vector space using the formula $k=2\mathsf{\pi }\sqrt{2m\left(E-E_0\right)}$, where E is the incident X-ray energy, E₀ is the energy of the absorption edge, m is the electron mass, and h is the Planck constant. In order to enhance the signal strength, the data is weighted by k². Figure 7b shows the al K-varying k² χ(k) of the sample. It can be found that with the distortion of the kaolinite crystal structure, the curve deviates from its original position, indicating the change in the structure at the selected atomic site.

Figure 7c presents the different Fourier transforms of the EXAFS oscillations for the studied samples, measured at the Si K-edge, with the k-range of the Hanning window from 3 to around 15. The simulated spectrum possesses most of the characteristics found in the experimental measurements, which shows that the calculated structure is consistent with the X-ray absorption technique. The peaks in the graph correspond to the Si/O coordination shell at the first nearest-neighbor position, around the Si atoms. For the radial structural function of Si (IV), Figure 7c shows that the atomic distance between Si-O is maintained at 1.32 Å (without phase correction), but with the distortion of the crystal structure, the coordination number of Si increases significantly, which may be due to the difference in the multiple scattering calculation results caused by the position change of the silicon atom in the center of the (SiO₄)-tetrahedron structure and the distance change between the silicon atom and the surrounding oxygen atoms. Corresponding to Figure 8, the height of the curve at 1.32 Å has a sudden change, which may correspond to the obvious change in the position of the Si atom. However, the change in the length of the Si-O bond is small, while the influence on the position of the peak is not reflected.

4. Conclusions

A thermal treatment was utilized for regulating the distortion of kaolin to obtain kaolinite with different degrees of distortion.

The process when the kaolinite crystal structure disappears and transforms into metakaolin during annealing was studied and characterized. The Materials Studio software was used to simulate the dehydroxylation process of two groups of kaolinite during annealing and the model was optimized. The XAFS results of the model were calculated by multiple scattering. The XAFS result for Si shows that the structural changes in the dehydroxylation process of kaolinite will lead to significant changes in the X-ray absorption coefficients of Si, especially the X-ray absorption spectra of Si, which have a certain regularity. With the progression of dehydroxylation, the characteristic peak position of the Si atom and (SiO₄)-tetrahedron shifts to a low-energy position, and the intensity increases abruptly. The results were Fourier-transformed into R-space, and it was observed that with the dehydroxylation, the k-space data of Si obviously deviated from the initial value. In particular, the phenomena in R-space corresponded to changes in the spatial structure around silicon atoms, which is caused by the lack of hydroxyl. The distortion of (SiO₄)-tetrahedron to a minimum-energy structure resulted in changes in the position of the Si atom and in the length of the Si-O bond.

This work is devoted to exploring the structural characteristics and the identification mechanism of kaolinite with different distortions, which may be helpful to identify kaolinite in coal gangue. The crystallographic structure of kaolinite is different from that of amorphous coal in its natural state, and this may make it easier for scientists to identify whether there is a crystal structure. However, it is necessary to study the low crystallinity and excessive impurities of kaolinite further in actual coal gangue.