In situ spectroscopic studies of decomposition of ZrSiO₄ during alkali fusion process using various hydroxides

Abstract

ZrSiO₄ powder synthesized by the sol–gel method is used to study the reaction mechanism of natural zircon mineral treated by an alkali fusion method. The reaction processes are analyzed by in situ Raman spectroscopy. Other characterization experiments using techniques, such as FTIR spectroscopy, TG-DTA and X-ray powder diffraction complement and verify the Raman spectroscopy study. The results reveal that hydroxylation–dehydration play important roles during the alkali-fusion process. Hydroxylation action breaks some bonds between Si and O in silicon–oxygen tetrahedral and releases zircon as ZrO₂ and silanol groups and then dehydration makes SiO₄ tetrahedron polymerize to the silicate with different lattice structure. The electron-donating ability of the O atom in molten alkali determines the number of Si–O bonds broken in SiO₄ tetrahedral and the lattice structure of silicate products. Finally, with further development of the reaction, the bridge-oxygen bonds of intermediate products are substitutes for non-bridging-oxygen bonds.

---

Introduction

Zircon (ZrSiO₄) is the most abundant zirconium ore and the main source for many industrial useful zirconium compounds, e.g. Zr, ZrO₂ and ZrC. These zirconium compounds are widely used in modern industries, such as nuclear energy, catalysts, catalyst supports, ceramics, fuel cells, pigments, and automotive gas sensors.^1–3 Zircon is one of the most chemically stable compounds because of the high coordination of bisdisphenoid ZrO₈ in a tetragonal structure with SiO₄ tetrahedrons. Therefore, more aggressive reaction conditions are required for the decomposition of the stable structure of silicates and release of the value metal from the silicate network.

The alkali-fusion process is an effective operation for extracting the valuable components from silicate-bearing ores. During this process, the strong binding between zirconium and silicon parts is disrupted in the compound by alkali attack. Therefore, it can lead to a high recovery of zirconium from the activated clinker by the subsequent acid-leaching process under moderate conditions. However, some investigators reported that there are some water-insoluble compounds, such as Na₂ZrSiO₅ and Na₄Zr₂Si₂O₁₂ when the less caustic Na₂CO₃ has been used. That means the chemical properties of alkalis play an important role in the decomposition of zircon and the deep utilization of the Zr resource in the alkali fusion process.

During the past few years, previous investigations have elucidated some aspects of the dynamics and thermodynamics of the alkali fusion processes. All of these studies are used to develop a systematic understanding of the joint process of alkali-roasting activation and acid leaching for the extraction of the valuable metal from zircon, but the studies about the mechanism of alkali fusion at the molecular level, particularly on the breakdown and reconfiguration of chemical bonds have been less frequently reported. The conventional research focuses on the starting materials and the final products of alkali fusion and seldom considers some short-life intermediates during the process. The big problem for these studies is to consider the phase constitute of the quenching sample after the alkali fusion as the same as that of the mixture during the alkali fusion. However, experimental evidences have proved that there are some differences between the microstructure of melt at high temperatures and after quenching.^4,5 In order to study the mechanism of alkali fusion in detail, it is necessary to use a method that can observe the chemical state changes of the lattice structure in situ at high temperature during the occurrence of the alkali reaction.

Raman spectroscopy measurement is an accurate, nondestructive test of the characterization of materials.^6,7 In the past few decades, Raman spectroscopy studies have been applied to analyze crystal and amorphous silicates both experimentally and theoretically.^8–10 In particular, many studies by Raman spectroscopy have revealed the structure of materials at different temperatures and pressures.^11–13 These studies show that changes in Raman mode frequencies with the elevated temperatures or pressures can provide information on the chemical bonding characteristics and the crystal structures in the reaction system.

In the present work, FTIR spectroscopy and TG-DTA measurements are used to study the chemical behavior of hydroxyls during the alkali fusion reaction. These techniques had been used for identification of the hydroxyl groups of silicon oxide thin films and nanoparticles.¹⁴ We report the research on the reactions of zircon with NaOH, KOH and Na₂CO₃, respectively, by the in situ Raman spectroscopy, FTIR spectroscopy, X-ray diffraction and TG-DTA measurements. The objective of the study is to obtain a comprehensive understanding of the mechanism of the zircon-alkali reaction, and reveal the reaction pathway of zircon under different reaction conditions.

Experimental

Minerals

The synthesis of zircon (ZrSiO₄) powder samples were prepared by the sol–gel method as described in the literature.¹⁵ The ZrSiO₄ powder with high purity was prepared by using ZrOCl₂ solution and fumed SiO₂ as the starting materials. Mixtures of tetraethyl orthosilicate (TEOS) and ZrOCl₂ solution were hydrolysed using a mixture of acid, water, and ethanol with a volume ratio of 1 : 3 : 7, respectively. After gelation, additional water was added and stirred to break up the gel and the liquid. The mixture was aged for 12 h and slowly dried, then calcined at 1000 °C for 1 h.

Experimental methods

Raman spectroscopy. The ZrSiO₄ and alkali were fully mixed and ground in a mortar, then put into a platinum crucible measuring 5 mm in diameter and 3 mm deep. The reactor mixture was placed in an electric furnace with a heating rate of 3 °C min⁻¹ for all experiments. The mixture of ZrSiO₄ and alkali was placed on the Raman spectrometer (Jobin Y′ von U1000). The high-temperature Raman spectroscopy analysis was carried out in time intervals of 5 min. The pulsed exciting light (355 nm) from a Q-switch pulsed THG-Nd:YAG laser was focused by an Olympus BH-2 microscope, and the Raman scattering light from the reactants was collected by a confocal-lens system, which included an intensified charge-coupled device (ICCD). Raman spectra are excited by a He Ne laser (532 nm) at a resolution of 2 cm⁻¹ and the spatial resolution about 1 μm in the range of 200 to 1200 cm⁻¹. Spectra are calibrated using the 520.5 cm⁻¹ line of a silicon wafer. In order to ensure that the correct spectra are obtained, the incident excitation radiation is scrambled.

FTIR spectroscopy. With the development of the reaction, the reactor was removed from the furnace at various temperatures and allowed to cool. The fusion product was leached with methanol to remove the unreacted NaOH. The sample was dried and measured using the KBr pellet technique. FTIR spectra were obtained using a Nicolet Nexus 740 FTIR spectrometer with a smart endurance single bounce diamond ATR cell. Spectra over the 4000–400 cm⁻¹ range were obtained by the co-addition of 32 scans with a resolution of 2 cm⁻¹. Spectroscopy manipulation, such as baseline adjustment, smoothing and normalization is performed using the Nicolet Omnic software (Nicolet Corporation, England, 2011).

X-ray diffraction analysis. The X-ray diffraction (XRD) patterns were collected with a D/Max-2500PC X-ray diffractometer. The Cu-Kα (0.15406 nm) radiation, 1° divergent slit, 0.15 mm receiving slit and 5.1° soller slit system are used in all measurements. The data acquisition is done in step-scan mode (step size 2θ = 0.02°, scan speed = 4° (2θ) per min) in angular range 2θ = 0°–90°. Structural refinement analyses of XRD data collected from zinc oxide powders were performed using the Search-Match software.

TG-DTA analysis. The change of mass and calorific effects of the reaction were analyzed by the method of TG-DTA analysis. The sample was put into a platinum crucible and heated in a nitrogen atmosphere at a flow rate 100 ml min⁻¹. TG-DTA analysis experiments were carried out on a Bähr STA 503 thermal analyzer with a heating rate of 15 °C min⁻¹, up to 700 °C.

Results and discussion

The characterization of prepared ZrSiO₄

The ZrSiO₄ powders obtained by a sol–gel method were characterized by X-ray powder diffraction. As shown at Fig. 1, the calcined product phase was found to be zircon (JCPDS no. 42-547). The diffraction peaks prove that the synthesis of ZrSiO₄ has good crystallinity and less spectroscopic interferences from impurities exist.

Fig. 1 X-ray diffraction patterns and lattice structure of the ZrSiO₄ calcined at 1000 °C for 4 h.

Fig. 2 shows a series of Raman spectra of ZrSiO₄ heated from 25 °C to 600 °C. Previous theoretical analysis indicated that there are 12 Raman active modes (2A_1g + 4B_1g + B_2g + 5E_g) for the zircon-type structure with the space group I₄₁/and (D¹⁹_4h).¹⁶ In our work, five Raman peaks were assigned as the internal vibration modes of SiO₄ tetrahedrons: 1008 cm⁻¹ (B_1g, υ₃, antisymmetric stretching), 975 cm⁻¹ (A_1g, υ₁, symmetric stretching), 641 cm⁻¹ (B_1g, υ₄, antisymmetric bending), 545 cm⁻¹ (E_g, υ₄, antisymmetric bending), and 439 cm⁻¹ (A_1g, υ₂, bending).¹⁷ The peaks at 182, 192, 207 and 215 cm⁻¹ are assigned to the vibration of the silicate network, which is induced by metal cations.¹⁸ It is clear that the width and intensity of all the bands keep invariant with the rise in temperature. These features are indicative of the angle and the lengths of the chemical bonds of ZrSiO₄ keeping nearly constant when the temperature is below 600 °C. The reason is that ZrSiO₄ is very stable as representative of nesosilicate, and the isolated SiO₄ tetrahedron cannot polymerize after normal heat treatment.^19–24 The Raman spectra of ZrSiO₄ show no distinct vibration modes associated with the Zr–O vibration of the ZrO₈ polyhedral at a higher wavenumber region than 200 cm⁻¹. This is an indication of weak or absent covalent bonding within the ZrO₈ polyhedral and in-between neighboring Zr⁴⁺ cations. As a result, the ZrO₈ polyhedral are bound together by coulombic forces and have shorter vibration lifetimes relative to those of the stable SiO₄ tetrahedron. This, in turn, implies that the key issue of structural transmission is the breakage and recombination of Si–O in the ZrSiO₄ lattice.

Fig. 2 Raman spectra of prepared ZrSiO₄ with the temperature rise up from 25 °C to 600 °C.

The reaction of ZrSiO₄ with NaOH

Raman spectroscopy measurement of NaOH fusion reaction. The variations in Raman patterns of ZrSiO₄ decomposed by NaOH with the rise in the reaction temperature are shown in Fig. 3 and 4. In the lower wavenumber region, we can observe the peaks at 353 and 439 cm⁻¹ (A_1g and υ₂, symmetric bending) corresponding to Q₀ in ZrSiO₄ broadening and shifting slightly to lower wavenumbers with the rise in temperature. Such red shifts in the frequencies of the stretching modes have been previously observed in Raman spectroscopic studies of inorganic crystals and melts²⁵ and are attributed to the weakening of the central atomic ligand bonding due to volume expansion. In the meantime, some new Raman peaks appear between 280 and 640 cm⁻¹. The spectra of zirconium compound shown in the literature prove that the position of new vibration peaks is similar to those of ZrO₂.²⁶ At the initial stage of the reaction, Raman peaks at 333, 381, 614 and 637 cm⁻¹ assigned to ZrO₂ (ref. 27) are observed on these samples. In particular, Raman patterns at 476 and 557 cm⁻¹ due to three-dimensional amorphous ZrO₂,²⁸ which were often observed at ZrO₂–SiO₂ glasses, can be detected. With the rise in temperature, Raman peaks at 287 and 466 cm⁻¹ appear to indicate that tetragonal ZrO₂ (ref. 29) might have formed when the temperature is above 400 °C. Compared with the spectra of reaction of ZrO₂ with NaOH, the bands above 400 °C imply that the reaction pathway of NaOH and zirconium oxide species released from the network of silicate is similar to that of the reaction of ZrO₂ and alkali. The final product is all zirconate. Raman spectroscopy studies indicate that the zirconium are separated from the silicate network probably in the form of amorphous oxides and transforms to those of tetragonal with the rise in temperature, and the zirconium oxide then reacted with alkali to produce zirconate.

Fig. 3 Raman spectra of ZrSiO₄ and ZrO₂ during the NaOH fusion process with the temperature rising from 25 °C to 600 °C in the 150–700 cm⁻¹ region (the molar ratio of NaOH to ZrSiO₄ is 6 : 1).

Fig. 4 Raman spectra of ZrSiO₄ during the NaOH fusion process with the temperature rising from 25 °C to 600 °C in the 600–1200 cm⁻¹ region (the molar ratio of NaOH and ZrSiO₄ is 6 : 1).

Fig. 4 shows the change of the vibration mode of Si–O during the NaOH fusion reaction. The broadening of peaks with rising temperature is usually induced by the stretching vibration of the secondary structural units in different coordination environments. Partial weakness of the Raman spectra for the sample when temperature rise from 200 °C to 300 °C has been often observed in the glass system calcined at various temperatures. No analogies have been discussed in the accessible literature on alkali fusion. However, similar phenomena were observed when studying the Raman spectroscopy of ceria doped by rare-earth cations.³⁰ In these systems, some ordering in the defects distribution in the lattice is able to cause the disappearance of the Raman peaks. This suggests that in a similar way, ordered defects-polysynthetic twins occur with reconfiguration of SiO₄ tetrahedron in an alkali fusion reaction.

In other studies of high temperature, 1008 cm⁻¹ (B_1g, υ₃, antisymmetric stretching) is considered a correlation to average bond distance of Si–O–Zr and crystal cell volume.²⁷ In the present work, the peak at 1008 cm⁻¹ shifts down to 984 cm⁻¹ and broadens with increasing temperature. An earlier Raman study has reported similar behavior in metamict zircon samples.³¹ These Raman changes have been used to estimate or determine the degree of radiation damage in natural zircon and other geological events that natural crystals might have experienced.³² It may imply that the alkali induced similar defective crystalline domains and structural disorders of crystalline grains in silicate arrays as alpha-decay radiation or metamictization does in natural zircon.³³ The decreasing bond strength of Si–O–Zr accompanying crystal cell volume expansion and structural disorders in lattice denote amorphous ZrO₂ from the network of silicate when reaction temperature is up to 300 °C.

When temperature reaches 300 °C, the stretching mode of the bond between silicon and non-bridging oxygen (Si–O⁻) shifts to the wavenumber assigned to the vibration of the bond between silicon and the bridging oxygen (Si–O–Si),³⁴ which appear at 1050 cm⁻¹. It indicates that the SiO₄ tetrahedron combines with each other by sharing an oxygen atom when zirconium oxide departs from the silicate network. The absence of other Si–O–Si Raman features may indicate that only the most localized modes around Si, such as Si–O stretch, may be present in the reaction system and the other mode of the Si sites, similar to the glass structure, has not taken place drastically in the fusion reaction. When the temperature is up to 400 °C, the vibration peaks at 837, 859, 867 cm⁻¹ assigned to Si–O⁻ appear again.³⁵ Concomitant with the disappearance of the Si–O–Si peak, the peaks of asymmetric stretching vibration of Na₄SiO₄ are clearly observed at 802 cm⁻¹ when the temperature is above 500 °C. It provides the most definitive evidence for the absence of any Si–O–Si linkages in the reaction system, and the structure of the silicate contains solely the isolated Q₀ species. The result shows the reaction continues up to the point where the bonds of Si–O–Si in the framework are broken thoroughly. Nesosilicate is finally obtained.

Raman measurement of OH⁻ in NaOH fusion reaction. Compared with pure ZrSiO₄, all Raman spectra of the reaction system exhibit an additional feature at high temperature. As is shown in Fig. 5, a weak peak at 970 cm⁻¹ and the Si–O–Si vibration mode appear simultaneously at about 300 °C. In the ZrO₂–SiO₂ glasses system, a weak Raman peak observed at about 970 cm⁻¹ has been assigned to Si–OH linkages.³⁶ In the FTIR spectra of the quench residues of NaOH fusion reaction at 300 °C, the peaks at 1350 and 1411 cm⁻¹ are assigned the bending vibration of coordinated and bridging hydroxyls, respectively.^37,38 Chernorukov and Kortikov³⁹ attributed these wavenumbers of bands to silanol, Si–OH. The spectroscopy studies indicate that silanol appears as intermediate products when the SiO₄ tetrahedrons polymerize at alkali fusion reaction at 300 °C.

Fig. 5 Raman and FTIR spectra of ZrSiO₄ and the products of alkali fusion at 300 °C (a: Raman spectra and b: FTIR spectra).

The spectroscopy analyses show that hydroxylation and dehydration play an important role in the structural transformation of silicate. We speculate that free OH⁻ in the molten NaOH breaks two Si–O bonds of the SiO₄ tetrahedron in the ZrSiO₄ lattice, and the Si atom then forms chemical bond with hydroxyl. This also explains why Zr departs from the silicate array in the form of oxide. The thermal process then leads to two neighboring Si–OH approaching each other and condensing to form a Si–O–Si bond. As a result, all Q₀ units transform into Q₂ units. A one-dimensional, two-connected silicate framework forms during the condensation process. The process of hydroxylation and dehydration condensation of the silicate is shown in Fig. 6.

Fig. 6 Proposed routes of hydroxylation and dehydration condensation.

Phase analysis of NaOH fusion reaction products. To determine the phase of the products, X-ray diffraction is used to detect the residues after reactions at various conditions. Fig. 7 shows XRD patterns of residues for NaOH fusion at different temperatures. ZrO₂ and Na₂SiO₃ can be detected from the XRD patterns as the reaction products when the alkali-fusion process was performed at 300 °C. After the temperature rising up to 600 °C, it can be observed that the diffraction peaks of Na₄SiO₄ and Na₂ZrO₃ are predominant, and the diffraction peaks of ZrO₂ become weak. The results of XRD suggest that the starting materials produce ZrO₂ and Na₂SiO₃ at an initial stage of NaOH fusion process and can be finally converted to Na₄SiO₄ and Na₂ZrO₃.

Fig. 7 X-ray diffraction patterns of residues from ZrSiO₄ after the NaOH fusion process at 300 °C (a) and 600 °C (b) for 1 h (the molar ratio of NaOH and ZrSiO₄ is 6 : 1).

Fig. 8 Schematic pathway of ZrSiO₄ decomposed by NaOH.

Above all, the schematic structural description of the pathway of ZrSiO₄ decomposed by NaOH is shown in Fig. 8. The zircon crystal cell consists of four ZrSiO₄ molecules. The coordination number of O is three and Si is four, as to construct SiO₄ tetrahedron, whereas Zr with the coordination number is eight construct dodecahedra. ZrO₈ dodecahedra occupy the symcenter of hexahedron-space made up by the SiO₄ tetrahedron. At the initial stage of the NaOH fusion reaction, Zr is separated from the silicate array in the form of oxide and polymerization of SiO₄ tetrahedron is achieved by the corner-sharing oxygen atom. The Si polyhedral forms a meandering chain of corner-share SiO₄ polyhedral parallel to [001], with the chain joined along [100] by edge-sharing NaO₈ dodecahedra. With the loading of alkali, NaO₈ dodecahedra insert between SiO₄ tetrahedrons to make Na and Si polyhedral stagger parallel to [001] by corner-sharing oxygen. On the other hand, ZrO₈ dodecahedra linked by edge-sharing in the ZrO₂ array are separated by alkali. Na atoms insert into the gap of the network of zirconium–oxygen polyhedral and make the coordination number of Zr decrease to six. In the final product lattice, Na is bounded in the network of ZrO₆ octahedron layer by edge-sharing along [111].

Fig. 9 Raman spectra of ZrSiO₄ during the KOH fusion process with temperature rising from 25 °C to 600 °C (the molar ratio of KOH to ZrSiO₄ is 6 : 1).

The reaction of ZrSiO₄ with KOH

Raman spectroscopy measurement of KOH fusion reaction. Fig. 9 shows the change of the vibration mode of ZrSiO₄ during KOH fusion reaction. Irrespective of the presence of K₂CO₃ obtained by the reaction of KOH and atmosphere, spectra exhibit the presence of different features compared with silicate observed in NaOH fusion reaction, particularly for the bands between 600 and 800 cm⁻¹. It is noteworthy that the Raman spectra for the sample disappear when temperature rise up to about 300 °C. This feature is similar to that of alkalis metal silicate glass and peak position shift significantly compared to NaOH fusion reaction. These spectra features indicate that disorder of SiO₄ tetrahedron network and the additional internal stress exist in lattice when temperature rise up to 300 °C.⁴⁰ We speculate that rapid decomposition of ZrSiO₄ induces vacancy distribution in the lattice and amorphization of zircon crystalline. Large quantity silicate units with unoccupied orbital and electric charge may exist and recompose simultaneously at fusion reaction. The high-wavenumber band between 800 and 1100 cm⁻¹ covers the symmetric vibrations of the Q_i (3 ≥ i ≥ 0) species. These peaks appear and enhance at the initial stage of reaction. In particular, the bending mode of Q₃ units give rise to the peak designated at 1053 cm⁻¹ and 547 cm⁻¹.⁴¹ The presence of Q₃ species proves that more Si–O⁻ bonds in SiO₄ tetrahedron are broken in KOH than NaOH fusion reaction. As a result, the SiO₄ tetrahedrons form layer network. When temperature rises up continuously, the intensity of bands at 868 and 915 cm⁻¹ assigned to Si–O stretch vibration mode in the silicate tetrahedron unit with two no bridge-oxygen (NBO),⁴² increase. The result in that excessive alkali can disrupt the network of Q₃ and make phyllosilicate decompose to chain units. The decomposition of intermediate product is gradual.

On the basis of the aforementioned results, it can be concluded that the pathway of zircon decomposed by KOH fusion is similar to the NaOH fusion reaction. Molten KOH breaks the Si–O⁻ bond in the ZrSiO₄ lattice and forms Si–O–Si through the mode of hydroxylation–dehydration condensation, but the base strength improves the extent of damage of the SiO₄ tetrahedron, resulting in different intermediate products. The molten KOH breaks three Si–O⁻ bonds of SiO₄ tetrahedron, and the Si atom forms a bond with OH⁻ in three directions, then silanol transforms to phyllosilicate by dehydration condensation of Si–OH at high temperatures. Thus, the structure of the products is a two-dimensional, three-connected framework. The structure of the silicate transforms from Q₀ to Q₃. The hydroxylation and dehydration process of the KOH fusion reaction is schematically shown in Fig. 10.

Fig. 10 Diagram of hydroxylation and dehydration process: (a) hydroxylation and (b) dehydration.

The phase analysis of KOH fusion reaction

Fig. 11 shows the X-ray diffraction pattern of the residues for the KOH fusion at different temperatures. The results show K₄ZrSiO₄ appears when the alkali-fusion process was performed at 300 °C. In ZrO₂ and similar compounds, the decomposition of OH⁻ absorbed on the surface of the lattice often formed oxygen vacancy in the lattice and made the valence state of the Zr ion change.⁴³ We infer that abundant OH⁻ forms chemical bonds at the surface of the ZrSiO₄ particle in KOH than in the NaOH fusion reaction, and the decomposition of adsorptive OH⁻ then made the valence state of Zr decrease at high temperature. The oxygen molecules diffuse into the interior and consequently, occupy oxygen vacancies to promote Zr²⁺ transform to Zr⁴⁺ again. That is why low valence Zr compounds disappear with rising temperature. When the temperature rises to 600 °C, K₂Si₂O₅ (a type of phyllosilicate) and K₂ZrO₄ can be recognized as final products of the alkali fusion.

Fig. 11 X-ray diffraction patterns of residues from ZrSiO₄ after the KOH fusion process at 300 °C (a) and 600 °C (b) for 1 h (the molar ratio of KOH to ZrSiO₄ is 6 : 1).

A schematic description of the phase transformation of ZrSiO₄ decomposed by KOH as deduced from the Raman spectra, XRD is shown at Fig. 12. The initial stage of the KOH fusion reaction is similar to that of NaOH. ZrO₂ is separated from the silicate array and SiO₄ tetrahedrons connect with each other by sharing apical O atoms in three-space directions. The Si polyhedral forms a six-membered ring, KO₆ hexahedron occupies the slack in the centricity of ring and between the layers by sharing a corner with the SiO₄ tetrahedron. On the other hand, ZrO₈ dodecahedra linked by edge-sharing in the ZrO₂ array are separated by alkali. K atoms insert into the gap of the network of zirconium–oxygen polyhedral and make the coordination number of Zr decrease to six. Finally, the K atom is bonded between layers of Zr–O polyhedron connected with each other by edge-sharing.

Fig. 12 Schematic pathway of ZrSiO₄ decomposed by KOH fusion.

The reaction of ZrSiO₄ with Na₂CO₃

Phase analysis of KOH fusion reaction. On the basis of the previous results, the fusion reaction of ZrSiO₄ and Na₂CO₃ can be carried out to study the effect of weaker alkali on the microstructure of intermediate products of alkali fusion. Temperatures depending on the Raman spectra of ZrSiO₄ during the Na₂CO₃ fusion reaction are shown in Fig. 13. Compared with the NaOH fusion reaction, some important differences should be emphasized, such as no vibration mode of Zr–O and Si–O–Si can be observed at the initial stage of the reaction. Instead, the peaks at 973 and 877 cm⁻¹ corresponding to new Q₀ appear immediate and enhanced with a rise in temperature.^44–47 This feature implies that the SiO₄ tetrahedron directly rearranges from the original isolated structure, and no intermediate products with the array of Si–O–Si appear as the NaOH and KOH fusion reaction. In addition, there is no release of the zirconium oxide unit from the silicate with the loading of alkaline oxide. A conclusion can be drawn that the decomposed interaction between alkaline carbonates and zircon is governed by various mechanisms attributed to NaOH and KOH. The chemical bond stability of the homogeneous elementary substance (Zr and O) is different when silicate is decomposed by alkali. A weaker base can change the coordination properties of cations by the loading of alkaline oxide but not completely break all of the Zr–O–Si bonds.

Fig. 13 Raman spectra of ZrSiO₄ during the Na₂CO₃ fusion process with the temperature rising from 25 °C to 900 °C (the molar ratio of Na₂CO₃ to ZrSiO₄ is 3 : 1).

The X-ray diffraction patterns of ZrSiO₄ decomposed by Na₂CO₃ at 900 °C at different times are shown in Fig. 14. As can be observed, only the phase of Na₂ZrSiO₅ was indexed in residue after alkali fusion. The schematic description of the phase transformation of ZrSiO₄ decomposed by Na₂CO₃ is shown in Fig. 15. With the loading of Na and O atoms, an eight-coordinated zirconium monomer is divided into a six-coordinate monomer and joined to a NaO₆ hexahedron by a sharing edge. SiO₄ tetrahedron pairs corner-linked to the ZrO₆ octahedral are isolated from the others so that the Si–O vibrations are well localized. The distances between oxygen and zirconium atoms vary from 1.97 Å to 2.17 Å, while the value is 2.18 Å and 2.23 Å in ZrSiO₄, which form a highly distorted square pyramid. The decreasing bond length of Zr–O makes Zr more difficult to depart from the silicate network. Phase analysis proves that the arrangement of Zr and O atoms is distorted, but the linkage of Zr–O–Si is not broken totally in the Na₂CO₃ fusion reaction. The Zr atom is still in the nesosilicate with the addition of oxygen to the lattice of silicate, while Na₂ZrSiO₅ is gained as the final product.

Fig. 14 X-ray diffraction patterns of residues from ZrSiO₄ after the Na₂CO₃ fusion process at 900 °C for 30 min (a), 1 h (b) and 2 h (c) (the molar ratio of Na₂CO₃ to ZrSiO₄ is 3 : 1).

Fig. 15 Schematic diagram of ZrSiO₄ decomposed by Na₂CO₃.

The effect of base strength on the transformation of silicate in molten alkali

The most important issues concerning the different approaches in the study of decomposition and re-composition of the silicate in alkali fusion is to evaluate and explain the various experimental conditions resulting in different observations. By the SHAB theory, all of the atoms and ions in the alkali fusion reaction belong to hard acids and bases. The larger the orbital energy difference of the lowest unoccupied molecular orbital (LUMO) of acid and the highest occupied molecular orbital (HOMO) of the base, the stronger is the chemical bond of reaction product. Because outer shell electron distribution of Zr and Si are 4d2 5s2 and 3s2 3p2, respectively, the orbital energy difference of LUMO of Si and HOMO of OH⁻ is larger than that of Zr and OH⁻, which induce Si–O⁻ to break down, and OH⁻ forms a chemical bond with the Si atom. It also means that the amount of fractured Si–O⁻ in Si tetrahedron determines the structure of condensation products. Our study shows the microstructure of alkali fusion products varies with regularly increasing the base strength. The simulated structure shows that the anionic SiO₄ tetrahedron in intermediate products are all held together by the metal cation, the preferential coordination of which is retained by the flexible orientation of the terminal oxygen of the SiO₄ tetrahedron in these alkaline rich suborthosilicate systems. The important difference in the structure is the connection mode of Si and O atom. The lattice diagram of the products prove that the structure of the NaOH fusion products belongs to a one-dimension space of silicate and that of the KOH fusion products is two, whereas the basic structural unit of Na₂CO₃ fusion reaction is still isolated, and Zr cannot be separated from silicate lattice. The reason can be revealed by the FTIR spectra of the ZrSiO₄ and residue obtained from ZrSiO₄ after NaOH, KOH and Na₂CO₃ fusion reactions shown in Fig. 16. Three vibration regions can reflect the variation law of the Si–O⁻ bands broken at different base strengths. FTIR peaks at 1081 and 975 cm⁻¹, corresponding to stretching vibration modes of the Si–O⁻ of SiO₄ tetrahedron, diminished gradually and were substituted by the peaks at 1013 cm⁻¹, corresponding to Si–O.⁴⁸ It indicates that the coordination mode of the Si–O changed completely. In contrast, the intensity of the Si–O⁻ bands diminished and did not disappear completely in reaction with Na₂CO₃. The variation of spectra shows that the Si–O stability of SiO₄ tetrahedron is inconsistent in different alkali and less Si–O⁻ break in a weaker molten base. From the view of a bonding property, the effective electron charge of an O atom in a molten state determines the number of Si–OH in a silanol molecule. The lower electronegativity of alkali metal cation makes the electron density of the O atom higher in KOH than that in NaOH, which can break more Si–O bonds of the SiO₄ tetrahedron to form Si–OH bonds. In contrast, the electronegativity of C is higher than H, the electron density of the O atom is lower in [CO₃]²⁻ than that in OH⁻, which leads to [CO₃]²⁻ being more difficult to approach to Si atom to break SiO₄ tetrahedron. In a word, the pathway of the reaction and the structures of products depend mainly on the electron-donating ability of the O atom in molten alkali. As the reaction medium becomes highly electron donating, more Si–O are broken and form crystals with more complicated structure.

Fig. 16 FTIR spectra of the residues from ZrSiO₄ after the different fusion processes.

Conclusions

Spectroscopic and TG-DTA analysis show that alkali breaks the Si–O bond in the SiO₄ tetrahedron through hydroxylation, and Zr departs from the array of silicate in the form of ZrO₂, then neighboring Si–OH groups condensate to form Si–O–Si bonds. The polymersilicate and ZrO₂ can gradually react with excessive alkali. Therefore, zirconate and nesosilicate is gained. The structure of the intermediate product rests in the electrons affording ability of alkali in the reaction. Stronger alkali can break the abundant Si–O⁻ bonds in the SiO₄ tetrahedron, ensuring a multi-dimensional silicate structure form.

Acknowledgements

The authors gratefully acknowledge financial support from the National Natural Science Foundation of China (Grant no. 51174055).

References

1. H. P. Scott, Q. Williams and E. Knittle, Phys. Rev. Lett., 2001, 88, 015506.
2. S. E. Jackson, N. J. Pearson, W. L. Griffin and E. A. Belousova, Chem. Geol., 2004, 211, 47–69.
3. H. Fujimaki, Contrib. Mineral. Petrol., 1986, 94, 42–45.
4. J. You, G. Jiang and K. Xu, J. Non-Cryst. Solids, 2001, 282, 125–131.
5. M. Zhang, E. K. Salje and R. C. Ewing, J. Phys.: Condens. Matter, 2002, 14, 3333.
6. H. Aguiar, E. Solla, J. Serra, P. González, B. León, N. Almeida, S. Cachinho, E. Davim, R. Correia and J. Oliveira, J. Non-Cryst. Solids, 2008, 354, 4075–4080.
7. M. Muniz-Miranda, C. Gellini and L. Bindi, Spectrochim. Acta, Part A, 2009, 73, 456–459.
8. H. P. Scott, Q. Williams and E. Knittle, Geophys. Res. Lett., 2001, 28, 1875–1878.
9. R. Syme, D. Lockwood and H. Kerr, J. Phys. C: Solid State Phys., 1977, 10, 1335.
10. B. Kolesov and J. Tanskaya, MRS Bull., 1996, 31, 1035–1044.
11. A. Kalampounias, N. Nasikas and G. Papatheodorou, J. Chem. Phys., 2009, 131, 114513.
12. N. Choudhury, S. Ghose, C. P. Chowdhury, C. Loong and S. Chaplot, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 58, 756.
13. M. B. Smirnov, S. V. Sukhomlinov and K. S. Smirnov, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 82, 094307.
14. G. Cerrato, S. Bordiga, S. Barbera and C. Morterra, Appl. Surf. Sci., 1997, 115, 53–65.
15. G. Chuah, Catal. Today, 1999, 49, 131–139.
16. P. Dawson, M. Hargreave and G. Wilkinson, J. Phys. C: Solid State Phys., 1971, 4, 240.
17. D. A. McKeown, I. S. Muller, H. Gan, I. L. Pegg and C. A. Kendziora, J. Non-Cryst. Solids, 2001, 288, 191–199.
18. A. Golubović, M. Šćepanović, A. Kremenović, S. Aškrabić, V. Berec, Z. Dohčević-Mitrović and Z. Popović, J. Sol-Gel Sci. Technol., 2009, 49, 311–319.
19. M. Zhang, E. K. Salje, I. Farnan, A. Graeme-Barber, P. Daniel, R. C. Ewing, A. M. Clark and H. Leroux, J. Phys.: Condens. Matter, 2000, 12, 1915.
20. D. McKeown, F. Galeener and G. Brown Jr, J. Non-Cryst. Solids, 1984, 68, 361–378.
21. P. McMillan, B. Piriou and A. Navrotsky, Geochim. Cosmochim. Acta, 1982, 46, 2021–2037.
22. F. Galeener and P. Sen, Phys. Rev. B: Solid State, 1978, 17, 1928.
23. S. A. Brawer and W. B. White, J. Non-Cryst. Solids, 1977, 23, 261–278.
24. N. Zotov and H. Keppler, Am. Mineral., 1998, 83, 823–834.
25. J. L. You, G. C. Jiang, H. Y. Hou, H. Chen, Y. Q. Wu and K. D. Xu, J. Raman Spectrosc., 2005, 36, 237–249.
26. S. W. Lee and R. Condrate Sr, J. Mater. Sci., 1988, 23, 2951–2959.
27. L. Nasdala, G. Irmer and D. Wolf, Eur. J. Mineral., 1995, 7, 471–478.
28. M. Zhang and E. K. Salje, J. Phys.: Condens. Matter, 2001, 13, 3057.
29. E. Balan, D. R. Neuville, P. Trocellier, E. Fritsch, J.-P. Muller and G. Calas, Am. Mineral., 2001, 86, 1025–1033.
30. D. W. Matson, S. K. Sharma and J. A. Philpotts, J. Non-Cryst. Solids, 1983, 58, 323–352.
31. G. Papatheodorou and S. Yannopoulos, in Molten Salts: From Fundamentals to Applications, Springer, 2002, pp. 47–106.
32. E. Anastassakis, B. Papanicolaou and I. Asher, J. Phys. Chem. Solids, 1975, 36, 667–676.
33. V. G. Keramidas and W. B. White, J. Am. Ceram. Soc., 1974, 57, 22–24.
34. I. R. Lewis and H. Edwards, Handbook of Raman spectroscopy: from the research laboratory to the process line, CRC Press, 2001.
35. V. Sadykov, T. Kuznetsova, G. Alikina, Y. V. Frolova, A. Lukashevich, Y. V. Potapova, V. Muzykantov, V. Rogov, V. Kriventsov and D. Kochubei, Catal. Today, 2004, 93, 45–53.
36. C.-C. Lin, S.-F. Chen, L.-g. Liu and C.-C. Li, J. Non-Cryst. Solids, 2007, 353, 413–425.
37. M. Gaune-Escard, Molten salts: from fundamentals to applications, Springer, 2002.
38. B. O. Mysen and J. D. Frantz, Chem. Geol., 1992, 96, 321–332.
39. N. Chernorukov and V. Kortikov, Russ. J. Gen. Chem., 2001, 71, 1669–1675.
40. Y. Q. Wu, G. C. Jiang, J. L. You, H. Y. Hou, H. Chen and K. Di Xu, J. Chem. Phys., 2004, 121, 7883–7895.
41. J. H. Giles, D. A. Gilmore and M. B. Denton, J. Raman Spectrosc., 1999, 30, 767–771.
42. D. A. Long and D. Long, Raman spectroscopy, McGraw-Hill, New York, 1977.
43. M. Marinšek, J. Maček and T. Meden, J. Sol-Gel Sci. Technol., 2002, 23, 119–127.
44. T. Rojac, M. Kosec, M. Połomska, B. Hilczer, P. Šegedin and A. Bencan, J. Eur. Ceram. Soc., 2009, 29, 2999–3006.
45. R. L. Frost, J. Čejka, M. L. Weier, W. Martens and J. T. Kloprogge, Spectrochim. Acta, Part A, 2006, 64, 308–315.
46. S.-m. Chang and R.-a. Doong, Chem. Mater., 2005, 17, 4837–4844.
47. H. Aguiar, J. Serra, P. González and B. León, J. Non-Cryst. Solids, 2009, 355, 475–480.
48. C. Gautam, A. K. Yadav and A. K. Singh, A review on infrared spectroscopy of borate glasses with effects of different additives, Int. Scholarly Res. Not., 2012, 2012, 1–17.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra12305k

---