03_Abstract (part 1) - Yiwei666/08_computional-chemistry-learning-materials- GitHub Wiki
1. An Ab Initio Molecular Dynamics Simulation of Liquid FeO–SiO2 Silicate System with Sulfur Dissolving
The desulfurization mechanism is of great significance to quality improvement in metallurgical process. In this work, the structural features, chemical and dynamical properties of the liquid FeO·SiO2 were calculated under 2000 K through ab initio molecular dynamics simulations. Further calculation of desulfurization was conducted based on the structural evolution information. The results showed that the liquid FeO·SiO2 is primarily constituted by Si–O and Fe–O bonds, with the former being strong covalent bonds, while the latter showing the feature of ionic bonding. Bader charges analysis indicated that Fe and O have a wide range of charge states, while that of Si is relatively concentrated. It is found that the sulfur atom that is incorporated into the liquid FeO·SiO2 tends to form a stable bonding structure with three iron atoms, and the Si–S bond seems to be unstable thus, unable to exist in the S-doped FeO·SiO2 silicate.
2. Comparison of desulfurization mechanism in liquid CaO-SiO2 and MnO-SiO2: An ab initio molecular dynamics simulation
In present study, systematic investigations of the CaO-SiO2 and MnO-SiO2 slag were performed of the evolution of the structure, and the sulfur dissolution mechanism at a temperature of 2000 K using ab initio molecular dynamics simulations. The results show that the structure and charge of CaO-SiO2 and MnO-SiO2 are very different. Firstly, Si-O has a strong polar bond while Ca-O and Mn-O show ionicity, and Mn-O has weaker ionicity than Ca-O. A small amount of Mn-Mn clusters are found in liquid MnO-SiO2. Secondly, charge distribution depicts that there is less charge around Ca, while there are relatively more charges around Mn. Bader charge analysis indicates that Mn and O have a broad valence distribution in MnO-SiO2 compared with CaO-SiO2. Thirdly, Sulfur prefers to form stable bonds with Mn atoms (Mn-S-Mn), whereas Si-S bonds are unstable and cannot be found in S-doped MnO-SiO2 silicate. However, in the CaO-SiO2 system, the S atom does not undergo rapid bond transitions. The study of the desulfurization mechanism shows that the uneven distribution of charge in MnO-SiO2 system will affect the transformation of oxygen types, resulting in the decrease of bridged oxygen and the increase of non-bridged oxygen. However, there is no charge effect in CaO-SiO2 system, and the non-bridged oxygen will be consumed in the desulfurization process, resulting in the decrease of non-bridged oxygen and the increase of bridged oxygen. This mechanism well explains the experimental results from a more microscopic perspective, which is of great significance to the research on the removal mechanism of S in the metallurgical industry.
3. Effect of alkaline oxides on aluminate slag structure by first principles calculation
To guide the research and development of CaO-Al2O3-based mold fluxes, the influence mechanism of 8 mol% alkaline oxides (Na2O, Li2O, BaO, MgO) on the multi-scale structure characteristics of aluminate slag (CaO/Al2O3 = 1.3) at 1873 K was analyzed by first-principles calculation. With the addition of alkaline oxides, the stability of Al-O and Ca-O bonds in slag system was Na2O-Al2O3-CaO (NAC) ≈ BaO-Al2O3-CaO (BAC) > Li2O-Al2O3-CaO (LAC) ≈ Al2O3-CaO (AC) > MgO-Al2O3-CaO (MAC), and the proportion of average coordination number for Al were AC > MAC > LAC > BAC > NAC. The alkaline cations played two roles of structure depolymerization and charge compensation, as the ability was ranked as NAC > BAC > MAC > LAC, and Na+ > Li+ > Ba2+ > Mg2+, respectively. At the electronic level, a small number of electrons were localized between O atoms, which surrounded Al to form a closed structure. Adding appropriate amount of Na2O/BaO was necessary to effectively regulate the low/non-reactive mold flux.
4. First-principles study on microstructure of CaO-Al2O3-B2O3 slag
The structure evolution of CaO-Al2O3-B2O3 slag at 1873 K with various CaO/Al2O3 ratio in 0.7 to 1.5 was studied by first-principles molecular dynamics. The bond stability was in the order of B-O > Al-O > Ca-O > O-O. In addition, the average coordination number of B and Al was observed as the increase in the CaO/Al2O3 ratio from 0.7 to 1.1, while decrease with further CaO/Al2O3 addition to 1.5. Combining the results of oxygen distribution and FTIR spectra, the slag polymerization degree was decreased with CaO/Al2O3 enhancement. Ca-O was ionic bond, while B-O and Al-O were charge-transfer bond, so that the structure evolutions of B-O and Al-O network were accompanied with charge transfer. Each O bonded with B formed a stable electron cloud connection, which was beneficial to the stability of B-O network structure, and there were few stable electron cloud connections between each O bonded to Al.
5. Stabilization mechanism of arsenic-sulfide slag by density functional theory calculation of arsenic-sulfide clusters
Stabilization of arsenic sulfur slag (As‒S slag) is of high importance to prevent the release of deadly As pollutants into environment. However, the molecular understanding on the stability of As‒S slag is missing, which in turn restricts the development of robust approach to solve the challenge. In this work, we investigated the structure-stability relationship of As‒S slag with adopting various As‒S clusters as prototypes by density functional theory (DFT). Results showed that the configuration of S multimers-covering-(As2S3)n is the most stable structure amongst the candidates by the analysis of energies and bonding characteristics. The high stability is explained by orbital composition that the 4p-orbital (As) binding with 3p-orbital (S) decreases energy level of highest occupied molecular orbital (HOMO). Inspired from the calculations, an excess-S-based hydrothermal method was successfully proposed and achieved to promote the stabilization of As‒S slag. Typically, the As concentration from the leaching test of stabilized As‒S slag is only 0.8 mg/L, which is much lower than the value from other stabilized slag.
6. Ab initio molecular dynamics assessment of thermodynamic and transport properties in (K,Li)Cl and (K, Na)Cl molten salt mixtures
Molten salt mixtures are integral part of highly important technological applications such as nuclear reactors. However, due to inherent difficulties associated with experiment at high temperatures and the intrinsic complexity of liquid-phase multi-component systems, understanding their properties at a molecular level remains a challenge. Here, we report on an ab initio molecular dynamics investigation on structural, electronic, transport, and thermal properties of two common molten salt mixtures, (K, Li)Cl and (K,Na)Cl, at five different compositions and three temperatures. Most of the properties were found to depend on both composition and temperature. While properties, like atomic charges, show additive behaviors, other properties, such as electrical conductivity, show considerable deviations from additivity. We shall show that the mixing of the molten salt mixtures is mainly driven by entropy, and that the KCl and LiCl mix better than KCl and NaCl. Our computational results are in general consistent with available experimental data. Comparison with available theoretical data is also provided.
7. Effects of La2O3 Addition into CaO-SiO2 Slag: Structural Evolution and Impurity Separation from Si-Sn Alloy
Boron (B) and phosphorus (P) are the most problematic impurities to be removed in the production of solar-grade silicon by the metallurgical process. In this work, the distribution of B and P between CaO-(La2O3)-SiO2 slags and Si-10 mass pct Sn melt was experimentally studied. B distribution coefficient increased from 2.93 in binary CaO-SiO2 slag to 3.33 and 3.65 with 2 and 10 mass pct La2O3 additions, respectively. In the followed acid-leaching experiments, the slag-treated Si-Sn alloys exhibited higher B and P removal than that of the initial alloy without slag treatment. Molecular dynamics simulations were performed to study the effect of La2O3 addition on the slag structural and transport properties. A novel oxygen classification method was proposed to distinguish the different structural roles of La and Ca in the CaO-La2O3-SiO2 system. It was found that La3+ prefers to stay in the depolymerized region, mostly connects with 6-7 non-bridging oxygen, and requires a weak charge compensation with Ca2+. Possible silicothermic reduction was evaluated to discuss the slag chemistry and the mass transfer between slag and metal phase. A thermodynamic model was derived to theoretically study the alloying effect on impurity distribution in slag refining where positive interaction coefficient and high alloying concentration were found most beneficial to improve the impurity removal.
8. Estimation of Activity Coefficients and Interaction Parameters of Solutes in Silicon Melts
Because the speed and cost performance of computing has been dramatically increased and computational science has been progressed concomitantly, their utilization should be considered also in thermodynamic property estimation. When thermodynamic property data are not available, their estimation by the use of computation might be useful as their first guess. From this viewpoint, the estimation of activity coefficients of solutes and interaction parameters among them in dilute silicon solution has been tried by the use of ab initio calculation and Monte Carlo simulation.
9. Influence of Glass Composition on the Luminescence Mechanisms of CdSe Quantum-Dot-Doped Glasses
In this work, we characterized the electronic structure of CdSe quantum dots embedded in a series of x Na2O, (1–x) SiO2 glass matrices (x = 0, 0.25, 0.33, and 0.5). We analyzed the impact of the glass matrix composition on both the atomic structure of the quantum dot (QD) and the QD/glass interface, as well as the luminescence mechanisms, using density functional theory calculations. The increase of Na2O content in the glass matrices was found to promote the formation of Cd–O and Se–Na interfacial bonds and disrupt the Cd–Se bonds network. In particular, we show that the glass composition directly affects the nature of the highest occupied molecular orbitals (HOMO). According to the atomic structure, band gap distribution, and density of states calculation, we find that there is significant reconstruction of the QD and that the picture sometimes proposed of a “pristine quantum dot” surrounded by glass is not realistic. The introduction of CdSe QD significantly decreased the HOMO–LUMO gap of the glass compared to pristine glasses, and the interfacial bonds greatly contributed to the frontier orbitals without forming midgap states. We propose a new energy diagram, quite different from the traditional model, to explain the luminescence of CdSe quantum-dot-doped glasses, originating from the intrinsic emission of this hybrid system {QD + glass}. These results improve our understanding of the luminescence of CdSe quantum-dot-doped glasses, explaining the reason for the poor quantum efficiency and broad emission linewidth compared with their colloidal counterparts.
10. Ab Initio Molecular Dynamics of CdSe Quantum-Dot-Doped Glasses
We have probed the local atomic structure of the interface between a CdSe quantum dot (QD) and a sodium silicate glass matrix. Using ab initio molecular dynamics simulations, we determined the structural properties and bond lengths, in excellent agreement with previous experimental observations. On the basis of an analysis of radial distribution functions, coordination environment, and ring structures, we demonstrate that an important structural reconstruction occurs at the interface between the CdSe QD and the glass matrix. The incorporation of the CdSe QD disrupts the Na–O bonds, while stronger SiO4 tetrahedra are reformed. The existence of the glass matrix breaks the stable 4-membered (4MR) and 6-membered (6MR) Cd–Se rings, and we observe a disassociated Cd atom migrated in the glass matrix. Besides, the formation of Se–Na and Cd–O linkages is observed at the CdSe QD/glass interface. These results significantly extend our understanding of the interfacial structure of CdSe QD-doped glasses and provide physical and chemical insight into the possible defect structure origin of CdSe QD, of interest to the fabrication of the highly luminescent CdSe QD-doped glasses.
11. Structure and density of basaltic melts at mantle conditions from first-principles simulations
The origin and stability of deep-mantle melts, and the magmatic processes at different times of Earth’s history are controlled by the physical properties of constituent silicate liquids. Here we report density functional theory-based simulations of model basalt, hydrous model basalt and near-MORB to assess the effects of iron and water on the melt structure and density, respectively. Our results suggest that as pressure increases, all types of coordination between major cations and anions strongly increase, and the water speciation changes from isolated species to extended forms. These structural changes are responsible for rapid initial melt densification on compression thereby making these basaltic melts possibly buoyantly stable at one or more depths. Our finding that the melt-water system is ideal (nearly zero volume of mixing) and miscible (negative enthalpy of mixing) over most of the mantle conditions strengthens the idea of potential water enrichment of deep-mantle melts and early magma ocean.
12. Prediction of crystal–melt partition coefficients from elastic moduli
MANY geochemical processes, such as crystallization of silicate magmas or planetary differentiation, require a knowledge of the way in which elements become partitioned between coexisting crystal and liquid phases1,2. But quantitative prediction of crystal/melt partition coefficients from thermodynamic principles has not previously been possible. By studying the partitioning of 15 elements between silicate minerals and their coexisting melts, we show here that the partitioning behaviour of any series of isovalent cations can be rationalized in terms of a simple model in which the size and elasticity of the crystal lattice sites play a critical role. We find that elasticity varies linearly with the formal charge of the cation. This model allows us to predict element partitioning behav-iour solely from the physical characteristics of the cation sites in the crystal.
13. Structural dynamics of basaltic melt at mantle conditions with implications for magma oceans and superplumes
Transport properties like diffusivity and viscosity of melts dictated the evolution of the Earth’s early magma oceans. We report the structure, density, diffusivity, electrical conductivity and viscosity of a model basaltic (Ca11Mg7Al8Si22O74) melt from first-principles molecular dynamics calculations at temperatures of 2200 K (0 to 82 GPa) and 3000 K (40–70 GPa). A key finding is that, although the density and coordination numbers around Si and Al increase with pressure, the Si–O and Al–O bonds become more ionic and weaker. The temporal atomic interactions at high pressure are fluxional and fragile, making the atoms more mobile and reversing the trend in transport properties at pressures near 50 GPa. The reversed melt viscosity under lower mantle conditions allows new constraints on the timescales of the early Earth’s magma oceans and also provides the first tantalizing explanation for the horizontal deflections of superplumes at ~1000 km below the Earth’s surface.
14. Constraints on the composition of the Earth's core from ab initio calculations
Knowledge of the composition of the Earth's core1,2,3 is important for understanding its melting point and therefore the temperature at the inner-core boundary and the temperature profile of the core and mantle. In addition, the partitioning of light elements between solid and liquid, as the outer core freezes at the inner-core boundary, is believed to drive compositional convection4, which in turn generates the Earth's magnetic field. It is generally accepted that the liquid outer core and the solid inner core consist mainly of iron1. The outer core, however, is also thought to contain a significant fraction of light elements, because its density—as deduced from seismological data and other measurements—is 6–10 per cent less than that estimated for pure liquid iron1,2,3. Similar evidence indicates a smaller but still appreciable fraction of light elements in the inner core5,6. The leading candidates for the light elements present in the core are sulphur, oxygen and silicon3. Here we obtain a constraint on core composition derived from ab initio calculation of the chemical potentials of light elements dissolved in solid and liquid iron. We present results for the case of sulphur, which provide strong evidence against the proposal that the outer core is close to being a binary iron–sulphur mixture7.
15. A solid-state electrolysis process for upcycling aluminium scrap
he recycling of aluminium scrap today utilizing a remelting technique downgrades the quality of the aluminium, and the final sink of this downgraded recycled aluminium is aluminium casting alloys1,2,3,4,5,6,7,8,9. The predicted increase in demand for high-grade aluminium as consumers choose battery-powered electric vehicles over internal combustion engine vehicles is expected to be accompanied by a drop in the demand for low-grade recycled aluminium, which is mostly used in the production of internal combustion engines2,7,10,11. To meet the demand for high-grade aluminium in the future, a new aluminium recycling method capable of upgrading scrap to a level similar to that of primary aluminium is required2,3,4,7,11. Here we propose a solid-state electrolysis (SSE) process using molten salts for upcycling aluminium scrap. The SSE produces aluminium with a purity comparable to that of primary aluminium from aluminium casting alloys. Moreover, the energy consumption of the industrial SSE is estimated to be less than half that of the primary aluminium production process. By effectively recycling aluminium scrap, it could be possible to consistently meet demand for high-grade aluminium. True sustainability in the aluminium cycle is foreseeable with the use of this efficient, low-energy-consuming process.
16. Element Partitioning: The Role of Melt Structure and Composition
We segregated coexisting gabbroic and granitic melts by centrifuging them at high pressures and temperatures and measured the trace element compositions of the melts by laser ablation inductively coupled plasma mass spectrometry. Our results demonstrate that the effect of melt structure contributes about one order of magnitude to crystal/melt partition coefficients. Partitioning of alkali and alkaline earth elements strongly depends on field strength: Amphoteric and lone pair electron elements partition into the polymerized granitic melt; and rare earth, transition, and high field strength elements coordinated by nonbridging oxygens partition remarkably similar into the gabbroic melt. A regular solution model predicts these effects.
17. Partition model for trace elements between liquid metal and silicate melts involving the interfacial transition structure: An exploratory two-phase first-principles molecular dynamics study
Separation of trace elements has become a major obstacle for preparation of high-purity metal materials. Effective design of the separation medium depends on accurate partitioning prediction of the trace elements in the two phases. However, development of a separation prediction model is difficult because of the limitations of recognition of the interfacial transition structures of the trace elements. Here, a partition model for trace elements between metal and silicate melts involving an interfacial transition structure was developed by exploratory two-phase first-principles molecular dynamics simulation. The results showed that the distribution strongly depends on the local coordination structure in the cluster (LCSC) of the impurity element, which is the key for investigating the interfacial transition structure. A computational strategy for the trace-element distribution ratio is proposed to quantify the contribution of the trace element at the interface. The LCSC partition model was demonstrated for trace element boron removal from silicon for molten silicon and silicate melts. The LCSC model gave a better predicted value of the experimental value than the traditional activity model. The new model assists in explaining the transformation mechanism of B atoms at the silicate–silicon interface at the atomic scale. The LCSC partition model allows prediction of the trace element partitioning behavior solely from first-principles calculations.
18. Carbon and other light element contents in the Earth’s core based on first-principles molecular dynamics
Carbon (C) is one of the candidate light elements proposed to account for the density deficit of the Earth’s core. In addition, C significantly affects siderophile and chalcophile element partitioning between metal and silicate and thus the distribution of these elements in the Earth’s core and mantle. Derivation of the accretion and core–mantle segregation history of the Earth requires, therefore, an accurate knowledge of the C abundance in the Earth’s core. Previous estimates of the C content of the core differ by a factor of ∼20 due to differences in assumptions and methods, and because the metal–silicate partition coefficient of C was previously unknown. Here we use two-phase first-principles molecular dynamics to derive this partition coefficient of C between liquid iron and silicate melt. We calculate a value of 9 ± 3 at 3,200 K and 40 GPa. Using this partition coefficient and the most recent estimates of bulk Earth or mantle C contents, we infer that the Earth’s core contains 0.1–0.7 wt% of C. Carbon thus plays a moderate role in the density deficit of the core and in the distribution of siderophile and chalcophile elements during core–mantle segregation processes. The partition coefficients of nitrogen (N), hydrogen, helium, phosphorus, magnesium, oxygen, and silicon are also inferred and found to be in close agreement with experiments and other geochemical constraints. Contents of these elements in the core derived from applying these partition coefficients match those derived by using the cosmochemical volatility curve and geochemical mass balance arguments. N is an exception, indicating its retention in a mantle phase instead of in the core.
19. Ni partitioning between metal and silicate melts: An exploratory ab initio molecular dynamics simulation study
Element partitioning is a key geochemical process. While partition coefficients between phases including melts have been measured in many experimental studies, new insight into the mechanisms of partitioning may be obtained by relating partitioning to melt structure. Here, we address this problem by exploring an ab initio molecular dynamics simulation approach. Combined with the thermodynamic integration method, these simulations provide a unique way to predict simultaneously thermodynamic properties related to element partitioning and information about the molecular structure of the melt. Thus, it should be possible not only to predict the partitioning of elements, but also to provide an explanation for this behavior based on atomic structures of the coexisting phases. Using this approach, we derive from first-principles the Ni partition coefficient between a metal and a silicate melt at 2500 K and ambient pressure, which is at least in qualitative agreement with experiment. Structural analysis of various (Mg,Fe,Ni)2SiO4 silicate and (Fe,Ni) metal melts reveals that the Ni partitioning is mainly determined by its structural environment in the silicate melt, whereas the coordination environments of Ni and Fe are almost indistinguishable in the metal melt. Possible strategies to improve the predictive power of the proposed approach are discussed.
20. Trace element partitioning between silicate melts – A molecular dynamics approach
Knowledge of trace element partition coefficients is crucial for our understanding of global element cycles. While a great number of experimental studies on mineral-melt partitioning have been performed in the past, the influence of melt structure on partitioning has mostly been considered empirically. This is mainly due to the lack of reliable structure models for typical melts at the relevant pressure and temperature conditions. Molecular dynamics simulations on the other hand may open a new window into this problem as they provide a unique approach to both structural and thermodynamic properties of minerals and melts. In this contribution, we employ first-principles and classical molecular dynamics simulations to (1) explore further a new approach to predict trace element partitioning between several silicate melts and (2) simultaneously investigate the structural controls of the observed partitioning. Specifically, we use a thermodynamic integration scheme to investigate the partitioning behavior of various trace elements (Y, La, As) in a granitic and gabbroic as well as two Ti-bearing melts and compare our data to experimental findings. Our results indicate that, similar to the lattice strain model, partitioning in melts as well seems to depend on an ideal coordination environment for each trace element and on how well this environment can be accommodated in a specific melt.
21. Molecular dynamics simulations of Y in silicate melts and implications for trace element partitioning
Element partitioning depends strongly on composition and structure of the involved phases. In this study, we use molecular dynamics simulations to investigate the local environment of Y as an exemplary trace element in four silicate melts with different compositions and thus varying degrees of polymerization. Based on these structural results, we propose a mechanism which explains the observed partitioning trends of Y and other rare-earth elements between crystals and melts or between two melts. With our computational approach, we found a systematic correlation between melt composition and Y coordination as well as Y―O bond lengths, a result which was corroborated by EXAFS spectroscopy on glasses with the same compositions as the simulated melts. Our simulations revealed, moreover, the affinity of Y for network modifiers as second-nearest neighbors (Ca in this study) and the tendency to avoid network formers (Si and Al). This is consistent with the observation that Y (and other rare-earth elements) in general prefer depolymerized to polymerized melts in partitioning experiments (see, e.g., Schmidt et al. (2006)). Furthermore, we used the method of thermodynamic integration to calculate the Gibbs free energy which governs Y partitioning between two exemplary melts. These more quantitative results, too, are in line with the observed partitioning trends.
22. Thermodynamic criteria of the end-of-life silicon wafers refining for closing the recycling loop of photovoltaic panels
The collected end-of-life (EoL) silicon wafers from the discharged photovoltaic (PV) panels are easily contaminated by impurities such as doping elements and attached materials. In this study, the thermodynamic criteria for EoL silicon wafers refining using three most typical metallurgical refining processes: oxidation refining, evaporation refining, and solvent refining were systemically and quantitatively evaluated. A total of 42 elements (Ag, Al, Au, B, Be, Bi, C, Ca, Ce, Co, Cr, Cu, Fe, Ga, Gd, Ge, Hf, In, La, Mg, Mn, Mo, Na, Nb, Ni, Os, P, Pb, Pd, Pt, Re, Ru, Sb, Sn, Ta, Ti, U, V, W, Y, Zn, Zr) that are likely to be contained in the collected EoL silicon-based PV panels were considered. The principal findings are that the removal of aluminum, beryllium, boron, calcium, gadolinium, hafnium, uranium, yttrium, and zirconium into the slag, and removal of antimony, bismuth, carbon, lead, magnesium, phosphorus, silver, sodium, and zinc into vapor phase is possible. Further, solvent refining process using aluminum, copper, and zinc as the solvent metals, among the considered 14 potential ones, was found to be efficient for the EoL silicon wafers refining. Particularly, purification of the phosphorus doped n-type PV panels using solvent metal zinc and purification of the boron doped p-type PV panels using solvent metal aluminum are preferable. The efficiency of metallurgical processes for separating most of the impurity elements was demonstrated, and to promote the recycling efficiency, a comprehensive management and recycling system considering the metallurgical criteria of EoL silicon wafers refining is critical.
23. Origins of structural and electronic transitions in disordered silicon
Structurally disordered materials pose fundamental questions1,2,3,4, including how different disordered phases (‘polyamorphs’) can coexist and transform from one phase to another5,6,7,8,9. Amorphous silicon has been extensively studied; it forms a fourfold-coordinated, covalent network at ambient conditions and much-higher-coordinated, metallic phases under pressure10,11,12. However, a detailed mechanistic understanding of the structural transitions in disordered silicon has been lacking, owing to the intrinsic limitations of even the most advanced experimental and computational techniques, for example, in terms of the system sizes accessible via simulation. Here we show how atomistic machine learning models trained on accurate quantum mechanical computations can help to describe liquid–amorphous and amorphous–amorphous transitions for a system of 100,000 atoms (ten-nanometre length scale), predicting structure, stability and electronic properties. Our simulations reveal a three-step transformation sequence for amorphous silicon under increasing external pressure. First, polyamorphic low- and high-density amorphous regions are found to coexist, rather than appearing sequentially. Then, we observe a structural collapse into a distinct very-high-density amorphous (VHDA) phase. Finally, our simulations indicate the transient nature of this VHDA phase: it rapidly nucleates crystallites, ultimately leading to the formation of a polycrystalline structure, consistent with experiments13,14,15 but not seen in earlier simulations11,16,17,18. A machine learning model for the electronic density of states confirms the onset of metallicity during VHDA formation and the subsequent crystallization. These results shed light on the liquid and amorphous states of silicon, and, in a wider context, they exemplify a machine learning-driven approach to predictive materials modelling.
24. Thermodynamics of Elements in Dilute Silicon Melts
Because high-purity silicon is one of the core materials for use in cleaner energy industry, silicon purification techniques have taken on greater importance. An understanding of the thermodynamics of impurity elements in silicon is therefore of great scientific and industrial importance. Experimental and assessment works on the thermodynamics of 23 impurity elements (Ag, Al, Au, B, Ca, Co, Cr, Cu, Fe, Mg, Mn, Mo, Nb, Ni, P, Pb, Sb, Ta, Ti, V, W, Zn, and Zr) in silicon melts are reviewed herein, and their activity coefficients in dilute silicon melts are discussed. The parameters suggested for use in assessing the liquid phase in each of the silicon binary systems are selected based on this discussion. The segregation coefficients of impurity elements calculated using the activity coefficients agree well with reported values and are used to evaluate silicon purification by directional solidification. The purpose of this paper is to provide fundamental and systemic thermodynamics knowledge for the development of silicon purification processes.