CALYPSO Structure Prediction Method And Applications

Crystal structure is a unique arrangement of atoms or molecules in a solid. Thestructure occupies a central and critical role in materials science, particularly whenestablishing a correspondence between material performance and its basic com-position since properties of a solid are intimately tied to its crystal structure.Experimentally, structural determination through X-ray diffraction technique has beendeveloped extremely well, leading to numerous crystal structures solved. However, ithappens frequently that experiments fail to determine structures due to the obtainedlow-quality x-ray diffraction data, particularly at extreme conditions (e.g., highpressure). Thus, the theoretical prediction of crystal structures with the only knowninformation of chemical composition independent of previous experimentalknowledge is greatly necessary. According to the principle of minimum energy, thestructure prediction method is to explore the potential energy surfaces and finds theglobal energy minimum. Specifically, the essence of structure prediction is todetermine the atomic arrangement of the global energy minimum with knowninformation of chemical composition conditions. However, this is extremely difficultas it basically involves in classifying a huge number of energy minima on the latticeenergy surfaces. Thus, John Maddox even published an article in Nature to questionthe predictive power provided with only the knowledge of chemical compositions.Recently, owing to significant progress in both computational power and basicmaterials theory, it is now possible to predict the crystal structure at0K using thequantum mechanical method. In our thesis, we have proposed a methodology forcrystal and two-dimentsional layer structures prediction based on the particle-swarm optimization (PSO) technique and developed a software package CALYPSO (Crystalstructure AnaLYsis by Particle Swarm Optimization) under the guidance of thesupervisor. We applied this method to the prediction of various systems under highpressure and two-dimenstional layer structure. The major works of our thesis are asfollows:1. A methodology (CALYPSO) for structure prediction based on the particle-swarmoptimization is proposed to predict the stable and meta-stable structures with theonly known information of chemical composition and external conditions (e.g.,pressure). The CALYPSO method is based on several major techniques (e.g.particle-swarm optimization algorithm, symmetry constraints on structuralgeneration, bond characterization matrix on elimination of similar structures,partial random structures per generation on enhancing structural diversity, penaltyfunction, and global and local PSO, etc.) for global structure minimization fromscratch. All of these techniques have been demonstrated to be critical to theprediction of global stable structure. We have implemented these techniques intoCALYPSO code. The CALYPSO has been successfully applied to the predictionof many known systems (e.g., elemental, binary, and ternary compounds) withvarious chemical-bonding environments (e.g., metallic, ionic, and covalentbonding). The high success rate demonstrates the reliability of this methodologyand illustrates the promise of CALYPSO as a major technique on crystal structuraldetermination.2. Water ice belongs to an important group of archetypal binary compounds with alarge abundance in the interiors of giant planets. Thus, the ultra-high pressurestudies on the structure properties of ice are of crucial importance for theunderstanding of the physics and chemistry of planetary interiors. Ice has a veryrich phase diagram. Till now, at least15solid phases have been identifiedexperimentally. According to chemical bonding, ice can be classified into threecategories: The first category is the water molecules, which connected byhydrogen bonds. The second category is an atomic crystal having the symmetrical hydrogen bonds with each O atom covalently bonded to four H atoms. The thirdcategory is super-ionic state of ice, which exists under high temperature andpressure conditions. The search for ice forms upon ultra-compression hastherefore attracted great attention. However, the conditions as occurring in thecores of giant planets are not directly experimentally accessible, as an alternativeab initio simulations are able to explore states of matters under such extremeconditions. We newly predicted two energetically favourable structures byCALYPSO method: a tetragonal I-42d and a monoclinic P21. The I-42d structurecontains two interpenetrating hydrogen bonded networks. Intriguingly, thestructural feature of O–H tetrahedrons is no longer preserved in P21structure.Instead, we find the surprising occurrence of alternate layers of OH and H3O units,implying the ionic nature of this solid.3. Human progress and development has always been marked by breakthroughs incontrol of materials, and every technical progress largely depends on thedevelopment of materials science. One age replaces another when the material isreplaced by another material, which presents more qualities. Pursuit of moderntechnology is a multi-tasking, speed, versatility and flexibility. Two-dimensionallayer materials meet these requirements to be the potential candidates because ofnovel properties, such as high thermal conductivity and high electron mobility.Furthermore, two-dimensional (2D) materials are of special importance becausethey are usually parent structures of one-dimensional nanotubes andzero-dimensional nanocages. Thus, the structural prediction of2D has become ahot issue.Most of the traditional experimental synthesis method of two-dimensionalstructure has no predictability for the product and results in a great waste ofresources. Therefore, we expand the CALYPSO methods to design the structuresof2D material for meeting requirements of industrial applications, and then guidethe experimental synthesis. Our method is successful in predicting the structuresof known2D materials, including single layer and multi-layer graphene, and some quasi-2D materials. The CALYPSO has been successfully applied to theprediction of many known two-dimensional systems. The high success ratedemonstrates the reliability of this methodology for two-dimensional structureprediction methods.The CALYPSO method has been applied to two-dimensional materials bysome researchers. Xiang et al. designed new types of BC and NC layeredcompound [see JACS133,16285(2011) and PRX2,011003(2012)] and Wu etal. predicted the B and B-Si two-dimensional layered structures [see ACS Nano6,7443(2012) and JPCL4,561,(2013)] using CALYPSO method. Furthermore,by use of this method, we predict a new family of mono-layered boron nitridestructures with different chemical compositions. The first-principles electronicstructure calculations reveal that the band gap of these N-rich BN systems can betuned from5.4eV to2.2eV by adjusting the composition，showing strongpotential for applications to electronics and optoelectronics.