从粉末绕射解析未知物之晶体结构AbinitioCrystalStructure.pdf

從粉末繞射解析未知物之晶體結構 Ａb initio Crystal Structure Determination from Powder Diffraction Data 許火順 國家同步輻射研究中心 摘要 從粉末繞射解析未知物的晶體結構仍是一件深具挑戰性的工作。以往的觀念認為無法或 很難從粉末繞射解出未知物結構，確實在 1987 年以前，只有屈指可數的幾個成功例子。其後 由於重要材料的開發對於結構的需求大增，同步輻射的蓬勃發展，實驗數據品質提昇，解晶 方法學的開發與眾多研究人員的投入，累積至今已有近千個成功的實例。粉末解晶的方法主 要分兩大方向：倒空間法 (Reciprocal space method) 與實空間法 (Direct space method)。前者 應用類似單晶解析的流程，後者由實空間原子排列為出發點，配合繁雜的計算與許多已知的 化學資訊，困難耗時但已成為粉末解晶的主流。本文將回顧粉末解晶的發展歷程，詳述粉末 解晶的程序與介紹數個實空間法的基本原理。 關鍵詞：粉末繞射，晶體結構，倒空間法，實空間法。 Abstract The ab initio crystal structure determination from powder diffraction data (SDPD) is remaining a challenge problem. Before 1987 SDPD is very difficult and only few examples successful. In addition to high demanding from material researchers, methodology and computer development make this work become very active. Now a day near 1000 powder samples prior unknown structure were solved by reciprocal space method and direct space method. The reciprocal method is more or less following single crystal structure determination processes, while the direct space method using different strategy. Direct space method constructs molecule structure into crystal unit cell and calculates its powder pattern. The main themes of direct space methods are using different computational methods to get calculate patterns and compare with experimental data and minimized its difference. These methods need extensive calculation and prior chemical information to arrange “molecular structure”. Many mathematic algorithms such as Monte Carlo, simulated annealing, genetic algorithm, topological method, and energy minimization work quite successfully. In this article I will review the processes of SDPD and the principle of direct space