New Methods for Predicting Mechanisms for Complex Heterogeneous Catalysts with Applications to Metal Oxide Functionalization of Alkanes

获得化学部化学催化项目的这一奖项后，来自加州理工学院化学系和化学工程系的William A. Goddard教授和他的同事Robert Nielsen将通过蒙特卡罗（MC）和分子动力学（MD）模拟来开发和应用反应力场（ReaxFF）。确定催化剂表面和非均相界面的原位原子尺度结构，配合量子力学（QM）研究反应性。多范式ReaxFF/MC/MD框架将用于解析丙烷和丙烯氨氧化的职业无序多金属氧化物的整体-占领超级单体结构。其中最突出的是用于丙烯氨氧化制丙烯腈的MoVNbTeO催化剂。结合高温反应动力学轨迹和关键中间体和过渡态的量子力学研究，将确定参与速率和选择性决定CH活化和碳杂原子成键步骤的金属元素。另外，通过模拟焦磷酸钒基和其他高氧化态V/P/O相的大电池模型的退火和煅烧，确定焦磷酸钒基（VO）2P2O7)在其表面储存氧原子用于丁烷14电子氧化制马来酸酐的化学机理。通过对现有催化剂的CH活化、自由基捕获、氨活化和氧活化功能的优化组合，可以发现新的多相多金属烃类功能化催化剂。这个计算框架已经在更简单的双金属氧化物上得到了验证，并将继续优化。开发工作将集中于(1)简化针对QM训练集的新元素组合的力场参数拟合；(2)将ReaxFF的MC应用扩展到大规范实现。催化和可持续能源消耗所需的应用中的界面现象发生在一个太大的范围内，成熟的基于量子力学的模拟无法发挥预测作用：半导体表面的钝化和带调谐，层状热电材料的晶界，半导体-金属和半导体有机金属溶剂界面，促进光合作用或光伏器件中光敏和催化组分之间的电荷转移。开发ReaxFF/MC/MD方法对经典但反应性分子动力学模拟的更广泛的科学影响将成为解决许多异相催化和界面科学应用中的限制因素的工具：确定操作条件下的原子尺度结构。多金属氧化物催化剂用于从丙烯生产商品化学品丙烯醛和丙烯腈，即使催化性能（即选择性和活性）的微小改进也会对能源消耗和产量产生重大影响。用丙烷取代丙烯的经济影响怎么强调也不为过。由于从丙烷中生成丙烯的成本约为10美分/磅，因此取消这一步骤有可能每年为美国化学工业节省数亿美元。此外，在戈达德教授一年的课程中，每个季度都会介绍小组当前研究的进展。说明原子相互作用的具体应用将在化学物理的第四节课中讨论；新的和已建立的计算技术将在方法讲座中讲授；通过小组研究开发的方法和软件由实验组的学生在学生各自领域的实践项目工作中应用。

With this award from the Chemical Catalysis Program of the Chemistry Division, Professor William A. Goddard and colleague Robert Nielsen from the Departments of Chemistry and Chemical Engineering at the California Institute of Technology will develop and apply reactive force field (ReaxFF), through Monte Carlo (MC) and molecular dynamics (MD) simulations, to determine the in situ atomic scale structure of catalyst surfaces and heterogeneous interfaces in coordination with quantum mechanical (QM) studies of reactivity. The multiparadigm ReaxFF/MC/MD framework will be used to resolve the integral-occupation supercell structures of occupationally disordered multimetal oxides responsible for propane and propene ammoxidation. Most prominent among these are MoVNbTeO catalysts for ammoxidation of propene to acrylonitrile. Combining high temperature reactive dynamics trajectories with quantum-mechanical studies of key intermediates and transition states, the metallic elements involved in rate- and selectivity determining CH activation and carbon-heteroatom bond-forming steps will be identified. Separately, the chemical mechanism by which vanadyl pyrophosphate ((VO)2P2O7) stores oxygen atoms at its surface for use in the 14-electron oxidation of butane to maleic anhydride will be determined by simulating the annealing and calcination of large unit cell models of the vanadyl pyrophosphate and other high oxidation state V/P/O phases. New heterogeneous multimetallic catalysts for hydrocarbon functionalization will be posited through the optimal combination of the CH activation, radical trapping, ammonia activation and oxygen activation functions of existing catalysts. This computational framework has been validated on simpler bimetallic oxides, and will continue to be optimized. Developmental work will focus on (1) streamlining the fitting of force-field parameters for new combinations of elements against QM training sets and (2) extending the MC application of ReaxFF to grand canonical implementations. Interfacial phenomena in catalysis and applications necessary for sustainable energy consumption occur at a regime too large for mature quantum mechanics-based simulations to play a predictive role: passivation and band tuning at semiconductor surfaces, grain boundaries in layered thermoelectric materials, semiconductor-metal and semiconductor organometallic solvent interfaces which facilitate charge transfer between photosensitive and catalytic components in photosynthetic or photovoltaic devices. The broader scientific impact of developing the ReaxFF/MC/MD approach to classical but reactive molecular dynamics simulations will be a tool for addressing a limiting factor in many heterogeneous catalysis and interfacial science applications: determining atomic-scale structures under operating conditions. Multimetal oxide catalysts are used to produce the commodity chemicals acrolein and acrylonitrile from propene, and even small improvements in catalytic performance (i.e., selectivity and activity) will have a substantial effect on energy consumption and yield. The economic impact of replacing propene with propane cannot be overstated. Since propene is generated from propane with an approximate cost of 10 ¢/lb, elimination of this step has the potential to save the US chemical industry several hundred million dollars per year. In addition, each quarter of Prof. Goddard's year long class is informed by progress in the group's current research. Applications which illustrate specific kinds of atomic interactions are discussed in the quarter of lectures in chemical physics; new and established computational techniques are taught in the quarter of lectures on methods; methods and software developed through the group's research are applied by students from experimental groups in the quarter of hands-on project work in students' respective fields.