Knowledge Management System Of Institute of process engineering,CAS
热加工技术有潜力将木质素转化为化学品、材料和燃料等高值化产品，热解是其中最基础的过程，深入认识木质素热解反应机理对发展木质素高效利用工艺非常重要。鉴于木质素结构和热解反应的复杂性，已有的实验方法和量子化学计算方法还难以获得木质素热解过程的全貌和动态演化细节，本论文利用大规模ReaxFF MD模拟新方法，采用基于GPU的高性能计算程序GMD-Reax直接模拟了11800−77200原子规模的木质素模型热解过程，并利用化学反应与可视化分析工具VARxMD深入挖掘了热解过程全貌及其背后反应机理，系统地研究了软木、硬木和硫酸盐木质素热解过程中的连接键、中间体、官能团及主要产物的演化行为，揭示了不同种类木质素热解行为的差异与共性规律。本论文取得的主要结果如下：（1）基于木质素不同的植物来源、不同的分离方法、不同的单体聚合观点构建了软木、硬木、硫酸盐三类木质素的7种分子模型：Adler和Freudenberg软木木质素支化高分子模型、Ralph-SL软木木质素线型高分子模型、Nimz硬木木质素支化高分子模型、Ralph-HL硬木木质素线型高分子模型、以及Marton和Crestini硫酸盐木质素支化高分子模型。这些3D模型规模均大于10,000原子，结构覆盖了7种含氧官能团和20种连接键。这些大规模木质素模型可以更好地描述木质素高分子的支化或线型结构的多样性，有助于考察木质素结构在热解反应过程中的相互影响。（2）提出了采用木质素单体间连接键和其所连接的两个苯环的组合结构作为识别特定连接键反应子结构的策略，获得了木质素中常见11种连接键整体结构的初始主导转化路径：Cα/β−O醚键断裂反应在α-O-4、β-O-4、α-O-4 & β-5、α-O-4 & β-O-4 & 5-5和β-1 & α-O-α连接键整体结构的初始转化路径中占主导地位（>85%）；Cα−O醚键断裂反应和苯环开环反应在β-β & γ-O-α连接键整体结构的初始转化路径中是两类势均力敌的主导反应路径（各约50%）；苯环开环反应是4-O-5、5-5和β-N连接键整体结构的初始主导转化路径（>95%）。此外，相邻连接键整体结构之间的反应活性会相互影响：α-O-4和β-O-4等反应活性较高的连接键发生醚键断裂后生成的苯氧基会促进其相邻的、反应活性较低的4-O-5等连接键整体结构发生键断裂反应或苯环开环反应。（3）模拟获得的苯环可以发生开环反应的结论与前人的量子力学计算和冲击波实验得到的结果一致，本论文进一步揭示了木质素热解过程中导致稳定的芳环发生开环反应的重要中间体及主要路径：①苯环在苯氧基的作用下被插入一个O原子后可转化为含氧7元杂环；②苯环在二环[3.1.0]己基结构的作用下发生3元环上的碳碳键断裂反应后可转化为5元脂环结构；③苯环在二环[3.1.0]己基结构的作用下发生5元环上的碳碳键断裂反应后可转化为3元脂环结构；④苯环在苄基结构的作用下被插入一个C原子后可转化为7元脂环结构。（4）获得了软木、硬木和硫酸盐木质素模型热解行为的差异与共性规律。差异体现为：①热重曲线的演化快慢与C−O醚键的含量密切相关，C−O醚键含量高的硬木木质素热反应性最高，其次为软木木质素，硫酸盐木质素的热反应性最低。②提出了依据C原子数区分的产物类型及其演化特征划分木质素热解阶段的策略，发现软木和硬木木质素热解过程具有显著的三阶段特征，而硫酸盐木质素热解过程仅呈现出后两个阶段。③不同种类木质素模型中β-O-4连接键结构上所含有的羰基、甲氧基及邻接α-O-4连接键等活性取代基的不同结构环境会影响其转化速率：活性取代基含量最高的Nimz硬木木质素转化速率最快；Adler软木和Freudenberg软木木质素次之；含量最低的Marton硫酸盐木质素转化速率最慢。类似的活性取代基效应也影响β-β & γ-O-α连接键整体结构的转化反应：含羰基（Cα=O）的Marton硫酸盐木质素及含醚键（Cα-O-4和−OCH3）的Nimz硬木木质素中其转化更早发生。当木质素中含有较多非活性Ph−(C)n−Ph键和较少活性PhO−C键时，会使碳碳型连接键整体结构的转化变慢。不同种类木质素热解的共性体现为：木质素单体的芳环结构在不同种类木质素热解反应过程中均可以在苯氧基、苯上含氧官能团或苄基的作用下转化生成含氧七元杂环、五元脂环或七元脂环等环结构中间体且各环结构的演化趋势相同。（5）考察了同一类木质素中仍然存在争议的支化和线型分子结构对木质素热解行为可能产生的影响，结果表明：不同木质素模型热解模拟获得的6种主要小分子产物的开始生成温度和最大生成量与木质素中相关官能团或连接键结构的含量密切相关，与其支化或线型高分子结构的关联性不大。（6）采用Solid-Py/SR-VUV-PI-TOF-MS实验装置进行了木质素热解实验研究，实验结果与模拟结果对比表明：实验检测到的16种C0−C3 产物在模拟中也能获得；不同木质素模型热解模拟获得的苯、苯酚及其衍生物的种类少于实验结果。更大模型的模拟可以更好地描述苯、苯酚及其衍生物等热解产物的演化行为，大规模硫酸盐木质素模型的热解模拟获得的乙烯、甲醛、甲醇、硫化氢、甲硫醇、苯、苯酚、邻甲基苯酚、邻苯二酚和愈创木酚等产物随温度的演化趋势和实验定性一致，且热解产物在不同温度下随时间的演化规律均呈现出生成速率随温度升高而增大、温度越高其产量越快达到最大值的趋势。尽管木质素的Solid-Py/SR-VUV-PI-TOF-MS实验和ReaxFF MD模拟的时间尺度和温度区间存在较大差距，但实验和模拟获得的热解产物随温度的演化趋势以及在不同温度下随时间的演化趋势定性一致。本论文的工作表明：大规模ReaxFF MD模拟方法不仅可以定性预测木质素热解产物随温度和时间的演化概貌，并且可通过所获得的反应路径对预测结果加以解释，从而可为木质素热解的高值化应用提供深入的分子反应机理的认识，为发展木质素高效利用的工艺提供理论支持。;Pyrolysis-based technologies are promising methods to convert lignin into biochemicals, biomaterials, and biofuels. Deep understanding of the molecular mechanisms involved in lignin pyrolysis is important for the development of the efficient lignin utilization technologies. Due to the complexity of lignin structure and pyrolysis process, it is still difficult to obtain the dynamic evolution scenario of the complex lignin pyrolysis reactions with the available experimental methods and QM calculation methods. By taking advantage of the new methodology of large scale ReaxFF MD simulations, this work investigates the complicated lignin pyrolysis process by simulations of large-scale lignin models containing 11800−77200 atoms. The simulations were performed using the GPU-enabled high performance code of GMD-Reax. The underlying reaction mechanisms of lignin pyrolysis were discovered with aid of VARxMD, a unique tool for visualization and reaction analysis for ReaxFF MD simulations. The evolution of linkages, intermediates, functional groups, and main products in the pyrolysis process of softwood, hardwood and Kraft lignin was systematically studied. The differences and similarities in the pyrolysis of different types of lignin were unraveled. The main results obtained in this work are as follows:(1) 7 molecular models of softwood, hardwood, and Kraft lignin were constructed with a coverage of different plant sources, different separation methods, and different viewpoints on monomer polymerization of lignin. The models include branched polymer models of Adler and Freudenberg for softwood lignin, linear polymer model of Ralph-SL for softwood lignin, branched polymer model of Nimz for hardwood lignin, linear polymer model of Ralph-HL for hardwood lignin, and branched polymer models of Marton and Crestini for Kraft lignin. These 3D models span 7 types of oxygen-containing functional groups and 20 types of linkages. All the models contain more than 10,000 atoms. The large-scale lignin models can provide a better description of the structure diversity of branched or linear lignin polymers, thus facilitating the investigation of the reactivity interactions between linkages or monomer units in lignin pyrolysis.(2) A reaction analysis strategy was proposed to identify the specific linkage involved reactions by using a combined query structure of one linkage together with its two connected monomer rings. The dominant reaction pathways in the initial conversion of the 11 linkages and their linked monomers were obtained. Cα/Cβ ether bond cracking is the dominant pathway for the consumption of α-O-4, β-O-4, α-O-4 & β-5, α-O-4 & β-O-4 & 5-5, and β-1 & α-O-α linkages and their linked monomer aryl rings. Both the Cα−O ether bond cracking and the monomer aryl ring opening are equally important for the consumption of the β-β & γ-O-α linkage and its linked monomer aryl rings. The aryl ring opening reactions are the dominant pathways for the consumption of other 4-O-5, 5-5, β-1, β-2, and β-5 linkages and their linked monomer aryl rings. In addition, the reactivity interactions of neighboring linkages and their monomer aryl rings were observed. The reactivity of a less reactive linkage (such as 4-O-5 linkage) and its monomer aryl rings can be activated by the phenoxy radicals generated by the ether bond breaking of its neighboring active linkages of α-O-4 or β-O-4.(3) It is found from ReaxFF MD simulations that the benzene ring can undergo ring-opening reactions, which is consistent with what suggested by previous QM calculations and shock wave experiments. This work further reveals for the first time the four pathways and three important intermediates that activate and convert the thermally stable aromatic rings into other ring structures in the pyrolysis of lignin. The aromatic rings can be converted into: (i) an oxygen-containing seven-membered heterocyclic ring induced by phenoxy radicals, (ii) a five-membered aliphatic ring activated by the bridged 3- and 5-membered aliphatic rings, (iii) a 3-membered aliphatic ring activated by the bridged 3- and 5-membered aliphatic rings, and into (iv) a 7-membered alicyclic ring with the activation of the bridged 3- and 6-membered aliphatic rings .(4) The differences and similarities in the pyrolysis behavior of softwood, hardwood and Kraft lignin models were further unraveled. The differences are: (i) The profile of the thermogravimetric curve is closely related to the content of C−O ether bonds. With the highest content of C−O ether bond, hardwood lignin has the highest thermal reactivity, lower of the softwood lignin, and the lowest thermal reactivity of Kraft lignin correspondingly. (ii) Using the proposed strategy to identify lignin pyrolysis stages based on the distinguished five lumped categories of lignin pyrolysates and their evolution characteristics, the pyrolysis process of softwood and hardwood lignin was found showing significant three-stage characteristics, while Kraft lignin pyrolysis exhibits only the latter two stages. (iii) Different structural environments of active substituents of carbonyl, methoxy, and adjacent α-O-4 linkages on β-O-4 linkage structures will affect its conversion rate in different types of lignin models. With the highest portions of active substituents, the Nimz hardwood lignin model has the fastest conversion rate of the β-O-4 linkages, faster for the Adler and Freudenberg softwood lignin models, and the slowest for Marton Kraft lignin model. Similar substituent effects also exist in the conversion of the β-β & γ-O-α linkages. The β-β & γ-O-α linkages containing carbonyl groups (Cα=O) in the Marton Kraft lignin model and the active ether bond (Cα-O-4 and −OCH3) in the Nimz hardwood lignin model begin to react earlier. In addition, it was found that lignin models containing more inactive Ph−(C)n−Ph bonds and less active PhO−C bonds have the slower conversion rate of carbon-carbon linkages and their linked monomers.The pyrolysis similarity of softwood, hardwood, and Kraft lignin lies in the aromatic ring structures of lignin monomers that can be converted into six-membered aliphatic ring, oxygen-containing seven-membered heterocyclic ring, five-membered aliphatic ring, or seven-membered aliphatic rings by the activation of phenoxy functional groups, oxygen-containing functional groups on aryl rings, or benzyl groups in the pyrolysis of different types of lignin. Moreover, similar evolution trends of the ring structure intermediates were observed in the pyrolysis of different types of lignin.(5) This work investigated the possible influence of the branched or linear polymer structures of the same type of lignin on their lignin pyrolysis behavior. The results showed that the temperature for the six main small molecule products begining to generate and their maximum amount produced are closely associated to the content of the functional groups or linkage structures involved in their formation. The branched or linear polymer structures does not seem to be significant in generation of these small molecule products.(6) The Solid-Py/SR-VUV-PI-TOF-MS experiments of lignin samples were also carried out in this work. The comparison between the experimental results and the simulation results showed that 16 kinds of C0−C3 products detected in the experiments can be obtained in the simulations. Less pyrolyzates of benzene, phenol and their derivatives were obtained from the pyrolysis simulations of different lignin models than from the experimental results. The ReaxFF MD simulations of larger-scale Kraft lignin model can better describe the evolution behavior of benzene, phenol, and its derivatives pyrolysis products. The consistent evolution trends with temperature were observed between the experiments and simulations for the C0−C3 pyrolyzates of ethylene, formaldehyde, methanol, hydrogen sulfide and methyl mercaptan, as well as the aromatic pyrolyzates of benzene, phenol, o-methylphenol, catechol, and guaiacol. The consistent evolution trends over time at different temperatures were also obtained between the experiments and ReaxFF MD simulations for CH2O, CH3OH, benzene, phenol, and guaiacol. Their generation rates increase with the increasing temperatures. The higher the temperature, the faster approaching to their maximum yields.Although the time scale and temperature range between the Solid-Py/SR-VUV-PI-TOF-MS experiments and the ReaxFF MD simulations of lignin pyrolysis are quite different, the consistent evolution trends of the pyrolysis products over temperature and reaction time were obtained between the experiments and simulations. This work demonstrates that the large scale ReaxFF MD simulation method is a promising method not only for qualitatively predicting the evolution trend of lignin pyrolyzates with temperature and time, but also for explaining what predicted with the unraveled underlying mechanisms. It can provide deep insight into the molecular reaction mechanisms and theoretical supports for the technology development in the effective valorisation of lignin.
|张婷婷. 不同种类木质素热解的反应分子动力学模拟[D]. 中国科学院大学,2020.|
|Files in This Item:|
|不同种类木质素热解的反应分子动力学模拟.（8048KB）||学位论文||限制开放||CC BY-NC-SA||Application Full Text|
|Recommend this item|
|Export to Endnote|
|Similar articles in Google Scholar|
|Similar articles in Baidu academic|
|Similar articles in Bing Scholar|
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.