Artificially Intelligent, Autonomous Microreactors for the Discovery of Polyolefin Catalysis

用于发现聚烯烃催化的人工智能自主微反应器

• 批准号: 1701393
• 负责人: Ryan Hartman
• 金额: $ 29.8万
• 依托单位: New York University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2020-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1701393&HistoricalAwards=false
• 关键词: Artificially; Intelligent; Autonomous; Microreactors; Discovery

The current technology gaps of laboratory-scale reaction systems have limited the commercial availability of: 1) automated microreactor platforms that deliver first principles kinetics, 2) automated microreactor platforms for catalyst activity screening, 3) online analytics that capture real-time concentration-response profiles, and 4) computational data analytics able to analyze, report, and recommend. Process development and reaction engineers could potentially implement such reaction systems to expedite their laborious research and development activities. The research project aims at closing these technology gaps in laboratory-scale catalyst screening and characterization to broadly accelerate materials development and polymer commercialization timelines. The study of artificially intelligent, autonomous microreactors (microAIRs) is proposed to address the technical challenges that currently limit the next-generation needs in catalyst discovery research. The governing hypothesis for the study is that microAIRs engineered with online analytics can accelerate, improve accuracy, and minimize the energy and environmental impacts during the iterative discovery of a next-generation olefin catalyst system. Successful testing of this hypothesis will address the principal need to combine real-time, multiphase microfluidics tracking and feedback algorithms with a non-invasive analytical method that can directly measure a reaction parameter. The ability of a reactor system to decipher phase behaviors, analyze the reaction progress, decide which catalyst is the most active, and identify accurate kinetic expressions are tremendous challenges that microAIRs can solve in order to more efficiently screen catalyst and discover kinetics. Opportunities exist to i) improve the accuracy of process kinetics, ii) establish input/output responses that fully fingerprint a catalyst in a business portfolio, and iii) improve the presentation of real-time analytics and accelerated decision making in catalyst discovery. Review of the state-of-the-art reveals that current technical challenges for laboratory-scale, homogeneous polyolefin catalytic systems include: 1) combinatorial challenge to fully fingerprint catalyst performance, 2) evaluation of large commercial libraries of catalyst systems, 3) engineering for control of chemical transport challenges in flow, 4) sensing with adaptive response to catalyst activity, 5) on-chip analytics coupled with unique reactor/mixer designs in flow, and 6) real-time data analysis with adaptive experimental design and execution. The proposed research, if successful, will broadly impact polymers manufacturing, and it will also introduce novel laboratory techniques for the discovery of new science. The project will also involve curriculum development activities and outreach to the community through NYU's incubator program.

实验室规模反应系统的当前技术差距限制了：1）自动化微反应器平台，可提供第一原理动力学，2）用于催化剂活动筛查的自动反应器平台，3）在线分析，这些分析可捕获实时浓度响应概况的在线分析，以及4）计算数据分析以分析，报告和推荐。 流程开发和反应工程师可以潜在地实施此类反应系统，以加快其艰苦的研发活动。 该研究项目旨在缩小实验室规模的催化剂筛查和表征的这些技术差距，以广泛加速材料开发和聚合物商业化时间表。 提出了对人工智能的自主微反应器（微反应器）的研究，以解决目前限制Catalyst Discovery Research中下一代需求的技术挑战。 该研究的管理假设是，在线分析设计的微型仪可以在迭代发现下一代烯烃催化剂系统期间加速，提高准确性并最大程度地减少能量和环境影响。 该假设的成功检验将解决与可以直接测量反应参数的非侵入性分析方法相结合的实时，多相微富集算法和反馈算法的主要需求。 反应堆系统破译相行为，分析反应进展，确定哪种催化剂是最活跃的能力，并确定准确的动力学表达是微观台上可以解决的巨大挑战，以便更有效地筛选催化剂并发现动力学。存在机会i）提高过程动力学的准确性，ii）建立在业务组合中充分指纹催化剂的输入/输出响应，以及iii）改善对催化剂发现中实时分析和加速决策的呈现。 对最先进的综述表明，当前针对实验室规模的均质聚烯烃催化系统的技术挑战包括：1）对完全指纹催化剂性能的组合挑战，2）评估催化系统的大型商业库的评估，3）3）工程在流动中对化学运输挑战的工程，以控制流动式的catsip instrips instrip，4）与CATALE STRIPTION的SENSTICS ONSTION，4）5）5）反应堆/混合器在流动中的设计，6）具有自适应实验设计和执行的实时数据分析。 拟议的研究（如果成功）将广泛影响聚合物制造，并且还将引入新颖的实验室技术以发现新科学。该项目还将通过NYU的孵化器计划涉及课程开发活动和向社区推广。

项目成果

Supervised machine learning for prediction of zirconocene-catalyzed α-olefin polymerization

用于预测二茂锆催化α-烯烃聚合的监督机器学习

• DOI: 10.1016/j.ces.2019.115224
• 发表时间: 2019
• 期刊: Chemical Engineering Science
• 影响因子: 4.7
• 作者: Rizkin, Benjamin A.;Hartman, Ryan L.
• 通讯作者: Hartman, Ryan L.

Combining automated microfluidic experimentation with machine learning for efficient polymerization design

• DOI: 10.1038/s42256-020-0166-5
• 发表时间: 2020-04-01
• 期刊: NATURE MACHINE INTELLIGENCE
• 影响因子: 23.8
• 作者: Rizkin, Benjamin A.;Shkolnik, Albert S.;Hartman, Ryan L.
• 通讯作者: Hartman, Ryan L.

Artificial Neural Network control of thermoelectrically-cooled microfluidics using computer vision based on IR thermography

• DOI: 10.1016/j.compchemeng.2018.11.016
• 发表时间: 2019-02-02
• 期刊: COMPUTERS & CHEMICAL ENGINEERING
• 影响因子: 4.3
• 作者: Rizkin, Benjamin A.;Popovich, Karina;Hartman, Ryan L.
• 通讯作者: Hartman, Ryan L.

Flow chemistry remains an opportunity for chemists and chemical engineers

• DOI: 10.1016/j.coche.2020.05.002
• 发表时间: 2020-09-01
• 期刊: CURRENT OPINION IN CHEMICAL ENGINEERING
• 影响因子: 6.6
• 作者: Hartman, Ryan L.

Using word embeddings in abstracts to accelerate metallocene catalysis polymerization research

• DOI: 10.1016/j.compchemeng.2020.107026
• 发表时间: 2020-10-04
• 期刊: COMPUTERS & CHEMICAL ENGINEERING
• 影响因子: 4.3
• 作者: Ho, David;Shkolnik, Albert S.;Hartman, Ryan L.
• 通讯作者: Hartman, Ryan L.