Automation and Machine Learning for Accelerated Polymer Characterization and Development: Past, Potential, and a Path ForwardClick to copy article linkArticle link copied!
- Peter A. Beaucage*Peter A. Beaucage*Email: [email protected]NIST Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United StatesMore by Peter A. Beaucage
- Duncan R. SutherlandDuncan R. SutherlandUniversity of Colorado Boulder, Boulder, Colorado 80309, United StatesMaterial Science and Engineering Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United StatesMore by Duncan R. Sutherland
- Tyler B. Martin*Tyler B. Martin*Email: [email protected]Material Science and Engineering Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United StatesMore by Tyler B. Martin
Abstract
Automation and machine learning techniques are poised to dramatically accelerate the development of new materials while simultaneously increasing our understanding of the physics and chemistry that underlie the formation of such materials. In particular, the convergence of accessible machine learning tools, the availability of high-quality data, and the advent of accessible experimental automation platforms have led to a number of closed-loop autonomous experimentation platforms or “self-driving labs”. Such platforms integrate robotic experimenters with AI-guided experiment planning to autonomously perform large numbers of experiments without human input. After briefly reviewing the state of the field and the broad classes of autonomous efforts, this perspective outlines several high-value focus areas for future ML-guided characterization efforts. Among many advantages, we expect that autonomous approaches will allow the systematic study of rare and nonequilibrium phenomena, provide dramatically greater measurement efficiency through targeting of cutting-edge, resource-intensive characterization, and enable a higher level of thinking and experimental planning for human investigators. Finally, we outline the principal barriers to realization of these advantages, including: (1) a lack of organizational structures and workforce development for the highly interdisciplinary programs needed; (2) funding and publication mechanisms that assign greater value to individual scientific results than foundational infrastructure development; and (3) a dearth of standards for open interchange of hardware, software, and data among the polymer community. We believe that we are in the early days of a once-in-a-generation shift in the way science is planned, executed, and evaluated, and we hope to provide a blueprint for the broader polymer community to take a leading role in this shift.
This publication is licensed for personal use by The American Chemical Society.
Introduction
Past (and Present)
Potential
Automation Enables the Routine, Rigorous Study of Statistics
Automation and Machine Learning Allow Targeting of Rare Measurements
Autonomous and Human-Machine Teaming Is a Force Multiplier
Working late into the night over many weeks, the scientist runs the 10th (or 100th or 1000th) iteration of a measurement/synthesis/simulation. Quick looks at the intermediate data during collection seem fine, but when they finally sit down to look at the full corpus of results, they discover that the instrument was miscalibrated, or the starting materials were contaminated, or their input files contained an error.
Path Forward
Training and Workforce Development Needs for the Autonomous Future
Open Standards for Hardware, Software, and Data Interchange
Physics-Informed Machine Learning Modeling
20th Century Funding for 21st Century Infrastructure
Open Hardware and Software Communities for Autonomous Polymer Science
Conclusion
Biographies
Acknowledgments
We thank Dr. Chelsea Edwards of the NIST Center for Neutron Research and Dr. Christopher Stafford of the NIST Materials Science and Engineering Division for helpful comments and discussions. Certain commercial equipment, instruments, materials, organizations, or software are identified in this paper to foster understanding. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose.
References
This article references 53 other publications.
- 1Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles; J3016_202104; SAE International, 2021.Google ScholarThere is no corresponding record for this reference.
- 2Meredith, C.; Smith, A. P.; Crosby, A. J.; Amis, E. J.; Karim, A. Combinatorial Methods for Polymer Science. In Encyclopedia of Polymer Science and Technology; Wiley, DOI: 10.1002/0471440264.pst069 .Google ScholarThere is no corresponding record for this reference.
- 3Hoogenboom, R.; Meier, M. A. R.; Schubert, U. S. Combinatorial Methods, Automated Synthesis and High-Throughput Screening in Polymer Research: Past and Present. Macromol. Rapid Commun. 2003, 24 (1), 15– 32, DOI: 10.1002/marc.200390013Google Scholar3https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXhs1Siu7o%253D&md5=0253f289489c86ddef3ed65b3e75f3edCombinatorial methods, automated synthesis and high-throughput screening in polymer research: Past and presentHoogenboom, Richard; Meier, Michael A. R.; Schubert, Ulrich S.Macromolecular Rapid Communications (2003), 24 (1), 15-32CODEN: MRCOE3; ISSN:1022-1336. (Wiley-VCH Verlag GmbH & Co. KGaA)A review describes polymer research related to high-throughput screening development. Combinatorial techniques, parallel experimentation and high-throughput methods represent a very promising approach in order to speed up the prepn. and investigation of new polymeric materials: a large variety of parameters can be screened simultaneously resulting in new structure/property relationships. The field of polymer research seems to be perfectly suited for parallel and combinatorial methods due to the fact that many parameters can be varied during synthesis, processing, blending as well as compounding. In addn., numerous important parameters have to be investigated, such as mol. wt., polydispersity, viscosity, hardness, stiffness and other application-specific properties. A no. of corresponding high-throughput techniques have been developed in the last few years and their introduction into the com. market further boosted the development. These combinatorial approaches can reduce the time-to-market for new polymeric materials drastically compared to traditional approaches and allow a much more detailed understanding of polymers from the macroscopic to the nanoscopic scale. Here we provide an overview of the present status of combinatorial and parallel polymer synthesis and high-throughput screening.
- 4Zaquen, N.; Rubens, M.; Corrigan, N.; Xu, J.; Zetterlund, P. B.; Boyer, C.; Junkers, T. Polymer Synthesis in Continuous Flow Reactors. Prog. Polym. Sci. 2020, 107, 101256, DOI: 10.1016/j.progpolymsci.2020.101256Google Scholar4https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXht1Wmur3I&md5=d157f9944158582c064dde6fb3eed7cePolymer Synthesis in Continuous Flow ReactorsZaquen, Neomy; Rubens, Maarten; Corrigan, Nathaniel; Xu, Jiangtao; Zetterlund, Per B.; Boyer, Cyrille; Junkers, TanjaProgress in Polymer Science (2020), 107 (), 101256CODEN: PRPSB8; ISSN:0079-6700. (Elsevier Ltd.)A variety of polymns. has long been performed in continuous flow reactors on an industrial scale; comparatively, on smaller scales, continuous polymn. methods have only gained significant attention in recent years. Yet, within the last decade, the field has moved from the rare occurrence of flow reactors to their abundant use today. A wide variety of polymer reactions have been performed in a continuous fashion on small and intermediate scales. The advantages of applying flow chem. principles for polymer reactions include increased reproducibility and synthetic precision, significant increases in reaction performances for photochem. reactions, the ability to couple reactors to create complex materials in a single reactor pass, as well as the unique combination of online monitoring and machine learning. In this review we give a comprehensive overview of polymer reactions being carried out in continuous flow reactors to date. The development of the field is discussed, concluding with the most recent examples on automated polymer synthesis, reactor telescoping and nanoparticle synthesis. Finally, the design of flow reactors is discussed to help newcomers contribute to the current and future developments in the field.
- 5Tamasi, M.; Kosuri, S.; DiStefano, J.; Chapman, R.; Gormley, A. J. Automation of Controlled/Living Radical Polymerization. Advanced Intelligent Systems 2020, 2 (2), 1900126, DOI: 10.1002/aisy.201900126Google ScholarThere is no corresponding record for this reference.
- 6Winicov, H.; Schainbaum, J.; Buckley, J.; Longino, G.; Hill, J.; Berkoff, C. E. Chemical process optimization by computer ─ a self-directed chemical synthesis system. Anal. Chim. Acta 1978, 103 (4), 469– 476, DOI: 10.1016/S0003-2670(01)83110-XGoogle ScholarThere is no corresponding record for this reference.
- 7MATSUDA, R.; ISHIBASHI, M.; TAKEDA, Y. Simplex Optimization of Reaction Conditions with an Automated System. CHEMICAL & PHARMACEUTICAL BULLETIN 1988, 36 (9), 3512– 3518, DOI: 10.1248/cpb.36.3512Google ScholarThere is no corresponding record for this reference.
- 8King, R. D.; Whelan, K. E.; Jones, F. M.; Reiser, P. G. K.; Bryant, C. H.; Muggleton, S. H.; Kell, D. B.; Oliver, S. G. Functional genomic hypothesis generation and experimentation by a robot scientist. Nature 2004, 427 (6971), 247– 252, DOI: 10.1038/nature02236Google Scholar8https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXjtFKjuw%253D%253D&md5=58c844f52c70fa8f9030707129cd6472Functional genomic hypothesis generation and experimentation by a robot scientistKing, Ross D.; Whelan, Kenneth E.; Jones, Ffion M.; Reiser, Philip G. K.; Bryant, Christopher H.; Muggleton, Stephen H.; Kell, Douglas B.; Oliver, Stephen G.Nature (London, United Kingdom) (2004), 427 (6971), 247-252CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)The question of whether it is possible to automate the scientific process is of both great theor. interest and increasing practical importance because, in many scientific areas, data are being generated much faster than they can be effectively analyzed. We describe a phys. implemented robotic system that applies techniques from artificial intelligence to carry out cycles of scientific experimentation. The system automatically originates hypotheses to explain observations, devises expts. to test these hypotheses, phys. runs the expts. using a lab. robot, interprets the results to falsify hypotheses inconsistent with the data, and then repeats the cycle. Here we apply the system to the detn. of gene function using deletion mutants of yeast (Saccharomyces cerevisiae) and auxotrophic growth expts. We built and tested a detailed logical model (involving genes, proteins and metabolites) of the arom. amino acid synthesis pathway. In biol. expts. that automatically reconstruct parts of this model, we show that an intelligent expt. selection strategy is competitive with human performance and significantly outperforms, with a cost decrease of 3-fold and 100-fold (resp.), both cheapest and random-expt. selection.
- 9Chan, E. M.; Xu, C.; Mao, A. W.; Han, G.; Owen, J. S.; Cohen, B. E.; Milliron, D. J. Reproducible, High-Throughput Synthesis of Colloidal Nanocrystals for Optimization in Multidimensional Parameter Space. Nano Lett. 2010, 10 (5), 1874– 1885, DOI: 10.1021/nl100669sGoogle Scholar9https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXksF2ksbo%253D&md5=5e6ac53a08b778fdb055aca3b651ff14Reproducible, High-Throughput Synthesis of Colloidal Nanocrystals for Optimization in Multidimensional Parameter SpaceChan, Emory M.; Xu, Chenxu; Mao, Alvin W.; Han, Gang; Owen, Jonathan S.; Cohen, Bruce E.; Milliron, Delia J.Nano Letters (2010), 10 (5), 1874-1885CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)While colloidal nanocrystals hold tremendous potential for both enhancing fundamental understanding of materials scaling and enabling advanced technologies, progress in both realms can be inhibited by the limited reproducibility of traditional synthetic methods and by the difficulty of optimizing syntheses over a large no. of synthetic parameters. Here, the authors describe an automated platform for the reproducible synthesis of colloidal nanocrystals and for the high-throughput optimization of phys. properties relevant to emerging applications of nanomaterials. This robotic platform enables precise control over reaction conditions while performing workflows analogous to those of traditional flask syntheses. The authors demonstrate control over the size, size distribution, kinetics, and concn. of reactions by synthesizing CdSe nanocrystals with 0.2% coeff. of variation in the mean diams. across an array of batch reactors and over multiple runs. Leveraging this precise control along with high-throughput optical and diffraction characterization, the authors effectively map multidimensional parameter space to tune the size and polydispersity of CdSe nanocrystals, to maximize the photoluminescence efficiency of CdTe nanocrystals, and to control the crystal phase and maximize the upconverted luminescence of lanthanide-doped NaYF4 nanocrystals. From these demonstrative examples, this automated synthesis approach will be of great utility for the development of diverse colloidal nanomaterials for electronic assemblies, luminescent biol. labels, electroluminescent devices, and other emerging applications.
- 10Nikolaev, P.; Hooper, D.; Webber, F.; Rao, R.; Decker, K.; Krein, M.; Poleski, J.; Barto, R.; Maruyama, B. Autonomy in materials research: a case study in carbon nanotube growth. npj Computational Materials 2016, 2 (1), 16031, DOI: 10.1038/npjcompumats.2016.31Google ScholarThere is no corresponding record for this reference.
- 11Upadhya, R.; Kosuri, S.; Tamasi, M.; Meyer, T. A.; Atta, S.; Webb, M. A.; Gormley, A. J. Automation and data-driven design of polymer therapeutics. Adv. Drug Delivery Rev. 2021, 171, 1– 28, DOI: 10.1016/j.addr.2020.11.009Google Scholar11https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXjtlaqtL4%253D&md5=12bc6c0804c41e0a1d6baabe63f27360Automation and data-driven design of polymer therapeuticsUpadhya, Rahul; Kosuri, Shashank; Tamasi, Matthew; Meyer, Travis A.; Atta, Supriya; Webb, Michael A.; Gormley, Adam J.Advanced Drug Delivery Reviews (2021), 171 (), 1-28CODEN: ADDREP; ISSN:0169-409X. (Elsevier B.V.)A review. Polymers are uniquely suited for drug delivery and biomaterial applications due to tunable structural parameters such as length, compn., architecture, and valency. To facilitate designs, researchers may explore combinatorial libraries in a high throughput fashion to correlate structure to function. However, traditional polymn. reactions including controlled living radical polymn. (CLRP) and ring-opening polymn. (ROP) require inert reaction conditions and extensive expertise to implement. With the advent of air-tolerance and automation, several polymn. techniques are now compatible with well plates and can be carried out at the benchtop, making high throughput synthesis and high throughput screening (HTS) possible. To avoid HTS pitfalls often described as "fishing expeditions," it is crucial to employ intelligent and big data approaches to maximize exptl. efficiency. This is where the disruptive technologies of machine learning (ML) and artificial intelligence (AI) will likely play a role. In fact, ML and AI are already impacting small mol. drug discovery and showing signs of emerging in drug delivery. In this review, we present state-of-the-art research in drug delivery, gene delivery, antimicrobial polymers, and bioactive polymers alongside data-driven developments in drug design and org. synthesis. From this insight, important lessons are revealed for the polymer therapeutics community including the value of a closed loop design-build-test-learn workflow. This is an exciting time as researchers will gain the ability to fully explore the polymer structural landscape and establish quant. structure-property relationships (QSPRs) with biol. significance.
- 12Reis, M.; Gusev, F.; Taylor, N. G.; Chung, S. H.; Verber, M. D.; Lee, Y. Z.; Isayev, O.; Leibfarth, F. A. Machine-Learning-Guided Discovery of F-19 MRI Agents Enabled by Automated Copolymer Synthesis. J. Am. Chem. Soc. 2021, 143 (42), 17677– 17689, DOI: 10.1021/jacs.1c08181Google ScholarThere is no corresponding record for this reference.
- 13Tamasi, M. J.; Patel, R. A.; Borca, C. H.; Kosuri, S.; Mugnier, H.; Upadhya, R.; Murthy, N. S.; Webb, M. A.; Gormley, A. J. Machine Learning on a Robotic Platform for the Design of Polymer-Protein Hybrids. Adv. Mater. 2022, 34, 2201809, DOI: 10.1002/adma.202201809Google Scholar13https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XhsFCisLbK&md5=b9c01ddd2dd5430c9f76c5b1ccbfffd2Machine Learning on a Robotic Platform for the Design of Polymer-Protein HybridsTamasi, Matthew J.; Patel, Roshan A.; Borca, Carlos H.; Kosuri, Shashank; Mugnier, Heloise; Upadhya, Rahul; Murthy, N. Sanjeeva; Webb, Michael A.; Gormley, Adam J.Advanced Materials (Weinheim, Germany) (2022), 34 (30), 2201809CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)Polymer-protein hybrids are intriguing materials that can bolster protein stability in non-native environments, thereby enhancing their utility in diverse medicinal, com., and industrial applications. One stabilization strategy involves designing synthetic random copolymers with compns. attuned to the protein surface, but rational design is complicated by the vast chem. and compn. space. Here, a strategy is reported to design protein-stabilizing copolymers based on active machine learning, facilitated by automated material synthesis and characterization platforms. The versatility and robustness of the approach is demonstrated by the successful identification of copolymers that preserve, or even enhance, the activity of three chem. distinct enzymes following exposure to thermal denaturing conditions. Although systematic screening results in mixed success, active learning appropriately identifies unique and effective copolymer chemistries for the stabilization of each enzyme. Overall, this work broadens the capabilities to design fit-for-purpose synthetic copolymers that promote or otherwise manipulate protein activity, with extensions toward the design of robust polymer-protein hybrid materials.
- 14Knox, S. T.; Parkinson, S. J.; Wilding, C. Y. P.; Bourne, R. A.; Warren, N. J. Autonomous polymer synthesis delivered by multi-objective closed-loop optimization. Polym. Chem. 2022, 13 (11), 1576– 1585, DOI: 10.1039/D2PY00040GGoogle Scholar14https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XkvFehtb4%253D&md5=9a3189dbbdedc30e4626e025dcc196d7Autonomous polymer synthesis delivered by multi-objective closed-loop optimizationKnox, Stephen T.; Parkinson, Sam J.; Wilding, Clarissa Y. P.; Bourne, Richard A.; Warren, Nicholas J.Polymer Chemistry (2022), 13 (11), 1576-1585CODEN: PCOHC2; ISSN:1759-9962. (Royal Society of Chemistry)Application of artificial intelligence and machine learning for polymer discovery offers an opportunity to meet the drastic need for the next generation high performing and sustainable polymer materials. Here, these technologies were employed within a computationally controlled flow reactor which enabled self-optimization of a range of RAFT polymn. formulations. This allowed for autonomous identification of optimum reaction conditions to afford targeted polymer properties - the first demonstration of closed loop (i.e. user-free) optimization for multiple objectives in polymer synthesis. The synthesis platform comprised a computer-controlled flow reactor, online benchtop NMR and inline gel permeation chromatog. (GPC). The RAFT polymn. of tert-Bu acrylamide (tBuAm), Bu acrylate (BuA) and Me methacrylate (MMA) were optimized using the Thompson sampling efficient multi-objective optimization (TSEMO) algorithm which explored the trade-off between molar mass dispersity (D) and monomer conversion without user interaction. The pressurised computer-controlled flow reactor allowed for polymn. in normally "forbidden" conditions - without degassing and at temps. higher than the normal b.p. of the solvent. Autonomous experimentation included comparison of five different RAFT agents for the polymn. of tBuAm, an investigation into the effects of polymn. inhibition using BuA and intensification of the otherwise slow MMA polymn.
- 15Rubens, M.; Vrijsen, J. H.; Laun, J.; Junkers, T. Precise Polymer Synthesis by Autonomous Self-Optimizing Flow Reactors. Angew. Chem., Int. Ed. 2019, 58 (10), 3183– 3187, DOI: 10.1002/anie.201810384Google Scholar15https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXit1yjsr7N&md5=155743831af59c7bd6ac81545294817dPrecise Polymer Synthesis by Autonomous Self-Optimizing Flow ReactorsRubens, Maarten; Vrijsen, Jeroen H.; Laun, Joachim; Junkers, TanjaAngewandte Chemie, International Edition (2019), 58 (10), 3183-3187CODEN: ACIEF5; ISSN:1433-7851. (Wiley-VCH Verlag GmbH & Co. KGaA)A novel continuous flow system for automated high-throughput screening, autonomous optimization, and enhanced process control of polymns. was developed. The computer-controlled platform comprises a flow reactor coupled to size exclusion chromatog. (SEC). Mol. wt. distributions are measured online and used by a machine-learning algorithm to self-optimize reactions towards a programmed mol. wt. by dynamically varying reaction parameters (i.e. residence time, monomer concn., and control agent/initiator concn.). The autonomous platform allows targeting of mol. wts. in a reproducible manner with unprecedented accuracy (<2.5 % deviation from pre-selected goal) for both thermal and light-induced reactions. For the first time, polymers with predefined mol. wts. can be custom made under optimal reaction conditions in an automated, high-throughput flow synthesis approach with outstanding reproducibility.
- 16Cao, L.; Russo, D.; Felton, K.; Salley, D.; Sharma, A.; Keenan, G.; Mauer, W.; Gao, H.; Cronin, L.; Lapkin, A. A. Optimization of Formulations Using Robotic Experiments Driven by Machine Learning DoE. Cell Reports Physical Science 2021, 2 (1), 100295, DOI: 10.1016/j.xcrp.2020.100295Google ScholarThere is no corresponding record for this reference.
- 17Vriza, A.; Chan, H.; Xu, J. Self-Driving Laboratory for Polymer Electronics. Chem. Mater. 2023, 35 (8), 3046– 3056, DOI: 10.1021/acs.chemmater.2c03593Google Scholar17https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXkslajt78%253D&md5=6f3abd3fa1bc1e249c1017761c49ab47Self-Driving Laboratory for Polymer ElectronicsVriza, Aikaterini; Chan, Henry; Xu, JieChemistry of Materials (2023), 35 (8), 3046-3056CODEN: CMATEX; ISSN:0897-4756. (American Chemical Society)A review. Owing to the chem. pluripotency and viscoelastic nature of electronic polymers, polymer electronics has shown unique advances in many emerging applications such as skin-like electronics, large-area printed energy devices and neuromorphic computing devices, but their development period is years-long. Recent advancements in automation, robotics, and learning algorithms have led to a growing no. of self-driving (autonomous) labs. that have begun to revolutionize the development and accelerated discovery of materials. In this perspective, we first introduce the current state of autonomous labs. Then we analyze why it is challenging to conduct polymer electronic research by autonomous lab. and highlight the needs. We further discuss our efforts in building an autonomous lab., namely Polybot, for the automated synthesis and characterization of electronic polymers and their processing and fabrication into electronic devices. Finally, we share our vision in using a self-driving lab. (SDL) for different types of polymer electronics research.
- 18Gongora, A. E.; Xu, B.; Perry, W.; Okoye, C.; Riley, P.; Reyes, K. G.; Morgan, E. F.; Brown, K. A. A Bayesian experimental autonomous researcher for mechanical design. Science Advances 2020, 6 (15), eaaz1708 DOI: 10.1126/sciadv.aaz1708Google ScholarThere is no corresponding record for this reference.
- 19Bateni, F.; Sadeghi, S.; Orouji, N.; Bennett, J. A.; Punati, V. S.; Stark, C.; Wang, J.; Rosko, M. C.; Chen, O.; Castellano, F. N.; Reyes, K. G.; Abolhasani, M. Smart Dope: A Self-Driving Fluidic Lab for Accelerated Development of Doped Perovskite Quantum Dots. Adv. Energy Mater. 2024, 14 (1), 2302303, DOI: 10.1002/aenm.202302303Google ScholarThere is no corresponding record for this reference.
- 20Lee, J.; Mulay, P.; Tamasi, M. J.; Yeow, J.; Stevens, M. M.; Gormley, A. J. A fully automated platform for photoinitiated RAFT polymerization. Digital Discovery 2023, 2 (1), 219– 233, DOI: 10.1039/D2DD00100DGoogle Scholar20https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXitFeju7nF&md5=2613a3e3ae1bdf2542c9791a49e6e147A fully automated platform for photoinitiated RAFT polymerizationLee, Jules; Mulay, Prajakatta; Tamasi, Matthew J.; Yeow, Jonathan; Stevens, Molly M.; Gormley, Adam J.Digital Discovery (2023), 2 (1), 219-233CODEN: DDIIAI; ISSN:2635-098X. (Royal Society of Chemistry)Oxygen tolerant polymns. including Photoinduced Electron/Energy Transfer-Reversible Addn.-Fragmentation Chain-Transfer (PET-RAFT) polymn. allow for high-throughput synthesis of diverse polymer architectures on the benchtop in parallel. Recent developments have further increased throughput using liq. handling robotics to automate reagent handling and dispensing into well plates thus enabling the combinatorial synthesis of large polymer libraries. Although liq. handling robotics can enable automated polymer reagent dispensing in well plates, photoinitiation and reaction monitoring require automation to provide a platform that enables the reliable and robust synthesis of various polymer compns. in high-throughput where polymers with desired mol. wts. and low dispersity are obtained. Here, we describe the development of a robotic platform to fully automate PET-RAFT polymns. and provide individual control of reactions performed in well plates. On our platform, reagents are automatically dispensed in well plates, photoinitiated in individual wells with a custom-designed lightbox until the polymns. are complete, and monitored online in real-time by tracking fluorescence intensities on a fluorescence plate reader, with well plate transfers between instruments occurring via a robotic arm. We found that this platform enabled robust parallel polymer synthesis of both acrylate and acrylamide homopolymers and copolymers, with high monomer conversions and low dispersity. The successful polymns. obtained on this platform make it an efficient tool for combinatorial polymer chem. In addn., with the inclusion of machine learning protocols to help navigate the polymer space towards specific properties of interest, this robotic platform can ultimately become a self-driving lab that can dispense, synthesize, and monitor large polymer libraries.
- 21Politi, M.; Baum, F.; Vaddi, K.; Antonio, E.; Vasquez, J.; Bishop, B. P.; Peek, N.; Holmberg, V. C.; Pozzo, L. D. A high-throughput workflow for the synthesis of CdSe nanocrystals using a sonochemical materials acceleration platform. Digital Discovery 2023, 2 (4), 1042– 1057, DOI: 10.1039/D3DD00033HGoogle ScholarThere is no corresponding record for this reference.
- 22Beaucage, P. A.; Martin, T. B. The Autonomous Formulation Laboratory: An Open Liquid Handling Platform for Formulation Discovery Using X-ray and Neutron Scattering. Chem. Mater. 2023, 35 (3), 846– 852, DOI: 10.1021/acs.chemmater.2c03118Google ScholarThere is no corresponding record for this reference.
- 23Rahmanian, F.; Fuchs, S.; Zhang, B.; Fichtner, M.; Stein, H. S. Autonomous millimeter scale high throughput battery research system. Digital Discovery 2024, 3 (5), 883– 895, DOI: 10.1039/D3DD00257HGoogle ScholarThere is no corresponding record for this reference.
- 24Kusne, A. G.; Yu, H.; Wu, C.; Zhang, H.; Hattrick-Simpers, J.; DeCost, B.; Sarker, S.; Oses, C.; Toher, C.; Curtarolo, S.; Davydov, A. V.; Agarwal, R.; Bendersky, L. A.; Li, M.; Mehta, A.; Takeuchi, I. On-the-fly closed-loop materials discovery via Bayesian active learning. Nat. Commun. 2020, 11 (1), 5966, DOI: 10.1038/s41467-020-19597-wGoogle Scholar24https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXisVOnsL3N&md5=138fb31700fa99c1962c048fe7829145On-the-fly closed-loop materials discovery via Bayesian active learningKusne, A. Gilad; Yu, Heshan; Wu, Changming; Zhang, Huairuo; Hattrick-Simpers, Jason; DeCost, Brian; Sarker, Suchismita; Oses, Corey; Toher, Cormac; Curtarolo, Stefano; Davydov, Albert V.; Agarwal, Ritesh; Bendersky, Leonid A.; Li, Mo; Mehta, Apurva; Takeuchi, IchiroNature Communications (2020), 11 (1), 5966CODEN: NCAOBW; ISSN:2041-1723. (Nature Research)Active learning-the field of machine learning (ML) dedicated to optimal expt. design-has played a part in science as far back as the 18th century when Laplace used it to guide his discovery of celestial mechanics. In this work, we focus a closed-loop, active learning-driven autonomous system on another major challenge, the discovery of advanced materials against the exceedingly complex synthesis-processes-structure-property landscape. We demonstrate an autonomous materials discovery methodol. for functional inorg. compds. which allow scientists to fail smarter, learn faster, and spend less resources in their studies, while simultaneously improving trust in scientific results and machine learning tools. This robot science enables science-over-the-network, reducing the economic impact of scientists being phys. sepd. from their labs. The real-time closed-loop, autonomous system for materials exploration and optimization (CAMEO) is implemented at the synchrotron beamline to accelerate the interconnected tasks of phase mapping and property optimization, with each cycle taking seconds to minutes. We also demonstrate an embodiment of human-machine interaction, where human-in-the-loop is called to play a contributing role within each cycle. This work has resulted in the discovery of a novel epitaxial nanocomposite phase-change memory material.
- 25Ament, S.; Amsler, M.; Sutherland, D. R.; Chang, M.-C.; Guevarra, D.; Connolly, A. B.; Gregoire, J. M.; Thompson, M. O.; Gomes, C. P.; van Dover, R. B. Autonomous materials synthesis via hierarchical active learning of nonequilibrium phase diagrams. Science Advances 2021, 7 (51), 12, DOI: 10.1126/sciadv.abg4930Google ScholarThere is no corresponding record for this reference.
- 26Lo, S.; Baird, S. G.; Schrier, J.; Blaiszik, B.; Carson, N.; Foster, I.; Aguilar-Granda, A.; Kalinin, S. V.; Maruyama, B.; Politi, M.; Tran, H.; Sparks, T. D.; Aspuru-Guzik, A. Review of low-cost self-driving laboratories in chemistry and materials science: the “frugal twin” concept. Digital Discovery 2024, 3 (5), 842– 868, DOI: 10.1039/D3DD00223CGoogle ScholarThere is no corresponding record for this reference.
- 27Martin, T. B.; Audus, D. J. Emerging Trends in Machine Learning: A Polymer Perspective. ACS Polymers Au 2023, 3 (3), 239– 258, DOI: 10.1021/acspolymersau.2c00053Google Scholar27https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXht1Cgu7o%253D&md5=39e8ea57891e174a0f5957e251f8de61Emerging Trends in Machine Learning: A Polymer PerspectiveMartin, Tyler B.; Audus, Debra J.ACS Polymers Au (2023), 3 (3), 239-258CODEN: APACCD; ISSN:2694-2453. (American Chemical Society)A review. In the last five years, there has been tremendous growth in machine learning and artificial intelligence as applied to polymer science. Here, we highlight the unique challenges presented by polymers and how the field is addressing them. We focus on emerging trends with an emphasis on topics that have received less attention in the review literature. Finally, we provide an outlook for the field and outline important growth areas in machine learning and artificial intelligence for polymer science and discuss important advances from the greater material science community.
- 28Heil, C. M.; Patil, A.; Dhinojwala, A.; Jayaraman, A. Computational Reverse-Engineering Analysis for Scattering Experiments (CREASE) with Machine Learning Enhancement to Determine Structure of Nanoparticle Mixtures and Solutions. ACS central science 2022, 8 (7), 996– 1007, DOI: 10.1021/acscentsci.2c00382Google ScholarThere is no corresponding record for this reference.
- 29Treece, B. W.; Kienzle, P. A.; Hoogerheide, D. P.; Majkrzak, C. F.; Losche, M.; Heinrich, F. Optimization of reflectometry experiments using information theory. J. Appl. Crystallogr. 2019, 52 (1), 47– 59, DOI: 10.1107/S1600576718017016Google ScholarThere is no corresponding record for this reference.
- 30Durant, J. H.; Wilkins, L.; Cooper, J. F. K. Optimizing experimental design in neutron reflectometry. J. Appl. Crystallogr. 2022, 55 (4), 769– 781, DOI: 10.1107/S1600576722003831Google Scholar30https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XitlWitr3K&md5=bc22f593ccc372fcb6d40b5c84fa430fOptimizing experimental design in neutron reflectometryDurant, James H.; Wilkins, Lucas; Cooper, Joshaniel F. K.Journal of Applied Crystallography (2022), 55 (4), 769-781CODEN: JACGAR; ISSN:1600-5767. (International Union of Crystallography)Using the Fisher information (FI), the design of neutron reflectometry expts. can be optimized, leading to greater confidence in parameters of interest and better use of exptl. time [Durant, Wilkins, Butler & Cooper (2021). J. Appl. Cryst.54, 1100-1110]. In this work, the FI is utilized in optimizing the design of a wide range of reflectometry expts. Two lipid bilayer systems are investigated to det. the optimal choice of measurement angles and liq. contrasts, in addn. to the ratio of the total counting time that should be spent measuring each condition. The redn. in parameter uncertainties with the addn. of underlayers to these systems is then quantified, using the FI, and validated through the use of expt. simulation and Bayesian sampling methods. For a 'one-shot' measurement of a degrading lipid monolayer, it is shown that the common practice of measuring null-reflecting water is indeed optimal, but that the optimal measurement angle is dependent on the deuteration state of the monolayer. Finally, the framework is used to demonstrate the feasibility of measuring magnetic signals as small as 0.01 μB per atom in layers only 20 Å thick, given the appropriate exptl. design, and that the time to reach a given level of confidence in the small magnetic moment is quantifiable.
- 31Kjær, E. T. S.; Anker, A. S.; Weng, M. N.; Billinge, S. J. L.; Selvan, R.; Jensen, K. M. Ø. DeepStruc: towards structure solution from pair distribution function data using deep generative models. Digital Discovery 2023, 2 (1), 69– 80, DOI: 10.1039/D2DD00086EGoogle ScholarThere is no corresponding record for this reference.
- 32Walsh, D. J.; Zou, W.; Schneider, L.; Mello, R.; Deagen, M. E.; Mysona, J.; Lin, T.-S.; de Pablo, J. J.; Jensen, K. F.; Audus, D. J.; Olsen, B. D. Community Resource for Innovation in Polymer Technology (CRIPT): A Scalable Polymer Material Data Structure. ACS Central Science 2023, 9 (3), 330– 338, DOI: 10.1021/acscentsci.3c00011Google Scholar32https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXjsVels7Y%253D&md5=f524fbe2207619b331eb96f85b233ed3Community Resource for Innovation in Polymer Technology (CRIPT): A Scalable Polymer Material Data StructureWalsh, Dylan J.; Zou, Weizhong; Schneider, Ludwig; Mello, Reid; Deagen, Michael E.; Mysona, Joshua; Lin, Tzyy-Shyang; de Pablo, Juan J.; Jensen, Klavs F.; Audus, Debra J.; Olsen, Bradley D.ACS Central Science (2023), 9 (3), 330-338CODEN: ACSCII; ISSN:2374-7951. (American Chemical Society)Polymeric materials are integral components of nearly every aspect of modern life. However, developing cheminformatic solns. for polymers has been difficult since they are large stochastic mols. with hierarchical structures spanning multiple length scales. Here we present the design for a general material data model that underpins the Community Resource for Innovation in Polymer Technol. (CRIPT) data ecosystem.
- 33Schneider, L.; Walsh, D.; Olsen, B.; de Pablo, J. Generative BigSMILES: an extension for polymer informatics, computer simulations & ML/AI. Digital Discovery 2024, 3 (1), 51– 61, DOI: 10.1039/D3DD00147DGoogle ScholarThere is no corresponding record for this reference.
- 34Lu, S.; Jayaraman, A. Pair-Variational Autoencoders for Linking and Cross-Reconstruction of Characterization Data from Complementary Structural Characterization Techniques. JACS Au 2023, 3 (9), 2510– 2521, DOI: 10.1021/jacsau.3c00275Google ScholarThere is no corresponding record for this reference.
- 35Sutliff, B. P.; Goyal, S.; Martin, T. B.; Beaucage, P. A.; Audus, D. J.; Orski, S. V. Correlating Near-Infrared Spectra to Bulk Properties in Polyolefins. Macromolecules 2024, 57 (5), 2329– 2338, DOI: 10.1021/acs.macromol.3c02290Google ScholarThere is no corresponding record for this reference.
- 36Edwards, C. E. R.; Lakkis, K. L.; Luo, Y.; Helgeson, M. E. Coacervate or precipitate? Formation of non-equilibrium microstructures in coacervate emulsions. Soft Matter 2023, 19 (45), 8849– 8862, DOI: 10.1039/D3SM00901GGoogle ScholarThere is no corresponding record for this reference.
- 37Lee, S.; Bluemle, M. J.; Bates, F. S. Discovery of a Frank-Kasper σ Phase in Sphere-Forming Block Copolymer Melts. Science 2010, 330 (6002), 349– 353, DOI: 10.1126/science.1195552Google Scholar37https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXht1Ois7fE&md5=03af9cd3e115135671009f5bf28607d4Discovery of a Frank-Kasper σ Phase in Sphere-Forming Block Copolymer MeltsLee, Sangwoo; Bluemle, Michael J.; Bates, Frank S.Science (Washington, DC, United States) (2010), 330 (6002), 349-353CODEN: SCIEAS; ISSN:0036-8075. (American Association for the Advancement of Science)Sphere-forming block copolymers are known to self-assemble into body-centered cubic crystals near the order-disorder transition temp. Small-angle x-ray scattering and transmission electron microscopy expts. on diblock and tetrablock copolymer melts have revealed an equil. phase characterized by a large tetragonal unit cell contg. 30 microphase-sepd. spheres. This structure, referred to as the sigma (σ) phase by Frank and Kasper more than 50 years ago, nucleates and grows from the body-centered cubic phase similar to its occurrence in metal alloys and is a crystal approximant to dodecagonal quasicrystals. Formation of the σ phase in undiluted linear block copolymers (and certain branched dendrimers) appears to be mediated by macromol. packing frustration, an entropic contribution to the interparticle interactions that control the sphere-packing geometry.
- 38Kim, K.; Arora, A.; Lewis, R. M.; Liu, M.; Li, W.; Shi, A.-C.; Dorfman, K. D.; Bates, F. S. Origins of low-symmetry phases in asymmetric diblock copolymer melts. Proc. Natl. Acad. Sci. U. S. A. 2018, 115 (5), 847– 854, DOI: 10.1073/pnas.1717850115Google Scholar38https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhtlKqt78%253D&md5=4eb24dfe1e67ab529452bf05ea9a6af6Origins of low-symmetry phases in asymmetric diblock copolymer meltsKim, Kyungtae; Arora, Akash; Lewis, Ronald M., III; Liu, Meijiao; Li, Weihua; Shi, An-Chang; Dorfman, Kevin D.; Bates, Frank S.Proceedings of the National Academy of Sciences of the United States of America (2018), 115 (5), 847-854CODEN: PNASA6; ISSN:0027-8424. (National Academy of Sciences)Cooling disordered compositionally asym. diblock copolymers leads to the formation of nearly spherical particles, each contg. hundreds of mols., which crystallize upon cooling below the order-disorder transition temp. (TODT). SCF theory (SCFT) reveals that dispersity in the block ds.p. stabilizes various Frank-Kasper phases, including the C14 and C15 Laves phases, which were accessed exptl. in low-molar-mass poly(isoprene)-b-poly(lactide) (PI-PLA) diblock copolymers using thermal processing strategies. Heating and cooling a specimen contg. 15% PLA above and below the TODT from the body-centered cubic (BCC) or C14 states regenerates the same cryst. order established at lower temps. This memory effect is also demonstrated with a specimen contg. 20% PLA, which recrystallizes to either C15 or hexagonally ordered cylinders (HEXC) upon heating and cooling. The process-path-dependent formation of cryst. order shapes the no. of particles per unit vol., n/V, which is retained in the highly structured disordered liq. as revealed by small-angle X-ray scattering (SAXS) expts. We hypothesize that symmetry breaking during crystn. is governed by the particle no. d. imprinted in the liq. during ordering at lower temp., and this metastable liq. is kinetically constrained from equilibrating due to prohibitively large free energy barriers for micelle fusion and fission. Ordering at fixed n/V is enabled by facile chain exchange, which redistributes mass as required to meet the multiple particle sizes and packing assocd. with specific low-symmetry Frank-Kasper phases. This discovery exposes universal concepts related to order and disorder in self-assembled soft materials.
- 39Biswas, A.; Liu, Y.; Creange, N.; Liu, Y.-C.; Jesse, S.; Yang, J.-C.; Kalinin, S. V.; Ziatdinov, M. A.; Vasudevan, R. K. A dynamic Bayesian optimized active recommender system for curiosity-driven partially Human-in-the-loop automated experiments. npj Computational Materials 2024, 10 (1), 29, DOI: 10.1038/s41524-023-01191-5Google ScholarThere is no corresponding record for this reference.
- 40U.S. National Science Foundation: Award Abstract # 2152210 Graduate Traineeship on Advances in Materials Science using Machine Learning. https://www.nsf.gov/awardsearch/showAward?AWD_ID=2152210 (accessed 2024-08).Google ScholarThere is no corresponding record for this reference.
- 41Adorf, C. S.; Ramasubramani, V.; Anderson, J. A.; Glotzer, S. C. How to Professionally Develop Reusable Scientific Software─And When Not To. Computing in Science & Engineering 2019, 21 (2), 66– 79, DOI: 10.1109/MCSE.2018.2882355Google ScholarThere is no corresponding record for this reference.
- 42Zhao, Z.; Chavez, T.; Holman, E. A.; Hao, G.; Green, A.; Krishnan, H.; McReynolds, D.; Pandolfi, R. J.; Roberts, E. J.; Zwart, P. H.; Yanxon, H.; Schwarz, N.; Sankaranarayanan, S.; Kalinin, S. V.; Mehta, A.; Campbell, S. I.; Hexemer, A. MLExchange: A web-based platform enabling exchangeable machine learning workflows for scientific studies. In 2022 4th Annual Workshop on Extreme-scale Experiment-in-the-Loop Computing (XLOOP); IEEE, 2022.Google ScholarThere is no corresponding record for this reference.
- 43Hugging Face. https://huggingface.co (accessed 2024-08).Google ScholarThere is no corresponding record for this reference.
- 44Pelkie, B. G.; Pozzo, L. D. The laboratory of Babel: highlighting community needs for integrated materials data management. Digital Discovery 2023, 2 (3), 544– 556, DOI: 10.1039/D3DD00022BGoogle ScholarThere is no corresponding record for this reference.
- 45McDannald, A.; Frontzek, M.; Savici, A. T.; Doucet, M.; Rodriguez, E. E.; Meuse, K.; Opsahl-Ong, J.; Samarov, D.; Takeuchi, I.; Ratcliff, W.; Kusne, A. G. On-the-fly autonomous control of neutron diffraction via physics-informed Bayesian active learning. Applied Physics Reviews 2022, 9 (2), 021408, DOI: 10.1063/5.0082956Google ScholarThere is no corresponding record for this reference.
- 46Audus, D. J.; McDannald, A.; DeCost, B. Leveraging Theory for Enhanced Machine Learning. ACS Macro Lett. 2022, 11 (9), 1117– 1122, DOI: 10.1021/acsmacrolett.2c00369Google ScholarThere is no corresponding record for this reference.
- 47Ethier, J. G.; Audus, D. J.; Ryan, D. C.; Vaia, R. A. Integrating theory with machine learning for predicting polymer solution phase behavior. Giant 2023, 15, 100171, DOI: 10.1016/j.giant.2023.100171Google ScholarThere is no corresponding record for this reference.
- 48Raissi, M.; Perdikaris, P.; Karniadakis, G. E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 2019, 378, 686– 707, DOI: 10.1016/j.jcp.2018.10.045Google ScholarThere is no corresponding record for this reference.
- 49Qu, C.; Schneider, B. I.; Kearsley, A. J.; Keyrouz, W.; Allison, T. C. Applying graph neural network models to molecular property prediction using high-quality experimental data. Artificial Intelligence Chemistry 2024, 2 (1), 100050, DOI: 10.1016/j.aichem.2024.100050Google ScholarThere is no corresponding record for this reference.
- 50University of Toronto Acceleration Consortium. https://acceleration.utoronto.ca/ (accessed 2024-06).Google ScholarThere is no corresponding record for this reference.
- 51CASE: The Community for Autonomous Scientific Experimentation. https://autonomous-discovery.lbl.gov/ (accessed 2024-06).Google ScholarThere is no corresponding record for this reference.
- 52Kiefer, K.; Rossa, L.; Wutzler, F.; Ektrom, N.; Pettersson, A.; Faulhaber, E.; Zolliker, M. An introduction to SECoP – the sample environment communication protocol. Journal of Neutron Research 2020, 21, 181– 195, DOI: 10.3233/JNR-190143Google ScholarThere is no corresponding record for this reference.
- 53Wierenga, R. P.; Golas, S. M.; Ho, W.; Coley, C. W.; Esvelt, K. M. PyLabRobot: An open-source, hardware-agnostic interface for liquid-handling robots and accessories. Device 2023, 1 (4), 100111, DOI: 10.1016/j.device.2023.100111Google ScholarThere is no corresponding record for this reference.
Cited By
This article has not yet been cited by other publications.
Article Views
Altmetric
Citations
Article Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.
Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.
The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated.
Recommended Articles
References
This article references 53 other publications.
- 1Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles; J3016_202104; SAE International, 2021.There is no corresponding record for this reference.
- 2Meredith, C.; Smith, A. P.; Crosby, A. J.; Amis, E. J.; Karim, A. Combinatorial Methods for Polymer Science. In Encyclopedia of Polymer Science and Technology; Wiley, DOI: 10.1002/0471440264.pst069 .There is no corresponding record for this reference.
- 3Hoogenboom, R.; Meier, M. A. R.; Schubert, U. S. Combinatorial Methods, Automated Synthesis and High-Throughput Screening in Polymer Research: Past and Present. Macromol. Rapid Commun. 2003, 24 (1), 15– 32, DOI: 10.1002/marc.2003900133https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXhs1Siu7o%253D&md5=0253f289489c86ddef3ed65b3e75f3edCombinatorial methods, automated synthesis and high-throughput screening in polymer research: Past and presentHoogenboom, Richard; Meier, Michael A. R.; Schubert, Ulrich S.Macromolecular Rapid Communications (2003), 24 (1), 15-32CODEN: MRCOE3; ISSN:1022-1336. (Wiley-VCH Verlag GmbH & Co. KGaA)A review describes polymer research related to high-throughput screening development. Combinatorial techniques, parallel experimentation and high-throughput methods represent a very promising approach in order to speed up the prepn. and investigation of new polymeric materials: a large variety of parameters can be screened simultaneously resulting in new structure/property relationships. The field of polymer research seems to be perfectly suited for parallel and combinatorial methods due to the fact that many parameters can be varied during synthesis, processing, blending as well as compounding. In addn., numerous important parameters have to be investigated, such as mol. wt., polydispersity, viscosity, hardness, stiffness and other application-specific properties. A no. of corresponding high-throughput techniques have been developed in the last few years and their introduction into the com. market further boosted the development. These combinatorial approaches can reduce the time-to-market for new polymeric materials drastically compared to traditional approaches and allow a much more detailed understanding of polymers from the macroscopic to the nanoscopic scale. Here we provide an overview of the present status of combinatorial and parallel polymer synthesis and high-throughput screening.
- 4Zaquen, N.; Rubens, M.; Corrigan, N.; Xu, J.; Zetterlund, P. B.; Boyer, C.; Junkers, T. Polymer Synthesis in Continuous Flow Reactors. Prog. Polym. Sci. 2020, 107, 101256, DOI: 10.1016/j.progpolymsci.2020.1012564https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXht1Wmur3I&md5=d157f9944158582c064dde6fb3eed7cePolymer Synthesis in Continuous Flow ReactorsZaquen, Neomy; Rubens, Maarten; Corrigan, Nathaniel; Xu, Jiangtao; Zetterlund, Per B.; Boyer, Cyrille; Junkers, TanjaProgress in Polymer Science (2020), 107 (), 101256CODEN: PRPSB8; ISSN:0079-6700. (Elsevier Ltd.)A variety of polymns. has long been performed in continuous flow reactors on an industrial scale; comparatively, on smaller scales, continuous polymn. methods have only gained significant attention in recent years. Yet, within the last decade, the field has moved from the rare occurrence of flow reactors to their abundant use today. A wide variety of polymer reactions have been performed in a continuous fashion on small and intermediate scales. The advantages of applying flow chem. principles for polymer reactions include increased reproducibility and synthetic precision, significant increases in reaction performances for photochem. reactions, the ability to couple reactors to create complex materials in a single reactor pass, as well as the unique combination of online monitoring and machine learning. In this review we give a comprehensive overview of polymer reactions being carried out in continuous flow reactors to date. The development of the field is discussed, concluding with the most recent examples on automated polymer synthesis, reactor telescoping and nanoparticle synthesis. Finally, the design of flow reactors is discussed to help newcomers contribute to the current and future developments in the field.
- 5Tamasi, M.; Kosuri, S.; DiStefano, J.; Chapman, R.; Gormley, A. J. Automation of Controlled/Living Radical Polymerization. Advanced Intelligent Systems 2020, 2 (2), 1900126, DOI: 10.1002/aisy.201900126There is no corresponding record for this reference.
- 6Winicov, H.; Schainbaum, J.; Buckley, J.; Longino, G.; Hill, J.; Berkoff, C. E. Chemical process optimization by computer ─ a self-directed chemical synthesis system. Anal. Chim. Acta 1978, 103 (4), 469– 476, DOI: 10.1016/S0003-2670(01)83110-XThere is no corresponding record for this reference.
- 7MATSUDA, R.; ISHIBASHI, M.; TAKEDA, Y. Simplex Optimization of Reaction Conditions with an Automated System. CHEMICAL & PHARMACEUTICAL BULLETIN 1988, 36 (9), 3512– 3518, DOI: 10.1248/cpb.36.3512There is no corresponding record for this reference.
- 8King, R. D.; Whelan, K. E.; Jones, F. M.; Reiser, P. G. K.; Bryant, C. H.; Muggleton, S. H.; Kell, D. B.; Oliver, S. G. Functional genomic hypothesis generation and experimentation by a robot scientist. Nature 2004, 427 (6971), 247– 252, DOI: 10.1038/nature022368https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXjtFKjuw%253D%253D&md5=58c844f52c70fa8f9030707129cd6472Functional genomic hypothesis generation and experimentation by a robot scientistKing, Ross D.; Whelan, Kenneth E.; Jones, Ffion M.; Reiser, Philip G. K.; Bryant, Christopher H.; Muggleton, Stephen H.; Kell, Douglas B.; Oliver, Stephen G.Nature (London, United Kingdom) (2004), 427 (6971), 247-252CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)The question of whether it is possible to automate the scientific process is of both great theor. interest and increasing practical importance because, in many scientific areas, data are being generated much faster than they can be effectively analyzed. We describe a phys. implemented robotic system that applies techniques from artificial intelligence to carry out cycles of scientific experimentation. The system automatically originates hypotheses to explain observations, devises expts. to test these hypotheses, phys. runs the expts. using a lab. robot, interprets the results to falsify hypotheses inconsistent with the data, and then repeats the cycle. Here we apply the system to the detn. of gene function using deletion mutants of yeast (Saccharomyces cerevisiae) and auxotrophic growth expts. We built and tested a detailed logical model (involving genes, proteins and metabolites) of the arom. amino acid synthesis pathway. In biol. expts. that automatically reconstruct parts of this model, we show that an intelligent expt. selection strategy is competitive with human performance and significantly outperforms, with a cost decrease of 3-fold and 100-fold (resp.), both cheapest and random-expt. selection.
- 9Chan, E. M.; Xu, C.; Mao, A. W.; Han, G.; Owen, J. S.; Cohen, B. E.; Milliron, D. J. Reproducible, High-Throughput Synthesis of Colloidal Nanocrystals for Optimization in Multidimensional Parameter Space. Nano Lett. 2010, 10 (5), 1874– 1885, DOI: 10.1021/nl100669s9https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXksF2ksbo%253D&md5=5e6ac53a08b778fdb055aca3b651ff14Reproducible, High-Throughput Synthesis of Colloidal Nanocrystals for Optimization in Multidimensional Parameter SpaceChan, Emory M.; Xu, Chenxu; Mao, Alvin W.; Han, Gang; Owen, Jonathan S.; Cohen, Bruce E.; Milliron, Delia J.Nano Letters (2010), 10 (5), 1874-1885CODEN: NALEFD; ISSN:1530-6984. (American Chemical Society)While colloidal nanocrystals hold tremendous potential for both enhancing fundamental understanding of materials scaling and enabling advanced technologies, progress in both realms can be inhibited by the limited reproducibility of traditional synthetic methods and by the difficulty of optimizing syntheses over a large no. of synthetic parameters. Here, the authors describe an automated platform for the reproducible synthesis of colloidal nanocrystals and for the high-throughput optimization of phys. properties relevant to emerging applications of nanomaterials. This robotic platform enables precise control over reaction conditions while performing workflows analogous to those of traditional flask syntheses. The authors demonstrate control over the size, size distribution, kinetics, and concn. of reactions by synthesizing CdSe nanocrystals with 0.2% coeff. of variation in the mean diams. across an array of batch reactors and over multiple runs. Leveraging this precise control along with high-throughput optical and diffraction characterization, the authors effectively map multidimensional parameter space to tune the size and polydispersity of CdSe nanocrystals, to maximize the photoluminescence efficiency of CdTe nanocrystals, and to control the crystal phase and maximize the upconverted luminescence of lanthanide-doped NaYF4 nanocrystals. From these demonstrative examples, this automated synthesis approach will be of great utility for the development of diverse colloidal nanomaterials for electronic assemblies, luminescent biol. labels, electroluminescent devices, and other emerging applications.
- 10Nikolaev, P.; Hooper, D.; Webber, F.; Rao, R.; Decker, K.; Krein, M.; Poleski, J.; Barto, R.; Maruyama, B. Autonomy in materials research: a case study in carbon nanotube growth. npj Computational Materials 2016, 2 (1), 16031, DOI: 10.1038/npjcompumats.2016.31There is no corresponding record for this reference.
- 11Upadhya, R.; Kosuri, S.; Tamasi, M.; Meyer, T. A.; Atta, S.; Webb, M. A.; Gormley, A. J. Automation and data-driven design of polymer therapeutics. Adv. Drug Delivery Rev. 2021, 171, 1– 28, DOI: 10.1016/j.addr.2020.11.00911https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXjtlaqtL4%253D&md5=12bc6c0804c41e0a1d6baabe63f27360Automation and data-driven design of polymer therapeuticsUpadhya, Rahul; Kosuri, Shashank; Tamasi, Matthew; Meyer, Travis A.; Atta, Supriya; Webb, Michael A.; Gormley, Adam J.Advanced Drug Delivery Reviews (2021), 171 (), 1-28CODEN: ADDREP; ISSN:0169-409X. (Elsevier B.V.)A review. Polymers are uniquely suited for drug delivery and biomaterial applications due to tunable structural parameters such as length, compn., architecture, and valency. To facilitate designs, researchers may explore combinatorial libraries in a high throughput fashion to correlate structure to function. However, traditional polymn. reactions including controlled living radical polymn. (CLRP) and ring-opening polymn. (ROP) require inert reaction conditions and extensive expertise to implement. With the advent of air-tolerance and automation, several polymn. techniques are now compatible with well plates and can be carried out at the benchtop, making high throughput synthesis and high throughput screening (HTS) possible. To avoid HTS pitfalls often described as "fishing expeditions," it is crucial to employ intelligent and big data approaches to maximize exptl. efficiency. This is where the disruptive technologies of machine learning (ML) and artificial intelligence (AI) will likely play a role. In fact, ML and AI are already impacting small mol. drug discovery and showing signs of emerging in drug delivery. In this review, we present state-of-the-art research in drug delivery, gene delivery, antimicrobial polymers, and bioactive polymers alongside data-driven developments in drug design and org. synthesis. From this insight, important lessons are revealed for the polymer therapeutics community including the value of a closed loop design-build-test-learn workflow. This is an exciting time as researchers will gain the ability to fully explore the polymer structural landscape and establish quant. structure-property relationships (QSPRs) with biol. significance.
- 12Reis, M.; Gusev, F.; Taylor, N. G.; Chung, S. H.; Verber, M. D.; Lee, Y. Z.; Isayev, O.; Leibfarth, F. A. Machine-Learning-Guided Discovery of F-19 MRI Agents Enabled by Automated Copolymer Synthesis. J. Am. Chem. Soc. 2021, 143 (42), 17677– 17689, DOI: 10.1021/jacs.1c08181There is no corresponding record for this reference.
- 13Tamasi, M. J.; Patel, R. A.; Borca, C. H.; Kosuri, S.; Mugnier, H.; Upadhya, R.; Murthy, N. S.; Webb, M. A.; Gormley, A. J. Machine Learning on a Robotic Platform for the Design of Polymer-Protein Hybrids. Adv. Mater. 2022, 34, 2201809, DOI: 10.1002/adma.20220180913https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XhsFCisLbK&md5=b9c01ddd2dd5430c9f76c5b1ccbfffd2Machine Learning on a Robotic Platform for the Design of Polymer-Protein HybridsTamasi, Matthew J.; Patel, Roshan A.; Borca, Carlos H.; Kosuri, Shashank; Mugnier, Heloise; Upadhya, Rahul; Murthy, N. Sanjeeva; Webb, Michael A.; Gormley, Adam J.Advanced Materials (Weinheim, Germany) (2022), 34 (30), 2201809CODEN: ADVMEW; ISSN:0935-9648. (Wiley-VCH Verlag GmbH & Co. KGaA)Polymer-protein hybrids are intriguing materials that can bolster protein stability in non-native environments, thereby enhancing their utility in diverse medicinal, com., and industrial applications. One stabilization strategy involves designing synthetic random copolymers with compns. attuned to the protein surface, but rational design is complicated by the vast chem. and compn. space. Here, a strategy is reported to design protein-stabilizing copolymers based on active machine learning, facilitated by automated material synthesis and characterization platforms. The versatility and robustness of the approach is demonstrated by the successful identification of copolymers that preserve, or even enhance, the activity of three chem. distinct enzymes following exposure to thermal denaturing conditions. Although systematic screening results in mixed success, active learning appropriately identifies unique and effective copolymer chemistries for the stabilization of each enzyme. Overall, this work broadens the capabilities to design fit-for-purpose synthetic copolymers that promote or otherwise manipulate protein activity, with extensions toward the design of robust polymer-protein hybrid materials.
- 14Knox, S. T.; Parkinson, S. J.; Wilding, C. Y. P.; Bourne, R. A.; Warren, N. J. Autonomous polymer synthesis delivered by multi-objective closed-loop optimization. Polym. Chem. 2022, 13 (11), 1576– 1585, DOI: 10.1039/D2PY00040G14https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XkvFehtb4%253D&md5=9a3189dbbdedc30e4626e025dcc196d7Autonomous polymer synthesis delivered by multi-objective closed-loop optimizationKnox, Stephen T.; Parkinson, Sam J.; Wilding, Clarissa Y. P.; Bourne, Richard A.; Warren, Nicholas J.Polymer Chemistry (2022), 13 (11), 1576-1585CODEN: PCOHC2; ISSN:1759-9962. (Royal Society of Chemistry)Application of artificial intelligence and machine learning for polymer discovery offers an opportunity to meet the drastic need for the next generation high performing and sustainable polymer materials. Here, these technologies were employed within a computationally controlled flow reactor which enabled self-optimization of a range of RAFT polymn. formulations. This allowed for autonomous identification of optimum reaction conditions to afford targeted polymer properties - the first demonstration of closed loop (i.e. user-free) optimization for multiple objectives in polymer synthesis. The synthesis platform comprised a computer-controlled flow reactor, online benchtop NMR and inline gel permeation chromatog. (GPC). The RAFT polymn. of tert-Bu acrylamide (tBuAm), Bu acrylate (BuA) and Me methacrylate (MMA) were optimized using the Thompson sampling efficient multi-objective optimization (TSEMO) algorithm which explored the trade-off between molar mass dispersity (D) and monomer conversion without user interaction. The pressurised computer-controlled flow reactor allowed for polymn. in normally "forbidden" conditions - without degassing and at temps. higher than the normal b.p. of the solvent. Autonomous experimentation included comparison of five different RAFT agents for the polymn. of tBuAm, an investigation into the effects of polymn. inhibition using BuA and intensification of the otherwise slow MMA polymn.
- 15Rubens, M.; Vrijsen, J. H.; Laun, J.; Junkers, T. Precise Polymer Synthesis by Autonomous Self-Optimizing Flow Reactors. Angew. Chem., Int. Ed. 2019, 58 (10), 3183– 3187, DOI: 10.1002/anie.20181038415https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXit1yjsr7N&md5=155743831af59c7bd6ac81545294817dPrecise Polymer Synthesis by Autonomous Self-Optimizing Flow ReactorsRubens, Maarten; Vrijsen, Jeroen H.; Laun, Joachim; Junkers, TanjaAngewandte Chemie, International Edition (2019), 58 (10), 3183-3187CODEN: ACIEF5; ISSN:1433-7851. (Wiley-VCH Verlag GmbH & Co. KGaA)A novel continuous flow system for automated high-throughput screening, autonomous optimization, and enhanced process control of polymns. was developed. The computer-controlled platform comprises a flow reactor coupled to size exclusion chromatog. (SEC). Mol. wt. distributions are measured online and used by a machine-learning algorithm to self-optimize reactions towards a programmed mol. wt. by dynamically varying reaction parameters (i.e. residence time, monomer concn., and control agent/initiator concn.). The autonomous platform allows targeting of mol. wts. in a reproducible manner with unprecedented accuracy (<2.5 % deviation from pre-selected goal) for both thermal and light-induced reactions. For the first time, polymers with predefined mol. wts. can be custom made under optimal reaction conditions in an automated, high-throughput flow synthesis approach with outstanding reproducibility.
- 16Cao, L.; Russo, D.; Felton, K.; Salley, D.; Sharma, A.; Keenan, G.; Mauer, W.; Gao, H.; Cronin, L.; Lapkin, A. A. Optimization of Formulations Using Robotic Experiments Driven by Machine Learning DoE. Cell Reports Physical Science 2021, 2 (1), 100295, DOI: 10.1016/j.xcrp.2020.100295There is no corresponding record for this reference.
- 17Vriza, A.; Chan, H.; Xu, J. Self-Driving Laboratory for Polymer Electronics. Chem. Mater. 2023, 35 (8), 3046– 3056, DOI: 10.1021/acs.chemmater.2c0359317https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXkslajt78%253D&md5=6f3abd3fa1bc1e249c1017761c49ab47Self-Driving Laboratory for Polymer ElectronicsVriza, Aikaterini; Chan, Henry; Xu, JieChemistry of Materials (2023), 35 (8), 3046-3056CODEN: CMATEX; ISSN:0897-4756. (American Chemical Society)A review. Owing to the chem. pluripotency and viscoelastic nature of electronic polymers, polymer electronics has shown unique advances in many emerging applications such as skin-like electronics, large-area printed energy devices and neuromorphic computing devices, but their development period is years-long. Recent advancements in automation, robotics, and learning algorithms have led to a growing no. of self-driving (autonomous) labs. that have begun to revolutionize the development and accelerated discovery of materials. In this perspective, we first introduce the current state of autonomous labs. Then we analyze why it is challenging to conduct polymer electronic research by autonomous lab. and highlight the needs. We further discuss our efforts in building an autonomous lab., namely Polybot, for the automated synthesis and characterization of electronic polymers and their processing and fabrication into electronic devices. Finally, we share our vision in using a self-driving lab. (SDL) for different types of polymer electronics research.
- 18Gongora, A. E.; Xu, B.; Perry, W.; Okoye, C.; Riley, P.; Reyes, K. G.; Morgan, E. F.; Brown, K. A. A Bayesian experimental autonomous researcher for mechanical design. Science Advances 2020, 6 (15), eaaz1708 DOI: 10.1126/sciadv.aaz1708There is no corresponding record for this reference.
- 19Bateni, F.; Sadeghi, S.; Orouji, N.; Bennett, J. A.; Punati, V. S.; Stark, C.; Wang, J.; Rosko, M. C.; Chen, O.; Castellano, F. N.; Reyes, K. G.; Abolhasani, M. Smart Dope: A Self-Driving Fluidic Lab for Accelerated Development of Doped Perovskite Quantum Dots. Adv. Energy Mater. 2024, 14 (1), 2302303, DOI: 10.1002/aenm.202302303There is no corresponding record for this reference.
- 20Lee, J.; Mulay, P.; Tamasi, M. J.; Yeow, J.; Stevens, M. M.; Gormley, A. J. A fully automated platform for photoinitiated RAFT polymerization. Digital Discovery 2023, 2 (1), 219– 233, DOI: 10.1039/D2DD00100D20https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXitFeju7nF&md5=2613a3e3ae1bdf2542c9791a49e6e147A fully automated platform for photoinitiated RAFT polymerizationLee, Jules; Mulay, Prajakatta; Tamasi, Matthew J.; Yeow, Jonathan; Stevens, Molly M.; Gormley, Adam J.Digital Discovery (2023), 2 (1), 219-233CODEN: DDIIAI; ISSN:2635-098X. (Royal Society of Chemistry)Oxygen tolerant polymns. including Photoinduced Electron/Energy Transfer-Reversible Addn.-Fragmentation Chain-Transfer (PET-RAFT) polymn. allow for high-throughput synthesis of diverse polymer architectures on the benchtop in parallel. Recent developments have further increased throughput using liq. handling robotics to automate reagent handling and dispensing into well plates thus enabling the combinatorial synthesis of large polymer libraries. Although liq. handling robotics can enable automated polymer reagent dispensing in well plates, photoinitiation and reaction monitoring require automation to provide a platform that enables the reliable and robust synthesis of various polymer compns. in high-throughput where polymers with desired mol. wts. and low dispersity are obtained. Here, we describe the development of a robotic platform to fully automate PET-RAFT polymns. and provide individual control of reactions performed in well plates. On our platform, reagents are automatically dispensed in well plates, photoinitiated in individual wells with a custom-designed lightbox until the polymns. are complete, and monitored online in real-time by tracking fluorescence intensities on a fluorescence plate reader, with well plate transfers between instruments occurring via a robotic arm. We found that this platform enabled robust parallel polymer synthesis of both acrylate and acrylamide homopolymers and copolymers, with high monomer conversions and low dispersity. The successful polymns. obtained on this platform make it an efficient tool for combinatorial polymer chem. In addn., with the inclusion of machine learning protocols to help navigate the polymer space towards specific properties of interest, this robotic platform can ultimately become a self-driving lab that can dispense, synthesize, and monitor large polymer libraries.
- 21Politi, M.; Baum, F.; Vaddi, K.; Antonio, E.; Vasquez, J.; Bishop, B. P.; Peek, N.; Holmberg, V. C.; Pozzo, L. D. A high-throughput workflow for the synthesis of CdSe nanocrystals using a sonochemical materials acceleration platform. Digital Discovery 2023, 2 (4), 1042– 1057, DOI: 10.1039/D3DD00033HThere is no corresponding record for this reference.
- 22Beaucage, P. A.; Martin, T. B. The Autonomous Formulation Laboratory: An Open Liquid Handling Platform for Formulation Discovery Using X-ray and Neutron Scattering. Chem. Mater. 2023, 35 (3), 846– 852, DOI: 10.1021/acs.chemmater.2c03118There is no corresponding record for this reference.
- 23Rahmanian, F.; Fuchs, S.; Zhang, B.; Fichtner, M.; Stein, H. S. Autonomous millimeter scale high throughput battery research system. Digital Discovery 2024, 3 (5), 883– 895, DOI: 10.1039/D3DD00257HThere is no corresponding record for this reference.
- 24Kusne, A. G.; Yu, H.; Wu, C.; Zhang, H.; Hattrick-Simpers, J.; DeCost, B.; Sarker, S.; Oses, C.; Toher, C.; Curtarolo, S.; Davydov, A. V.; Agarwal, R.; Bendersky, L. A.; Li, M.; Mehta, A.; Takeuchi, I. On-the-fly closed-loop materials discovery via Bayesian active learning. Nat. Commun. 2020, 11 (1), 5966, DOI: 10.1038/s41467-020-19597-w24https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXisVOnsL3N&md5=138fb31700fa99c1962c048fe7829145On-the-fly closed-loop materials discovery via Bayesian active learningKusne, A. Gilad; Yu, Heshan; Wu, Changming; Zhang, Huairuo; Hattrick-Simpers, Jason; DeCost, Brian; Sarker, Suchismita; Oses, Corey; Toher, Cormac; Curtarolo, Stefano; Davydov, Albert V.; Agarwal, Ritesh; Bendersky, Leonid A.; Li, Mo; Mehta, Apurva; Takeuchi, IchiroNature Communications (2020), 11 (1), 5966CODEN: NCAOBW; ISSN:2041-1723. (Nature Research)Active learning-the field of machine learning (ML) dedicated to optimal expt. design-has played a part in science as far back as the 18th century when Laplace used it to guide his discovery of celestial mechanics. In this work, we focus a closed-loop, active learning-driven autonomous system on another major challenge, the discovery of advanced materials against the exceedingly complex synthesis-processes-structure-property landscape. We demonstrate an autonomous materials discovery methodol. for functional inorg. compds. which allow scientists to fail smarter, learn faster, and spend less resources in their studies, while simultaneously improving trust in scientific results and machine learning tools. This robot science enables science-over-the-network, reducing the economic impact of scientists being phys. sepd. from their labs. The real-time closed-loop, autonomous system for materials exploration and optimization (CAMEO) is implemented at the synchrotron beamline to accelerate the interconnected tasks of phase mapping and property optimization, with each cycle taking seconds to minutes. We also demonstrate an embodiment of human-machine interaction, where human-in-the-loop is called to play a contributing role within each cycle. This work has resulted in the discovery of a novel epitaxial nanocomposite phase-change memory material.
- 25Ament, S.; Amsler, M.; Sutherland, D. R.; Chang, M.-C.; Guevarra, D.; Connolly, A. B.; Gregoire, J. M.; Thompson, M. O.; Gomes, C. P.; van Dover, R. B. Autonomous materials synthesis via hierarchical active learning of nonequilibrium phase diagrams. Science Advances 2021, 7 (51), 12, DOI: 10.1126/sciadv.abg4930There is no corresponding record for this reference.
- 26Lo, S.; Baird, S. G.; Schrier, J.; Blaiszik, B.; Carson, N.; Foster, I.; Aguilar-Granda, A.; Kalinin, S. V.; Maruyama, B.; Politi, M.; Tran, H.; Sparks, T. D.; Aspuru-Guzik, A. Review of low-cost self-driving laboratories in chemistry and materials science: the “frugal twin” concept. Digital Discovery 2024, 3 (5), 842– 868, DOI: 10.1039/D3DD00223CThere is no corresponding record for this reference.
- 27Martin, T. B.; Audus, D. J. Emerging Trends in Machine Learning: A Polymer Perspective. ACS Polymers Au 2023, 3 (3), 239– 258, DOI: 10.1021/acspolymersau.2c0005327https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXht1Cgu7o%253D&md5=39e8ea57891e174a0f5957e251f8de61Emerging Trends in Machine Learning: A Polymer PerspectiveMartin, Tyler B.; Audus, Debra J.ACS Polymers Au (2023), 3 (3), 239-258CODEN: APACCD; ISSN:2694-2453. (American Chemical Society)A review. In the last five years, there has been tremendous growth in machine learning and artificial intelligence as applied to polymer science. Here, we highlight the unique challenges presented by polymers and how the field is addressing them. We focus on emerging trends with an emphasis on topics that have received less attention in the review literature. Finally, we provide an outlook for the field and outline important growth areas in machine learning and artificial intelligence for polymer science and discuss important advances from the greater material science community.
- 28Heil, C. M.; Patil, A.; Dhinojwala, A.; Jayaraman, A. Computational Reverse-Engineering Analysis for Scattering Experiments (CREASE) with Machine Learning Enhancement to Determine Structure of Nanoparticle Mixtures and Solutions. ACS central science 2022, 8 (7), 996– 1007, DOI: 10.1021/acscentsci.2c00382There is no corresponding record for this reference.
- 29Treece, B. W.; Kienzle, P. A.; Hoogerheide, D. P.; Majkrzak, C. F.; Losche, M.; Heinrich, F. Optimization of reflectometry experiments using information theory. J. Appl. Crystallogr. 2019, 52 (1), 47– 59, DOI: 10.1107/S1600576718017016There is no corresponding record for this reference.
- 30Durant, J. H.; Wilkins, L.; Cooper, J. F. K. Optimizing experimental design in neutron reflectometry. J. Appl. Crystallogr. 2022, 55 (4), 769– 781, DOI: 10.1107/S160057672200383130https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB38XitlWitr3K&md5=bc22f593ccc372fcb6d40b5c84fa430fOptimizing experimental design in neutron reflectometryDurant, James H.; Wilkins, Lucas; Cooper, Joshaniel F. K.Journal of Applied Crystallography (2022), 55 (4), 769-781CODEN: JACGAR; ISSN:1600-5767. (International Union of Crystallography)Using the Fisher information (FI), the design of neutron reflectometry expts. can be optimized, leading to greater confidence in parameters of interest and better use of exptl. time [Durant, Wilkins, Butler & Cooper (2021). J. Appl. Cryst.54, 1100-1110]. In this work, the FI is utilized in optimizing the design of a wide range of reflectometry expts. Two lipid bilayer systems are investigated to det. the optimal choice of measurement angles and liq. contrasts, in addn. to the ratio of the total counting time that should be spent measuring each condition. The redn. in parameter uncertainties with the addn. of underlayers to these systems is then quantified, using the FI, and validated through the use of expt. simulation and Bayesian sampling methods. For a 'one-shot' measurement of a degrading lipid monolayer, it is shown that the common practice of measuring null-reflecting water is indeed optimal, but that the optimal measurement angle is dependent on the deuteration state of the monolayer. Finally, the framework is used to demonstrate the feasibility of measuring magnetic signals as small as 0.01 μB per atom in layers only 20 Å thick, given the appropriate exptl. design, and that the time to reach a given level of confidence in the small magnetic moment is quantifiable.
- 31Kjær, E. T. S.; Anker, A. S.; Weng, M. N.; Billinge, S. J. L.; Selvan, R.; Jensen, K. M. Ø. DeepStruc: towards structure solution from pair distribution function data using deep generative models. Digital Discovery 2023, 2 (1), 69– 80, DOI: 10.1039/D2DD00086EThere is no corresponding record for this reference.
- 32Walsh, D. J.; Zou, W.; Schneider, L.; Mello, R.; Deagen, M. E.; Mysona, J.; Lin, T.-S.; de Pablo, J. J.; Jensen, K. F.; Audus, D. J.; Olsen, B. D. Community Resource for Innovation in Polymer Technology (CRIPT): A Scalable Polymer Material Data Structure. ACS Central Science 2023, 9 (3), 330– 338, DOI: 10.1021/acscentsci.3c0001132https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3sXjsVels7Y%253D&md5=f524fbe2207619b331eb96f85b233ed3Community Resource for Innovation in Polymer Technology (CRIPT): A Scalable Polymer Material Data StructureWalsh, Dylan J.; Zou, Weizhong; Schneider, Ludwig; Mello, Reid; Deagen, Michael E.; Mysona, Joshua; Lin, Tzyy-Shyang; de Pablo, Juan J.; Jensen, Klavs F.; Audus, Debra J.; Olsen, Bradley D.ACS Central Science (2023), 9 (3), 330-338CODEN: ACSCII; ISSN:2374-7951. (American Chemical Society)Polymeric materials are integral components of nearly every aspect of modern life. However, developing cheminformatic solns. for polymers has been difficult since they are large stochastic mols. with hierarchical structures spanning multiple length scales. Here we present the design for a general material data model that underpins the Community Resource for Innovation in Polymer Technol. (CRIPT) data ecosystem.
- 33Schneider, L.; Walsh, D.; Olsen, B.; de Pablo, J. Generative BigSMILES: an extension for polymer informatics, computer simulations & ML/AI. Digital Discovery 2024, 3 (1), 51– 61, DOI: 10.1039/D3DD00147DThere is no corresponding record for this reference.
- 34Lu, S.; Jayaraman, A. Pair-Variational Autoencoders for Linking and Cross-Reconstruction of Characterization Data from Complementary Structural Characterization Techniques. JACS Au 2023, 3 (9), 2510– 2521, DOI: 10.1021/jacsau.3c00275There is no corresponding record for this reference.
- 35Sutliff, B. P.; Goyal, S.; Martin, T. B.; Beaucage, P. A.; Audus, D. J.; Orski, S. V. Correlating Near-Infrared Spectra to Bulk Properties in Polyolefins. Macromolecules 2024, 57 (5), 2329– 2338, DOI: 10.1021/acs.macromol.3c02290There is no corresponding record for this reference.
- 36Edwards, C. E. R.; Lakkis, K. L.; Luo, Y.; Helgeson, M. E. Coacervate or precipitate? Formation of non-equilibrium microstructures in coacervate emulsions. Soft Matter 2023, 19 (45), 8849– 8862, DOI: 10.1039/D3SM00901GThere is no corresponding record for this reference.
- 37Lee, S.; Bluemle, M. J.; Bates, F. S. Discovery of a Frank-Kasper σ Phase in Sphere-Forming Block Copolymer Melts. Science 2010, 330 (6002), 349– 353, DOI: 10.1126/science.119555237https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXht1Ois7fE&md5=03af9cd3e115135671009f5bf28607d4Discovery of a Frank-Kasper σ Phase in Sphere-Forming Block Copolymer MeltsLee, Sangwoo; Bluemle, Michael J.; Bates, Frank S.Science (Washington, DC, United States) (2010), 330 (6002), 349-353CODEN: SCIEAS; ISSN:0036-8075. (American Association for the Advancement of Science)Sphere-forming block copolymers are known to self-assemble into body-centered cubic crystals near the order-disorder transition temp. Small-angle x-ray scattering and transmission electron microscopy expts. on diblock and tetrablock copolymer melts have revealed an equil. phase characterized by a large tetragonal unit cell contg. 30 microphase-sepd. spheres. This structure, referred to as the sigma (σ) phase by Frank and Kasper more than 50 years ago, nucleates and grows from the body-centered cubic phase similar to its occurrence in metal alloys and is a crystal approximant to dodecagonal quasicrystals. Formation of the σ phase in undiluted linear block copolymers (and certain branched dendrimers) appears to be mediated by macromol. packing frustration, an entropic contribution to the interparticle interactions that control the sphere-packing geometry.
- 38Kim, K.; Arora, A.; Lewis, R. M.; Liu, M.; Li, W.; Shi, A.-C.; Dorfman, K. D.; Bates, F. S. Origins of low-symmetry phases in asymmetric diblock copolymer melts. Proc. Natl. Acad. Sci. U. S. A. 2018, 115 (5), 847– 854, DOI: 10.1073/pnas.171785011538https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhtlKqt78%253D&md5=4eb24dfe1e67ab529452bf05ea9a6af6Origins of low-symmetry phases in asymmetric diblock copolymer meltsKim, Kyungtae; Arora, Akash; Lewis, Ronald M., III; Liu, Meijiao; Li, Weihua; Shi, An-Chang; Dorfman, Kevin D.; Bates, Frank S.Proceedings of the National Academy of Sciences of the United States of America (2018), 115 (5), 847-854CODEN: PNASA6; ISSN:0027-8424. (National Academy of Sciences)Cooling disordered compositionally asym. diblock copolymers leads to the formation of nearly spherical particles, each contg. hundreds of mols., which crystallize upon cooling below the order-disorder transition temp. (TODT). SCF theory (SCFT) reveals that dispersity in the block ds.p. stabilizes various Frank-Kasper phases, including the C14 and C15 Laves phases, which were accessed exptl. in low-molar-mass poly(isoprene)-b-poly(lactide) (PI-PLA) diblock copolymers using thermal processing strategies. Heating and cooling a specimen contg. 15% PLA above and below the TODT from the body-centered cubic (BCC) or C14 states regenerates the same cryst. order established at lower temps. This memory effect is also demonstrated with a specimen contg. 20% PLA, which recrystallizes to either C15 or hexagonally ordered cylinders (HEXC) upon heating and cooling. The process-path-dependent formation of cryst. order shapes the no. of particles per unit vol., n/V, which is retained in the highly structured disordered liq. as revealed by small-angle X-ray scattering (SAXS) expts. We hypothesize that symmetry breaking during crystn. is governed by the particle no. d. imprinted in the liq. during ordering at lower temp., and this metastable liq. is kinetically constrained from equilibrating due to prohibitively large free energy barriers for micelle fusion and fission. Ordering at fixed n/V is enabled by facile chain exchange, which redistributes mass as required to meet the multiple particle sizes and packing assocd. with specific low-symmetry Frank-Kasper phases. This discovery exposes universal concepts related to order and disorder in self-assembled soft materials.
- 39Biswas, A.; Liu, Y.; Creange, N.; Liu, Y.-C.; Jesse, S.; Yang, J.-C.; Kalinin, S. V.; Ziatdinov, M. A.; Vasudevan, R. K. A dynamic Bayesian optimized active recommender system for curiosity-driven partially Human-in-the-loop automated experiments. npj Computational Materials 2024, 10 (1), 29, DOI: 10.1038/s41524-023-01191-5There is no corresponding record for this reference.
- 40U.S. National Science Foundation: Award Abstract # 2152210 Graduate Traineeship on Advances in Materials Science using Machine Learning. https://www.nsf.gov/awardsearch/showAward?AWD_ID=2152210 (accessed 2024-08).There is no corresponding record for this reference.
- 41Adorf, C. S.; Ramasubramani, V.; Anderson, J. A.; Glotzer, S. C. How to Professionally Develop Reusable Scientific Software─And When Not To. Computing in Science & Engineering 2019, 21 (2), 66– 79, DOI: 10.1109/MCSE.2018.2882355There is no corresponding record for this reference.
- 42Zhao, Z.; Chavez, T.; Holman, E. A.; Hao, G.; Green, A.; Krishnan, H.; McReynolds, D.; Pandolfi, R. J.; Roberts, E. J.; Zwart, P. H.; Yanxon, H.; Schwarz, N.; Sankaranarayanan, S.; Kalinin, S. V.; Mehta, A.; Campbell, S. I.; Hexemer, A. MLExchange: A web-based platform enabling exchangeable machine learning workflows for scientific studies. In 2022 4th Annual Workshop on Extreme-scale Experiment-in-the-Loop Computing (XLOOP); IEEE, 2022.There is no corresponding record for this reference.
- 43Hugging Face. https://huggingface.co (accessed 2024-08).There is no corresponding record for this reference.
- 44Pelkie, B. G.; Pozzo, L. D. The laboratory of Babel: highlighting community needs for integrated materials data management. Digital Discovery 2023, 2 (3), 544– 556, DOI: 10.1039/D3DD00022BThere is no corresponding record for this reference.
- 45McDannald, A.; Frontzek, M.; Savici, A. T.; Doucet, M.; Rodriguez, E. E.; Meuse, K.; Opsahl-Ong, J.; Samarov, D.; Takeuchi, I.; Ratcliff, W.; Kusne, A. G. On-the-fly autonomous control of neutron diffraction via physics-informed Bayesian active learning. Applied Physics Reviews 2022, 9 (2), 021408, DOI: 10.1063/5.0082956There is no corresponding record for this reference.
- 46Audus, D. J.; McDannald, A.; DeCost, B. Leveraging Theory for Enhanced Machine Learning. ACS Macro Lett. 2022, 11 (9), 1117– 1122, DOI: 10.1021/acsmacrolett.2c00369There is no corresponding record for this reference.
- 47Ethier, J. G.; Audus, D. J.; Ryan, D. C.; Vaia, R. A. Integrating theory with machine learning for predicting polymer solution phase behavior. Giant 2023, 15, 100171, DOI: 10.1016/j.giant.2023.100171There is no corresponding record for this reference.
- 48Raissi, M.; Perdikaris, P.; Karniadakis, G. E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 2019, 378, 686– 707, DOI: 10.1016/j.jcp.2018.10.045There is no corresponding record for this reference.
- 49Qu, C.; Schneider, B. I.; Kearsley, A. J.; Keyrouz, W.; Allison, T. C. Applying graph neural network models to molecular property prediction using high-quality experimental data. Artificial Intelligence Chemistry 2024, 2 (1), 100050, DOI: 10.1016/j.aichem.2024.100050There is no corresponding record for this reference.
- 50University of Toronto Acceleration Consortium. https://acceleration.utoronto.ca/ (accessed 2024-06).There is no corresponding record for this reference.
- 51CASE: The Community for Autonomous Scientific Experimentation. https://autonomous-discovery.lbl.gov/ (accessed 2024-06).There is no corresponding record for this reference.
- 52Kiefer, K.; Rossa, L.; Wutzler, F.; Ektrom, N.; Pettersson, A.; Faulhaber, E.; Zolliker, M. An introduction to SECoP – the sample environment communication protocol. Journal of Neutron Research 2020, 21, 181– 195, DOI: 10.3233/JNR-190143There is no corresponding record for this reference.
- 53Wierenga, R. P.; Golas, S. M.; Ho, W.; Coley, C. W.; Esvelt, K. M. PyLabRobot: An open-source, hardware-agnostic interface for liquid-handling robots and accessories. Device 2023, 1 (4), 100111, DOI: 10.1016/j.device.2023.100111There is no corresponding record for this reference.