Using genetic algorithms to systematically improve the synthesis conditions of Al-PMOF

Abstract

The synthesis of metal-organic frameworks (MOFs) is often complex and the desired structure is not always obtained. In this work, we report a methodology that uses a joint machine learning and experimental approach to optimize the synthesis conditions of Al-PMOF (Al2(OH)2TCPP) [H2TCPP = meso-tetra(4-carboxyphenyl)porphine], a promising material for carbon capture applications. Al-PMOF was previously synthesized using a hydrothermal reaction, which gave a low throughput yield due to its relatively long reaction time (16 hours). Here, we use a genetic algorithm to carry out a systematic search for the optimal synthesis conditions and a microwave-based high-throughput robotic platform for the syntheses. We show that, in just two generations, we could obtain excellent crystallinity and yield close to 80% in a much shorter reaction time (50 minutes). Moreover, by analyzing the failed and partially successful experiments, we could identify the most important experimental variables that determine the crystallinity and yield.

Leveraging Machine Learning for Metal–Organic Frameworks: A Perspective

“…A genetic algorithm was used by Domingues et el. to systematically search for the optimal synthesis conditions of Al-PMOF, and excellent crystallinity and yield close to 80% were shown in a short reaction time in just two generations . Kitamura et al used a data-driven approach to visually map the previously reported synthesis conditions for anionic lanthanide-based MOFs, revealed the existence of unexplored search spaces, and then synthesized a series of new MOFs .…”

“…to systematically search for the optimal synthesis conditions of Al-PMOF, and excellent crystallinity and yield close to 80% were shown in a short reaction time in just two generations. 78 Kitamura et al used a data-driven approach to visually map the previously reported synthesis conditions for anionic lanthanidebased MOFs, revealed the existence of unexplored search spaces, and then synthesized a series of new MOFs. 79 The importance of abandoned synthesis data was also emphasized by Hu et al and shown to benefit the ML predictions of C 2 H 2 , C 2 H 4 , and CO 2 adsorption in anion-pillared MOFs.…”

Accelerated Discovery of Metal–Organic Frameworks for CO2 Capture by Artificial Intelligence

“…Although Al-PMOF was proven to be exceptional in CO 2 capture and was successfully synthesized, it had low yields (38%–88%) and required a long reaction time (16 h). To tackle this inefficiency problem, synthesis conditions for Al-PMOF were optimized using the parameters of both failed and partially successful experiments …”

“…The good samples (green) have crystallinity scores larger than 6 out of 10 with yields larger than 50%, while the poor and bad samples (yellow/orange and brown) have crystallinity scores less than 6 out of 10 and yields less than 50%. Readapted with permission from ref .…”

“…To tackle this inefficiency problem, synthesis conditions for Al-PMOF were optimized using the parameters of both failed and partially successful experiments. 62 Figure 5 shows the main experimental variables such as the power of the microwave used for synthesis, reaction temperature and time, the concentration of substrates, and the type of organic solvent. According to these variables, 25 different synthetic pathways were identified, and through experiments, two different generations were employed by genetic algorithm to optimize both the crystallinity and the yield, which define the success of a valid and stable MOF synthesis.…”