Tyler O'Neal, Staff Editor CHEMISTRY

Predicting reaction results: Machines learn chemistry

Chemists and computer scientists develop artificial intelligence / Study published in 'Chem'

Everyday life without artificial intelligence is barely conceivable in today's world. Countless applications in areas such as autonomous driving, foreign language translations or medical diagnostics have found their way into our lives. In chemical research, too, great efforts are being made to apply artificial intelligence (AI), also known as machine learning, effectively. These technologies have already been used to predict the properties of individual molecules, making it easier for researchers to select the compound to be produced.

This production, known as synthesis, usually involves considerable effort as there are many possible synthesis routes to producing a target molecule. Since the success of each individual reaction depends on numerous parameters, it is not always possible, even for experienced chemists, to predict whether a reaction will take place - and even less how well it will work. In order to remedy this situation, a team of chemists and computer scientists from the University of Münster (Germany) has joined forces and developed an AI tool which has now been published in the journal "Chem". {module INSIDE STORY}

Background and method:

"A chemical reaction is a highly complex system", explains Frederik Sandfort, PhD student at the Institute of Organic Chemistry and one of the lead authors of the publication. "In contrast to the prediction of properties of individual compounds, a reaction is the interaction of many molecules and thus a multidimensional problem," he adds. Moreover, there are no clearly defined "rules of the game" which, as in the case of modern chess computers, simplify the development of AI models. For this reason, previous approaches to accurately predicting reaction results such as yields or products are mostly based on a previously gained understanding of molecular properties. "The development of such models involves a great deal of effort. Moreover, the majority of them are highly specialized and cannot be transferred to other problems," Frederik Sandfort adds.

The focus of the work presented was therefore on a general applicability of the programme, so that other chemists can easily use it for their own work. To ensure this, the model is based directly on molecular structures. "Every organic compound can be represented as a graph, in principle as an image," explains Marius Kühnemund, another author, from the field of computer science. "On such graphs, simple structural queries - comparable to the question of colours or shapes in photo - can be made in order to capture the so-called chemical environment as accurately as possible."

The combination of many such successive queries results in a so-called molecular fingerprint. These simple number sequences have long been used in chemoinformatics to find structural similarities and are well suited for computer-aided applications. In their approach, the authors use a large number of such fingerprints to represent the chemical structure of each molecule as accurately as possible. "In this way, we have been able to develop a robust system that can be used to predict completely different reaction results," adds Marius Kühnemund, "The same model can be used to predict both yields and stereoselectivities, which is unique".

The authors demonstrated that their programme can be applied easily and allows accurate predictions, especially in combination with modern robotics, by using a data set that was not originally created for machine learning. "This data set contains only relative sales of the starting materials and no exact yields," Frederik Sandfort explains. "For exact yields, calibrations have to be created. However, due to the high effort involved, this is rarely done in reality".

The team will continue to develop their programme further and equip it with new functions in the future. Prof. Frank Glorius is confident: "When it comes to evaluating large amounts of complex data, computers are fundamentally superior to us. However, our goal is not to replace synthetic chemists with machines, but to support them as effectively as possible. Models based on artificial intelligence can significantly change the way we approach chemical syntheses. But we are still at the very beginning."

Arizona materials scientist needs computational chemistry to produce cheaper infrared plastic lenses

Tyler O'Neal, Staff Editor CHEMISTRY

The new material could bring consumers affordable access to consumer-grade infrared detectors in products such as autonomous cars and in-home thermal imaging for security or fire protection

Five years ago, when the University of Arizona materials scientist Jeffrey Pyun presented his first generation of an orange-tinted plastic lens to optical scientist Robert Norwood, he responded, "This isn't the '60s. No one wants orange glasses, man." CAPTION A sample of the polymer material. CREDIT Mikayla Mace{module In-article}

In the years since a team led by Puyn has refined the material and created the next generation of lenses. The plastic, a sulfur-based polymer forged from waste generated by refining fossil fuels, is incredibly useful for lenses, windows and other devices requiring the transmission of infrared light, or IR, which makes heat visible.

"IR imaging technology is already used extensively for military applications such as night vision and heat-seeking missiles," said Pyun, a professor in the Department of Chemistry and Biochemistry who leads the lab that developed the polymer. "But for consumers and the transportation sector, cost limits high-volume production of this technology."

The new lens material could make IR cameras and sensor devices more accessible to consumers, according to Norwood, a professor in the James C. Wyant College of Optical Sciences. Potential consumer applications include economical autonomous vehicles and in-home thermal imaging for security or fire protection.

The new polymers are stronger and more temperature resistant than the first-generation sulfur plastic developed in 2014 that was transparent to mid-IR wavelengths. The new lenses are transparent to a wider spectral window, extending into the long-wave IR, and are far less expensive than the current industry standard of metal-based lenses made of germanium, an expensive, heavy, rare and toxic material.

Because of germanium's many drawbacks, Tristan Kleine, a graduate student in Puyn's lab and first author on the paper, identified a sulfur-based plastic as an attractive alternative. However, the ability to make IR-transparent plastics is a tricky business.

The components that give rise to useful optical properties, such as sulfur-sulfur bonds, also compromise the strength and temperature resistance of the material. Moreover, the inclusion of additional organic molecules to give the material strength resulted in reduced transparency since nearly all organic molecules absorb IR light, Kleine said.

To overcome the challenge, Kleine - in collaboration with chemistry graduate student Meghan Talbot and chemistry and biochemistry professor Dennis Lichtenberger - used computational simulations to design organic molecules that were not IR-absorbing and predicted transparency of candidate materials.

"It could have taken years to test these materials in the laboratory, but we were able to greatly accelerate new materials design using this method," Kleine said.

Germanium requires temperatures greater than 1,700 degrees Fahrenheit to melt and shape, but because of its chemical makeup, the sulfur polymer lenses can be shaped at a much lower temperature.

"A major advantage of these new sulfur-based plastics is the ability to readily process these materials at much lower temperatures than germanium into useful optical elements for cameras or sensors, while still maintaining good thermomechanical properties to prevent cracking or scratches," Pyun said. "This new material has just checked so many boxes we couldn't before."

"Its reliability is essentially equivalent to optical polymers that are routinely used for eyeglasses," Norwood added.

The team is partnering with Tech Launch Arizona to translate the research into a viable technology.

"Humans light up like a Christmas tree in IR," Pyun said. "So, as we think about the Internet of Things and human-machine interfaces, the use of IR sensors is going to be a really important way to detect human behavior and activity."