Categories: world

Machine-learning research locks up the potential energy savings potential of the molecular carrier

Credit: Oregon State UniversityNanoscopic cages can play a major role in reducing energy consumption in science and industry, and machine learning research at Oregon State University aims to accelerate the use of these remarkable molecules. The porous organic cage molecules studied at OSU can selectively capture gas molecules, which potentially enable large energy savings in the numerous gas separations performed in the chemical sector. "These porous molecular solids are like fungi that aspirate gases discriminatory," said Cory Simon, Associate Professor of Chemical Engineering and the corresponding author of a study published in ACS Central Science . Together, separation and purification of chemical mixtures is responsible for more than 1 0 percent of the world's energy consumption. Porous cage molecules have nanosed cavities that are built into their structure and gas molecules are attracted to and captured in these cavities via adsorption. "But each cage adsorbs some gases easier than others and this property potentially makes the cages useful for separating gas mixtures more energy efficiently," says Simon. There are, however, tho countless of these cage molecules that can be synthesized to make and with one of them and testing its properties takes months in the lab and hundreds of different chemical separations are required in industry, hence the need for a calculation opportunity to sort through the possibilities and find the best molecule for the current work. Simon utilized the idea that the form of [19659005] Simon and students Arni Sturluson, Melanie Huynh and Arthur York employed an "unanswered" machine…



Credit: Oregon State University

Nanoscopic cages can play a major role in reducing energy consumption in science and industry, and machine learning research at Oregon State University aims to accelerate the use of these remarkable molecules.

The porous organic cage molecules studied at OSU can selectively capture gas molecules, which potentially enable large energy savings in the numerous gas separations performed in the chemical sector.

“These porous molecular solids are like fungi that aspirate gases discriminatory,” said Cory Simon, Associate Professor of Chemical Engineering and the corresponding author of a study published in ACS Central Science .

Together, separation and purification of chemical mixtures is responsible for more than 1

0 percent of the world’s energy consumption.

Porous cage molecules have nanosed cavities that are built into their structure and gas molecules are attracted to and captured in these cavities via adsorption.

“But each cage adsorbs some gases easier than others and this property potentially makes the cages useful for separating gas mixtures more energy efficiently,” says Simon.

There are, however, tho countless of these cage molecules that can be synthesized to make and with one of them and testing its properties takes months in the lab and hundreds of different chemical separations are required in industry, hence the need for a calculation opportunity to sort through the possibilities and find the best molecule for the current work.

Simon utilized the idea that the form of [19659005] Simon and students Arni Sturluson, Melanie Huynh and Arthur York employed an “unanswered” machine learning method to categorize and group burm molecules based on their hole shapes and thus adsorption properties. means that the computer did the learning about form / property relationships on their own;

“Only show data to the algorithm, and it automatically finds the pattern structure in data,” says Simon.

The researchers used an exercise data set of 74 experimentally synthesized porous organic cage molecules each calculated for calculation, resulting in a 3-D “porosity” image of each and the same as an image generated by a CT scan.

“Based on these 3-D images, we took inspiration from a facial recognition algorithm, self-fitting, to group cages with similarly shaped voids,” he said. “By using the singular value resolution, we encoded the 3-D images on the cages in lower dimensional vectors.”

Simon explains the process of analogy to people’s faces.

“Imagine having to map everyone’s face at one point in a two-dimensional scatter plot while keeping as much information as possible about faces,” he says. “So each face is described with just two numbers, and the same faces are grouped near the scatter plot. In essence, singular value decomposition performed this coding, but for porous cage molecules. “

The research showed that Learning coding captures the main properties of the cavities of porous cages and can predict properties of cages related to the cavity.

” Our methods can applied to learn latent performances of voids in other classes of porous materials and forms of molecules in general, “said Simon.


Explore further:
New technology developed to separate complex molecular mixtures

More information:
Arni Sturluson et al., Egencages: Learning a latent space of porous cage molecules, ACS Central Science (2018). DOI: 10,1021 / acscentsci.8b00638

Journal Reference:
ACS Central Science

Provided by:
Oregon State University

Share
Published by
Faela