(Invited) Using Computational High-Throughput Screening and Machine Learning for the Data Design of High Performance Nanoporous Materials
Bibliographic record
Abstract
Metal organic frameworks (MOFs) are a diverse class of materials built from organic and inorganic building units that can self assemble to form nanoporous frameworks. The hallmark feature of MOFs is the potential to tune the materials for any given application, which arises from the seemingly endless number of combinations of building units one can construct the materials from. To overcome this combinatorial design challenge we have developed a number of in silico screening tools and applied it to the development of new materials for gas separation processes such as post-combustion CO2 capture. First we have constructed a database of millions hypothetical materials using a novel structure spawning algorithm in which MOF structures can be generated from any topological periodic net (~1800 are experimentally known). Using molecular simulation techniques, we have screened the hypothetical materials for their gas adsorption properties, which provides the locations of the guest-host binding sites in each material. From the best performing materials, a similarity analysis was performed on over 100,000 binding sites to determine if the best performing materials share any common features in their binding sites. For the application of post-combustion CO2 capture, we found that two aromatic rings separated by 7.2 Ang provide an ideal, hydrophobic binding pocket that selectively adsorbs CO2. Our experimental collaborators were then able to synthesize 2 new MOFs with the targeted binding pockets and experimentally confirm that CO2 binds in them as predicted. To date these MOFs are amongst the best materials for selective CO2 capture under realistic humid combustion gas streams. Time permitting, we will also present how we have used deep learning models of nanoporous materials to rapidly predict the low-pressure adsorption properties of MOFs. Under these conditions, the chemistry of the pores is important and the geometric features such as the pore size, are not good predictors of the performance of the materials. Using a distance-between-features descriptor that accounts for the local chemistry of the materials, we have been able to develop accurate models that can predict the adsorption properties of nanoporous materials. Models were trained and tested on large, diverse databases of over 330,000 MOFs giving correlation R2 values on the test sets of >0.95, while geometric features provide R2 values of 0.71.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.007 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".