Finding reaction conditions in fewer runs
Could we find the best reaction conditions with fewer experiments?
When setting up a new reaction, we think it is common to vary temperature, concentration and time over many experiments. Using recorded yields, we want to approach good conditions in fewer runs, while taking safety limits into account. We would like to share ways to cut down the trial and error with other labs.
Reaction conditions and yield records
We aim to connect and clean the records, then have agents compute the next conditions and verify each other.
Recommended conditions, within safety limits
| Field | Chemistry |
|---|---|
| Status | Planned |
| Input | Reaction conditions · yield records |
| Output | Recommended reaction conditions |
| Delivery | Joint research |