| 研究課題/領域番号 |
21H01925
|
|---|
| 研究機関 | 北海道大学 |
|---|
研究代表者 |
LIST BENJAMIN 北海道大学, 化学反応創成研究拠点, 特任教授 (80899253)
|
|---|
| 研究分担者 |
SIDOROV PAVEL 北海道大学, 化学反応創成研究拠点, 准教授 (30867619)
辻 信弥 北海道大学, 化学反応創成研究拠点, 特任助教 (30873575)
GIMADIEV TIMUR 北海道大学, 化学反応創成研究拠点, 博士研究員 (30874838) [辞退]
長田 裕也 北海道大学, 化学反応創成研究拠点, 特任准教授 (60512762)
|
|---|
| 研究期間 (年度) |
2021-04-01 – 2024-03-31
|
| キーワード | Organocatalysis / Machine learning / Theoretical chemistry / Automated synthesis |
|---|
| 研究実績の概要 |
Based on the original proposal, we focus on the understanding of the behavior of the catalysts and the construction of the framework for the development of the desired alkene activation.
To get better insights into the behavior of the catalysts, especially the effects of the nitrogen substituents of IDPi catalysts, we picked up the Mukaiyama-aldol reaction as an example and performed a computational study using the GRRM program because a peculiar switch of enantioselectivity depending on the nitrogen substituents was observed. This work is reported in the literature (J. Am. Chem. Soc. 2021, 143, 14475-14481).
For the development of a more efficient working protocol, we have developed a semi-automated seamless platform from the screening to machine learning, including a newly-developed molecular fragment descriptor. By applying the developed method to an intramolecular hydroalkoxylation reaction, the selectivities of an array of untested catalysts can be predicted. Some of them were experimentally validated, and it was proven that the prediction works pretty well. More importantly, highly selective catalysts were discovered through this working process. The manuscript is currently being prepared.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
1: 当初の計画以上に進展している
理由
The computational study on the enantio- and diastereoselective Mukaiyama-aldol reactions provided insights into the unexpected effects of the nitrogen substituents; the installed SO2CF2H groups offer intramolecular hydrogen bonding interactions to stabilize the conformations, having a narrower catalyst pocket.
Regarding the new screening method development, by using a synthesis robot, a semi-automatic screening protocol from the screening to an analysis by chiral SFC is established. The data is directly uploaded to the server as a csv file, which can be used for machine learning studies. The newly developed fragment descriptor enabled predicting higher selective catalyst structures from training data comprised of only moderately selective catalysts.
|
| 今後の研究の推進方策 |
Based on the established working protocol, we will restart investigating the intermolecular hydroalkoxylation reaction. Since the target transformation is an intermolecular one, there is another variable in the nucleophile part, which would induce higher complexity in the screening, and therefore this approach would be highly effective compared to the conventional ones.
In parallel to the multi-dimensional screening approach, a computational analysis of the reaction mechanism will be conducted using the GRRM program, which should provide a relationship between the energy profile and the catalyst-substrate complex structures.
|
