How Small Chemical Biases Can Build Pure Mirror-Image Molecules

The 2026 Nobel Prize in Chemistry has gone to Henri B. Kagan and Kenso Soai for work on non-linear effects and autocatalysis in asymmetric organic synthesis. Nobel Prize press release

The decision was announced on 7 October 2026. At the time, Kagan was affiliated with Université Paris-Sud in France and Soai with Tokyo University of Science in Japan.

Kagan published the first result in 1986. He found a way to steer a reaction toward an excess of one mirror-image form over the other. In asymmetric synthesis, that bias is the task. The two forms, called enantiomers, contain the same atoms joined in the same order and differ only as a left hand differs from a right hand.

Before Kagan, chemists expected proportionality. A slightly biased control chemical would give a slightly biased product. Kagan showed the relationship could be non-linear. The enrichment of the product did not track the enrichment of the controlling agent in a straight line. Selectivity could be amplified or suppressed.

Soai extended the idea to self-replication. In a 1995 paper, he described the first reaction with the potential to become homochiral, meaning made of one handedness only. By 2003, he had a reaction that produced only one of two possible mirror images. The product serves as a catalyst for its own formation. A small initial imbalance grows until symmetry breaks.

The broader context here will be familiar to engineers. Linear transfer is easy to picture. Non-linear transfer plus positive feedback is different. The outcome is not a weaker copy of the input. It is a sharper version of it.

In my view, that explains the pairing. Kagan showed that selection need not be proportional. Soai showed that it need not be reapplied from outside at each step. One result complicates the link between chiral influence and chiral product. The other closes the loop and lets the product build up its own handedness, a logic familiar from gain and thresholds in electronics and software.

Looking at what this means for practice, the issue is leverage. Chiral information is costly to create and easy to lose. With a non-linear effect, a partly enriched controller can still yield a highly enriched product. With autocatalysis, that enriched product becomes the controller. Chemists can start with bias and let reaction dynamics finish the work, rather than starting with purity.

Taking the longer view, the timeline from 1986 through 1995 and 2003 to 2026 is part of the story. Basic selectivity effects take time to isolate, repeat, and generalize. The Academy often waits until a mechanism is seen as a principle, not a single example. Non-linear effects and autocatalysis fit that pattern. They are behaviors that can show up across reaction types once chemists know to look.

What this enables over the long term is cleaner control of molecular handedness. Single-enantiomer synthesis can rely less on exhaustive separation and more on design of the reaction network. That is a hopeful direction, closer to programming outcome through dynamics than forcing it through brute-force purification.