The Soai reaction: how a tiny excess becomes one hand, and why life is homochiral

Kenso Soai found a reaction whose product copies its own handedness, turning a millionth of a percent excess into almost pure single-handed molecules. How it works, the triggers that tip it, and what it says about the origin of life's one-handedness.

Kagan showed in 1986 that pairs of chiral molecules can magnify a catalyst’s purity (part 2). Kenso Soai of the Tokyo University of Science went one step further: he found a reaction in which the product is its own catalyst, so each new molecule helps make more molecules of the same hand. Batch after batch, an almost undetectable excess grows until nearly every molecule has the same handedness. This last part of the series explains how the Soai reaction works, what can tip it one way or the other, and why it changed how scientists think about the origin of life’s one-handedness. You can run the reaction in our chiral amplification lab and check the statistics with the Normal lab.

In short

  • In the Soai reaction a zinc compound adds to an aldehyde to make a chiral alcohol, and that alcohol catalyses its own formation, favouring its own hand.
  • Because mixed pairs of the two hands are idle (Kagan’s idea again), the major hand wins faster and faster. An excess of about 0.00005 % was amplified to more than 99.5 % in three batches.
  • With no chiral help at all, the reaction still ends up almost pure, but which hand wins changes from run to run: true symmetry breaking.
  • Life’s own one-handedness needs some process like this. Soai’s reaction is the first real chemical proof that such processes exist.

A 1953 idea waiting for a reaction

In 1953 the physicist Charles Frank (best known for his work on crystals) asked how a world of molecules made in equal mirror-image amounts could end up one-handed, as life is. He wrote down a minimal recipe with two ingredients:

  1. Autocatalysis: each hand of a molecule speeds up the making of more of its own hand.
  2. Mutual antagonism: the two hands somehow cancel or block each other.

On paper, any small chance excess then runs away until one hand takes over, like two rival populations competing for one food supply. But for forty years nobody had a real reaction that did this. Most chemists thought it was a mathematical curiosity.

1995: the reaction that copies itself

Soai worked on adding zinc compounds (dialkylzincs) to aldehydes, the same chemistry in which Noyori had found strong amplification with an amino alcohol catalyst (part 2). In 1990 his group showed that a chiral alcohol made this way could act as the catalyst for its own formation, though with low selectivity. By 1995 they had found the right molecule: a pyrimidyl alkanol, made by adding diisopropylzinc to pyrimidine-5-carbaldehyde. Its zinc alkoxide catalyses the same addition and favours its own hand, strongly. Starting from a product with a small excess, each new batch, seeded with some of the last one, came out purer. The Nature paper (Soai, Shibata, Morioka and Choji, 1995) was titled “Asymmetric autocatalysis and amplification of enantiomeric excess of a chiral molecule”.

With an improved version (an alkynyl group on the pyrimidine), the group later took an excess of about 0.00005 % (five parts in ten million more of one hand) to above 99.5 % in just three batches. Run it yourself:

Interactive simulation: the Soai reaction batch by batch (needs JavaScript).

Why does it amplify? Pairs again

Kinetic studies, notably by Donna Blackmond and co-workers in the early 2000s, showed that the working catalyst is not a single product molecule but a dimer of two zinc alkoxides (and possibly bigger clusters). That brings in Kagan’s picture: RR dimers make R, SS dimers make S, and the mixed RS dimer is inactive and relatively stable. So the minor hand is tied up in idle mixed pairs, and the active catalyst is purer than the product as a whole: Frank’s “mutual antagonism”, provided by Kagan’s reservoir. Because the rate of making R goes roughly as the amount of RR dimers, which grows faster than the amount of R, even a chance pairing (K = 4 in our model) amplifies slowly; a favoured, idle mixed pair (large K) amplifies fast. Our lab‘s autocatalysis mode is this dimer model in its simplest form: a teaching model, not a full description of the real system, which has bigger aggregates too.

What tips the balance

If a reaction amplifies tiny excesses, then any tiny chiral influence at the start should decide the final hand. Soai’s group spent the next decades testing this, and the list of triggers is astonishing:

  • Circularly polarised light, which is itself handed, used to make the first trace of product.
  • Quartz crystals (1999): quartz grows in left- and right-handed crystals. Adding powdered d-quartz gave one hand of product, l-quartz the other.
  • Chiral crystals of otherwise achiral compounds, such as sodium chlorate.
  • Isotopes (2009, Science): molecules whose only chirality comes from swapping one carbon-12 atom for carbon-13, a difference of one neutron, were enough to choose the product’s hand. Later work did the same with hydrogen, nitrogen and oxygen isotopes.

Each trigger shows how sensitive the reaction is. But the most striking experiment used no trigger at all.

2003: breaking the symmetry by chance

In 2003 Soai’s group ran the reaction without adding anything chiral. Statistically, a supposedly 50:50 mixture of N molecules is never exactly 50:50: the excess fluctuates by about 1/√N. For a laboratory sample of around 10²⁰ molecules that is about one part in ten billion. The autocatalysis amplified whatever excess happened to be there. Each run ended up strongly enriched in one hand, but which hand varied from run to run, roughly half and half: exactly what symmetry breaking by chance should look like. The Royal Swedish Academy of Sciences, announcing the prize, highlights that apart from life itself, nobody had made only one of the two mirror images from nothing before. Watch it in a row of flasks:

Interactive simulation: absolute asymmetric synthesis in many flasks (needs JavaScript).

So why is life one-handed?

Nobody knows yet, but Soai’s work changed the question. There are two parts to any answer:

  1. A first small excess. Candidates include circularly polarised light in star-forming clouds (which is observed, for example in the Orion nebula), the tiny energy difference between mirror-image molecules caused by the weak nuclear force (real, but extremely small), selection on chiral mineral surfaces such as quartz, magnetic surfaces acting on electron spin, or just chance. Some meteorites, such as the Murchison meteorite that fell in Australia in 1969, carry small excesses of L-amino acids of a few percent, so some excess existed before life.
  2. Amplification. Whatever the first excess, it had to be magnified to near 100 %. The Soai reaction is the first proof that a chemical reaction can do that, from a trace to almost pure, under simple conditions. It is not itself a prebiotic reaction (it needs a reactive zinc compound in a dry solvent), but it shows the principle is real. Other amplifiers have been found since, such as “Viedma ripening” (2005), in which grinding a slurry of crystals that interconvert between hands drives them all to one hand, and Kagan-type pairing effects in sublimation and crystallisation of amino acids.

Checked with the tools

1. Three batches, as Soai reported. In the chiral amplification lab‘s autocatalysis mode (5 % of each batch kept as seed, mixed pairs idle), K = 50 takes 0.00005 % → 1.3 % → 89 % → 99.5 %: the same three-batch climb. With K = 4 (pairs by chance) the excess grows only slowly. The model has one free knob, K; we did not fit it to Soai’s numbers, we just looked for a round value that behaves like the experiment.

Batch after batch: 0.00005 % ee grows to almost one handBatch after batch: 0.00005 % ee grows to almost one hand0.00010.0010.010.111010001234Batchee (%, log scale)99.5 %K = 4K = 20K = 50K = 400fibub.com
Chiral amplification lab, autocatalysis mode: each batch seeded with 5 % of the previous product, mixed pairs idle. With K = 50 the model follows the three-batch climb Soai's group reported (0.00005 % to above 99.5 %). A teaching model, not a fit to their data.

2. Inside one batch: an S-shaped rise. Starting from 1 % ee, the excess climbs slowly at first and then quickly as the catalyst purifies itself, ending near 88 %. This S-shape is the fingerprint of autocatalysis, the same shape as logistic growth in the Population growth models tool: growth that feeds on itself until the food (here the starting aldehyde) runs out.

Inside one batch: the excess grows as the reaction runsInside one batch: the excess grows as the reaction runs020406080020406080100Conversion of the starting material (%)ee of all product so far (%)fibub.com
Chiral amplification lab model, one batch from a 1 % excess with a 5 % seed. The S-shaped rise is the signature of autocatalysis: the purer the catalyst gets, the faster it makes its own hand.

3. How big is a chance excess? For N molecules of random hand, the excess has a standard deviation of 1/√N and stays below 1.96/√N in 95 % of samples (the normal approximation to the binomial, as in the Normal lab). For 10⁴ molecules that is about 1 %; for 10²⁰ about 10⁻⁸ %. The Soai reaction can amplify excesses at least this small, which is why its 2003 runs could choose a hand by chance.

The excess that chance alone givesThe excess that chance alone gives10^-810^-610^-410^-210^010^210^610^1010^1410^1810^22Number of molecules N (log scale)Chance excess (%, log scale)Typical (1 standard deviation)95 % of samples stay belowfibub.com
Binomial statistics (standard deviation of the excess = 1/√N; 95 % point from the normal quantile 1.96, stats maths behind the Normal distribution lab). Our own calculation.
Reported Our model or calculation
Amplification in three batches about 0.00005 % to above 99.5 % K = 50, 5 % seed: 0.00005 % → 1.3 % → 89 % → 99.5 %
Runs with no chiral additive each run strongly enriched, hand varies run to run 24 simulated flasks: all end almost pure, R and S about equally often
Chance excess in 10²⁰ molecules about 10⁻⁸ % (1 standard deviation)

How we made the graphs

scripts/blog-figures/nobel-chemistry-2026.mjs calls soaiRounds() and soaiBatch() from src/lib/sci/chiralamp.ts (the Chiral amplification lab’s autocatalysis model: product molecules pair into RR, SS and RS with equilibrium constant K; RR makes R and SS makes S; RS is idle) and normInv() from src/lib/sci/stats.ts for the 95 % point. Each batch adds fresh starting material to a 5 % seed of the last product and runs, in small steps, until the starting material is used up. The “reported” numbers are those quoted in Soai’s papers and in the prize announcement; the model is ours.

How we test the maths

tests/sci.mjs checks the autocatalysis model on every change: idle mixed pairs (K = 50) must amplify more than chance pairing (K = 4); K = 50 with a 5 % seed must stay below 5 % after the first batch, pass 80 % after the second and 99.4 % after the third. The pairing shares are checked against the equilibrium they must satisfy (K = z²/(xy)), and against hand-worked cases such as RR : SS : RS = 1 : 1 : 2 for a racemic mixture at K = 4.

Check it yourself

  1. Open the chiral amplification lab and choose Autocatalysis (Soai).
  2. Keep the defaults: Starting ee (%) 0.00005, Seed share of each batch (%) 5, Mixed-pair constant K 50, Batches 4. The table should show about 1.3 %, 89 %, 99.5 % and then above 99.9 %, and Batches to reach 99 % = 3.
  3. Set K to 4. The excess still grows, but slowly: chance pairing is not enough for fast amplification.
  4. Set K back to 50 and the seed to 50 %. More of the old product dilutes each batch, so the climb is slower: this is why Soai used small seeds.
  5. Open the Normal lab and find the value that 95 % of a standard normal stays below on both sides (1.96). Divide by √N for N = 10,000: about 0.0196, or 2 %, the 95 % limit on a chance excess among ten thousand molecules.

What you can use it for

  • Teaching kinetics: autocatalysis, S-shaped curves and symmetry breaking in one model, with the chiral amplification lab and the logistic curves of Population growth models.
  • Origin-of-life discussions: estimate how big a chance excess is for a given number of molecules, and how much amplification is needed.
  • Research planning: see how seed size and pairing strength change how many batches an amplification needs.
  • The other 2026 science Nobel series: neutrinos from deep space, the Physics series, and brain cells switched with light, the Medicine series.

Sources

  • Nobel Prize in Chemistry 2026, announced 7 October 2026: Henri B. Kagan and Kenso Soai, 'for the discovery of non-linear effects and autocatalysis in asymmetric organic synthesis'. NobelPrize.org and the Royal Swedish Academy of Sciences (kva.se), which dates Soai's key papers to 1995 and 2003; ASBMB Today (2026).
  • Frank FC (1953), 'On spontaneous asymmetric synthesis', Biochimica et Biophysica Acta 11: 459-463.
  • Soai K, Shibata T, Morioka H, Choji K (1995), 'Asymmetric autocatalysis and amplification of enantiomeric excess of a chiral molecule', Nature 378: 767-768.
  • Soai K, Osanai S, Kadowaki K, Yonekubo S, Shibata T, Sato I (1999), 'd- and l-Quartz-promoted highly enantioselective synthesis of a chiral organic compound', J. Am. Chem. Soc. 121: 11235-11236.
  • Soai K et al. (2003), 'Asymmetric synthesis of pyrimidyl alkanol without adding chiral substances by the addition of diisopropylzinc to pyrimidine-5-carbaldehyde in conjunction with asymmetric autocatalysis', Tetrahedron: Asymmetry 14: 185-188.
  • Kawasaki T, Matsumura Y, Tsutsumi T, Suzuki K, Ito M, Soai K (2009), 'Asymmetric autocatalysis triggered by carbon isotope (13C/12C) chirality', Science 324: 492-495.
  • Blackmond DG (2004), 'Asymmetric autocatalysis and its implications for the origin of homochirality', PNAS 101: 5732-5736.
  • Viedma C (2005), 'Chiral symmetry breaking during crystallization: complete chiral purity induced by nonlinear autocatalysis and recycling', Phys. Rev. Lett. 94: 065504.
  • Cronin JR, Pizzarello S (1997), 'Enantiomeric excesses in meteoritic amino acids', Science 275: 951-955.

If you are an author or publisher and want a credit or a passage changed or removed, tell us and we will act quickly.

Tools in this article

Read next

  • Kagan's non-linear effect: when an impure catalyst gives a pure product Nobel series ·

    In 1986 Henri Kagan showed that a chiral catalyst's purity need not pass straight to its product. How the effect works, the models behind it, why chemists use it to read reaction mechanisms, and how to fit it to your own data.

  • Mirror-image molecules and the race to make just one Nobel series ·

    Why molecules come in left- and right-handed forms, why life and medicines care which one you have, and how chemists learned to make just one. The background to the 2026 Chemistry Nobel, with animations and checks you can run.

  • Channelrhodopsin: the protein that is both a light sensor and a gate Nobel series ·

    How Georg Nagel, Peter Hegemann and Ernst Bamberg put two algal genes into frog eggs and human kidney cells and found a channel opened by light: channelrhodopsin-1 in 2002 and channelrhodopsin-2 in 2003. The evidence, the physics of the photon and the pore, rebuilt with fibub tools.

  • How IceCube was built and found neutrinos from deep space Nobel series ·

    5,160 light sensors frozen into the South Pole ice, a flood of background a million times larger than the signal, and the 2013 discovery of neutrinos from beyond the solar system. Play with an event display and check the statistics yourself.

  • How nerve cells signal, and the green alga that taught them to see light Nobel series ·

    The electric language of nerve cells, why scientists wanted a way to switch single kinds of brain cell on and off, and the decades-long hunt for the light sensor of a tiny green alga. The roots of the 2026 Nobel Prize in Physiology or Medicine.