Skip to the content
Browse the library

How to read a reaction mechanism

Each curly arrow means one pair of electrons moved: the tail sits where the pair is now, the head where it goes. Nothing else about the drawing is arbitrary, which is why a mechanism can be checked rather than memorised.

What an arrow is claiming

An arrow starts at electrons — a lone pair or a bond — and never at a positive charge or an atom. It ends where those electrons finish: at an atom that is gaining a lone pair, or between two atoms that are gaining a bond.

Two arrows in the same step mean two pairs moved at once, which is what makes a step concerted rather than sequential. A carbon that gains a bond must lose one in the same step, because it cannot hold five.

Checking one rather than believing it

Count the atoms on each side; they must match. Count the charge; it must balance, though a step can differ by one if a proton was transferred by a base too ordinary to have been drawn. Check that no carbon has five bonds at any point.

Those three checks catch most mechanisms that are wrong, including most that look plausible. They are also the checks this platform runs: a pathway that does not conserve atoms, or does not arrive at the recorded product, is not offered at all.

Where this stops being simple

  • A mechanism drawn here is rule-derived: a documented pathway for that reaction class, applied to your structures and checked that it arrives. It is not a measurement of what the electrons did, and where two pathways both arrive, drawing one is not evidence against the other.
  • Arrows are a model. They describe bonding changes well and say nothing about timing, energy, or the fact that a real transition state is a continuum rather than a snapshot.

Everything here is open without an account. If you want to try the idea rather than read it, the mechanisms run on structures you supply and prediction will tell you what it thinks a flask would give, with the evidence it is reasoning from attached.