A runner on the anchor leg of a relay race takes the baton and runs on. If the baton could carry two different starting points inside it, the runner would set off in two directions at once. That is roughly how a new protocol from a quantum optics lab works: a single electron transfers its quantum superposition to a levitating nanoparticle, and the particle begins moving in two opposite directions simultaneously.

A nanoparticle with a radius of 50 nanometers hangs in an optical trap, cooled almost to a complete standstill. An electron prepared in a time-bin state flies toward it: it is in a superposition of two options, arriving a little earlier or a little later. At the moment of collision the electron sticks to the particle, and a short electric pulse lasting 100 nanoseconds gives the particle a kick. The direction of the kick depends on exactly when the electron arrived: earlier means a push to the left, later means a push to the right.

After the kick the particle ends up in a superposition of two coherent states with opposite momenta. This is the "Schrödinger kitten": the dust grain moves left and right at the same time.

Half an oscillation period later (5 microseconds), the particle is measured with a precision of 85 picometers. By repeating the experiment at different phases of the electron state, the interference pattern is reconstructed. At a phase of π the Wigner function becomes negative at the origin of phase space, which is a signature of nonclassicality. For realistic parameters the interference visibility reaches 50%, and when the kick is doubled it drops to 10%. The main enemy is photon recoil from the trapping laser, but negativity persists for roughly 5π/12 of a period, and with a threefold reduction in decoherence it lasts throughout the entire protocol.

Previously, creating macroscopic superpositions required either expanding the wave packet or using nonlinear potentials. Both routes sharply amplified decoherence, the loss of quantum coherence through interaction with the environment. The new approach sidesteps this obstacle: the difficulty is shifted onto the well-developed techniques of electron interferometry. A single electron, whose mass is $10^{12}$ times smaller than the particle's mass, acts as a microscopic agent that transfers a quantum state to a mesoscopic object.

In this scheme the relay baton is the quantum superposition itself. An electron that has not yet decided when to arrive passes that uncertainty on to the particle. If the protocol can be implemented, it will make it possible to create quantum superpositions of objects at least five times more massive than the current record. That would open the door to experimental tests of collapse models and quantum gravity, and would also yield ultrasensitive force sensors.