The transport problem behind continuous neutral-atom reloading

Atom-accel • 3d ago Neutral-atom quantum processors lose atoms over time, so long-running devices need a way to bring in fresh atoms without disturbing stored qubits. One approach, demonstrated recently by Chiu et al. ("Continuous operation of a coherent 3,000-qubit system," arXiv:2506.20660v2), is to cool atoms somewhere else and move them into the processor region with optical-lattice conveyor belts. I have been building a particle-based classical model of one part of that problem: the handoff between two crossed optical-lattice conveyors. The animation shows transfer from an incoming conveyor to an outgoing conveyor. The upper-right panel is a top view of the highlighted zoom box, with the vertical scale exaggerated by 50x. The lower panels show cloud projections and directional temperature components during the handoff. The cloud heats from about 20 uK to about 120 uK during the handoff, roughly a 6x increase. Here "temperature" means directional kinetic temperature inferred from momentum variance: T_i = sigma_{p,i}^2 / (m k_B). The surprising part is that the big visible sloshing does not appear to be the main heating channel. It is mostly from mismatch between the incoming cloud and the equilibrium distribution of the outgoing lattice. The heating seems to come mostly from the tightly confined direction during transfer. This is the point I am trying to show: equipartition is about energy, not spatial amplitude. A weakly confined direction can move a lot without holding most of the kinetic energy, while a tightly confined direction can look almost stationary but still be dynamically hot. Does the animation illustrate that point clearly? fefetornado • 2d ago I don't get your point, or maybe i get it but it is just totally obvious. Changing confinement (changing trap) doesn't change temperature by itself. The particles have the same energy in both cases, they will just bounce differently in a deep and narrow trap compared to a large weak one. The energy (or temperature if you prefer to call it that way) is the same. Usually, what makes the temperature change going from one trap to another is the scattering induced by the trap light. Atoms scatter photons which makes them gain kinetic energy Maybe i didn't really get your point though Atom-accel • 23h ago You’re right that simply changing the confinement of a static trap does not automatically heat the atoms, but that is not quite what is happening here. During the handoff, the potential experienced by the atoms is explicitly changing with time, so their mechanical energy is not conserved. The changing optical potential can do work on the atoms and excite their motion. That added energy does not immediately correspond to a new thermodynamic temperature, though. Without collisions, the cloud can remain anisotropic and out of equilibrium, with different amounts of kinetic energy in different directions. In this simulation I approximate atom–atom collisions using the Direct Simulation Monte Carlo method. Those collisions redistribute the added energy between directions and allow the cloud to rethermalize, which is what produces the final temperature increase. There is also an important distinction between two physically different kinds of optical trapping that often gets blurred together. A magneto-optical trap works through near-resonant photon scattering. Red-detuned, counterpropagating laser beams produce a velocity-dependent force through the Doppler shift, while a magnetic-field gradient produces a position-dependent Zeeman shift. The atoms repeatedly absorb photons from the laser beams and spontaneously emit them in random directions. That process provides both trapping and cooling, but it also produces a lot of near-resonant scattered light. An optical lattice or optical tweezer works differently. The trapping light is far detuned from resonance and produces an AC Stark shift, or equivalently an induced-dipole potential. To a good approximation this is a conservative force: the atoms move in the spatially varying light-shift potential without repeatedly absorbing and spontaneously emitting photons. There is still some off-resonant scattering, so it is not literally zero, but it can be made much smaller than in a MOT. A bare optical-dipole trap does not provide cooling; it simply confines the atoms. With enough laser power, though, optical lattices and tweezers can provide very strong confinement. So the heating in this simulation is not from photon scattering. It comes from the time-dependent handoff between the two lattice potentials. My main point was that the spatial amplitude of the motion is a poor indicator of where that energy is. The weakly confined direction can show large, dramatic-looking sloshing, while substantial energy can be stored in the tightly confined lattice direction as motion that is spatially tiny but very fast. Continue this thread Continue this thread fefetornado • 2d ago I don't get your point, or maybe i get it but it is just totally obvious. Changing confinement (changing trap) doesn't change temperature by itself. The particles have the same energy in both cases, they will just bounce differently in a deep and narrow trap compared to a large weak one. The energy (or temperature if you prefer to call it that way) is the same. Usually, what makes the temperature change going from one trap to another is the scattering induced by the trap light. Atoms scatter photons which makes them gain kinetic energy Maybe i didn't really get your point though Atom-accel • 23h ago You’re right that simply changing the confinement of a static trap does not automatically heat the atoms, but that is not quite what is happening here. During the handoff, the potential experienced by the atoms is explicitly changing with time, so their mechanical energy is not conserved. The changing optical potential can do work on the atoms and excite their motion. That added energy does not immediately correspond to a new thermodynamic temperature, though. Without collisions, the cloud can remain anisotropic and out of equilibrium, with different amounts of kinetic energy in different directions. In this simulation I approximate atom–atom collisions using the Direct Simulation Monte Carlo method. Those collisions redistribute the added energy between directions and allow the cloud to rethermalize, which is what produces the final temperature increase. There is also an important distinction between two physically different kinds of optical trapping that often gets blurred together. A magneto-optical trap works through near-resonant photon scattering. Red-detuned, counterpropagating laser beams produce a velocity-dependent force through the Doppler shift, while a magnetic-field gradient produces a position-dependent Zeeman shift. The atoms repeatedly absorb photons from the laser beams and spontaneously emit them in random directions. That process provides both trapping and cooling, but it also produces a lot of near-resonant scattered light. An optical lattice or optical tweezer works differently. The trapping light is far detuned from resonance and produces an AC Stark shift, or equivalently an induced-dipole potential. To a good approximation this is a conservative force: the atoms move in the spatially varying light-shift potential without repeatedly absorbing and spontaneously emitting photons. There is still some off-resonant scattering, so it is not literally zero, but it can be made much smaller than in a MOT. A bare optical-dipole trap does not provide cooling; it simply confines the atoms. With enough laser power, though, optical lattices and tweezers can provide very strong confinement. So the heating in this simulation is not from photon scattering. It comes from the time-dependent handoff between the two lattice potentials. My main point was that the spatial amplitude of the motion is a poor indicator of where that energy is. The weakly confined direction can show large, dramatic-looking sloshing, while substantial energy can be stored in the tightly confined lattice direction as motion that is spatially tiny but very fast. Continue this thread Atom-accel • 23h ago You’re right that simply changing the confinement of a static trap does not automatically heat the atoms, but that is not quite what is happening here. During the handoff, the potential experienced by the atoms is explicitly changing with time, so their mechanical energy is not conserved. The changing optical potential can do work on the atoms and excite their motion. That added energy does not immediately correspond to a new thermodynamic temperature, though. Without collisions, the cloud can remain anisotropic and out of equilibrium, with different amounts of kinetic energy in different directions. In this simulation I approximate atom–atom collisions using the Direct Simulation Monte Carlo method. Those collisions redistribute the added energy between directions and allow the cloud to rethermalize, which is what produces the final temperature increase. There is also an important distinction between two physically different kinds of optical trapping that often gets blurred together. A magneto-optical trap works through near-resonant photon scattering. Red-detuned, counterpropagating laser beams produce a velocity-dependent force through the Doppler shift, while a magnetic-field gradient produces a position-dependent Zeeman shift. The atoms repeatedly absorb photons from the laser beams and spontaneously emit them in random directions. That process provides both trapping and cooling, but it also produces a lot of near-resonant scattered light. An optical lattice or optical tweezer works differently. The trapping light is far detuned from resonance and produces an AC Stark shift, or equivalently an induced-dipole potential. To a good approximation this is a conservative force: the atoms move in the spatially varying light-shift potential without repeatedly absorbing and spontaneously emitting photons. There is still some off-resonant scattering, so it is not literally zero, but it can be made much smaller than in a MOT. A bare optical-dipole trap does not provide cooling; it simply confines the atoms. With enough laser power, though, optical lattices and tweezers can provide very strong confinement. So the heating in this simulation is not from photon scattering. It comes from the time-dependent handoff between the two lattice potentials. My main point was that the spatial amplitude of the motion is a poor indicator of where that energy is. The weakly confined direction can show large, dramatic-looking sloshing, while substantial energy can be stored in the tightly confined lattice direction as motion that is spatially tiny but very fast. oofos_deletus • 2d ago Can someone explain to me in simple terms what am I looking at? (New to this stuff myself) Sorry, something went wrong when loading this video. View in app 0xB01b • 2d ago Looks good