A dark spot in laser light helps a trapped ion cool faster
Cornell’s chip-guided beams cooled several kinds of atomic motion together, tackling a delay that can slow trapped-ion computers.
Reading tools for text size, copying, printing and your reading list
Researchers at Cornell have demonstrated a way to cool a trapped atom faster by placing it in a dark spot between laser beams. The experiment addresses an easily overlooked part of quantum computing: before delicate calculations can work reliably, the atoms doing them must stop jiggling so much.[1][2]
Using light guided through a chip, the team cooled three kinds of motion in a single charged calcium atom, or ion, together. A 150-microsecond cooling step brought the targeted motion closer to its lowest-energy state than a 500-microsecond step using the comparison technique. These are cooling times for one atom, not complete-program runtimes.[1]
Cornell highlighted the work on November 12. The peer-reviewed paper appeared in PRX Quantum on November 9; an earlier version was first posted in December 2025 and revised in September 2026. The experiment was already publicly described before this week’s journal publication.[1][2][3]
In a trapped-ion computer, electrically charged atoms hold quantum information. Their motion also matters for controlling interactions between them. Moving ions around a processor, rearranging them or simply operating the apparatus can add unwanted motion, requiring repeated cooling.[1]
The Cornell approach uses a standing wave: an alternating pattern of brighter and darker regions formed by overlapping light beams. At a dark region, called a node, the light intensity is minimal but changes sharply with position. That spatial variation permits interactions that remove motional energy while suppressing unwanted excitation.[1]
Optical pathways built into the trap chip keep the light pattern stable. The researchers combined it with electromagnetically induced transparency cooling, a technique that uses the atom’s response to multiple laser fields to select useful interactions. They compared it with a traveling-wave arrangement, tuning both approaches for the same target motion.[1]
The distinction between cooler and all the way cold matters. After the standing-wave step, the two radial motions had average excitation levels of roughly 0.05 and 0.06 quanta, the discrete units of vibration. Motion along the trap’s axis remained higher, at about 0.8. The team had cooled all three, but not to identical endpoints.[1]
That broader reach is useful alongside the shorter pulse. A method that handles several motions together could reduce the need for separate cooling steps. The experiment does not establish how much time a complete processor would save: that depends on how often its ions need cooling, how they are moved and what other operations occupy its schedule.
There are practical obstacles to carrying the result further. The experiment used one ion, and not every required beam traveled through the chip. Imperfect polarization, residual light at the dark spot and unwanted heating limited performance. At higher light intensities, position drift also constrained the cooling rate.[1]
The next useful tests are whether the method retains those benefits with several ions and during repeated processor operations. Those measurements would connect an improved laboratory step to how much less time a quantum computer could spend getting ready to calculate.