Exploring the Lasers that Power Quantum Computers

Lasers are essential infrastructure for neutral-atom and other atom-based quantum computers: they cool and position atoms, manipulate their qubit states, and provide a way to read those states. That dependence creates a scaling problem. Quantum-device chips can be small and contain thousands of devices, while the lasers serving them may cost hundreds of thousands of dollars, fill large boxes in a separate room, and connect through a single fiber-optic cable. Shrinking the quantum chip alone will not make a complete system compact or inexpensive if its lasers remain bulky and costly.

In neutral-atom systems, tightly focused laser beams called optical tweezers hold individual atoms, while other lasers can cool them. Doppler cooling slows atoms on average by having them absorb red-detuned light; beams from multiple directions can further reduce their motion, an arrangement Guillermo tentatively calls optical molasses. A tweezer’s intensity gradient creates a potential well that draws an atom toward the beam’s brightest point. Different tasks require different wavelengths: the examples given for rubidium are about 850 nanometers for tweezers and about 780 nanometers for cooling, while state-interface lasers may use violet or more infrared light.

Lasers can also drive transitions between encoded qubit states, including a bit flip or a superposition, and can support operations such as a CNOT gate, where one qubit’s state determines whether another is flipped. For fluorescence readout, resonant light causes an atom in the relevant state to absorb photons and emit light spontaneously; optics and a detector or camera collect that fluorescence. Detecting it indicates one state, while its absence indicates the other. The measurement process itself changes a sufficiently small quantum system and collapses a qubit’s superposition without requiring a conscious observer. Readout can retrieve an algorithm’s result or be used during a computation, but measuring generally risks disrupting entanglement that needs to be preserved.

Brightlight Photonics is working to shrink the titanium sapphire, or Ti:sapphire, laser used as a workhorse in quantum optics. A typical bulk system can be about the size of a large microwave and cost more than $100,000; the laser chip itself can be roughly one millimeter square, about the size of a grain of rice. It is a separate component from the quantum-computing chip, packaged with electrical and optical inputs and outputs into a plug-and-play module only a few centimeters at most. The speaker says the chip-scale laser replaced a $100,000 system in a PhD experiment, and that experience helped prompt Brightlight’s effort to commercialize it.

The design replaces a bulk crystal and mechanically tuned mirrors with integrated Ti:sapphire waveguides, avoiding free-space mirrors and moving parts. Sapphire bonding, thinning, photolithography, and plasma etching form the waveguides, which guide light like miniature optical fibers and can route it across a chip. Coupling light into them is demanding: a conventional fiber core can be hundreds of microns across, whereas a chip waveguide may be around one micron or hundreds of nanometers wide. Integrated photonics can form a laser cavity with a looped waveguide, with tuning provided by heaters and thermo-optic effects or by electro-optic and piezo-optic effects. The bulk Ti:sapphire crystal also generates enough heat to make cooling challenging; air cooling may not suffice, and water cooling or external chillers may be needed.

Public roadmaps for useful quantum computers range from tens of thousands to millions of physical qubits, but that does not imply one laser per qubit, since optical elements can split a high-power laser into many beams. Atom-based systems still need lasers with varied wavelengths, power levels, and noise or stability characteristics. Josh says Brightlight’s Ti:sapphire platform could meet the vast majority, but not necessarily all, of those needs. Chip-scale waveguides concentrate light intensely, so thermal effects or damage can limit their power compared with free-space lasers; system designs may keep heat-producing lasers separate from the atoms. Brightlight is developing wafer-scale dies for batch production, with per-laser costs potentially falling to hundreds or tens of dollars if production volume supports it. The company plans to ship packaged evaluation prototypes to customers and partners later that year, while acknowledging that reliability and scaling work remains; near-term users include research and academic groups, quantum-system integrators, and laser manufacturers through collaborations.