Two laser filters help load the right barium ions

Choosing atoms from a mixed source

Before trapped ions can store quantum information, researchers must load them into a trap. That becomes harder when the desired isotope is a minority of the available atoms. Isotopes share an element’s chemical identity but have different neutron counts, subtly shifting the frequencies of light they absorb.

Barium-137 and barium-135 are stable isotopes with nuclear spin, giving them internal energy structures useful for encoding quantum information. Yet they constitute only about 11.23% and 6.59% of natural barium, respectively. A loading method must distinguish them from their more abundant neighbors while capturing ions quickly enough to be practical.

In a November 2026 arXiv preprint, Zichun Zhou and colleagues report an experimental approach that combines two isotope-sensitive excitation steps with a final ionizing photon. Their central idea is “relayed selectivity”: the second excitation can reject unwanted isotopes that slipped through the first.

Three photons, two opportunities to discriminate

The researchers heated naturally abundant barium in an oven and directed neutral atoms through overlapping laser beams inside a linear Paul trap, which confines charged particles using electric fields.

First, a laser at 791 nanometers excites a narrow atomic transition. The desired isotope’s resonance is sufficiently separated from competing lines to provide strong initial discrimination. However, atomic motion and laser intensity broaden the response, leaving some overlap with unwanted isotopes.

A second laser, at 554 nanometers, transfers the excited atoms to a higher energy level. Its resonances also differ between isotopes, providing another opportunity for selection. Finally, a 532-nanometer photon removes an electron, producing a positively charged barium ion that the trap can capture.

The final beam was already available in the apparatus for controlling ion states. None of these three loading steps requires ultraviolet light, avoiding one known source of unwanted electrical charging on trap surfaces. The implementation nevertheless requires additional cooling and auxiliary beams to capture and stabilize the newly formed ions.

What the experiment measured

Scanning the first laser’s frequency produced a fitted peak loading rate of 5.2 ± 0.3 barium-137 ions per second under the reported conditions. One example image showed a chain of 51 barium-137 ions loaded in ten seconds.

The second excitation offered two accessible resonances for each target isotope. For barium-137, fitted peak rates were 4.8 ± 0.3 and 8.5 ± 0.4 ions per second. Corresponding barium-135 rates were 2.6 ± 0.2 and 3.4 ± 0.2 ions per second. These are apparatus-specific loading rates, not processing speeds for a quantum computer.

The team identified ions using cooling and repumping light tuned to the selected isotope. Unwanted trapped ions could appear as dark gaps in otherwise fluorescent ion chains. Across the barium-137 spectroscopy scans, all 2,834 counted ions were identified as the target isotope; the barium-135 second-step scans counted 668 target ions. No unwanted dark ions were observed in those crystallized samples.

Predicted purity is not measured purity

To estimate discrimination beyond what those counts could establish, the authors modeled the overlap of isotope-dependent spectral lines. Their calculation retained the isotope populations and intermediate-state contributions selected by the first laser when evaluating the second step.

For barium-137, the modeled unwanted-isotope excitation fraction fell from about 0.0594% after the first step to roughly 0.000068% after both resonant steps: an approximately 870-fold reduction. The latter corresponds to about 0.68 unwanted excitations per million in the modeled response.

That number is not a directly measured trapped-ion contamination rate. The absence of detected contaminants among a few thousand ions cannot establish sub-part-per-million purity. The estimate also depends on the adopted line positions, widths and excitation assumptions. Importantly, this purity benchmark uses the more selective second-step branch, not the branch giving the highest reported loading rate.

A loading improvement, with boundaries

Increasing the final beam’s intensity increased loading approximately linearly over the tested range, without resolved saturation. That supports further optimization, but does not establish unlimited gains at higher power.

The authors also calculate favorable discrimination for radioactive barium-133 from a synthetic source; they do not demonstrate loading that isotope here. Extensions to other elements likewise remain proposals.

The demonstrated advance is narrower and useful: successive resonances can combine fast loading with strong isotope discrimination from an ordinary mixed source. It improves a preparation step for ion experiments, rather than demonstrating a larger quantum processor or better quantum logic.