Preprint

Laser cavity reaches one watt of deep-ultraviolet light near 230 nm

Preprint: Researchers report a high-power continuous-wave source, but say coating lifetime remains a central challenge.

A laboratory laser system has delivered up to 1.0 watt of continuous-wave light near 230 nanometres, a result the researchers describe as the highest reported power from direct second-harmonic generation below 237 nanometres. The result came from a normal-incidence, anti-reflection-coated BBO cavity, one of three designs tested for the deep-ultraviolet stage.

The work appears as arXiv version 1 dated 26 August 2026. The paper also reports application demonstrations, including AlF fluorescence over an approximately 6-gigahertz ultraviolet frequency scan and a cadmium magneto-optical trap containing 2 × 10^5 atoms.

Three routes to the same difficult wavelength

The DUV stage used three BBO cavity variants: spherical focusing through a Brewster-cut crystal, elliptical focusing through a Brewster-cut crystal, and spherical focusing through a crystal with an anti-reflection coating at normal incidence.

The underlying 926-nanometre laser source was a VECSEL. Its external blue-generation stage produced close to 4 watts at 463 nanometres, with cavity efficiency peaking at 94% and external efficiency reaching 88%. The near-4-watt result was obtained near maximum VECSEL operation, while stability measurements used reduced power.

To compare the focusing choices, the researchers used a Boyd-Kleinman factor generalized to elliptical focusing. The calculation provided a common basis for comparing the spherical and elliptical arrangements.

The highest power came with a trade-off

The normal-incidence coated cavity reached 51% cavity efficiency and 44% external efficiency when locked. Those figures were reported for the anti-reflection-coated design alongside its 1.0-watt near-230-nanometre output.

A spherical-focus Brewster-cut cavity reached up to 700 milliwatts of deep-ultraviolet output, with 34% cavity efficiency and 27% external efficiency. Its long-term run used slightly lower operating powers, so the stability record was not a test at the maximum reported output.

The broader UVQuanT programme operated 14 continuous-wave deep-ultraviolet cavities in 12 laser systems at six European laboratories over four years. The long-term comparison drew on sustained output from six cavities grouped by geometry, making it a multi-site operating record rather than a controlled head-to-head trial.

Stability depended on more than the crystal

In a continuous 70-hour run of the spherical Brewster-cut cavity, circulating blue power stayed constant while delivered ultraviolet power declined slowly. Inspection found localized defects on the cavity output window and an external dichroic mirror, so the fall in delivered power could not simply be assigned to the BBO crystal.

The study also tested a way to slow damage in ultraviolet windows. At an intensity of 11.5 watts per square centimetre, standard windows reached a 5% transmission decay after 24 minutes, while ion-beam-etched windows had an extrapolated decay time of 192 minutes, an eightfold improvement. The longer figure was extrapolated, and the decay was not a single-exponential process.

The coated design therefore carries a clear engineering trade-off: it reached the reported power and efficiency marks, but coating lifetime remains unresolved. Hydrocarbon photochemistry on a BBO facet remained plausible in the study, but unconfirmed.

A different focus lowers the local load

At fixed circulating power, elliptical focusing reduced peak fundamental intensity sixfold in the reported comparison. In a beam-quality calculation, the second-moment values were 2.47 and 1.00 for the two axes of the spherical focus, compared with 1.09 and 1.00 for the elliptical version.

That geometry did not deliver the highest efficiency in the scan and lock comparison. Cavity second-harmonic-generation efficiency was 35% in the scan and 24% when locked, while external efficiency was 30% and 18%, respectively. The authors describe power-dependent mode overlap and impedance matching as possible factors, rather than identifying one unique loss mechanism.

The comparison does not isolate elliptical focusing as the sole cause of any stability difference because focusing, temperature and other operating conditions changed together. Nor does it show that one geometry is universally best across crystals, laboratories or applications.

The optical chain was tested in two demonstrations

The 227.5-nanometre laser recorded AlF laser-induced fluorescence over an approximately 6-gigahertz deep-ultraviolet frequency scan. This was a system demonstration, not a comparative test of scientific performance between the cavity designs.

In a second demonstration, a dispenser-loaded cadmium magneto-optical trap contained the reported 2 × 10^5 atoms and was used to validate the complete deep-ultraviolet optical chain under continuous operation. The trap experiment was diagnostic and does not establish performance for every species or experiment requiring light near 230 nanometres.

A capable platform with open engineering questions

The measurements show that several VECSEL-based cavity designs can generate substantial continuous-wave deep-ultraviolet power near 230 nanometres. The coated geometry set the reported power mark, the spherical Brewster-cut geometry combined high output with a 70-hour operating record, and the elliptical design reduced calculated local intensity.

The evidence covers laboratory laser and optical-component measurements, a multi-site operating record, and AlF and cadmium demonstrations. It supports the tested systems and components, but the broader record combines different sources, crystals, sites, operators and applications, while long-term conclusions rely mainly on descriptive output traces and diagnoses of where degradation appeared.

The practical questions left by the tests include how long the anti-reflection coating can withstand high-power continuous-wave ultraviolet exposure, whether moving or dithering the crystal can extend the life of exposed positions, and whether the findings can be reproduced across more crystals, laboratories and operating conditions.

The UVQuanT collaboration was funded by the Horizon Europe Framework Programme (101080164). S. Truppe acknowledged funding from the European Research Council (949119) and the Engineering and Physical Sciences Research Council (UKRI2226).

Paper data and sources

Original title: Robust watt-level continuous-wave deep-ultraviolet lasers near 230 nm
Authors: J. Cai, M. Stoepper, P. Agarwal et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text

Versions and corrections

  1. Published automatically after legal-source, freshness, evidence, and independent-verification gates passed.