An independently grown SBO crystal converts 296.8 nm to 148.4 nm
The paper reports an implementation of continuous-wave vacuum-ultraviolet generation at 148.4 nm. The mechanism is single-pass second-harmonic generation: 296.8 nm light enters a strontium tetraborate (SBO, SrB4O7) crystal, and part of it is converted to half the wavelength, 148.4 nm.
The crystal in this work was grown by the top-seeded solution growth method, not obtained from the same source as earlier SBO results. That matters because the paper frames its contribution as an independent implementation, testing whether a separately grown crystal can support the same conversion.
The crystal measured 2.5 x 4.6 x 7.6 mm3 along the crystallographic a, b, and c axes. Up to 580 mW of 296.8 nm fundamental light was focused into it, propagating approximately along the a axis with polarization nominally aligned with the c axis, and with a 1/e2 intensity waist radius of about 25 micrometers at the crystal center.
Quadratic power dependence under vacuum separates VUV from stray pump light
Second-harmonic generation is a nonlinear process, so its contribution should scale with the square of the fundamental power. The authors used that expectation as a diagnostic. They measured the photomultiplier count rate as a function of incident fundamental power with the detection chamber either evacuated or filled with air.
Under vacuum, the count rate contained a quadratic contribution. When air was admitted, the count rate dropped substantially and what remained was approximately linear in fundamental power. Because air strongly attenuates 148.4 nm radiation while transmitting residual 296.8 nm light over the detection path, the authors interpret the quadratic part as the VUV second harmonic and the surviving linear part as pump-related background.
The detection chain was built to make that distinction credible. Three VUV bandpass filters preceded a solar-blind CsI photomultiplier tube. Each filter had a manufacturer-specified transmission of about 21.4% at 148.4 nm, while its measured transmission at 296.8 nm was about 4 x 10-4. Combined with the photomultiplier response, the authors state this provides roughly twelve orders of magnitude of nominal rejection of 296.8 nm light relative to 148.4 nm light for equal incident optical powers.
The 1.7 nW figure is an estimate, not a directly measured power
The headline number is an estimated VUV output power of approximately 1.7 nW at the crystal exit, for an incident fundamental power of approximately 580 mW. It is important to see how that number was produced, because the paper does not report a direct power measurement at 148.4 nm.
The vacuum count-rate data were fitted with a quadratic polynomial: a quadratic term attributed to second-harmonic generation, a linear term describing residual pump-related background, and a constant term for dark counts and power-independent background. The fitted quadratic coefficient was a = 7.5 counts per second per milliwatt squared. At 580 mW, the corresponding quadratic contribution is about 2.5 x 106 counts per second.
That count rate was converted to power using an estimated overall detection efficiency from a VUV photon emerging from the crystal to a registered count. The authors used a photomultiplier quantum efficiency of about 0.222 at 148.4 nm, an MgF2 window transmission of about 0.90, and individual filter transmissions of about 0.214, while taking photoelectron collection and pulse-discrimination acceptance and optical collection efficiency as unity. This gives an overall detection efficiency of about 1.96 x 10-3, or roughly 6.8 x 10-16 W per count per second. The 1.7 nW estimate follows from those assumptions, so it inherits their uncertainty.
Conversion varies by orders of magnitude across the crystal
The paper maps the SHG signal while translating the crystal perpendicular to the incident beam. The resulting transverse position map shows strong spatial inhomogeneity, with the signal varying by several orders of magnitude across the mapped region. The strongest signals occurred near the upper edge of the map.
For the subsequent parameter scans, the authors selected the position (x, y) = (5.7, 4.5) mm, marked on the map. This spatial sensitivity is consistent with the crystal's irregular domain structure: the conversion depends on where in the crystal the beam interacts, not only on the total input power.
The practical consequence is that a single output number for this kind of crystal is not a fixed property of the material alone. It reflects a chosen interaction region, and the paper's scans show that choosing a favorable region is part of the experiment.
Incidence angle, polarization, and temperature each shift the signal
At the selected position, the authors varied the incidence angle by tilting the crystal mount. The SHG signal showed a nonmonotonic angular dependence, with a pronounced maximum near -3 degrees within the scanned range.
Polarization was varied with a half-wave plate upstream of the SHG chamber. The signal was largest near waveplate angles of 0 degrees and 90 degrees, corresponding to fundamental polarization along the crystal c axis, and decreased to approximately one quarter at other settings.
Crystal temperature was also scanned, over a measured range of roughly 301 K to 306 K, and the normalized SHG signal changed across that range. Together these scans identify favorable settings of incidence angle, polarization, and temperature, and they show that the conversion is sensitive to operating conditions rather than being fixed by the crystal alone.
A separately grown crystal reaches a comparable estimated output
The paper positions its result against earlier single-pass continuous-wave SHG in SBO. In that prior work, the reported output power was 1.3 +0.7/-0.6 nW for an incident fundamental power of 325 mW, and the measurements also revealed strong dependence of conversion performance on beam position, incidence angle, and crystal temperature.
This study used a higher incident fundamental power, approximately 580 mW, and estimates approximately 1.7 nW at the crystal exit. The two numbers are not a like-for-like efficiency comparison, because they come from different crystals, different incident powers, and an estimated rather than directly measured output in this work. What the comparison supports is narrower and still useful: an independently grown SBO crystal, prepared by top-seeded solution growth, can produce continuous-wave 148.4 nm light at an estimated power in the same nanowatt range.
The paper also notes that other routes to continuous-wave generation near 148 nm exist, including resonance-enhanced four-wave mixing in cadmium vapour and cavity-enhanced SHG in periodically poled barium magnesium fluoride, and that crystal-based schemes differ in nonlinear medium, phase-matching strategy, optical implementation, and conversion performance. The specific appeal of random quasi-phase matching in SBO, as described here, is that it exploits the irregular domain structure formed during crystal growth and avoids the need for a deliberately patterned domain structure.
Domain-containing regions show strain and split diffraction peaks
The crystal was characterized structurally before use. During growth, local perturbations, particularly temperature fluctuations, can induce wedge-shaped inclusions along the c-axis growth direction and promote sheet-like nonlinear domains predominantly oriented perpendicular to the crystallographic a axis. Pronounced iridescence was observed around internal cracks in the as-grown crystal when viewed along the a axis.
Two samples were cleaved for comparison: one from the domain-containing region along the c-axis growth direction, used for structural characterization and frequency conversion, and one from the domain-deficient region along the a-axis growth direction. X-ray diffraction patterns for the (0k0) reflections showed the same principal reflections in both samples, indicating the same underlying crystal phase.
The domain-containing sample, however, showed a shift toward lower diffraction angles together with clear peak splitting. The authors describe these features as consistent with local strain and structural variations associated with the domain-containing region. In particular, local tensile strain near domain walls could increase the interplanar spacing relative to the domain interior, which by Bragg's law shifts the corresponding diffraction peak toward lower angle. Variations in lattice spacing or crystallographic orientation among different regions may also contribute to the observed peak splitting.
The useful target is a favorable region, not just a better crystal
For anyone assessing SBO as a route to continuous-wave 148.4 nm sources, the most decision-relevant result in this paper is not the 1.7 nW estimate by itself. It is the combination of that estimate with the spatial and parametric scans: the conversion signal varies by orders of magnitude across the crystal, peaks near a specific incidence angle of about -3 degrees, is largest with fundamental polarization along the c axis, and changes with crystal temperature between roughly 301 K and 306 K.
That pattern sets the practical priority. Improving this approach means finding and holding a favorable interaction region and operating point, and it means treating the crystal's domain structure as part of the source design rather than as an uncontrolled detail. The paper's stated motivation is practical guidance for SBO-based continuous-wave VUV sources for nuclear spectroscopy and clock applications, where narrow-linewidth radiation near 148.4 nm would enable sustained interrogation of the 229Th nuclear resonance and feedback stabilization of the probe frequency.
The boundary is equally clear. The output power is inferred from a fitted quadratic contribution and an estimated detection efficiency, not measured directly, and the result comes from one independently grown crystal at one set of operating conditions. The paper's own framing is that characterizing independently grown SBO crystals is what will assess the reproducibility and practical utility of this approach, so the next question is whether other crystals show the same spatial and parametric behavior, not whether this single crystal's 1.7 nW estimate can be generalized.
Sources used in this answer
Continuous-wave 148.4-nm generation in top-seeded solution-grown strontium tetraborate
Single-pass continuous-wave second-harmonic generation of 296.8 nm light in a top-seeded solution-grown strontium tetraborate crystal produced 148.4 nm VUV light, with an estimated output power of approximately 1.7 nW at the crystal exit for approximately 580 mW of incident fundamental power.