Magnetic Fields Turned Into an Ally for Laser Enrichment of Gadolinium Isotopes

For decades, engineers who enriched gadolinium with lasers treated magnetic fields as an enemy. Stray fields mixed up the magnetic sublevels of atoms in ways that eroded the exquisite polarization-based selectivity on which atomic vapor laser isotope separation, or AVLIS, depends. A new theoretical study from researchers at Lawrence Livermore National Laboratory flips that logic on its head: instead of suppressing magnetic fields, the authors show through detailed density-matrix simulations that a modest, deliberately oriented field of roughly 10 to 20 gauss can substantially boost the photoionization of the even gadolinium isotopes gadolinium-160 and gadolinium-152, roughly doubling the isotopic cut achievable in a single pass compared with zero-field operation.

Enriched gadolinium-160 is the starting material for producing terbium-161, a clinically promising theranostic radionuclide generated in reactors by neutron capture and beta decay. High-purity terbium-161 production has been demonstrated with gadolinium-160 targets enriched beyond 97.5 percent, and the quantities required range from about 10 milligrams for research-scale targets to projected deployments exceeding 10 grams for terabecquerel-scale production. Enriched gadolinium-152, meanwhile, serves as a precursor for gadolinium-153, a short-lived radionuclide with a 241-day half-life used in brachytherapy sources and as an x-ray and gamma-ray calibration source.

Gadolinium presents a peculiar separation challenge. Because it does not form a suitable volatile compound, the industrial workhorses of isotope separation, gas centrifuges and aerodynamic methods, are off the table. That leaves electromagnetic isotope separation and AVLIS as the practical options. Electromagnetic separation can deliver very high isotopic assays, but it is constrained by low material utilization and limited throughput, often yielding only milligram- to gram-scale quantities per year for rare isotopes, with reported source efficiencies of roughly 0.01 to 5 percent. AVLIS, by contrast, combines high isotopic selectivity with the potential for far higher throughput; previous demonstrations at Lawrence Livermore included producing several kilograms of enriched gadolinium-157 and hundreds of kilograms of low-enriched uranium-235.

Most published laser isotope separation work on gadolinium has targeted the odd isotopes gadolinium-155 and gadolinium-157, which possess exceptionally large thermal neutron capture cross sections of 62.2 and 239.8 kilobarns and are valued as burnable neutron poisons in reactors. Those schemes typically use polarization-selective photoionization built on a J = 2 to 2 to 1 to 0 excitation ladder, and in that context magnetic fields have been treated as a nuisance: field-induced mixing between magnetic sublevels degrades polarization selectivity and lets even isotopes leak into the ion stream. Earlier experiments quantified this sensitivity, revealing Hanle-type oscillations and a strong dependence of even-isotope ionization on field-driven precession, and subsequent studies established threshold criteria for preserving odd-isotope selectivity.

The new study reverses the objective. Rather than minimizing magnetic fields to protect odd-isotope enrichment, the researchers explored using a controlled external field as an active design parameter to enhance the photoionization of even isotopes, whose recovery in the standard ladder is capped near 40 percent. Focusing on a previously reported excitation scheme at 575.190, 581.037 and 623.347 nanometers, they examined how magnetic-field orientation and amplitude, laser polarization and laser fluence jointly influence the photoionization efficiency of gadolinium-160 and gadolinium-152.

The modeling is technically detailed. Three resonant laser fields drive the angular-momentum ladder, terminating in an autoionizing state above the ionization threshold, with spontaneous decay and ionization included as incoherent processes in a Lindblad-equation density-matrix framework. Simulations cover 31 velocity bins averaged over a Gaussian velocity distribution at 1500 degrees Celsius to capture Doppler shifts, and the lasers are modeled as transform-limited pulses with a 50-nanosecond full-width-at-half-maximum Gaussian envelope in a counter-propagating geometry that yields nearly Doppler-free excitation between the first two steps. The framework was independently validated against the historical SHUX photoionization model used at Lawrence Livermore during the 1980s and 1990s, agreeing in predicted state populations to within a few percent across test cases spanning polarizations, pulse energies, field orientations and nuclear spins. SHUX itself was historically benchmarked against experimental AVLIS measurements, adding a layer of experimental grounding to the simulations.

The orientation results are striking. At a mere 1 gauss, comparable to Earth’s field of about 0.5 gauss, changing only the laser polarization configuration altered the even-isotope photoionization probability by more than an order of magnitude. Ionization of the spin-zero isotopes is forbidden under all-parallel linear polarization without a magnetic field, and the strongest photoionization occurred when the field was oriented transverse to both the laser propagation direction and the dominant polarization axis, a configuration that maximizes the sigma-plus and sigma-minus coupling components in the atomic frame and accelerates redistribution among magnetic sublevels. When the field aligned parallel to the propagation direction or a polarization axis, that coupling was suppressed and ionization minimized, behavior consistent with the earlier Hanle-type observations but now exploited rather than avoided.

Field amplitude matters too, and the physics has a clean timescale argument. Zeeman splitting introduces Larmor precession, and efficient redistribution among magnetic sublevels requires the Larmor frequency to be comparable to or exceed the inverse of the laser-atom interaction time. For spin-zero gadolinium isotopes in the ground state, that frequency is 3.71 megahertz per gauss, which matched against the 50-nanosecond pulse suggests a minimum field of about 5.4 gauss. The simulations bear this out: at roughly 5 gauss, ionization remains incomplete even at high fluence, but as the field rises from 5 to about 15 to 20 gauss the ionization fraction climbs sharply toward a broad plateau, with diminishing returns beyond roughly 20 gauss and no significant degradation even at 40 gauss. Fluence sweeps showed that certain polarization configurations, notably sigma-plus, sigma-plus and x-linear, achieved both higher peak ionization fractions and lower saturation thresholds, because the field-assisted mixing keeps population from accumulating in dark sublevels.

Translated into separation metrics, the gains are most consequential for throughput rather than purity. With a 15-gauss field oriented along one transverse axis and the favored polarization scheme, theoretical product assays exceeding 90 percent for gadolinium-160 enriched from natural feed are achievable over broad parameter regions when non-selective pickup is a modest 1 percent and the stripping efficiency factor is high, around 0.8; even at 10 percent pickup, substantial enrichment remains possible. Residual off-resonant photoionization of neighboring isotopes stays small, rising from roughly 0.02 percent of laser-generated photoions at the lowest fluences to about 0.5 percent at the highest. Maximum separation factors reached as high as 470 for gadolinium-160 and 610 for gadolinium-152, which constitutes only about 0.2 percent of natural gadolinium. The decisive contrast appears in the isotopic cut, the fraction of target atoms harvested in a single pass: at high stripping efficiency, cuts exceeding 60 to 70 percent are predicted over a wide fluence range, whereas with zero field the achievable assay changes only marginally but the isotopic cut drops by approximately a factor of two.

The authors are careful to frame these as theoretical predictions grounded in a validated simulation framework, with results limited to photoionization dynamics and separation metrics derived from assumed pickup and stripping-efficiency parameters, and the data available from the corresponding author on request. Even so, the conceptual shift is significant. Magnetic fields, long treated solely as a source of selectivity degradation in polarization-selective AVLIS, can instead be leveraged to improve throughput for even-isotope enrichment, particularly for rare isotopes such as gadolinium-152 where product cut is the dominant constraint. Because the gains persist over broad ranges of field amplitude and orientation, the predicted operating regime appears robust enough for large-scale photozones, and the combination of high assay and substantial single-pass cut suggests that AVLIS could become competitive with electromagnetic isotope separation for producing even gadolinium isotopes, a capability that reactor-based analyses of terbium-161 production explicitly identify as critical for sustaining clinical supply.

Subject of Research: Simulation study of magnetic-field-assisted laser isotope separation for gadolinium-152 and gadolinium-160

Article Title: Laser isotope separation of Gd-152 and Gd-160

Article References: Robben, K. C., Rotter, M. D., Haynam, C. A., Johnson, M. A., Carson, C. F., & Isselhardt, B. H. (2026). Laser isotope separation of Gd-152 and Gd-160. Applied Physics B, 132(9), Article 100. https://doi.org/10.1007/s00340-026-08712-2

Image Credits: AI Generated

DOI: 10.1007/s00340-026-08712-2

Keywords: laser isotope separation, gadolinium, AVLIS, magnetic field, photoionization, density-matrix simulation, gadolinium-160, gadolinium-152, isotopic cut, medical isotopes, Zeeman splitting, enrichment