Blue Laser Fusion beam layouts aim for highly uniform direct drive

Category: Inertial, Lasers, Simulations

View from inside a laser fusion target chamber, with red, blue and yellow light streaks converging on a bright central point and diagnostic hardware around the edges.

Inside the OMEGA target chamber, whose 60-beam geometry is the paper’s benchmark

(Image courtesy of Blue Laser Fusion)

Blue Laser Fusion researchers report that a 60-beam layout built from the rotational symmetry group of the icosahedron gives about an order of magnitude lower calculated nonuniformity than the OMEGA 60-beam pointing geometry when beam profiles are smooth, and that optimised layouts with more beams stay robust against random power imbalance over the range studied. The advantage narrows for flat-top beams, and the paper, published in Nuclear Fusion at the end of October and co-authored by company co-founder Shuji Nakamura, covers illumination geometry only, with hydrodynamic validation still to come.

Why beam geometry governs low-order nonuniformity

Irradiation nonuniformity seeds hydrodynamic instabilities, distorts implosion symmetry and reduces stagnation pressure, and with a finite number of beams, uniform drive becomes a geometry problem related to the Thomson problem of placing points on a sphere. Low-degree asymmetry already limits ignition-regime experiments at the National Ignition Facility, where simulations show low-order flux asymmetries degrading stagnation and yield, and diagnostics and simulations at OMEGA show it can strongly limit cryogenic implosion performance.

Skupsky and Lee separated nonuniformity into a single-beam factor and a geometrical factor set only by beam directions and weights, showing that suitable arrangements suppress selected low-degree modes. Schmitt showed in an idealised setting that absolutely uniform illumination is possible when the single-beam profile is shaped appropriately, McGeoch and Obenschain optimised configurations on that framework with approximate hyper-Gaussian profiles, and Murakami and collaborators treated beam directions as mutually repelling points on a sphere. The paper describes the StarDriver concept as a complementary strategy, cutting global nonuniformity through beamlet multiplication and strong spatial and temporal incoherence, which works against high-degree structure while systematic suppression of low-degree modes needs structural control of the angular beam distribution. In the paper’s framework, symmetry-forbidden degrees cancel for any common axisymmetric beam profile, whereas Schmitt’s construction required a specific profile.

Icosahedral symmetry forbids fifteen harmonic degrees

The paper applies all 60 proper rotations of the icosahedral group to a reference beam and expands the drive pattern in spherical harmonics. The twelve vertices of the icosahedron form a tight spherical 5-design, which removes every degree up to 5, and exact invariance under the full group removes further degrees, leaving 1 to 5, 7 to 9, 11, 13, 14, 17, 19, 23 and 29 with no power. These exclusions hold for delta-function and finite-width axisymmetric beams alike, as long as beams in each complete 60-beam orbit carry equal power. Larger layouts, labelled IcN with N = 60K, start from K generic seed directions expanded by the same rotations, and the seeds were tuned by random perturbation from many starting points, with no claim of global optimality.

Smooth beams open the gap with OMEGA

For a von Mises-Fisher profile at a normalised spot diameter of 0.8 times the target diameter, the optimised 60-beam layout gives an RMS nonuniformity of about 0.041 per cent, against about 0.33 per cent for the OMEGA 60-beam geometry, as annotated in the paper’s Figure 2. OMEGA’s antipodal beam pairs already cancel every odd degree, but even-degree modes around 8 and 10 survive and contribute appreciably to the residual, while the icosahedral layout’s dominant residual components sit near degrees 10 and 12. A flat-top profile leaves substantial high-degree power, so the two layouts perform similarly, and near a normalised spot size of 0.7 the icosahedral layout calculates slightly worse.

The comparison isolates beam-axis geometry. OMEGA’s on-target uniformity also depends on measured focal spots, beam balance, smoothing by spectral dispersion, cross-beam energy transfer and plasma response, none of which the model reproduces.

More beams reduce sensitivity to power imbalance

At equal beam power and a normalised spot size of 0.8, moving from 60 to 120 beams cuts the calculated nonuniformity to roughly 1/256th of its value, and moving from 120 to 240 beams roughly halves it again. The authors relate the equal-power behaviour to the spherical t-design layouts Shvydky and colleagues proposed for direct drive in 2023, then randomise beam powers by between ±1 and ±20 per cent. At that spot size the residual rises roughly in proportion to the imbalance, and it falls roughly as 1/N at low to intermediate beam counts, steepening towards 1/N² above 1,000 beams, faster than the 1/√N expected for randomly placed beams. Imbalance generates modes from degree 1 to 9, with peaks near degrees 2 and 3 that are the main reason nonuniformity rises at large spot sizes, and the paper presents the redundancy of large beam counts as a practical advantage for reactor-scale lasers, where small beam-to-beam power variations are difficult to eliminate completely.

What a thousandfold seed reduction would take

The paper links drive nonuniformity to ablation pressure, which scales with laser intensity to the power 0.71 at 351 nm, so a small fractional intensity perturbation passes into the pressure at about 0.71 times its size. Follett and colleagues found in an OMEGA-based surrogate design model that a tenfold cut in the Rayleigh-Taylor seed can raise target gain by about 50 per cent, that recovering one-dimensional performance would take a thousandfold cut, and that upgrades including raising the drive beam count from 60 to 720 can bring comparable improvements.

Starting from the OMEGA-level nonuniformity of about 2 per cent that the paper cites from Sampat and colleagues, a thousandfold reduction lands near 2 × 10⁻⁵, a level the calculations reach at about 1,920 beams with ±1 per cent power imbalance, provided laser imprint dominates the seed and scales approximately linearly with the irradiation nonuniformity. Target roughness, offset, beam mistiming, stalk perturbations and polarisation effects sit outside the model.

Hydrodynamic validation is the next test

The paper concludes that full icosahedral symmetry, smooth beam profiles and enough beams can reduce irradiation nonuniformity to levels relevant to ignition-class or reactor-class implosions, and presents the work as an illumination-geometry design principle rather than a full assessment of implosion performance. Hydrodynamic validation comes next, along with target imperfections such as out-of-roundness and surface roughness, and the model leaves out intrabeam speckle, mutual coherence between beams and optical-field statistics.

The results extend a 2025 Optics Express study from the same group that presented a direct-drive reactor concept with more than about 500 beams, adding an explicit spectral analysis, a direct comparison with the OMEGA geometry and scaling under power imbalance. Blue Laser Fusion is developing a pulsed laser based on optical enhancement cavities, which it is pursuing for directed-energy defence and for gigawatt-scale fusion energy.

This article is based on: Sunahara A, Raj G, Cohen T, Pattison P M and Nakamura S 2027 Icosahedral-group beam layouts for highly uniform direct-drive laser fusion. Nuclear Fusion 67 016018. DOI: 10.1088/1741-4326/aeb431

Stay ahead in the fusion revolution explore more breakthroughs from leading innovators in clean energy technology.