Sub-µm helium microscopy without giving up working distance

A paper from the group has been accepted in Ultramicroscopy, reporting a helium atom beam just 340 nm across at a working distance of 850 µm. That is a sixfold improvement on the previous Cambridge configuration, and it closes most of the gap with the highest-resolution scanning helium microscopes reported anywhere.

Scanning helium microscopy (SHeM) images a surface by rastering it beneath a beam of neutral helium atoms. Because the atoms carry only 10–70 meV and never penetrate past the outer electron density, the technique is entirely non-destructive and works equally well on insulators, delicate biological material and samples that would charge or degrade under an electron beam. The catch has always been the beam itself. In a pinhole instrument, the beam spreads as soon as it leaves the aperture, so a short working distance buys a small spot. Short-working-distance instruments had reached 315–350 nm; large-working-distance designs like the Cambridge SHeM, which gain depth of field, modularity and diffraction capability from the extra space, had been stuck above 1 µm.

The new work applies constrained atom-optical optimisation to the whole incident optics rather than shrinking any one component. The result is a design point at roughly 710 µm working distance where the three contributions to beamwidth — geometric projection of the pinhole, the demagnified source, and diffraction — are all comparable, at 68%, 64% and 36% of the predicted width. The instrument now sits in a genuinely optimised regime rather than one dominated by simple geometric broadening.

Three hardware changes made that design realisable:

• A new sample chamber with an adjustable bellows, allowing the source–pinhole distance to be varied between about 130 mm and 400 mm. It was set to 385 mm, giving an angular source size of 2.18 × 10⁻⁴ rad, close to the optimised target.
• A compact, 3D-printed pinhole-plate carrying a 470 nm pinhole, focused-ion-beam milled into a 50 nm silicon nitride membrane on a custom 1 mm scaffold. Standard 3 mm membranes would have crowded out the detector.
• A much larger detector aperture, raising the collected solid angle from 0.094 sr to 1.44 sr. Smaller pinhole and longer source distance together cost about ×18 in raw signal; the aperture redesign recovers roughly ×8 of that in contrast-to-noise terms, leaving a practical net penalty of about 2.25.

Measured beamwidths track the prediction across working distances from roughly 550 µm to 1100 µm, so the large depth of field that motivates the geometry in the first place survives the upgrade.

The imaging results make the point better than the numbers. Isolated Clostridioides difficile cells, around 1 µm in diameter and casting shadows as small as 500 nm, are resolved individually — not possible with any previous large-working-distance pinhole helium microscope. A Pseudomonas aeruginosa biofilm shows the string-like formaldehyde crystal patterns clearly. Eroded polycrystalline diamond, which previously showed only grooves above 2 µm, now reveals sub-µm structure between the larger facets. And reusable PPE mask fabric, insulating and highly three-dimensional, is resolved deep into the material as well as at its surface.

The paper is candid about what comes next. Further passive optimisation of this architecture will yield little; substantial gains will need a different approach, most likely active focusing with Fresnel zone plates. The immediate follow-up is combining this beam size with the group’s micro-diffraction work, which will require redesigning the detection optics currently tuned for diffuse topographic contrast.

The work was supported by EPSRC grants EP/R008272/1 and EP/R008051/1, with additional support from CoRDE (EP/T00634X/1) and IAA award EP/X525686/1, and was carried out with Ionoptika Ltd., the University of Glasgow and the ISIS Facility.

Read the paper: doi.org/10.1016/j.ultramic.2026.114446.