PUFFINS

Telescope and spectropolarimeter

The driving science requirement for PUFFINS is to measure a polarization fraction of 0.2% with an error of 0.02% at 2175 Å. The optical design avoids reflections with large angles of incidence prior to the polarimeter optics to minimize errors due to instrumental polarization.

The last dedicated ultraviolet spectropolarimetry observations were made 30 years ago, by WUPPE aboard the Space Shuttle. PUFFINS picks up where it left off.

Image: NASA — the Astro observatory aboard Space Shuttle Columbia, carrying WUPPE

Design choices

The choice of a protective aluminum coating (Al+MgF₂) on all mirrors, a UV-sensitive scintillator coating on the detector, and a high, optimized grating efficiency provide an average throughput of 11.5% for unpolarized light. All the subsystems of PUFFINS have been selected to be at or above Technology Readiness Level 6 (TRL 6). PUFFINS builds on the heritage of WUPPE, using an MgF₂ Wollaston prism and a rotating half-wave MgF₂ retarder.

Spectropolarimeter and guider

The light from the target star is focused on the spectrograph slit, which functions as a field stop for the spectropolarimeter arm. The slit mask is tilted to reflect off-axis stars to a visible-wavelength imaging guide camera, which feeds back to the spacecraft bus at a 1 Hz cadence. The payload-integrated guider delivers 1.5″ (1σ) pointing stability and 2.75″ (1σ) pointing accuracy.

Optomechanical layout of the PUFFINS payload, from the primary mirror through the spectropolarimeter to the detector.
Optomechanical design of PUFFINS payload Credit: Anche et al. 2025, SPIE.

Detector

Science and guide cameras use the same CMOS sensor with 3.76 µm pixels, read noise under 1.5 e⁻ and low dark current, in an integrated sensor-and-computer package with flight heritage. The science sensor carries a scintillator coating that converts incoming UV photons down to visible wavelengths, where quantum efficiency is high — around 40% across the band.

Instrument parameters

The effective area curve for PUFFINS calculated by multiplying the light-collecting area of the M1 (considering the obscuration from the secondary mirror (M2) and the spider structure) with the Al+MgF₂ reflectivity (6 reflecting surfaces), the projected grating efficiency from Horiba J-Y (averaged over both polarizations, ∼ 55–60%), and the detector quantum efficiency with the UV-sensitive scintillator coating. The spectral resolution achievable across all terms in the optical performance budget is shown alongside it, demonstrating sufficient margin between the requirements and the expected instrument performance.

Effective area and spectral resolution plotted across the PUFFINS bandpass, shown against the science requirement.
Effective area and spectral resolution across the PUFFINS bandpass. Credit: Anche et al. 2025, SPIE.
Modelled performance
Instrumental polarization< 0.1%
Linear-to-circular crosstalk0.02%
Average throughput11.5% (unpolarized light)
Mirror reflectivity> 85% (Al + MgF₂)
Grating1300 lines/mm, −1 order, > 60% efficiency
Spatial resolution5.4″ (1σ, total)
Spectral resolution150 at 1800 Å
Technology readinessAll subsystems ≥ TRL 6

Values from the instrument paper, Anche et al. 2025. arXiv:2509.02994

Two panels for a B3 star over a 66-hour exposure: polarization fraction against wavelength on the left, and polarimetric signal-to-noise against wavelength on the right, staying above the science requirement across the band.
Projected p and pSNR performance of PUFFINS for a B3-type star at V = 10.5. Credit: Anche et al. 2025, SPIE.

Preliminary simulated performance

For a representative target — a B3 star at V = 10.5 — PUFFINS reaches the required polarimetric signal-to-noise on a 0.2% signal in about 66 hours of integration. Across the full target list, the survey closes in roughly eight months.