Pix.PAN: Dual Use in Lunar & Sky Surveys

• Pix.PAN is a dual-use system that serves as a compact charged-particle spectrometer in LunPAN and a pixel-level image-processing pipeline in Pan-STARRS.
• The LunPAN component employs precise silicon-pixel tracking and a Halbach-array magnet to achieve a maximum detectable rigidity of ~10 GV with <1% misidentification rates.
• The Pan-STARRS pipeline features robust detrending, warping, stacking, and differencing methods to deliver artifact-free, calibrated images for wide-field survey science.

Pix.PAN (“Pixel Processing Analysis Node”) refers to two prominent yet contextually distinct instrumentation and algorithmic pipelines in contemporary research: (1) a compact charged-particle magnetic spectrometer central to the LunPAN lunar energetic particle mission (Hulsman et al., 12 Nov 2025), and (2) the pixel-level image-processing pipeline for the Pan-STARRS1 imaging surveys (Waters et al., 2016). The following article details both, clarifying their architectures, technical workflows, and operational roles in their respective domains.

1. Pix.PAN as a Particle Spectrometer in the LunPAN Mission

In the context of space instrumentation, Pix.PAN denotes the primary magnetic spectrometer for the LunPAN (Lunar Particle Analyzer Network) mission. Designed for deployable measurements of high-energy charged particles in low lunar orbit, its core scientific function is to provide precision spectroscopy for Galactic Cosmic Rays (GCR), Solar Energetic Particles (SEP), and lunar albedo (secondary) particles over the 100 MeV–10 GeV range (Hulsman et al., 12 Nov 2025).

Spectrometer Structure

• Magnetic Geometry: Pix.PAN utilizes a cylindrical Halbach-array architecture within a 16 × 14 × 14 cm³ envelope, total mass ≈10 kg. Its two identical permanent magnet sectors (5 cm diameter, ≈1 cm thick) yield a nominal bore field $B = 0.4$ T. Magnet sectors are separated by 3 cm, accommodating the central tracking station.
• Tracking System: Three tracking stations (one upstream, one between, one downstream of the magnets), each with two closely spaced (1 cm gap) silicon-pixel layers, give six precise space-points across instrument length $L \approx 14$ cm.
• Maximum Detectable Rigidity (MDR): MDR for a Z=1 particle in the combined “Mode 2” (six-plane) spectrometer geometry is calculated using the sagittal arc approximation and the relation $R = p/(Ze) = B\rho$; with layer spatial resolution (σ ≈ 3 μm), this yields an MDR ≈10 GV.

Silicon Pixel Sensors

• Sensor Layering: Six layers (300 μm thick, range 200–400 μm tested), each up to 60 × 50 mm², based on quad-assembled Timepix4 ASICs (55 × 55 μm² pixel cells).
• Readout: Timepix4 readout ASICs provide per-pixel timestamping (down to 200 ps), time-over-threshold energy measurement, and per-layer spatial resolution (σ_x,y ≈ 3–5 μm at CERN test beams).
• Pitch Adaptation: Long-thin sensors (13.75 × 1746 μm²) wire-bonded to the fine-pitch Timepix4 matrix using integrated adapters.

Measurement and Identification Algorithms

• Track Fitting: Trajectory clusters are centroided per plane; global track reconstruction uses Kalman filtering or segmented circle (broken-line) fits to extract bending radius (ρ), momentum (via $p = ZeB\rho$), and charge.
• dE/dx-Based Charge ID: Per-layer energy deposition follows a Bethe-Bloch parametrization for heavy-ion discrimination.
• Multiple Scattering Modeling: Highland’s formula sets angular covariance and the practical low-energy momentum resolution floor.
• Particle ID: Joint likelihood (or $\chi^2$) incorporating curvature, energy loss, and cluster morphology; misidentification rate <$1\%$ (Z=1↔Z=2, $p>200$ MeV/$c$), electron/proton separation >$10^3$.

Performance Parameters

Parameter	Value
Energy range	100 MeV – 10 GeV
MDR (Z=1)	~10 GV
$\Delta p/p$ (electrons)	<12% (few MeV–few GeV)
$\Delta p/p$ (protons)	<40% (~100 MeV–few GeV)
Angular resolution	σ$_\theta$ ≈ 50 μrad
Geometrical factor	≈15 cm² sr
Detection efficiency	>90% for 200 MeV–5 GeV protons
Background rejection	misID <1% (Z=1↔Z=2)
Data rate (raw, 100 km LLO)	~35 kB/s
Modes (4/6-plane)	~3 kB/s / ~0.8 kB/s

Calibration and Radiation Hardness

• On-orbit calibration leverages Timepix4 internal test-pulse injection and cosmic-ray muon scans.
• Radiation mitigation: Timepix4’s 65 nm CMOS provides single-event latch-up and SEU resistance; 5 mm Al shielding, periodic annealing, and quadrant redundancy for scattering-induced damage and performance heterogeneity.

2. Pix.PAN as the Pan-STARRS Pixel-Processing Pipeline

In large-scale optical survey astronomy, “Pix.PAN” denotes the image-level reduction and calibration pipeline for the GPC1 imager in the Pan-STARRS1 (PS1) survey (Waters et al., 2016). This pipeline encompasses primary tasks of detrending, warping, stacking, and differencing, forming the backbone for delivery of science-grade images and catalogs in DR1 and DR2.

Processing Stages

• Detrending (“chip” stage): Removes detector signatures (bias, dark current, gain variations, nonlinearity, fringing, CTE, artifacts). Bias/dark correction uses overscan and partially temperature-dependent, multi-coefficient pixel-based models, with $D_{\rm model}(x,y;t,T)$ parameterized in $t$ (exposure) and $T$ (temp).
• Warping (“warp” stage): Reprojects exposures from native (chip) geometry to a common sky-oriented grid (skycell-based, RINGS.V3 tessellation), using locally linear affine transforms per $128 \times 128$ block and preserving surface brightness via Jacobian scaling. Image data are interpolated using a Lanczos3 kernel.
• Stacking (“stack” stage): Flux-conservatively coadds normalized warps, matching PSFs (using multi-component Gaussian ISIS kernels) and applying robust pixelwise outlier rejection (two-component Gaussian mixture if $n\geq 6$, “Olympic” weighted mean otherwise), yielding deep stacks, masks, variance, and metadata maps.
• Difference Imaging: Template and target images are PSF-matched and subtracted in both warp–stack and warp–warp modes, supporting variability/transient detection.

3. Mathematical Models and Algorithms

Key formulas and calibration algorithms are central to both contexts:

Pix.PAN Spectrometer (LunPAN) (Hulsman et al., 12 Nov 2025):

• Rigidity: $R = p/(Ze) = B\rho$
• Sagitta: $S = L^2/(8\rho)$
• Highland formula for multiple scattering: $$\theta_0 = (13.6\,\text{MeV})/(\beta p c) \,z\sqrt{x/X_0}\left[1+0.038\ln(x/X_0)\right]$$
• Bethe–Bloch energy loss: $$-dE/dx=K\,Z^2/\beta^2[\ln(2m_e c^2\beta^2\gamma^2/I^2)-2\beta^2]$$

Pan-STARRS Pixel Pipeline (Waters et al., 2016):

• Dark model: $$D_{\rm model}(x,y;t,T)=a_0(x,y)+a_1(x,y)t+a_2(x,y)Tt+a_3(x,y)T^2t$$
• Flat-field and variance: $\sigma^2(x,y)=I(x,y)/g+{\rm Noisemap}^2(x,y)$
• Warping interpolation: $$I_{\rm warp}(u,v)=\sum_{i,j}I_{\rm chip}(x_i,y_j)\,L_3(x-x_i)\,L_3(y-y_j)$$
• Stacking combination: $I_{\rm stack}={\sum_i W_i I_i}/{\sum_i W_i}$, $\sigma^2_{\rm stack}=1/{\sum_i 1/\sigma_i^2}$

4. Instrument Calibration and Data Integrity

• Pix.PAN (LunPAN): Pedestal/gain calibration via internal test pulses; cosmic-ray surveys validate layer alignment and timing. Radiation effects are mitigated with shielding and annealing.
• Pan-STARRS pipeline: Per-epoch master frames (dark, skyflat, fringe) constructed with sigma-clipping; dynamic mask generation handles a full taxonomy of optical, electronic, and cosmic-ray artifacts. Cross-frame photometric and astrometric calibration leverages dithered science exposures and a database (“ubercal”) approach.

5. Workflow, Data Products, and System Throughput

Pix.PAN (LunPAN):

• Output products include time-tagged particle event lists detailing track parameters (momentum, species, incidence angle), detection efficiencies, and energy spectra for each particle class encountered in the lunar orbital environment.

Pan-STARRS pipeline:

• Process flow comprises registration, chip-level detrending, WCS calibration, skycell warping, multi-exposure stacking, and differencing, with explicit stage boundaries and database-driven parallelization (hundreds of nodes, ~10M CPU-hours for a full PV3 reprocessing).
• Data products include science, variance, mask, stack, and difference images, as well as exposure/time/weight maps per skycell×filter, enabling both static reference and time-domain science at PB scale.

6. Domains of Application and Scientific Significance

Pix.PAN, in both LunPAN spectrometry and Pan-STARRS imaging, addresses critical scientific needs:

• LunPAN-Pix.PAN: Fills existing data gaps in lunar and deep-space radiation environments, enabling advancements in space physics, lunar geology, space weather forecasting, and risk assessment for future lunar exploration (Hulsman et al., 12 Nov 2025).
• Pan-STARRS Pix.PAN pipeline: Provides uniform, artifact-corrected, astrometrically and photometrically calibrated images and deep stacks supporting wide-field survey science, transient and variable object discovery, and archival data mining (Waters et al., 2016).

The dual usage of “Pix.PAN” highlights how modern research couples hardware innovation and large-scale pipelines to generate high-fidelity measurements—whether of high-energy cosmic particles or astronomical photons—for subsequent scientific inference, modeling, and exploration.