Description

Classes:

jwst.pfpc.pfpc_step.PFPCStep

Alias:

pfpc

Overview

The pfpc (point-fixed pattern correction) step is an alternative method of spectral extraction that combines spectra derived from individual dithered observations rather than extracting a single spectrum from a dither-combined resampled data product. During this process, corrections are applied to account for known and stable fixed-pattern noise at each dither position caused by flat field uncertainties, sampling artifacts, and (in the case of MIRI MRS) residual fringes.

Programs that observe bright point sources can sometimes obtain higher signal-to-noise ratio spectra using the PFPC approach compared to the regular x1d data products. The PFPC method can be particularly helpful for MIRI MRS, which is affected by large spectral fringes caused by multiple reflections within the MIRI detectors and optical elements (see discussion on JDox).

Wavelength-dependent correction vectors for each supported mode and dither position are derived from observations of stars or asteroids at standard dither positions in each instrument configuration of interest and stored in a PFPC reference file. The derived corrections can be applied to point source observations taken at the same dither positions and configurations, as long as a target acquisition was performed to accurately center the source.

This step is currently available for MIRI MRS exposures only. It is incorporated into the calwebb_spec3 pipeline, after the pixel_replace step is performed. If a correction is successfully applied, the step produces a set of spectra in pfpc format, one per band. These spectra are saved as a side product of the pipeline, and are not propagated through any further processing steps. If no correction can be applied, no pfpc products are produced.

Algorithm

PFPC corrections must be applied to each dither position separately, so they require an extracted spectrum for each exposure. The input for the step is a set of cleaned cal exposures for a single source. Each exposure must be processed with a set of standard pipeline steps, and configured with default recommended parameters for which the PFPC corrections are appropriate.

For each input exposure, the correction process is:

  1. Check that the exposure can be corrected. If any of the following conditions are not met, no further processing is performed:

    • The PFPC reference file must contain a correction matching the observed spectral band and dither pattern.

    • The exposure must be a point source (SRCTYPE = "POINT").

    • The exposure must have been obtained using target acquisition to ensure accurate source positioning.

    It is also recommended that the input exposure has been processed through the calwebb_spec2 pipeline using the same CRDS context as used for the pfpc step to ensure consistent application of the flatfield, photom, fringe, and fixed pattern corrections. If the calibration file names stored in the reference file table do not match the input, a warning is issued, but processing will proceed.

  2. Run the cube_build step to create a resampled spectral cube for each input channel and band.

  3. Run the extract_1d step on each cube to extract a spectrum for each channel and band.

  4. Run the spectral_leak step on the extracted spectra at each dither position, to correct channel 3A spectra for light leakage from channel 1B.

  5. Match a correction vector from the reference file table to the input and interpolate it onto the extracted wavelengths.

  6. Divide the spectral flux and surface brightness and their associated errors by the correction vector.

After all spectra are corrected, they are averaged with sigma clipping across all dither positions to create one final spectrum for each band. Residual fringes are further corrected with the fit_residual_fringes_1d() function. The output spectra are stored in a format identical to the x1d product, with the PFPC-corrected flux and surface brightness respectively stored in FLUX and SURF_BRIGHT columns in a binary table, and the PFPC+residual fringe corrected spectra stored in RF_FLUX and RF_SURF_BRIGHT columns.

References

The PFPC algorithm and reference files are based on work by K. Gordon and D. Law, “JWST MIRI Medium Resolution Spectrometer Point-fixed Pattern Corrections: Cleaner and Higher Signal-to-noise Spectra of Point Sources” (2026, AJ, 172(4), 204).

Step Arguments

The pfpc step has no step-specific arguments.