The James Webb Space Telescope Absolute Flux Calibration. V. Near-Infrared Camera Wide Field Slitless Spectroscopy

TL;DR

This work delivers a complete absolute flux and wavelength calibration for JWST/NIRCam WFSS by building a field-dependent geometric model of spectral traces and a high-order wavelength solution across both modules A and B and grisms R and C. The core methodology combines a two-dimensional polynomial trace geometry P_{n,m}(x,y,t) with a wavelength relation tied to the source position, calibrated against SMP LMC 58 and anchored to NIRSpec reference spectra, while flux calibration uses the standard star P330E with careful background and aperture corrections. Key results include trace predictions accurate to better than 0.1 pixel, wavelength residuals of 0.65–0.91 Å for +1 order and ~0.5 Å for +2 order, and an absolute flux accuracy of 3% for +1 order across all configurations. The calibration framework establishes robust, cross-consistent references for NIRCam WFSS data products and will underpin ongoing pipeline improvements and reliable dispersed spectroscopy with JWST.

Abstract

We present the absolute flux and wavelength calibration of the James Webb Space Telescope (JWST) Near-Infrared Camera (NIRCam) Wide Field Slitless Spectroscopy (WFSS) mode. Each of NIRCam's two modules (A and B) provides independent long wavelength (LW) grism spectroscopy over the 2.4-5.0 micron range, with orthogonally oriented R and C grisms. Using commissioning and calibration data from programs 01076, 01536, 01537, 01538, 01479, 01480, 04449, 04498, 06606, and 06628, we have measured the field-dependent geometry and wavelength dispersion of both first and second order spectra across the full detector area. The trace geometry was modeled using two-dimensional third-order polynomials that reproduce the observed spectral positions with an RMS accuracy better than 0.1 pixel. Wavelength calibration, derived from observations of the planetary nebula SMP LMC 58, achieves a precision of 0.65-0.91A for the +1 orders and 0.5A for the +2 orders. Absolute flux calibration, established from observations of the G-type star standard P330E, provides a consistent sensitivity function across all grisms and modules with an absolute flux accuracy of 3\%. The resulting calibration framework defines the geometric, wavelength, and photometric reference for all NIRCam WFSS observations and ensures cross-consistency between modules and grism orientations. These calibrations form the basis for accurate slitless spectroscopy with NIRCam and will support ongoing improvements to the JWST calibration pipeline and data products.

Figures (31)

• Figure 1: This figure shows the source positions (black dots) along with the measured segments of the dispersed +1 order traces that were used to calibrate their shape. The coronagraph footprint above the detector is indicated by a black square, with its semi-transparent subtraction shown in dark gray. The approximate extent of the pick-off mirror (POM), which defines the effective field of view of the WFSS mode, is shown in light gray. Our measurements provide good spatial coverage across both the detector and the effective field of view. The spectral trace segments are shown in red and include data obtained with both F322W2 and F444W filters. For an object located at the position indicated by the large black star, the extents of the F322W2 and F444W traces are highlighted with thick blue and red lines, respectively.
• Figure 2: This figure is analogous to Figure \ref{['Trace_FOV']}, but shows the +2 order spectra. Only F322W2 data are included, as the +2 order of the F444W spectra falls outside the detector.
• Figure 3: All traces shown in Figure \ref{['Trace_FOV']} for Module A and Grism R, plotted relative to the source position $(x_0, y_0)$. The dispersed traces exhibit a wide range of shapes and curvature.
• Figure 4: Impact of the polynomial order on the residuals of the trace geometry fit using increasingly higher-order $P_{n,m}(x,y,t)$ polynomials. As shown, using a simple first-order trace with no field dependence ($P_{0,1}(x,y,t)$) results in large residuals of several pixels, as the variations in trace geometry are not captured. A second-order trace without field dependence ($P_{0,2}(x,y,t)$; second panel) also fails to model the geometry accurately. Only by including field dependence (third and fourth panels from the left) are the residuals significantly reduced to a small fraction of a pixel, uniformly across the field of view.
• Figure 5: Contour plots of the mean residuals between our fitted model and the observations for the +1 order traces. As shown, there is little to no field dependence in the RMS of the fit, the spatial coverage is uniform, and the mean error remains within a small fraction of a pixel across the entire detector.