An accurate measurement of the spectral resolution of the JWST Near Infrared Spectrograph

TL;DR

This work provides a robust, wavelength-dependent calibration of the JWST/NIRSpec spectral resolution for fixed-slit and integral-field modes by fitting H and He nebular lines in the planetary nebula SMP LMC 58, corrected for the nebula's intrinsic expansion velocity measured with VLT/X-shooter. The authors model the line-spread function as Gaussian and derive a two-parameter, inverse-linear dependence of the instrumental dispersion on wavelength, yielding $\sigma_{\rm inst}(\lambda)$ and corresponding $R(\lambda)$ across multiple disperser–filter configurations. They show that in-flight resolutions exceed pre-launch JDox estimates by 11–53% (FS) and 1–24% (IFS), with the resolution increasing with wavelength as expected, and provide a publicly available parametric description to enable high-precision kinematic analyses. These results improve the reliability of velocity-dispersion measurements from NIRSpec data in cosmology and galaxy evolution studies.

Abstract

The spectral resolution ($R \equiv λ/ Δλ$) of spectroscopic data is crucial information for accurate kinematic measurements. In this letter, we present a robust measurement of the spectral resolution of the JWST's Near Infrared Spectrograph (NIRSpec) in fixed slit (FS) and integral field spectroscopy (IFS) modes. Due to the similarity of the utilized slit dimension if the FS mode to that of the shutters in the multi-object spectroscopy (MOS) mode, our resolution measurements in the FS mode can also be used for the MOS mode in principle. We modeled H and He lines of the planetary nebula SMP LMC 58 using a Gaussian line spread function (LSF) to estimate the wavelength-dependent resolution for multiple disperser and filter combinations. We corrected for the intrinsic width of the planetary nebula's H and He lines due to its expansion velocity by measuring it from a higher-resolution X-shooter spectrum. We find that NIRSpec's in-flight spectral resolutions exceed the pre-launch estimates provided in the JWST User Documentation by 11-53% in the FS mode and by 1-24% in the IFS mode across the covered wavelengths. We recover the expected trend that the resolution increases with the wavelength within a configuration. The robust and accurate LSFs presented in this letter will enable high-accuracy kinematic measurements using NIRSpec for applications in cosmology and galaxy evolution.

Figures (3)

• Figure 1: Post-pixelized Gaussian profile fits (orange) to prominent emission-line groups (black) in the NIRSpec IFS spectra of the PN SMP LMC 58. The IFS-mode G140 disperser is illustrated here as an example among the nine combinations analyzed in this letter. The top row corresponds to the G140M/F100LP configuration (i.e., medium resolution), and the bottom row to G140H/F100LP (i.e., high resolution). The principal line within each group is annotated in the corresponding panel. The individual line wavelengths fitted in each group are marked with dashed blue lines. All of these lines are fitted simultaneously to directly constrain the parameters in $\sigma_{\rm inst}^\prime(\lambda)$ from Eq. (\ref{['eq:parametrization']}).
• Figure 2: Measurement of the PN's expansion velocity. Left-hand panel: Measurement of the X-shooter resolution (black points with error bars) from arc-lamp lines with the wavelength dependence modeled with a quadratic function (blue line, with the shaded region showing 1$\sigma$ uncertainty). Right-hand panel: The H$\alpha$ line of the PN in the X-shooter spectra (emerald points). The error bars are too small to be noticeable here. The best-fit post-pixelized Gaussian profile is shown in red, while the width of the X-shooter LSF, as determined by the best fit in the left-hand panel, is shown in blue for comparison. Both sets of illustrated datapoints are simultaneously fit to infer the PN's expansion velocity $\sigma_{\rm PN} = 6.90 \pm 0.49$ km s$^{-1}$.
• Figure 3: Comparison of our measured resolution curves in the FS (emerald) and IFS (orange) modes with the pre-launch estimates from JDox (dashed grey lines). The shaded region around each resolution curve signifies the 1$\sigma$ credible region. The in-flight resolutions are 11--53% higher than the pre-launch estimates in the FS mode, and 1--24% higher in the IFS mode (across the covered wavelengths).