NIRISS GR150 Grisms
The JWST NIRISS GR150C and GR150R grisms produce spectra for wavelengths between 0.8 and 2.2 μm with a resolving power in first order of R ≈ 150. They are used with blocking filters to enable wide field slitless spectroscopy (WFSS).
The NIRISS GR150C and GR150R grisms produce spectra of essentially all objects in the field of view with a resolving power of of R ≡ λ/Δλ ≈ 150 over the wavelength range between 0.8 and 2.2 μm. When used with a blocking filter, these grisms enable the WFSS mode of NIRISS. Figure 1 shows the GR150R grism before it was mounted in the filter wheel.
The GR150C and GR150R are identical optical elements. Their physical properties are summarized in Table 1, and leads to the following optical performance characteristics:
Dispersion: in the first order, the grisms provide linear dispersion of 0.00478 μm/pixel.
Resolving power: The NIRISS point spread function is undersampled at the wavelengths covered by the GR150 grisms. Consequently, the resolving power for a point source is defined by the wavelength increment over two pixels, which is 0.0095 μm. As shown in Figure 2, the resolving power is directly proportional to wavelength. Under many circumstances, the resolving power will actually be dictated by the angular size and geometry of the astronomical source.
Throughput: Figure 3 shows the throughput for the 1st and 2nd order of the GR150 grisms based on the properties listed in Table 1. These curves have been scaled to agree with measured values of the transmission at a few discrete wavelengths.
Table 1. Physical properties of the NIRISS GR150 grisms
Diameter of ruled area
|Resolution element||0.0095 μm|
Although the properties of the GR150C and GR150R grisms are identical, they are mounted in the filter wheel so that the spectra they produce are dispersed in orthogonal directions on the detector; see Figure 4. In the coordinate system used by the JWST calibration pipeline, the GR150C grism disperses in the "slow" direction of detector readout, while the GR150R grism disperses in the "fast" readout direction. This geometry is indicated in Figure 5. Access to spectra obtained in both directions helps to combat problems with blending, especially in crowded fields; and maximizes the chance that a "clean" spectrum can be extracted from the data.
Main Article: NIRISS Filters
The GR150 grisms are used with a blocking filter, which isolates a specific range of wavelengths. As a result, the length of the trace of the first order spectrum on the detector is reduced, which lessens the degree of overlapping and blending in crowded fields.
Table 2 summarizes the bandpasses associated with the available blocking filters; the range of redshifts of Lyman α emission that fall within the bandpass; the length of the spectral trace in pixels on the detector; and the relative offset in pixels between the direct image of the source and the beginning of its spectral trace (assuming there have been no intervening motions of the observatory). This information is presented graphically in Figure 6, which shows the extent and offsets associated with multiple orders of the GR150C grism when observed through different blocking filters.
Table 2. Length and location of the first order spectra produced by the GR150 grisms
Central wavelength (μm)
Extent of 1st order (pixels)
Offset from direct image (pixels)
† Defined by the 50% transmission points of the blocking filter.
Out-of-focus "ghost" images produced by multiple reflections within the optical trains of NIRISS have been detected in deep exposures of bright sources obtained during ground-test campaigns. The "double stack" filters, which are composed of two optical elements, are especially prone to multiple internal reflections. Figures 7 and 8 show the families of ghosts that have been identified in deep exposures that combine double stack filters with the GR150C and GR150R grism, respectively. The strengths of the ghosts associated with the 0th and 1st (m = +1) orders are indicated as a fraction of the flux contained in the parent order. In general, the flux in the ghost image is less than 1% of flux in the parent order, which suggests that ghosts will only be detectable for the brightest sources in a typical exposure. The locations and strengths of the ghosts associated with a particular bright source can be predicted reliably.