Space Telescope Science Institute
ACS Instrument Handbook Cycle 19
Table of Contents Previous Next Index Print

Advanced Camera for Surveys Instrument Handbook for Cycle 19 > Chapter 6: Polarimetry, Coronagraphy and Prism/Grism Spectroscopy > 6.3 Grism/Prism Spectroscopy

6.3 Grism/Prism Spectroscopy
The ACS filter wheels include four dispersing elements for low resolution slitless spectroscopy over the field of view of the three ACS channels. One grism (G800L) provides low resolution spectra over the 5500  to 10,500  range for both the WFC and HRC. A prism (PR200L) in the HRC covers the range 1700  to beyond 3900  (although reliable wavelength and flux calibration is only guaranteed to 3500 ). In the SBC a LiF prism covers the wavelength range 1150  to ~1800  (PR110L) and a CaF2 prism is useful over the 1250  to ~1800  range (PR130L). Table 6.3 summarizes the essential features of the four ACS dispersers in the five available modes. The grism provides first order spectra with almost constant dispersion as a function of wavelength but with second order overlap beyond about 10,000 ; the prisms however have non-linear dispersion with maximum resolution at shorter wavelengths but much lower resolving power at longer wavelengths. The two-pixel resolution is listed for each grism or prism at a selected wavelength in Table 6.3. The pixel scale for the prism spectra is given at the selected wavelength. The tilt of the spectra to the detector X axis (close to the spacecraft V2 axis) is also listed.
Table 6.3: Optical parameters of ACS dispersers.
Wavelength range ()
Tilt1 (deg)

Tilt with respect to the positive X-axis of the data frame.

The dispersion varies over the field by 11%; the tabulated value refers to the field center.

The dispersion varies over the field by 2%; the tabulated value refers to the field center.

The G800L grism and the WFC provide two-pixel resolving power from 69 (at 5500 ) to 131 (at 10,500 ) for first order spectra over the whole accessible field of 202 x 202 arcseconds. Table 6.3 lists the linear dispersion, but a second order dispersion solution provides a better fit. Figure 6.15 shows the wavelength extent and sensitivity for the zeroth, first, and second order spectra when used with the WFC. Figure 6.16 shows the same plot in pixel extent. The 0 position refers to the position of the direct image and the pixel size is 0.05 arcseconds. Note that there is contamination of the 1st order spectrum above 10,000  by the second order. The total power in the zeroth order is 2.5% of that in the first order, so locating the zeroth order may not be an effective method of measuring the wavelengths of weak spectra. The default method will be to obtain a matched direct image-grism pair. There is also sensitivity of about one percent of the first order in the third and fourth orders, and about half a percent in the negative orders. The full extent of the spectrum of a bright source (orders -2, -1, 0, 1, 2, 3) is 1200 pixels (60 arcseconds). The higher orders are not in focus and the spectral resolution of these orders is therefore less than what would be predicted from the nominally higher dispersion.
Figure 6.15: Sensitivity versus wavelength for WFC G800L.
Figure 6.16: Sensitivity versus pixel position for WFC G800L.
Figure 6.17 shows the full spectrum extent for a 60 second exposure of the white dwarf GD153 (V = 13.35) with the individual orders indicated. When bright objects are observed, the signal in fainter orders may be mistaken for separate spectra of faint sources and in crowded fields, many orders from different objects can overlap. The wavelength solution is field dependent on account of the tilt of the grism to the optical axis, and the linear dispersion varies by 11% from center to corner. This field dependence has been calibrated to allow wavelength determination to better than 0.5 pixels over the whole field.
Figure 6.17: Full dispersed spectrum for white dwarf GD153 with WFC/G800L. The numbers indicate the different grism orders.
When used with the HRC, the G800L grism provides higher spatial resolution (0.028 arcseconds) pixels than the WFC and also higher spectral resolution. However, the spectra are tilted at −38 to the detector X axis. Figure 6.18 shows the wavelength extent and sensitivity in the zero, first and second orders, with the pixel extent shown in Figure 6.19. Figure 6.20 shows the observed spectrum of the standard star GD153. Again there is contamination of the first order spectrum by the second order beyond 9500 . The total extent of the spectrum (orders -1 and +2) in Figure 6.20 covers about 70% of the 1024 detector pixels. In addition, a much greater number of spectra will be formed by objects situated outside the HRC direct image, or will have their spectra truncated by the chip edges, than for the WFC. The variation of the grism dispersion over the HRC field is about 2% from center to corner and has been calibrated.
Figure 6.18: Sensitivity versus wavelength for HRC G800L.
Figure 6.19: Sensitivity versus pixel position for HRC G800L.
Figure 6.20: Full dispersed spectrum of white dwarf GD153 with HRC/G800L. The numbers indicate the different grism orders.
The maximum dispersion of the prism is 5.3  at 1800 . At 3500 , the dispersion drops to 105 /pix and is 563 /pix at 5000 . The result is a pile-up of the spectrum to long wavelengths with 8 pixels spanning 1500 . For bright objects, this effect can lead to blooming of the HRC CCD from filled wells; the overfilled pixels bleed in the detector Y direction, and thus can affect other spectra. Figure 6.21 shows the sensitivity versus wavelength for PR200L and the wavelength extent of the pixels is indicated. The variation of the dispersion across the detector for PR200L amounts to about 4% (corner to corner) at 2000 . The tilt of the prism causes a deviation of about 300 pixels between the position of the direct object and the region of the dispersed spectrum on the CCD, while simultaneously vignetting this same amount from the low-x side of the HRC image. We have therefore defined an additional prism aperture with a recentered reference point. Because of this and the real deflection caused by the prism, the telescope has to be shifted by 7.4 arcseconds to place the target at the center. The use of this special aperture and the associated pointing correction takes place automatically when the prism PR200L is chosen. Users specify HRC as the aperture in conjunction with the filter PR200L. The current calibration of the wavelength solution used in the aXe data reduction software (see Section 6.3.7) assumes the use of these apertures. The numbers in Figure 6.21 indicate the resolving power (R) and the offset between the direct and prism images in pixels (Δx) as functions of wavelength. These offsets and the wavelength calibration take into account the use of the two different pointings for direct and prism observations.
Figure 6.21: Sensitivity versus wavelength for HRC/PR200L. The numbers indicate the resolving power (R) and the offset from the direct image in pixels (Δx) as functions of wavelength.
The PR110L prism is sensitive to below 1200  and includes the geo-coronal Lyman-alpha line, so it is subject to high background. The dispersion at Lyman-alpha is 2.6  per pixel. Figure 6.22 shows the sensitivity with wavelength and the wavelength width of the pixels. The long wavelength efficiency decline of the CsI MAMA detector beyond ~1800  occurs before the long wavelength pile-up. However, prism observations of flux standards and a G-type star indicate that the system throughput is higher by a factor of approximately 1000 at 4000  with respect to the curve in Figure 4.12. For stars redder than about F-type this will result in significantly increased counts between 2000  and 6000 , peaking at about 3500 . For a G-type star this peak can be a factor of ~3 times larger than the maximum count rate in the UV. The dispersion at 1800  is 21.6 /pixel. The detected counts in the spectrum must be within the MAMA Bright Object Protection Limits (BOP) (see Section 4.6). These limits must include the contribution of the geo-coronal Lyman-alpha flux per SBC pixel. The numbers in Figure 6.22 indicate the resolving power (R) and the offset from the direct image in pixels (Δx) as functions of wavelength. As for the HRC, the prisms deflect the SBC image, although this time the vignetting is on the +x part of the image. Users specify SBC as the aperture in conjunction with the filter PR110L and the correct prism aperture will automatically be used. The current wavelength calibration used in aXe (see Section 6.3.7) assumes these default apertures. In Figure 6.23 a direct image and a PR110L exposure of the standard star WD1657+343 are shown, as an example.
Figure 6.22: Sensitivity versus wavelength for SBC PR110L. The numbers indicate the resolving power (R) and the offset from the direct image in pixels (Δx) as functions of wavelength.
Figure 6.23: Sum of direct (F122M) and PR110L prism exposure of the standard star WD1657+343. The direct image exposure was scaled prior to combination for representation purposes. The cutout shown covers 400 x 185 pixels, where the wavelength increases from left to right for the dispersed image.
The short wavelength cut-off of the PR130L prism at 1250  excludes the geocoronal Lyman-alpha line, making it the disperser of choice for faint object detection in the 1250  to 1800  window. The dispersion varies from 1.65 /pixel at 1250  to 20.2 /pixel at 1800 . Figure 6.24 shows the sensitivity versus wavelength, with the resolving power (R) and the offset from the direct image in pixels (Δx) as functions of wavelength. Bright Object Protection considerations similar to the case of PR110L also apply to the use of this prism, except that the background count rate is lower (see Section 4.6). As for the other prisms, the direct and dispersed images use different apertures with a small angle maneuver between them.
Figure 6.24: Sensitivity versus wavelength for SBC/PR130L. The numbers indicate the resolving power (R) and the offset from the direct image in pixels (Δx) as functions of wavelength.
The normal observing technique for all ACS spectroscopy is to obtain a direct image of the field followed by the dispersed grism/prism image. This combination allows the wavelength calibration of individual target spectra by reference to the corresponding direct images. For WFC and HRC, the scheduling system automatically inserts a default direct image for each specified spectroscopic exposure, which for G800L consists of a 3 minute F606W exposure, and for PR200L, a 6 minute F330W exposure. If the observer wishes to eliminate the default image, the supported Optional Parameter AUTOIMAGE=NO can be specified. Then a direct image with a different filter and/or exposure time can be specified manually, or no direct image at all in the case of repeated spectroscopic exposures within an orbit, or if no wavelength calibration is required. For the SBC prisms, there is no default direct image because of the Bright Object Protection requirements (Section 7.2); the direct image must always be specified manually, and it must satisfy the BOP limits, which will be more stringent than for the dispersed image.
Because of the offset between the direct imaging and prism aperture definitions, the SAME POS AS option will generally not have the desired effect for prism spectroscopy (PR110L, PR130L, and PR200L). Users who wish to specify offsets from the field center by means of the POS-TARG option should do so by explicitly specifying the same POS-TARG for the direct imaging and prism exposures.
Table 6.4 lists the V detection limits for the ACS grism/prism modes for unreddened O5 V, A0 V, and G2 V stars, generated using the Exposure Time Calculator. WFC and HRC values used the parameters CR-SPLIT=2 and GAIN=2. An average sky background was used in these examples. However, limiting magnitudes are sensitive to the background levels; for instance, the limiting magnitude of an A0 V in the WFC using the F606W filter changes by 0.4 magnitudes at the background extremes.
Table 6.4: V detection limits for the ACS grism/prism modes.
Wavelength of reference ()
O5 V (Kurucz model)
A0 V (Vega)
G2  V (Sun)
Chapter 9 provides details of the calculations. Depending on the wavelength region, the background must also be taken into account in computing the signal-to-noise ratio. The background at each pixel consists of the sum of all the dispersed light in all the orders from the background source. For complex fields, the background consists of the dispersed spectrum of the unresolved sources; for crowded fields, overlap in the spectral direction and confusion in the direction perpendicular to the dispersion may limit the utility of the spectra.
The ACS Exposure Time Calculator supports all the available spectroscopic modes of the ACS and is available for more extensive calculations at:
The current version employs the on-orbit determinations of the dispersion solution and sensitivity where available.
For more detailed, two-dimensional simulations of ACS slitless spectra, an IRAF/PyRAF package called aXeSIM is available. aXeSIM generates slitless images and their associated direct images using object shapes and spectra given as input. In the most primitive form, aXeSIM uses Gaussians as object shapes and direct imaging magnitudes as “spectra”, however, more realistic object shapes (e.g. a PSF from TinyTim) and high resolution spectra can be provided. aXeSIM is described in Kuemmel, Kuntschner & Walsh (2007, ST-ECF Newsletter 43, 8) and Kuemmel et al. (2009, PASP 121, 59).
Please refer to the ACS Web page and for the aXeSIM software and related information.
Since there is no slit in the ACS, the Point Spread Function of the target modulates the spectral resolution. In the case of extended sources it is the extension of the target in the direction of dispersion which sets the achievable resolution. Simulations show that for elliptical sources, the spectral resolution depends on the orientation of the long axis of the target to the dispersion direction and is described in more detail in ACS ISR 2001-02. The dispersion of the grisms and prisms is well characterized, but for the wavelength zero point it is important to know the position of the target in the direct image. For the grisms, the zeroth order will generally be too weak to reliably set the wavelength zero point. Given the typical spacecraft jitter, wavelength zero points to 0.4 pixels should be routinely achievable using the direct image taken just before or after the slitless spectrum image.
The jitter information can be used to obtain more accurate coordinates for the center of the FOV. These in turn allow one to determine better relative offsets between the direct and the spectroscopic images.
The wavelength extent of each pixel for the WFC and HRC G800L modes in the red is small enough that fringing can modulate the spectra. For the HRC, the peak-to-peak fringe amplitude is about 30% at 9500 , and is about 25% for the WFC chips. Modelling of the fringing for WFC and HRC chips is described in ACS ISR 2003-12. In practice, for observations of point sources, and even more so for extended sources, the detectable fringing is significantly reduced by the smoothing effect that the PSF, together with the intrinsic object size, has on the spectrum in the dispersion direction. In ACS ISR 2008-01 it is shown that the error due to fringing amounts to less than 0.1% for continuum sources and can therefore be neglected. For narrow emission lines fringing can cause in principle line flux variations of 12% and more. For more realistic scenarios as in the case of emission lines in a Wolf Rayet star we measure variations of order 4%.
Because the STSCI pipeline does not provide an extracted spectral count rate vs. wavelength, an extraction software package, aXe, is available to extract, wavelength calibrate, flat field, and flux calibrate ACS grism and prism spectra. Full details can be found in Kuemmel et al. 2009, PASP 121, 59 and at:

Table of Contents Previous Next Index Print