8.6 Slitless Spectroscopy with Spatial Scanning

Spatial scanning of stellar spectra using the IR detector creates the potential for spectrophotometry of higher precision than possible via staring mode. By spreading a stellar spectrum perpendicular to its dispersion (Figure 8.11), more photons can be collected per exposure, and the exposure times can be longer without saturating the detector. This strategy results in higher signal-to-noise observations and an increased duty cycle. The most prevalent scientific application is transit spectroscopy, in which a time series of stellar spectra are obtained before, during, and after an exoplanet transit or eclipse; observations of this type with one-sigma precision of ~20 ppm in ~20 nm spectral bins from two transits have been reported (Knutson et al. 2014). Results for exoplanet transit spectroscopy have been published by numerous investigators; recent examples using WFC3 spatial scanning include Alam et al. 2022, Alderson et al. 2022, Barat et al. 2023, and Barclay et al. 2023. 

The WFC3 team has introduced and will continue to update the Transiting Exoplanets List of Space Telescope Spectroscopy (TrExoLiSTS, formerly called ExoCat; see WFC3 ISR 2022-09) that summarizes the existing WFC3/IR spatial scanning observations of time series observations acquired during primary transits, secondary eclipses, and phase curves. Information presented includes the observational parameters along with the direct image, white light curve, and the measured drift of the spectrum in the field of view. The goal of presenting this catalog is to facilitate the preparation of WFC3 exoplanet programs by providing observers with information that they can use to optimize their target selection and to prevent target duplications.

In the time series, the integrated flux of a WFC3/IR spectrum exhibits a hook-shaped rise in each HST orbit due to charge trapping within the detector pixels. Mitigation of the “hook” has been attempted in observational techniques (WFC3 ISR 2014-14) and in post-observation modeling (e.g., Zhou et al. 2017).

Spatial scanning is available with either WFC3 detector, UVIS or IR. However, overlap of spectral orders and curved spectral traces will compromise scanning’s utility for the UVIS grism. In this section, we assume that for IR spectroscopy, the observer desires the +1st order spectrum. Spatial scans are discussed elsewhere in this handbook (for UVIS imaging in Section 6.11.3, for IR imaging in Section 7.10.4) and in WFC3 ISR 2012-08. The latter is particularly relevant for anyone preparing a Phase II proposal.

Potential benefits of spatial scanning are 1) reducing overhead for time-series of short exposures due to detector operations required before and after each exposure, 2) avoiding saturation for very bright stars, 3) improved spectrophotometry, as collecting more photons per HST orbit improves the signal-to-noise ratio. We note that the main disadvantages are 1) STScI pipelines do not work on spatially-scanned IR data so the observer will need to reduce IR data themselves even to produce simple images, and 2) astronomical sources will overlap more often than with staring-mode observations, especially for spectra.

The scan rate can be any real number between 0.0 and 7.84 arcsec s^-1. Without FGS control, referred to as “gyro control”¹, rates as high as 7.5 arcsec s^-1 have been demonstrated. Under fine-guidance-sensor (FGS) control, rates between 0.0 and 4.8 arcsec s^-1 are supported for exposures with a single scan line. Due to a software limitation, boustrophedonic (serpentine) scans at rates greater than 1 arcsec s^-1 must be executed under gyro control. Observing under gyro control should be avoided if at all possible because the increased drift on the detector significantly degrades the measurement precision.

WFC3 ISR 2012-08 provides recommended scan rates for WFC3 IR spectroscopy of bright stars as well as formulae to predict appropriate scan rates. For stars with H-band apparent brightnesses fainter than H = 4.1 mag, G141 spectra can be unsaturated with scan rates less than or equal to 4.8 arcsec s^-1, which is the maximum achievable under FGS control. FGS control is recommended for time-series applications such as exoplanet transit spectroscopy, in order to keep the spectrum from drifting on the detector from one scan to the next, during an HST orbit. For stars in the range 4.1 > H > 3.7 mag, WFC3 G141 spectra can be obtained with less than 25,000 DN/pixel only under gyro control, i.e. with rates between 4.8 and 7.8 arcsec s-1. For stars brighter than H = 3.7 mag, spectral orders other than the +1st must be used to avoid saturation of the IR detector; e.g., observations of Vega in Visit 1 of HST calibration program 12336 used the -1st order (WFC3 ISR 2014-15), thus demonstrating that stars as bright as V~H~0.0 mag can be observed.

Python code and web tools are available to assist observers in designing spatial scanning observations for exoplanet transits. The PandExo code (Batalha et al. 2017) was extensively tested using WFC3 observations, and provides realistic noise estimates for simulated time-series observations based on system parameters, an input planetary model, and various observation parameters (e.g., number of transits/eclipses, number of spectrophotometric channels, size of the observation start window) and various WFC3 parameters (e.g., disperser, subarray size, N_SAMP, SAMP_SEQ, scan direction). The simulations can also account for the WFC3 ramp effects described above. The GUI for PandExo is integrated into ExoCTK (Exoplanet Characterization Toolkit) and can be used to make one-off calculations in a web form. PandExo_HST can be locally installed to perform more detailed calculations on one's own computer.

The sample sequence SPARS5, introduced during Cycle 23, has time steps intermediate between those of RAPID and SPARS10. As for all WFC3/IR sample sequences, the time intervals depend on the size of the aperture (see Sections 7.7.3 and 7.7.4). SPARS5 was designed to allow the efficiency and uninterrupted time series of SPARS10 and yet be short enough in cadence to better isolate a target exoplanet-host star from a nearby stellar companion in spatially-scanned observations using an IR grism. Better isolation may enable the observer to specify wider orient ranges, which may increase scheduling opportunities, which are often rare for these time-constrained observations. We expect that SPARS5 will be advantageous especially for G141 or G102 grism observations of stars brighter than approximately 7th magnitude in H band, scanned at approximately 1 arcsec per second or faster. The subarray apertures are especially useful for IR grism observations of bright stars with the sample sequences RAPID, SPARS5, and SPARS10 because of their short time steps and efficient orbit packing. The combination of SPARS10 with a 512 × 512 aperture is not currently supported (and therefore has a lower quality dark reference file), but spectroscopic spatial scan observers who need it for improved observing efficiency and consequent increases in signal-to-noise may send a justification for available mode use to the Contact Scientist or the Program Coordinator.

APT provides a diagram to assist observers planning spatial scan observations as shown in Figure 8.12. The line connecting the green and red arrows corresponds to where the target’s direct image would appear. In grism observations, the first order spectrum appears at larger X coordinate values than the direct image (see Figures 8.5 and 8.7). By design of the GRISM128, GRISM256 and GRISM512 subarrays, a POSTARGX = 0 centers the first order spectrum in each subarray in the X coordinate. Because the diagram is only approximate and because of on-going developments in APT and spacecraft operations, users of fast spatial scans (~1 arcsec/sec or faster) should consult with their contact scientist to optimize the POSTARGY value.

¹ HST uses gyros at all times to maintain spacecraft pointing, regardless of whether the FGS are used also. Here, the phrase “gyro control” means “without FGS control”.