Instrument Science Reports

We discuss the analysis of two sets of optical/near-IR spectra of the exoplanet host star 55 Cnc, obtained with the Space Telescope Imaging Spectrograph (STIS) and grating G750L in spatial scanning mode, in order to assess the performance of that relatively new observing mode for studies of transiting exoplanets. Standard pipeline reductions of the CCD spectral images were augmented by custom procedures for removing both cosmic rays and the strong fringing seen at wavelengths longer than about 7000 Å. Both total (“white-light”) fluxes and the fluxes for some narrower wavelength intervals were extracted from the processed spectral images. Apart from slight (∼400 ppm) orbit-to-orbit offsets between the relative fluxes in each set, the patterns exhibited by the flux values within each orbit are very similar. The systematic differences in the fluxes are somewhat smaller than those seen in archival STIS spectra of 55 Cnc obtained in so-called “stare mode”, in which the CCD is deliberately saturated at a fixed pointing. A parameterized detrending method similar to those commonly used to remove instrumental effects from time series observations of exoplanet host stars was then applied to the extracted fluxes. For the total fluxes, the scatter about the detrending models is of order 30-40 ppm – comparable to the best precision previously obtained for time series photometry with HST. The scatter is somewhat larger for the narrower wavelength bins – particularly at the longer wavelengths where the CCD is less sensitive; the defringing does reduce the scatter by 15-20% at the longer wavelengths, however. The depth of the transit of the super-Earth 55 Cnc e (∼ 450 ppm for the total flux) is consistent with previously obtained values. Both the scan-mode and the stare-mode observations of 55 Cnc e appear to indicate an unexpected (and variable?) increase in the transit radius Rp/Rs between 0.55 and 1.0 μm (by >40% for the scan-mode data). While these data are somewhat limited, they do suggest that spatial scanning with the STIS CCD can provide high-quality optical/near-IR spectra of the brighter exoplanet hosts.

We investigate the wavelength calibration accuracy of STIS CCD spectra as a function of detector position, focusing on the E1/E2 pseudo-apertures. Comparison between measured lamp line centroids and laboratory wavelengths shows that, while the standard calstis solution is stable at the detector center, significant offsets are present toward the CCD edges and increase with time. We used a test version of calstis in which the wavelength calibration step employs a row-selected cross-correlation, leading to improved wavelength calibration accuracy for edge extractions. Motivated by these results, we implemented an update to calstis4 that applies a row-selected wavecal procedure to observations taken with the E1/E2 pseudo-apertures, while preserving the standard procedure for nominal extractions. The updated pipeline was then used to reprocess the entire STIS archive in MAST. Validation tests confirm improved agreement between nominal and E1/E2 spectra without significant side effects for more than 97% of the datasets in the MAST archive. The main exceptions are the majority of G230MB and a few G230LB and G430M datasets, where the lower signal-to-noise at the CCD edge makes the lamp lines harder to detect and cosmic-ray residuals can dominate the cross-correlation, leading to incorrect wavelength shifts. A Jupyter Notebook is also provided to enable similar corrections in case the cross-correlation fails, for extractions at positions other than E1/E2, or for multiple extractions along the slit. Finally, we discuss the potential impact of the wavelength calibration accuracy on science results.

It has been previously studied that the STIS CCD exhibits a slow rotation, evident from changes in spectral traces and flat-field evolution. In this Instrument Science Report, we present an independent spectroscopic analysis of the CCD rotation angle and a comparison with previous studies. We describe the methodology used to derive rotation angles and rates across different gratings and central wavelengths, and report the corresponding median rotation rates over time. The results are broadly consistent within uncertainties with earlier measurements, indicating the CCD rotation has the same impact on all the CCD gratings. We also discuss key assumptions limiting our analysis. These findings contribute to the ongoing monitoring of STIS CCD rotation and its implications for calibration accuracy.

We describe stistools.barycentric correction, a new Python utility for calculating barycentric timing corrections for HST/STIS observations. This tool replaces the deprecated stsdas.hst_calib.stis.odelaytime IRAF function that was previously used for HST barycentric corrections. Our new utility uses astropy for conversion between time formats and standards and introduces a new way to calculate HST’s position through JPL Horizons, replacing the need to download separate HST orbital ephemeris files. Here, we describe the methods used in the new utility, the tests that were carried out to verify its accuracy, and explain some of the complexities involved in determining light travel times to accuracies down to a millisecond for HST. We also summarize the current understanding of the absolute accuracy of STIS time stamps.

The STIS team monitors the time dependent sensitivity (TDS) of each grating with one from a set of three secondary CALSPEC standard stars: GRW+70D5824, AGK+81D266, and BD+28D4211. Here, we use the three primary CALSPEC White Dwarf standard stars, dubbed the standard star “triad” (GD71, GD153, G191B2B), as an independent set of standards to verify the accuracy of STIS TDS corrections derived from the TDS monitoring stars, increasing the sample for each STIS L-mode from one up to three or four standard stars. We focus on triad star observations using the STIS L-mode gratings (e.g., G140L, G230L, etc.) with the same configuration as our standard TDS monitoring programs, and compare the triad observations to the TDS pipeline trends. Our analysis indicates the relative net count rates inferred from the triad standards agree with the TDS trends derived from the TDS monitoring stars with average residuals < 2% across the full wavelength range of STIS, suggesting our current TDS L-mode trends are reliable and robust. We note that the dispersion in the residuals does vary with wavelength, with the NUV showing the lowest spread (± 0.32% at 2400-2500 Å) and the NIR the largest (± 1.32% at 9500-9900 Å); however, this scatter is also seen in our measurements of the TDS monitoring stars and is more indicative of other instrumental effects. Our findings rule out long term deviations, such as variability in our TDS monitoring stars, within measurement uncertainties.

We discuss the behavior of the default “wavecal” spectra obtained together with most STIS spectroscopic exposures, which are needed for proper wavelength calibration of the science data. Because the Pt/Cr-Ne lamps used for the wavecals have been fading (especially at the shortest wavelengths), some changes in the default lamp and/or exposure time have been implemented in recent years to maintain accurate calibrations. To assess whether additional changes might be appropriate, we examine the trends in the SHIFTA1 and SHIFTA2 values derived from the wavecals (the x and y offsets of the spectral image on the detector), we re-visit the wavelength-dependent fading of the lamps, and we perform simulations to estimate the exposure times that would be needed to obtain accurate SHIFTA values. While the current wavecals do appear to yield reasonable SHIFTA, increases in the default exposure times for some of the shortest-wavelength settings would help to ensure reliable wavelength zero points as the lamps continue to fade.

We evaluate uncertainty calculations in the calstis pipeline for data in the low-count regime. Due to the low dark rate and read-noise free nature of MAMA detectors, observations of UV-dim sources can result in exposures with 0 or 1 counts in some pixels. In this regime, the “root-N” approximation widely used to calculate uncertainties breaks down, and one must compute Poisson confidence intervals for more accurate uncertainty calculations. The CalCOS pipeline was updated in 2020 to account for these low-count uncertainties. Here, we assess how STIS observations are currently affected by this phenomenon, describe a new Jupyter notebook exploring the issue, and introduce a new utility, stistools.poissonerr, to manually calculate Poisson confidence intervals for 1D STIS spectra. Additionally, we describe a related software bug in the stistools.inttag utility, which splits TIME-TAG data into sub-exposures. This newly fixed bug serves as a useful case-study for the proper use of Poisson confidence intervals.

The sensitivities of STIS first-order, medium resolution modes were redetermined from on-orbit observations and CALSPEC models (version 11) of the primary white-dwarf spectrophotometric standard stars G191-B2B, GD 71, and GD 153. The sensitivity of an additional configuration was updated by comparing observations of the secondary standard BD +75°325 with the STIS low-resolution spectrum that has been calibrated consistently with the version 11 models. The procedures used to derive the sensitivities and verify the PHOTTAB reference files prior to their activation in CRDS (on May 1, 2025) are described. Results are presented in graphical form in an extensive appendix. Issues and uncertainties are discussed briefly, along with recommendations for future work.

Re-derivation of the sensitivities of all of the Space Telescope Imaging Spectrograph (STIS) observing modes were required after major updates were introduced to the model atmospheres of the three primary standard stars. The new predicted continuum fluxes were up to 2–3% different from the models used to originally calibrate STIS. This work focuses on the re-derivation of spectral sensitivities for the five STIS low-resolution (L-mode) gratings: G140L, G230L, G230LB, G430L, and G750L, which span wavelengths from the far-ultraviolet through the near infrared. Updated photometric throughput tables were delivered to the Calibration Reference Data System (CRDS) on April 7, 2022 and April 14, 2023, which triggered a recalibration of all historical STIS datasets taken with these modes. The sensitivities derived from each of the standard stars typically agree with one another to better than 1%, though discrepancies as large as 1.5% are found in spectral regions most impacted by hydrogen absorption.

We discuss the suite of STIS calibration programs executed during HST Cycle 30, covering the period 2022 Nov 07 through 2023 Nov 05. For each of the 19 current regular calibration programs, we provide brief descriptions of the objectives, observations, analysis procedures, and results – with comparisons to the results from previous cycles and to desired accuracies, as well as references to more detailed analyses of the calibration data. Many of these calibration programs produce routine reference file deliveries or demonstrate the continuing applicability of existing reference files for processing STIS observations. This ISR provides a brief snapshot of the current instrument performance, similar to those given in annual reports for Cycles 7–10 and 17–21. Two Appendices briefly discuss the state of the onboard calibration lamps and the ongoing major effort to revise the flux calibration for the many STIS spectroscopic and imaging modes.

The STIS team re-derived on-orbit sensitivity curves for the echelle modes for post-servicing mission 4 observations using the standard DA white dwarf G 191-B2B. These new updates relied on the recent CALSPECv11 models, which introduced improvements in the fluxes of the primary standard stars of the order of ~1-3% depending on the wavelength of interest. As part of this effort, the team also released new blaze shift coefficients and echelle ripple tables. We present a detailed description of the procedures followed in the derivation of these new throughputs and the accompanying updates.

The STIS echelle gratings can be used with a variety of different central wavelength settings. “Secondary” wavelength settings, designed to cover select absorption or emission lines, have not been calibrated as precisely as their primary mode counterparts. In particular, secondary echelle mode traces (and subsequent extraction regions) have been previously defined using straight line fits to each spectral order. In this work, we define a new general method for defining echelle traces that utilizes Gaussian process regression and accounts for the detailed curvature of each order across the detector. Across a variety of echelle grating and central wavelength settings, we find that this method can improve flux throughput by ~4% especially near wavelengths located close to the edge of the detector. We have used this method to provide new traces and update reference files for 9 different echelle modes for both pre- and post-Servicing Mission 4 (SM4; in 2009) observations.

Recent improvements to stellar atmospheric models have merited updated flux calibration for high priority STIS observing modes. Specifically, in the FUV and NUV, continuum differences of 1-3% are present between the newest models (CALSPECv11) and previous models (CALSPECv04-v07). As a result of these improvements the STIS team has derived updated sensitivity curves and blaze shift coefficients for a variety of echelle modes in order to meet targeted flux accuracies. The first series of echelle sensitivity updates primarily targeted post-Servicing Mission 4 (SM4; in 2009) observations. In this ISR, we investigate instead applying a simple scaling (derived from the ratio of new vs old CALSPEC model continua) to the previously determined throughputs of STIS echelle modes. This alternative approach has a straightforward implementation and provides reasonable accuracy, especially in cases where available calibration data are lacking (e.g., pre-SM4 era). Adopting this scaling approach, we delivered pre-SM4 throughput updates for 8 echelle modes, resulting in typical improvements of 0.5-2.4% across the FUV and NUV.

For observations of supernovae, kilonovae, tidal disruption events, and other non-repeatable observations, it is important the science data is taken successfully within a specific time window. Part of obtaining that data is often centering objects in the aperture to a higher accuracy than is available from Hubble Space Telescope’s (HST’s) blind pointing. On the HST Space Telescope Imaging Spectrograph (STIS) the sequence of exposures responsible for this centering is the target acquisition or STIS ACQ sequence, and it is most often placed only at the beginning of a visit. Unfortunately, STIS ACQ sequences will fail if the observatory experiences issues locating guide stars in time for the start of the required exposures. If the guide stars are located at a later point in the visit, the remaining science exposures can be taken but the pointing might not be as accurate as is required. This work discusses both the frequency of this issue and the feasibility of placing redundant or “safety” STIS ACQ sequences in a multi-orbit visit to regain the desired pointing accuracy in an affected visit. To do so we select a subset of all 113 STIS ACQ sequences from September 2018 to September 2023 which have experienced this issue. We find that this problem occurs in ∼5% of the total STIS ACQ sequences taken during that time period, with a recent increase in the rate to ∼9% from March to September 2023. Since the observatory goes through periods of better or worse pointing performance, this recent increased failure rate is not guaranteed to continue. For those failed visits which span multiple orbits, ∼39% never obtain a lock on the guide stars and thus take no data. Of the multi-orbit visits that do recover the guide stars, the majority (∼78%) do so before the beginning of science exposures in the second orbit. We also provide advice for users on how to make a risk assessment based on the analysis presented here.

The correction formula for Charge Transfer Efficiency (CTE) that is used in the HSTCAL CALSTIS pipeline CCD data reduction has not been significantly updated since 2006. Correcting for CTE losses is crucial to the goal of 1% precision in the STIS spectrophotometric fluxes that are the basis for all HST and JWST flux calibrations. Precision in the CTE correction is especially relevant for the faintest flux standards, where the amount of correction can exceed 20%. The comparison of new datasets of STIS spectra and ACS photometry of faint stars reveals the required updates to the parameters of the STIS CTE correction formula. After replacing the original with the new parameters, the change in the spectral energy distribution (SED), i.e. the flux decrease in physical units, for a very faint star NGC2506-G31 ranges from 4 to 6% over most of the G430L and G750L spectral wavelength ranges. No observations of the faint stars with G230LB or the medium dispersion modes were made; but the CTE correction depends only on signal level and should apply to all CCD spectroscopic modes. With the new formulation for the STIS CTE correction the STIS, ACS, and WFC3 flux measures are in accord at the 1% level, not only for the primary standards and other stars in the neighborhood of V=13 and brighter, but also now for fainter stars in the V=16 range.