Updating the Sensitivity Curves of the STIS Echelles (Post-SM4)

The STIS team has updated the on-orbit sensitivity curves for the instrument's echelle modes following Servicing Mission 4 by utilizing improved CALSPECv11 flux models for the G 191-B2B standard star and releasing new blaze shift coefficients and ripple tables.

The Gist

Imagine the Hubble Space Telescope as a giant, cosmic camera that has been taking pictures of the universe for decades. One of its most powerful lenses is called STIS (the Space Telescope Imaging Spectrograph). Think of STIS not just as a camera, but as a prism that takes starlight and splits it into a rainbow of colors (a spectrum) so astronomers can read the chemical "fingerprints" of stars and galaxies.

However, like any camera, STIS isn't perfect. Over time, its "lens" gets a little dusty, its focus shifts, and the way it measures the brightness of light changes. To get accurate scientific data, astronomers need to constantly update the "instruction manual" (calibration) for how the camera sees the world.

This new report, STIS 2024-04, is essentially a major software and hardware update for the STIS prism, specifically for the high-resolution "zoom" modes (called Echelle modes) used after the telescope's last major repair mission in 2009.

Here is the breakdown of what they did, using some everyday analogies:

1. The "Standard Candle" Problem (The New Ruler)

To measure how bright a star is, astronomers need a "standard candle"—a star whose true brightness is known perfectly, so they can compare everything else to it. For Hubble, this standard star is G 191-B2B, a white dwarf that acts like a cosmic lightbulb.

• The Old Way: For years, the team used an old model (CALSPECv07) of this lightbulb. It was like using a ruler that was accurate to the nearest centimeter.
• The New Way: In 2020, scientists updated the model of this star (CALSPECv11). It's like swapping that ruler for a laser-measured one accurate to the nearest millimeter. They found that the star was actually slightly brighter (by 1–3%) in certain colors than we thought.
• The Result: Because the "ruler" changed, every single measurement Hubble took with this prism needed to be re-measured to match the new, more precise standard.

2. The "Shifting Spotlight" (Blaze Shifts)

The STIS prism works by bouncing light off a special grating (like a CD surface). This grating has a "sweet spot" where it reflects light most efficiently, called the blaze.

• The Problem: Over time, this sweet spot drifts. Imagine a spotlight on a stage that slowly moves a few inches to the left every year. If you don't adjust your camera, the actor (the star) will look dimmer or brighter than they actually are because the light is hitting the wrong part of the lens.
• The Fix: The team calculated exactly how much this "spotlight" has shifted over the last 15 years. They created new mathematical rules (coefficients) to tell the computer: "Hey, in 2010, the light was here; in 2024, it's moved there. Adjust the brightness accordingly."
• The Surprise: They found that around 2018, the speed of this shift changed slightly (like a car changing gears). They had to create two different sets of rules: one for before 2018 and one for after.

3. The "Static Noise" (Ripple Tables)

When light hits the detector, it doesn't just create a smooth rainbow; it creates a faint, wavy pattern of noise (like static on an old TV). This is called the ripple.

• The Fix: The team created a new "noise map" (Ripple Table) to subtract this static from the data. It's like putting on noise-canceling headphones to hear the music clearly. By removing this static more accurately, the final picture of the star is much cleaner.

4. The "Edge Cases" (Recovering Lost Orders)

The prism splits light into many different "orders" (like different pages of a book). Some of the pages at the very top and bottom of the detector were previously ignored because they were hard to read or seemed too noisy.

• The Fix: With the new, cleaner "noise maps" and better rulers, the team realized they could actually read those edge pages now. They recovered several "lost" pages of the spectrum, giving astronomers more data than they had before.

Why Does This Matter?

Before this update, if you measured the brightness of a star, you might have been off by 2–3%. That's like guessing the weight of a person and being off by 5 pounds.

After this update:

• The measurements are now accurate to within 1%.
• The "noise" is gone.
• The "spotlight" is perfectly aligned.

In Summary:
The STIS team took the telescope's high-resolution prism, updated its "ruler" to match a new, more precise standard star, fixed the "drifting spotlight" that had moved over the last decade, and cleaned up the "static noise." The result is a crystal-clear view of the universe that allows astronomers to measure the cosmos with unprecedented precision. It's a massive upgrade that ensures the data we get from Hubble today is as reliable as possible for the next generation of discoveries.

Technical Summary

Here is a detailed technical summary of Instrument Science Report STIS 2024-04, titled "Updating the Sensitivity Curves of the STIS Echelles (Post-SM4)."

1. Problem Statement

The Space Telescope Imaging Spectrograph (STIS) on the Hubble Space Telescope (HST) requires precise flux calibration to convert observed counts into physical flux units. Previous calibrations for the STIS echelle modes (E140M, E140H, E230M, E230H) relied on the CALSPECv07 standard star models, specifically for the white dwarf G 191-B2B.

In 2020, the CALSPECv11 database was released, introducing significant updates to the atmospheric models of primary standard stars (GD 71, GD 153, and G 191-B2B). Key changes included:

• The inclusion of metal-line-blanketed model atmospheres for G 191-B2B, which redistributed flux as a function of wavelength.
• Updates to the absolute fluxes of primary standards by approximately 1–3%.
• A re-examination of the Vega spectral flux.

These updates necessitated a complete re-derivation of STIS echelle sensitivity curves, throughput files, and related calibration parameters to ensure data taken after Servicing Mission 4 (post-SM4) is consistent with the new absolute flux standards. Additionally, previous calibrations had issues with time-dependent shifts in the blaze function and the exclusion of certain spectral orders at the detector edges.

2. Methodology

The STIS team re-derived the calibration parameters using a multi-step iterative process based on archival observations of G 191-B2B (Program IDs 11866 and 15381) and BD+28°4211.

A. Data Sources and Reduction

• Observations: Data from Cycle 17 (PID 11866, 2009–2010) and a 2018 calibration program (PID 15381) were used.
• Calibration Software: Data were reduced using CALSTIS v3.4.2.
• Reference Models: The CALSPECv11 model for G 191-B2B (g191b2b_mod_011.fits) served as the absolute flux reference. Unlike the previous v07 model, v11 includes metal absorption lines and accounts for the star's radial velocity (22.0 km/s).

B. Derivation Steps

1. Sensitivity Curves ($S(\lambda)$):
   • NET counts from standard star exposures were divided by the interpolated CALSPECv11 model SED.
   • Masking: Regions affected by stellar absorption lines, interstellar medium (ISM) features, Lyman $\alpha/\beta$, and detector vignetting were masked to derive a clean continuum sensitivity.
   • Fitting: Multi-node quadratic splines were fitted to the masked data to generate smooth sensitivity curves.
2. Ripple Function (RIPTAB):
   • Sensitivity fits were normalized to create updated ripple functions for each spectral order.
   • These functions generate 2D scattered light models used to subtract background noise from raw images.
   • An iterative loop was performed: update RIPTAB $\rightarrow$ re-calibrate data $\rightarrow$ re-derive sensitivity $\rightarrow$ update RIPTAB again to ensure self-consistency.
3. Throughput Conversion (PHOTTAB):
   • Sensitivities were corrected for time-dependent sensitivity changes (using TDSTAB).
   • Converted from aperture-specific units to infinite aperture/box efficiency units using standard reference files and HST collecting area corrections.
4. Blaze Shift Coefficients:
   • The team derived new time-dependent coefficients ($BSHIFT\_VS\_T$ and $BSHIFT\_OFFSET$) for the blaze function equation to correct for shifts in the wavelength scale over time.
   • E230M: A subtle inflection in the linear trend of pixel shifts was detected around 2018.3. Separate coefficients were derived for periods before and after this date. An "extra" pixel shift specific to G 191-B2B observations was identified and removed to improve relative flux agreement.
   • E230H: New coefficients were derived for the primary setting (2263). For the secondary setting (2713), an empirical approach was used to optimize the offset term.
5. Recovery of Edge Orders:
   • Spectral orders previously excluded from PHOTTAB files due to their location on detector edges were re-included. While background subtraction is slightly more uncertain for these orders (due to one background region falling off the detector), the flux calibration is now available for them.

3. Key Contributions

• CALSPECv11 Integration: The first comprehensive update of STIS echelle sensitivities utilizing the metal-line-blanketed models of the CALSPECv11 library.
• Blaze Shift Refinement:
  • Discovery of a non-linear trend (inflection) in E230M blaze shifts around 2018.3, leading to the delivery of two sets of coefficients with different USEAFTER dates.
  • Identification and correction of a systematic "extra" pixel shift in G 191-B2B observations that had previously limited relative flux accuracy by ~2%.
• Expanded Coverage: Re-inclusion of previously excluded spectral orders at the top and bottom edges of the detector into the calibration reference files.
• New Reference Files: Release of updated PHOTTAB (throughput), RIPTAB (ripple), and BLAZETAB (blaze shift) files for the prioritized echelle settings.

4. Results and Validation

• Residual Analysis: The new calibration was validated by comparing calibrated STIS spectra of G 191-B2B against the CALSPECv11 model.
  • Accuracy: The average percent residuals for all updated modes are within 1% (standard deviation ~1–2%), a significant improvement over the previous 2–3% residuals.
  • FUV/NUV Accuracy: Estimated absolute flux calibration accuracies are 6% for the Far-UV (FUV) and 4% for the Near-UV (NUV), accounting for instrumental uncertainties and model differences.
• Flux Agreement:
  • Figure 2 and Figure 4 demonstrate that the new calibration significantly reduces flux mismatches (up to ~10% previously) in overlapping spectral regions of adjacent orders.
  • The removal of the "extra" G 191-B2B shift improved the relative flux agreement for E230M/2707 observations.
• Consistency: Testing against Time-Dependent Sensitivity (TDS) monitoring data showed flux variations of $\lesssim$ 5%, well within expected photometric accuracy limits.

5. Significance

This report represents a critical update for the HST STIS instrument, ensuring that scientific data taken post-Servicing Mission 4 is aligned with the most current and accurate astronomical flux standards.

• Scientific Impact: The ~1–3% flux corrections and improved relative flux agreement (reducing errors from ~2–3% to ~1%) are crucial for precise abundance studies, spectral energy distribution (SED) modeling, and comparative studies of astronomical objects.
• Operational Impact: The identification of the 2018.3 inflection point in E230M blaze shifts prevents systematic errors in long-term monitoring programs. The inclusion of edge orders maximizes the usable spectral range of the instrument.
• Legacy: These updates ensure the longevity and reliability of STIS data, allowing it to remain a premier tool for high-resolution spectroscopy in the UV regime.

Note: These updates apply specifically to post-SM4 data. Pre-SM4 data requires a different calibration strategy due to different spatial and spectral offsets, which is addressed in a separate report (Siebert et al., 2024).