🔭 astrophysics

Recalibrating the Sensitivities of the STIS First-Order, Medium-Resolution Modes

This paper details the recalibration of STIS first-order, medium-resolution sensitivities using on-orbit observations and updated CALSPEC models, describing the verification procedures for new PHOTTAB reference files activated in CRDS on May 1, 2025, while presenting results, discussing uncertainties, and offering recommendations for future work.

Original authors: Alex Fullerton

Published 2026-02-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Alex Fullerton

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Great Telescope Tune-Up: A Story of STIS and the "Sensitivity" Fix

Imagine the Hubble Space Telescope as a legendary, high-end camera that has been taking pictures of the universe for decades. One of its most important lenses is called STIS (the Space Telescope Imaging Spectrograph). Think of STIS not just as a camera, but as a prism. Instead of taking a picture of a star, it breaks the star's light apart into a rainbow (a spectrum) so scientists can read the chemical "fingerprint" of the star.

However, over time, even the best lenses get a little dusty, and the sensors get a bit tired. To make sure the colors in the rainbow are accurate, scientists need to know exactly how sensitive the camera is at every single color of light. This is called the "sensitivity curve."

This report, written by Alex Fullerton in 2025, is the story of how they went back and re-calibrated (re-tuned) this sensitivity for a specific set of settings on STIS.

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

1. The Problem: The "Old Map" vs. The "New GPS"

For years, scientists used an "old map" (models from 2005–2006) to interpret the light coming from stars. But recently, astronomers updated the "GPS" (the CALSPEC models) for the standard stars they use as reference points.

The Analogy: Imagine you are trying to measure the length of a room. For 20 years, you used a ruler that you thought was exactly 12 inches long. But then, you realized your ruler was actually slightly bent or stretched. You also realized the "standard foot" used by the government had been redefined.
The Result: The new "GPS" (Version 11 of the models) showed that the old "ruler" was off by up to 3% in some places. That might sound small, but in astronomy, 3% is a huge difference when you are trying to measure the temperature of a star or the amount of oxygen in a galaxy.

2. The Stars: The "Perfect Rulers"

To fix the telescope, you need a perfect ruler to compare it against. STIS uses three specific white dwarf stars (G191-B2B, GD 71, and GD 153) as its "Perfect Rulers." These stars are like the "standard kilogram" kept in a vault in France; we know exactly how much light they should be emitting at every color.

The Challenge: The telescope has many different "settings" (called CENWAVEs and gratings). It's like having a camera with 61 different zoom lenses and filters. The scientists had to check every single one of these 61 settings to see if the "Perfect Ruler" looked the same as the new GPS said it should.

3. The Process: Cleaning the Lens and Re-mapping

The scientists didn't just guess; they went through a rigorous process:

Gathering the Data: They went into the Hubble archive (a giant digital library) and pulled out all the old photos of these "Perfect Ruler" stars.
The "Fringe" Problem: For the reddest colors of light (long wavelengths), the telescope's detector (a CCD) creates a weird pattern of interference, like ripples in a pond. This is called "fringing."
- The Metaphor: Imagine taking a photo of a streetlight through a window covered in rain. The rain creates streaks that mess up the image. The scientists had to write special software to "dry off the window" (a process called defringing) so they could see the true light of the star.
The Math: They compared the light the telescope actually saw against the light the star should have emitted according to the new GPS. The difference between the two told them exactly how to adjust the telescope's settings.
Smoothing it Out: They used a mathematical tool called a spline (think of it as a flexible ruler that bends to fit a curve) to connect all the dots and create a smooth, perfect sensitivity curve for each setting.

4. The Result: A New "Instruction Manual"

Once they calculated the new curves, they updated the telescope's internal "instruction manual" (called PHOTTAB files).

The Activation: On May 1, 2025, these new instructions were turned on. Now, whenever a scientist points Hubble at a star using these specific settings, the computer automatically applies the new, more accurate math.
The Payoff: The data coming out of Hubble is now accurate to within about 2%. Before this, some measurements might have been off by a bit more.

5. The "Loose Ends"

The report also admits that they couldn't fix everything.

The Missing Data: For a few specific settings (about 19 of them), they didn't have enough good data from the "Perfect Ruler" stars.
The Metaphor: It's like trying to calibrate a specific zoom lens on a camera, but you lost the lens cap and never took a test photo with it.
The Solution: They left those specific settings on their old, potentially inaccurate settings. They recommend that if a scientist really needs to use those specific settings, they should be warned that the data might be a bit "stale."

Summary

In short, this paper is a maintenance log for one of the most important tools in astronomy. The team realized their "ruler" had changed, so they went back, cleaned off the "rain" on the lens, and re-measured everything against the new standard.

Because of this work, every time a scientist uses the Hubble Space Telescope to study the chemical makeup of the universe in the coming years, they can trust that the colors in their data are true, accurate, and ready to reveal the secrets of the cosmos.

1. Problem Statement

The Spectroscopic capabilities of the Hubble Space Telescope's Space Telescope Imaging Spectrograph (STIS) in its first-order, medium-resolution (M-mode) configurations rely on sensitivity curves derived from spectrophotometric standard stars.

Outdated Models: The previous calibration (last updated in 2005–2006) relied on CALSPEC model versions 4 and 5 for the primary white dwarf standards (G191-B2B, GD 71, GD 153).
Model Evolution: CALSPEC models have been significantly updated to Version 11 (Bohlin, Hubeny, & Rauch 2020), incorporating improved physics such as line blanketing in the UV. These updates introduce flux differences of up to 3% in certain wavelength regions compared to previous models.
Data Limitations: The existing sensitivity curves were derived from older data sets with varying signal-to-noise ratios (S/N) and lacked consistency with the new flux standards. Additionally, specific instrumental configurations (e.g., G140M "up" vs. "down" due to a 1999 grating tilt change) relied on different standard stars, creating potential systematic offsets.
Fringing: At longer wavelengths (CENWAVE > 7000 Å), interference fringes in the CCD detector complicated the reliable derivation of sensitivities.

2. Methodology

The authors performed a comprehensive recalibration of 61 primary and 19 secondary M-mode configurations using a Python-based pipeline that re-implemented the procedures from Proffitt (2006).

Key Steps:

Data Assembly: Relevant observations from MAST were retrieved, focusing on primary standards (G191-B2B, GD 71, GD 153) and the secondary standard (BD +75°325).
Reprocessing: Data were reprocessed through calstis with time-dependent sensitivity (TDS) and charge-transfer efficiency (CTE) corrections omitted. This allowed for the calculation of a correction factor ( $f_{corr}$ ) to normalize the count rates to the "beginning of life" (BoL) of the instrument.
Fringe Correction: For G750M configurations with CENWAVE > 7000 Å, a specialized fringe-correction procedure was applied using stistools.defringe to mitigate CCD interference patterns.
Sensitivity Calculation:
- Sensitivity $S(t)$ was calculated by dividing the corrected count-rate spectrum ( $C_{std}$ ) by the updated CALSPEC Version 11 model spectrum ( $F_{std}$ ).
- Multiple observations for the same configuration were merged, weighted by the inverse square of their errors.
Fitting: A weighted cubic spline fit was applied to the merged data to generate smooth sensitivity curves.
- Masking: Regions with strong spectral features (e.g., Lyman $\alpha$ ) were masked to prevent biasing the fit.
- Outlier Rejection: Sigma-clipping (3 iterations at $3\sigma$ ) removed outliers caused by resolution mismatches between data and models.
Reference File Generation: The new sensitivity curves were converted to throughputs and inserted into PHOTTAB reference files within the Calibration Reference Data System (CRDS).

3. Key Contributions

CALSPEC Version 11 Integration: The report establishes the first set of STIS M-mode sensitivities fully consistent with the latest CALSPEC models, ensuring flux accuracy across the 1150–10,000 Å range.
Unified Calibration: It addresses the historical split between "up" and "down" G140M configurations by utilizing data from both eras (GD 71 for "up" and G191-B2B for "down") to create consistent sensitivity curves where possible.
Secondary Standard Utilization: For the configuration G430M/4781, where no primary standard observations exist, the authors successfully utilized the secondary standard BD +75°325 (calibrated against a version 11-consistent L-mode spectrum) to derive a new sensitivity curve.
Fringe Mitigation: The implementation of a robust defringing procedure for long-wavelength CCD data significantly improved the reliability of sensitivity curves in the near-infrared.

4. Results

Coverage: The recalibration covers 61 primary and 19 secondary M-mode configurations across all STIS gratings (G140M, G230M, G230MB, G430M, G750M).
Accuracy: The median S/N of the input data is 47.7 (ranging from 9 to 142). The authors estimate the resulting flux accuracy to be ~2%, with potential localized excursions near Lyman $\alpha$ .
Verification:
- Internal Consistency: Overlapping wavelength regions between adjacent CENWAVE settings show excellent flux continuity (e.g., Figures 3–7).
- Model Agreement: Reprocessed standard star spectra using the new PHOTTAB files match the CALSPEC v11 models within the expected uncertainties.
- Cross-Check: Comparisons between G191-B2B and BD +75°325 (e.g., Figure 8) show good agreement, validating the use of secondary standards where necessary.
Unchanged Modes: Several secondary modes (indicated in bold in Table 1) were not updated due to a lack of suitable standard star observations. These retain their previous, potentially outdated calibrations.

5. Significance and Recommendations

Legacy Value: This recalibration ensures that the STIS archive remains scientifically robust for the remainder of the Hubble mission and for future legacy studies, correcting systematic biases introduced by outdated stellar models.
Operational Impact: The new PHOTTAB reference files were activated in CRDS on May 1, 2025, immediately improving the calibration of all new and re-reduced STIS M-mode data.
Future Work:
- The authors recommend acquiring new high-S/N (target ~100) observations for the remaining uncalibrated secondary modes.
- Observing multiple standard stars for modes covering Lyman $\alpha$ is suggested to better constrain model uncertainties.
- A comparison of "beginning of life" vs. "end of life" sensitivities is proposed to enhance the integrity of the long-term calibration record.

In summary, this report represents a critical update to the STIS spectrophotometric calibration, aligning the instrument's performance with modern stellar atmosphere models and ensuring the continued scientific utility of Hubble's medium-resolution spectroscopy.