Pushing the limits of the Cosmic Origin Spectrograph (COS) with an optimized background correction

Svea Hernandez,Andrei Igoshev,Jo Taylor,David Sahnow,Logan Jones

2024-06-11

Abstract:Observations utilizing the ultraviolet capabilities of the Cosmic Origin Spectrograph (COS) onboard the Hubble Space Telescope are of unique value to the astronomy community. Spectroscopy down to 900 A with COS has enabled new science areas. However, contrary to the situation at longer wavelengths, these observations are limited by detector background noise. The background correction currently applied by the standard calibration pipeline (CalCOS) is not optimized for faint targets, limiting the scientific value of low signal-to-noise observations. In this work we investigate a possible dependence of the variations of the dark rate in both segments of the COS far-ultraviolet (FUV) detector on time, detector high voltage (HV), and solar activity. Through our analysis we identified a number of detector states (on a configuration basis, e.g., HV and segment) characterizing the spatial distribution of dark counts, and created superdarks to be used in an optimized 2-dimensional (2D) background correction. We have developed and tested Another COS Dark Correction (ACDC), a dedicated pipeline to perform a 2D background correction based on statistical methods, producing background-corrected and flux-calibrated spectra. While our testing of ACDC showed an average improvement in S/N values of ~10%, in a few cases the improvements in S/N reached 60% across the whole wavelength range of individual segments.

Instrumentation and Methods for Astrophysics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: How to improve the signal - to - noise ratio (S/N) of faint celestial body observations in the far - ultraviolet (FUV) band by the Cosmic Origins Spectrograph (COS) on the Hubble Space Telescope (HST) through optimizing the background correction method. Specifically, when dealing with low - signal - to - noise ratio targets, the existing standard calibration pipeline (CalCOS) has insufficiently accurate background noise correction, which limits its scientific value. Therefore, this paper aims to develop a new background correction method to subtract background noise more accurately and thus improve the quality of observational data.

### Specific description of the problem

1. **Influence of background noise**:
- At shorter wavelengths (< 1130 Å), the sensitivity of COS drops significantly, making observations in these bands mainly limited by detector background noise.
- The background correction method currently used by the standard calibration pipeline (CalCOS) is not optimized enough for faint targets, resulting in limited scientific value for low - signal - to - noise ratio observations.

2. **Limitations of existing correction methods**:
- CalCOS determines the background contribution by calculating the average count in a predefined area, but this method may cause the dark current at the target location to be different from that in the predefined background area, resulting in over - correction or under - correction.
- In addition, differences between pixels may also lead to over - correction or under - correction of the dark current in the final spectrum.

3. **Objectives of improvement**:
- Develop a new background correction method that can model and subtract the expected dark current more accurately in two - dimensional (2D) scientific exposures.
- Improve the accuracy of low - signal - to - noise ratio (S/N ≲ 5) observations, especially in cases where background noise is limiting.

### Solution

The author proposes a new software tool named "Another COS Dark Correction (ACDC)", which achieves more accurate background correction through the following steps:

1. **Data analysis and model construction**:
- Collected and analyzed all COS FUV dark current exposure data from June 2009 to September 2023, with a total of 7,190 exposures.
- Identified the influence of factors such as different time periods, high - voltage (HV) settings, and solar activity on the dark current distribution.
- Created multiple "superdarks" to characterize background noise in different states.

2. **Multi - state background correction**:
- Adopted a data - driven approach to divide different states of the detector into several classes (2 - 10 states) and model the dark current distribution by linearly combining these states.
- Used the maximum - likelihood estimation method to optimize the background correction coefficients to ensure that the predicted value of the dark current for each pixel is closest to the actual observed value.

3. **Performance evaluation**:
- Tested the effect of ACDC on a series of low - signal - to - noise ratio observational data, and the results showed that the average signal - to - noise ratio was improved by about 10%, and in some cases even reached a 60% improvement.
- In particular, in the A - segment exposures with "glow" structures, the performance of ACDC was particularly remarkable.

### Conclusion

By optimizing the background correction method, ACDC can subtract background noise more accurately, significantly improve the signal - to - noise ratio of COS for faint celestial body observations in the far - ultraviolet band, and thus enhance the scientific value of these observations.