Question and Answer

Question ID

0044

Language

English

Assigned Member

K. Fukushima

Category

resolve analysis

Name

Hayden Hall

Institute

Wayne State University

Question

Question detail: In doing our spectral analysis of the Hp-only events of ObsID:201073010 on the low-mass X-ray binary Serpens X-1, we have come across a discrepancy between the spectral shape of data from different subarrays on the Resolve detector. Above 8 keV, there appears to be a spectral change between the inner and outer pixel regions of the detector at greater than a 30% level (our target was centered on pixels 0,17,18,35). The central 4 pixels produce a spectrum that is spectrally harder than the outer rings. What could be causing this spectral shape difference?

We checked if the subarray response files may be the culprit for the discrepancy, however, when fitting each spectrum with the same response files, the differences persisted. This leads us to believe the difference in spectral shape is not due to the response files of the subarrays.

The first plot in the attached pdf shows the ratio plot from each detector region with its associated response files. The following four plots again show the ratio but now all data sets share the same response files to investigate whether/how the response files may change spectral shape or fitting.

Answer from

K. Fukushima

Institute

JAXA/ISAS

Answer

This is recognised as the "Pixel-to-Pixel Spectral Differences" issue.
The Resolve spectra obtained from various pixels occasionally have significantly distinct spectral shapes in the broadband range. The origin of this phenomenon is currently under investigation, although no consensus has been reached at this moment. Additional information is summarised on the webpage "Things to Watch Out for with XRISM Data Processing and Analysis" (RSL-4: RMF): https://heasarc.gsfc.nasa.gov/docs/xrism/analysis/ttwof/index.html.

Considering the source's brightness and the substantial variation in Grade fractions across the array, one potential issue is the ARFs: although the effective area for each pixel should be adjusted according to the Hp fraction per pixel, the current software is incapable of processing this, resulting in potentially inaccurate ARFs for sub-array regions. For variable sources, the issue is more complex since the Hp fraction might fluctuate due to a time-dependent distortion of the PSF.

The following are the recommendations from the XRISM Science Data Center:

1. With the current software what you can do is make ARFs & RMFs for the 4 *individual* central pixels. To get the relative normalizations correct, you will have to make the RMFs based on event files with only single pixels, and PI selections (3 to 10 keV), to accurately get the Hp fractions (and do it with and without Ls events included). Do these 4 spectra and fits provide consistent results with each other?

2. Restrict spectral fitting to below ~8 keV if you are using more than one pixel in a spectral group.

Status

Accept 2025-07-17

Close 2025-08-02