Question and Answer

Question ID

0046

Language

English

Assigned Member

S. Ogawa

Category

resolve analysis

Name

Maxime Parra

Institute

Ehime University

Question

Dear XRISM Helpdesk,

Following the release of Heasoft 6.36, I have reanalyzed observations of a bright source I am working on with the new tasks, and notably rslmkrsp, to create appropriately weighted responses for cases with spatially dependent grade fractions. I am doing this for several pixel groups in my observation, which I then fit together with different models, as there are several sources in the FoV.

In Heasoft 6.35.1, the responses were consistent between the different pixel groups.

In Heasoft 6.36, the normalizations are higher, but also (worryingly) inconsidtent: I get up to +/-30% normalization difference between the center pixels of my PSF and some further pixels, and this difference varies wildly for different arf choices.

I wonder if this may be due to the edge pixelss being dominated by the branching ratios of the central source.

I can provide the details of the runs of rslmkrsp I used, but I believe I applied only standard options. I'm making Hp+Mp products, using X RMF sizes.

I have several ideas on how to mitigate the error, such as restricting the rslmkrsp grade computation range to the band where a given source dominates when making the response of that source. However, I do not know if this is scientifically valid, and the band where each source dominates is not the band that is most central to my scientific analysis.

Are there some guidelines on how to combine rslmkrsp optimally when spatial-spectral mixing is prominent in bright source observations ?

In addition, would it be possible to get put in contact with the specific XRISM team members working on the tool ?

I would like to present the observation, spectra, and my scientific objectives in details, in order to get the most appropriate answer.

Answer from

Institute

Answer

Status

Accept 2025-10-03