🔭 astrophysics

XRISM spectroscopy of a crowded Galactic center region -- I. Disentangling the sources in the field of view

This paper presents the first study in a series analyzing XRISM, XMM-Newton, and NuSTAR observations of the crowded Galactic center region to develop robust methodologies for disentangling the overlapping X-ray emissions from the transient MAXI J1744-294, the outbursting neutron star AX J1745.6-2901, and various diffuse background sources.

Original authors: Maxime Parra, Kai Matsunaga, Shifra Mandel, Kaya Mori, Hideki Uchiyama, Masayoshi Nobukawa, Tahir Yaqoob, Takayuki Hayashi, Misaki Mizumoto, Shinya Yamada, Megumi Shidatsu, Paul A. Draghis, Efrain Gat

Published 2026-03-31

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Maxime Parra, Kai Matsunaga, Shifra Mandel, Kaya Mori, Hideki Uchiyama, Masayoshi Nobukawa, Tahir Yaqoob, Takayuki Hayashi, Misaki Mizumoto, Shinya Yamada, Megumi Shidatsu, Paul A. Draghis, Efrain Gatuzz, John A. Tomsick, Charles J. Hailey, Chichuan Jin, Benjamin Levin, Gabriele Ponti, Mark Reynolds

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 Cosmic Cocktail Party: Untangling a Messy X-Ray Signal

Imagine you are at a massive, crowded cocktail party in the center of a galaxy. The room is packed with thousands of people (stars), a few VIPs (black holes and neutron stars), and a thick fog of smoke (diffuse gas) that fills the entire space. Everyone is talking at once, shouting, whispering, and singing different songs.

Now, imagine you are a sound engineer trying to record a specific conversation between two guests: MAXI J1744-294 (a black hole that just started shouting) and AX J1745.6-2901 (a neutron star nearby that is also shouting).

The problem? You are using a microphone (the XRISM telescope) that is incredibly sensitive to pitch (energy) but has a very wide, blurry lens (angular resolution). When you point it at the VIPs, the microphone picks up their voices, but it also picks up the voices of everyone standing right next to them, plus the echo of the room itself.

This paper is the "Sound Engineering Manual" for that specific night at the party. It explains how the team figured out how to separate the two VIPs' voices from the chaotic background noise so they could actually hear what they were saying.

The Setting: The Galactic Center

The Galactic Center is the busiest neighborhood in our Milky Way. It's home to Sgr A*, the supermassive black hole that acts like the party host. Around it, there are:

Supernova Remnants (Sgr A East): Like the lingering smell of a fireworks display from a party that happened years ago.
Diffuse Gas (GCXE): The "fog" of hot gas that fills the room, glowing with its own light.
Thousands of other stars: The background chatter.

In March 2025, a new "guest" arrived: MAXI J1744-294, a black hole that suddenly went into "outburst" mode (it started eating a lot and glowing very brightly). The astronomers wanted to study its "voice" (specifically, the iron lines in its X-ray spectrum) to understand what was happening.

The Problem: The "Blurry Lens"

The XRISM telescope is amazing. It can hear the tiniest differences in pitch (energy resolution), which is like being able to distinguish between a violin and a flute perfectly. However, it can't tell exactly where the sound is coming from very well (poor angular resolution).

Because the telescope's "lens" is blurry, the light from the black hole (MAXI J1744-294) gets mixed with:

The neutron star (AX J1745.6-2901) right next door.
The glowing gas (Sgr A East) behind them.
The general background glow of the galaxy.

It's like trying to hear one person speak while standing in the middle of a stadium where everyone else is shouting, and your microphone is so sensitive it picks up the echo from the walls too.

The Solution: Two Tricks to Clean the Signal

The team used two main strategies to untangle this mess, acting like a detective using two different types of clues.

1. The "Big vs. Small" Region Strategy

They looked at the data in two different ways:

The "Big" Region (The Crowd Approach): They grabbed a large chunk of the sky. This gave them a lot of data (a loud signal), but it was very "dirty" because it included a lot of the background noise. To fix this, they used Empirical Modeling. Think of this as saying, "We don't know exactly what every background noise is, so let's just mathematically describe the shape of the noise and subtract it." It's like using noise-canceling headphones that learn the pattern of the crowd and cancel it out without needing to know who is saying what.
The "Small" Region (The VIP Approach): They zoomed in on just the four brightest pixels right next to the black hole. This area was "cleaner" (less background noise) but had less total data. Here, they used Physical Modeling. They tried to build a physical model of the background (e.g., "This part of the noise is definitely from the supernova remnant, and this part is from the hot gas") and subtracted those specific ingredients.

By comparing the results from both the "Big" and "Small" approaches, they could be sure they weren't just making up the answer. If both methods gave the same result, they knew they were on the right track.

2. The "Time Travel" Strategy

The team realized they couldn't just look at the 2025 data alone. They needed a "control group."

They went back in time to March 2024, when they had observed the same patch of sky before the black hole (MAXI J1744-294) started shouting.
In 2024, the black hole was quiet. So, the 2024 data showed them exactly what the background noise (the gas, the other stars, the neutron star) looked like without the new black hole.
They used this 2024 "background map" to subtract the noise from the 2025 data. It's like taking a photo of a messy room, then taking a photo of the same room after someone dropped a giant pile of confetti. By comparing the two, you can isolate exactly where the confetti is.

The Result: Clearing the Air

By using these clever tricks, the team successfully separated the signal of the black hole from the neutron star and the background gas.

They created a "clean" version of the black hole's spectrum.
They created a "clean" version of the neutron star's spectrum.
They created a "clean" map of the gas in the galaxy.

Why This Matters

This paper is the foundation. It doesn't tell you the final story of what the black hole is doing (that's coming in "Paper II"). Instead, it explains how they cleaned the data so that the next papers can tell the story.

Without this work, the black hole's voice would have been lost in the noise, and astronomers might have thought the background gas was part of the black hole, or vice versa.

In short: This paper is the recipe for how to filter out the static on a radio so you can finally hear the music. It shows that even in the most crowded, messy parts of the universe, with the right tools and a little bit of clever math, we can isolate individual stories from the cosmic chaos.

1. Problem Statement

The Galactic Center (GC) is an extremely crowded X-ray region containing the supermassive black hole Sgr A*, numerous accreting compact objects, and complex diffuse emission. The paper addresses the specific challenge of analyzing the X-ray transient MAXI J1744-294 (located $\sim18''$ from Sgr A*) using data from the XRISM satellite (launched 2023), which offers high spectral resolution but limited angular resolution.

Key Challenges:

Spatial-Spectral Mixing: XRISM's Point Spread Function (PSF) is large relative to the pixel size, causing significant blending of emission from multiple sources within a single pixel.
Source Contamination: The field of view (FoV) contains:
- The target transient MAXI J1744-294 (Black Hole candidate).
- The nearby neutron star low-mass X-ray binary AX J1745.6-2901, which was also in outburst during the observation.
- Diffuse emission from the supernova remnant Sgr A East.
- The Galactic Center X-ray Emission (GCXE) and unresolved point sources.
Dust Scattering Halos (DSH): Dust around the sources scatters X-rays, altering the spectral shape, particularly for off-axis regions.
Instrumental Limitations: The need to distinguish intrinsic source features from background contamination and instrumental effects (e.g., pile-up, gain shifts) in a crowded field.

2. Methodology

The authors employed a multi-instrument approach combining XRISM (Resolve and Xtend), XMM-Newton, and NuSTAR (data from a simultaneous campaign in March 2025) with archival data from a 2024 Performance Verification (PV) phase observation.

A. Data Reduction and Calibration

Software: Used Heasoft 6.36 and XRISM CALDB 11 (avoiding CALDB 12 due to known optical axis tilt errors).
Resolve (Microcalorimeter):
- Addressed event grade branching ratios (High, Middle, Low resolution) due to high count rates.
- Generated custom Response Matrix Files (RMF) and Ancillary Response Files (ARF) using rslmkrsp to account for spatial variations in event grades.
- Applied strict proximity and rise-time cuts to remove noise.
Xtend (CCD Imager):
- Removed "flickering pixels" using a flat-top filter.
- Verified negligible pile-up for the target in Burst mode.
Region Selection:
- Defined "Big" (21 pixels) and "Small" (4 brightest pixels) regions for MAXI J1744-294 to test the trade-off between Signal-to-Noise Ratio (SNR) and contamination.
- Defined specific regions for AX J1745.6-2901 and background estimation based on PSF fraction modeling.

B. PSF Modeling and Background Subtraction

PSF Fraction Mapping: Used ray-tracing simulations (xaxmaarfgen) to calculate the contribution of each source (MAXI J1744-294, AX J1745.6-2901, Sgr A East, GCXE) to every pixel in the FoV.
Background Modeling: Leveraged the 2024 PV observation (where MAXI J1744-294 was inactive) to model the underlying diffuse emission and the contribution of AX J1745.6-2901.
Dust Scattering: Acknowledged that while DSH affects the continuum, its impact on narrow emission lines in XRISM is minimal. Physical models for DSH were applied to XMM-Newton/NuSTAR data but not yet fully adapted for XRISM.

C. Spectral Modeling Strategies

The authors utilized two complementary approaches to model the diffuse emission and AX J1745.6-2901 contamination:

Empirical Modeling (for "Big" Region):
- Used a single absorbed powerlaw for the continuum.
- Modeled emission/absorption lines using Gaussian and Lorentzian profiles without strict physical constraints.
- Effective for averaging out PSF mixing in large regions.
Physical Modeling (for "Small" Region):
- Modeled Sgr A East and GCXE separately using overionized plasma models (bvvrnei in XSPEC) and powerlaws.
- Used photoionization models for AX J1745.6-2901 (imported from previous work) but switched to an empirical model for the DDT observation due to anomalous line profiles.
- Allowed for spatial variations in spectral brightness.

3. Key Contributions

Robust Background Subtraction Framework: Developed a rigorous method to disentangle overlapping sources in XRISM data by combining high-resolution archival imaging (Chandra) with XRISM's spectral capabilities.
Dual-Approach Validation: Demonstrated that both empirical (line-by-line) and physical (plasma model) approaches yield consistent results for the underlying diffuse emission, validating the reliability of the background subtraction.
PSF Fraction Quantification: Provided detailed maps of PSF fractions for multiple sources across the XRISM FoV, enabling precise extraction of source spectra despite the instrument's limited angular resolution.
Calibration Refinement: Addressed specific XRISM pipeline issues, including energy scale shifts and resolution degradation at high count rates, confirming these effects are negligible for the specific count rates observed in this campaign.

4. Results

Successful Disentanglement: The study successfully isolated the spectral contributions of MAXI J1744-294, AX J1745.6-2901, and the diffuse background (Sgr A East and GCXE).
Background Characterization:
- The "Big" region empirical model achieved a global C-statistic/d.o.f. of 33214/31850 with no significant residuals in the 2–10 keV band.
- The "Small" region physical model achieved a C-statistic/d.o.f. of 32379/31970.
- The column density ( $N_H$ ) for AX J1745.6-2901 was refined to $2.47 \times 10^{23} \text{ cm}^{-2}$ in the physical model, differing from the fixed $3 \times 10^{23} \text{ cm}^{-2}$ used in the empirical model, likely due to dust scattering halo effects.
Line Feature Recovery: The modeling confirmed that narrow emission lines (specifically Fe K $\alpha$ complexes) could be recovered with high fidelity, as the background contamination was effectively modeled and subtracted.
Instrument Performance: Confirmed that XRISM/Resolve can handle crowded fields if accompanied by careful PSF modeling and multi-epoch background observations.

5. Significance

Foundation for Future Science: This paper serves as the technical prerequisite (Paper I) for subsequent studies (Paper II and III) that will analyze the intrinsic physics of MAXI J1744-294 (e.g., outflow velocities, wind properties) and the interstellar medium.
Benchmark for Crowded Fields: It establishes a benchmark for analyzing XRISM data in the Galactic Center and other crowded regions, demonstrating that high-spectral-resolution science is possible even when angular resolution is limited, provided rigorous background modeling is applied.
Methodological Advancement: The combination of empirical and physical modeling, coupled with detailed PSF fraction mapping, offers a template for future X-ray missions dealing with complex, blended source environments.
Scientific Output: The work highlights the potential of XRISM Director's Discretionary Time (DDT) observations to uncover unexpected phenomena in complex astrophysical environments.