Primary-source closure: Ponti Table 4 and Ueda disk-fit domain

Published

July 13, 2026

Result

This audit closes two previously incomplete primary-source checks for the current absorbed 0.5–2.0 keV Figure 3. The exact publisher full text and version-of-record PDF were retrieved on 2026-07-13T16:20:29Z.

  • Ponti et al. (2023) Table 4 directly supplies the disjoint CGM-component surface brightnesses used in the eFEDS conversion.
  • Ueda et al. (2022) must be described with two nested domains: the parent 130-sightline sample and the narrower Table 3 row from which the adopted disk normalization was taken. The current numerical projection remains valid for all 14 M31 fields, but the narrower fit selection is now stated explicitly.

Ponti et al. (2023): eFEDS Table 4

Citation: G. Ponti et al., A&A 674, A195, DOI 10.1051/0004-6361/202243992.

Primary sources:

Table 4 is in Sect. 9.2 on physical PDF page 16, labeled “A195, page 16 of 21.” It labels the quantity as component surface brightness in units of 10^-13 erg cm^-2 s^-1 deg^-2. Its note states that each left/right pair is the same model under negligible/high SWCX, respectively.

CGM component Negligible SWCX (left) High SWCX (right)
0.3–2.0 keV 29.7 20.5
0.3–0.6 keV 24.1 15.6
0.6–2.0 keV 5.6 4.9

Sect. 5.4 defines the left branch by completely neglecting SWCX. The right branch fixes a high-contamination scenario from the e3–e1 difference spectrum, whose anchor is (6.9 +/- 1.5)e-13 erg cm^-2 s^-1 deg^-2 in 0.4–0.6 keV. It is not an independent Table-4 detection. The Table-4 SWCX value 7.6 is instead integrated over 0.3–0.6 keV, so the two numbers are not contradictory.

Sect. 9 calls the totals the mean surface brightness observed by eROSITA. Table 4 is therefore an observed/best-fitting sky decomposition, not a second intrinsic or absorption-corrected table. The paper models the neutral-column distribution with disnht, identifies the LHB as unabsorbed, and applies the full Galactic absorption to the CXB. Using the tabulated CGM value as an attenuated line-of-sight component is an inference from that model structure; the Table 4 note itself does not repeat the word “absorbed” or separately define an intrinsic CGM flux.

The primary products disagree on the disnht width: the PDF and arXiv source give 0.117, whereas publisher HTML Equation 1 renders 0.177. The current reconstruction uses 0.117, matching both the version-of-record PDF and arXiv source. CXB-PL is an alternative steep-power-law fit component, not an extra term to add to the Table-4 total.

The current conversion uses only disjoint published bands. For each fixed SWCX scenario it evaluates the corresponding published kT/Z CGM model to obtain F(0.5--0.6)/F(0.3--0.6), then adds the published 0.6–2.0 value. The resulting absorbed 0.5–2.0 points are:

  • negligible SWCX: 0.4065102009
  • high SWCX: 0.2912081084

in Figure-3 units of 10^-15 erg cm^-2 s^-1 arcmin^-2. Their span is a discrete scenario bracket, not a confidence interval.

Ueda et al. (2022): parent domain versus adopted fit row

Citation: M. Ueda et al., PASJ 74, 1396–1414, DOI 10.1093/pasj/psac077.

Primary sources:

Three source-defined selections

  1. Parent observational sample: the abstract states, “130 Suzaku observations at 75° < l < 285° and |b| > 15°,” obtained from 2005–2015.
  2. High-latitude uniform subset: the abstract separately describes data before the end of 2009 at |b| > 35° and 105° < l < 255°; this supports the quoted nearly uniform kT = 0.22 keV and EM = 2e-3 cm^-6 pc. It is not the disk-normalization fit row used here.
  3. Adopted disk-fit row: Table 3 gives 2005--2009, |l| > 105°, N=36, n_e0 = (3.4 +/- 0.1)e-3 cm^-3, fixed z0 = 2.7 kpc, and chi2/d.o.f. = 36/35. The adjacent text states that R0=7.0 kpc and z0=2.7 kpc were fixed; the unrestricted 2005–2009 disk fit has chi2/d.o.f. = 389/63, whereas excluding |l| < 105° gives the good fit.

Within the parent longitude window, the adopted Table 3 angular selection is approximately 105° < l < 255°. The 14 M31 fields span 119.508° <= l <= 122.303° and 19.888° <= |b| <= 23.167°; therefore all are inside both the parent sample and adopted disk-fit selection. They are not in the separate |b| > 35° high-latitude uniform subset, and the manuscript no longer conflates these selections.

Ueda et al. define EM_halo = integral n_e n_H ds in cm^-6 pc, identify n_e, n_H, and s as electron density, hydrogen density, and line-of-sight distance, and label the Table 3 normalization as n_e0. The MWH EM is an intrinsic plasma integral inferred through an absorbed phabs spectral model; it is not itself an observed surface brightness.

The paper’s disk equation prints exp(-z/z0) rather than exp(-|z|/z0) and calls n a generic number density, while Table 3 reports central electron density. It does not explicitly provide the n_e/n_H conversion, the internal observer transform from (l,b,s) to (R,z), or the adopted Solar radius. The following are therefore current projection augmentations, not verbatim Ueda definitions:

  • impose reflection symmetry with exp(-|z|/z0);
  • adopt n_e/n_H = 1.2;
  • adopt a Solar Galactocentric radius of 8.2 kpc and the explicit LOS geometry.

With those additions, the implementation uses n_e = 3.4e-3 exp(-R/7.0 kpc) exp(-|z|/2.7 kpc) cm^-3 and n_H = n_e/1.2.

The resulting 14-field absorbed 0.5–2.0 keV projection is unchanged:

  • field range: 0.2894675688--0.3206666288
  • observed inverse-variance estimator: 0.3119042867
  • conditional M31 complement: 0.6528222800 (displayed as 0.653)

The footprint range and the marginal n_e0 sensitivity are not a joint posterior interval.

Retrieved-source hashes

These hashes identify the exact retrieved primary-source bytes used in this audit; publisher content may be updated later.

Retrieved object SHA-256
Ponti full HTML 50353bf63624fd9e0662197b045f09147c4e3b34df2a551aba8f73bce773193d
Ponti Table 4 HTML 1af5d064f4226a3c29d94a521faa731c07c19a0cdfec2be77d9e81b4a75294d9
Ponti version-of-record PDF 7a41cc588c93c81a4271453b6e21e8aebd19a62adf86754fae4ad408c610d8db
Ueda full HTML e6a2ccdd8c20018505d82afeed58e1a4a98c807012e02cbbe3018af1bfba62be
Ueda version-of-record PDF 686e636e53e190ed3d33c4a1371c22b5dcc7bc5a035993c6813ff3d0f40ded0f
Ueda repeated PDF retrieval 1a1ad3fb99292d30cdf6f4dadf34499e830f0eef7c523b59c0385ba69731db87
Ueda pdftotext -layout from both PDF downloads 990f2297a73270b28c93c6719d85f2f65b6cd91cc7f66a981e47062fe0efd86e

The two Ueda PDF downloads have the same size and byte-identical extracted text but differ in container SHA-256, consistent with publisher-side PDF metadata variation. The text hash above closes the scientific-content check.