REVIEW 3 major objections 5 minor 9 references
A NICER Look at Strong X-ray Obscuration in the Seyfert-2 Galaxy NGC 4388
T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Obscuring gas around NGC 4388 halved in six years, new X-ray spectra show.
desk verdict NICER gives a clean, well-modeled column density for NGC 4388, but the headline variability versus NuSTAR rests on a 16-month average and is not yet secured. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The argument rests on two independent spectral model families fitted to the same 105.6 ks NICER spectrum. The first, pexmon, is built on a cold, neutral reflection slab (pexrav) with Fe and Ni K lines whose strengths are tied to the reflection fraction by atomic physics. The second, mytorus, is a self-consistent torus model with linked absorption, scattering, and line-emission components; the version used here decouples the line-of-sight (near-side) and face-on (far-side) views so the obscuring geometry and the reflecting geometry can be constrained separately. Both model constructions include the same XSTAR photoionized absorber in front of the intrinsic cutoff power law, three mekal plasma components for the soft X-ray emission, and a small scattered polar component. Agreement between the two families on $N_H$ is the load-bearing result; a relativistic blurring function (rdblur) is then added to the pexmon model to turn the Fe K line width into lower limits on the inner radius of the neutral reflector.
What would settle it
Re-observe NGC 4388 with NuSTAR and NICER simultaneously: if NuSTAR again measures a column near $5\times 10^{23}$ cm$^{-2}$ while NICER measures about $2.6\times 10^{23}$ cm$^{-2}$, the reported variability is an artifact of cross-calibration or model assumptions; if both instruments agree on the lower value, genuine variability is confirmed.
Extended reading notes
Core claim
Using a summed, time-averaged NICER spectrum of NGC 4388 with a net exposure of 105.6 ks, the paper finds that the intrinsic power-law continuum from the central engine passes through a Compton-thin neutral absorber with column density $N_H = 2.67^{+0.02}_{-0.03}\times 10^{23}$ cm$^{-2}$ under pexmon and $N_H = 2.64\pm 0.03\times 10^{23}$ cm$^{-2}$ under mytorus. These values are formally consistent, and both models find a small reflection fraction relative to the direct continuum, ruling out a Compton-thick line of sight. Compared with $N_H \simeq 4$--$6.5\times 10^{23}$ cm$^{-2}$ measured from 2013 NuSTAR data, the NICER value is about half as large, and the paper argues this difference is genuine variability because the same model families and procedures were used. The neutral Fe K$\alpha$ line is resolved at $\sigma = 40^{+10}_{-10}$ eV, corresponding to a velocity broadening around $1200^{+400}_{-400}$ km s$^{-1}$; interpreted as Keplerian motion, this places the line-forming region at roughly $10^4\,GM/c^2$, inside the optical BLR. The data also require a photoionized (XSTAR) absorber at the $4\sigma$ level, similar to warm absorbers seen in Seyfert-1 galaxies, and three soft thermal plasma components describe the low-energy spectrum.
Load-bearing premise
The models assume the obscuring and reflecting gas is cold and neutral with a single uniform column along the line of sight, so if the gas is partly ionized, as it would likely be if it sits in the broad line region, the reported column density could be systematically biased.
Editorial extensions
If this is right
- With the line of sight Compton-thin, the strong Fe K edge and apparent reflection cannot come from a Compton-thick torus; the mytorus fits favor scattering inside the near-side absorber, though a degenerate far-side contribution remains.
- Because the 2013 NuSTAR measurements and the 2018-2019 NICER measurements were made with consistent model families, the factor-of-two difference is evidence that the column density varies on timescales of years, consistent with the much faster variability reported for this source in 2004.
- A resolved neutral Fe K$\alpha$ line with $\sigma \simeq 40$ eV and implied velocity $\sim 1200$ km s$^{-1}$ is consistent with an origin in the optical broad line region, tying the line to gas much closer than a parsec-scale torus.
- Relativistic blurring fits place lower limits on the inner radius of the neutral reflector from $r \geq 270\,GM/c^2$ to $r \geq 4500\,GM/c^2$ depending on inclination, allowing the obscuration to begin well inside the classical torus.
- Simulations in the paper show XRISM/Resolve should detect dynamical line broadening in 100 ks if the reflector sits at $1600\,GM/c^2$, and Athena/X-IFU should detect factor-of-two column changes in 10 ks.
Reading between the lines
- If the obscuring gas is part of the optical BLR, then the Seyfert-1/Seyfert-2 distinction in NGC 4388 is a line-of-sight alignment through the same close-in cloud population, rather than a difference in the presence of a parsec-scale torus; future spectropolarimetry or Fe K reverberation could test this directly.
- A natural monitoring experiment is to track $N_H$ with NICER on weeks-to-months timescales; if the gas is clumpy BLR material, the column should show stochastic jumps and possibly brief Compton-thick excursions, which would distinguish cloud crossings from a smooth radial gradient.
- The paper's assumed radial ordering, with ionized absorber interior to the neutral gas, could be tested by catching a continuum flare: the warm absorber's ionization parameter should respond before the neutral column changes, yielding a direct distance estimate for each component.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents a NICER X-ray spectral analysis of the Seyfert-2 galaxy NGC 4388 using a time-averaged 0.6–10 keV spectrum with 105.6 ks of net exposure accumulated between 2017 December and 2019 March. The authors fit the spectrum with two independent model families, pexmon and mytorus, each combined with an XSTAR ionized absorber, scattered power-law components, and three mekal plasma components. Both models give consistent line-of-sight neutral column densities, N_H ≈ 2.6 × 10^23 cm^-2, which are Compton-thin and broadly consistent with earlier Suzaku and XMM-Newton measurements but about half of the column reported from a 2013 NuSTAR observation; this difference is interpreted as genuine variability. The Fe K-alpha line is nominally resolved with a width corresponding to ~1200 km/s, suggesting an origin in the optical broad line region. The paper also reports ionized absorption in the Fe K band, derives lower limits on the Fe K production radius using rdblur, and presents simulated XRISM and Athena spectra as future tests of the inferred geometry.
Significance. If the central result holds, the paper provides a clean, well-characterized measurement of a Compton-thin, variable absorber in a nearby Seyfert-2 galaxy, with geometric inferences that link the obscuring/reflecting gas to the broad line region. The use of two independent physical model families with statistically consistent column densities is a genuine strength, as are the reported MCMC uncertainties and the explicit comparison with previous Suzaku, XMM-Newton, and NuSTAR analyses. The simulated XRISM/Athena spectra give falsifiable predictions for future high-resolution X-ray observatories. The paper is transparent about the limitations of the cold-neutral-gas assumption and about calibration-related residuals, which is commendable.
major comments (3)
- [§2 and §4, Fig. 1] The central variability claim—that the obscuring column declined by roughly a factor of two between the 2013 NuSTAR epoch and the NICER epoch—is not yet secured because the NICER column is fitted to a 105.6 ks spectrum averaged over ~16 months. The paper itself notes rare high-flux intervals in the light curve and states that the hardness-ratio errors are large. If even a small fraction of the exposure caught an unveiling low-column state of the kind reported by Elvis et al. (2004), a single-zone absorber fitted to the summed spectrum will return an effective N_H biased below the dominant-state value, mimicking a genuine decline. The hardness-ratio check in Fig. 1 is not a substitute for time-resolved spectral fitting because the hardness ratio is not a direct column-density diagnostic and has large uncertainties. To support the 'genuine variability' conclusion, the authors should either fit spectra extracted from low-flux and high-flux intervals separately, add a two-zone absorption component to the time-averaged model, or present a quantitative demonstration that the observed sporadic high-flux intervals contribute negligibly to the fitted N_H. Absent that, the abstract and Section 4 should be rephrased to describe the difference as apparent or tentative.
- [§4, Kamraj et al. (2017) and Masini et al. (2016)] The abstract states that the robustness of the variability is 'reinforced by the use of consistent models and procedures,' but the quoted NuSTAR measurements are not model-identical to the present fits. Kamraj et al. (2017) used pexrav, which lacks the self-consistent Fe K lines included in pexmon, and their mytorus fit used a single scattering/line component and a single inclination. Masini et al. (2016) used a mytorus implementation that is similar but not exactly the same as the one in this paper. Because pexmon/pexrav and decoupled mytorus geometries differ, part of the measured N_H difference could reflect model systematics rather than source variability. The wording in Section 4 ('could reflect a genuine reduction') is appropriately cautious, but the abstract and title are not. The authors should either re-fit the 2013 NuSTAR data directly with the exact models used here, or explicitly frame the variability as tentative pending a uniform reanalysis.
- [§3.2, §3.3, and final paragraph of §4] The final paragraph appropriately acknowledges that pexmon and mytorus assume cold, neutral gas and that the gas may not be entirely neutral if it resides in the BLR. This is not a peripheral caveat: the quoted N_H values are neutral-equivalent columns, and a partially ionized absorber or reflector can bias the inferred column. The paper reports that exploratory xillver fits are equally good and that replacing the neutral absorber with zxipcf gives significantly worse fits, but those tests do not quantify how much the best-fit N_H changes under partial ionization, and the xillver and zxipcf geometries are not equivalent to the pexmon/mytorus setups. Given that the Fe K line width places the gas near or within the BLR, the systematic error on N_H from the ionization assumption could be substantially larger than the statistical errors of ±0.03 × 10^23 cm^-2. The authors should either add a quantitative systematic-error estimate for N_H under a partially ionized absorption/reflection model, or explicitly label the measured quantity as a 'neutral-equivalent column density' throughout the abstract and conclusions.
minor comments (5)
- [§3.2] In the model equation, 'contant[6]' should be 'constant[6]'.
- [§3.3] The text cites 'Vasudevan & Fabian (2006)' for the bolometric correction range, but the reference list contains Vasudevan & Fabian (2007) and (2009); please correct the citation year.
- [Table 1] The units of F_{0.6-10} and F_{0.6-10,unabs} are missing the exponent on cm: they should be erg cm^-2 s^-1.
- [§3.4 and Table 2] Please state how the 90% confidence lower limits in Table 2 were computed (e.g., Δχ^2 or MCMC percentile), since the table gives only the χ^2 value at the boundary.
- [§3.2] The sentence 'The pexmon model is not formally acceptable in a statistical sense' is important; please quantify the calibration-residual contribution to χ^2 (e.g., the χ^2 contribution of the 2–3 keV band) so readers can judge the impact on parameter uncertainties.
Circularity Check
No circularity: the reported NH is a fitted parameter from two independent spectral models, and the future-mission simulations are clearly extrapolations from the best-fit model, not inputs.
full rationale
The paper's central numbers (NH = 2.67e23 cm^-2 with pexmon; NH = 2.64e23 cm^-2 with mytorus) are obtained by fitting independent spectral models to the NICER data; neither model defines the fitted column in terms of the claimed conclusion, and the agreement between pexmon and mytorus is a cross-check, not a construction. The variability claim compares the NICER fit to earlier NuSTAR/Suzaku/XMM values; the authors re-fit with 'consistent models and procedures' to reduce systematic offsets, which is a comparison of independent measurements, not a self-referential prediction. The XSTAR table model is cited to Kallman & Bautista (2001) with Miller et al. (2015, 2016) as additional references, but the absorber's detection is driven by the data and the model is externally standard; no load-bearing argument reduces to those self-citations. The XRISM/Athena spectra in Figure 6 are simulations 'based on the best-fit mytorus model' and are explicitly offered as testable predictions, not as evidence for the fit. The paper itself flags its main limitations ('Both pexmon and mytorus assume cold, neutral gas'; the hardness-curve errors are large and the analysis is confined to the time-averaged spectrum), but these are ordinary model-systematic and temporal-mixing concerns about a fitted quantity, not cases where a derived result is equivalent to its input by definition. I therefore find no circular step.
Assumptions & free parameters
free parameters (12)
- N_H,LOS (neutral line-of-sight column density) =
2.67e23 cm^-2 (pexmon), 2.64e23 cm^-2 (mytorus)
- N_H,LOS,ion (ionized absorber column) =
1.7e22 cm^-2 (pexmon), 1.8e22 cm^-2 (mytorus)
- log xi (ionization parameter) =
3.44 (pexmon), 3.4 (mytorus)
- Gamma (photon index) =
1.54 (pexmon), 1.56 (mytorus)
- KPL (power-law normalization) =
3.2e-2 (pexmon), 3.96e-2 (mytorus)
- Rrefl (reflection fraction) =
0.113 (pexmon)
- fSC+lines (scattered + line normalization) =
0.62 (mytorus)
- Fe K alpha line sigma =
40 eV (Gaussian fit)
- mekal plasma temperatures and normalizations =
kT1=0.29, kT2=0.81, kT3=1.8 keV; norm1=8.5e-5, norm2=9.5e-5, norm3=2.0e-4 (pexmon)
- AFe (iron abundance) =
2.0 solar (fixed)
- Ecut (cutoff energy) =
500 keV (fixed)
- Inclination angles (theta_inc, theta_torus) =
5 deg/5 deg (pexmon), 90 deg/0 deg (mytorus)
assumptions (8)
- domain assumption The intrinsic X-ray continuum from the central engine is described by a power-law with exponential cutoff (cutoffpl) or a simple power-law (zpowerlaw for mytorus).
- domain assumption The obscuring material is cold and neutral, modeled by pexmon and mytorus slabs.
- ad hoc to paper The ionized absorber XSTAR model assumes a UV/X-ray flux ratio of 10:1 and a luminosity of 1e44 erg/s for the SED.
- domain assumption The low-energy diffuse emission is described by collisional ionization equilibrium (mekal) plasma.
- domain assumption For the line width interpretation, the broadening is Keplerian and the accretion disk is viewed at a specified inclination.
- ad hoc to paper For radius lower limits, the rdblur model assumes a Schwarzschild metric, emissivity index q=3, and outer radius 1e6 GM/c^2.
- domain assumption Galactic foreground absorption is fixed at NH=4.0e21 cm^-2.
- domain assumption Distance 18 Mpc and black hole mass 8.4e6 M_sun are adopted from the literature (Sorce et al. 2014; Kuo et al. 2011).
Cite this review
Pith. "Pith review of A NICER Look at Strong X-ray Obscuration in the Seyfert-2 Galaxy NGC 4388." pith.science (2026). https://pith.science/paper/I2CEBQAE
@misc{pith2026190808023,
author = {Pith},
title = {Pith review of: A NICER Look at Strong X-ray Obscuration in the Seyfert-2 Galaxy NGC 4388},
year = {2026},
howpublished = {\url{https://pith.science/paper/I2CEBQAE}},
note = {Machine review of arXiv:1908.08023}
}
read the original abstract
We present an analysis of the time-averaged spectrum of the Seyfert-2 active galaxy NGC 4388, obtained by NICER. The intrinsic strength of the reflection spectrum in NGC 4388, the large collecting area and favorable pass band of NICER, and a net exposure of 105.6 ks yielded an exceptionally sensitive spectrum. Using two independent families of models, the intrinsic spectrum from the central engine is found to be highly obscured but not Compton-thick. Enforcing physical self-consistency within each model, the independent treatments give formally consistent results: N_H = 2.67 (-0.03,+0.02) E+23 cm^-2 or N_H = 2.64 (-0.03, +0.03) E+23 cm^-2. Past measurements made with Suzaku and XMM-Newton are in broad agreement with these column density values. A more recent measurement with NuSTAR (in late 2013) recorded a column density about twice as large; the robustness of this variability is reinforced by the use of consistent models and procedures. The neutral Fe K-alpha line in the NICER spectrum is nominally resolved and consistent with an origin in the optical broad line region (BLR). The data also require ionized absorption in the Fe K band, similar to the "warm absorbers" detected in Seyfert-1 active galactic nuclei (AGN). The low-energy spectrum is consistent with a set of ionized plasma components. We discuss these findings and note that the geometric inferences that derive from this analysis can be tested with XRISM and Athena.
Reference graph
Works this paper leans on
-
[1]
48 1 . 52 1 . 56 1 . 60 Γ FIG . 3.— Confidence contours from the MCMC analysis related to th e fits with pexmon (see Table 1 and the text). The progressively lighter blue hu es indicate the 1σ , 2 σ , and 3 σ levels of confidence. The final panel in each row depicts the no rmalized one-dimensional probability density function fo r the parameter of interest. N...
-
[2]
8 1 . 6 2 . 4 3 . 2 4 . 0 NH, ion(× 1022)
-
[3]
2 3 . 6 4 . 0 4 . 4 log(ξion)
-
[5]
52 1 . 56 1 . 60 Γ 1 2 3 4 NH, ion(× 1022)
-
[6]
5 26 . 0 26 . 5 27 . 0 NH, LOS(× 1022)
-
[7]
60 Γ 1 2 3 4 NH, ion(× 1022)
-
[8]
25 3 . 50 3 . 75 4 . 00 log(ξion) FIG . 5.— Confidence contours from the MCMC analysis related to th e fits with mytorus (see Table 1 and the text). The progressively lighter blue hues indicate the 1σ , 2 σ , and 3 σ levels of confidence. The final panel in each row depicts the no rmalized one-dimensional probability density function fo r the parameter of int...
work page 2011
-
[9]
Note that the Fe K α line is revealed as a doublet
The simulated spectrum was binned to require a signal-to- noise ratio of 10. Note that the Fe K α line is revealed as a doublet. The spectrum in black is consistent with a distant, parsec-scale torus, whereas the spectrum in red (with broader lines) assumes emission from r = 1600 GM/ c2, a radius allowed by the NICER spectrum and consistent with prior var...
Show all 9 references
-
[25]
6 26. 0 26. 4 26. 8 27. 2 NH, LOS(× 1022)
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.