Available via license: CC BY 4.0
Content may be subject to copyright.
Investigation of the Nonthermal X-Ray Emission from the Supernova Remnant CTB 37B
Hosting the Magnetar CXOU J171405.7-381031
Chanho Kim
1
, Jaegeun Park
1
, Hongjun An
1
, Kaya Mori
2
, Stephen P. Reynolds
3
, Samar Safi-Harb
4
, and
Shuo Zhang
5
1
Department of Astronomy and Space Science, Chungbuk National University, Cheongju, 28644, Republic of Korea; hjan@cbnu.ac.kr
2
Columbia Astrophysics Laboratory, 550 West 120th Street, New York, NY 10027, USA
3
Physics Department, NC State University, Raleigh, NC 27695, USA
4
Department of Physics and Astronomy, University of Manitoba, Winnipeg, MB R3T 2N2, Canada
5
Department of Physics and Astronomy, Michigan State University, East Lansing, MI 48824, USA
Received 2024 September 13; revised 2024 November 13; accepted 2024 November 13; published 2024 December 11
Abstract
We present a detailed X-ray investigation of a region (S1)exhibiting nonthermal X-ray emission within the
supernova remnant (SNR)CTB 37B hosting the magnetar CXOU J171405.7−381031. Previous analyses modeled
this emission with a power law (PL), inferring various values for the photon index (Γ)and absorbing column density
(N
H
). Based on these, S1 was suggested to be an SNR shell, a background pulsar wind nebula, or an interaction
region between the SNR and a molecular cloud. Our analysis of a larger data set favors a steepening (broken or
curved PL)spectrum over a straight PL, with the best-fit broken power-law (BPL)parameters of Γ=1.23 ±0.23 and
2.24 ±0.16 below and above a break at 5.57 ±0.52 keV, respectively. However, a simple PL or srcut model
cannot be definitively ruled out. For the BPL model, the inferred N
H
=(4.08 ±0.72)×10
22
cm
−2
towards S1 is
consistent with that of the SNR, suggesting a physical association. The BPL-inferred spectral break ΔΓ ≈1 and hard
Γcan be naturally explained by a nonthermal bremsstrahlung (NTB)model. We present an evolutionary NTB model
that reproduces the observed spectrum, which indicates the presence of subrelativistic electrons within S1. However,
alternate explanations for S1, an unrelated PWN or the SNR shock with unusually efficient acceleration, cannot be
ruled out. We discuss these explanations and their implications for gamma-ray emission from CTB37B and describe
future observations that could settle the origin of S1.
Unified Astronomy Thesaurus concepts: Supernova remnants (1667);Magnetars (992);Gamma-ray sources (633);
X-ray sources (1822);Non-thermal radiation sources (1119);High energy astrophysics (739)
1. Introduction
High-energy cosmic rays nearing PeV energies have been
suggested to originate from Galactic sources such as supernova
remnants (SNRs)and pulsar wind nebulae (PWNe). PWNe are
believed to primarily accelerate leptons, while SNRs are
thought to be responsible for hadron acceleration. Evidence for
energetic hadrons in several SNRs comes from their gamma-
ray spectra (e.g., M. Ackermann et al. 2013).
Three primary radiation mechanisms involving energetic
leptons or hadrons are thought to be responsible for gamma-ray
emission from astrophysical objects. Inverse-Compton (IC)
scattering refers to the process where electrons boost the energy
of low-energy photons, such as those from the cosmic
microwave background or interstellar radiation field (ISRF),
to TeV energies. Additionally, nonthermal bremsstrahlung
(NTB)radiation emitted by energetic electrons can contribute
to the observed gamma-ray emission (e.g., R. A. Chevalier
1999; P. Slane et al. 2015). On the other hand, the hadronic
process entails the collision of high-energy protons, accelerated
by SNR shocks or through interaction between an SNR and a
molecular cloud (MC), with a dense surrounding medium (e.g.,
A. M. Bykov et al. 2000). These collisions give rise to neutral
pions, which subsequently decay into MeV–TeV gamma rays.
These concurrent leptonic and hadronic processes can
coexist within a given source. Therefore, definitively
identifying “hadronic”acceleration requires careful consid-
eration of the aforementioned radiation mechanisms to rule
out a purely leptonic origin for the gamma-ray emission via
IC and/or NTB processes. This necessitates an approach
involving the analysis of the multiwavelength image and
spectral energy distribution (SED)and ultimately, the
application of emission models to the observational data
(e.g.,S.P.Reynolds2008).
CTB 37B (G348.7+0.3)
6
is an SNR harboring the bright
magnetar CXOU J171405.7−381031 (hereafter J1714)with a
spin period of 3.8 s, surface dipole magnetic-field strength of
B
s
=4.8 ×10
14
G, and spin-down power of
E
4.2
SD
=´
10 erg s
34 1-(J. P. Halpern & E. V. Gotthelf 2010; T. Sato
et al. 2010). The estimated distance to and age of the SNR are
8–13 kpc and 650–6200 yr, respectively (R. Nakamura et al.
2009; W. W. Tian & D. A. Leahy 2012; H. Blumer et al. 2019).
The SNR has been well detected across various wavelengths,
including radio (N. E. Kassim et al. 1991), X-ray (R. Nakamura
et al. 2009), GeV (S. Abdollahi et al. 2020), and TeV bands
(F. Aharonian et al. 2008a). Its radio emission emanates from a
shell-like structure east of the magnetar (Figure 1). Diffuse
X-ray emission was detected surrounding the magnetar; this
X-ray emission region is mostly contained within the radio
shell. The GeV and TeV emissions exhibit significant spatial
overlap with both radio and X-ray regions.
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 https://doi.org/10.3847/1538-4357/ad938c
© 2024. The Author(s). Published by the American Astronomical Society.
Original content from this work may be used under the terms
of the Creative Commons Attribution 4.0 licence. Any further
distribution of this work must maintain attribution to the author(s)and the title
of the work, journal citation and DOI.
6
http://snrcat.physics.umanitoba.ca/SNRtable.php
1
Previous TeV observations (e.g., Figure 1; F. Aharonian
et al. 2008a)of the SNR favored hadronic processes over
leptonic ones. This is because leptonic scenarios would require
an unrealistically low magnetic-field strength (B)of ∼1μGor
an unexpected cutoff in the electron distribution at around
40 TeV. H. Zeng et al. (2017)proposed a lepto-hadronic model
where hadronic interactions dominate the TeV emission. Their
model necessitates a high gas density within the shell
(10 cm
−3
)to match the supernova (SN)energy budget (e.g.,
E
10 erg s
SN 51 1
=-). This value significantly exceeds the
density inferred from models of the SNR shell’s overall
X-ray emission (e.g., 0.5 cm
−3
; F. Aharonian et al. 2008a;
H. Blumer et al. 2019). The origin of this discrepancy in gas
density estimations remains unclear in their work.
While the X-ray spectrum of the diffuse emission around
J1714 was found to be thermal (F. Aharonian et al. 2008a),
R. Nakamura et al. (2009)observed nonthermal hard power-
law (PL)emission in the south of the magnetar, and subsequent
investigations by Chandra and XMM-Newton resolved this PL
emission to originate from a compact region at ∼4′south of
J1714 (S1 in Figure 1; see also H. Blumer et al. 2019;
E. V. Gotthelf et al. 2019). These previous studies reported
inconsistent spectral properties for S1, including photon index
(Γ)and absorbing column density (N
H
; Section 2.4).Asa
result, the origin of S1ʼs emission remains uncertain. While
R. Nakamura et al. (2009)attributed it to the SNR shell,
E. V. Gotthelf et al. (2019)proposed a background source.
H. Blumer et al. (2019)considered both an SNR–MC
interaction and an unrelated PWN as possible explanations.
The potential association of this nonthermal emission with the
SNR holds significant implications for the radiation processes
at play within CTB 37B.
This nonthermal X-ray region S1 could contribute to the
total TeV flux through processes like hadronic interactions or
NTB (e.g., S. P. Reynolds 2008; P. Slane et al. 2015). Based on
hard X-ray spectra with Γ<2 within the SNRs IC 443 and
W49B, S. Zhang et al. (2018)and T. Tanaka et al. (2018)
suggested NTB emission from them. Moreover, T. Tanaka
et al. (2018)explored the possibility of gamma-ray emission
through the NTB process in W49B. However, previous studies
on the broadband SED of CTB 37B have not considered this
possibility.
In this paper, we investigate the S1 emission using both
archival and newly acquired X-ray data from XMM-Newton
and NuSTAR. Our analysis methods and results are presented
in Section 2. We interpret the X-ray data under the framework
of the NTB scenario in Section 3. In Section 4, we discuss the
implications of our analysis results.
2. X-Ray Data Analysis
We focus on the S1 emission identified at ∼4′south of
J1714. This source was detected in four XMM-Newton and
two NuSTAR observations. NuSTAR detected S1 with high
significance (10σ)in the 10–20 keV band, confirming its
spectrally hard X-ray emission. While a Chandra observation
also captured this nonthermal source, its emission is very faint
and situated across the chip gap. Consequently, we have
excluded these Chandra data from our analysis. The X-ray data
sets analyzed in this study are listed in Table 1.
2.1. Data Reduction
The XMM-Newton observations were processed using the
pipeline tasks emproc and epproc of the XMMSAS
software (version 20230412_1735). Particle flare contamina-
tion was removed from each observation following the standard
procedures. The source was detected by the metal oxide
semiconductor (MOS)detector in only one observation, as the
remaining three employed the small-window mode for MOS.
The source was well detected in all four PN data sets. The
NuSTAR observations were processed using the nupipe-
line script integrated in HEASOFT v6.32. While the source
was well detected in the NuSTAR observations (e.g.,
Figure 2(a)), its proximity to stray-light patterns hindered
reliable background estimation from the in-flight data,
necessitating a careful examination of the background during
analysis. The net exposure times after these initial cleaning
steps are presented in Table 1.
Figure 1. A composite image of combined SUMSS (843 MHz; white
contours), Herschel SPIRES (red), XMM-Newton (1–8 keV; green)and
NuSTAR (3–20 keV; blue)data of CTB 37B. The magnetar is marked as
J1714 (cyan cross), and our main target “S1”is denoted by a white ellipse. The
NuSTAR image was truncated to remove stray-light contamination (e.g.,
Figure 2(a)). The magenta ellipse and the yellow circle show the GeV and TeV
counterparts detected by Fermi-LAT (magenta; 68% positional uncertainty)
and H.E.S.S. (yellow; 1σextension), respectively. The images were smoothed
and logarithmically scaled to improve legibility.
Table 1
X-Ray Data Used in This Study
Observatory Date Obs. ID Exposure
(MJD)(ks)
XMM-Newton 55273 0606020101 99/51
a
XMM-Newton 55999 0670330101 8
b
XMM-Newton 57655 0790870201 24
b
XMM-Newton 57806 0790870301 17
b
NuSTAR 57654 30201031002 79/78
a
NuSTAR 60234 40901004002 80/78
a
Notes.
a
MOS/PN and FPMA/FPMB for XMM-Newton and NuSTAR, respectively.
b
PN only.
2
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
2.2. Image Analysis
We used the eimageget script of the Science Analysis
System (SAS)for each XMM-Newton observation to create a
background-subtracted and vignetting-corrected image.
7
These
images were combined to produce a 1–8 keV image. NuSTAR
images in the range 3–20 keV were generated with back-
ground-subtracted simulations using nuskybgd (D. R. Wik
et al. 2014)and exposure corrections applied. We then
combined these NuSTAR images with the XMM-Newton
image alongside an IR image measured by Herschel SPIRES.
These energy bands were selected to optimize signal-to-noise
ratio. We additionally displayed radio contours obtained from
the SUMSS data (T. Mauch et al. 2003). The resultant radio-to-
X-ray composite image is shown in Figure 1. Notably, the radio
contours overlap well with the XMM-Newton image, and the
IR emission seems to delineate the radio SNR shell to the east.
The X-ray image, encompassing J1714, the SNR emission
surrounding it, and the southern nonthermal emission (“S1”),
appears to exhibit an overall morphology similar to the radio
shell. In both the XMM-Newton and NuSTAR data, S1
manifests as extended emission with
R
1~
¢
. The radio and
X-ray emissions significantly overlap with the GeV and TeV
emissions measured by Fermi-Large Area Telescope
(4FGL J1714.1−3811; S. Abdollahi et al. 2020)and H.E.S.S
(HESS J1713−381; H.E.S.S. Collaboration et al. 2018).Itis
worth noting that the GeV emission remains unresolved and
seems to originate from within the SNR shell, whereas the TeV
emission exhibits extension with a Gaussian width (σ)of 5 5,
covering a large region containing both the shell and S1.
2.3. Timing Analysis for J1714
Although the magnetar J1714 is not our main target and its
emission was heavily contaminated by stray light (bright
regions except for J1714 and S1 in Figure 2(a)) in the
NuSTAR observations, we carry out a timing analysis with
the new data to ascertain if there have been any significant
changes in the magnetar’s rotation since the last measurement
(J. P. Halpern & E. V. Gotthelf 2010;T.Satoetal.2010;
E. V. Gotthelf et al. 2019). During the new observation, the
magnetar was observed at 5~
¢
off-axis position near the edge
of the detector, resulting in a distorted event distribution and
reduced counts (Figure 2(a)). Moreover, due to the bright
stray-light pattern overlapping with J1714 in the FPMB data,
we relied solely on the FPMA data for the timing analysis. We
selected source events within a 50″×30″(radii)elliptical
region, barycenter-corrected their arrival times using (R.A.,
decl.)=(258°. 5239057, −38°. 1752758)(J2000),andcon-
ducted an H test (O. C. de Jager et al. 1989)around the
expected period based on previous measurements. The spin
period derivative (
P
)was held fixed at a previously reported
value of 5 ×10
−11
ss
−1
.
Our analysis successfully detected pulsations at low
energies (1.6–5keV), revealing a period of P=3.852450
(5)s on MJD 60233 (Figures 2(b)and (c)). Unfortunately,
owing to limited statistics resulting from a large off-axis
angle and elevated background due to stray light, we were
unable to confirm the previously reported phase reversal of
the pulse profile at higher energies (>6keV; E. V. Gotthelf
et al. 2019)with the new data. However, the obtained Pvalue
aligns with the trend presented in E. V. Gotthelf et al. (2019);
by comparing our result with the previous Chandra measure-
ment on MJD 54856, we estimated an average
P
of 6 ×
10
−11
ss
−1
, falling within the previously reported range of
(5–7)×10
−11
ss
−1
.
2.4. Spectral Analysis of the Emission from S1
Several previous studies have extensively characterized the
X-ray spectrum of S1 (R. Nakamura et al. 2009; H. Blumer
et al. 2019; E. V. Gotthelf et al. 2019). R. Nakamura et al.
(2009)analyzed a large region encompassing S1 in Suzaku
X-ray Imaging Spectrometer (XIS)data, wherein they
employed a model to fit the spectrum while accounting for
contamination from J1714 and the thermal shell of the SNR.
Regarding the S1 emission, they derived Γ=1.5 ±0.4 for
N
3.5 10 cm
H0.7
0.5 22 2
=´
-
+-using the angr abundances
(E. Anders & N. Grevesse 1989). While this approach provided
insights into the broader region, it was susceptible to
Figure 2. (a)3–20 keV NuSTAR FPMA (left)and FPMB (right)images made
with the 2023 observation. While J1714 appears to be heavily contaminated by
stray light in the FPMB image, it was detected outside the contamination in
FPMA. We smoothed and logarithmically scaled the images to enhance
legibility. (b)and (c)Result of our pulsation search (b)and background-
subtracted 1.6–5 keV pulse profile (c)of J1714 measured using the 2023
NuSTAR data.
7
https://www.cosmos.esa.int/web/xmm-newton/sas-thread-images
3
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
contamination from other sources. XMM-Newton’s high
angular resolution facilitated a more precise measurement of
S1ʼs spectrum, minimizing contamination from other sources.
H. Blumer et al. (2019)analyzed XMM-Newton MOS data
(Obs. ID 0670330101)and reported Γ=1.3 ±0.3 and
N
3.1 10 cm
H0.8
0.9 22 2
=´
-
+-for S1, while E. V. Gotthelf et al.
(2019)analyzed the same MOS data jointly with NuSTAR
spectra (Obs. ID 30201031002),finding a steeper Γof
2
.2 0.5
0.6
-
+
and higher absorption with N
H
=(11 ±4)×10
22
cm
−2
.It
should be noted that these N
H
values were determined
employing the wilms abundance model (J. Wilms et al. 2000).
Given the ongoing debate surrounding the origin and
spectral characteristic of the S1 emission, we acquired a new
NuSTAR observation and reanalyzed the existing XMM-
Newton and NuSTAR data (Table 1). While the previous
XMM-Newton studies utilized only the MOS data, the source
was well detected by XMM-Newton PN. Expanding the data
set to include the PN data can potentially improve the previous
characterization of the S1 spectrum. The source spectra were
extracted using an elliptical region of 60″×40″(radii)centered
at (R.A., decl.)=(258°. 5423, −38°. 2471)from both XMM-
Newton and NuSTAR data, as depicted in Figures 1and 2(a).
While the in-flight data around the source region can
effectively represent the background in the XMM-Newton
data, the complex stray-light pattern in the NuSTAR data (e.g.,
Figure 2(a)) poses challenges for background estimation. To
address this issue, we conducted nuskybgd simulations to
estimate the NuSTAR background, utilizing source-free
regions while excluding the stray-light patterns. These simula-
tions provided estimates of background contributions to the
source-region spectra. For XMM-Newton data, we extracted
background from 45″radius circles located 150″west of the
source, for both MOS and PN data. These background regions
were chosen to be on the same detector chips as the source,
avoiding S1 and the SNR shell. Alternative background regions
(south or west of S1)were tested, and we found that the results
do not alter significantly (see below).
We initially performed independent fits to the XMM-Newton
and NuSTAR spectra. We collected 6700/3400 and 3300/
1600 counts within the source/background regions from the
XMM-Newton (0.3–10 keV; all observations combined)and
NuSTAR data (3–20 keV; all observations combined).As
reported by H. Blumer et al. (2019), the XMM-Newton data
favor a hard PL model with Γ=1.35 ±0.17 and
N
H
=(4.38 ±0.65)×10
22
cm
−2
. Conversely, the NuSTAR
data are well fit by a softer PL model with Γ=2.06 ±0.09
for a fixed N
H
of 4.38 ×10
22
cm
−2
. Optimizing N
H
for the
NuSTAR fit results in a softer Γof 2.31 ±0.18 with a higher
N
H
value of (8.46 ±2.88)×10
22
cm
−2
. For Galactic absorp-
tion, we employed the wilms abundances and vern cross
section (D. A. Verner et al. 1996). We checked systematic
effects on the Γmeasurements due to background selection by
employing five different background regions for each data set.
Depending on the background selection, the Γvalues inferred
from the XMM-Newton and NuSTAR data varied by ∼0.04
and ∼0.05, respectively. These are smaller than the statistical
uncertainties.
The difference in the Γvalues obtained from the XMM-
Newton and NuSTAR analyses suggests a potential spectral
break or curvature. Consequently, we jointly fit the combined
XMM-Newton and NuSTAR spectra using both an absorbed
PL and a broken PL (BPL)model (Figure 3). To consider the
possibility of synchrotron emission from the SNR shock in the
cutoff regime, we also employed the srcut model (S. P. Rey-
nolds & J. W. Keohane 1999). This analysis included two MOS
spectra (MOS1 and MOS2), four PN spectra, and four
NuSTAR spectra (FPMA and FPMB). A cross-normalization
factor was applied to each data set, with the value fixed to 1 for
the MOS1 spectrum. The best-fit values for these factors were
found to be consistent with 1 within uncertainties. Detailed
values and uncertainties for the spectral parameters are
presented in Table 2. Systematic uncertainties due to back-
ground selection were estimated and reported in the table.
The PL model prefers a large Γof 1.95 ±0.09 with a high
N
H
=6.52 ±0.52 cm
−2
. These parameters are statistically
consistent with those reported by E. V. Gotthelf et al. (2019).
The N
H
value differs at a >3σlevel from those measured
toward the SNR shell (e.g.,
4
.3 10 cm
;
0.1
0.2 22 2
´
-
+-H. Blumer
et al. 2019)and J1714 ((3.6–4.0)×10
22
cm
−2
; E. V. Gotthelf
et al. 2019). Despite an adequate fit to the data
(χ
2
/dof =524/549), the residuals of the PL model exhibit a
small downward slope with increasing energy and an excess at
5–6 keV where the BPL model predicts a break. The srcut
model, with the previously reported αvalue of 0.3 (although
this is for the entire SNR; N. E. Kassim et al. 1991), also yields
an acceptable fit to the data without overpredicting the
measured 1 GHz flux density of the entire SNR. Similar to
the PL model, this model requires a high N
H
. Changing αto a
larger value, e.g., 0.5 as observed in other radio SNRs, results
in an E
brk
of ∼10 keV, a flux density of ∼mJy and an N
H
of
6.2 ×10
22
cm
−2
.
F-tests indicate that the BPL model provides statistically
significant improvements over the PL and the srcut models
with F-test probabilities of 5 ×10
−4
and 5 ×10
−3
, respec-
tively. This finding remains valid even after excluding a few
PN observations wherein a significant portion of S1 fell on bad
pixels. The BPL fit indicates that the N
H
value toward S1 is
consistent with those for the SNR shell and J1714, and the
spectrum exhibits a break at 6 keV with ΔΓ ≈1.
3. NTB Model for the Nonthermal X-Ray Emission from S1
Building on our description of S1ʼs X-ray emission with a
BPL model, we investigate the origin of this nonthermal
component. The association of S1 with the SNR shell may
suggest particle acceleration resulting from the interaction of an
MC with the SNR shock (H. Blumer et al. 2019). While not
explicitly identified in previous work, MC 3251, located at an
estimated distance of 8.1–8.6 kpc (M.-A. Miville-Deschênes
et al. 2017), is a plausible candidate for the MC. Theoretical
studies have demonstrated that such shock propagation in an
MC can accelerate thermal electrons in weakly ionized regions
to nonthermal energies (A. M. Bykov et al. 2000). The
observed hard photon index and strong ΔΓ ≈1 spectral break
(Section 2.4)favor the NTB process as the dominant X-ray
emission mechanism for S1. In this section, we construct an
evolutionary NTB model and apply it to the X-ray emission
from S1.
3.1. NTB Emission from Energetic Electrons
NTB emission from electrons with energies higher than
those in the background plasma has been proposed as a source
of hard X-ray emission in SNRs by several authors (e.g.,
T. Tanaka et al. 2018; S. Zhang et al. 2018). Here, we
4
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
summarize some of the properties of this process. We consider
a background plasma composed of ions (density n
i
)and
electrons (density n
e,b
)in thermal equilibrium at temperature T.
To this background, we add a suprathermal electron population
with kinetic energies E
ke
?E
th
∼kT and density n
e
, satisfying
()nn n1.2 . 1
eb e i,+=
The suprathermal electrons interact with background electrons
and ions on different timescales (e.g., L. Spitzer 1978)and
radiate bremsstrahlung photons. Electrons will share energy
among themselves on a timescale
()
⎜⎟
⎛
⎝⎞
⎠
tn
E
m
1.24 10 2 ,2
ee
ee e b e
18
,
ke
32
l
@´-
where λ
ee
is the Coulomb logarithm, typically ∼30 for
E
ke
=10 keV, kT =1 eV, and n
e,b
=80 cm
−3
(as we find
below). This timescale also approximates the characteristic
cooling time for high-energy electrons due to energy transfer to
the thermal electron pool (i.e.,
tEE
ee ke ke
@
).
In the nonrelativistic regime, the electron energy evolution
can be determined analytically (J. Vink 2008):
() () ()Et E nt01.1610 . 3
ee e bke 1.5 ke 1.5 5 ,
l=-´
-
This describes the evolving energy of an electron with initial
energy E
ke
(0)?kT interacting with a much larger pool of
background electrons. This energy transfer is commonly
referred to as Coulomb losses although energy remains in the
fluid. In the absence of any additional acceleration processes
such as turbulent acceleration, the electron energy distribution
evolves over time as lower-energy (suprathermal)electrons
cool and successively disappear into the background plasma.
Consequently, for the electron distribution of age t, there is an
energy E
C
(E
ke
with t=t
ee
in Equation (3)) below which the
suprathermal electrons are steeply depleted. This depletion
produces a break in the distribution at E
C
that rises with time.
Figure 4(a)illustrates this effect, showing the evolution of the
electron distribution for a PL index s=1.5 and n
e,b
=81 cm
−3
.
Electrons also scatter off ions with a typical timescale
t
pe
=(n
e,b
/n
i
)t
ee
(L. Spitzer 1978). While these scatterings
produce bremsstrahlung photons, the radiative energy losses
are significantly smaller than those due to Coulomb interactions
with background electrons (V. Petrosian 2001). An individual
electron with E
ke
emits a bremsstrahlung photon spectrum that
is independent of energy up to hE
max ke
n~. Photons of energy
E
γ
≡hνare typically produced by electrons with energies
several times larger. Consequently, the photon spectrum N(E
γ
)
closely resembles the electron distribution.
3.2. NTB Modeling of the X-Ray Spectrum of S1
We consider a simplified scenario in which the SN that
produced the shell SNR CTB 37B and J1714 sent a blast wave
into a small region of higher density (S1). This shock-
accelerated electrons to ∼100 keV, adequate to produce
bremsstrahlung photons up to ∼20 keV. The shock injected a
total amount of energy W
e,S1
to electrons, proportional to the
transverse extent of S1 (f
Ω,S1
; solid angle fraction of S1):
()WfE,4
eS S
,1 ,1 SN
h=W
where ηis the fraction of shock energy converted to electron
acceleration. We describe the electron distribution with a
Maxwellian with kT ∼1eV (thermal background)with an
Figure 3. 1–20 keV X-ray spectra of the S1 region measured by XMM-Newton (black)and NuSTAR (red), and the best-fitPL(a),srcut (b), and BPL models (c).
The bottom panels display residuals after subtracting the best-fit model from the data.
Table 2
Results of Joint Fits of the XMM-Newton and NuSTAR Spectra
Model Energy Range N
H
Γ/α
a
E
brk
Γ
2
b
Flux
c
χ
2
/dof
(keV)(10
22
cm
−2
)(keV)
PL 0.3–20 keV 6.52 ±0.52 ±0.11 1.95 ±0.09 ±0.03 LL3.18 ±0.20 ±0.01 524/549
srcut 0.3–20 keV 5.91 ±0.45 ±0.04 0.30 4.31 ±1.70 ±0.05 L35.00 ±8.11 ±1.33 520/549
BPL 0.3–20 keV 4.08 ±0.72 ±0.07 1.23 ±0.23 ±0.02 5.57 ±0.52 ±0.06 2.24 ±0.16 ±0.06 2.79 ±0.21 ±0.02 510/547
Notes. The statistical and systematic uncertainties, reported as the first and second errors, respectively, are at the 1σconfidence level.
a
X-ray photon index (Γ)for the PL and BPL models, and radio spectral index (α)for the srcut model. We held αfixed at 0.3.
b
Photon index above the break energy E
brk
.
c
Absorption-corrected 2–10 keV flux in units of 10
−13
erg s
−1
cm
−2
for the PL and BPL models, and flux density at 1 GHz in units of μJy for the srcut model.
5
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
attached nonthermal tail (shock-accelerated population). This
postshock nonthermal electron distribution
()
() ()
dN
dE dt
nV s
EE E
1,5
eeS
ss
s
ke
1
ke,min
1
ke,max
1age
ke
t
=-
-
--
-
with V
S1
and τ
age
being the volume and lifetime of S1, evolves
due to Coulomb losses as described by Equation (3). We ignore
the possibility of turbulent reacceleration behind the shock.
To obtain the total photon spectrum at a given age, we
integrate over the shocked region using a one-dimensional
geometry. Assuming the transverse area of the source A
(60″×60″corresponding to 2.6 ×2.6 pc
2
for an assumed
distance of 9 kpc), we divide the emitting region (V
S1
)into
discrete volumes ΔV=Adz where dz =v
sh
dt and v
sh
is the
shock velocity; we assume it to be 900 km s
−1
as inferred for
the SNR shock (H. Blumer et al. 2019). We calculate E
ke
(t), the
electron distribution, and the emission spectrum for each
volume (time step dt). Summing over these spectra yields the
total spectrum for the source age τ
age
.
We use energetic considerations to proceed. First, we assume
the total energy in accelerated particles (nonthermal electrons
and ions)is small enough that the test-particle result for the
particle distribution for diffusive shock acceleration applies.
This yields s=3/2 for an initial shocked-particle distribution
N(p)∝p
−4
. For s<2, most energy is at the higher end.
Therefore, the total energy constraint primarily affects
(
)E
0
ke,max . The (
)E
0
ke,max value determines the cutoff energy
in the electron distribution (Equation (3)), which corresponds to
the energy above which the photon spectrum steepens.
Consequently, high values of (
)E
0
ke,max overpredict the
observed X-ray spectrum above the spectral break at ∼6 keV.
To reproduce the observed X-ray break, the initially most
energetic electrons (having ()
E
0 120 keV;
ke,max =Table 3)
should cool to ∼30 keV. Equation (3)then gives
n
e,b
τ
age
∼3300 cm
−3
yr. A low n
e
, limited by the SN energy
budget, necessitates a high n
i
to explain the observed flux since
the NTB flux scales as ∝n
i
n
e
V
S1
. This leads to a high n
e,b
(Equation (1)) and a correspondingly short τ
age
.
We optimized model parameters to reproduce the X-ray
spectrum, assuming a fixed SN energy injection W
e,S1
(Equation (4)) and s=1.5. Additionally,
E
ke,min was fixed
to ensure a smooth connection between the PL and thermal
distributions. For the given W
e,S1
and s, the other parameters
except for n
i
(and thus n
e,b
; Equation (1)) are well constrained;
n
i
exhibits flexibility within a broad range. This is because the
emission, which is proportional to n
i
, is counterbalanced
by the effects of n
e,b
(∼n
i
)since emission (cooling)
timescale is inversely proportional to n
e,b
:t
cool
≈E
ke
(0)
1.5
/
(1.16 ×10
−5
λ
ee
n
e,b
)(Equation (3).
This cooling timescale has observational consequences.
Under continuous shock acceleration over τ
age
, the X-ray flux
of S1 would initially rise for t
cool
because the injection rate
exceeds the cooling rate during the initial phase. The emission
then remains stationary for the rest of its age, τ
age
−t
cool
.We
adjusted n
e,b
such that this stationary period is longer than the
∼15 yr period of relatively constant observed X-ray flux
(Table 1); a larger n
e,b
is also acceptable as it extends this
period. Table 3shows a sample set of parameters.
It is important to note that alternative choices for the
parameter values can also provide acceptable fits to the data
due to parameter covariance (especially with W
e,S1
).
Figure 4. (a)Evolution of an electron distribution over 55 yr for s=1.5 and initial n
e
=39 cm
−3
and n
e,b
=81 cm
−3
(Table 3).n
e
decreases to ∼0 over τ
age
, while n
e,
b
increase to 120 cm
−3
as the nonthermal electrons cool (see text). The gray solid curve displays the summed distribution of the background (n
e,b
; Maxwellian with
kT =1eV)and injected (n
e
;PL)electrons. The other solid curves display the time evolution of the PL distribution (young to old from red to purple), and the black
dashed line shows the sum of the distributions (scaled with dt).(b)X-ray emission SED (data points)of S1 and our NTB model computation (dashed line).
Table 3
Parameters for the NTB Model in Figure 4
Parameter Symbol Value
SN energy (10
51
erg)
E
SN 1
a
Solid angle fraction of S1 f
Ω,S1
0.013
a
Energy conversion efficiency η0.1
a
Injected energy (10
48
erg)W
e,S1
1.3
a
Index of electron distribution s1.5
a
Minimum energy of electrons (eV)
E
ke,min 3.5
a
Maximum energy of electrons (keV)
E
ke,max 120
Injected electron density (cm
−3
)n
e
39.1
b
Background ion density (cm
−3
)n
i
100
Background electron density (cm
−3
)n
e,b
80.9
b
Lifetime of S1 (yr)τ
age
55
Notes.
a
Fixed.
b
Initial values. These values evolve with time as injected electrons (n
e
)cool
and transition into the background population (n
e,b
). See Section 3.1 for details.
6
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
Consequently, the specific values reported in Table 3represent
just one possible solution within a range of possibilities. We
present further discussions in Section 4.3.
4. Discussion
The origin of the nonthermal X-ray emission from S1 has
been controversial. R. Nakamura et al. (2009)proposed that
these X-rays share a common origin with the radio shell
emission of CTB 37B based on their similar spectral indices.
Alternatively, H. Blumer et al. (2019)suggested that a shock
interaction with a nearby MC or a PWN unassociated with
CTB 37B could be responsible. E. V. Gotthelf et al. (2019)
attributed the emission to a background PWN based on the high
value of N
H
that they inferred.
Leveraging improved photon statistics facilitated by XMM-
Newton PN and NuSTAR data, we obtained a refined
measurement of the S1 spectrum. Our analysis of the S1
X-ray spectrum revealed that three models (PL, srcut, and
BPL)provide adequate fits, with the F-test favoring the BPL
model. None of the three could be definitively excluded,
leaving an association between S1 and CTB 37B inconclusive.
All three explanations have significant difficulties, as we
describe below.
4.1. Scenario 1: Emission from a PWN Unrelated to CTB 37B
If the true emission spectrum of S1 is a PL, the inferred
N
H
of 6.5 ×10
22
cm
−2
is incompatible with those for
the SNR and J1714 at a >3σlevel (Table 1). This suggests
that S1 is not physically associated with the SNR, aligning with
the interpretation of a background PWN proposed by
E. V. Gotthelf et al. (2019). The properties of S1 (L
X
=3.6 ×
10
33
(d/10 kpc)
2
erg s
−1
and Γ∼2)would be quite typical for a
PWN (see Figure 5 of O. Kargaltsev et al. 2013). In this
scenario, the nondetection by XMM-Newton or Chandra of a
central point source, potentially a middle-aged pulsar, is
somewhat puzzling since existing empirical correlations
between luminosities of pulsars and their PWNe (O. Kargalt-
sev & G. G. Pavlov 2008; X.-H. Li et al. 2008)suggest that
pulsars should be as luminous as their PWNe. However, X-ray
PWNe are sometimes without detected pulsars. While most of
the 91 X-ray PWNe tabulated in O. Kargaltsev et al. (2013)
have a point source(s)within them, pulsations have not been
detected in 15 of them, making the association between the
PWN and the point source(s)unclear. Moreover, the putative
pulsar may be observationally faint due to strong absorption if
its emission is spectrally soft. Deeper X-ray observations with
future instruments like AXIS and HEX-P (C. S. Reynolds et al.
2023; K. K. Madsen et al. 2024)could help resolve this issue.
4.2. Scenario 2: Synchrotron Emission from the SNR Shock
Highly relativistic electrons, possibly accelerated by inter-
action between S1 and the SNR shock, are capable of
generating X-rays via synchrotron radiation. In this case, one
might expect a slowly cutting-off spectrum that can be
described by the simple srcut model, as is seen in other
remnants (e.g., Tycho; L. A. Lopez et al. 2015). However, the
roll-off photon energies reported in Table 2(5–10 keV,
depending on radio properties)would be the highest ever
observed for an SNR. When electron acceleration is limited by
synchrotron losses, the characteristic roll-off photon energy is
given by
() ()
⎛
⎝⎞
⎠
huR21000 km s keV 6
gJrolloff shock
1
2
1
nh~-
-
(e.g., S. P. Reynolds 2008). Here η
g
≡λ
mfp
/r
g
is the
“gyrofactor,”the electron mean-free path in units of its
gyroradius, and R
J
is a geometric factor reflecting potential
variations in acceleration rate as a function of the shock
obliquity angle θ
Bn
between the shock velocity and upstream
magnetic field (J. R. Jokipii 1987). J. R. Jokipii (1987)shows
that as η
g
increases, acceleration can proceed much faster in
perpendicular shocks (θ
Bn
∼π/2)than in parallel ones though
in an increasingly narrow range of θ
Bn
near π/2. For large η
g
,
R
J
varies as g
2
h
-, so Equation (6)shows that higher roll-offs
could be obtained, invoking this effect with large values of η
g
.
The relatively low shock velocity (∼900 km s
−1
; H. Blumer
et al. 2019)means that a very large value of η
g
would be
required. IXPE (M. C. Weisskopf et al. 2022)observations of
SNRs revealed small θ
Bn
values in young SNRs (P. Slane et al.
2024), potentially indicating a modest η
g
for S1. However,
examples of tangential (large θ
Bn
)magnetic fields also exist in
older SNRs (e.g., D. A. Prokhorov et al. 2024). Polarization
angle measurements for CTB 37B and S1 are crucial to validate
the feasibility of this scenario.
Nevertheless, the srcut model implies a higher N
H
than
estimates for the SNR and J1714 at the 3σlevel, casting doubt
on the association between S1 and CTB 37B.
4.3. Scenario 3: NTB Emission from S1
The BPL model suggests a different scenario. The
consistency between N
H
toward S1 and that measured for the
SNR and J1714, along with the absence of a point-like source
within S1, argues for a physical association with CTB 37B.
Furthermore, the hard spectral index below 6 keV and the
observed degree of the spectral break (ΔΓ ≈1)favor an NTB
interpretation over synchrotron radiation, as synchrotron
emission from PWNe is typically softer (Γ≈2)with smaller
ΔΓ values (e.g., 0.5; A. Bamba et al. 2022). The NTB
scenario aligns with the suggestion of H. Blumer et al. (2019)
regarding an interaction between the SNR shock and an MC.
While our NTB model provided a successful explanation for
the X-ray measurements, this NTB scenario also has some
difficulties upon closer examination.
These difficulties stem primarily from the inefficiency of
NTB radiation compared to Coulomb losses (see V. Petrosian
& W. E. East 2008, for details)and the limited energy budget
of the SN. Only a small fraction (∼10
−5
; V. Petrosian 2001)of
electron energy contributes to NTB emission. Consequently,
the available SN energy of W
e,S1
∼10
48
erg can sustain the
observed flux of F
2−10 keV
∼3×10
−13
erg s
−1
cm
−2
for only
∼100 yr, significantly shorter than the estimated SN age of
650–6200 yr (R. Nakamura et al. 2009; H. Blumer et al. 2019).
Our NTB model reflects this constraint by assuming that all
available SN energy is dedicated to electron acceleration,
resulting in τ
age
=55 yr for S1 age. However, it remains
unclear whether the interaction process accelerates only
electrons (and not protons). If the SN energy is shared with
protons, the age estimate decreases, potentially conflicting with
the observed stability of the X-ray flux over ≈13.5 yr (Table 1).
Additionally, an asymmetric SN explosion (e.g., J1714ʼs
7
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
proximity to the western shell; Figure 1)could have injected
less energy into S1, leading to a reduced W
e,S1
and a smaller
τ
age
. We note that a similar requirement of a young age arises
from the rather low value of break energy E
brk
, necessitating a
small value of n
e,b
τ
age
, as already remarked in Section 3.2.
It is important to acknowledge that our model considered
only Coulomb cooling for the evolution of the electron
distribution. In reality, additional processes, including turbu-
lence and magnetic-field interactions (e.g., A. M. Bykov et al.
2000), likely play significant roles. If these processes induce
sufficiently rapid acceleration, the nonthermal PL tail (respon-
sible for NTB emission)may persist for extended periods
(V. Petrosian & W. E. East 2008), potentially mitigating the
aforementioned challenges. Further investigation into these
complexities is warranted to develop a more comprehensive
model for the nonthermal emission from S1, including
predictions for Fe Kαflux from excitation by nonthermal
electrons, which XRISM (XRISM Science Team 2020)
can test.
4.4. Possibility of TeV Emission from S1
Our analysis suggests three potential origins for the
nonthermal X-ray emission from S1: an unassociated PWN
(Scenario 1), the SNR shock accelerating relativistic, synchro-
tron-emitting electrons (Scenario 2), or the SNR shock
accelerating suprathermal but nonrelativistic electrons produ-
cing NTB (Scenario 3). Scenarios 1 and 2 predict IC emission
at TeV energies or higher although precise flux estimates are
challenging.
In Scenario 1, the S1 size of
R
1»
¢
(∼4 pc for an assumed
distance of 13 kpc scaled by N
H
)may indicate a middle-aged
PWN. Such a PWN could exhibit a TeV flux comparable to its
X-ray flux (e.g., J. Park et al. 2023), which is an order of
magnitude lower than the measured gamma-ray flux of the
SNR shell (peaking at ∼100 GeV). IC emission from the
electrons emitting synchrotron photons at ∼10 keV would
appear at TeV energies, as observed in other middle-aged
PWNe. Similarly, the high roll-off energies of 5–10 keV
inferred from the srcut model (Scenario 2)imply electron
energies of 30 TeV for an assumed B=100 μG within the
SNR shock (e.g., H. Zeng et al. 2017). These electrons could
upscatter ISRF (e.g., 30 K blackbody)to TeV.
In Scenario 3, S1 is unlikely to produce a significant gamma-
ray flux on its own due to the limited energy budget (i.e., lack
of >TeV electrons). However, the surrounding SNR shell
could potentially contribute high-energy particles to S1 via
energetic proton diffusion. Similar scenarios involving particle
escape from SNR shells and interaction with nearby clouds
have been proposed for sources like SNR W28 (F. Aharonian
et al. 2008b)and dark accelerators (S. Gabici et al. 2009).
While the small solid angle coverage of S1 (f
Ω,S1
≈1.3%)
limits the number of protons reaching the cloud, the high gas
density within S1 (estimated ion density of 100 cm
−3
; Table 3)
could still potentially lead to detectable gamma-ray emission. It
is important to note that proton diffusion is energy dependent
(F. A. Aharonian & A. M. Atoyan 1996), preferentially
enriching S1 with higher-energy protons from the shell.
In summary, TeV emission from S1 is possible under any
of the three scenarios for the hard X-ray emission. In
particular, S1 may manifest as a distinct “high-energy”
TeV source distinguishable from the SNR shell itself. Future
TeV observatories like the Cherenkov Telescope Array
(M. Actis et al. 2011)may have the resolving power to
reveal such a source.
5. Summary
Despite our investigation, the origin of the nonthermal X-ray
emission from S1 remains uncertain. The X-ray data favor the
BPL spectrum over the PL and srcut ones, but the latter two
cannot be definitively ruled out. These spectral models suggest
three different scenarios (Sections 4.1–4.3)for the S1 emission,
and we find that all three scenarios have significant problems.
The NTB scenario (BPL model; Scenario 3)requires an
unrealistically short source age, and its overall energetics are
problematic. The synchrotron explanation (srcut model;
Scenario 2)requires dramatically more rapid electron accelera-
tion than has been documented in other sources although in
principle such acceleration cannot be ruled out. The unrelated-
PWN explanation (PL model; Scenario 1)probably has the
fewest fatal flaws, requiring only a somewhat unlikely spatial
coincidence between a fairly conventional X-ray PWN and the
CTB 37B shell. The various scenarios are testable: a detection
of a central point source with deep Chandra observations could
lend credence to “unassociated PWN”scenario. The NTB
scenario predicts a decline in the X-ray flux over the next
decades due to Coulomb cooling (but see Section 4.3). The
keV electrons producing NTB should also excite atomic lines,
most prominently Fe Kα; detecting such emission from S1,
e.g., with XRISM, could strengthen the case for NTB, while
more stringent upper limits could weaken it. Further X-ray
observations are essential to distinguish among these scenarios.
Acknowledgments
This work used data from the NuSTAR mission, a project led
by the California Institute of Technology, managed by the Jet
Propulsion Laboratory, and funded by NASA. We made use of
the NuSTAR Data Analysis Software (NuSTARDAS)jointly
developed by the ASI Science Data Center (ASDC, Italy)and
the California Institute of Technology (USA). J.P. acknowledges
support from Basic Science Research Program through the
National Research Foundation of Korea (NRF)funded by the
Ministry of Education (RS-2023-00274559). This research was
supported by the National Research Foundation of Korea (NRF)
grant funded by the Korean Government (MSIT)(NRF-
2023R1A2C1002718). S.S.H.ʼs research is primarily supported
by the Natural Sciences and Engineering Research Council of
Canada (NSERC)through the Canada Research Chairs and the
Discovery Grants programs.
Facilities: XMM (F. Jansen et al. 2001), NuSTAR
(F. A. Harrison et al. 2013).
Software: HEAsoft (v6.31; NASA High Energy Astrophy-
sics Science Archive Research Center (Heasarc),2014), XMM-
SAS (20211130_0941; C. Gabriel 2017), XSPEC (v12.12;
K. A. Arnaud 1996).
ORCID iDs
Chanho Kim https://orcid.org/0000-0003-0226-9524
Jaegeun Park https://orcid.org/0000-0002-9103-506X
Hongjun An https://orcid.org/0000-0002-6389-9012
Kaya Mori https://orcid.org/0000-0002-9709-5389
Stephen P. Reynolds https://orcid.org/0000-0002-5365-5444
Samar Safi-Harb https://orcid.org/0000-0001-6189-7665
Shuo Zhang https://orcid.org/0000-0002-2967-790X
8
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
References
Abdollahi, S., Acero, F., Ackermann, M., et al. 2020, ApJS,247, 33
Ackermann, M., Ajello, M., Allafort, A., et al. 2013, Sci,339, 807
Actis, M., Agnetta, G., Aharonian, F., et al. 2011, ExA,32, 193
Aharonian, F., Akhperjanian, A. G., Barres de Almeida, U., et al. 2008a, A&A,
486, 829
Aharonian, F., Akhperjanian, A. G., Bazer-Bachi, A. R., et al. 2008b, A&A,
481, 401
Aharonian, F. A., & Atoyan, A. M. 1996, A&A, 309, 917
Anders, E., & Grevesse, N. 1989, GeCoA,53, 197
Arnaud, K. A. 1996, in ASP Conf. Ser. 101, Astronomical Data Analysis Software
andSystemsV,ed.G.H.Jacoby&J.Barnes(San Francisco, CA: ASP),17
Bamba, A., Shibata, S., Tanaka, S. J., et al. 2022, PASJ,74, 1186
Blumer, H., Safi-Harb, S., Kothes, R., Rogers, A., & Gotthelf, E. V. 2019,
MNRAS,487, 5019
Bykov, A. M., Chevalier, R. A., Ellison, D. C., & Uvarov, Y. A. 2000, ApJ,
538, 203
Chevalier, R. A. 1999, ApJ,511, 798
de Jager, O. C., Raubenheimer, B. C., & Swanepoel, J. W. H. 1989, A&A,
221, 180
Gabici, S., Aharonian, F. A., & Casanova, S. 2009, MNRAS,396, 1629
Gabriel, C. 2017, in Proc. of the X-Ray Universe 2017 Symp., ed.
J.-U. Ness & S. Migliari (Darmstadt: ESA),84
Gotthelf, E. V., Halpern, J. P., Mori, K., & Beloborodov, A. M. 2019, ApJ,
882, 173
Halpern, J. P., & Gotthelf, E. V. 2010, ApJ,725, 1384
Harrison, F. A., Craig, W. W., Christensen, F. E., et al. 2013, ApJ,770, 103
H.E.S.S. Collaboration, Abdalla, H., Abramowski, A., et al. 2018, A&A,612, A2
Jansen, F., Lumb, D., Altieri, B., et al. 2001, A&A,365, L1
Jokipii, J. R. 1987, ApJ,313, 842
Kargaltsev, O., & Pavlov, G. G. 2008, in AIP Conf. Ser. 983, 40 Years of
Pulsars: Millisecond Pulsars, Magnetars and More, ed. C. Bassa et al.
(Melville, NY: AIP),171
Kargaltsev, O., Rangelov, B., & Pavlov, G. 2013, in The Universe Evolution:
Astrophysical and Nuclear Aspects, ed. I. Strakovsky & L. Blokhintsev
(New York: Nova Science Publishers),359
Kassim, N. E., Baum, S. A., & Weiler, K. W. 1991, ApJ,374, 212
Li, X.-H., Lu, F.-J., & Li, Z. 2008, ApJ,682, 1166
Lopez, L. A., Grefenstette, B. W., Reynolds, S. P., et al. 2015, ApJ,814,
132
Madsen, K. K., García, J. A., Stern, D., et al. 2024, FrASS,11, 1357834
Mauch, T., Murphy, T., Buttery, H. J., et al. 2003, MNRAS,342, 1117
Miville-Deschênes, M.-A., Murray, N., & Lee, E. J. 2017, ApJ,834, 57
Nakamura, R., Bamba, A., Ishida, M., et al. 2009, PASJ,61, S197
NASA High Energy Astrophysics Science Archive Research Center (Heasarc)
2014, HEAsoft: Unified Release of FTOOLS and XANADU, Astrophysics
Source Code Library, ascl:1408.004
Park, J., Kim, C., Woo, J., et al. 2023, ApJ,945, 66
Petrosian, V. 2001, ApJ,557, 560
Petrosian, V., & East, W. E. 2008, ApJ,682, 175
Prokhorov, D. A., Yang, Y.-J., Ferrazzoli, R., et al. 2024, arXiv:2410.20582
Reynolds, C. S., Kara, E. A., Mushotzky, R. F., et al. 2023, Proc. SPIE,12678,
126781E
Reynolds, S. P. 2008, ARA&A,46, 89
Reynolds, S. P., & Keohane, J. W. 1999, ApJ,525, 368
Sato, T., Bamba, A., Nakamura, R., & Ishida, M. 2010, PASJ,62, L33
Slane, P., Bykov, A., Ellison, D. C., Dubner, G., & Castro, D. 2015, SSRv,
188, 187
Slane, P., Ferrazzoli, R., Zhou, P., & Vink, J. 2024, Galax,12, 59
Spitzer, L. 1978, Physical Processes in the Interstellar Medium (New York:
Wiley)
Tanaka, T., Yamaguchi, H., Wik, D. R., et al. 2018, ApJL,866, L26
Tian, W. W., & Leahy, D. A. 2012, MNRAS,421, 2593
Verner, D. A., Ferland, G. J., Korista, K. T., & Yakovlev, D. G. 1996, ApJ,
465, 487
Vink, J. 2008, A&A,486, 837
Weisskopf, M. C., Soffitta, P., Baldini, L., et al. 2022, JATIS,8, 026002
Wik, D. R., Hornstrup, A., Molendi, S., et al. 2014, ApJ,792, 48
Wilms, J., Allen, A., & McCray, R. 2000, ApJ,542, 914
XRISM Science Team 2020, arXiv:2003.04962
Zeng, H., Xin, Y., Liu, S., et al. 2017, ApJ,834, 153
Zhang, S., Tang, X., Zhang, X., et al. 2018, ApJ,859, 141
9
The Astrophysical Journal, 977:163 (9pp), 2024 December 20 Kim et al.
