Available via license: CC BY 4.0
Content may be subject to copyright.
Draft version June 14, 2023
Typeset using L
A
T
E
Xtwocolumn style in AASTeX631
Hard X-ray observation and multiwavelength study
of the PeVatron candidate pulsar wind nebula “Dragonfly”
Jooyun Woo ,1Hongjun An ,2Joseph D. Gelfand ,3Charles J. Hailey ,1Kaya Mori ,1
Reshmi Mukherjee ,4Samar Safi-Harb ,5and Tea Temim 6
1Columbia Astrophysics Laboratory, 550 West 120th Street, New York, NY 10027, USA
2Department of Astronomy and Space Science, Chungbuk National University, Cheongju, 28644, Republic of Korea
3NYU Abu Dhabi, PO Box 129188, Abu Dhabi, United Arab Emirates
4Department of Physics and Astronomy, Barnard College, 3009 Broadway, New York, NY 10027, USA
5Department of Physics and Astronomy, University of Manitoba, Winnipeg, MB R3T 2N2, Canada
6Princeton University, 4 Ivy Ln, Princeton, NJ 08544, USA
ABSTRACT
We studied the PeVatron nature of the pulsar wind nebula G75.2+0.1 (“Dragonfly”) as part of
our NuSTAR observational campaign of energetic PWNe. The Dragonfly is spatially coincident with
LHAASO J2018+3651 whose maximum photon energy is 0.27 PeV. We detected a compact (radius 1′)
inner nebula of the Dragonfly without a spectral break in 3 −20 keV using NuSTAR. A joint analysis
of the inner nebula with the archival Chandra and XMM-Newton observations yields a power-law
spectrum with Γ = 1.49 ±0.03. Synchrotron burnoff is observed from the shrinkage of the NuSTAR
nebula at higher energies, from which we infer the magnetic field in the inner nebula of 24 µG at 3.5
kpc. Our analysis of archival XMM data and 13 years of Fermi-LAT data confirms the detection of
an extended (∼10′) outer nebula in 2 −6 keV (Γ = 1.82 ±0.03) and non-detection of a GeV nebula,
respectively. Using the VLA, XMM, and HAWC data, we modeled a multi-wavelength spectral energy
distribution of the Dragonfly as a leptonic PeVatron. The maximum injected particle energy of 1.4
PeV from our model suggests that the Dragonfly is likely a PeVatron. Our model prediction of the low
magnetic field (2.7 µG) in the outer nebula and recent interaction with the host supernova remnant’s
reverse shock (4 kyrs ago) align with common features of PeVatron PWNe. The origin of its highly
asymmetric morphology, pulsar proper motion, PWN-SNR interaction, and source distance will require
further investigations in the future including a multi-wavelength study using radio, X-ray, and gamma-
ray observations.
1. INTRODUCTION
Pulsar wind nebulae (PWNe) of energetic (spin-down
luminosity ˙
E > 1036 erg/s) middle-aged (characteris-
tic age τ= 10 −100 kyr) pulsars are often associated
with very-high-energy (VHE, above 1 TeV) sources (e.g.,
H.E.S.S. Collaboration et al. (2018)). Many of them
are luminous above a hundred TeV without a hint of a
spectral cutoff (e.g., Abeysekara et al. (2020) and Su-
doh et al. (2021)). Recently, the higher energy regime
of their spectra was unveiled by the Large High Altitude
Air Shower Observatory (LHAASO), the first gamma-
ray observatory sensitive to PeV-energy gamma rays,
Corresponding author: Jooyun Woo
jw3855@columbia.edu
and their detection of 14 Galactic ultra-high-energy
(UHE, above 100 TeV) sources (Cao et al. (2021a), Aha-
ronian et al. (2021), and Cao et al. (2021b)). The high-
est photon energies detected from these sources range
from several hundred TeV to above 1 PeV: irrefutable
evidence of particle acceleration above 1 PeV in both
hadronic (neutral pion decay) and leptonic (inverse
Compton scattering) cases. Identifying these Galactic
“PeVatrons” is the key to the origin of the highest-
energy Galactic cosmic rays observed on the Earth (in
hadronic case) as well as a better understanding of
the particle acceleration, radiation, and transportation
mechanism (in both hadronic and leptonic case).
The majority of the LHAASO sources are spatially co-
incident with middle-aged energetic PWNe, well-known
leptonic particle accelerators. Our NuSTAR observa-
tional campaign of energetic PWNe aims to explore
arXiv:2306.07347v1 [astro-ph.HE] 12 Jun 2023
2Woo et al.
the extreme nature of such PWNe (Mori et al. (2022)).
Broadband hard X-ray observations with NuSTAR pro-
vide a unique window to the highest end of their par-
ent particle spectra by resolving their synchrotron ra-
diation without contamination from thermal radiation.
Combined with modeling the multi-wavelength (MW)
spectral energy distribution (SED) of the PWNe over 20
decades of energy range, it allows deducing the key phys-
ical parameters that define the systems, such as the max-
imum particle energy and magnetic field. Our NuSTAR
observation and MW SED modeling have functioned
as powerful probes of PWNe as energetic leptonic cos-
mic ray accelerators in our Galaxy (e.g., Burgess et al.
(2022), Park et al. (2023a), and Park et al. (2023b)).
G75.2+0.1 (“Dragonfly”) is one of the eight tar-
get PWNe of our NuSTAR observational campaign
and is likely associated with LHAASO J2018+3651.
The Dragonfly is powered by PSR J2021+3651 (RA
= 20:21:05.40, Dec = +36:51:04.5) first discovered by
Roberts et al. (2002) as a radio pulsar with a rota-
tion period P∼
=104 ms. The radio observation of the
pulsar was motivated by the detection of an uniden-
tified X-ray source AX J2021.1+3651 (Roberts et al.
(2001)), which was a follow-up observation of an uniden-
tified gamma-ray source GeV J2020+3658 (Hartman
et al. (1999)). As a middle-aged pulsar whose char-
acteristic age τ≡P/2˙
P∼17 kyr, PSR J2021+3651
is still energetic, with ˙
E∼3.4×1036 erg s−1. PSR
J2021+3651 is detected in X-ray as a soft (mostly) ther-
mal (kTBB = 0.16 ±0.02 keV) point source by Chandra
(Hessels et al. (2004) and Van Etten et al. (2008)). The
authors of both works reported the detection of X-ray
pulsations to be insignificant. The PWN G75.2+0.1 of
PSR J2021+3651 was first observed in X-ray by Hes-
sels et al. (2004) and was named the “Dragonfly” by
Van Etten et al. (2008) for its double-torus structure.
Fermi -LAT (Fermi ) observation by (Abdo et al. (2009))
detected GeV pulsations from PSR J2021+3651, yet its
spectrum sharply cuts off below 10 GeV with no evi-
dence of higher energy emission from the PWN.
PSR J2021+3651 and the Dragonfly are located in the
Cygnus region, an active star-forming region. The first
TeV gamma-ray source detected in spatial coincidence
with PSR J2021+3651 and the Dragonfly was MGRO
J2019+37 (Abdo et al. (2007)). MGRO J2019+37 is the
second brightest TeV source in the northern hemisphere
after the Crab Nebula and largely extended (circular
2D Gaussian with σ= 0.32◦±0.12◦). Since its de-
tection, numerous observations in different wavebands
have been carried out as attempts to identify the origin
of such high energy emissions. Roberts et al. (2008)
observed the region with the VLA in radio (20 cm)
and the XMM-Newton (XMM) in soft X-rays. Both
observations revealed a more comprehensive picture of
G75.2+0.1 beyond the substructures seen by Chandra
−a conical diffuse nebula pivoted at PSR J2021+3651
that extends out to ∼20′(radio) and ∼10′(soft X-ray)
on the west with decreasing surface brightness. In this
work, the entire structure of the PWN is referred to as
the Dragonfly.
Aliu et al. (2014) used the Very Energetic Radia-
tion Imaging Telescope Array System (VERITAS) to
resolve MGRO J2019+37 into two separate sources:
VER J2019+368 and VER J2016+371. While VER
J2016+371 is dominated by low-energy (below 1 TeV)
emission near a supernova remnant (SNR) CTB 87,
VER J2019+368 (RA = 20:19:25, Dec = 36:48:14, ellip-
tical 2D Gaussian with major-axis σmaj = 0.34◦±0.03◦
and minor-axis σmin = 0.13◦±0.02◦) is bright above
1 TeV. With additional 120 hours of data, Abeysekara
et al. (2018) reported that VER J2019+368 may be re-
solved into two source candidates, VER J2020+368∗and
VER J2018+367∗. The High-Altitude Water Cherenkov
gamma-ray observatory (HAWC) found the high-energy
emission from VER J2019+368 to be significant even
above 56 TeV and named the source eHWC J2019+368
(Abeysekara et al. (2020)). Its significant detection
above 100 TeV by LHAASO with the maximum pho-
ton energy 0.27 ±0.02 PeV (Cao et al. (2021a)) con-
firms that one or more PeVatrons of Galactic origin are
present in this region. This extreme Galactic source,
namely LHAASO J2018+3651, is spatially coincident
with multiple possible cosmic ray accelerators, includ-
ing a Wolf-Rayet (WR) star WR 141, H II region Sh
2-104, PSR J2021+3651 and the Dragonfly.
In this work, we aim to evaluate the Dragonfly’s po-
tential as a leptonic PeVatron. We report the first hard
X-ray observation of the Dragonfly using NuSTAR. We
analyze the archival Chandra and XMM data and 13
years of Fermi data on the Dragonfly. We combine the
spectra of the Dragonfly extracted from our analyses
with the radio and TeV spectra from the previous works
to model the MW SED of the Dragonfly. We discuss the
common features of PeVatron PWNe, source distance,
and magnetic field.
2. X-RAY DATA ANALYSIS
We analyzed two sets of archival Chandra data
(Obs ID 8502, 34 ks, 2006 Dec 25, and Obs ID
7603, 60 ks, 2006 Dec 29), one set of archival XMM
data (Obs ID 0674050101, 135 ks, 2012 Apr 17),
one set of new NuSTAR data (Obs ID 40660004002,
61 ks, 2021 May 19). We processed the Chandra
data using the chandra repro task in CIAO 4.13 (Fr-
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 3
uscione et al. (2006)) and the calibration database
CALDB 4.9.5. We processed NuSTAR data using the
nuproducts task in NuSTAR Data Analysis Software
package (NuSTARDAS v2.0.0) contained within HEASOFT
6.28 and the NuSTAR calibration database (CALDB
version 20210315). We processed the XMM Euro-
pean Photon Imaging Camera (EPIC) MOS data using
the emchain and emfilter tasks in the XMM-Newton
Extended Source Analysis Software (XMM-ESAS) pack-
age contained within the XMM-Newton Science Analy-
sis System (SAS v20.0.0). The net exposure after re-
moving soft proton (SP) flares is 85 ks. The XMM EPIC
pn data was not used since it was obtained in small win-
dow mode (one single CCD) and hence is inappropriate
for observing a large diffuse nebula that extends over
multiple CCDs.
2.1. Timing analysis
Hessels et al. (2004) reported a marginal (3.7σ)
detection of X-ray pulsations in 0.5−3 keV from
PSR J2021+3651 using a Chandra data in continuous-
clocking mode and contemporaneous radio ephemeris.
The same authors reported significant timing noise and
a possibility of large glitches in PSR J2021+3651. We
attempted to search for hard X-ray pulsations from
PSR J2021+3651 using the NuSTAR data. We applied
an astrometric correction on the pulsar position to the
cleaned event files using the Chandra data analyzed in
this work. We applied a barycentric correction to these
event files for the corrected pulsar position using the
barycorr task in NuSTARDAS. We used extractor to
select position- and timing-corrected events within the
r= 30′′ circular region around PSR J2021+3651 cor-
responding to the half-power diameter (HPD) of NuS-
TAR. We generated binned light curves from the se-
lected events in 3 −6, 6 −20, and 3 −20 keV bands
(bin size = 1 ms) using the timing analysis software
HENDRICS 7.0 (Bachetti (2018)). We used the light
curves to create power spectra with the timing analysis
software Stingray v1.1 (Huppenkothen et al. (2019)).
No significant frequency features were found. Given
the lack of contemporaneous pulsar ephemeris, we per-
formed Z2
n(n= 2) searches around the radio pulsar fre-
quency and frequency derivative found by Roberts et al.
(2002). This search did not yield a significant detection
of pulsations.
2.2. Imaging analysis
Van Etten et al. (2008) resolved the substructures of
the PSR J2021+3651 and the Dragonfly using Chan-
dra. Such substructures include pulsar jets, 20′′ ×10′′
double-torus inner nebula, a bow shock standoff, and a
20h21m07s06s05s04s03s
36°51'30"
15"
00"
50'45"
Right Ascension
Declination
Inner
nebula
Jet
20h21m30s00s20m30s
36°57'
54'
51'
48'
Right Ascension
Declination
Arc
Figure 1. Merged (Obs ID 8502 and 7603) and exposure-
corrected Chandra images of the Dragonfly in 2−6 keV. The
scales were adjusted for better legibility. PSR J2021+3651
is marked as a cross (X) in both images. Top: 1′×1′image
after Gaussian smoothing with σ= 1.5 pixel = 0.7′′. The
20′′ ×10′′ inner nebula and the pulsar jet stretching out to
∼30′′ from the pulsar are marked with dashed lines. Bottom:
21′×11′image after Gaussian smoothing with σ= 3 pixel
= 3.0′′. The arc in length ∼7.7′is traced with a dotted line.
The extent of the outer nebula seen by XMM is marked as
a dashed line.
peculiar “arc” stretching toward the east of the pulsar
(dotted line in the bottom figure of Figure 1). An outer
nebula with a size much larger than the inner nebula
seen by Chandra was discovered by Roberts et al. (2008)
using XMM. Zabalza et al. (2010) used XMM observa-
tions covering the region further west to that of Roberts
et al. (2008) and constrained the size of the outer neb-
ula to 10 −15′to the west of PSR J2021+3651. Mizuno
et al. (2017) not only confirmed the western extent of
the outer nebula measured by Zabalza et al. (2010) us-
ing Suzaku but also claimed the emission seen by XMM
on the east of PSR J2021+3651 including the “arc” is
part of the outer nebula.
In this section, we discuss the X-ray morphology of
the Dragonfly seen by Chandra , XMM, and NuSTAR.
4Woo et al.
36°57'
54'
51'
48'
Declination
3-6 keV
20h21m30s00s20m30s
36°57'
54'
51'
48'
Right Ascension
Declination
6-20 keV
Figure 2. Merged (FPMA and FPMB) and exposure- and
vignetting-corrected 21′×11′NuSTAR images after Gaussian
smoothing with σ= 1.5 pixels = 4.7′′ . The scales were
adjusted for better legibility. PSR J2021+3651 and WR 141
are marked with a cross (X) and a plus (+), respectively.
The extent of the outer nebula seen by XMM is marked with
a dashed line.
We investigate a change in the morphology of the hard
X-ray nebula in two different energy ranges: soft band
(3 −6 keV) and hard band (6−20 keV). We present the
XMM image of the Dragonfly to study the morphology
of the outer nebula and briefly discuss the nature of the
“arc.” Chandra images in 2 −6 keV are compared to
the NuSTAR and XMM images in similar energy ranges
(3 −6 keV and 2 −6 keV, respectively). A detailed
description of the Chandra image can be found in Van
Etten et al. (2008).
2.2.1. Chandra image
We merged the two Chandra observations (Obs ID
8502 and 7603) using the merge obs task in CIAO to
create an exposure-corrected image in 2 −6 keV.
Figure 1shows smaller (inner nebula and jet) and
larger (outer nebula and arc) structures of the Drag-
onfly. The inner nebula is centered at PSR J2021+3651
and axis-symmetric along the jet. Its size is 20′′ along
the major axis and 10′′ along the minor axis in diame-
ter. The jet is measured to extend out to 30′′ from the
pulsar. The observations covered only part of the outer
nebula seen by XMM (dashed line in the bottom panel
of Figure 1), yet it is clearly visible. The arc continues
to the edge of the FOV, measuring 7.7′in length.
2.2.2. NuSTAR image
20h21m15s10s05s00s
36°53'
52'
51'
50'
Right Ascension
Declination
r = 1'
3-6 keV
20h21m15s10s05s00s
Right Ascension
r = 1'
6-20 keV
Figure 3. 4′×4′NuSTAR images of the inner nebula. The
images were created following the same procedures described
in the caption of Figure 2. PSR J2021+3651 is marked with a
cross (X). A dashed circle of radius 1′is shown as a reference.
We created images for both focal plane modules
(FPMA and FPMB) in the soft band (3 −6 keV) and
the hard band (6−20 keV) using extractor. The corre-
sponding exposure maps after vignetting correction were
created using nuexpomap task in NuSTARDAS. We com-
bined the FPMA and FPMB images and corrected the
exposure using XIMAGE to create Figure 2.
A bright emission is detected in both energy bands at
the location of PSR J2021+3651 (marked with a cross
(X) in the figure) and the surrounding region (inner neb-
ula). The west of PSR J2021+3651 is contaminated by
a stray light background, so it is difficult to estimate the
emission from the faint outer nebula. WR 141 (marked
with a plus sign (+) in the figure) becomes significantly
dimmer in the hard band. To examine the detailed
morphology of the inner nebula, we created zoomed-in
images (Figure 3). The emission is roughly symmetric
about the pulsar in both energy bands. The nebula fits
well in a radius 1′circle, while it shows an apparent
decrease in size in the hard band (6 −20 keV). We fit-
ted PSF-convolved models to the images using Sherpa
(Freeman et al. (2001)), a fitting and modeling applica-
tion in CIAO. Both images are fitted with a constant
background and a single 2D Gaussian. The FWHM
of the Gaussian is 26.5′′ ±3.2′′ for the soft band and
15.2′′ ±2.0′′ for the hard band.
2.2.3. XMM image
XMM is the only instrument whose image captures
the entirety of the outer nebula in the X-ray band.
To study this large diffuse emission, we first removed
the contamination of the outer nebula by bright point
sources in the FOV, such as PSR J2021+3651, WR 141,
and a star USNO-B1.0 1268-0448692. We created Swiss
cheese masks for MOS1 and MOS2 images that reduce
the surface brightness of point sources to 20% of the
surrounding background using the cheese task. These
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 5
20h22m00s21m30s00s20m30s
37°00'
36°55'
50'
45'
40'
Right Ascension
Declination
Outer
nebula
Ring
Figure 4. Merged (MOS1 and MOS2), QPB-subtracted,
exposure-corrected, and smoothed XMM image in 2 −6 keV.
PSR J2021+3651 and WR 141 are marked with a cross (X)
and a plus (+) sign, respectively. The extent of the outer
nebula ∼10′is marked with dashed lines. A “ring” is marked
with a dotted line.
masks were applied to the cleaned event files using the
mos-spectra task to create MOS1 and MOS2 images
of the entire FOV in 2 −6 keV. The quiescent particle
background (QPB) image was generated for the entire
FOV in the same energy range using the mos back task.
Residual SP contamination was found to be negligible
(see §2.3.3). No significant instrumental or SWCX back-
ground is present in the energy range of our analysis,
and no significant stray light background was observed
in the image. Therefore, after combining the MOS1 and
MOS2 images (comb task), we subtracted only the QPB
image, corrected the exposure, and adaptively smoothed
it using the adapt task to create Figure 4.
Significant emissions are present on the east and west
of PSR J2021+3651 (marked with X). The emission
on the east of PSR J2021+3651 (“ring”-like structure,
marked with dotted line) shows no low-energy counter-
part in the radio (VLA L band) observation by (Roberts
et al. (2008)). Mizuno et al. (2017) claimed this ring-
like structure to be part of the Dragonfly. On the other
hand, Van Etten et al. (2008) and Jin et al. (2023) de-
tected a bow-shock structure from the inner nebula of
the Dragonfly in the X-ray (Chandra ) and radio (VLA
C and L band) observations, respectively. Such detec-
tions indicate a supersonic motion of PSR J2021+3651
toward the east, in which case it is unlikely to expect
PWN emission ahead of the bow shock formed by the
pulsar. Possible origins of the emission on the west of
Figure 5. VLA 20cm image from Roberts et al. (2008). The
radio nebula extends out to >20′from PSR J2021+3651.
The XMM contours are overlaid in blue. The permission for
the use of the image was acquired from AIP Publishing via
RightsLink®.
the pulsar are discussed in the last paragraph of this
section.
The emission on the west of PSR J2021+3651 (“outer
nebula”, marked with dashed line) extends out to ∼10′
with decreasing surface brightness. This X-ray nebula
is spatially coincident with the first half of the radio
nebula, as shown in Figure 5. The radio nebula extends
further out to >20′(Roberts et al. (2008)), whose flux
was used for modeling the SED of the Dragonfly (see
§4). For consistency, we analyze the X-ray counterpart
of the radio nebula, namely the outer nebula, and use
its spectrum for SED modeling.
A ring-like structure is centered at WR 141 and has
radius ∼5′. The “arc” seen by Chandra comprises the
lower part of this ring. Mizuno et al. (2017) claimed that
the “ring” is part of the PWN based on the similar spec-
tral index between the “ring” and the “outer nebula.”
Barkov et al. (2019) explained the “arc” as a “kinetic
jet”: pulsar wind particles that escaped into the ISM
due to magnetic reconnection between the PWN and
the ISM and became visible in a high >10 µG ISM
magnetic field. A similar filamentary emission ahead of
the main body of the PWN (“outer nebula”) was ob-
served in the “Snail” PWN, whose “prongs” may be the
result of the interaction between the PWN and the re-
verse shock of its host SNR (Temim et al. (2015)). We
propose that the “ring” in our XMM image is possibly
6Woo et al.
associated with WR 141. WR stars are known to have
strong stellar winds that can create a bubble of several
parsecs in radius (Weaver et al. (1977)). This bubble
is often observed as a ring-shaped nebula and can be
visible in X-ray (e.g., Toal´a et al. (2012)). The parallax
of 0.5024 mas in Gaia DR3 (Collaboration et al. (2022))
implies a 2.0 kpc distance to WR 141. This yields the
radius of the bubble = 2.9 pc. Part of the “ring” was
also seen in Hαphotometry by Law et al. (2002), which
the authors postulated to be part of the ring nebula
photoionized by WR 141.
2.3. Spectral analysis
We present a spectral analysis of PSR J2021+3651
and its PWN using Chandra, XMM, and NuSTAR data.
We first characterize the pulsar spectrum with Chan-
dra taking advantage of its fine angular resolution (HPD
<0.5′′). We analyze the spectrum of the inner nebula
taking into account the contribution of the pulsar by in-
dividually and jointly fitting the Chandra , XMM, and
NuSTAR spectra. We use the XMM data to study the
spectrum of the outer nebula. All the spectral models
for X-ray analysis presented in this work were multi-
plied by a cross-normalization factor (const) to adjust
relative normalization between different detectors and
instruments.
2.3.1. Pulsar spectrum
We used Chandra data to analyze the spectrum of
PSR J2021+3651. We extracted the source spectra
from a circular region with radius 2′′ centered at PSR
J2021+3651, and the background spectra from an an-
nulus around PSR J2021+3651 with radii 2 −4′′ using
the specextract command in CIAO. The source spec-
tra were binned to have at least 3σsignificance over the
background in each bin. We began by fitting an ab-
sorbed power law (const*tbabs*pow) to the spectra in
0.5−7 keV where the source emission dominates over
the background. The abundance table was set to wilm
(Wilms et al. (2000)) for all the X-ray spectral analyses
presented in this work. This model gives a reasonable
fit (χ2/d.o.f = 130/152) with the best-fit Γ = 1.73+0.13
−0.12
and NH= (0.26 ±0.05) ×1022 cm−2. However, this
best-fit NHis 3 times smaller than the NHfound from
the Chandra spectra of the inner nebula. When the NH
was fixed to the best-fit value found from the Chan-
dra inner nebula spectra (0.76 ×1022 cm−2, see §2.3.2),
the fit quality became worse (χ2/d.o.f = 190/152) with
much softer Γ = 2.87 ±0.15. This is because the pul-
sar has significant emission in both below and above 3
keV. Initially, a small NHwas favored to explain the
emission below 3 keV. Later, the Γ was significantly
softened to compensate for the larger NH, leaving the
10 14
10 13
E2dN/dE (erg/s/cm2)
1 5
E (keV)
2.5
0.0
2.5
Figure 6. Chandra ObsID 8502 and 7603 (black and red
crosses, respectively) spectra of PSR J2021+3651. The back-
ground is dominant outside of 0.5−7 keV. The source spectra
were extracted from a circular region with radius 2′′ centered
at the pulsar. The background spectra were extracted from
an annulus region with radii 2 −4′′ centered at the pulsar.
emission above 3 keV poorly fitted. We added a black
body component to fit the emission below 3 keV while
the power law component explains the emission above 3
keV (const*tbabs*(bbod+pow)). NHis highly degen-
erate with the black body temperature and the power
law index, so we fixed NHto the Chandra value of the
inner nebula. This gives the best-fit kT = 0.13 ±0.01
keV and Γ = 1.63 ±0.17 with χ2/d.o.f = 123/151. We
used this model as the initial pulsar component when
jointly fitting the Chandra spectra of the inner neb-
ula and the XMM and NuSTAR spectra of the pul-
sar and the inner nebula (see §2.3.2). We iteratively
fit the Chandra pulsar spectra by changing the NHto
the best-fit value found from the joint fit of the Chan-
dra, XMM, and NuSTAR spectra. The pulsar model
converged to kT = 0.11 ±0.01 keV and Γ = 1.77 ±0.17
with χ2/d.o.f = 123/151. The best-fit parameters are
comparable to Van Etten et al. (2008) (Γ = 1.73+1.15
−1.02),
kT = 0.16 ±0.02 keV, and NH= 0.67 ×1022 cm−2)
considering the degeneracy between kT and NH. The
unabsorbed flux of PSR J2021+3651 in 3 −10 keV is
F3−10 = (1.20+0.18
−0.17)×10−13 erg/s/cm2.
2.3.2. Inner nebula spectrum
We individually and jointly fitted the Chandra, XMM,
and NuSTAR spectra of the inner nebula. The Chan-
dra spectra were extracted from an annulus region with
radii 2 −20′′ centered at PSR J2021+3651. The XMM
spectra were extracted from circular regions with ra-
dius 40′′ using the xmmselect task in SAS. Response
files were generated using rmfgen and arfgen tasks.
NuSTAR spectra were extracted from circular regions
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 7
10 15
10 14
10 13
10 12
E2dN/dE (erg/s/cm2)
1 5 10 20
E (keV)
2.5
0.0
2.5
Figure 7. Chandra ObsID 8502 (black) and 7603 (red),
XMM MOS1 (green) and MOS2 (blue), NuSTAR FPMA
(magenta) and FPMB (cyan) spectra of the inner nebula.
The background is dominant outside of 0.5−7 keV for Chan-
dra, 0.5−8 keV for XMM, and 3 −20 keV for NuSTAR. The
source spectra were extracted from an annulus region with
radii 2 −20′′ for Chandra, a circular region with radius 40′′
for XMM, and a circular region with radius 1′for NuSTAR,
all centered at PSR J2021+3651. The background spectra
were extracted from a 2′×2′box in a source-free region.
For XMM and NuSTAR, the pulsar spectra were subtracted.
The best-fit models are displayed as solid lines in both plots.
with radius 1′using nuproducts task in NuSTARDAS.
The sizes of the source region for different instruments
were determined considering the PSF sizes of the in-
struments ((HPD 1′′ for Chandra , 34′′ for XMM, and
58′′ for NuSTAR) and the cross-normalization term be-
tween the source spectra. The background spectra for
all three instruments were taken from a 2′×2′box in a
nearby source-free region. Fitting was performed in the
energy range where the source emission dominates over
the background (0.5−7 keV for Chandra, 0.5−8 keV for
XMM, 3−20 keV for NuSTAR). All spectra were binned
such that the source counts have at least 3σsignificance
above the background counts in each bin.
We first modeled the Chandra spectra of the inner
nebula using an absorbed power law (const*tbabs*pow)
to find the best-fit NH= (0.76 ±0.06) ×1022 cm−2. Us-
ing this NH, the best-fit model for the pulsar was found
(kT = 0.13 ±0.01 keV, Γ = 1.63 ±0.17, see §2.3.1).
This pulsar component was included and held fixed in
the model for the XMM and NuSTAR spectra of the
inner nebula. The NHfor both instruments were held
fixed to the value found from the Chandra-only fit. The
best-fit Γ for Chandra and XMM are in good agreement
(1.25 ±0.06 and 1.35 ±0.03, respectively), while the
NuSTAR spectra give much softer Γ = 1.73 ±0.07. We
jointly fitted the Chandra , XMM, and NuSTAR spectra
10 12
2 × 10 12
3 × 10 12
4 × 10 12
6 × 10 12
E2dN/dE (erg/s/cm2)
3 4 5
E (keV)
2.5
0.0
2.5
Figure 8. XMM MOS1 and MOS2 (green and blue, respec-
tively) spectra of the outer nebula. The source spectra were
extracted from the dashed polygon in Figure 4. The Line
and continuum backgrounds are dominant below 2 keV and
above 6 keV, respectively. The best-fit models are displayed
as solid lines.
to constrain the model for the inner nebula more tightly
and to test the presence of a spectral break. The it-
erative fitting of the pulsar spectrum was performed in
parallel (see §2.3.1). An absorbed power law model with
NH= (0.96 ±0.04) ×1022 cm−2and Γ = 1.49 ±0.03 ex-
plains the spectra well (χ2/d.o.f = 710/705). Adding a
break to the power law does improve the fit (F test prob-
ability = 0.002); however, the break energy Ebreak =
6.02+0.75
−1.19 keV near the borderline between the Chandra
and XMM vs. NuSTAR energy ranges is suspect. We
concluded that the hint of spectral break might origi-
nate from the imperfect cross-calibration between the
different instruments. The unabsorbed flux of the in-
ner nebula is F3−10 = (5.31 ±0.21) ×10−13 erg/s/cm2.
The best-fit Γ for the inner nebula from the joint fit is
comparable with Van Etten et al. (2008) (1.45 ±0.09).
2.3.3. Outer nebula spectrum
We used the XMM data to analyze the outer nebula
spectrum. The cleaned event files and the Swiss cheese
masks (see §2.2.3) were processed with the mos-spectra
task to extract the source spectra from the dashed poly-
gon in Figure 4. We followed the procedures described
in the manual for the use of XMM-ESAS1to carefully
estimate the background in such a large source region
(∼10′). First, the QPB spectra were generated for the
same region using the mos back task. Second, the back-
ground spectrum below 2 keV contains multiple instru-
mental and solar wind charge exchange (SWCX) lines.
1https://heasarc.gsfc.nasa.gov/FTP/xmm/software/xmm-esas/
xmm-esas- v13.pdf
8Woo et al.
Table 1. Summary of X-ray spectral analysis results
Region Instrument†Energy NHkT ΓF3−10 χ2/d.o.f
(keV) (1022 cm−2) (keV) (10−13 erg/s/cm2)
Pulsar C 0.5−7 0.96‡0.11 ±0.01 1.77 ±0.17 1.20+0.18
−0.17 128/151
C 0.5−7 0.76 ±0.06 −1.26 ±0.06 6.80+0.32
−0.31 345/349
Inner X 0.5−8 0.76‡−1.35 ±0.03 5.58+0.18
−0.17 246/240
nebula N 3 −20 0.76‡−1.73 ±0.07 5.50 ±0.22 106/114
C+X+N 0.5−20 0.96 ±0.04 −1.49 ±0.03 5.31 ±0.21 710/705
Outer nebula X 2 −6 0.96‡−1.82 ±0.03 32.53 ±0.69 509/541
†C = Chandra, X = XMM, N = NuSTAR.‡held fixed.
We chose 2 −6 keV for the energy range of our anal-
ysis to avoid modeling too many background compo-
nents. The continuum background dominates over 6
keV. Third, we attempted to model the remaining back-
ground components on Xspec: SP residuals and the cos-
mic X-ray background (CXB). The QPB spectra were
loaded as background spectra. The source spectra were
binned to have at least 3σsignificance over the QPB
spectra in each bin. The background from SP residuals
was modeled with a power law using unitary response
matrices, but none of the model parameters were con-
strained. Therefore we assumed that the contribution
from residual SP is insignificant and excluded it from the
model. The CXB was modeled with an absorbed power
law (const*tbabs*pow). All of its model parameters
were fixed to the canonical values (Γ = 1.41, normal-
ization=11.6 photons/keV/s/cm2/sr at 1 keV, De Luca,
A. & Molendi, S. (2004)) to circumvent the degeneracy
between the two power-law components (CXB and the
outer nebula). We used the Galactic hydrogen column
density2at the center of the source extraction region
(1.13 cm−2) as the NHfor the CXB.
The outer nebula was modeled with an absorbed
power law (const*tbabs*pow). The NHis not con-
strained in the energy range of this analysis (2 −6 keV).
We fixed the NHto the best-fit value found from the
Chandra, XMM, and NuSTAR joint fit of the inner
nebula (0.96 ×1022 cm−2). The best-fit model with
Γ = 1.82 ±0.03 yields a reasonable fit (χ2/d.o.f =
509/541). The unabsorbed flux of the outer nebula in
2−10 keV is F2−10 = (4.20 ±0.07) ×10−12 erg/s/cm2.
The best-fit Γ is clearly harder than Mizuno et al. (2017)
(Γ = 2.10 ±0.12), yet the flux value is comparable
(F2−10 ∼4.1×10−12 erg/s/cm2for the PWN-west).
3. FERMI ANALYSIS
The gamma-ray pulsations of PSR J2021+3651 was
first detected by AGILE (Halpern et al. (2008)) and later
2https://www.swift.ac.uk/analysis/nhtot/
confirmed by Fermi (Abdo et al. (2009)). The pulsar is
registered in the most recent Fermi -LAT source cata-
log (4FGL-DR3, Fermi-LAT collaboration et al. (2022))
as 4FGL J2021.1+3651. We analyzed 13-year Fermi
data (August 2008 −October 2021, MET 239557417 −
656813666) to detect the GeV emission from the Drag-
onfly. We selected SOURCE class and FRONT+BACK
type events (evclass=128, evtype=3) and used the in-
strument response functions (IRFs) P8R3 SOURCE V3.
The 90◦zenith angle cut and the filter expression
DATA QUAL>0 && LAT CONFIG==1 were applied. The re-
gion of interest (ROI) is a 10◦×10◦box region centered
at 4FGL J2021.1+3651.
We performed a binned likelihood analysis (spatial
bin = 0.1◦, energy bin = 8 bins per decade) using
Fermipy v1.0.1 (Wood et al. (2017)). The ROI model
includes the 4FGL-DR2 sources (gll psc v27.fit,Bal-
let et al. (2020)) within a 30 ×30 box region centered
on 4FGL J2021.1+3651, the Galactic diffuse emission
model (gll iem v07.fits), and the isotropic emission
model (iso P8R3 SOURCE V3 v1.txt)3. We used the
optimize() and fit() methods in Fermipy to optimize
the model in 100 MeV−300 GeV. For the fit() method,
the parameters of bright nearby sources (within 5◦of
4FGL J2021.1+3651 and TS (test statistics) >25) were
left free. After fitting, the residual map was visually in-
spected, and a standard normal distribution was fitted
to the residual significance histogram to ensure that the
residuals are statistical fluctuations.
4FGL J2021.1+3651 is modeled with PLSuperExp-
Cutoff24. The best-fit parameters of 4FGL
J2021.1+3651 agreed with those of 4FGL-DR2 within
1σerror. Since the emission from 4FGL J2021.1+3651
cuts off in the 10−30 GeV range, we created a √TS
map of the ROI above 30 GeV to avoid contamination
3https://fermi.gsfc.nasa.gov/ssc/data/access/lat/
BackgroundModels.html
4https://fermi.gsfc.nasa.gov/ssc/data/analysis/scitools/
source models.html
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 9
by the pulsar and investigate any possible diffuse emis-
sion from the PWN. We did not find any excess in the
vicinity of 4FGL J2021.1+3651 that can be attributed
to the emission from the PWN. This result confirms the
non-detection of the GeV PWN in the vicinity of PSR
J2021+3651 from the previous studies of the off-pulse
data (Abdo et al. (2009)).
GeV gamma rays are IC upscattered photons off the
low energy electrons that emit synchrotron radiation in
radio −infrared. Given the large (>20′) size of the
radio nebula, a putative GeV nebula may be largely ex-
tended and too faint to be significantly detected over the
background. The large (∼1◦) size of the IC nebula in
the VHE range (eHWC J2019+368) may also indicate a
largely extended GeV nebula. Acero et al. (2013) calcu-
lated upper limits for a GeV PWN of PSR J2021+3651
assuming a size of MGRO J2019+37. Di Mauro et al.
(2021) used an ICS template with the best-fit diffusion
coefficient for eHWC J2019+368 to place GeV upper
limits of the PWN. Both works resulted in GeV upper
limits similar to the flux of eHWC J2019+368.
4. MW SED MODELING
Figure 9shows the MW counterparts of the Drag-
onfly overlaid on the HAWC significance map. PSR
J2021+3651 is located at the Eastern edge of the
extended TeV source eHWC J2019+368. Its PWN,
the Dragonfly, extends toward the centroid of eHWC
J2019+368. We model the SED of the Dragonfly us-
ing these MW data to investigate the Dragonfly as a
leptonic PeVatron.
While our NuSTAR observation allowed an in-depth
study of the inner nebula, the faint emission from the
outer nebula was not detected due to the limited sen-
sitivity. Instead, we used our XMM analysis result of
the outer nebula presented in §2.3.3. The radio spec-
trum and the GeV upper limits were taken from Roberts
et al. (2008) and Di Mauro et al. (2021) (“IEM-4FGL”),
respectively. In the TeV band, three independent
flux measurements by VERITAS (VER J2019+368),
HAWC (eHWC J2019+368), and LHAASO (LHAASO
J2018+3651) are available. Cao et al. (2021a) did
not provide detailed spectral information of LHAASO
J2018+3651 except for its flux at 100 TeV. Abeysekara
et al. (2018) and Abeysekara et al. (2020) provide the
spectrum of VER J2019+368 and eHWC J2019+368,
respectively, over three decades of energy. While VER
J2019+368 and eHWC J2019+368 exhibit similar source
size, the flux of VER J2019+368 reported in Abeysekara
et al. (2018) was extracted from a region smaller than
the source size, yielding a 2−3 times lower flux than that
of eHWC J2019+368 in the overlapping energy range
20h24m21m
38°00'
37°30'
00'
Right Ascension
Declination
PSR J2021+3651
Sh 2-104
eHWC J2019+368
Dragonfly (XMM)
Dragonfly (VLA)
Figure 9. HAWC significance map (2◦×2◦) centered at
the centroid of eHWC J2019+368. The eHWC J2019+368
flux extraction region is marked with a dashed circle (radius
0.5◦). PSR J2021+3651 and Sh 2-104 are marked with a
star and a triangle, respectively. The shaded region is the
XMM spectrum extraction region of the outer nebula. The
dotted line shows the extent of the outer nebula seen by
VLA. The angular resolution of HAWC varies depending on
the energy and zenith angle. The approximate size of the
68% containment region at 56 TeV (radius 0.2◦) is marked
with a solid white circle at the top left corner.
The HAWC image was obtained from the 3HWC survey
public data (https://data.hawc-observatory.org/datasets/
3hwc-survey/fitsmaps.php), and the VLA nebular extent was
estimated from Roberts et al. (2008).
(see also Albert et al. (2021)). Therefore, we used the
spectrum of eHWC J2019+368 from Abeysekara et al.
(2020) for SED modeling in this work.
The distance estimates of PSR J2021+3651 vary
widely depending on different distance measures.
Roberts et al. (2002) used the dispersion measure (DM)
∼371 pc cm−3to put the pulsar at ≥10 kpc on the outer
edge of the Galaxy, although they left a possibility of a
nearer distance in case of a contribution from excess gas
in the Cygnus region. Van Etten et al. (2008) suggested
3−4 kpc based on various arguments, such as the X-ray
spectral fit to a neutron star atmosphere model and the
gamma-ray efficiency of the pulsar. Abdo et al. (2009)
estimated a distance ≥4 kpc based on the pulsar rota-
tion measure (RM). Kirichenko et al. (2015) suggested
1.8+1.7
−1.4kpc using the interstellar extinction and distance
relation. The 1.8 kpc distance was adopted by Mizuno
et al. (2017), Fang et al. (2020), and Albert et al. (2021)
for their MW SED modeling.
10 Woo et al.
Table 2. Best-fit SED model parameters using Naima
α2.4
Ecut 0.9 PeV
Magnetic field 1.6 µG
IR temperature 26 K
IR energy density 1.0 eV/cm3
Total particle energy†6.1d2
3.5×1049 erg
†d3.5is the distance to the Dragonfly scaled to the
nominal distance of 3.5 kpc.
We start with preliminary modeling of the MW SED
using Namia (Zabalza (2015)), a generic model for non-
thermal radiation from relativistic particles. We do not
make assumptions about the distance or evolutionary
history of the PWN at this stage. Our purpose for
this preliminary modeling is to provide a basic under-
standing of the current status of the PWN and initial
estimation of the input parameters for a more sophis-
ticated model, namely the dynamical model (Gelfand
et al. (2009)). Then we move on to MW SED modeling
using the dynamical model to acquire insight into the
dynamical evolution of the Dragonfly over its lifetime
while the interactions between the PWN, SNR, and ISM
are accounted for.
4.1. Naima
Naima allows us to characterize the current particle
population using a minimal number of parameters with-
out introducing any physical assumptions on the evolu-
tionary history of the system.
For a leptonic particle accelerator, synchrotron (SC)
emission and emission via inverse Compton (IC) scatter-
ing off the input seed photon fields (cosmic microwave
background (CMB) and interstellar dust emission (in-
frared, or IR)) are calculated based on a particle distri-
bution model. We vary the model parameters of a single
particle distribution so that the SC and IC spectra are
consistent with the observed flux in radio, X-ray, and
TeV gamma-ray bands. The minimum particle energy
(Emin) and the reference energy (E0) were fixed to 1
MeV and 1 TeV, respectively. The best-fit parameters
are summarized in Table 2, and the best-fit SED model
is plotted with the MW data and residuals in Figure 10.
The TeV spectrum exhibits a smooth cutoff after 20
TeV. This cutoff is better explained by an exponential
cutoff power law distribution of particles, dN/dE =
A(E/E0)−αe−E /Ecut , than a simple power law with a
sharp cutoff at Emax. The best-fit particle index is
α= 2.4. Adding an IR field gives a better fit than the
CMB-only model, although the IR field energy density
tends to grow indefinitely to an unphysical value. There-
10 14
10 13
10 12
10 11
E2dN/dE (erg/s/cm2)
Total
CMB
SSC
IR
Sync
VLA
XMM-Newton
Fermi-LAT
HAWC
10 710 410 11021051081011 1014
E (eV)
2.5
0.0
2.5
Figure 10. The best-fit SED model from Naima. The model
parameters are given in Table 2. Synchrotron flux (Sync),
inverse Compton flux from the CMB (CMB), infrared emis-
sion from the interstellar dust grains (IR), synchrotron self-
Compton component (SSC), and total flux are plotted. The
SSC flux level is very low and located below the lower bound
of the y-axis. Radio, X-ray, and TeV flux data points, and
GeV upper limits are overlaid. The residuals are plotted
in terms of significance ((data−model) / (1σuncertainty of
data)).
fore we fixed the IR field energy density to the average
cosmic ray energy density (1 eV/cm3,Cummings et al.
(2016)). This yields the best-fit IR field temperature T
= 26 K, magnetic field B = 1.6 µG, and cutoff energy
Ecut = 0.9 PeV. This SED model shows a good fit to
the MW data as seen in the residuals plotted in Figure
10; however, the narrow IC peak resolved by HAWC is
difficult to explain with physically reasonable model pa-
rameters. Such a narrow peak originates from the flux
point in the lowest energy bin, the range in which air
imaging Cherenkov telescopes (IACTs), such as VERI-
TAS, are more sensitive. A deeper VERITAS observa-
tion and more accurate flux measurement of the region
have been proposed for our future work to resolve the
IC spectrum of the Dragonfly better.
The best-fit cutoff energy Ecut = 0.9 PeV strongly
suggests that the Dragonfly is likely a PeVatorn. The
cutoff energy is greater than 0.3 PeV by Albert et al.
(2021) and 0.4 PeV by Fang et al. (2020) mainly due
to the difference in the X-ray spectra used in each work
(see §4). The magnetic field inside the PWN is at the
level of interstellar magnetic field (∼3µG, Jansson &
Farrar (2012), as one may infer from the low X-ray to
gamma-ray luminosity ratio.
The IR field temperature is higher than the IR emis-
sion from cold dust grains (∼15 K). A possible source
of this warm dust emission is an H II region Sh2-104
(marked with an inverted triangle in Figure 9). Lo-
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 11
cated at 4 ±0.5 kpc from the Earth, Sh2-104 is visi-
ble in radio through X-ray (Deharveng, L. et al. (2003),
Rod´on et al. (2010), Xu et al. (2017), and Gotthelf et al.
(2016)). Sh 2-104 hosts an ultra-compact H II (UCHII)
region on the eastern periphery of its dense molecular
shell. As strong candidates for active star formation
(Deharveng, L. et al. (2003) and Xu et al. (2017)), Sh2-
104 and the UCHII region each contain a stellar clus-
ter, MASS J20174184+3645264 and IRAS 20160+3636,
respectively, ionizing the regions (Deharveng, L. et al.
(2003) and Paredes et al. (2009)). Rod´on et al. (2010)
used Herschel observations to estimate the dust temper-
ature in Sh2-104 to be 20 −30 K on the exterior. Our
best-fit IR field temperature (26 K) lies in this temper-
ature range. Since the lower bound of the distance to
Sh2-104 (Deharveng, L. et al. (2003)) and the upper
bound of the distance to PSR J2021+3651 Kirichenko
et al. (2015) coincide (3.5 kpc), we adopt 3.5 kpc as a
nominal distance to the Dragonfly hereafter and scale
relevant parameters to this distance whenever possible.
4.2. Dynamical model
Taking the result from Naima as a starting point, we
fit the dynamical model to the MW data of the Dragon-
fly. The dynamical model is a time-evolutionary model
for a composite system of a PWN and its host SNR.
The model assumes a spherical single-zone system whose
SNR is in a free-expansion or Sedov-Taylor phase. The
model evolves a PWN and its SNR from their birth
to the true age of the system, calculating the interac-
tion between them and with the surrounding interstel-
lar medium (ISM). The model output includes a pulsar
wind particle distribution, its synchrotron and inverse
Compton emission spectrum, and the dynamics of a sys-
tem (e.g., a radius of a PWN, a radius of an SNR forward
and reverse shock, and a magnetic field inside a PWN)
at each evolutionary phase.
The dynamical model evolves the particle distribution
inside a PWN via three mechanisms: continuous particle
injection of a fraction of the pulsar spin-down luminos-
ity, adiabatic energy loss due to the expansion of the
PWN, and radiative energy loss due to synchrotron and
inverse Compton emission. The spin-down luminosity ˙
E
of a PWN at its age tis formulated as
˙
E(t) = ˙
E01 + t
τsd −p+1
p−1
,(1)
where ˙
E0is the initial spin-down luminosity, pis the
pulsar braking index, and τsd is the characteristic pulsar
spin-down timescale. τsd is related to a pulsar’s charac-
teristic age τand true age tage as
tage =2τ
p−1−τsd.(2)
Table 3. SED model parameters using the dynamical model
Source distance 3.5 kpc
Input
ESN 1.0×1051 erg
Mej 7.2M⊙
nIS M 0.03 cm−3
Braking index p2.5
Age 16 kyr
ηB0.008
Emax 1.4 PeV
Particle index α2.4
IR field temperature 9.9 K
IR field energy density 1.4 eV cm−3
Output
tRS 12 kyr
RPWN 9.5 pc
Magnetic field 2.7 µG
Total particle energy 3.9×1048 erg
A fraction of the spin-down luminosity ηB˙
Eis injected
into the PWN as magnetic fields, while the rest of the
spin-down luminosity, (1 −ηB)˙
E, is injected as particles
(electrons and positrons). The particle injection spec-
trum is defined within the energy range between Emin
(minimum energy of the injected particle, fixed to 0.1
GeV in this work) and Emax (maximum energy of the
injected particle). The magnetic field is assumed to be
homogeneous throughout the volume of the PWN, and
thus it decreases as the PWN expands and increases
as the PWN is crushed by the collision with the SNR
reverse shock. The radiative loss changes accordingly
−synchrotron emission is much stronger than inverse
Compton emission in the early stage or post-collision
era of the PWN, whereas inverse Compton flux becomes
comparable with synchrotron flux as the PWN ages.
Adiabatic loss is most severe when a PWN freely ex-
pands against only ram pressure from the unshocked
SN ejecta in the free-expansion phase. Once the PWN
collides with the SNR reverse shock and starts encoun-
tering the pressure from shocked ejecta, the expansion
of the PWN slows down until the compression starts,
during which the PWN undergoes adiabatic heating.
The dynamical evolution of a system is calculated
based on input parameters related to a supernova (SN),
SNR, and surroundings, such as SN explosion energy
(ESN ), ejecta mass inside an SNR (Mej ), and ISM den-
sity just outside an SNR forward shock (nISM ). These
parameters determine the pressure just outside a PWN
and at the location of a reverse shock. A PWN size
changes such that the pressure from the ejecta is in bal-
ance with the pressure inside a PWN, which comprises
magnetic pressure and pressure from particles as a rela-
tivistic ideal gas. When a reverse shock reaches a PWN,
12 Woo et al.
10 14
10 13
10 12
10 11
E2dN/dE (erg/s/cm2)
VLA
XMM-Newton
Fermi-LAT
HAWC
10 710 410 11021051081011 1014
E (eV)
2.5
0.0
2.5
(a)
0 2000 4000 6000 8000 10000 12000 14000 16000
Time (year)
0.0
2.5
5.0
7.5
10.0
12.5
15.0
17.5
20.0
Distance (pc)
100
101
102
103
104
105
B field (uG)
R
PWN
R
RS
B field
(b)
10 610 21021061010 1014
E (eV)
10 18
10 16
10 14
10 12
10 10
10 8
E2dN/dE (erg/s/cm2)
1600
3200
4800
6400
8000
9600
11200
12800
14400
16000
Time (yr)
(c)
1041061081010 1012 1014 1016
E (eV)
1020
1023
1026
1029
1032
1035
1038
E2dN/dE (erg)
1600
3200
4800
6400
8000
9600
11200
12800
14400
16000
Time (yr)
(d)
Figure 11. (a) The SED model for the Dragonfly from the dynamical model. The model parameters are given in Table 3.
Radio, X-ray, and TeV flux data points are overlaid. The residuals are plotted in terms of significance ((data−model)/(1σ
uncertainty of data)). (b) Time evolution of the PWN radius (Rpwn, blue solid line), SNR reverse shock radius (RRS, orange
dashed line), and magnetic field inside the PWN (B field, green dotted line) over the true age from our model (16 kyr). The
PWN collided with the SNR reverse shock at tRS = 12 kyr. (c)(d) Time evolution of the radiation (c) and particle (d) spectrum
of the Dragonfly.
the pressure experienced by a PWN increases dramat-
ically. This leads to a rapid decrease in the size of a
PWN and, consequently, a sharp increase in the mag-
netic field inside a PWN. Synchrotron loss is extreme
at this point. Once a PWN is highly compressed, and
its pressure exceeds the ejecta pressure, the PWN starts
re-expanding, and hence the magnetic field inside the
PWN starts decreasing. As a PWN ages, its size grows
as large as several parsecs or more, and its magnetic
field becomes as low as a few µG. The resulting spec-
trum yields a similar level of SC and IC flux, a typical
spectrum observed in middle-aged PWNe.
We aimed to find a set of model parameters that re-
produces not only the MW spectrum but also the ob-
served size of the Dragonfly. The Dragonfly displays a
highly asymmetric morphology, although the dynamical
model assumes a spherical system. Given this limitation
of the model, we focus on characterizing the evolution
of the PWN properties averaged over the entire system
rather than their spatial dependency. We approximated
a nominal radius of the PWN to be 10′so that its spher-
ical volume roughly matches the physical volume of the
outer nebula. The radio nebula size ∼20′was used for
this calculation −the lowest energy particles seen in the
radio band would have a much longer lifetime than its
synchrotron cooling time and thus best reflect the true
extent of the PWN. The angular size of 10′is equivalent
to RPWN = 10 pc at the nominal distance d= 3.5 kpc.
The input and output parameters of the dynamical
model are listed in Table 3. Figure 11 shows the SED
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 13
Table 4. Comparison between PeVatron pulsar wind nebulae
G75.23+0.12 (“Dragonfly”) G18.5-0.4 (“Eel”)aG106.65+2.96 (“Boomerang”)b
TeV counterpart eHWC J2019+368 HAWC J1826-128 HAWC J2227+610
LHAASO J2018+3651 LHAASO J1825-1326 LHAASO J2226+6057
Pulsar
Name PSR J2021+3651 PSR J1825-1256 PSR J2220+6114
τ(kyr) 17 14 10
˙
E(1036 erg/s) 3.4 3.6 22
Distancec(kpc) 0.4 −12 (3.5) 3.5 0.8 −7.5 (7.5)
PWN
Emax (PeV) 1.4 4.6 3.3
Magnetic field (µG) 2.7 0.6 2.2
True age (kyr) 16 4.6 3.3
tRS (kyr) 12 −1.5
aBurgess et al. (2022) and references therein. bPope et al. (submitted to ApJ) and references therein.
cFor the sources with a wide range of distance estimates, the distance used for SED modeling is given in parentheses.
model plotted with the MW data and residuals, the
time evolution of the dynamical parameters (magnetic
field, PWN radius RPWN, and SNR reverse shock radius
RRS ), radiation and particle spectra.
The maximum particle energy Emax = 1.4 PeV pro-
vides strong evidence that the Dragonfly is a PeVatron
PWN. The true age of 16 kyr was found to be slightly
younger than its characteristic age τ= 17 kyr. The true
age found by our model is much older than 7 kyr found
by Albert et al. (2021). This difference can be attributed
most likely to the assumed source distance (3.5 kpc in
this work, 1.8 kpc in Albert et al. (2021)), as well as to
the SED models and the X-ray spectra (see §4). The
low magnetic field (2.7 µG) is consistent with that from
Naima,Mizuno et al. (2017), Fang et al. (2020), and Al-
bert et al. (2021). The low magnetic fraction ηB= 0.008
contributes to this low magnetic field.
Our model predicts that the PWN expanded to ∼20
pc, collided with the SN reverse shock 4 kyrs ago (tRS =
12 kyr), and has been shrinking since then to reach the
current size ∼10 pc (Figure 11 (b)). Relatively low
ISM density nISM = 0.03 cm−3drove a slow reverse
shock, allowing the PWN to grow large enough to reach
the reverse shock even before it started heading back
toward the PWN. Combined with the substantial ejecta
mass Mej = 7.2M⊙, this low ISM density may indicate
that the host SNR of the Dragonfly evolved into the
wind-blown bubble of a massive progenitor star with an
extremely low density (below 0.001 cm−3) during the
first few kyrs of its lifetime (Dwarkadas (2005)).
The particle index α= 2.4 is consistent with Naima.
The IR field temperature (9.9 K) falls below the range
of the dust temperature in Sh2-104 (see §4.1). Using the
braking index p = 2.5 and τsd = 6.8 kyr of our model,
the total particle energy (3.9×1048 erg) is 50% of the
total injected energy over the true age (16 kyr) of the
Dragonfly.
5. DISCUSSION
5.1. PeVatron pulsar wind nebulae
We compare three PeVatron PWNe studied in our
NuSTAR observational program of energetic PWNe:
G75.23+0.12 (“Dragonfly”, this work), G18.5-0.4
(“Eel”, Burgess et al. (2022)), and G106.65+2.96
(“Boomerang”, submitted to ApJ). All three PWNe
were modeled with the dynamical model. Key facts and
the model parameters of the three PWNe are summa-
rized in Table 4. The common features of the three
PeVatron PWNe are the following:
1. The maximum particle energy is greater than 1
PeV.
2. The source morphology is highly asymmetric and
energy-dependent. The pulsar is located on the
edge of the extended radio and soft X-ray nebulae
and is offset from the centroid of its TeV counter-
parts.
3. The magnetic field strength inside the PWNe is
low <3µG.
4. Compact hard X-ray nebula was detected up to 20
keV by NuSTAR. The nebular size is much smaller
than the lower-energy nebula (radio and soft X-
ray).
5. The compact hard-Xray nebular spectrum does
not exhibit a spectral break or cutoff. A syn-
chrotron burnoff is observed from the shrinkage
of its size at higher energies.
6. The dynamical model predicts that the PWN col-
lided with the reverse shock of the host SNR “re-
14 Woo et al.
cently” (the Dragonfly and Boomerang), or such a
collision is about to happen (the Eel).
2and 3are known properties of bright TeV PWNe
(e.g., Kargaltsev et al. (2013) and Torres (2017)). While
asymmetric morphology bears a few different possibili-
ties (fast pulsar velocity, asymmetric SNR reverse shock
−PWN interaction, or a combination of both effects
−see §5.2), energy dependency of morphology can be
attributed to particle transport and cooling. Particles
that emit synchrotron radiation in the radio band (par-
ticle energy Eγ∼1 GeV in the interstellar magnetic
field ∼3µG) have cooling times much longer than the
age of the PWN and hence transport to large distances
away from the pulsar without losing much of their en-
ergies. Particles that radiate in the hard X-ray band
(Eγ∼100 TeV), on the other hand, have much shorter
cooling times than the PWN age. Such particles can
travel only to short distances before cooling down to
lower energies, resulting in 4(see §5.3). Therefore, only
freshly injected highly energetic particles contribute to
the compact hard X-ray nebula. Relic particles, after
cooling, exhibit larger extents in lower energies.
Looking at an IC spectrum, relic particles with ener-
gies Eγ∼10 TeV upscatter the CMB photons to TeV
energies. Such particles can be dim in the synchrotron
spectrum due to a lower magnetic field farther away
from the pulsar, explaining the offset of the TeV emis-
sion from the pulsar. 3is necessary for this reason and
manifests as the observed low X-ray to gamma-ray lu-
minosity ratios. GeV-emitting particles (Eγ∼10 GeV)
are expected to form even fainter SC and IC nebulae
due to their lower energies and larger distances trav-
eled. No GeV nebulae were detected for the PeVatron
PWNe except for the “tail” region of the Boomerang
whose emission is attributed to its parent SNR (e.g.,
Fang et al. (2022)).
6provides a hint as to how particles are accelerated
in PeVatorn PWNe. Ohira et al. (2018) proposed using
a Monte Carlo simulation that particles may be acceler-
ated to 1 PeV during the compression of a PWN by the
collision with the reverse shock of its host SNR.
Some of the above properties are in contrast to those
of other TeV PWNe in our NuSTAR observational cam-
paign, such as G313.54+0.23 (“K3” or “Kookaburra”,
Park et al. (2023b)) and G313.3+0.1 (“Rabbit”, Park
et al. (2023a)). These southern sources are invisible to
HAWC and LHAASO, the telescopes that operate in
the highest energy regime (>100 TeV). Their NuSTAR
hard X-ray nebulae are extended (radius ∼3′at the
source distance ∼5.6 kpc for both PWNe), and the
nebular sizes do not change significantly with energy.
Multi-zone SED modeling using the spatially resolved
NuSTAR spectra along with MW flux data yielded the
maximum particle energies below 1 PeV for both PWNe.
The PWNe have bright GeV counterparts whose spectra
connect smoothly to the spectra of their TeV counter-
parts. Future gamma-ray observatories in the southern
hemisphere, the Southern Wide-field Gamma-ray Obser-
vatory (SWGO) and Cherenkov Telescope Observatory
(CTAO) −South, will play a crucial role in studying
the true energetics of these PWNe and their relation to
the MW observations.
5.2. Distance and proper motion
The distance of the Dragonfly is relevant to not only
its brightness but also its physical size, and hence the
proper motion of PSR J2021+3651. Like many other
TeV PWNe (e.g., H.E.S.S. Collaboration et al. (2018)),
the Dragonfly is offset from the center of the nebula and
its TeV counterparts. Such highly asymmetric morphol-
ogy is often attributed to a fast proper motion of the
pulsar. Aliu et al. (2014) estimated the transverse ve-
locity of PSR J2021+3651 to be 840d5t−1
17 km/s, where
d5is the distance of the Dragonfly scaled to 5 kpc and
t17 is the age of PSR J2021+3651 scaled to its charac-
teristic age of 17 kyr, in case it was born at the end of
the radio nebula. Albert et al. (2021) estimated it to be
∼1,300 km/s in case PSR J2021+3651 is located at 1.8
kpc and was born 7 kyrs ago at the location of HAWC
J2019+368. Van Etten et al. (2008) and Jin et al. (2023)
claimed the detection of a bow shock structure on the
East side of the Dragonfly, yet noted a possibility of
at most a mildly supersonic motion of PSR J2021+3651
considering the well-preserved substructures of the inner
nebula.
Given the true age of 16 kyr from our model, the
Dragonfly may be too young to have escaped its host
SNR and form a bow shock in the ISM (e.g., Gaensler
& Slane (2006)). Instead, the bow shock structure with
the well-defined inner nebula and the asymmetric PWN
morphology could be explained by an asymmetric in-
teraction between the PWN and the reverse shock of
its host SNR due to an ambient density gradient (e.g.,
Temim et al. (2015) and Temim et al. (2017)). In this
case, the orientation of the bow shock does not necessar-
ily align with the direction of the pulsar’s proper motion
(Kolb et al. (2017)). There is no known dense object on
the East of PSR J2021+3651, and the host SNR has
not been detected. This mystery could be solved by a
deep and expansive radio observation that covers a large
(∼1◦) region to search for the faint host SNR of PSR
J2021+3651. Here, we focus on discussing the proper
motion of PSR J2021+3651 that may have caused the
asymmetric morphology of the Dragonfly.
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 15
The angular separation between PSR J2021+3651
and the centroid of eHWC J2019+368 is ∼16′. If
PSR J2021+3651 was born near the centroid of eHWC
J2019+368and traveled to the current location at a con-
stant speed, the corresponding transverse velocity of
the pulsar is vpsr = 996d3.5t−1
16 km/s, where t16 is the
true age scaled to 16 kyr. This is above the aver-
age pulsar velocity (540 km/s, Verbunt et al. (2017)),
but not exceptionally high (Kargaltsev et al. (2017)).
In this case, measuring the pulsar proper motion of
∼0.06′′d3.5t−1
16 /yr may not be feasible unless PSR
J2021+3651 is significantly closer than 3.5 kpc or sig-
nificantly younger than 16 kyr. Another Chandra ob-
servation of PSR J2021+3651, nearly 20 years after the
last observation, to detect the pulsar motion could pro-
vide insight into the source distance and age; however,
their degeneracy will still need to be disentangled. Our
future work will combine new radio (VLA) and X-ray
(Chandra) observation with an energy-dependent mor-
phology study using VERITAS and Fermi-LAT to place
tight constraints on the source distance and evolution-
ary history of the Dragonfly.
5.3. Magnetic field
For particles with a synchrotron lifetime shorter than
the age of the system, the distance that a particle can
travel is determined by its synchrotron lifetime rather
than the system age. In the vicinity of the pulsar
where the magnetic field is strong, and the particles
are transported mainly by energy-independent advec-
tion, the PWN size in different energy bands should be
proportional to the synchrotron lifetime of the electrons
(e.g., Tang & Chevalier (2012)). This is demonstrated
by the changing nebula size in energy observed with
NuSTAR. A synchrotron lifetime tsync can be defined
as a time scale that an electron with Lorentz factor γ
loses all of its energy Eγvia synchrotron radiation in
magnetic field strength B. A synchrotron spectrum of a
single electron is highly peaked around its critical fre-
quency νcrit(γ)∝BE2
γ. A rough estimation of a syn-
chrotron lifetime using this information yields
tsync =Eγ
Psynch ∼γmc2
γ2B2∼1
pB3νcrit(γ).(3)
Assuming a constant average magnetic field and ad-
vection velocity in the region, the ratio of a syn-
chrotron lifetime between particles emitting 3 keV pho-
tons and those emitting 6 keV photons is calculated
as p6 keV/3 keV = 1.4. This ratio is indeed compa-
rable to the ratio of the nebula size in two different
energy bands: FWHM(3−6 keV)/FWHM(6−20 keV)=
26.5′′/15.2′′ = 1.3.
Comparing the nebula sizes in the two energy bands
also allows placing an upper limit of the magnetic field
inside the compact nebula. The inner nebula detected
by NuSTAR is located well outside the termination
shock (2−3 smaller than the torii ∼10′′ (Van Etten
et al. (2008))), where the advection velocity can be ap-
proximated to the overall PWN expansion velocity (e.g.,
Porth et al. (2016)). The expansion velocity of the Drag-
onfly has not been measured, yet some other PWNe were
estimated to expand at ∼1,000 km/s (e.g., Porth et al.
(2016), Reynolds et al. (2018), and Vorster et al. (2013)).
Eq. (6) in Reynolds et al. (2018) gives the time it takes
for an electron to lose half its energy via synchrotron ra-
diation (t1/2). For example, an electron that was emit-
ting 12 keV photons t1/2years ago has cooled down by
now to emit 3 keV photons (νcrit ∝E2). Assuming that
this electron traveled from the edge of the hard-band
nebula (FWHM = 15.2′′) to the edge of the soft-band
nebula (FWHM = 26.5′′) at velocity vad =1,000 km/s,
the constant average magnetic field inside the compact
nebula yields B= 24d−3/2
3.5µG. Compared with the 2.7
µG in the outer nebula, the much stronger magnetic field
for the outer nebula was anticipated from the compact
size of the hard X-ray inner nebula. This magnetic field
estimate is consistent with the inner nebula magnetic
field estimated by Van Etten et al. (2008) (∼20 µG as-
suming a dipolar field) and Jin et al. (2023) (∼22 µG
assuming equipartition between the magnetic field and
particle energy).
6. SUMMARY
As part of our NuSTAR observational campaign of
energetic PWNe, we studied the X-ray properties of the
Dragonfly PWN and its viability as a leptonic PeVatron.
Our NuSTAR observation detected a compact (r = 1′)
inner nebula of the Dragonfly in 3 −20 keV. The size of
this nebula decreases at higher energies, indicating syn-
chrotron burnoff in a strong (∼24 µG) magnetic field
near its pulsar PSR J2021+3651. The large diffuse outer
nebula of the Dragonfly is observed in soft X-ray (∼10′)
and radio (∼2′). We used these outer nebula spectra
along with the TeV spectrum of eHWC J2019+368 to
model the MW SED of the Dragonfly. The dynamical
model yields the maximum particle energy of 1.4 PeV,
and a low magnetic field (2.7 µG) averaged over the
outer nebula in contrast to the high magnetic field in the
inner nebula. At a nominal distance of 3.5 kpc, this 16-
kyr-old PWN was found to have collided with the SNR
reverse shock 4 kyrs ago. The highly asymmetric and
energy-dependent morphology of the Dragonfly implies
a fast proper motion of its pulsar (∼1,000 km/s) and/or
inhomogeneity in the ISM that initiated an asymmetric
16 Woo et al.
SNR −PWN interaction. Our future work will investi-
gate these scenarios and provide a deeper understanding
of particle transport in such an evolved system using ra-
dio, X-ray, and gamma-ray observations.
The Dragonfly shares common features with other
PWNe in our NuSTAR observational campaigns with
the maximum particle energies above 1 PeV −the Eel
(Burgess et al. (2022)) and Boomerang (Pope et al. sub-
mitted to ApJ). These features include a compact hard
X-ray inner nebula undergoing synchrotron burnoff, a
large diffuse outer nebula in lower energy, and an ab-
sence of a GeV nebula. Opposite patterns are observed
in two of our target PWNe, the K3 (Park et al. (2023b))
and Rabbit (Park et al. (2023a)). These PWNe exhibit
extended hard X-ray nebulae without a sign of syn-
chrotron burnoff, energy-insensitive morphologies, and
bright GeV nebulae. The best-fit multi-zone models
of the two PWNe yield the maximum particle ener-
gies below 1 PeV, while the PWNe are invisible to the
current UHE observatories. The next-generation UHE
observatories in the southern hemisphere (SWGO and
CTAO−South) will enable us to study the true energet-
ics of the PWNe and its relation to the MW observa-
tions.
We thank Mattia Di Mauro for providing the GeV up-
per limits. We acknowledge Ruo-Yu Shang, Eric Got-
thelf, and Jordan Eagle for their helpful discussions. We
thank the referee for carefully reading our manuscript
and providing valuable comments. Support for this work
was partially provided by NASA through NuSTAR Cy-
cle 6 Guest Observer Program grant NNH19ZDA001N.
HA acknowledges support from the National Research
Foundation of Korea (NRF) grant funded by the Korean
Government (MSIT) (NRF-2023R1A2C1002718).
Facilities: NuSTAR,Chandra, XMM-Newton,
Fermi -LAT
Software: NuSTARDAS (v2.0.0), CIAO (4.13; Frus-
cione et al. (2006)), SAS (v20.0.0; Gabriel et al. (2004)),
Fermipy (v1.0.1; Wood et al. (2017)), HEASoft (Nasa
High Energy Astrophysics Science Archive Research
Center (Heasarc) (2014)), HENDRICS (7.0; Bachetti
(2018)), Stingray (v1.1; Huppenkothen et al. (2019)),
Xspec (Arnaud (1996)), Naima (Zabalza (2015))
REFERENCES
Abdo, A. A., Allen, B., Berley, D., et al. 2007, ApJL, 658,
L33, doi: 10.1086/513696
Abdo, A. A., Ackermann, M., Ajello, M., et al. 2009, The
Astrophysical Journal, 700, 1059,
doi: 10.1088/0004-637x/700/2/1059
Abeysekara, A. U., Archer, A., Aune, T., et al. 2018, The
Astrophysical Journal, 861, 134,
doi: 10.3847/1538-4357/aac4a2
Abeysekara, A. U., Albert, A., Alfaro, R., et al. 2020, Phys.
Rev. Lett., 124, 021102,
doi: 10.1103/PhysRevLett.124.021102
Acero, F., Ackermann, M., Ajello, M., et al. 2013, ApJ,
773, 77, doi: 10.1088/0004-637X/773/1/77
Aharonian, F., An, Q., Axikegu, et al. 2021, Phys. Rev.
Lett., 126, 241103, doi: 10.1103/PhysRevLett.126.241103
Albert, A., Alfaro, R., Alvarez, C., et al. 2021, ApJ, 911,
143, doi: 10.3847/1538-4357/abecda
Aliu, E., Aune, T., Behera, B., et al. 2014, ApJ, 788, 78,
doi: 10.1088/0004-637X/788/1/78
Arnaud, K. A. 1996, in Astronomical Society of the Pacific
Conference Series, Vol. 101, Astronomical Data Analysis
Software and Systems V, ed. G. H. Jacoby & J. Barnes,
17
Bachetti, M. 2018, HENDRICS: High ENergy Data
Reduction Interface from the Command Shell,
Astrophysics Source Code Library, record ascl:1805.019.
http://ascl.net/1805.019
Ballet, J., Burnett, T. H., Digel, S. W., & Lott, B. 2020,
Fermi Large Area Telescope Fourth Source Catalog Data
Release 2, arXiv, doi: 10.48550/ARXIV.2005.11208
Barkov, M. V., Lyutikov, M., Klingler, N., & Bordas, P.
2019, MNRAS, 485, 2041, doi: 10.1093/mnras/stz521
Burgess, D. A., Mori, K., Gelfand, J. D., et al. 2022, The
Astrophysical Journal, 930, 148,
doi: 10.3847/1538-4357/ac650a
Cao, Z., Aharonian, F. A., An, Q., et al. 2021a, Nature,
594, 33, doi: 10.1038/s41586-021-03498-z
Cao, Z., Aharonian, F., An, Q., et al. 2021b, ApJL, 917,
L4, doi: 10.3847/2041-8213/ac0fd5
Collaboration, G., Vallenari, A., Brown, A., & et al., T. P.
2022, Astronomy & Astrophysics,
doi: 10.1051/0004-6361/202243940
Cummings, A. C., Stone, E. C., Heikkila, B. C., et al. 2016,
The Astrophysical Journal, 831, 18,
doi: 10.3847/0004-637X/831/1/18
De Luca, A., & Molendi, S. 2004, A&A, 419, 837,
doi: 10.1051/0004-6361:20034421
Hard X-ray observation and multiwavelength study of the Dragonfly nebula 17
Deharveng, L., Lefloch, B., Zavagno, A., et al. 2003, A&A,
408, L25, doi: 10.1051/0004-6361:20031157
Di Mauro, M., Manconi, S., Negro, M., & Donato, F. 2021,
Phys. Rev. D, 104, 103002,
doi: 10.1103/PhysRevD.104.103002
Dwarkadas, V. V. 2005, The Astrophysical Journal, 630,
892, doi: 10.1086/432109
Fang, J., Wen, L., Yu, H., & Chen, S. 2020, MNRAS, 498,
4901, doi: 10.1093/mnras/staa2703
Fang, K., Kerr, M., Blandford, R., Fleischhack, H., &
Charles, E. 2022, Phys. Rev. Lett., 129, 071101,
doi: 10.1103/PhysRevLett.129.071101
Fermi-LAT collaboration, :, Abdollahi, S., et al. 2022, arXiv
e-prints, arXiv:2201.11184.
https://arxiv.org/abs/2201.11184
Freeman, P., Doe, S., & Siemiginowska, A. 2001, in Society
of Photo-Optical Instrumentation Engineers (SPIE)
Conference Series, Vol. 4477, Astronomical Data
Analysis, ed. J.-L. Starck & F. D. Murtagh, 76–87,
doi: 10.1117/12.447161
Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006, in
Society of Photo-Optical Instrumentation Engineers
(SPIE) Conference Series, Vol. 6270, Observatory
Operations: Strategies, Processes, and Systems, ed. D. R.
Silva & R. E. Doxsey, 62701V, doi: 10.1117/12.671760
Gabriel, C., Denby, M., Fyfe, D. J., et al. 2004, in
Astronomical Society of the Pacific Conference Series,
Vol. 314, Astronomical Data Analysis Software and
Systems (ADASS) XIII, ed. F. Ochsenbein, M. G. Allen,
& D. Egret, 759
Gaensler, B. M., & Slane, P. O. 2006, Annual Review of
Astronomy and Astrophysics, 44, 17,
doi: 10.1146/annurev.astro.44.051905.092528
Gelfand, J. D., Slane, P. O., & Zhang, W. 2009, The
Astrophysical Journal, 703, 2051,
doi: 10.1088/0004-637x/703/2/2051
Gotthelf, E. V., Mori, K., Aliu, E., et al. 2016, The
Astrophysical Journal, 826, 25,
doi: 10.3847/0004-637x/826/1/25
Halpern, J. P., Camilo, F., Giuliani, A., et al. 2008, The
Astrophysical Journal, 688, L33, doi: 10.1086/594117
Hartman, R. C., Bertsch, D. L., Bloom, S. D., et al. 1999,
ApJS, 123, 79, doi: 10.1086/313231
H.E.S.S. Collaboration, Abdalla, H., Abramowski, A., et al.
2018, A&A, 612, A2, doi: 10.1051/0004-6361/201629377
Hessels, J. W. T., Roberts, M. S. E., Ransom, S. M., et al.
2004, ApJ, 612, 389, doi: 10.1086/422408
Huppenkothen, D., Bachetti, M., Stevens, A. L., et al. 2019,
ApJ, 881, 39, doi: 10.3847/1538-4357/ab258d
Jansson, R., & Farrar, G. R. 2012, ApJ, 757, 14,
doi: 10.1088/0004-637X/757/1/14
Jin, R., Ng, C.-Y., Roberts, M. S. E., & Li, K.-L. 2023, The
Astrophysical Journal, 942, 100,
doi: 10.3847/1538-4357/aca656
Kargaltsev, O., Pavlov, G. G., Klingler, N., & Rangelov, B.
2017, Journal of Plasma Physics, 83, 635830501,
doi: 10.1017/S0022377817000630
Kargaltsev, O., Rangelov, B., & Pavlov, G. G. 2013,
Gamma-ray and X-ray Properties of Pulsar Wind
Nebulae and Unidentified Galactic TeV Sources.
https://arxiv.org/abs/1305.2552
Kirichenko, A., Danilenko, A., Shternin, P., et al. 2015, The
Astrophysical Journal, 802, 17,
doi: 10.1088/0004-637x/802/1/17
Kolb, C., Blondin, J., Slane, P., & Temim, T. 2017, ApJ,
844, 1, doi: 10.3847/1538-4357/aa75ce
Law, D. R., DeGioia-Eastwood, K., & Moore, K. L. 2002,
ApJ, 565, 1239, doi: 10.1086/337922
Mizuno, T., Tanaka, N., Takahashi, H., et al. 2017, The
Astrophysical Journal, 841, 104,
doi: 10.3847/1538-4357/aa7201
Mori, K., An, H., Burgess, D., et al. 2022, in 37th
International Cosmic Ray Conference, 963,
doi: 10.22323/1.395.0963
Nasa High Energy Astrophysics Science Archive Research
Center (Heasarc). 2014, HEAsoft: Unified Release of
FTOOLS and XANADU, Astrophysics Source Code
Library, record ascl:1408.004. http://ascl.net/1408.004
Ohira, Y., Kisaka, S., & Yamazaki, R. 2018, Monthly
Notices of the Royal Astronomical Society, 478, 926,
doi: 10.1093/mnras/sty1159
Paredes, J. M., Mart´ı, J., Ishwara-Chandra, C. H., et al.
2009, A&A, 507, 241, doi: 10.1051/0004-6361/200912448
Park, J., Kim, C., Woo, J., et al. 2023a, The Astrophysical
Journal, 945, 66, doi: 10.3847/1538-4357/acba0e
—. 2023b, The Astrophysical Journal, 945, 33,
doi: 10.3847/1538-4357/acb1b0
Porth, O., Vorster, M. J., Lyutikov, M., & Engelbrecht,
N. E. 2016, Monthly Notices of the Royal Astronomical
Society, 460, 4135, doi: 10.1093/mnras/stw1152
Reynolds, S. P., Borkowski, K. J., & Gwynne, P. H. 2018,
ApJ, 856, 133, doi: 10.3847/1538-4357/aab3d3
Roberts, M. S. E., Hessels, J. W. T., Ransom, S. M., et al.
2002, ApJL, 577, L19, doi: 10.1086/344082
Roberts, M. S. E., Romani, R. W., & Kawai, N. 2001, The
Astrophysical Journal Supplement Series, 133, 451,
doi: 10.1086/320352
18 Woo et al.
Roberts, M. S. E., Brogan, C., Ransom, S., et al. 2008, in
American Institute of Physics Conference Series, Vol.
1085, American Institute of Physics Conference Series,
ed. F. A. Aharonian, W. Hofmann, & F. Rieger, 328–331,
doi: 10.1063/1.3076673
Rod´on, J. A., Zavagno, A., Baluteau, J. P., et al. 2010,
A&A, 518, L80, doi: 10.1051/0004-6361/201014609
Sudoh, T., Linden, T., & Hooper, D. 2021, Journal of
Cosmology and Astroparticle Physics, 2021, 010,
doi: 10.1088/1475-7516/2021/08/010
Tang, X., & Chevalier, R. A. 2012, The Astrophysical
Journal, 752, 83, doi: 10.1088/0004-637X/752/2/83
Temim, T., Slane, P., Kolb, C., et al. 2015, ApJ, 808, 100,
doi: 10.1088/0004-637X/808/1/100
Temim, T., Slane, P., Plucinsky, P. P., et al. 2017, The
Astrophysical Journal, 851, 128,
doi: 10.3847/1538-4357/aa9d41
Toal´a, J. A., Guerrero, M. A., Chu, Y.-H., et al. 2012, The
Astrophysical Journal, 755, 77,
doi: 10.1088/0004-637X/755/1/77
Torres, D. F., ed. 2017, Modelling Pulsar Wind Nebulae
(Springer International Publishing),
doi: 10.1007/978-3-319-63031- 1
Van Etten, A., Romani, R. W., & Ng, C. Y. 2008, ApJ,
680, 1417, doi: 10.1086/587865
Verbunt, F., Igoshev, A., & Cator, E. 2017, A&A, 608, A57,
doi: 10.1051/0004-6361/201731518
Vorster, M. J., Tibolla, O., Ferreira, S. E. S., & Kaufmann,
S. 2013, The Astrophysical Journal, 773, 139,
doi: 10.1088/0004-637X/773/2/139
Weaver, R., McCray, R., Castor, J., Shapiro, P., & Moore,
R. 1977, ApJ, 218, 377, doi: 10.1086/155692
Wilms, J., Allen, A., & McCray, R. 2000, ApJ, 542, 914,
doi: 10.1086/317016
Wood, M., Caputo, R., Charles, E., et al. 2017, in
International Cosmic Ray Conference, Vol. 301, 35th
International Cosmic Ray Conference (ICRC2017), 824.
https://arxiv.org/abs/1707.09551
Xu, J.-L., Xu, Y., Yu, N., et al. 2017, The Astrophysical
Journal, 849, 140, doi: 10.3847/1538-4357/aa8ee0
Zabalza, V. 2015, Proc. of International Cosmic Ray
Conference 2015, 922
Zabalza, V., Paredes, J. M., Mart´ı, J., S´anchez-Sutil, J. R.,
& Mu˜noz-Arjonilla, A. J. 2010, in Astronomical Society
of the Pacific Conference Series, Vol. 422, High Energy
Phenomena in Massive Stars, ed. J. Mart´ı, P. L.
Luque-Escamilla, & J. A. Combi, 186
