of 82
A blue ring nebula from a stellar merger several thousand
years old
Keri Hoadley
1
,
10
,
12
, D. Christopher Martin
1
,
12
, Brian D. Metzger
2
,
3
,
12
, Mark Seibert
4
,
12
, Andrew
McWilliam
4
, Ken J. Shen
5
, James D. Neill
1
, Gudmundur Stefansson
6
,
7
,
8
,
11
, Andrew Monson
7
,
8
&
Bradley E. Schaefer
9
1
Cahill Center for Astrophysics, California Institute of Technology, 1216 East California Boule-
vard, Mail Code 278-17, Pasadena, California 91125, USA.
2
Department of Physics, Columbia Astrophysics Laboratory, Columbia University, New York, NY
10027, USA.
3
Center for Computational Astrophysics, Flatiron Institute, New York, NY 10010, USA
4
The Observatories of the Carnegie Institution for Science, 813 Santa Barbara Street, Pasadena,
CA 91101, USA.
5
Department of Astronomy & Theoretical Astrophysics Center, University of California at Berke-
ley, 501 Campbell Hall, Berkeley, CA 94720, USA.
6
Department of Astrophysical Sciences, Princeton University, 4 Ivy Lane, Princeton, NJ 08540,
USA
7
Department of Astronomy & Astrophysics, The Pennsylvania State University, 525 Davey Lab,
University Park, PA 16802, USA
8
Center for Exoplanets & Habitable Worlds, University Park, PA 16802, USA
9
Department of Physics & Astronomy, Louisiana State University, Baton Rouge, LA 70820, USA.
10
David & Ellen Lee Caltech Prize Postdoctoral Fellow in Experimental Physics
1
arXiv:2011.09589v1 [astro-ph.SR] 19 Nov 2020
11
Henry Norris Russell Fellow
12
Authors have contributed equally
Stellar mergers are a brief-lived but common phase in the evolution of binary star systems[1,
2]. Among the many astrophysical implications of these events include the creation of atypi-
cal stars (e.g. magnetic stars[3], blue stragglers[4], rapid rotators[5]), interpretation of stellar
populations[6], and formation channels of LIGO-detected compact object mergers[7]. Al-
though stellar mergers are thus commonly invoked phenomena, observations of these events
and details of their evolution remain elusive. While a handful of stellar mergers have been
directly observed in recent years[8, 9], the central remnants of these events remain shrouded
by an opaque shell of dust and molecules[10], making it impossible to observe their final state
(e.g. as a single merged star or a tighter surviving binary[11]). Here we report observations
of an unusual, ring-shaped ultraviolet nebula and the star at its center, TYC 2597-735-1. The
nebula shows two opposing fronts, suggesting a bipolar outflow from TYC 2597-735-1. TYC
2597-735-1’s spectrum and proximity above the Galactic plane suggest it is an old star, yet it
shows abnormally low surface gravity and a detectable long-term luminosity decay, unchar-
acteristic for its evolutionary stage. TYC 2597-735-1 also exhibits H-alpha emission, radial
velocity variations, enhanced ultraviolet radiation, and excess infrared emission, common
signposts of dusty circumstellar disks[12], stellar activity[13], and accretion[14]. The com-
bined observations, paired with stellar evolution models, suggests TYC 2597-735-1 merged
with a lower-mass companion several thousand years ago. TYC 2597-735-1 provides a look
at an unobstructed stellar merger found at an evolutionary stage between its dynamic onset
2
and the theorized final equilibrium state, directly inferring how two stars merge into a single
star.
The “blue” ring nebula (Figure 1(a)) is a rare far-ultraviolet emitting object discovered by
the
Galaxy Evolution Explorer
[15]. The nebula has not been observed in any other part of the
electromagnetic spectrum to date (Extended Data (ED) Figure 1; see Supplementary Information
(SI)). The nebula is ring-shaped and smooth, extending
8
across the sky at a slightly inclined (15
degrees), face-on view. Like other extended, far-ultraviolet sources[16], molecular hydrogen (H
2
),
which fluoresces throughout the far-ultraviolet (
λ <
1700
̊
A; ED Figure 2), is responsible for the
nebular emission. The total luminosity of the nebula,
3
×
10
33
erg s
1
(see SI), yields an upper
limit of 10
44
H
2
molecules fluorescing per second. With a 15% chance of destroying H
2
in the
fluorescence process, the rate of H
2
destruction in the nebula is
.
2.5
×
10
14
solar masses (M
)
per second. Combined with the age of the nebula (see below), we estimate a minimum nebular
mass
M
BRN
&
0
.
004
M
(or
4 Jupiter masses; see SI).
Lining the western edge of the ultraviolet nebula appears a thin, faint shock filament seen
in near-ultraviolet, far-ultraviolet, and H
α
emission (Figure 1(b,c,d)). Ground-based H
α
imaging
shows the filament is part of a more extended shock system (Figure 1(d)), comprised of two over-
lapping, circular rings, which are offset by
3
on the sky (Figure 1(e,f)). We measure the velocity
of the shock using Keck/LRIS multi-slit spectroscopy, finding the two circular rings expanding
in opposite directions directions with v
shock
±
400 kilometers per second. Neutral hydrogen (H
α
,
H
β
, etc.) emits in these filaments, with line ratios suggesting formation in a non-radiative shock
3
created as the outflow sweeps up and heats interstellar gas (see SI).
The large-scale geometry and inference of dual shock fronts expanding in opposite directions
indicate a bipolar outflow originating from a star at the center of the nebula, TYC 2597-735-
1 (Figure 1(f)). TYC 2597-735-1 is located 1.9 kiloparsecs away[17] at 1.5 kiloparsecs above the
Galactic plane, suggesting the nebula extends
4 parsecs. Pairing the physical extent of the nebula
with its velocity, we limit the age of the nebula to
<
5,000 years (see SI).
Spectral synthesis models of TYC 2597-735-1’s Keck/HIRES optical spectra find its stellar
luminosity
L
?
'
110
L
, effective temperature
T
ef f
5850
K, surface gravity
g
600
cm/s
2
, and
stellar radius
R
?
11
R
(see SI). These spectroscopic parameters suggest TYC 2597-735-1 has a
mass around 1 – 2.1 M
and has evolved off the main sequence. Notably, TYC 2597-735-1 appears
puffier for its temperature than other evolved stars of similar luminosity (ED Figure 3).
The surface iron abundance of TYC 2597-735-1 is sub-solar (
[Fe
/
H]
∼−
0
.
9
dex), while its
α
-element abundances appear enhanced (
[
α/
Fe]
+0
.
4
dex; see SI). Its parallax, proper motions,
and radial velocity indicate disk-like kinematics (UVW
7.8, 11.3, 36.0 km/s), based on solar
peculiar motions. These observations support TYC 2597-735-1’s membership in the thick-disk
population[18].
TYC 2597-735-1 also displays prominent H
α
line emission, excess far-ultraviolet flux, and
radial velocity variations. The H
α
emission varies in both line shape and amplitude on timescales
of days, showing an enhanced blue-shifted component (ED Figure 4). The observed far-ultraviolet
4
magnitude of TYC 2597-735-1 is over 6 orders of magnitude brighter than is expected from
synthetic stellar models (Figure 2(a)). Radial velocity measurements of TYC 2597-735-1 find
200 meters per second Doppler shift, yet exclude the presence of a binary companion with mass
&
0
.
01
M
in tight orbit around TYC 2597-735-1 (
a
.
0.1 astronomical units; ED Figure 5).
Altogether, these signatures point to heightened stellar surface activity at TYC 2597-735-1.
Additionally, TYC 2597-735-1 emits excess infrared radiation (Figure 2(a); see SI), a tell-tale
sign of a dusty circumstellar disk[12]. A disk-like geometry, which lies in a plane perpendicular
to our line of sight and the symmetric axis of the nebula, is favored because it lacks evidence
of circumstellar reddening, as measured from TYC 2597-735-1’s optical spectrum (see SI). The
combination of infrared emission and surface activity paint a picture where TYC 2597-735-1 is
actively accreting material from a disk of gas and dust extending to several astronomical units (see
SI). Other systems with similar observable traits (e.g., T Tauri stars[14, 19] and AGB stars with
accretion disks[20]), are often interpreted as actively accreting material from circumstellar disks.
Any viable explanation for TYC 2597-735-1 and its ultraviolet nebula should account for the
unusual properties of the star, its circumstellar environment, and the fast outflow launched from
its vicinity
<
5,000 years ago. Protostellar systems exhibit many of the same characteristics as
TYC 2597-735-1, including H
α
emission and excess ultraviolet flux[21]; however, its isolation
from star-forming environments and proximity above the Galactic plane strongly disfavor it being
a young protostar. At the other end of the stellar evolution spectrum, TYC 2597-735-1’s stellar
properties do not match those of “post-red giant” systems, which are more luminous (
>
1,000 L
).
5
While evolved stars are known to host dusty disks and expel massive stellar winds[22], the velocity
of the ultraviolet nebula (v
shock
400 km/s) is
&
10
times faster than those measured from (post-
)asymptotic giant branch stars and pre-planetary nebulae (v
wind
10’s km/s)[23]. Radial velocity
measurements exclude a short-period companion orbiting TYC 2597-735-1 capable of ejecting a
collimated, bipolar outflow with its observed velocity (see SI), thus disfavoring classical novae,
cataclysmic variability, or other mass-transfer interactions with a surviving compact object.
The mass ejection event responsible for this ultraviolet nebula and the unusual present state
of TYC 2597-735-1 appear most consistent with a binary star merger. To test this scenario and
estimate the initial state of the system, we use the stellar evolution code MESA[24] to explore
the impact a stellar merger has on long-term stellar properties[25] (see SI). We find that a low-
mass companion (
M
c
0
.
1
M
) reasonably reproduces TYC 2597-735-1’s effective temperature,
luminosity, and surface gravity at a post-merger age of
t
age
1
,
000
years, accounting for TYC
2597-735-1’s position in
T
ef f
-log
g
space (ED Figure 3). These models also predict that TYC
2597-735-1 was
0
.
1
B-mag brighter a century ago, which we observe in historical DASCH
archive records of TYC 2597-735-1 (see SI, ED Figure 6).
The best-fit models are those where the merger happens after the primary begins evolving
off its main sequence (ED Figure 7). Such timing may not be coincidental if the companion was
dragged into the star through tidal interaction (Figure 3(a)): the timescale for tidal orbital decay
depends sensitively on the primary radius (
τ
tide
R
5
?
)[26], accelerating the orbital decay as the
primary reaches its sub-giant phase. Indeed, tidally induced mergers may explain the dearth of
6
short-period planets around evolved A-stars[27].
The events leading up to and following the merger of TYC 2597-735-1 and its companion
shape the system we see today. As TYC 2597-735-1 and its companion approached sufficiently
closely, the former overflowed its Roche lobe onto the latter, initiating the merger (Figure 3(b)).
Numerical simulations demonstrate that the earliest phase of the merger process results in the ejec-
tion of matter through the outer
L
2
Lagrange point in the equatorial plane of the binary[28]. Most
of this matter remains gravitationally bound around the star system, forming a circumbinary disk.
The companion, unable to accommodate the additional mass, is dragged deeper into the primary’s
envelope in a runaway process[11] (Figure 3(c)). This delayed dynamical phase is accompanied by
the ejection of a shell of gas. A portion of the ejected material is collimated by the circumbinary
disk into a bipolar outflow[29]. The balance of the mass lost during primary-companion interac-
tions remains as a circumstellar disk, which spreads out and cools over time, eventually reaching
sufficiently low temperatures to form dust.
We see evidence of this relic disk around TYC 2597-735-1 today as an infrared excess. A
simple analytic model, which follows the spreading evolution of the gaseous disk due to internal
viscosity over the thousands of years since the merger (see SI), is broadly consistent with both the
present-day gas accretion rate (estimated from H
α
; see SI) and a lower limit on the present-day disk
mass obtained by fitting the infrared spectral energy distribution of TYC 2597-735-1 (
M
disk
,
dust
&
5
×
10
9
M
; see SI). Accretion of disk material onto TYC 2597-735-1 could account for its
observed stellar activity (e.g., H
α
emission and far-ultraviolet excess). Angular momentum added
7
to the envelope of the star by the merger, and subsequent accretion of the disk, would also increase
TYC 2597-735-1’s surface rotation velocity. Indeed, we find the de-projected surface rotation
velocity of TYC 2597-735-1 to be
25
km/s (see SI, ED Figure 8), larger than expected for a star
which has just evolved off the main sequence (
v <
10 km/s)[30].
The bipolar ejecta shell expands away from the stellar merger, cooling and satisfying within
weeks the conditions for molecular formation and solid condensation (Figure 3(d)), as seen di-
rectly in the observed ejecta of luminous red novae[31]. The mass ejected, inferred from merger
simulations and modeling the light curves of luminous red novae, is typically
0.01-0.1
M
[32],
consistent with our lower mass limit of the ultraviolet nebula (0.004
M
).
As the nebula expands and sweeps up interstellar gas, a reverse shock crosses through the
ejecta shell, heating electrons in its wake. These electrons excite the H
2
formed in the outflow,
which fluoresces in the far-ultraviolet (Figure 3(e)). Although dust is almost always observed in
the ejecta of stellar mergers[31], we find no evidence of dust in the ultraviolet nebula (e.g., ultra-
violet/optical reddening). We speculate that either the dust was destroyed in the reverse shock[33]
or that the bipolar ejecta shell has thinned out sufficiently such that dust is currently undetectable.
Consistent with the latter, this ultraviolet nebula marks the oldest observed stellar merger to date,
being
&
3
10
times older than the previously oldest stellar merger candidate, CK Vulpeculae
(1670), which is still shrouded by dust[34]. The system was caught at an opportune time - old
enough to reveal the central remnant, yet young enough that the merger-generated nebula has not
dissolved into the interstellar medium.
8
The discovery of an ultraviolet nebula introduces a new way of identifying otherwise hidden
late-stage stellar mergers. With 1 – 10 of these objects expected to be observable in the Milky Way
(see SI), future far-ultraviolet telescopes may uncover more late-stage stellar mergers. TYC 2597-
735-1 poses a unique opportunity to study post-merger morphology as the only known merger sys-
tem not enshrouded by dust. For example, the close separation of the initial stellar binary
which
shares properties broadly similar to those of protoplanetary disks
could form “second generation”
planets[35]. However, given the relatively short time (
.
10
8
years) until TYC 2597-735-1 reaches
the end of its nuclear burning life, the potential window of habitability for such planets could be
drastically reduced compared to main sequence stars.
Supplementary Information
is linked to the online version of the paper at www.nature.com/nature.
Acknowledgements
This research is based on observations made with the
Galaxy Evolution Explorer
,
obtained from the MAST data archive at the Space Telescope Science Institute, which is operated by the As-
sociation of Universities for Research in Astronomy, Inc., under NASA contract NAS 5–26555. Some of the
data presented herein were obtained at the W. M. Keck Observatory, which is operated as a scientific partner-
ship among the California Institute of Technology, the University of California and the National Aeronautics
and Space Administration. This research has made use of the Keck Observatory Archive (KOA), which is
operated by the W. M. Keck Observatory and the NASA Exoplanet Science Institute (NExScI), under con-
tract with the National Aeronautics and Space Administration. The Observatory was made possible by the
generous financial support of the W. M. Keck Foundation. The authors wish to recognize and acknowledge
the very significant cultural role and reverence that the summit of Maunakea has always had within the in-
digenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from
9
this mountain. Some of the data presented herein were obtained at the Palomar Observatory. This research
has made use of the NASA/IPAC Infrared Science Archive, which is operated by the Jet Propulsion Labora-
tory, California Institute of Technology, under contract with the National Aeronautics and Space Adminis-
tration. We especially thank Vicky Scowcroft for obtaining
Spitzer
/IRAC photometry of TYC 2597-735-1.
Funding for APASS has been provided by the Robert Martin Ayers Sciences Fund. The DASCH data from
the Harvard archival plates was partially supported from National Science Foundation grants AST-0407380,
AST-0909073, and AST-1313370. The American Association of Variable Star Observers has been critically
helpful for finder charts, comparison star magnitudes, and recruiting skilled observers, including Sjoerd
Dufoer, Kenneth Menzies, Richard Sabo, Geoffrey Stone, Ray Tomlin, and Gary Walker. These results are
based on observations obtained with the Habitable-zone Planet Finder Spectrograph on the Hobby-Eberly
Telescope. These data were obtained during HPF’s Engineering and Commissioning period. We thank the
Resident astronomers and Telescope Operators at the HET for the skillful execution of our observations
of our observations with HPF. The authors would like to thank Caleb Ca
̃n
as for providing an indepen-
dent verification of the HPF SERVAL pipeline using a CCF-based method to calculate the RVs, which
resulted in fully-consistent RVs to the SERVAL-based RVs presented here. The Hobby-Eberly Telescope is
a joint project of the University of Texas at Austin, the Pennsylvania State University, Ludwig-Maximilians-
Universit
̈
at M
̈
unchen, and Georg-August Universit
̈
at Gottingen. The HET is named in honor of its principal
benefactors, William P. Hobby and Robert E. Eberly. The HET collaboration acknowledges the support
and resources from the Texas Advanced Computing Center. This work was partially supported by funding
from the Center for Exoplanets and Habitable Worlds. The Center for Exoplanets and Habitable Worlds
is supported by the Pennsylvania State University, the Eberly College of Science, and the Pennsylvania
Space Grant Consortium. This research made use of
photutils
and
astropy
, community-developed
core
Python
packages for Astronomy, and Modules for Experiments in Stellar Astrophysics (MESA). The
10
authors thank Prof. Armando Gil de Paz for obtaining the narrow-band filter H
α
imagery; Prof. John John-
son for commissioning TYC 2597-735-1 RV measurements as part of the California Planet Finder (CPF)
program; and Prof. Andrew Howard for spearheading Keck/HIRES spectra and performing the primary RV
reduction on all HIRES data. KH thanks Lynne Hillenbrand and Erika Hamden for productive discussions
regarding critical aspects of this work. BDM acknowledges support from the Hubble Space Telescope (
#
HST-AR-15041.001-A) and from the National Science Foundation (
#
80NSSC18K1708). KJS received
support from the NASA Astrophysics Theory Program (NNX17AG28G). GS and A. Monson acknowledge
support from NSF grants AST-1006676, AST-1126413, AST-1310885, AST-1517592, AST-1310875, AST-
1907622, the NASA Astrobiology Institute (NAI; NNA09DA76A), and PSARC in our pursuit of precision
radial velocities in the NIR with HPF. We acknowledge support from the Heising-Simons Foundation via
grant 2017-0494 and 2019-1177. Computations for this research were performed on the Pennsylvania State
University’s Institute for Computational & Data Sciences (ICDS). GS acknowledges support by NASA HQ
under the NASA Earth and Space Science Fellowship (NESSF) Program through grant NNX16AO28H.
Author Contributions
Authors K. Hoadley and B. D. Metzger organized and wrote the main body of the
paper. Authors K. Hoadley and M. Seibert performed the data reduction and analysis of the
GALEX
data,
investigated the source of the ultraviolet emission, quantified the mass the far-ultraviolet nebula, and led the
the analysis of TYC 2597-735-1’s H
α
emission and variability. Author B. D. Metzger lead all theoretical
and analytic interpretation efforts of the ultraviolet nebula origins and TYC 2597-735-1 in the context of
stellar mergers and present-day luminous red novae. Authors D. C. Martin and M. Seibert spearheaded the
GALEX
program that led to the detection of the ultraviolet nebula in 2004 and all subsequent follow-up
observations of the nebula with
GALEX
. Both contributed to the overall interpretation of the observational
data. Author D. C. Martin contributed to the organization and manifestation of the manuscript. Author
11
M. Seibert lead the radial velocity analysis, the interpretation and analysis of the infrared excess in TYC
2597-735-1’s spectral energy distribution, and modeled the TYC 2597-735-1 spectral energy distribution
(both stellar and dust infrared excess components). M. Seibert coordinated all ground-based observations of
the blue ring nebula and TYC 2597-735-1 at Palomar Observatory and W. M. Keck Observatory.
Author A. McWilliam derived the physical parameters and performed the model atmosphere chemical abun-
dance analysis of TYC 2597-735-1. A. McWilliam participated in the discussion of observations, analysis,
and interpretation that led to the manifestation of this work. Author K. J. Shen performed the MESA
calculations and participated in the discussion of observations, analysis, and interpretation that led to the
manifestation of this work.
Author J. D. Neill handled the data analysis and reported the subsequent result of the velocity structure of
the H
α
shock observed with Keck/LRIS. J. D. Neill participated in the discussion of observations, analysis,
and interpretation that led to the manifestation of this work.
Author G. Stefansson performed the HET/HPF radial velocity and differential line width indicator extrac-
tions and provided expertise on the interpretation of the combined radial velocity data sets.
Author A. Monson coordinated HET/HPF observations and both performed and reduced all TMMT B-band
observations.
Author B. E. Schaefer extracted and analyzed the long-term light curve data from 1897.5–2019.9.
Competing Interests
The authors declare that they have no competing financial interests.
Correspondence
Correspondence and requests for materials should be addressed to Dr. Keri Hoadley
(email: khoadley@caltech.edu).
Data availability
All
GALEX
imaging and grism data of TYC 2597-735-1 and its ultraviolet nebula are
12
publicly available for download from the Mikulski Archive for Space Telescopes (MAST) in raw and re-
duced formats. The archive can be accessed either at
http://galex.stsci.edu/GalexView/
or
https://mast.stsci.edu/portal/Mashup/Clients/Mast/Portal.html
. All Keck-
LRIS and Keck-HIRES data for TYC 2597-735-1 are publicly available for download from the Keck Ob-
servatory Archive (KOA). KOA can be accessed at
https://koa.ipac.caltech.edu/cgi-bin/
KOA/nph-KOAlogin
. TYC 2597-735-1 raw photometric lightcurve frames, plates, and lightcurves from
1895 – 1985 are publicly available and downloaded as a part of the DASCH: Digital Access to a Sky
Century at Harvard program. The DASCH data base can be accessed at
https://projects.iq.
harvard.edu/dasch
. More recent photmetry for the lightcurve construction were obtained by inde-
pendent observers and the American Association of Variable Star Observers (AAVSO). These data have
been uploaded to a publicly available repository; please contact Corresponding Author for a link to the
repository. All other photometric data for TYC 2597-735-1 was obtained from publicly archived ground-
and space-based imaging and surveys, stored on the SIMBAD Astronomical Database (
http://simbad.
u-strasbg.fr/simbad/
) and the NASA/IPAC Infrared Science Archive (
https://irsa.ipac.
caltech.edu/frontpage/
). The Hobby-Eberly Telescope does not have a publicly-available archive
to access Habitable Planet Finder (HPF) data. We provide spreadsheets with the relevant data products from
the HPF campaign for TYC 2597-735-1 in a publicly available repository:
https://github.com/
oglebee-chessqueen/BlueRingNebula.git
Code availability
We used and Modules for Experiments in Stellar Astrophysics (MESA)[24] for a por-
tion of our analysis. While MESA is readily available for public use, we use a custom sub-routine and MESA
inline code to produce the TYC 2597-735-1 merger evolution model presented in this work. We provide both
in a public repository:
https://github.com/oglebee-chessqueen/BlueRingNebula.git
.
13
Use the ATLAS9 pre-set grid of synthetic stellar spectra[36] to fit the TYC 2597-735-1 spectral energy distri-
bution to representative stellar spectra. All synthetic stellar spectra are publicly available:
https://www.
stsci.edu/hst/instrumentation/reference-data-for-calibration-and-tools/astronomical-catalogs/
castelli-and-kurucz-atlas
. Portions of our analysis used Python packages
photutils
[37] and
astropy
[38].
14