A Candidate Electromagnetic Counterpart to the Binary Black Hole Merger
Gravitational Wave Event S190521g
a
M.J. Graham,
1
K.E.S. Ford,
2, 3, 4
B. McKernan,
2, 3, 4
N.P. Ross,
5
D. Stern,
6
K. Burdge,
1
M. Coughlin,
7, 8
S.G. Djorgovski,
1
A.J. Drake,
1
D. Duev,
1
M. Kasliwal,
1
A.A. Mahabal,
1
S. van Velzen,
9, 10
J. Belecki,
11
E.C. Bellm,
12
R. Burruss,
11
S.B. Cenko,
13, 14
V. Cunningham,
9
G. Helou,
15
S.R. Kulkarni,
1
F.J. Masci,
16
T. Prince,
1
D. Reiley,
11
H. Rodriguez,
11
B. Rusholme,
16
R.M. Smith,
11
and M.T. Soumagnac
17, 18
1
Cahill Center for Astronomy
&
Astrophysics, California Institute of Technology,
1200 E. California Blvd., Pasadena, CA 91125, USA
†
2
Department of Science, CUNY-BMCC, 199 Chambers St., New York, NY 10007, USA
3
Department of Astrophysics, American Museum of Natural History,
Central Park West, New York, NY 10028, USA
4
Physics Program, The Graduate Center, CUNY, New York, NY 10016, USA
5
Institute for Astronomy, University of Edinburgh,
Royal Observatory, Blackford Hill, Edinburgh EH9 3 HJ, UK
6
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109, USA
7
Division of Physics, Mathematics, and Astronomy,
California Institute of Technology, Pasadena, CA 91125, USA
8
School of Physics and Astronomy, University of Minnesota, Minneapolis, Minnesota 55455, USA
9
Department of Astronomy, University of Maryland, College Park, MD 20742, USA
10
Center for Cosmology and Particle Physics, New York University, NY 10003, USA
11
Caltech Optical Observatories, California Institute of Technology, Pasadena, CA 91125, USA
12
DIRAC Institute, Department of Astronomy, University of Washington, 3910 15th Ave. NE, Seattle, WA 98195, USA
13
Astrophysics Science Division, NASA Goddard Space Flight Center, MC 661, Greenbelt, MD 20771, USA
14
Joint Space-Science Institute, University of Maryland, College Park, MD 20742, USA
15
IPAC, California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125, USA
16
IPAC, California Institute of Technology, 1200 E. California Blvd., Pasadena, CA 91125, USA
17
Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA
18
Department of Particle Physics and Astrophysics,
Weizmann Institute of Science, Rehovot 76100, Israel
(Dated: June 29, 2020)
We report the first plausible optical electromagnetic (EM) counterpart to a (candidate) binary
black hole (BBH) merger. Detected by the Zwicky Transient Facility (ZTF), the EM flare is consis-
tent with expectations for a kicked BBH merger in the accretion disk of an active galactic nucleus
(AGN) [1], and is unlikely (
< O
(0
.
01%)) due to intrinsic variability of this source. The lack of
color evolution implies that it is not a supernovae and instead is strongly suggestive of a constant
temperature shock. Other false-positive events, such as microlensing or a tidal disruption event,
are ruled out or constrained to be
< O
(0
.
1%). If the flare is associated with S190521g, we find
plausible values of: total mass
M
BBH
∼
100
M
, kick velocity
v
k
∼
200 km s
−
1
at
θ
∼
60
◦
in a disk
with aspect ratio
H/a
∼
0
.
01 (i.e., disk height
H
at radius
a
) and gas density
ρ
∼
10
−
10
g cm
−
3
.
The merger could have occurred at a disk migration trap (
a
∼
700
r
g
;
r
g
≡
GM
SMBH
/c
2
, where
M
SMBH
is the mass of the AGN supermassive black hole). The combination of parameters implies
a significant spin for at least one of the black holes in S190521g. The timing of our spectroscopy
prevents useful constraints on broad-line asymmetry due to an off-center flare. We predict a repeat
flare in this source due to a re-encountering with the disk in
∼
1
.
6 yr (
M
SMBH
/
10
8
M
) (
a/
10
3
r
g
)
3
/
2
.
a
At the time of writing, LIGO has not yet officially confirmed this event. We still refer to it in this paper using the S-* naming syntax
to acknowledge this.
arXiv:2006.14122v1 [astro-ph.HE] 25 Jun 2020
2
Introduction.—
The Laser Interferometer Gravitational wave (GW) Observatory (LIGO) is now detecting binary
black hole (BBH) mergers at a high rate in the local (
z <
1) Universe [2]. The two main channels to BBH mergers
are believed to be field binary star evolution [e.g., 3, 4] and dynamical encounters. Dynamical mergers can occur in
globular clusters [5, 6], galactic nuclei [7–9], and in gas disks in galactic nuclei [10–18]. Mergers involving
>
50
M
black holes (BHs) are unlikely to involve field binary stars [19]. Rather, massive mergers suggest a dynamical origin,
likely in a deep potential where kicked merger products can be retained [20]. Several massive mergers may have already
been detected, including GW170929 [21] and GW170817A [22] (not to be confused with the binary neutron star merger
GW170817). A dynamical origin for these mergers implies a much larger number of lower mass mergers from the
same channel. Electromagnetic (EM) counterparts are hard to generate in the absence of gas. EM counterparts to
supermassive BBH mergers in gas disks are well studied [e.g. 23–25], but stellar-origin BBH mergers in active galactic
nucleus (AGN) disks can also yield a significant, detectable EM counterpart [1].
The Zwicky Transient Facility (ZTF) is a state-of-the-art time-domain survey employing a 47 deg
2
field-of-view
camera on the Palomar 48-inch Samuel Oschin Schmidt telescope [26, 27]. A public survey covers the visible northern
sky every three nights in
g
- and
r
-bands to
∼
20
.
5 mag [28]. Other observing programs cover smaller areas to greater
depth, with higher cadence, or with an additional
i
-band filter. Alerts are generated in real time for all
≥
5
σ
transient
detections from difference imaging, and those from the public survey are issued to the community [29].
Searching for counterparts.—
For the 21 LIGO BBH merger triggers in observing run O3a (2019 April 1 - September
30), we identified possible AGN which lay within the 90% confidence limit region and within the 3
σ
limits of the
marginal distance distribution integrated over the sky. AGN were identified from the Million Quasar Catalog v6.4
[30]. Any flare associated with the BBH merger should present within a few days to weeks [1] and so we determined
the subset of AGN which were associated with a ZTF alert
≤
60 days post-LIGO trigger. Here we present our
most promising EM counterpart to a BBH GW event based on a Bayesian changepoint analysis (Graham et al., in
preparation).
The event S190521g was observed by both LIGO detectors and the VIRGO detector at 2019-05-21 03:02:29 UTC
with a false alarm rate of 3
.
8
×
10
−
9
Hz (FAR = 1
/
8
.
3 yr) [31]. It has a luminosity distance of 3931
±
953 Mpc and
was classified as a BBH merger with 97% certainty. ZTF observed 48% of the 765 deg
2
90% localization region of
S190521g (half of the localization region is in the southern sky). Alert ZTF19abanrhr (see Fig. 1), first announced
∼
34 days after the GW event and associated with AGN J124942.3+344929 at
z
= 0
.
438 (hereafter J1249+3449),
was identified as potentially interesting. The AGN is located at the 78% spatial contour and 1
.
6 (0
.
7)
σ
from the peak
marginal (conditional) luminosity distance. If we convolve the marginal distance distribution for the LIGO event [32]
with the quasar luminosity function [33] and assume a survey depth of 20.5 mag and a flare probability of 10
−
4
per
quasar (see below), we would expect to find 10
−
5
events in the area and timeframe considered.
From a fit to the H
β
line profile of the AGN, using the QSFit routine [34], we find the mass of the central super-
massive black hole (SMBH) spans
M
SMBH
= [1
,
10]
×
10
8
M
and therefore the pre-flare luminosity is
L
bol
/L
Edd
=
[0
.
02
−
0
.
23] relative to the Eddington luminosity. From the ZTF lightcurve, J1249+3449 varied by only a few percent
of its mean flux level (
∼
19
.
1 mag in
g
-band) over the 15 months prior to S190521g. A flare peaking
∼
50 days after
the GW trigger elevated the flux by
∼
0
.
3 mag (equivalent to
∼
10
45
erg s
−
1
) for
∼
50 days, assuming a typical quasar
bolometric correction factor [35]. The total energy released by the flare is therefore
O
(10
51
erg).
False positives.—
We consider and rule out, or at least constrain, several possible causes of the ZTF19abanrhr
flaring event, such as AGN variability, a supernova, microlensing, and the tidal disruption of a star by an SMBH.
AGN are intrinsically variable, often on quite short timescales [37, 38]. However, from Fig. 2, this AGN has had
a relatively constant luminosity for a year around the flare. We applied models consisting of a generic flare profile
(Gaussian rise, exponential decay) superimposed on a linear luminosity model to ZTF lightcurves of all detected
sources in the larger
WISE
-selected R90 catalogue of 4.5 million high-probability quasar candidates, of which 2.5
million are within the area of sky covered by ZTF and 603,000 are spectroscopically confirmed quasars [39]. We
exclude 2912 known blazars and select objects where the flare model is strongly preferred over the linear model (i.e.,
change in the Bayesian information criterion ∆
BIC >
10), the flare is detected in both
g
- and
r
-bands, has at least a
25% increase in flux, and lasts
≥
20 days in the observed frame. This gives 393 events, of which 209 produced a ZTF
alert (the remaining 182 were
<
5
σ
detections above background, and therefore did not produce alerts).
AGN variability is commonly described statistically as a damped random walk (DRW) process [40, 41]. If the
flare is consistent with this then the same parameterized DRW model (within the confidence limits on the model
parameters) should describe the time series with and without the flare [42]. Applying this constraint to both
g
- and
r
-band data reduces the number of flares similar to ZTF19abanrhr (i.e., not attributable to regular AGN activity
with greater than 3
σ
confidence) to 13. Graham et al. (in preparation) provides more details on the search and the
full identified sample. In summary, this analysis shows that the probability of a flare + linear model randomly fitting
any given ZTF AGN lightcurve is
∼
5
×
10
−
6
.
3
0
h
21
h
18
h
15
h
12
h
9
h
6
h
3
h
0
h
0°
30°
60°
60°
30°
0°
-30°
-60°
-60°
-30°
S190521g
5°
E
N
50%
50%
90%
90%
90%
0
1000
2000
3000
4000
5000
6000
7000
Distance (Mpc)
0.0000
0.0001
0.0002
0.0003
0.0004
0.0005
Probability (Mpc
1
)
FIG. 1. Left panel: A Mollweide projection of the 50% and 90% LIGO localization regions for S190521g (with 44%/56% in
the northern/southern hemisphere) and the location of ZTF19abanrhr (within the 78% contour). ZTF covered 48% of the
90% region and contours at declination
<
−
30
◦
indicate southern hemisphere regions not covered by ZTF. Right panel: The
marginal luminosity distance distribution integrated over the sky (dotted blue line) for S190521g as well as the conditional
distance distribution (black line) at the position of ZTF19abanrhr. The red line corresponds to the luminosity distance of
ZTF19abanrhr, assuming a Planck15 cosmology [36].
Fig. 3 shows that a decade-long baseline reveals evidence for more significant variability in J1249+3449. Note that
these data, from the Catalina Real-time Transient Survey [CRTS; 43], are noisier than ZTF (a result of a 0.7-m survey
telescope vs. a 1.2-m survey telescope), and are binned at 15 day intervals for clarity in the plot. Using the DRW
model parameters from the CRTS data, which characterize the overall variability of the source, we simulated the
observed ZTF light curve 250,000 times and find an equivalent flare (i.e., matching the selection criteria described
above) in four instances. The event is thus very unlikely to arise from AGN activity in this particular source (i.e.,
∼
O
(0
.
002%). Similarly, to address the look-elsewhere effect, we produced 1000 simulations of the full sample of 3255
AGN in the 90% three-dimensional localization region of S190521g using their CRTS DRW parameterizations and
ZTF time sampling. We find a comparable AGN flare in just five simulations, i.e.,
O
(0
.
5%) chance of a false positive,
prior to visual inspection.
Supernovae can occur in AGN [e.g., 44], although the rate is likely small (
>
2
×
10
−
7
AGN
−
1
yr
−
1
in the
WISE
sample). Even with a
O
(10
51
erg) energy output, we expect rise times of
O
(20
−
50) days and a decay time or plateau
of
∼
100
−
200 days [45]. The flare in Fig. 2 lasts 40 days observed-frame, or only 28 days rest-frame which is a poor
match to supernova lightcurves. In addition, supernovae evolve in color over time [46] whereas this flare is uniform
with color over time, suggestive of a shock or accretion, rather than a supernova. We therefore rule out a supernova
as a likely false positive.
Microlensing, with an expected rate of
O
(10
−
4
) per AGN [47], is uniform in color at restframe UV/optical bands
and is also expected for AGN. However, the expected characteristic timescale for microlensing is
O
(yrs) [47], which
is inconsistent with the several week ZTF19abanrhr flare. Assuming a
M
lens in the source galaxy, we require the
lens to orbit at
∼
1kpc at 200km s
−
1
in order to match the timescale (
∼
2
×
10
6
s) and magnification (
∼
1
.
4) of
this event; assuming a population of
O
(10
10
) stars in appropriate orbits, geometric considerations produce a rate of
O
(10
−
5
) events yr
−
1
AGN
−
1
.
Tidal disruption events (TDEs) also occur in AGN. Stellar disruptions can occur around the central SMBH in a
galaxy, but only for
M
SMBH
∼
<
10
8
M
[for a non-spinning SMBH; 48]. TDEs can also occur around small BHs in
AGN disks, but as neutron star (NS) or white dwarf (WD) disruptions. EM counterparts to BH-NS tidal disruptions
in AGN disks at
z <
0
.
5 should span
∼
[4
,
113] (
f
AGN
/
0
.
1) yr
−
1
where
f
AGN
is the fraction of BBH mergers expected
4
18.8
19.0
19.2
g
18.6
18.8
19.0
r
58200
58300
58400
58500
58600
58700
58800
58900
MJD
0.0
0.2
0.4
g - r
FIG. 2. ZTF
g
-band photometry,
r
-band photometry, and
g
−
r
color for J1249+3449 over the past 25 months. The flare
beginning MJD
∼
58650 represents a 5
σ
departure from the ZTF baseline for this source. The flare emission is fit according
to the model described in the text and assuming a linear model for the source continuum behaviour over time. The dashed
vertical line corresponds to the S190521g trigger time.
18.8
19.0
19.2
19.4
g
54000
55000
56000
57000
58000
59000
MJD
18.6
18.8
19.0
19.2
r
FIG. 3. Lightcurve for J1249+3449, including an additional decade of CRTS photometry (binned at 15 day intervals). ZTF
data is binned in 3 day intervals, with
g
- and
r
-band data corrected to the CRTS photometric system using median offsets of
0.52 mag for
g
-band and 0.34 mag for
r
-band.
from the AGN channel [49]. The expected integrated total energy of such events is
O
(10
52
erg) [50], an order of
magnitude more powerful than ZTF19abanrhr. Such an event would also produce a GW signal unlike what was
observed based on the inferred chirp mass
M
c
discussed below for S190521g, and the absence of any other reported
LIGO triggers with an appropriate spatial and temporal coincidence). BH-WD disruptions lead to underluminous
Type Ia SN with integrated energy 10
49
−
51
erg, generally less luminous than ZTF19abanrhr, and decay over a year,
and so are ruled out [51].
Testing the candidate counterpart.—
We can derive an approximate mass for any reported GW event from the
distance (
d
L
) and sky area (
A
90
, the 90% confidence interval for sky area) reported in the public GW event alerts.
Specifically,
A
90
∝
SNR
−
2
[e.g. 52] and SNR
∝
M
5
/
6
c
d
−
1
L
[53]. Deriving the proportionality constant for a 3-detector
system for
A
90
∝
SNR
−
2
from GW190412 [54], we estimate SNR
∼
8
.
6 for S190521g. Assuming equal mass components
for this rough calculation, that ZTF19abanrhr is related to S190521g, and using a binary NS range of 110 Mpc (LIGO
Hanford) to determine detector sensitivity during the S190521g detection, we estimate a source-frame total mass for
M
BBH
∼
150
M
(roughly accurate to a factor of 2, O(100
M
), and plausibly in the upper mass gap).
5
Gravitational radiation from merging unequal mass BBH carries linear momentum, so the BBH center of mass
recoils [55, 56]. For a BBH merger product kicked with velocity
v
k
in an AGN disk, gravitationally bound gas
(
R
bound
< GM
BBH
/v
2
k
) attempts to follow the BH of mass
M
BBH
, but collides with the surrounding disk gas, producing
a bright off-center hotspot at UV/optical wavelengths [1]. The radius of gravitationally bound gas is
R
bound
R
H
= 0
.
34
(
q
10
−
6
)
2
/
3
(
a
10
3
r
g
)
−
1
(
v
k
200 km s
−
1
)
−
2
(1)
where
R
H
=
a
(
q/
3)
1
/
3
is the Hill radius of the BH,
a
is the BH orbit semi-major axis in units of
r
g
≡
GM
SMBH
/c
2
,
and
q
=
M
BBH
/M
SMBH
is the mass ratio of the BBH to the central SMBH. The total energy delivered to the bound
gas is
E
b
= 1
/
2
M
b
v
2
k
= 3
/
2
N k
B
T
b
where
M
b
=
N m
H
is the mass of the bound gas expressed as
N
atoms of
Hydrogen (mass
m
H
),
k
B
is the Boltzmann constant, and
T
b
is the average temperature of the post-shock gas.
E
b
is
E
b
= 3
×
10
45
erg
(
ρ
10
−
10
g cm
−
3
)(
M
BBH
100
M
)
3
(
v
k
200 kms
−
1
)
−
4
.
(2)
The dynamical time in the source-frame associated with the ram pressure shock (or the time for the merger remnant
to cross the sphere of bound gas) is
t
ram
=
R
bound
/v
k
=
GM
BBH
/v
3
k
or
t
ram
∼
20 day
(
M
BBH
100
M
)
(
v
k
200 km s
−
1
)
−
3
(3)
or
∼
29 days observed-frame for the same parameterization given the redshift of J1249+3449. The luminosity increase
for this process should scale roughly as sin
2
(
π
2
t
t
ram
)
until
t > t
ram
, when the kicked BH leaves behind the gas that
was gravitationally bound at
t
= 0.
E
b
is inadequate to explain ZTF19abanrhr, though it induces a delay time (
t
ram
)
before the dominant luminosity-producing process can begin.
The BH leaves behind bound gas after
t
ram
and enters unperturbed disk gas at
t > t
ram
. Nearby gas is accelerated
around the BH, producing a shocked Bondi tail [e.g., 57–59] which both acts as a drag on the BH and accretes onto it.
We approximate the Bondi-Hoyle-Lyttleton (BHL) luminosity as
L
BHL
=
η
̇
M
BHL
c
2
where
η
is the radiative efficiency
and
̇
M
BHL
=
4
πG
2
M
2
BBH
ρ
v
3
rel
,
(4)
with
v
rel
=
v
k
+
c
s
and
c
s
is the gas sound speed. As the BH is decelerated,
̇
M
BHL
increases. Since
̇
M
BHL
is super-
Eddington typically, not all of the gas in
̇
M
BHL
may end up accreted, but we assume the shock emerges after gas
reprocessing with luminosity
L
BHL
≈
2
.
5
×
10
45
erg s
−
1
(
η
0
.
1
)
(
M
BBH
100
M
)
2
×
(
v
k
200 km s
−
1
)
−
3
(
ρ
10
−
10
g cm
−
3
)
(5)
where we assume
c
s
∼
50 km s
−
1
. Bondi drag slows down the kicked BH from initial kinetic energy 1
/
2
M
BBH
v
2
k
. The
drag force is
̇
M
BHL
v
k
and is equal to
M
BBH
v
k
/t
dec
where
t
dec
is the source-frame deceleration timescale
t
dec
= 224 yr
(
v
k
200 km s
−
1
)
3
(
ρ
10
−
10
g cm
−
3
)
−
1
×
(
M
BBH
100
M
)
−
1
.
(6)
Strong kicks (
v
k
>
1000 km s
−
1
) are possible under specific binary arrangements [60, 61], but as
v
k
→
50 km s
−
1
,
t
dec
∼
3
.
5yr. However, if the event is kicked at an angle
θ
to the mid-plane (
θ
= 0
◦
is in the disk mid-plane and
θ
= 90
◦
is straight up out of the disk), then the EM signature ends when the merged BH exits the disk. The
6
source-frame time for the EM signature to end is
t
end
≈
67 day
(
v
k
200 km s
−
1
)
−
1
(
a
700
r
g
)
×
(
M
SMBH
10
8
M
)(
H/a
0
.
01
)
1
sin (
θ/
60
◦
)
(7)
where
a
∼
700
r
g
is a plausible migration trap location [12], and
H/a
∼
[10
−
3
,
0
.
1] is the disk aspect ratio (i.e., disk
height
H
at radius
a
), with
ρ
∼
O
(10
−
10
)g
/
cm
3
appropriate at that radius [62, 63].
For any EM signature generated below the disk photosphere, the signal will emerge on the photon diffusion timescale
(
t
diff
) which is
t
diff
= 8 day
(
τ
100
)
(
H/a
0
.
01
)(
a
700
r
g
)(
M
SMBH
10
8
M
)
(8)
in the source-frame,
τ
is the optical depth to the midplane (assumed event location). We can treat photon diffusion
from the shocked hot-spot by convolving the shock lightcurve with a Maxwell-Boltzmann distribution with mean time
t
diff
. This has the effect of smearing out the actual emergent lightcurve from the disk surface. We plot the resulting
flare model fit to the ZTF lightcurve in Fig. 2, assuming a linear model for the source continuum. We note that a
kicked black hole merger remnant will produce a roughly constant temperature shock, and this is consistent with the
lack of color evolution for this flare. If ZTF19abanrhr is not an EM counterpart to S190521g, any flare model must
account for this observation.
Parameter estimation.—
For either the ram pressure shock or the BHL shock, given even modest optical depth,
the shape of the observed lightcurve will be dominated by the Maxwell-Boltzmann distribution. From the EM data
we find a best fit
t
diff
= 38
+2
−
1
day (observed frame) and a
t
delay
= 23
+1
−
1
day (observed frame). We also find a best fit
t
end
= 80 day (observed, corresponding to
∼
57 day rest-frame). We also find the total energy released in the flare
(
∼
10
51
erg). By inspection, the
g
−
r
color implies the temperature of the observed flare is too low to permit strong
kicks (
v
k
>
1000 km s
−
1
), and given the relatively brief duration of the flare (
t
flare
∼
40 day in the observed frame,
corresponding to
∼
28 day rest-frame), we must assume the event ends due to the merger remnant exiting the disk
rather than deceleration. Finding
M
BBH
=
O
(100
M
) from the GW data enables us to make order of magnitude
estimates for several system parameters from the EM measurements.
Assuming
M
BBH
∼
100
M
and
t
ram
∼
t
delay
, we estimate
v
k
∼
200 km s
−
1
from eqn. 3 (note
v
k
∝
M
1
/
3
BBH
). The total
energy released corresponds to
t
flare
L
BHL
, so
L
BHL
∼
10
45
erg s
−
1
. Thus,
ρ
∼
10
−
10
g cm
−
3
from eqn. 5, assuming
the energy release is dominated by the BHL shock. With
t
end
∼
80 day (=
v
k
H/
sin
θ
), if we assume the merger
happened near where we would expect a migration trap to occur (i.e.,
a
∼
700
r
g
), then we find an approximate (but
degenerate) combination of
H/a
∼
0
.
01 and
θ
∼
60
◦
for
M
SMBH
∼
10
8
M
.
M
BBH
and
v
k
are the best constrained
parameters, to factors of
∼
2. But, since the uncertainty in
M
SMBH
spans approximately an order of magnitude, the
other parameters estimated above are also uncertain to an order of magnitude.
Other tests of S190521g association.—
A kicked BBH merger in an AGN disk will yield an off-center disk flare,
producing an asymmetric illumination of the AGN broad line region (BLR) clouds. Depending on the flare luminosity,
location, and sightline to the observer, an asymmetric broad line profile will develop within a light-crossing time of
the BLR (
R
BLR
), and decay over
t
flare
[1]. Unfortunately, the first spectrum of this AGN was taken on UT 2020
January 25, or
∼
200 days after the trigger (see Fig. 4). Since the BLR light-crossing time is typically a few weeks,
any line broadening effect is no longer present. Therefore we cannot put useful limits on the off-center nature of the
flare ZTF19abanrhr.
A modest recoil kick velocity
v
k
corresponds to a small perturbation of the BBH Keplerian orbital velocity
v
∼
10
4
km s
−
1
(
a/
10
3
r
g
)
−
1
/
2
.
v
k
is not large enough to escape the AGN. Therefore, in approximately half an orbital
period, the kicked BBH orbit must re-encounter the disk. So, if ZTF19abanrhr is associated with S190521g, we predict
a similar flare (driven by Bondi accretion) in this source on a timescale of 1
.
6 yr (
M
SMBH
/
10
8
M
) (
a/
10
3
r
g
)
3
/
2
.
A massive merger in an AGN disk implies a hierarchical origin for at least the primary BH and therefore a high
likelihood of significant spin, depending on the merger mass ratio [18, 64–66]. Thus we predict that S190521g includes
a significant spin component with the primary BH, and a modest kick velocity [67, 68].
Discussion.—
If we associate ZTF19abanrhr with S190521g, the flare energy is mostly powered by a Bondi accretion
tail, which implies a constant color with time, consistent with our data. For a disk thicker than the Hill sphere of the
merged BBH, the delay between the GW event and the EM counterpart is
∼
t
diff
, the photon diffusion time, which
depends on the AGN disk density (
ρ
) and height (
H
). The temperature measured at the surface of the disk will be
7
FIG. 4. Spectra of J124942.3+344929, the AGN associated with ZTF19abanrhr from SDSS (UT 2006 January 30) and Keck (UT
2020 January 25). Other than fading by
∼
30%, there are no strong spectral changes over the intervening decade (rest-frame).
lower than the shock temperature, while the rise and decline times will increase, preserving the total energy emitted.
The strength of this signal (equation 5) depends on the BBH mass squared (
M
2
BBH
), the recoil kick velocity to the
negative three power (
v
−
3
k
), and the AGN disk gas density (
ρ
). So the brightest EM counterparts are for modestly
kicked, large mass BBH mergers in dense gas disks. In anticipation of future small GW error volumes, SMBH mass
estimates are needed in as many AGN as possible to constrain the EM follow-up cadence for individual AGN. Other
EM generating events will also occur in AGN disks [49] and may correspond to peculiar flares observed in several
AGN [42, 50].
Conclusions.—
We present the first plausible EM counterpart to a BBH merger in an AGN disk. We can rule
out most false-positive models at high (99
.
9%) confidence, and the energetics and color evolution are suggestive of a
constant temperature shock, consistent with a kicked BBH merger remnant. We predict a similar repeat flare in this
source when the kicked BBH re-encounters the disk on timescale 1
.
6 yr (
M
SMBH
/
10
8
M
) (
a/
10
3
r
g
)
3
/
2
. EM campaigns
that trigger follow-up on GW alerts should monitor AGN on multiple cadences, from days to weeks, in order to search
for EM counterparts in the AGN channel.
We thank the referees for useful, timely comments that have improved this manuscript. MJG is supported by the
NSF grants AST-1518308 and AST-1815034, and the NASA grant 16-ADAP16-0232. KESF & BM are supported by
NSF AST-1831415 and Simons Foundation Grant 533845. KESF & BM acknowledge extremely useful conversations
with Mordecai-Mark MacLow and Pierre Marchand. The work of DS was carried out at the Jet Propulsion Laboratory,
California Institute of Technology, under a contract with NASA. MMK acknowledges the GROWTH project funded
by the National Science Foundation under Grant No 1545949. MC is supported by NSF PHY-2010970. Based on
observations obtained with the Samuel Oschin Telescope 48-inch and the 60-inch Telescope at the Palomar Observatory
as part of the Zwicky Transient Facility project. ZTF is supported by the National Science Foundation under Grant
No. AST-1440341 and a collaboration including Caltech, IPAC, the Weizmann Institute for Science, the Oskar Klein
Center at Stockholm University, the University of Maryland, the University of Washington, Deutsches Elektronen-
Synchrotron and Humboldt University, Los Alamos National Laboratories, the TANGO Consortium of Taiwan, the
University of Wisconsin at Milwaukee, and Lawrence Berkeley National Laboratories. Operations are conducted by
COO, IPAC, and UW. The ZTF forced-photometry service was funded under the Heising-Simons Foundation grant
12540303 (PI: Graham).
†
mjg@caltech.edu
[1] B. McKernan, K. E. S. Ford, I. Bartos,
et al.
, Astrophys. J. Lett.
884
, L50 (2019), arXiv:1907.03746 [astro-ph.HE].
[2] B. P. Abbott
et al.
, Phys. Rev. X
9
, 031040 (2019).
[3] K. Belczynski, T. Bulik, C. L. Fryer,
et al.
, Astrophys. J.
714
, 1217 (2010), arXiv:0904.2784 [astro-ph.SR].
[4] S. E. de Mink and I. Mandel, Mon. Notices Royal Astron. Soc
460
, 3545 (2016), arXiv:1603.02291 [astro-ph.HE].
8
[5] C. L. Rodriguez, C.-J. Haster, S. Chatterjee, V. Kalogera, and F. A. Rasio, Astrophys. J. Lett.
824
, L8 (2016),
arXiv:1604.04254 [astro-ph.HE].
[6] C. L. Rodriguez, S. Chatterjee, and F. A. Rasio, Phys. Rev. D
93
, 084029 (2016), arXiv:1602.02444 [astro-ph.HE].
[7] F. Antonini, Astrophys. J.
794
, 106 (2014), arXiv:1402.4865 [astro-ph.GA].
[8] F. Antonini and F. A. Rasio, Astrophys. J.
831
, 187 (2016), arXiv:1606.04889 [astro-ph.HE].
[9] G. Fragione, N. W. C. Leigh, and R. Perna, Mon. Notices Royal Astron. Soc
488
, 2825 (2019), arXiv:1903.09160 [astro-
ph.GA].
[10] B. McKernan, K. E. S. Ford, W. Lyra, and H. B. Perets, Mon. Notices Royal Astron. Soc
425
, 460 (2012), arXiv:1206.2309
[astro-ph.GA].
[11] B. McKernan, K. E. S. Ford, B. Kocsis, W. Lyra, and L. M. Winter, Mon. Notices Royal Astron. Soc
441
, 900 (2014),
arXiv:1403.6433 [astro-ph.GA].
[12] J. M. Bellovary, M.-M. Mac Low, B. McKernan, and K. E. S. Ford, Astrophys. J. Lett.
819
, L17 (2016), arXiv:1511.00005
[astro-ph.GA].
[13] I. Bartos, B. Kocsis, Z. Haiman, and S. M ́arka, Astrophys. J.
835
, 165 (2017), arXiv:1602.03831 [astro-ph.HE].
[14] N. C. Stone, B. D. Metzger, and Z. Haiman, Mon. Notices Royal Astron. Soc
464
, 946 (2017), arXiv:1602.04226 [astro-
ph.GA].
[15] B. McKernan, K. E. S. Ford, J. Bellovary,
et al.
, Astrophys. J.
866
, 66 (2018), arXiv:1702.07818 [astro-ph.HE].
[16] A. Secunda, J. Bellovary, M.-M. Mac Low,
et al.
, Astrophys. J.
878
, 85 (2019), arXiv:1807.02859 [astro-ph.HE].
[17] Y. Yang, I. Bartos, V. Gayathri,
et al.
, Phys. Rev. Lett.
123
, 181101 (2019), arXiv:1906.09281 [astro-ph.HE].
[18] B. McKernan, K. E. S. Ford, R. O’Shaugnessy, and D. Wysocki, Mon. Notices Royal Astron. Soc
494
, 1203 (2020),
arXiv:1907.04356 [astro-ph.HE].
[19] S. E. Woosley, Astrophys. J.
836
, 244 (2017), arXiv:1608.08939 [astro-ph.HE].
[20] D. Gerosa and E. Berti, Phys. Rev. D
100
, 041301 (2019), arXiv:1906.05295 [astro-ph.HE].
[21] K. Chatziioannou, R. Cotesta, S. Ghonge,
et al.
, Phys. Rev. D
100
, 104015 (2019), arXiv:1903.06742 [gr-qc].
[22] B. Zackay, L. Dai, T. Venumadhav, J. Roulet, and M. Zaldarriaga, arXiv e-prints , arXiv:1910.09528 (2019),
arXiv:1910.09528 [astro-ph.HE].
[23] T. Bogdanovi ́c, B. D. Smith, S. Sigurdsson, and M. Eracleous, ApJS
174
, 455 (2008), arXiv:0708.0414 [astro-ph].
[24] E. M. Rossi, G. Lodato, P. J. Armitage, J. E. Pringle, and A. R. King, Mon. Notices Royal Astron. Soc
401
, 2021 (2010),
arXiv:0910.0002 [astro-ph.HE].
[25] L. R. Corrales, Z. Haiman, and A. MacFadyen, Mon. Notices Royal Astron. Soc
404
, 947 (2010), arXiv:0910.0014 [astro-
ph.HE].
[26] E. C. Bellm, S. R. Kulkarni, M. J. Graham,
et al.
, PASP
131
, 018002 (2019), arXiv:1902.01932 [astro-ph.IM].
[27] M. J. Graham, S. R. Kulkarni, E. C. Bellm,
et al.
, PASP
131
, 078001 (2019), arXiv:1902.01945 [astro-ph.IM].
[28] E. C. Bellm, S. R. Kulkarni, T. Barlow, U. Feindt, M. J. Graham, A. Goobar, T. Kupfer, C.-C. Ngeow, P. Nugent, E. Ofek,
T. A. Prince, R. Riddle, R. Walters, and Q.-Z. Ye, PASP
131
, 068003 (2019), arXiv:1905.02209 [astro-ph.IM].
[29] M. T. Patterson, E. C. Bellm, B. Rusholme, F. J. Masci, M. Juric, K. S. Krughoff, V. Z. Golkhou, M. J. Graham, S. R.
Kulkarni, G. Helou, and Zwicky Transient Facility Collaboration, PASP
131
, 018001 (2019), arXiv:1902.02227 [astro-
ph.IM].
[30] E. W. Flesch, arXiv e-prints , arXiv:1912.05614 (2019), arXiv:1912.05614 [astro-ph.GA].
[31] LIGO/VIRGO Collaboration
et al.
, GCN Circular
24621
(2019).
[32] L. P. Singer, H.-Y. Chen, D. E. Holz, W. M. Farr, L. R. Price, V. Raymond, S. B. Cenko, N. Gehrels, J. Cannizzo, M. M.
Kasliwal, S. Nissanke, M. Coughlin, B. Farr, A. L. Urban, S. Vitale, J. Veitch, P. Graff, C. P. L. Berry, S. Mohapatra, and
I. Mandel, ApJS
226
, 10 (2016), arXiv:1605.04242 [astro-ph.IM].
[33] P. F. Hopkins, G. T. Richards, and L. Hernquist, Astrophys. J.
654
, 731 (2007), arXiv:astro-ph/0605678 [astro-ph].
[34] G. Calderone, L. Nicastro, G. Ghisellini, M. Dotti, T. Sbarrato, F. Shankar, and M. Colpi, Mon. Notices Royal Astron.
Soc
472
, 4051 (2017), arXiv:1612.01580 [astro-ph.HE].
[35] J. C. Runnoe, M. S. Brotherton, and Z. Shang, Mon. Notices Royal Astron. Soc
422
, 478 (2012), arXiv:1201.5155 [astro-
ph.CO].
[36] Planck Collaboration
et al.
, Astron. Astrophys.
594
, A13 (2016), arXiv:1502.01589 [astro-ph.CO].
[37] D. Stern, B. McKernan, M. J. Graham,
et al.
, Astrophys. J.
864
, 27 (2018), arXiv:1805.06920 [astro-ph.GA].
[38] N. P. Ross, K. E. S. Ford, M. Graham,
et al.
, Mon. Notices Royal Astron. Soc
480
, 4468 (2018), arXiv:1805.06921
[astro-ph.GA].
[39] R. J. Assef, D. Stern, G. Noirot, H. D. Jun, R. M. Cutri, and P. R. M. Eisenhardt, ApJS
234
, 23 (2018), arXiv:1706.09901
[astro-ph.GA].
[40] B. C. Kelly, J. Bechtold, and A. Siemiginowska, Astrophys. J.
698
, 895 (2009), arXiv:0903.5315 [astro-ph.CO].
[41] J. Moreno, M. S. Vogeley, G. T. Richards, and W. Yu, PASP
131
, 063001 (2019), arXiv:1811.00154 [astro-ph.IM].
[42] M. J. Graham, S. G. Djorgovski, A. J. Drake, D. Stern, A. A. Mahabal, E. Glikman, S. Larson, and E. Christensen, Mon.
Notices Royal Astron. Soc
470
, 4112 (2017), arXiv:1706.03079 [astro-ph.GA].
[43] A. J. Drake, S. G. Djorgovski, A. Mahabal, E. Beshore, S. Larson, M. J. Graham, R. Williams, E. Christensen, M. Catelan,
A. Boattini, A. Gibbs, R. Hill, and R. Kowalski, Astrophys. J.
696
, 870 (2009), arXiv:0809.1394 [astro-ph].
[44] R. J. Assef, J. L. Prieto, D. Stern, R. M. Cutri, P. R. M. Eisenhardt, M. J. Graham, H. D. Jun, A. Rest, H. A. Flewelling,
N. Kaiser, R. P. Kudritzki, and C. Waters, Astrophys. J.
866
, 26 (2018), arXiv:1807.07985 [astro-ph.GA].
[45] D. Kasen and L. Bildsten, Astrophys. J.
717
, 245 (2010), arXiv:0911.0680 [astro-ph.HE].
[46] R. J. Foley, N. E. Sanders, and R. P. Kirshner, Astrophys. J.
742
, 89 (2011), arXiv:1107.3555 [astro-ph.CO].
9
[47] A. Lawrence, A. G. Bruce, C. MacLeod, S. Gezari, M. Elvis, M. Ward, S. J. Smartt, K. W. Smith, D. Wright, M. Fraser,
P. Marshall, N. Kaiser, W. Burgett, E. Magnier, J. Tonry, K. Chambers, R. Wainscoat, C. Waters, P. Price, N. Metcalfe,
S. Valenti, R. Kotak, A. Mead, C. Inserra, T. W. Chen, and A. Soderberg, Mon. Notices Royal Astron. Soc
463
, 296
(2016), arXiv:1605.07842 [astro-ph.HE].
[48] M. J. Rees, Nature (London)
333
, 523 (1988).
[49] B. McKernan, K. E. S. Ford, and R. O’Shaughnessy, arXiv e-prints , arXiv:2002.00046 (2020), arXiv:2002.00046 [astro-
ph.HE].
[50] G. Cannizzaro, M. Fraser, P. G. Jonker,
et al.
, Mon. Notices Royal Astron. Soc , 180 (2020), arXiv:2001.07446 [astro-ph.HE].
[51] S. Rosswog, E. Ramirez-Ruiz, and W. R. Hix, Astrophys. J.
695
, 404 (2009), arXiv:0808.2143 [astro-ph].
[52] C. P. L. Berry, I. Mandel, H. Middleton,
et al.
, Astrophys. J.
804
, 114 (2015), arXiv:1411.6934 [astro-ph.HE].
[53] L. S. Finn and D. F. Chernoff, Phys. Rev. D
47
, 2198 (1993), arXiv:gr-qc/9301003 [gr-qc].
[54] e. a. Abbott, R., arXiv e-prints , arXiv:2004.08342 (2020), arXiv:2004.08342 [astro-ph.HE].
[55] M. Campanelli, C. Lousto, Y. Zlochower, and D. Merritt, Astrophys. J. Lett.
659
, L5 (2007), arXiv:gr-qc/0701164 [gr-qc].
[56] J. A. Gonz ́alez, M. Hannam, U. Sperhake, B. Br ̈ugmann, and S. Husa, Phys. Rev. Lett.
98
, 231101 (2007), arXiv:gr-
qc/0702052 [gr-qc].
[57] E. C. Ostriker, Astrophys. J.
513
, 252 (1999), arXiv:astro-ph/9810324 [astro-ph].
[58] A. Antoni, M. MacLeod, and E. Ramirez-Ruiz, Astrophys. J.
884
, 10.3847/1538-4357/ab3466 (2019), arXiv:1901.07572v1
[astro-ph.HE].
[59] E. Tejeda and A. Aguayo-Ortiz, Mon. Notices Royal Astron. Soc
487
, 3607 (2019), arXiv:1906.04923 [astro-ph.HE].
[60] Y. Zlochower and C. O. Lousto, Phys. Rev. D
92
, 024022 (2015), arXiv:1503.07536 [gr-qc].
[61] J. Healy and C. O. Lousto, Phys. Rev. D
95
, 024037 (2017), arXiv:1610.09713 [gr-qc].
[62] E. Sirko and J. Goodman, Mon. Notices Royal Astron. Soc
341
, 501 (2003), arXiv:astro-ph/0209469 [astro-ph].
[63] T. A. Thompson, E. Quataert, and N. Murray, Astrophys. J.
630
, 167 (2005), arXiv:astro-ph/0503027 [astro-ph].
[64] E. Berti and M. Volonteri, Astrophys. J.
684
, 822 (2008), arXiv:0802.0025 [astro-ph].
[65] D. Gerosa and E. Berti, Phys. Rev. D
95
, 124046 (2017), arXiv:1703.06223 [gr-qc].
[66] M. Fishbach and D. E. Holz, Astrophys. J. Lett.
851
, L25 (2017), arXiv:1709.08584 [astro-ph.HE].
[67] D. Gerosa and C. J. Moore, Phys. Rev. Lett.
117
, 011101 (2016), arXiv:1606.04226 [gr-qc].
[68] V. Varma, M. Isi, and S. Biscoveanu, Phys. Rev. Lett.
124
, 101104 (2020), arXiv:2002.00296 [gr-qc].