GCN Circular 20755

V. Savchenko (ISDC, University of Geneva, CH)
on behalf of the INTEGRAL group:
S. Mereghetti (IASF-Milano, Italy),
C. Ferrigno ((ISDC, University of Geneva, CH),
E. Kuulkers (ESTEC/ESA, The Netherlands),
A. Bazzano (IAPS-Roma, Italy), E. Bozzo,
T. J.-L. Courvoisier (ISDC, University of Geneva, CH)
S. Brandt (DTU - Denmark) R. Diehl (MPE-Garching, Germany)
L. Hanlon (UCD, Ireland) P. Laurent (APC, Saclay/CEA, France)
A. Lutovinov (IKI, Russia) J.P. Roques (CESR, France)
R. Sunyaev (IKI, Russia) P. Ubertini (IAPS-Roma, Italy)

We investigated serendipitous INTEGRAL observations carried out at the
time of the LIGO/Virgo burst candidate G275404.  The satellite was
pointing at RA=17:44:43 Dec=-25:57:27 (in the direction of the
Galactic Center region), close to the low-probability area of LIGO
localization. About 2% of the probability was contained in the field
of view of INTEGRAL IBIS and SPI. Depending on the location within the
LIGO 90% localization region, as well as the assumed counterpart
spectrum and duration, the best upper limit is set by the
anti-coincidence shield of the spectrometer on board of INTEGRAL
(SPI/ACS), the anti-coincidence shield of the IBIS instrument
(IBIS/Veto), or by the imaging coded mask instruments (IBIS and SPI).
The combination of these instruments covered the full LIGO 90%
confidence region and provided stringent constraints on the flux of a
possible electromagnetic counterpart in the energy range covered by
the INTEGRAL instruments.

The INTEGRAL Burst Alert System (IBAS) did not identify any unusual
transients in coincidence with the LIGO/Virgo trigger. The IBAS
inspects both ISGRI Field of View and all-sky SPI-ACS light curve.

The IBAS trigger closest in time to the GW candidate ("Weak" ISGRI
trigger 7706) happened at UTC 2017-02-25 11:32:15.56 and it's location
(RA=270.48, Dec=-25.08, 3.73 arcmin radius 68% containment,
statistical uncertainty only) was outside the 90% probability
containment of the LIGO localization region.

The nearest excess over the background rate identified by SPI-ACS
offline pipeline happened 453 seconds after the LIGO/Virgo trigger. It
has an SNR of 5.5, duration of 100 ms, and is compatible with cosmic
ray origin.

We investigated the SPI-ACS, IBIS/Veto, and IBIS/ISGRI light curves
between -500 and +500 s from the trigger time (2017-02-25 18:30:21.3
UTC) on temporal scales from 0.1 to 100 s, and found no evidence for
any deviation from the background.  We estimate combined typical
3-sigma upper limits of 4.3e-7 erg/cm2 (75-2000 keV) for 8s duration
assuming Band model parameters alpha=-1, beta=-2.5, and E_ peak = 300
keV.  To derive a limit for a typical short burst with 1 s duration,
we use a harder cutoff power law spectrum with a slope of -0.5 and an
Epeak = 500 keV: we find a limiting fluence of 1.5e-7 erg/cm2 (75-2000
keV) at 3 sigma c.l. These limits assume a perpendicular direction of
the burst to the INTEGRAL pointing direction, optimal for SPI-ACS
sensitivity. However the extent of the region with optimal response
depends on the possible source spectrum: we perform a detailed
calculation only for a cutoff powerlaw spectrum with a slope of -0.5
and an Epeak = 500 keV: we estimate that 30% of the LIGO localization
probability region is covered with a range of sensitivity from optimal
for SPI-ACS (mentioned above) to 50% worse. About 2% of the LIGO
localization in the field of view of IBIS and SPI is covered with at
least factor 2 better sensitivity.

The SPI/ACS light curves, binned at 50 ms, are derived from 91
independent detectors with different lower energy thresholds (mainly
between 50 keV and 150 keV) and an upper threshold at about 100
MeV. The ACS response varies substantially as a function of the source
incident angle with an optimal effective area of about 6000 cm2 at 1
MeV.