J/MNRAS/508/3569 Catalogue of ULX's line features (Kosec+, 2021)
Ionized emission and absorption in a large sample of ultraluminous
X-ray sources.
Kosec P., Pinto C., Reynolds C.S., Guainazzi M., Kara E., Walton D.J.,
Fabian A.C., Parker M.L., Valtchanov I.
<Mon. Not. R. Astron. Soc. 508, 3569-3588 (2021)>
=2021MNRAS.508.3569K 2021MNRAS.508.3569K (SIMBAD/NED BibCode)
ADC_Keywords: X-ray sources ; Accretion ; Black holes ; Stars, neutron ;
Spectroscopy ; Space velocities ; Energy distributions ;
Equivalent widths
Keywords: accretion, accretion discs - X-rays: binaries
Abstract:
Most ultraluminous X-ray sources (ULXs) are thought to be powered by
super- Eddington accretion on to stellar-mass compact objects.
Accretors in this extreme regime are naturally expected to ionize
copious amounts of plasma in their vicinity and launch powerful
radiation-driven outflows from their discs. High spectral resolution
X-ray observations [with reflection grating spectrometer (RGS)
gratings onboard XMM-Newton] of a few ULXs with the best data sets
indeed found complex line spectra and confirmed such extreme
(0.1-0.3c) winds. However, a search for plasma signatures in a large
ULX sample with a rigorous technique has never been performed, thereby
preventing us from understanding their statistical properties such as
the rate of occurrence, to constrain the outflow geometry, and its
duty cycle. We developed a fast method for automated line detection in
X-ray spectra and applied it to the full RGS ULX archive, rigorously
quantifying the statistical significance of any candidate lines.
Collecting the 135 most significant features detected in 89
observations of 19 objects, we created the first catalogue of spectral
lines detected in soft X-ray ULX spectra. We found that the detected
emission lines are concentrated around known rest-frame elemental
transitions and thus originate from low- velocity material. The
absorption lines instead avoid these transitions, suggesting they were
imprinted by blueshifted outflows. Such winds therefore appear common
among the ULX population. Additionally, we found that spectrally hard
ULXs show fewer line detections than soft ULXs, indicating some
difference in their accretion geometry and orientation, possibly
causing overionization of plasma by the harder spectral energy
distributions of harder ULXs.
Description:
In our study of ULXs, we used and developped an automated spectral
method to detect absorption and emission line features of ULX objects.
In this work, we study all ULXs with high enough quality X-ray data,
and search their high-resolution spectra for any emission or
absorption lines, while quantifying the true statistical significance
of any features detected. All the detected features are included in a
single catalogue, thus creating the first catalogue of spectral line
detections in ULXs. The catalogue will allow us to make the first
statistical comparison of spectral lines observed in ULXs, important
to obtain model-independent diagnostics on the line significance, rate
of occurrence and nature. It will also be crucial for future
observations and more sensitive missions such as XRISM and Athena, by
highlighting promising ULXs to observe in future observational
campaigns.
This work builds upon our first systematic search for spectral
features in ULXs (described in Kosec et al. 2018MNRAS.473.5680K 2018MNRAS.473.5680K),
where a smaller sample (10 sources) was studied using the traditional
Gaussian search method. Most of the detected significances were
quantified only tentatively. We performed the fully rigorous search
with simulations on the four most promising sources.
Here, the sample is expanded to include all suitable ULX data sets (18
sources in total). As it is prohibitively expensive to search the full
sample with the current automated methods (assessing the detection
significances rigorously), we have developed a new, fast search
method, employing cross-correlation to search X-ray spectra for
spectral features. We apply the method to scan ULX spectra for
Gaussian lines. However, the method can in principle be used to search
the spectrum of any astrophysical source from any instrument for any
spectral feature of interest,
(Please refer to the section 1 Introduction).
The high-spectral resolution X-ray instrument in current use is the
reflection grating spectrometer (RGS; den Herder et al.
2001A&A...365L...7D 2001A&A...365L...7D) onboard XMM-Newton (Jansen et al.
2001A&A...365L...1J 2001A&A...365L...1J). Its collecting area is significantly higher than
that of HETG, but its energy band is narrower (0.35-1.8 keV RGS
bandpass versus 0.4-10 keV HETG). However, this is the energy band
which contains nitrogen, oxygen, neon, magnesium, and iron transitions
and is thus of great importance (Kaastra et al. 2008SSRv..134..155K 2008SSRv..134..155K).
These elements normally provide the strongest lines in X-ray binary
and AGN spectra unless the elements are too ionized. The RGS is
therefore the main instrument of this study.
Secondly, the imaging CCD-based and silicon-drift X-ray instruments
such as EPIC (Struder et al. 2001A&A...365L..27T 2001A&A...365L..27T; Turner et al.
2001A&A...365L..27T 2001A&A...365L..27T) onboard XMM-Newton, ACIS onboard Chandra
(Weisskopf et al. 2000SPIE.4012....2W 2000SPIE.4012....2W), NICER (Gendreau et al.
2016SPIE.9905E..1HG), and eROSITA (Predehl et al. 2021A&A...647A...1P 2021A&A...647A...1P)
offer a spectral resolution of roughly 100 eV. This is too poor to
resolve individual emission or absorption lines unless they are
isolated, particularly in the soft X-ray band (0.3-2 keV). We
therefore do not search data from X-ray CCD-based instruments for
narrow spectral features in this work. Nevertheless, we use
XMM-Newton EPIC data to constrain the broad-band continua of ULXs in
the 0.3-10 keV energy range. We select all ULX observations with good
enough quality RGS data for our sample. The criteria for a good
quality data set have previously been defined by Kosec et al.
(2018MNRAS.473.5680K 2018MNRAS.473.5680K) and we describe it in the section 2 The ULX
sample. The final sample contains 16 ULXs and 2 super-Eddington pulsars.
As explained and tested in the section 3 The croos-correlation method
and 4 results, we follows steps of the cross-correlation analysis to
produce a high quality catalogue of the strongest detected features
(true significance , TS > 1σ) containing 135 spectral lines, of
which 82 are emission and 53 are absorption lines. Statistical and
physical results are presented in the table.dat. Additionnaly, the
source.dat regroups the 18 ULX objects names as well as their
positional data.
All of the data underlying this article are publicly available from
ESA's XMM-Newton Science Archive
(https://www.cosmos.esa.int/web/XMM-Newtonlxsa) and NASA's HEASARC
archive (https://heasarc.gsfc.nasa.gov/). More, explanatins on
catalogue structure (i.e appendix A), the cross correlation method
steps (i.e appendix B Data reduction, Continuum modelling,
Pre-filtering the data, Generating real and simulated residual
spectra, Generating spectral models, Cross-correlation, Collecting
results). Finally, the statistics of each individual objects are
available in the appendix C.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
source.dat 67 18 List of the objects used in this study
table.dat 229 135 The 135 strongest detected spectral line
features
--------------------------------------------------------------------------------
Byte-by-byte Description of file: source.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 14 A14 --- Name Object name (source)
16- 17 I2 h RAh Hour of right ascension (J2000)
19- 20 I2 min RAm Minute of right ascension (J2000)
22- 26 F5.2 s RAs Second of right ascension (J2000)
28 A1 --- DE- Sign of declination (J2000)
29- 30 I2 deg DEd Degree of declination (J2000)
32- 33 I2 arcmin DEm Arcminute of declination (J2000)
35- 38 F4.1 arcsec DEs Arcsecond of declination (J2000)
40- 67 A28 --- SName Simbad name
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 14 A14 --- Name Object name (source) (1)
16- 25 A10 --- Approach Name of the approach based on the
observations (approach)
27- 36 E10.5 0.1nm Lambda Observational wavelength of the line
feature (wavelength)
38- 47 E10.5 keV Energy Observational photon energy of the line
feature (energy)
49- 58 E10.5 km/s Width Velocity width at which the line shows the
strongest cross-correlation (turb_vel)
60- 69 E10.5 --- pval The true-false positive rate of the line
feature (true_pval)
71- 81 E11.5 --- Sig Significance of the line feature (true_sig)
83- 93 E11.5 --- RC The renormalized correlation of the line
feature (ren_cor)
95-104 E10.5 --- spval The single trial false positive rate of the
line feature (s_pval)
106-116 E11.5 --- STS The single trial significance (STS)
118-127 E10.5 --- dcstat The ΔC-stat fit improvement value
obtained upon directly fitting the feature
in the spex fitting package (del_cstat) (2)
129-139 E11.5 ph/cm2/s Pflux The photon flux of the line feature
(phot_flux)
141-150 E10.5 ph/cm2/s e_Pflux Mean lower errorbar of Pflux (e_photflux)
152-161 E10.5 ph/cm2/s E_Pflux Mean upper errorbar of Pflux (E_photflux)
163-173 E11.5 erg/cm2/s Eflux Line energy flux (en_flux)
175-184 E10.5 erg/cm2/s e_Eflux Mean lower errorbar of Eflux (e_enflux)
186-195 E10.5 erg/cm2/s E_Eflux Mean upper errorbar of Pflux (E_enflux)
197-207 E11.5 keV EW The line equivalent width (eq_width)
209-218 E10.5 keV e_EW Mean lower errorbar of Eflux (e_eqwidth)
220-229 E10.5 keV E_EW Mean upper errorbar of Eflux (E_eqwidth)
--------------------------------------------------------------------------------
Note (1): Here the list of the 18 objects with the number of detected lines :
CircinusULX5 number of detected line feature : 4
HolmbergIIX1 number of detected line feature : 11
HolmbergIXX1 number of detected line feature : 2
IC342X1 number of detected line feature : 5
M33X8 number of detected line feature : 1
NGC1313X1 number of detected line feature : 19
NGC1313X2 number of detected line feature : 7
NGC247ULX number of detected line feature : 14
NGC300ULX1 number of detected line feature : 4
NGC4559X7 number of detected line feature : 6
NGC5204X1 number of detected line feature : 3
NGC5408X1 number of detected line feature : 15
NGC55ULX number of detected line feature : 22
NGC5643X1 number of detected line feature : 2
NGC6946X1 number of detected line feature : 5
NGC7793P13 number of detected line feature : 3
RXJ0209d6m7427 number of detected line feature : 10
SMCX3 number of detected line feature : 2
Note (2): Such methods have recently been developed and successfully applied to
detect outflows in a few ULXs and in active galactic nuclei. These
methods directly fit the X-ray spectra in spectral fitting packages
such as spex (Kaastra et al. 1996uxsa.conf..411K).
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Luc Trabelsi [CDS] 14-Aug-2024