MNRAS 519, 5848–5858 (2023) https://doi.org/10.1093/mnras/stad002 Advance Access publication 2023 January 5 Multiwavelength scrutiny of X-ray sources in dwarf galaxies: ULXs versus AGNs Erica 1 Thygesen , Richard M. Plotkin ,2 , 3 ‹ Roberto Soria ,4 , 5 , 6 7 Amy E. Reines, Jenny E. Greene,8 Gemma E. Anderson ,9 Vivienne F. 10 2 1 11 Baldassare , Milo G. Owens, Ryan T. Urquhart, Elena Gallo, James C. A. Miller-Jones ,9 Jeremiah D. Paul 2 and Alexandar P. Rollings2 1 Center for Data Intensive and Time Domain Astronomy, Department of Physics and Astronomy, Michigan State University, East Lansing, MI 48824, USA 2 Department of Physics, University of Nevada, Reno, NV 89557, USA 3 Nevada Center for Astrophysics, University of Nevada, Las Vegas, NV 89154, USA 4 College of Astronomy and Space Sciences, University of the Chinese Academy of Sciences, Beijing 100049, China 5 INAF – Osservatorio Astrofisico di Torino, Strada Osservatorio 20, I-10025 Pino Torinese, Italy 6 Sydney Institute for Astronomy, School of Physics A28, The University of Sydney, Sydney, NSW 2006, Australia 7 eXtreme Gravity Institute, Montana State University, Bozeman, MT 59717, USA 8 Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544, USA 9 International Centre for Radio Astronomy Research, Curtin University, GPO Box U1987, Perth, WA 6845, Australia 10 Department of Physics and Astronomy, Washington State University, Pullman, WA 99163, USA 11 Department of Astronomy, University of Michigan, 1085 S University, Ann Arbor, MI 48109, USA Accepted 2022 December 22. Received 2022 December 9; in original form 2022 October 18 A B S T R A C T Owing to their quiet evolutionary histories, nearby dwarf galaxies (stellar masses 9 M   3 × 10 M ) have the potential to teach us about the mechanism(s) that ‘seeded’ the growth of supermassive black holes, and also how the first stellar mass black holes formed and interacted with their environments. Here, we present high spatial resolution observations of three dwarf galaxies in the X-ray ( Chandra ), the optical/near-infrared ( Hubble Space Telescope ), and the radio (Karl G. Jansky Very Large Array). These three galaxies were previously identified as hosting candidate active galactic nuclei on the basis of lower resolution X-ray imaging. With our new observations, we find that X-ray sources in two galaxies (SDSS J121326.01 + 543631.6 and SDSS J122111.29 + 173819.1) are off-nuclear and lack corresponding radio emission, implying they are likely luminous X-ray binaries. The third galaxy (Mrk 1434) contains two X-ray sources (each with L X ≈ 1040 erg s−1 ) separated by 2.8 arcsec, has a low metallicity [12 + log(O/H) = 7.8], and emits nebular He II λ4686 line emission. The northern source has spatially coincident point-like radio emission at 9.0 GHz and extended radio emission at 5.5 GHz. We discuss X-ray binary interpretations (where an ultraluminous X-ray source blows a ‘radio bubble’) and active galactic nucleus interpretations 5 (where an ≈ 4 × 10 M  black hole launches a jet). In either case, we find that the He II emission cannot be photoionized by the X-ray source, unless the source was ≈30–90 times more luminous several hundred years ago. Key words: stars: black holes – galaxies: dwarf – radio continuum: galaxies – X-rays: galaxies. 2 1 M T b 1 a g M s g g D k 2 B b S c K g 2 t G a  Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023 I N T RO D U C T I O N here is abundant evidence that supermassive black holes (SMBHs; 06  M BH  109 M ) ubiquitously exist at the centres of large alaxies (e.g. Kormendy & Ho 2013 ), some of which accrete and hine as active galactic nuclei (AGNs). Some lower mass dwarf alaxies 9 (which we define by stellar masses M   3 × 10 M ) are nown to host nuclear black holes (e.g. Filippenko & Ho 2003 ; arth et al. 2004 ; Reines et al. 2011 ; Reines, Greene & Geha 2013 ; chramm et al. 2013 ; Moran et al. 2014 ; Sartori et al. 2015 ; Ho & im 2016 ; Mezcua et al. 2016 , 2018 ; Pardo et al. 2016 ; Chen et al. 017 ; Chilingarian et al. 2018 ; Nguyen et al. 2019 ; Baldassare, eha & Greene 2020 ; Mart ́ınez-Palomera et al. 2020 ; Cann et al.t ( E-mail: rplotkin@unr.edu V Pub021 ; Schutte & Reines 2022 ), with some mass estimates as low as BH ≈ 104 M  (e.g. Baldassare et al. 2015 ; Woo et al. 2019 ). These lack holes represent the lower mass end of the SMBH population, nd we refer to them here as ‘massive black holes’ (mBHs; 104  BH  106 M ). An actively accreting mBH can affect how dwarf alaxies provide feedback to their larger scale environments (e.g. ashyan et al. 2018 ; Trebitsch et al. 2018 ; Mezcua, Suh & Civ a no 019 ), and more generally, mBHs represent a phase that nuclear lack holes must pass through as they grow to SMBH sizes ov e r osmological time-scales (e.g. Volonteri 2010 ). Given that dwarf alaxies have had relatively quiet evolutionary histories, constraining he fraction of dwarf galaxies hosting mBHs in the local Universe, long with the mBH mass distribution, may provide constraints on he mechanism(s) that formed the first black holes in the Universe e.g. Ricarte & Natarajan 2018 ; Inayoshi, Visbal & Haiman 2020 ; olonteri, Habouzit & Colpi 2021 ). The fraction of dwarf galaxies© 2023 The Author(s) lished by Oxford University Press on behalf of Royal Astronomical Society X-ray sources in dwarf galaxies 5849 h T c L G r p a w b e g s s f t i 2 t I o e 2 P a 2 s O m ( e 0 f ( e w ( X s ( n a s c X m t M f b L p s f u i o r p e n s t a a 1 a h t 0 r M i B a S e e w L t e o ( o r w u s p u M m b d G e p f u o e o s d F 1 a X Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023osting an mBH is still relatively unknown, with current empirical onstraints implying  30 −50 per cent occupation (Miller et al. 2015 ; allo & Sesana 2019 ; Greene, Strader & Ho 2020 ). Stellar mass black holes ( M BH ≈ 10 M ) and neutron stars are lso observed within some dwarf galaxies in the form of X-ray inaries (XRBs). XRBs serve as probes of stellar populations within alaxies, with the number and/or luminosity of XRBs expected to cale with the star formation rate, stellar mass, and metallicity of he host galaxy (e.g. Grimm, Gilfanov & Sunyaev 2003 ; Gilfanov 004 ; Linden et al. 2010 ; Mineo et al. 2014 ; Lehmer et al. 2021 ). ntriguingly, lower metallicity galaxies appear to contain an excess f luminous XRBs compared to solar-metallicity galaxies (Prestwich t al. 2013 ; Brorby, Kaaret & Prestwich 2014 ; Douna et al. 2015 ; onnada, Brorby & Kaaret 2020 ; Lehmer et al. 2021 ), which may be consequence of lower metallicity progenitor stars having weaker tellar winds, and therefore producing black hole remnants that are ore numerous and/or more massive (e.g. Heger et al. 2003 ; Mapelli t al. 2010 ). Besides tracing stellar populations, the energy output rom XRBs can also provide feedback to their host galaxies. For xample, line emission from the high-ionization He II λ4686 line χ ion = 54.4 eV) has been observed from some ultraluminous X-ray ources (ULXs),1 which is often interpreted as an X-ray photoionized ebula (Pakull & Angebault 1986 ; Moon et al. 2011 ). Extrapolating uch ULX phenomenology in the local Universe to higher redshifts, RBs could have contributed to the X-ray heating of the intergalactic edium during the Epoch of Reionization and Cosmic Dawn (e.g. irabel et al. 2011 ; Ponnada et al. 2020 ). Thus, characterizing oth the XRB and mBH populations in nearby dwarf galaxies, articularly as a function of host galaxy metallicity, is important or understanding the formation of the first black holes and galaxies n the Universe. X-ray observations are commonly used to identify accreting com- act objects, since hard X-ray emission (  1–2 keV) is a universal sig- ature of accretion. How ev er, in sev eral cases, it is very challenging o determine the mass of an accreting object via X-ray observations lone. In particular, both a rapidly accreting XRB and a weakly ccreting mBH/SMBH can have comparable X-ray luminosities in he 1039 −1041 erg s−1 range, and they can also display similar X- ay spectral shapes (below ≈50 keV). Folding in multiwavelength nformation is therefore essential for differentiating between rapidly ccreting XRBs and weakly accreting mBHs/SMBHs. It is well stablished that weakly accreting SMBHs (i.e. low-luminosity AGNs ith L bol  0.01 L Edd, where L bol is the bolometric luminosity and 38 −1 Edd = 1.3 × 10 [ M /M ] erg s is the Eddington luminosity) BH  mit compact, usually unresolved, radio emission with a flat spectrum f ∝ ν−α ν , where f ν is the radio flux density at frequency ν, and the adio spectral index α = 0 for a flat spectrum; Ho 2008 ). Such nresolved, flat spectrum radio emission is usually interpreted as a artially self-absorbed synchrotron jet (Blandford & K ̈onigl 1979 ). eanwhile, rapidly accreting XRBs do not launch jets that would e detectable beyond distances of a few Mpc (Fender, Belloni & allo 2004 ). Thus, the presence of unresolved radio emission has the otential to exclude hard X-ray sources as rapidly accreting XRBs. In this paper, we present high spatial resolution X-ray ( Chandra ), ptical/near-infrared ( Hubble Space Telescope , HST ), and radio bservations (Karl G. Jansky Very Large Array, VLA) of three nearby warf galaxies that each host at least one hard X-ray point source. We define ULXs as having X-ray luminosities L > 1039 −1 2 X erg s . ULXs re most commonly interpreted as super-Eddington neutron star or black hole 3 RBs (see e.g. Feng & Soria 2011 ; Kaaret, Feng & Roberts 2017 ). these three targets were initially identified as AGN candidates by emons et al. ( 2015 ), but with the caveat that the positions of their X- ay sources were poorly determined. From the multiwavelength data resented here, we better locate the positions of the X-ray sources ithin these three galaxies, and we attempt to constrain the nature of ach source (i.e. XRB or mBH). In Section 2 , we detail our sample election and data reduction. We outline our results in Section 3 , ollowed by a discussion in Section 4 . Our conclusions are presented n Section 5 . Unless stated otherwise, uncertainties are reported at he 68 per cent confidence level. OBSERVATI ONS A N D DATA R E D U C T I O N .1 Sample ur three targets were selected from the surve y by Lemons et al. 2015 ) who cross-matched ∼44 000 nearby dwarf galaxies ( z < .055) from the NASA–Sloan Atlas2 to the Chandra Source Catalog CSC Release 1.1; Evans et al. 2010 ). They identified 19 galaxies ith hard X-ray point sources (2–7 keV), of which 10 contained an -ray source positionally consistent with the galaxy optical centre given positional uncertainties, we note that not every galaxy has well-defined nucleus). They presented these 10 galaxies as AGN andidates.3 Chandra prov ides ex quisite spatial resolution ( ≈0.4 arcsec) for argets located at the telescope’s aimpoint, but the resolution degrades or sources located fa rther awa y. Of the 10 AGN candidates in emons et al. ( 2015 ), they found that four galaxies contain X-ray ources that are far enough from the aimpoint to have large positional ncertainties ( > 5 arcsec, which is comparable to the projected size f the entire dwarf galaxy). Of these four galaxies, three contained X- ay sources with hard X-ray luminosities > 3 σ ( > 1.2 dex) larger than xpected from the galaxy-wide contribution from XRBs, given the tellar mass and star formation rate of each galaxy (see sections 4.3 nd 4.4 of Lemons et al. 2015 ). These three galaxies include Mrk 434 ( z = 0.00747), SDSS J121326.01 + 543631.6 ( z = 0.00797, ereafter SDSS J1213), and SDSS J122111.29 + 173819.1 ( z = .00699, hereafter SDSS J1221; see Table 1 ). Of particular interest, rk 1434 is a metal-poor galaxy [12 + log(O/H) = 7.8; Shirazi & rinchmann 2012 ] and its optical spectrum from the Sloan Digital ky Surve y (SDSS; York et al. 2000 ) shows nebular He II line mission (Shirazi & Brinchmann 2012 ). To better constrain the locations of the X-ray sources relative o their host galaxies, we obtained new Chandra X-ray and HST ptical/near-infrared observations for these three galaxies. We also btained new VLA radio observations for one target, SDSS J1213, hile archiv a l VLA data were already ava ilable for the other two ources. We adopt distances for each galaxy based on their redshifts, sing H 0 = 73 km s−1 Mpc−1 , except for SDSS J1221, which is a ember of the Virgo cluster (VCC 459). For this galaxy, we use a istance of 16.1 Mpc based on the Tully–Fisher relation (Kashibadze t al. 2020 ). For all three galaxies, we adopt star formation rates rom Lemons et al. ( 2015 ), which are based on (dust-corrected) far- ltraviolet and infrared luminosities and the relationships from Hao t al. ( 2011 ) and Kennicutt & Evans ( 2012 ). For SDSS J1221, we cale the star formation rate from Lemons et al. ( 2015 ) to 16.1 Mpc. or stellar mass estimates, following Reines & Volonteri ( 2015 ), we ht tp://www.nsat las.org/ Since publication of Lemons et al. ( 2015 ), there is new theoretical evidence hat mBHs do not need to reside in the nucleus (e.g. Bellovary et al. 2019 ). MNRAS 519, 5848–5858 (2023) 5850 E. Thygesen et al. Table 1. Properties of the three dwarf galaxies in our sample. Column 1: o galaxy names. The full designations of the second and third galaxies are p SDSS J121326.01 + 543631.6 and SDSS J122111.29 + 173819.1. Column 9 2: distances to each galaxy, assuming H 0 = 73 km s−1 Mpc−1 for Mrk f 1434 and SDSS J1213, and using the Tully–Fisher relation for SDSS w J1221 (Kashibadze, Karachentsev & Karachentseva 2020 ). Column 3: stellar masses, following the methodology of Reines & Volonteri ( 2015 t ). Column 4: logarithm of star formation o rates, based on far- ultrav iolet and infrared luminosities (Hao et al. 2011 ; Kennicutt & Evans 2012 ). Column p 5: metallicities when available in the literature [taken from Shirazi a & Brinchmann ( 2012 ) for Mrk 1434 and Zhao, Gao & Gu ( 2013 ) for SDSS ( J1221]. u Name D log M  log SFR 12 + log(O/H) w (Mpc) (M ) −1 (M  yr ) w (1) (2) (3) (4) (5) p Mrk 1434 30.7 6.6 −0.9 7.8 T SDSS J1213 32.7 7.3 n −2.2 – SDSS J1221 w 16.1 8.0 −1.5 8.3 i b u b R c s 2 ( W t w ( A W w t c v ( fl v p t l t 2 b W H f M t F i f C o r o s f p a e A c a h ( t w S S X i J a C  w p g t f 4 https:// cxc.harvard.edu/ cal/ ASPECT/celmon/ d 5 https://hst-docs.stsci.edu/drizzpac h 6 Note that we aligned HST images to the Gaia frame and the Chandra X-ray p images to the SDSS frame because we generally found a larger number of t common HST / Gaia sources versus common HST /SDSS sources (and vice s ve rsa for Chandra ). Compared to the statistical uncertainty on each Chandra M Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023se the colour-dependent mass-to-light ratios from Zibetti, Charlot & ix ( 2009 ). .2 Chandra e obtained new Chandra observations (Cycle 17; PI Plotkin) ith each galaxy centred at the aimpoint of the S3 chip of the dvanced CCD Imaging Spectrometer (Garmire et al. 2003 ). Data ere telemetered in VFAINT mode. Chandra data reduction was arried out using the Chandra Interactive Analysis of Observations CIAO ) software version 4.13 (Fruscione et al. 2006 ) and CALDB 4.9.5. The Chandra data were reprocessed using chandra repro o create new level 2 event files and bad pixel files, and to apply the atest calibration files. We then searched for background flares using he deflare script, and we did not find any periods with elev a ted ackground levels. Next, we aligned the event file astrometry to the SDSS reference rame. We first excluded areas on each X-ray image occupied by he dwarf galaxy, so that our astrometric corrections would not be nfluenced by sources within each target galaxy. We then filtered each handra image to 0.5–7.0 keV and ran wavdetect to identify X- ay point sources, adopting wavelet scales of 1, 2, 4, 8, and 16, etting sigthresh to 10−6 (i.e. approximately one false positive er chip), and using a point spread function map (at 2.3 keV) with an nclosed count fraction (ecf) of 0.9. The relatively large ecf was hosen to help filter out weak X-ray sources, which would not ave sufficient positional accuracy for astrometric alignment. We hen cross-matched X-ray sources identified by wavdetect to the DSS catalogue using wcs match . We found only two common -ray/optical sources for Mrk 1434, zero common sources for SDSS 1213, and one common source for SDSS J1221. Thus, we applied translational astrometric correction for Mrk 1434 (  x = 0.97, y = 1.32 pixels) and for SDSS J1221 (  x = 0.01,  y = 0.96 ixels) using wcs update . No astrometric correction was applied o SDSS J1213. We next re-ran wavdetect on the aligned event files (filtered rom 0.5 to 7 keV, now including each target dwarf galaxy) to etermine positions in the aligned reference frame of X-ray sources osted by each dwarf galaxy. We used the same wavdetect arameters as abov e , ex cept we used ecf = 0.3 when generating he point spread function map to allow the detection of fainter point ources. wavdetect identified two X-ray sources in Mrk 1434,NRAS 519, 5848–5858 (2023) ne source in SDSS J1213, and one source in SDSS J1221. The ositions of each X-ray source are listed in Table 2 . We estimated 5 per cent uncertainties of each X-ray position based on the distance rom the telescope aimpoint and the number of counts detected by avdetect , following equation (5) in Hong et al. ( 2005 ). Note that his 95 per cent positional uncertainty represents the statistical error n each source. For SDSS J1213 in particular, where we could not erform an astrometric alignment of the Chandra image, there is an dditional systematic uncertainty that could be as large as 2 arcsec although 0.8 arcsec is more typical).4 We then measured the number of counts from each X-ray source sing srcflux . We adopted circular apertures centred at each avdetect position with radii of 5 pixe ls, ex cept for Mrk 1434, hich contains two X-ray sources, where we adopted radii of 2.5 ixels to av o id the regions from each X-ray source from ov e rlapping. he number of background counts per pixel was estimated from earby source-free regions of each image. These measurements ere performed in both broad (0.5–7.0 keV) and hard (2.0–7.0 keV) mages, and we detected 19–73 counts from each source in the broad- and and 8–23 counts in the hard band. All X-ray detections (in all ands) are significant at the > 99 per cent level according to the onfidence tables in Kraft, Burrows & Nousek ( 1991 ). Finally, spectra were extracted for each X-ray source using pecextract and fit using an absorbed power-law model tbabs ∗powerlaw ) in the Interactive Spectral Interpretation Sys- em v1.6.2 ( ISIS ; Houck & Denicola 2000 ), adopting Cash statistics Cash 1979 ) given the relatively low number of counts per source. e initially left the column density as a free parameter. How ev er, for hree X-ray sources N H converged to zero, in which case we froze the alue to the Galactic column density and refit the spectrum. Model uxes were calculated using the cflux convolution model. Spectral arameters and model fluxes are reported in Table 3 . .3 Hubble Space Telescope e observed each galaxy with the Wide Field Camera 3 aboard ST for one orbit per galaxy (PI Plotkin; programme 14356). For rk 1434 and SDSS J1221 we observed in both the F 110 W and 606 W filters (with the IR and UVIS channels, respectively), and or SDSS J1213, which is a fainter galaxy, we took observations nly in the F 110 W filter. Observations in each filter were taken v e r four dither positions, and we used the IRSUB512 subarray or Mrk 1434 and SDSS J1221. Total exposure times in each filter re listed in Table 4 . Data were downloaded from the Mikulski rchive for Space Telescopes , and individual exposures were aligned nd combined using AstroDrizzle in the DRIZZLEPAC software Hack, Denchev a & Fruchter 2013).5 The F 110 W drizzled images ere created with plate scales 0.06 arcsec pix−1 for Mrk 1434 and DSS J1221, and 0.09 arcsec pix−1 for SDSS J1213. All F 606 W mages have plate scales 0.03 arcsec pix−1 . We aligned the HST astrometry to the Gaia Data Release 2 (Gaia ollaboration 2018 ) reference frame using the tweakreg task ithin AstroDrizzle (after excluding sources falling within each alaxy).6 For Mrk 1434, the corrections resulted in astrometric shifts X-ray sources in dwarf galaxies 5851 Table 2. Details of Chandra observations. Column 1: name of X-ray source. Column 2: Chandra obsID. Column 3: date of observation. Column 4: exposure time. Columns 5 and 6: right ascension and declination of each X-ray source. Column 7: radius of the 95 per cent positional uncertainty of each Chandra source, based on equation (5) of Hong et al. ( 2005 ). Column 8: aperture corrected net count rate (in counts ks−1 ) in the broad X-ray band (0.5–7.0 keV). Aperture corrections of 0.90, 0.95, and 0.96 were used for Mrk 1434, SDSS J1213, and SDSS J1221, respectively. Column 9: aperture corrected net count rate (in counts −1 ks ) in the hard band (2.0–7.0 keV). Aperture corrections of 0.87, 0.93, and 0.93 were used for Mrk 1434, SDSS J1213, and SDSS J1221, respectively. Source obsID Date Exp time Right ascension Declination p err Net rate (0.5–7.0 keV) Net rate (2.0–7.0 keV) (ks) (J2000) (J2000) (arcsec) (ks−1 −1 ) (ks ) (1) (2) (3) (4) (5) (6) (7) (8) (9) Mrk 1434 X-N 18059 2016 Jan 26 5.0 10:34:10.19 + 58:03:49.0 0.35 8.00 ± 2.22 2.98 ± 1.40 Mrk 1434 X-S 18059 2016 Jan 26 5.0 10:34:10.11 + 58:03:46.3 0.36 6.83 ± 2.04 2 .06+ 1 . 36 +−0 . 94 SDSS J1213 0 . 89 18060 2016 Aug 04 7.0 12:13:26.12 + 54:36:34.1 0.38 2.78 ± 1.10 1 . 15 −0 . 60 SDSS J1221 18061 2016 Feb 13 7.0 12:21:11.00 + 17:38:18.0 0.33 10.82 ± 2.11 3.46 ± 1.24 Table 3. Chandra spectral parameters, fluxes, and luminosities. Column 1: name of X-ray source. Column 2: column density. Column 3: best-fitting photon index. Column 4: best-fitting Cash statistic/degrees of freedom. Columns 5 and 6: logarithms of the unabsorbed model X-ray flux and luminosity from 0.5 to 10 keV, estimated using the cflux convolution model. Columns 7 and 8: logarithms of the unabsorbed model X-ray flux and luminosity from 2 to 10 keV, estimated using the cflux convolution model. Source N H C-stat/d.o.f Broad (0.5–10.0 keV) Hard (2.0–10.0 keV) log flux log luminosity log flux log luminosity (1020 cm−2 ) (erg −1 −2 −1 −1 −2 −1 s cm ) (erg s ) (erg s cm ) (erg s ) (1) (2) (3) (4) (5) (6) (7) (8) Mrk 1434 X-N <56.9a 1.3 ± 0.4 16.0/13 −12.8 ± 0.1 40.2 ± 0.1 −13.0 ± 0.2 40.1 ± 0.2 Mrk 1434 X-S 0.6b 1.7 ± 0.4 10.3/13 −13.1 ± 0.1 40.0 ± 0.1 −13.3 ± 0.2 39.8 ± 0.2 SDSS J1213 1.4b 1.3 ± 0.5 5.8/12 −13.3 ± 0.2 39.8 ± 0.2 −13.5 ± 0.2 39.6 ± 0.2 SDSS b J1221 2.7 1.6 ± 0.3 27.1/32 −12.8 ± 0.1 39.6 ± 0.1 −13.0 ± 0.1 39.5 ± 0.1 a Best-fitting column density N H = 8.0 × 1020 cm−2 , reported as an upper limit (95 per cent confidence level) because the uncertainty on the best-fitting value extends down to the Galactic value of 0.6 × 1020 cm−2 . b Column density frozen to the Galactic value during fitting, taken from Dickey & Lockman ( 1990 ). Table 4. Summary of HST observations. Column 1: galaxy name. Column 2 2: date of observations. Column 3: filters used for observations. Column 4: exposure times in the F 110 W / F 606 W filters, respectively, when both filters M were used. All observations were taken through HST Proposal ID 14356. 1 J Source Date Filter Exp. time o (min) S (1) (2) (3) (4) t t Mrk 1434 2016 Apr 16 F 110 W / F 606 W 8.6/30.9 SDSS J1213 2016 Apr 16 F 110 W 43.7 SDSS J1221 2016 Apr 9 F 110 W T / F 606 W 8.6/26.9 o a b p 1 d F t fi f S s t f h t e s 5 ( t p a i u Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023y  x = 1.8,  y = 0.0 pixels (from two common sources) and  x = .9,  y = 2.2 pixels (from nine common sources) in the F 110 W and 606 W filters, respective ly. Fo r SDSS J1213, we shifted the F 110 W lter by  x = 0.6,  y = 2.8 pixels (five common sources). Finally, for DSS J1221 we could not identify enough common sources between he HST image and the Gaia catalogue in the F 110 W filter (which as a smaller field of view). So, we only aligned the F 606 W filter to he Gaia frame, shifting by  x = 0.2,  y = 5.3 pixels (four common ources), and we then aligned the F 110 W filter to the F 606 W filter via three common sources between the two HST filters). t i 5 osition (0.3–0.4 arcsec), we do not expect a meaningful offset between f the bsolute astrometry of SDSS versus Gaia , such that systematic uncertainties 0 n our astrometric alignments are dominated by the small number of sources f sed to apply the corrections. .4 Very Large Array rk 1434 and SDSS J1221 both had archiv a l data sets (PI Satyapal, 4A-358) from the VLA, while new data were obtained for SDSS 1213 for this study (PI Plotkin, SH0563). All three galaxies were bserved in the most extended A configuration. Both Mrk 1434 and DSS J1221 observations were from 4.5 to 6.5 GHz ( C band) and 8 o 10 GHz ( X band), while SDSS J1213 was observed only from 8 o 12 GHz. The Common Astronomy Software Applications ( CASA ; CASA eam et al. 2022 ) software package version 5.1 was used to carry ut standard data reduction. We used 3C 286 to perform delay nd bandpass calibrations, and to set the flux density scale. Nearby hase calibrators (see Table 5 ) were observed to solve for the time- ependent complex gain solutions. Imaging was performed using the ask tclean , using two Taylor terms ( nterms = 2 ) to account or the wide fractional bandwidth and natural weighting to maximize ensitiv ity. We achiev ed root-mean-square (rms) sensitivities ranging rom 3.7 to −1 8.7 μJy bm in each observing band (see Table 5 ). The only X-ray source for which we found coincident radio mission is Mrk 1434 X-N, where we found radio detections at both .5 and 9.0 GHz within the X-ray error circle. We used imfit to fit wo-dimensional Gaussians in the image plane (at each frequency) o calculate the size of the radio structure, and to measure peak and ntegrated flux densities. As discussed further in Section 3.1 , the .5 GHz emission is slightly extended (with integrated flux density int = 0.191 ± 0.036 mJy) while the 9.0 GHz is point like ( f peak = .036 ± 0.009 mJy). The centroids of the radio emission at each requency are offset by 0.20 arcsec ± 0.07 arcsec. For the other twoMNRAS 519, 5848–5858 (2023) 5852 E. Thygesen et al. Table 5. Summary of VLA observations. Column 1: galaxy name. Column 2: VLA programme ID. Column 3: date of observation. Column 4: name of phase calibrator. Column 5: central frequency of each observation. Column 6: bandwidth of each observation. Column 7: the time spent integrating on each galaxy. Column 8: the size of the (elliptical) synthesized beam along the major and minor axes. Column 9: rms noise of each image. Source Programme Date Phase calibrator ν ν τ θbm σ rms (arcsec × (J2000) (GHz) (GHz) (min) arcsec) ( μJy bm−1 ) (1) (2) (3) (4) (5) (6) (7) (8) (9) Mrk 1434a 14A-358 2014 Feb 24 1035 + 564 5.5 2.0 8.5 0.45 × 0.38 8.7 Mrk 1434b 14A-358 2014 Feb 24 1035 + 564 9.0 2.0 8.5 0.27 × 0.24 8.6 SDSS J1213 SH0563 2016 Sep 30 1219 + 482 10.0 4.0 39.5 0.28 × 0.23 3.7 SDSS J1221 14A-358 2014 Feb 26 1158 + 248 5.5 2.0 25.8 0.42 × 0.38 5.9 SDSS J1221 14A-358 2014 Feb 26 1158 + 248 9.0 2.0 26.0 0.25 × 0.23 5.8 a Extended radio emission detected near Mrk 1434 X-N at 5.5 GHz, with f int = 0.191 ± 0.036 mJy and f peak = 0.054 ± 0.009 mJy bm−1 . The centroid of emission is located at RA = 10h 34m 10s . 1867 s ± 0 . 0042 and Dec. = 58◦03′ 49′.′ 1481 ± 0′′ . 0763. b Point-like radio emission detected near Mrk 1434 X-N at 9.0 GHz, −1 with f peak = 0.036 ± 0.009 mJy bm . The emission is located at RA = 10h 34m 10s . 2045 ± 0s . 0039 and Dec. = 58◦ 03′ ′ 49 .′ 2883 ′′ ± 0 . 0460. g e t 9 r t w f p a t 3 r I s f Z o ( f S 0 fl 3 c M s F i ( t h T 1 p ( N b a 1 t s C g i o z e b 0 r t d p fi a q T 1 e m q ( c T l X i c ( 3 l L S e p M Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023alaxies, we place 3 σ rms limits on their radio flux densities. We note hat SDSS J1221 displays radio emission aligned with a likely H II egion towards the eastern outskirts of the galaxy that is not associated ith X-ray emission, so we do not discuss that radio emission in this aper. RESULTS n the following subsections, we present the multiwavelength results or each galaxy, deferring discussions regarding the possible nature f each X-ray source to Section 4 . Composite HST images are shown or each galaxy in Fig. 1 , including the locations of X-ray sources. .1 Mrk 1434 rk 1434 hosts two X-ray sources separated by 2.8 arcsec (see ig. 1 a), both of which are classified as ULXs: the northern source Mrk 1434 X-N), which is located towards the galactic nucleus, as an unabsorbed hard X-ray luminosity L 2 –10 keV = ( 1 . 2 ± 0 . 6 ) × 040 erg s−1 , and the southern source (Mrk 1434 X-S) has L 2 –10 keV = 5 . 8 ± 0 .2) × 1039 erg s−1 . The X-ray spectra of each source are fit y power-law models with photon indices of = 1.3 ± 0.4 for Mrk 434 X-N and = 1.7 ± 0.4 for Mrk 1434 X-S. Neither source how s ev idence for significant intrinsic absorption. It is unlikely that either hard X-ray source is a superposed fore- round/background object. Given the density and flux distribution f hard X-ray sources in the cosmic X-ray background (see e.g. quation 2 of Moretti et al. 2003 ), we expect to only find 0.001 and .003 hard X-ray sources with 2–10 keV fluxes similar (or brighter) han Mrk 1434 X-N and Mrk 1434 X-S, respectively, within the rojected size of the galaxy (which we conserva tiv ely approximate s a circle with a 20 arcsec radius). Radio emission is detected only from the northern source, Mrk 434 X-N. At 5.5 GHz, the emission is extended with major and inor axis full width at half-maxima of 1.1 arcsec × 0.6 arcsec 160 pc × 90 pc), respective ly, cov e ring ≈3.5 synthesized beams. he centroid of the 5.5 GHz emission is 0.16 arcsec from the -ray position (for reference, the 95 per cent Chandra error cir- le is 0.35 arcsec), and the integrated luminosity is L 5.5 GHz,int = 1.2 ± 0.2) × 1036 erg s−1 . At 9.0 GHz, we detect a point source ocated 0.32 arcsec from the X-ray position, with a peak luminosity = (3.7 35 −1 9.0 GHz,peak ± 0.8) × 10 erg s . We do not detect any xtended radio structures at 9.0 GHz, thereby indicating that theNRAS 519, 5848–5858 (2023) mission seen at 5.5 GHz has a steep radio spectrum (our 5.5 and .0 GHz radio maps have similar sensitivities; see Table 5 ). Note hat extended emission is not simply resolved out at the higher radio requency, since the smallest baselines of the VLA in A configuration re sensitive to structures up to ≈5 arcsec at 9.0 GHz, which is larger han the ≈1 arcsec angular size of the 5.5 GHz emission. At 9.0 GHz, the chance of a random alignment of a background adio point source falling within the Chandra error circle is very mall. Integrating the differential source counts tabulated by de otti et al. ( 2010 ) at 8.4 GHz, and assuming a flat radio spectrum as expected if the 9.0 GHz emission is from a compact jet; see ection 4.1.2 −5 ), we expect only ≈3 × 10 sources with f peak > .036 mJy within the X-ray error circle. The chance of a statistical uctuation as large as 0.036 mJy (i.e. 4 σ rms) within the X-ray error ircle (which contains ≈240 pixels in the radio map) is also very mall ( p −5 = 3 × 10 ). Thus, we believe the 9.0 GHz emission s indeed physically associated with the galaxy. How ev er, we note hat the radio source lies towards the edge of the X-ray error circle. hus, even though the radio source formally falls within the Chandra ositional uncertainty, its association specifically with Mrk 1434 X- is less clear, particularly after considering that the Chandra X-ray strometry of Mrk 1434 was aligned to the optical frame using only wo common X-ray/SDSS sources. Finally, we note that towards the south-west of the 0.35 arcsec handra X-ray error circle of Mrk 1434 X-N, there is an optical/near- nfrared source that appears red in the HST composite image (see the oom-in of Fig. 1 a). If that source is a background quasar it may also e responsible for the X-ray and/or radio emission. How ev er, the andom alignment of such a background quasar is very unlikely, as escribed below. The AB magnitude of the HST source in the F 606 W lter is 18.8, which we convert to SDSS i ≈ 18.7 assuming a typical uasar spectrum (Vanden Berk et al. 2001 ). We then consider SDSS ype 1 quasar counts from 0.3 < z < 3.5 (Richards et al. 2006 ; Ross t al. 2013 ), and we find only a negligible number of background uasars ( ≈6 × 10−7 ) are likely to fall within the Chandra X-ray ircle by random chance (note that the random alignment of a radio- oud or a Type 2 quasar would be even rarer). That source is likely ntrinsic to the galaxy. .2 SDSS J1213 and SDSS J1221 DSS J1213 and SDSS J1221 each contain a single hard X-ray oint source near the outskirts of each galaxy (Fig. 1 b and c). The X-ray sources in dwarf galaxies 5853 Figure 1. (a) Composite HST image of Mrk 1434 in the F 606 W (blue/green) and the F 110 W (red) filters. The locations of the two X-ray point sources are shown as red cross hairs, with the dashed red circles illustrating the sizes of the 95 per cent positional errors from Chandra . The zoom-in of the centre of the galaxy shows the location of Mrk 1434 X-N relative to the radio emission, where yellow contours show the extended 5.5 GHz radio emission (1.1 arcsec × 0.6 arcsec; contours drawn at 3, 4, 5 × σ rms) and the magenta contours show the unresolved emission at 9.0 GHz (contours drawn at 3, 4 × σ rms). The sizes of the VLA synthesized beams are 0.45 arcsec × 0.38 arcsec (5.5 GHz) and 0.27 arcsec × 0.24 arcsec (9.0 GHz), respectively. Note that the SDSS spectroscopic fibre, from which the nebular He II emission is detected, has a diameter of 3 arcsec and is placed at the centre of the galaxy. (b) HST image of SDSS J1213 in the F 110 W filter, with the location of the X-ray source marked by the red cross hair and dashed circle. (c) HST composite image of SDSS J1221 in the F 606 W (blue/green) and the F 110 W (red) filters, with the location of the X-ray source marked by the red cross hair and dashed circle. In all images, north is up and east is to the left. h S ( s ( o e 4 a g 4 3 A J p a c h p o 4 s I m f F f L p f w e p i S 0 r u Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023ard (2–10 keV) X-ray luminosities of the sources are L 2 –10 keV = 4 .3 ± 2.4) × 1039 and (2.9 ± 0.8) × 1039 erg −1 s , respectively Table 2 ), such that both sources are classified as ULXs. The chance f a superposed foreground/background object is negligible (we xpect only 0.005 hard X-ray background sources for SDSS J1213 nd 0.001 sources for SDSS J1221; Moretti et al. 2003 ). Neither alaxy contains radio emission within the Chandra X-ray circles to σ rms upper limits of < 1.4 × 1035 erg s−1 at 10.0 GHz for SDSS 1213, and to limits of <3.0 × 1034 and <4.9 × 1034 erg s−1 at 5.5 nd 9.0 GHz, respectively, for SDSS J1221. DISCUSSION n the following subsections, we discuss possible interpretations or the X-ray sources in our sample of three dwarf galaxies. We ocus primarily on Mrk 1434 since it exhibits the most complex henomenology (i.e. two X-ray sources, one of which is coincident ith radio emission). We provide arguments for/against XRB inter- retations in Section 4.1.1 and for/against AGN interpretations in ection 4.1.2 . In Section 4.1.3 , we discuss whether the observed X- ay flux is sufficient to explain He II line emission observed in theDSS spectrum of Mrk 1434. A discussion on the nature of the X-ray ources in the other two galaxies is presented in Section 4.2 . .1 Mrk 1434 .1.1 XRB interpretations s shown in Section 3.1 , both X-ray sources in Mrk 1434 are hysically associated with the galaxy and luminous enough to be lassified as ULXs. The observed X-ray luminosity, how ev er, is igher than expected from the luminous tail of the galaxy’s XRB opulation. The luminosities of both X-ray sources are abov e the cut- ff of the low-mass XRB luminosity function (e.g. Gilfanov 2004 ), o in the following we only consider high-mass XRBs using the etallicity-dependent luminosity function from Lehmer et al. ( 2021 ). or Mrk 1434, with 12 + log(O/H) = 7.8 and SFR = 0.12 M  yr−1 , ehmer et al. ( 2021 ) predict a total 0.5–8.0 keV X-ray luminosity (i.e. rom all X-ray point sources) of L 0.5–8.0 keV = ( 1 . 7 ± 0 39 . 15 ) × 10 rg s−1 (where the error bar represents the 68 per cent confidence nterval provided by Lehmer et al. 2021 ). They also predict only .03+ 0 . 04 40 −1 −0.02 ULXs with L 0.5−8.0 keV > 10 erg s . For reference, the nabsorbed 0.5–8.0 keV model luminosities of Mrk 1434 X-N and MNRAS 519, 5848–5858 (2023) 5854 E. Thygesen et al. M Table 6. mBH mass estimates and limits. Column 1: galaxy name. Column e 2: logarithm of the hard X-ray luminosity. Column 3: logarithm of the radio U luminosity at 5 GHz, assuming a flat radio spectrum. For Mrk 1434 X-N, e this luminosity is based on the unresolved emission detected at 9 GHz. For all other X-ray sources, limits are placed as 3 σ rms. Column 4: logarithm of c the black hole mass (or limit) if X-ray sources are weakly accreting mBHs, based on the Fundamental Plane of black hole activity (G ̈ultekin et al. 2019 ). b Uncertainties on log M BH are ≈1 dex. t I Source log L 2 –10 keV log L 5 GHz log M BH X (erg s−1 ) (erg −1 s ) (M ) r (1) (2) (3) (4) r Mrk 1434 X-N 40.1 ± 0.4 35.3 ± 0.1 5.6 i Mrk 1434 X-S 39.8 ± 0.3 < 35.2 < 5.6 2 SDSS J1213 39.6 ± 0.4 < 34.9 < 5.3 r SDSS J1221 39.5 ± 0.2 < 34.4 < 5.0 ( U e M 9 w t f ( b r l s 0 I b w p w N e r A P b l f t e e α f b a ( S c r e X d c l t r a d 4 I M 4 t A T v c o 1 X r a ( d i A i X E w p w 5 r p t o c w X t t M Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023rk 1434 X-S are (1.3 ± 0.4) × 1040 and (0.8 ± 0.2) × 1040 rg s−1 , respectively. Thus, the combined X-ray luminosity of both LXs is ≈10 times higher than expected relative to the Lehmer t al. ( 2021 ) luminosity function, which is significant even after onsidering uncertainties and intrinsic scatter. Even though the above suggests that it is statistically unlikely for oth sources to be XRBs, small number statistics could influence he abov e arguments, and it is worth ex ploring XRB interpretations. n particular, the extended 5.5 GHz radio emission from Mrk 1434 -N could represent a ‘ULX bubble’, as similar types of extended adio structures have been observed from other ULXs, making the adio emission a signature of a ULX outflow shocking the nearby nterstellar environment (e.g. Pakull, Soria & Motch 2010 ; Soria et al. 010 , 2021 ; Cseh et al. 2012 ; Urquhart et al. 2019 ). If the 5.5 GHz adio emission is indeed a ULX bubble, then with L 5.5 GHz,int = 1.2 ± 0.2) 36 −1 × 10 erg s it would represent the most luminous LX bubble yet observed by a factor of ≈6 (Pakull et al. 2010 ; Soria t al. 2010 , 2021 ). Meanwhile, the projected size of ≈160 pc × 0 pc (1.1 arcsec × 0.6 arcsec) in diameter is fairly typical compared o other ULX bubbles, where diameters range from ≈25 to 350 pc Soria et al. 2021 ; also see table 1 of Berghea et al. 2020 , and eferences therein). Taking the peak flux density of the 5.5 GHz tructure, and extrapolating to 1 GHz assuming a spectral index α = .7, the intensity of the radio bubble in Mrk 1434 X-N would be ≈ 6 × 10−16 erg s−1 cm−2 1 GHz Hz−1 sr−1 , which is relatively large ut reasonable compared to other ULX radio bubbles with similar hysical sizes (see fig. 5 of Berghea et al. 2020 ). Although a ULX bubble is one interpretation of the 5.5 GHz mission, we stress that it is not a unique (or necessary) explanation. dopting SFR = 0.12 M  yr−1 for Mrk 1434 and the relation etween star formation rate and the 1.4 GHz specific luminosity rom Kennicutt 36 & Evans ( 2012 ), we expect L 5.5 GHz,SF ≈ 3.9 × 10 rg s−1 (we convert from 1.4 to 5.5 GHz assuming a spectral index = 0.7). Considering that the intrinsic scatter on the conversion etween SFR and radio luminosity is of the order of ±0.3 dex Murphy et al. 2011 ), the observe d ex tended structure at 5.5 GHz ould be produced entirely by star formation processes. Since the xtended radio structure at 5.5 GHz is not detected at 9.0 GHz, the ominant radio emission mechanism in such a scenario would most ikely be synchrotron radiation with a steep spectrum from supernova emnants. Note that our data exclude free–free radio emission from n H II region, which would produce a flat spectrum that would be etectable at 9.0 GHz. .1.2 AGN interpretations GN can also produce extended radio emission, which is another iable explanation for the 5.5 GHz radio structure. How ev er, in light f the discussion in the previous subsection that a super-Eddington RB is also capable of producing the observed extended emission t 5.5 GHz, the resolved radio complex does not provide useful iagnostics for attempting to discriminate between XRB versus GN. Since the X-ray spectra of Mrk 1434 X-N and Mrk 1434 -S ( = 1.3 ± 0.4 and = 1.7 ± 0.4, respectively) are consistent ith low-luminosity AGNs (Younes et al. 2011 ; Yang et al. 2015 ), e focus the following discussion on AGN scenarios with Eddington atios L bol/ L Edd  0.01. For such weakly accreting AGNs, we expect o observe unresolve d radio emission from a partially self-absorbed ompact jet (Ho 2008 ). By combining X-ray and radio luminosities, e can then make crude estimates on black hole masses by appealing o the Fundamental Plane of black hole activity (Merloni, Heinz & diNRAS 519, 5848–5858 (2023) atteo 2003 ; Falcke, K ̈ording & Markoff 2004 ). For Mrk 1434 X-N, e then interpret the unresolved 9.0 GHz radio emission as arising rom a compact jet, and we utilize the Fundamental Plane regression y G ̈ultekin et al. ( 2019 ), og(M /108 BH M ) = ( 0 . 55 ± 0 . 22 ) + (1.09 ± 0.10) log(L /1038 erg s−1 5 GHz ) − ( 0 . 59 ± 0. 40 −1 16 ) log ( L 2 / 10 erg s ) , (1) –10 keV hich has an intrinsic scatter ≈1 dex. We estimate that Mrk 1434 X-N ould have M BH ≈ 4 × 105 M if powered by an mBH (see Table 6 ).  ote that we assume a flat radio spectrum to convert the observed adio luminosity at 9.0 to 5.0 GHz for use in the Fundamental lane (we cannot use our 5.5 GHz radio map to estimate the 5 GHz uminosity because we do not have enough signal to noise to attempt o decompose a point source embedded within the extended radio mission observed at 5.5 GHz). Similarly, the lack of radio emission rom Mrk 1434 X-S implies M 5 BH  4 × 10 M  (where we adopt 3 σ rms upper limit, based on the observed σ rms near Mrk 1434 X- in our 5.5 GHz image). These mass estimates imply Eddington atios ( L 2–10 keV/L Edd) of ≈2 × 10−4 and 1 × 10−4 for Mrk 1434 -N and Mrk 1434 X-S, respectively, which, assuming bolometric orrections of ≈10, are consistent with Eddington ratios for which he Fundamental Plane can be applied (see e.g. Plotkin et al. 2012 ). .1.3 On the origin of nebular He II emission n the following, we determine whether the X-ray emission from rk 1434 is a strong enough source of photoionization to explain he strength of the He II emission in the SDSS spectrum of Mrk 1434. he observed He II line flux is F λ4686,obs = (7.5 ± 0.1) × 10−16 erg s−1 m−2 , which translates to a photon flux of N λ4686,obs = (1.8 ± 0.1) × 0−4 photons s−1 cm−2 . Every photon emitted in the He II line equires 5.2 ionizing photons incident on singly ionized helium Pakull & Angebault 1986 ). Given the ionization potential of singly onized helium ( χ ion = 54.4 eV), and considering that the photoion- zation cross-section −3 has a steep E ph dependence on photon energy, ph, then producing the observed SDSS He II line flux requires a hoton flux in the extreme ultraviolet (54–300 eV) of N 54 –300 eV = .2N −4 −1 −2 λ4686,obs = ( 9 . 1 ± 0 . 1 ) × 10 photons s cm . Note that this hoton flux is an underestimate because we have not corrected the bserved SDSS line flux for extinction. The 3 arcsec SDSS spectroscopic fibre is centred near Mrk 1434 -N, such that if the He II emission arises from photoionization by he X-ray source, we expect the emission to be dominated by Mrk X-ray sources in dwarf galaxies 5855 1 o u 1 r m w ( e m s h T o o b H w a b h p s fl e f f 2 2 f N s ( ( 2 b s X s 1 m e l w f T X N w i o t N f o b f t t T X t s u p S s e w W c t h o a s f I M a b o e t p p B 1 s r B m P w s t m a 7 m i t Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 20434 X-N. We do not have direct measurements on the extreme ltraviolet flux from 54 to 300 eV, so we extrapolate the Chandra X- ay spectrum into the extreme ultraviolet. Our best-fitting power-law odel predicts a photon flux of 0.3+ 2 . 5 −0.2 × 10−4 photons −1 s cm−2 note the large range in uncertainty because we are extrapolating the odel to energies lower than the Chandra X-ray band). Thus, while igh-energy radiation from Mrk 1434 X-N may contribute to some f the He II photoionization, the observed X-ray source is too faint, y a factor of ≈30, to supply all of the photoionizing photons. If e assume a thermal X-ray emission model ( tbabs ∗diskbb ), it ecomes even more difficult for the X-ray source to explain the He II hotoionization, as the extrapolated 54–300 eV extreme ultraviolet ux becomes ≈90 times too faint. Adding a contribution of photons rom Mrk 1434 X-S would only increase the abov e photon flux by a actor of ≈2, for either spectral model. There is currently no evidence for significant X-ray variability rom Mrk 1434 ov e r the past 1–2 decades. Coincidentally, the SDSS pectrum and the archiv a l Chandra observa tion from Lemons et al. 2015 , Chandra obsID 3347) were both taken in May 2002 (separated y ≈2 weeks). The archiv a l data from 2002 show nearly identical -ray luminosities ( log L −1 2 –10 keV = 40 . 1 and 39.9 erg s for Mrk 434 X-N and Mrk 1434 X-S, respectively; see table 2 of Lemons t al. 2015 ) compared to the Chandra observations presented here, hich were taken nearly 14 yr later (see Table 2 of this paper). here are also two X-ray detections of Mrk 1434 in the third XMM– ewton serendipitous source catalogue (3 XMM ; Rosen et al. 2016 ) n 2007 and 2008. Both X-ray sources are blended together due o XMM–Newton ’s poorer spatial resolution. Comparing the XMM– ew ton fluxe s to the combined fluxe s of both sources in the Chandra bserva tions, X-ray va riability is smaller than a factor of ≈2 ov e r the our observa tions. How ev e r, considering the light trave ltime between he X-ray source and the ionized medium, it is feasible that Mrk 1434 -N was more active in the past. The projected radius of the SDSS pectroscopic fibre is 730 light-years, and we cannot exclude the ossibility that Mrk 1434 X-N was ≈30–90 times more luminous everal hundred years ago, which appears to be on the only viable ay for the He II emission to be powered by X-ray photoionization. If the extended radio emission is produced by an outflow shocking he interstellar medium, then one must also consider the possibility f the He II emission being produced by ionization from a radiative hock (e.g. Dopita & Sutherland 1995 ). According to the MAPPINGS II libraries of line ratios for radiative shocks (Allen et al. 2008 ), ssuming a shock velocity of 300 km s−1 , we expect the luminosity f the He II λ4686 emission line L −4 λ4686 ≈ 4 × 10 L rad, where L rad is he total radiative luminosity of the shock.7 Assuming that the kinetic ower required to inflate a bubble P kin ≈ 77/27 L rad (Weaver et al. 977 ), then explaining the observed He II line via shock ionization equires an outflow with P kin ≈ 6 × 1041 erg −1 s . We do not have a reliable method to independently estimate kin (especially considering that other emission lines in the SDSS pectrum are dominated by star formation). How ev er, for an order of agnitude estimate, we calculate the minimum synchrotron energy t a A S c Given the low metallicity of Mrk 1434, we adopt the MAPPINGS III odel grid with Small Magellanic Cloud abundances. We 8 also assume an nterstellar medium density of 1 cm−3 and equipartition of magnetic and t hermal pressures. v 23f the 5.5 GHz radio emission, which is W 52 min ≈ 2 × 10 erg (Longair 994 ).8 A 300 −1 5 km s shock wo uld take ≈3 × 10 yr to inflate a bubble ith a 160 pc diameter, such that the average power stored in internal nergies of the synchrotron emitting structure is P̄ 39 min ≈ 2 × 10 erg −1 (i.e. the average power in particles and in the magnetic field). hus, an outflow would need to carry 102 times more power in rder for a shock to be the sole ionization source of the observed e II emission line. Of course, P̄ min is a minimum energy estimate, nd the power in bu bbles/cav ities carved out by kinetic outflows as sometimes been observed to be larger, sometimes by factors of everal hundreds (e.g. Ito et al. 2008 ), such that the abov e does not xclude the possibility of shock ionization. For comparison, the ULX NGC 6946 MF16 (Roberts & Colbert 003 ) has a luminous and compact radio bubble (Berghea et al. 020 ), which suggests a relatively powerful outflow. Adopting the GC 6946 MF16 bubble line flux in the [Fe II ] λ16440 emission line 4.2 × 10−15 −1 −2 erg s cm ) and a distance of 7.8 Mpc (Long et al. 020 ), the MAPPINGS III libraries for a 300 km s−1 shock (with olar abundances) suggest a kinetic power of P 40 kin ≈ 7 × 10 erg −1 . Thus, the kinetic power of NGC 6946 MF16 (i.e. one of the ost powerful known ULX radio bubbles) is an order of magnitude ower than the power required for shock ionization to be responsible or the observed strength of the He II emission line near Mrk 1434 -N. Thus, if the He II line is powered by shock ionization, then it ould represent one of the most powerful bubbles carved by a ULX utflow yet observed. Intriguingly, Mrk 1434 is one member of a population of 182 star- orming galaxies with nebular He II emission that were identified y Shirazi & Brinchmann ( 2012 ). The ratios of He II /H β relative o [N II ] λ6584/H α are inconsistent with AGN for these galaxies. ypically, when an AGN is absent, Wolf–Rayet stars are considered he primary stellar population capable of producing enough extreme ltraviolet flux abov e the 54 eV He II ionization edge. How ev er, hirazi & Brinchmann ( 2012 ) inspected the SDSS spectra for broad mission features indicative of Wolf–Rayet stars, and they found no olf–Rayet signatures in the spectrum of Mrk 1434. Thus, without oncrete evidence that Mrk 1434 X-N was indeed brighter several undred years ago to power the He II emission via photoionization, nd/or lacking a reliable estimate of the kinetic power of an outflow or shock ionization, the source of extreme ultraviolet photons in rk 1434 remains a mystery. Another plausible explanation could e photoionization from extreme ultraviolet photons emitted by xotic stellar populations (like rapidly rotating stars) in metal- oor environments (see the discussion in section 6 of Shirazi & rinchmann 2012 ). It is very plausible that several of the above cenarios contribute towards producing the He II line, and Shirazi & rinchmann ( 2012 ) recov e red a heterogeneous population (multiple echanisms may even contribute to producing the He II emission ithin a single galaxy). For example, Senchyna et al. ( 2020 ) conclude hat X-ray photoionization cannot explain nebular He II emission cross a sample of nearly a dozen metal-poor galaxies. Meanwhile, here are several well-established examples of X-ray sources that re indeed sufficient to power nebular He II emission (e.g. Pakull & ngebault 1986 ; Moon et al. 2011 ; Schaerer, Fragos & Izotov 2019 ; immonds, Schaerer & Verhamme 2021 ). Further observational onstraints, ideally via systematic X-ray surve y s of metal-poor dwarf We adopt L 36 −1 5.5 ≈ 10 erg s , a bubble diameter of ≈160 pc, and an ion o electron energy ratio of η = 40. We note that W ∝ 4/7 min η , and the proper alue of η is not well constrained. MNRAS 519, 5848–5858 (2023) 5856 E. Thygesen et al. g L t f i t s s e 4 4 A O t e ( e d b O s a w t i w L h i t m r 1 w t r L o S e h ( T t 5 w r A C ( g S X e i n 3 m s a l a s A s s p a ( w C ≈ t o s l e 5 V c W r g p a J w a t o 1 ( p o t w 1 c t 9 s m w m ≈ m M Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023alaxies under high spatial resolution, are required to understand he level to which ULXs contribute extreme ultraviolet radiation n metal-poor galaxies, which has implications for understanding ources of ionization and heating of the intergalactic medium in the arly Universe. .2 SDSS J1213 and SDSS J1221 ur new Chandra observations confirm the conclusion of Lemons t al. ( 2015 ) that both X-ray sources are more luminous than xpected from the XRB populations in each galaxy, as described elow. Unlike for Mrk 1434, the luminosities of both X-ray ources in SDSS J1213 and SDSS J1221 are low enough that e should consider both high-mass and low-mass XRBs. Follow- ng Lemons et al. ( 2015 ), we therefore adopt the relation from ehmer et al. ( 2010 ), which predicts the hard X-ray luminos- ty from low-mass and high-(mass XRBs as a) function of stellar ass and star formation rate: LXRB −1 2 –10 keV(/ erg s = ( 9 .) 05 ± 0 . 37 ) × 028 ( M / M ) + ( 1 . 62 ± 0 . 22 ) × 1039 SFR/ M  yr−1 , with an in- rinsic scatter of ±0.34 dex. The Lehmer et al. ( 2010 ) relation predicts XRB 37 37 −1 2 –10 keV = 1 . 2 × 10 and 5.6 × 10 erg s for SDSS J1213 and DSS J1221, respectively. The predicted luminosities are ≈3 times igher if we instead adopt the calibrations in Lehmer et al. ( 2019 ). hus, the observed X-ray luminosities are ≈120–360 and ≈17– 0 times higher than expected, for SDSS J1213 and SDSS J1221, espectively.9 In light of recent theoretical motiv a tion for ‘wandering’ mBHs Bellovary et al. 2019 , 2021 ; also see e.g. Mezcua & Dom ́ınguez ́anchez 2020 ; Reines et al. 2020 ; Greene et al. 2021 ; Sargent t al. 2022 , for observational searches), an X-ray source being ‘off- ucleus’ does not on its own preclude the possibility of an accreting BH. It is possible that these sources are mBHs launching jets that re either (a) beneath our radio detection limit or (b) very extended nd ‘resolved out’ by the VLA when it is in its most extended configuration. The largest angular scale to which the VLA is ensitive to radio emission at our observing frequencies ( X -band) nd configuration (A) is 5.3 arcsec, such that our VLA observations ould not detect flux from extended jets larger than ≈850 and 410 pc for SDSS J1213 and SDSS J1221, respectively. On the ther hand, the radio cores of weakly accreting AGNs (bolometric uminosities L bol < 0.01 L Edd) have flat radio spectra and are compact nough that their radio emission should not be ‘resolved out’ at LA resolutions (see e.g. Orienti & Prieto 2010 ). Thus, if only onsidering mBHs in the weak accretion regime, we can use our adio upper limits in conjunction with the Fundamental Plane to lace mass limits of 5 5 M BH < 2 × 10 and < 1 × 10 M  for SDSS 1213 and SDSS J1221, respectively. Requiring L bol < 0.01 L Edd, and ssuming X-ray bolometric corrections of 10, then places lower limits n black hole masses of  3 × 104 (SDSS J1213) and 4  2 × 10 M  SDSS J1221). Thus, there is a relatively narrow range of mass where ur VLA observations could ‘miss’ the compact radio jet from a eakly accreting mBH. Note that our radio limits do not place useful onstraints on the possibility of a more rapidly accreting mBH with e m The Lehmer et al. ( c 2010 ) relation is calibrated to galaxies with approximately olar metallicities. The metallicity of SDSS J1213 is unknown, and the i etallicity of SDSS J1221 is log(O/H) + 12 = 8.3 (Zhao et al. 2013 ). If e adopt the metallicity-dependent Lehmer et al. ( 2021 ) relation for high- ass XRBs, the X-ray luminosity of the X-ray source 1 in SDSS J1221 is still 20 times higher than expected for a galaxy with its star formation rate and a etallicity. q NRAS 519, 5848–5858 (2023) bol > 0.01 L Edd, which would correspond to 4 a mass M BH  10 M  or both sources. Nev ertheless, ev en though our data do not exclude he possibility of mBHs, Occam’s razor probably suggests that the implest and most likely scenario is that these are luminous XRBs. .3 An update to Lemons et al. ( 2015 ) fter considering the abov e multiwave length observations, all 10 of he dwarf galaxy AGN candidates identified by Lemons et al. ( 2015 ) via hard X-ray emission) now have sufficient spatial resolution to etermine whether the X-ray sources indeed reside in galactic nuclei. ur study reduces their number of AGN candidates to 7–8 (adopting n AGN definition that requires nuclear sources). It is very unlikely hat any of these seven to eight nuclear sources are chance alignments ith foreground/background X-ray emitting objects. Adopting the ard (2–10 keV) X-ray fluxes and X-ray position error circles of he nuclear candidates from table 2 of Lemons et al. ( 2015 ), and eplacing the X-ray flux and positional uncertainty of Mrk 1434 X-N ith the values presented here, the Moretti et al. ( 2003 ) cosmic X- ay background predicts only 0.003 sources to fall within the nuclei f the eight possible nuclear mBH candidates. Obtaining seven to ight viable AGN candidates is a significant result, considering that a) the Lemons et al. ( 2015 ) dwarf galaxy surve y was archiv a l and herefore serendipitous in nature, and (b) the three dwarf galaxies ith follow-up presented here represent three of their most unlikely GN candidates (given the poor spatial resolution of their archiv a l handra data). Lemons et al. ( 2015 ) found X-ray sources in 19 alaxies in total (i.e. the remaining 11–12 galaxies host off-nuclear -ray sources, most likely XRBs). Thus, if a luminous X-ray source s detected in a dwarf galaxy, our study (very roughly) implies a 0–40 per cent chance10 that it could be a nuclear mBH, which upports the viability of using X-ray surve y s to identify mBHs in ow-mass galaxies, as long as the surve y is performed with sufficient ensitivity and spatial resolution. We stress the importance of high patial resolution X-ray observations. For example, Mrk 1434 was reviously identified as an AGN from an XMM –Newton survey Birchall, Watson & Aird 2020 ), while our higher spatial resolution handra observation clearly resolves the ‘nuclear’ X-ray source into wo distinct sources (and even then, it remains unclear whether either ource is indeed an accreting mBH). SUMMARY A N D C O N C L U S I O N S e have presented a multiwavelength study of three nearby dwarf alaxies that host ULXs. Two galaxies in our sample, SDSS J1213 nd SDSS J1221, each contain single off-nuclear X-ray sources that e suspect are luminous XRBs. The third galaxy, Mrk 1434, hosts wo X-ray sources separated by 2.8 arcsec. The northern source (Mrk 434 X-N) also displays extended radio emission at 5.5 GHz and oint-like radio emission at 9.0 GHz. It remains unclear whether he X-ray sources in Mrk 1434 are XRBs or AGNs (especially Mrk 434 X-N), although either scenario is intriguing. If they are XRBs, hen the combined X-ray luminosity of both sources is larger than xpected for a galaxy with Mrk 1434’s star formation rate and (low) etallicity. Furthermore, the extended radio emission at 5.5 GHz ould then represent the most luminous ‘ULX bubble’ ever observed n the radio, although we stress that the 5.5 GHz radio emission0 This number is an upper limit, and it neglects biases inherent to an rchiv a l/serendipitous surve y , which is out of the scope of this paper to uantify. X-ray sources in dwarf galaxies 5857 c B o t B n B s a B B o B 1 C a C a C i C m C C A D W d t D A D N D o b E t F t F F T F U F N H p G I G R G R F f G F G b G i G C G t c H C H D H T A H r H T H - A H R I I A K K A A K B K B K B L 1 Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana State University Library user on 23 March 2023an also be attributed entirely to star formation within the galaxy, r to an AGN jet. Regardless of the correct scenario, we find that he line emission from He II in Mrk 1434 is inconsistent with a ebula being powered by the central X-ray source, unless the central ource underwent a period of higher activity several hundred years go, or if the nebula is shock ionized by an outflow that is an order f magnitude more powerful than yet observed from a ULX. If Mrk 434 X-N is an AGN, then the 9.0 GHz radio emission may represent compact synchrotron jet from a low-luminosity AGN power by n mBH with M ≈ 4 × 105 BH M . We conclude by stressing the mportance of high spatial resolution observations when performing ultiwavelength searches for mBHs in dwarf galaxies. C K N OW L E D G E M E N T S e thank the anonymous referee for helpful comments that improved his manuscript. Support for this work was provided by the National eronautics and Space Administration through Chandra Award umber GO6-17079X issued by the Chandra X-ray Center, which is perated by the Smithsonian Astrophysical Observatory for and on ehalf of the National Aeronautics Space Administration under con- ract NAS8-03060. This research is based on observations made with he NASA/ESA Hubble Space Telescope obtained from the Space elescope Science Institute, which is operated by the Association of niversities for Research in Astronomy, Inc., under NASA contract AS5-26555. These observations are associated with programme ST-GO-14356. Support for programme no. HST-GO-14356 was rovided by NASA through a grant from the Space Telescope Science nstitute, which is operated by the Association of Universities for esearch in Astronomy, Inc., under NA SA contract NA S5-26555. MP and JDP acknowledge support from the National Science oundation under grant number 2206123. RS acknowledges support rom grant number 12073029 from the National Natural Science oundation of China (NSFC). AER acknowledges support provided y NASA through EPSCoR grant number 80NSSC20M0231. GEA s the recipient of an Australian Research Council Discov e ry Early areer Researcher Award (project number DE180100346) funded by he Australian Gov e rnment. This research made use of ASTROPY,11 a ommunity-developed core PYTHON package for astronomy (Astropy ollaboration 2013 , 2018 ). ATA AVAILABILITY he data underlying this article are available in the Chandra Data rchive under ObsIDs 18059, 18060, and 18061 ( https://cda.ha var d.edu/chaser /), in the Barbara A. Mikulski Archive for Space elescopes under programme ID 14356 ( dx.doi.org/10.17909/3bxp zt07 ), and in the National Radio Astronomy Observatory Data rchive under programmes 14-358 and SH0563 ( data.nrao.edu ). EFERENCES llen M. G., Grov e s B. A., Dopita M. A., Sutherland R. S., Ke w ley L. J., 2008, ApJS , 178, 20 stropy Collaboration, 2013, A&A , 558, A33 stropy Collaboration, 2018, AJ , 156, 123 aldassare V. F., Reines A. E., Gallo E., Greene J. E., 2015, ApJ , 809, L14 aldassare V. F., Geha M., Greene J., 2020, ApJ , 896, 10 arth A. J., Ho L. C., Rutledge R. E., Sargent W. L. W., 2004, ApJ , 607, 90 L 1 http://www. astropy. org L ellovary J. M., Cleary C. E., Munshi F., Tremmel M., Christensen C. R., Brooks A., Quinn T. R., 2019, MNRAS , 482, 2913 ellovary J. M. et al., 2021, MNRAS , 505, 5129 erghea C. T., Johnson M. C., Secrest N. J., Dudik R. P., Hennessy G. S., El-khatib A., 2020, ApJ , 896, 117 irchall K. L., Watson M. G., Aird J., 2020, MNRAS , 492, 2268 landford R. D., K ̈onigl A., 1979, ApJ , 232, 34 rorby M., Kaaret P., Prestwich A., 2014, MNRAS , 441, 2346 ann J. M. et al., 2021, ApJ , 912, L2 ASA Team et al., 2022, PASP, 134, 114501 ash W., 1979, ApJ , 228, 939 hen C.-T. J. et al., 2017, ApJ , 837, 48 hilingarian I. V., Katkov I. Y., Zolotukhin I. Y., Grishin K. A., Beletsky Y., Boutsia K., Osip D. J., 2018, ApJ , 863, 1 seh D. et al., 2012, ApJ , 749, 17 ashyan G., Silk J., Mamon G. A., Dubois Y., Hartwig T., 2018, MNRAS , 473, 5698 e Zotti G., Massardi M., Negrello M., Wall J., 2010, A&AR, 18, 1 ickey J. M., Lockman F. J., 1990, ARA&A , 28, 215 opita M. A., Sutherland R. S., 1995, ApJ , 455, 468 ouna V. M., Pellizza L. J., Mirabel I. F., Pedrosa S. E., 2015, A&A , 579, A44 vans I. N. et al., 2010, ApJS , 189, 37 alcke H., K ̈ording E., Markoff S., 2004, A&A , 414, 895 ender R. P., Belloni T. M., Gallo E., 2004, MNRAS , 355, 1105 eng H., Soria R., 2011, New Astron. Rev. , 55, 166 ilippenko A. V., Ho L. C., 2003, ApJ , 588, L13 ruscione A. et al., 2006, in Silva D. R., Doxsey R. E., eds, Proc. SPIE Conf. Ser. Vol. 6270, Observatory Operations: Strategies, Processes, and Systems. SPIE, Bellingham, p. 62701V aia Collaboration, 2018, A&A , 616, A1 allo E., Sesana A., 2019, ApJ , 883, L18 armire G. P., Bautz M. W., Ford P. G., Nousek J. A., Ricker G. R., Jr, 2003, in Truemper J. E., Tananbaum H. D., eds, Proc. SPIE Conf. Ser. Vol. 4851, X- Ray and Gamma-Ray Telescopes and Instruments for Astronomy. SPIE, Bellingham, p. 28 ilfanov M., 2004, MNRAS , 349, 146 reene J. E., Strader J., Ho L. C., 2020, ARA&A , 58, 257 reene J. E. et al., 2021, ApJ , 917, 17 rimm H.-J., Gilfanov M., Sunyaev R., 2003, MNRAS , 339, 793 ̈ultekin K., King A. L., Cackett E. M., Nyland K., Miller J. M., Di Matteo T., Markoff S., Rupen M. P., 2019, ApJ , 871, 80 ack W. J., Denchev a N., Fruchter A. S., 2013, in Friedel D. N., ed., ASP Conf. Ser. Vol. 475, Astronomical Data Analysis Software and Systems XXII. Astron. Soc. Pac, San Francisco, p. 49 ao C.-N., Kennicutt R. C., Johnson B. D., Calzetti D., Dale D. A., Moustakas J., 2011, ApJ , 741, 124 eger A., Fryer C. L., Woosley S. E., Langer N., Hartmann D. H., 2003, ApJ , 591, 288 o L. C., 2008, ARA&A , 46, 475 o L. C., Kim M., 2016, ApJ , 821, 48 ong J., van den Berg M., Schlegel E. M., Grindlay J. E., Koenig X., Laycock S., Zhao P., 2005, ApJ , 635, 907 ouck J. C., Denicola L. A., 2000, in Manset N., Veillet C., Crabtree D., eds, ASP Conf. Ser. Vol. 216, Astronomical Data Analysis Software and Systems IX. Astron. Soc. Pac., San Francisco, p. 591 nayoshi K., Visbal E., Haiman Z., 2020, ARA&A , 58, 27 to H., Kino M., Kawakatu N., Isobe N., Yamada S., 2008, ApJ , 685, 828 aaret P., Feng H., Roberts T. P., 2017, ARA&A , 55, 303 ashibadze O. G., Karachentsev I. D., Karachentsev a V. E., 2020, A&A , 635, A135 ennicutt R. C., Evans N. J., 2012, ARA&A , 50, 531 ormendy J., Ho L. C., 2013, ARA&A , 51, 511 raft R. P., Burrows D. N., Nousek J. A., 1991, ApJ , 374, 344 ehmer B. D., Alexander D. M., Bauer F. E., Brandt W. N., Goulding A. D., Jenkins L. P., Ptak A., Roberts T. P., 2010, ApJ , 724, 559 ehmer B. D. et al., 2019, ApJS , 243, 3 ehmer B. D. et al., 2021, ApJ , 907, 17 MNRAS 519, 5848–5858 (2023) 5858 E. Thygesen et al. L R L R R L R L R R R M R S M S M M S M S S M S M S M S S M S M T M U M V V M V M W N W O P Y P P Y P Y Z P Z P R T M Downloaded from https://academic.oup.com/mnras/article/519/4/5848/6972382 by Montana Semons S. M., Reines A. E., Plotkin R. M., Gallo E., Greene J. E., 2015, ApJ , 805, 12 inden T., Kalogera V., Sepinsky J. F., Prestwich A., Zezas A., Gallagher J. S., 2010, ApJ , 725, 1984 ong K. S., Blair W. P., Winkler P. F., Lacey C. K., 2020, ApJ , 899, 14 ongair M. S., 1994, High Energy Astrophysics. Vol. 2: Stars, the Galaxy and the Interstellar Medium. Cambridge University Press, Cambridge, UK, p. 410 apelli M., Ripamonti E., Zampieri L., Colpi M., Bressan A., 2010, MNRAS , 408, 234 art ́ınez-Palomera J., Lira P., Bhalla-Ladd I., F ̈orster F., Plotkin R. M., 2020, ApJ , 889, 113 erloni A., Heinz S., di Matteo T., 2003, MNRAS , 345, 1057 ezcua M., Dom ́ınguez S ́anchez H., 2020, ApJ , 898, L30 ezcua M., Civ a no F., Fabbiano G., Miyaji T., Marchesi S., 2016, ApJ , 817, 20 ezcua M., Civ a no F., Marchesi S., Suh H., Fabbiano G., Volonteri M., 2018, MNRAS , 478, 2576 ezcua M., Suh H., Civ a no F., 2019, MNRAS , 488, 685 iller B. P., Gallo E., Greene J. E., Kelly B. C., Treu T., Woo J.-H., Baldassare V., 2015, ApJ , 799, 98 ineo S., Gilfanov M., Lehmer B. D., Morrison G. E., Sunyaev R., 2014, MNRAS , 437, 1698 irabel I. F., Dijkstra M., Laurent P., Loeb A., Pritchard J. R., 2011, A&A , 528, A149 oon D.-S., Harrison F. A., Cenko S. B., Shariff J. A., 2011, ApJ , 731, L32 oran E. C., Shahinyan K., Sugarman H. R., V ́elez D. O., Eracleous M., 2014, AJ , 148, 136 oretti A., Campana S., Lazzati D., Tagliaferri G., 2003, ApJ , 588, 696 urphy E. J. et al., 2011, ApJ , 737, 67 guyen D. D. et al., 2019, ApJ , 872, 104 rienti M., Prieto M. A., 2010, MNRAS , 401, 2599 akull M. W., Angebault L. P., 1986, Nature , 322, 511 akull M. W., Soria R., Motch C., 2010, Nature , 466, 209 ardo K. et al., 2016, ApJ , 831, 203 lotkin R. M., Markoff S., Kelly B. C., K ̈ording E., Anderson S. F., 2012, MNRAS , 419, 267 onnada S., Brorby M., Kaaret P., 2020, MNRAS , 491, 3606 restwich A. H., Tsantaki M., Zezas A., Jackson F., Roberts T. P., Foltz R., Linden T., Kalogera V., 2013, ApJ , 769, 92 eines A. E., Volonteri M., 2015, ApJ , 813, 82 NRAS 519, 5848–5858 (2023) eines A. E., Siv a koff G. R., Johnson K. E., Brogan C. L., 2011, Nature , 470, 66 eines A. E., Greene J. E., Geha M., 2013, ApJ , 775, 116 eines A. E., Condon J. J., Darling J., Greene J. E., 2020, ApJ , 888, 36 icarte A., Natarajan P., 2018, MNRAS , 481, 3278 ichards G. T. et al., 2006, AJ , 131, 2766 oberts T. P., Colbert E. J. M., 2003, MNRAS , 341, L49 osen S. R. et al., 2016, A&A , 590, A1 oss N. P. et al., 2013, ApJ , 773, 14 argent A. J., Johnson M. C., Reines A. E., Secrest N. J., van der Horst A. J., Cigan P. J., Darling J., Greene J. E., 2022, ApJ , 933, 160 artori L. F., Schawinski K., Treister E., Trakhtenbrot B., Koss M., Shirazi M., Oh K., 2015, MNRAS , 454, 3722 chaerer D., Fragos T., Izotov Y. I., 2019, A&A , 622, L10 chramm M. et al., 2013, ApJ , 773, 150 chutte Z., Reines A. E., 2022, Nature , 601, 329 enchyna P., Stark D. P., Mirocha J., Reines A. E., Charlot S., Jones T., Mulchaey J. S., 2020, MNRAS , 494, 941 hirazi M., Brinchmann J., 2012, MNRAS , 421, 1043 immonds C., Schaerer D., Verhamme A., 2021, A&A , 656, A127 oria R., Pakull M. W., Broderick J. W., Corbel S., Motch C., 2010, MNRAS , 409, 541 oria R., Pakull M. W., Motch C., Miller-Jones J. C. A., Schwope A. D., Urquhart R. T., Ryan M. S., 2021, MNRAS , 501, 1644 rebitsch M., Volonteri M., Dubois Y., Madau P., 2018, MNRAS , 478, 5607 rquhart R. et al., 2019, MNRAS , 482, 2389 anden Berk D. E. et al., 2001, AJ , 122, 549 olonteri M., 2010, A&AR, 18, 279 olonteri M., Habouzit M., Colpi M., 2021, Nat. Rev. Phys. , 3, 732 eaver R., McCray R., Castor J., Shapiro P., Moore R., 1977, ApJ , 218, 377 oo J.-H., Cho H., Gallo E., Hodges-Kluck E., Le H. A. N., Shin J., Son D., Horst J. C., 2019, Nat. Astron. , 3, 755 ang Q.-X., Xie F.-G., Yuan F., Zdziarski A. A., Gierli ́nski M., Ho L. C., Yu Z., 2015, MNRAS , 447, 1692 ork D. G. et al., 2000, AJ , 120, 1579 ounes G., Porquet D., Sabra B., Reeves J. N., 2011, A&A , 530, A149 hao Y., Gao Y., Gu Q., 2013, ApJ , 764, 44 ibetti S., Charlot S., Rix H.-W., 2009, MNRAS , 400, 1181 his paper has A been typeset from a TE X/L T E X file prepared by the author. tate University Library user on 23 March 2023