Resolved HI in two ultra-diffuse galaxies from contrasting non-cluster environments
T. C. Scott, Chandreyee Sengupta, P. Lagos, Aeree Chung, O. Ivy Wong
aa r X i v : . [ a s t r o - ph . GA ] F e b MNRAS , 1– ?? () Preprint 10 February 2021 Compiled using MNRAS L A TEX style file v3.0
Resolved HI in two ultra–di ff use galaxies from contrastingnon-cluster environments T. C. Scott ⋆ , Chandreyee Sengupta, , P. Lagos , Aeree Chung , O. Ivy Wong , , Institute of Astrophysics and Space Sciences (IA), Rua das Estrelas, 4150-762 Porto, Portugal Purple Mountain Observatory, No.8 Yuanhua Road, Qixia District, Nanjing 210034, China Department of Astronomy, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, Republic of Korea CSIRO Astronomy & Space Science, PO Box 1130, Bentley, WA 6102, Australia ICRAR-M468, University of Western Australia, Crawley, WA 6009, Australia ARC Centre of Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D), Australia
ABSTRACT
We report on the first resolved H i observations of two blue ultra-di ff use galaxies (UDGs)using the Giant Metrewave Radio Telescope (GMRT). These observations add to the so-far limited number of UDGs with resolved H i data. The targets are from contrasting non–cluster environments: UDG–B1 is projected in the outskirts of Hickson Compact Group 25and Secco–dI–2 (SdI–2) is an isolated UDG. These UDGs also have contrasting e ff ective radiiwith R e of 3.7 kpc (similar to the Milky Way) and 1.3 kpc respectively. SdI–2 has an unusuallylarge M HI M ∗ ratio = i observation. Both galaxies displayH i morphological and kinematic signatures consistent with a recent tidal interaction, whichis also supported by observations from other wavelengths, including optical spectroscopy.Within the limits of the observations’ resolution our analysis indicates that SdI–2 is darkmatter-dominated within its H i radius and this is also likely to be the case for UDG–B1.Ourstudy highlights the importance of high spatial and spectral resolution H i observations for thestudy of the dark matter properties of UDGs. Key words: galaxies: individual: UDG–B1 and Secco–dI–2 – galaxies: ISM – galaxies: in-teractions – galaxies: kinematics and dynamics – radio lines: galaxies
Recent studies (van Dokkum et al. 2015; Koda et al. 2015;Yagi et al. 2016) have reported over 1000 extended di ff use galaxiesin and surrounding the Coma galaxy cluster (z = µ g ) of ∼
24 – 26 magarcsec − , e ff ective radii ( R e ) ∼ ∼ × M ⊙ and were designated by van Dokkum et al. (2015)as ultra–di ff use galaxies (UDGs). These UDGs have a median R e similar to L* spirals ( R e ∼ ∼ − M ⊙ (Lagos et al.2011, and references therein). While faint, extended, low surfacebrightness galaxies (LSBs) are not a recent discovery, the ComaUDGs reveal their higher abundance in dense environments(van der Burg et al. 2017). Compared to classical LSBs, UDGs areoptically fainter and often more extended (Yagi et al. 2016) with awide range of optical R e . Conselice (2018), argues UDGs are partof an earlier reported population of Low–Mass Cluster Galaxies ⋆ [email protected] The e ff ective radius of a galaxy is the radius at which half of the totallight is emitted (LMCGs) and attributes their large diameters to interactions withthe cluster environment and expected them to be dark matter(DM) dominated. On the face of it, the reported tight correla-tion between the abundance of UDGs and the mass of the hostclusters and groups (van der Burg et al. 2017; Román & Trujillo2017) supports environmentally driven formation scenarios, suchas those proposed by Baushev (2016); Yozin & Bekki (2015);Carleton et al. (2018). Janssens et al. (2019) reported on theabundance and distribution of UDGs in 6 masive clusters at z = / LMCGs (referred from hereon as UDGs) turn out to be a subset of classical LSBs, as suggestedby Tanoglidis et al. (2020), or separate class(es) of their own, is © The Authors
Scott et al. likely to depend on the so-far unresolved question of how LSBsand UDGs form. Attention is turning toward the DM content andits distribution in UDGs because of its critical role in the formationand evolution of their observable baryonic components. Beyondthe cluster environment, where significant ram pressure strippingof a UDGs H i is not expected, H i studies become feasible.Resolved H i in particular is well suited to both the detection ofrecent interactions and determination of DM properties and as aresult, can assist in answering these questions.In this paper, we present resolved Giant Metrewave RadioTelescope (GMRT) H i observations of two blue (g – i < < = ± < g – i < ff ering definitions in the literatureof a UDG, in particular, the criteria for their R e , e.g. (Yagi et al.2016, > > > R e = ± R e is above the R e of ∼ R e of 1.5 kpc for UDGs adopted by many au-thors, e.g. Forbes et al. (2020a) who would consider SdI–2 as a“small UDG" or “LSB dwarf" rather than a UDG. The R e of SdI–2 lies only slightly below the R e R e lies within uncertainties of three of the UDGsin the Leisman et al. (2017) sample of HI rich UDGs. Addition-ally R e measurement can change significantly with the optical band(Forbes et al. 2020b). SdI–2’s R e (measured at 1.3 ± R e lower limit and consistency with the previousclassifications of SdI–2 as a UDG, we adopt its classification as aUDG. We use optical images from the Sloan Digital Sky Survey(SDSS), IAC Stripe 82 and Pan-STARRS1 (Flewelling et al. 2016)as well as UV imaging from Galaxy Evolution Explorer ( GALEX )in this paper. UDG–B1 also has an SDSS spectrum that we utiliseas part of our analysis.Section 2 gives details of the GMRT observations, with obser-vational results in section 3. A discussion follows in section 4 witha summary and concluding remarks in section 5. To aid compar-isons in this paper we adopt distances to UDG–B1 and SdI–2 of88 Mpc and 40 Mpc, respectively from Spekkens & Karunakaran(2018) and Papastergis et al. (2017). We also adopt their angularscales of 1 arcmin ∼
25 kpc and 1 arcmin ∼
11 kpc, for UDG–B1and SdI–2, respectively. All α and δ positions referred to through-out this paper are J2000.0. UDG–B1 and SdI–2 were observed in H i using the GMRT on 2018March 8th and 9th and selected observational parameters are de-tailed in Table 2. H i in UDG–B1 has previously been detected Table 1.
Properties of UDG–B1 and SdI–2
Property a Units UDG–B1 SdI–2 V optical b [km s − ] 6445 ±
12 2549 ± V HI W [km s − ] ± ± W [km s − ] ±
15 80 ± RA c [h:m:s] 03 20 21.1 11 44 33.8DEC [d:m:s] -01 10 12.0 -00 52 00.9Distance d [Mpc] 88 40 ± R e e [kpc] 3.7 ± ± µ g (0) f [mag arcsec ] 24.0 > + log(O / H) g ± ± S HI (GMRT) [Jy km s − ] ± ± M HI (GMRT) [10 M ⊙ ]
15 2.6 M ∗ h [10 M ⊙ ] 2.2 ± i ± a From the literature as noted or if in bold typeface from this work. b SdI–2 from Bellazzini et al. (2017). c RA and DEC positions for UDG–B1are from Román & Trujillo (2017)and SdI–2 are from (Bellazzini et al. 2017). d See section 1. e UDG–B1 from Román & Trujillo (2017), SdI–2 fromBellazzini et al. (2017). f UDG–B1 from Trujillo et al. (2017). g For UDG–B1 from a SDSS spectrum of the brightestSF knot see Table 3 and SdI–2 from (Bellazzini et al. 2017). h UDG–B1 from Román & Trujillo (2017), SdI–2 fromPapastergis et al. (2017). i UDG–B1 from Román & Trujillo (2017). SdI–2 fromSDSS photometry. within a VLA D–configration field (Spekkens & Karunakaran2018; Borthakur et al. 2010) centred on the HCG 25 group. TheUDG–B1 H i was unresolved in the VLA 70 ×
50 arcsec FWHMsynthised beam, which provided the motivation to reobserve UDG–B1 with the GMRT to obtain the first ever resolved mapping of itsH i . The GMRT data was reduced and analyzed using the standardreduction procedures with the Astronomical Image Processing Sys-tem ( aips ) software package. The flux densities are on the scale ofBaars et al. (1977), with uncertainties of ∼ aips , task imagr was used to convert the uv domain data to H i image cubes. Finallyintegrated H i and velocity field maps were extracted from the im-age cubes using the aips task momnt . To study the H i distributionin detail, image cubes with di ff erent resolutions were produced byapplying di ff erent ‘tapers’ to the data with varying uv limits. De-tails of the final low, medium and high resolution maps are given inTable 2. UDG–B1 is projected ∼ opt ) is 89 km s − higher than the mean ve-locity for HCG 25, V opt = − (NED) with a dispersion( σ v ) = − (Tovmassian et al. 1999). UDG-B1 is projectedwithin the typical group R of 500 kpc (Román & Trujillo 2017), MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments UDG-B1
UGC 2690 UDG -B1
Figure 1. Wide–angle region SW of HCG 25 with the GMRT low resolution (44.82 ′′ × ′′ ) H i contour (white) at 0.39 × atoms cm − and mediumresolution (26.46 ′′ × ′′ ) H i contours (green) at 1.46, 2.73, 3.6, 5.5 11.0 and 13.8 × atoms cm − . These contours are overlaid on an IAC stripe82 composite g,r, i image showing H i detections in UGC 2690 and UDG –B1. The ellipses at the bottom left indicate the size and orientation of the GMRTsynthesized beams. where R is the radius enclosing an overdensity of >
200 withrespect to the critical density of the universe. Its projected positionand velocity are therefore consistent with UDG–B1 being an out-skirts member of HCG 25 so past interactions between the UDG–B1 and HCG 25 group members cannot be ruled out. H i was de-tected in UDG-B1 with the GMRT at a velocity (V HI ) = ± − and in HCG 25 member UGC 2690 (V opt = − ,W = ± − ), projected 6.6 arcmin NE and within 165km s − of UDG–B1, see Figure 1. Further details of the H i detectedin UGC 2690 are presented in Appendix A.The left–hand panels of Figure 2 show (top to bottom) the low(44.82 ′′ × ′′ ), medium (26.46 ′′ × ′′ ) and high (22.73 ′′ × ′′ ) resolution GMRT UDG–B1 integrated H i map contoursoverlaid on an IAC Stripe 82 g, r, i composite image. The high res-olution H i map contours reveal significant asymmetry in the H i morphology, with the bulk of the high column density H i o ff set by ∼
20 arcsec (8 kpc) from optical centre. For UGC–B1 the GMRTflux density (S HI ) = − which converts to an M HI ∼ × M ⊙ . Spekkens & Karunakaran (2018) have previously re-ported Green Bank Telescope (GBT) single dish and Very Large The H i velocity was determined from the mid point in the W velocityrange, i.e. V HI W as definded in Reynolds et al. (2020). Array (VLA) H i detections for UGC–B1 with S HI = ± − from the VLA data. At the adopted distance of 88 Mpc thisVLA S HI implies M HI = × M ⊙ , in good agreement withthe M HI derived from the GMRT. The VLA flux density sensitiv-ity was ∼ − (Borthakur et al. 2010), quite similar toGMRT medium and high resolution maps. For the GBT spectrumSpekkens & Karunakaran (2018) reported a higher peak signal tonoise ratio (15.2) than from the VLA spectrum (9.2). At first sightthis may seem like some flux loss for the interferometric observa-tions. However, while the GBT pointing was at the position UDG–B1 the GBT FWHP beam is ∼ i spectra of UDG–B1 andUGC 2690, attenuated to 40% of its actual flux, produces a spec-trum consistent with the GBT spectrum. So, for the analysis in thispaper we adopt the GMRT H i mass. Based on this H i mass and the M ∗ from Table 2 the M HI M ∗ ratio for UGC–B1 is 6.8.For SdI–2, Figure 3 (Top row) shows the low (39.95 ′′ × ′′ ) and medium (24.50 ′′ × ′′ ) resolution GMRT inte-grated H i map contours overlaid on a smoothed Pan–STARRS g,r, i composite image. Figure 3 shows the main body of the galaxyhas an H i extent of ∼ i tail like MNRAS , 1– ????
20 arcsec (8 kpc) from optical centre. For UGC–B1 the GMRTflux density (S HI ) = − which converts to an M HI ∼ × M ⊙ . Spekkens & Karunakaran (2018) have previously re-ported Green Bank Telescope (GBT) single dish and Very Large The H i velocity was determined from the mid point in the W velocityrange, i.e. V HI W as definded in Reynolds et al. (2020). Array (VLA) H i detections for UGC–B1 with S HI = ± − from the VLA data. At the adopted distance of 88 Mpc thisVLA S HI implies M HI = × M ⊙ , in good agreement withthe M HI derived from the GMRT. The VLA flux density sensitiv-ity was ∼ − (Borthakur et al. 2010), quite similar toGMRT medium and high resolution maps. For the GBT spectrumSpekkens & Karunakaran (2018) reported a higher peak signal tonoise ratio (15.2) than from the VLA spectrum (9.2). At first sightthis may seem like some flux loss for the interferometric observa-tions. However, while the GBT pointing was at the position UDG–B1 the GBT FWHP beam is ∼ i spectra of UDG–B1 andUGC 2690, attenuated to 40% of its actual flux, produces a spec-trum consistent with the GBT spectrum. So, for the analysis in thispaper we adopt the GMRT H i mass. Based on this H i mass and the M ∗ from Table 2 the M HI M ∗ ratio for UGC–B1 is 6.8.For SdI–2, Figure 3 (Top row) shows the low (39.95 ′′ × ′′ ) and medium (24.50 ′′ × ′′ ) resolution GMRT inte-grated H i map contours overlaid on a smoothed Pan–STARRS g,r, i composite image. Figure 3 shows the main body of the galaxyhas an H i extent of ∼ i tail like MNRAS , 1– ???? () Scott et al.
Table 2.
GMRT observational and map parameters
UDG-B1
Rest frequency 1420.4057 MHzObservation Date 2018 Mar 9Integration time 10.0 hrsprimary beam 24 ′ at 1420.4057 MHzLow resolution beam–FWHP 44.82 ′′ × ′′ , PA = ◦ Medium resolution beam–FWHP 26.46 ′′ × ′′ , PA = ◦ High resolution beam–FWHP 22.73 ′′ × ′′ , PA = -9.61 ◦ RA (pointing centre) 03 h m s DEC (pointing centre) – 01 ◦ ′ ′′ SdI–2
Rest frequency 1420.4057 MHzObservation Date 2018 Mar 8Integration time 10.0 hrsprimary beam 24 ′ at 1420.4057 MHzLow resolution beam–FWHP 39.95 ′′ × ′′ , PA = – 10.63 ◦ Medium resolution beam–FWHP 24.50 ′′ × ′′ , PA = ◦ RA (pointing centre) 11 h m s DEC (pointing centre) – 00 ◦ ′ ′′ extension was also detected to the NE of the main H i body reach-ing ∼ i map. H i in this extended region has a column den-sity maximum > × atoms cm . The main H i body of SdI–2 has optical, NUV and FUV (GALEX) and NIR (WISE 3.4 µ mand 4.6 µ m) counterparts, see Figure 3 – Lower right panel. ForSdI–2 the GMRT S HI = ± − which at the adopteddistance of 40 Mpc gives M HI ∼ × M ⊙ . A previous H i detection for SdI–2 using the E ff elsberg 100 m single dish tele-scope gave S HI = − which converts to M HI = × M ⊙ (Papastergis et al. 2017), in good agreement with the M HI derived from the GMRT. For the analysis in this paper we adopt theGMRT H i mass. Based on the GMRT H i mass and the M ∗ from(Papastergis et al. 2017), the M HI M ∗ ratio for SdI–2 is 28.9, which ismuch higher than is typical for UDGs irrespective of R e , see Fig-ure 5. As reported in Papastergis et al. (2017) this high M HI M ∗ ratio isconsistent with the trend for isolated UDGs to show significantlyhigher HIM ∗ ratios than UDGs in denser environments. It shouldbe remembered that this plot only considers relatively isolated H i UDGs and does not include cluster galaxies which are expected tobe H i deficient. That said we would expect the higher R e region ofthe plot to favour high DM halo spin (Amorisco & Loeb 2016) orstar formation SF driven extended disc (e.g. Di Cintio et al. 2017)formation models. Conversely, environment density driven forma-tion scenarios should favour the lower M HI / M ∗ ratio regions of theplot, but more H i observations of cluster galaxies are needed to ex-plore this parameter space. A NED search revealed no companionsprojected within a radius of 30 arcmin ( ∼
330 kpc) and a velocityrange of ±
500 km s − . However, the SdI–2 head and tail H i mor-phology is suggestive of a recent interaction. Given this and that thenearerst galaxy cluster (ZwCl 1141.7-0158B) is projected ∼ − away, we can reasonably rule out ram pres-sure stripping as the origin of the one–sided H i tail. This leaves aninteraction with a low mass satellite or gas cloud within 7 × yr as a likely explanation for the H i detached region. The abovetimescale is based on Holwerda et al. (2011), who predict the H i perturbations from even major mergers will only remain detectablein H i for a maximum of 0.4 to 0.7 Gyr. In summary the H i mor-phologies of both UGG–B1 and SdI–2 indicate a recent interaction. The evidence for and against a recent interactions is discussed fur-ther in Section 4.2. Based on UDG–B1’s GMRT H i spectrum, its V HI = ± − and W = ±
15 km s − . This velocity agrees withinthe uncertainties with the V HI = ± − and W = ± − from the VLA (Spekkens & Karunakaran 2018). Figure 2(top right panel) shows the UDG–B1 H i low resolution velocityfield, with H i detected in the velocity range 6404 km s − to 6512km s − . The H i velocity field shows a NE to W gradient, whichindicates that the H i disc major axis is approximately perpendicu-lar to the N – S optical axis and the change in the position angleof the iso–velocity contours from the NE to W suggests a warpingof the disc, although with the caveat that the spatial resolution islow. We attempted to determine the rotational velocity (V rot ) froma 3D model fit to the UDG-B1 H i emission in the medium resolu-tion cube using bbarolo (Di Teodoro & Fraternali 2015). Unfortu-nately, the small number of beams across the H i disc and the lowsignal to noise ratio (S / N) meant this attempt failed. Instead usinginclination (i = ◦ ) and W derived from the medium resolutionH i cube and equation 1 we estimated V rot =
148 km s − . v rot = W sin ( i ) (1)where, sin ( i ) = vut − (cid:16) ba (cid:17) − q a, b are the lengths of the major and minor H i axes, 43.3 arc-sec and 41.46 arcsec respectively from the medium resolution H i map and q = W = q w raw − δ s where w raw is the W measured from the spectrum and δ s is acorrection for instrumental broadening and turbulence (10 km s − ).However, this inclination corrected V rot estimate of 148km s − is anomalously high by a factor of 2 to 3 for a UDG ofits H i mass, even allowing for the significant uncertainties in theH i major and minor axes because synthised beam is large relativeto axes. So, as a check we also calculated V rot using Equation 1but replacing the H i axial ratio (b / a) with the optical axial ratio of0.46 from Román & Trujillo (2017). This substitution resulted ina revised inclination of 68.6 ◦ and V rot =
48 km s − . A visual as-sessment of the medium resolution H i map in Figure 1 indicatesthe H i axial ratio is significantly lower than its optical counterpart.It is also important to note the calculations above assume that theH i is dynamical equilibrium, but if UDG–BI has su ff ered a recentinteraction, as the analysis of the H i morphology suggests, the ap-parently large H i V rot of 148 km s − is more likely an interactioninduced kinematic artefact rather than an usually large H i rotationvelocity. The implications of the two V rot estimates for calculationof UDG–B1’s dynamical mass are discussed in section 4.1.Measurent of SdI–2’s GMRT H i spectrum gives V HI = ± − and W = ± − . V HI = − and W =
69 km s − were reported for the galaxy in Papastergis et al. (2017)using observations from the E ff elsberg 100 m telescope. SdI–2’svelocity field in Figure 3 (lower right panel) shows two distinct MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments kinematic regions; the main H i body of the galaxy, which has alength of ∼ ′′ (14 kpc) and displays a regular velocity gradient in-dicating a possible rotating H i disc ( ∼ − to 2618 km s − )and a morphologically detached H i extension to the NE with a nar-row range of velocities from ∼ − to 2611 km s − . H i velocities in the detached region are similar to those in NE of themain galaxy body The systematic change in PA ( ∼ ◦ ) of the iso-velocity contours in the main body, indicates the H i disc is signifi-cantly warped. Based on the H i major and minor axes of the mainbody of the galaxy in medium resolution H i map we estimated theinclination of the H i disc at 67 ◦ . Using this inclination and W from the medium resolution SdI–2 cube and estimated V rot = − using equation 1. We also attempted to use bbarolo to fit a3D model to the GMRT H i medium resolution cube which gave afit with an inclination ∼ ◦ and V rot ∼
35 km s − , but the low num-ber of beams across the disc ( ∼
3) and low S / N ratio in the cubemeant we could not place a strong reliance on the fit, but the incli-nation and V rot from the two methods are in reasonable agreement.The PV diagram (Figure 4) is from a cut (PA = ◦ ) oriented alongthe H i disc’s major axis. The PV diagram is consistent with a bulkof H i consisting of a rotating disc with the detached NE region onlyappearing in a narrow range of velocities ( ∼ − to 2611km s − ). Figure 4 also shows low level emission extending NE frommain H i body toward the detached H i region and indicating the twoH i structure are likely to be associated. This is discussed further insection 4.2. Ideally a UDG’s dynamical mass would be determined using its H i rotation curve which in the field or in groups can probe a galaxy’sdynamical mass well beyond the radius of its stellar disk. By fit-ting a H i rotation curve to DM halo models the mass and massprofile of the DM halo can be estimated. But in clusters becauseof the absence of H i due to e ffi cient H i stripping by the clusterenvironment, dynamical and DM halo mass estimates to date aretypically based on globular cluster studies. (e.g. Beasley & Trujillo2016; Beasley et al. 2016; van Dokkum et al. 2018, 2019b,a) Thisapproach is supported by the scaling relation between the num-ber of globular clusters and the host galaxy’s halo virial mass(Burkert & Forbes 2020), although the impact of the cluster on aUDGs globular population remain unclear (Forbes et al. 2020a). Inthe case of the Coma cluster UDG DF–44, which has the largestknown DM halo mass of M = – 10 , stellar kinematics pro-vided the dynamical mass (van Dokkum et al. 2019c). There are asmall number of Coma UDGs, like DF–44, which have high num-bers of globular clusters per unit stellar mass implying extremelymassive DM halos with two thirds of Coma UDGs ( R e > > M ⊙ based on globular cluster–DM halomass scaling relations Forbes et al. (2020a). These halo masses areconsistent with the failed Milky Way mass galaxies. Earlier glob-ular cluster-based studies of galaxy cluster UDGs had indicatedthat while high dynamical to stellar mass ratios were common, e.g.VCC 1287 and DF 17, typically those studies implied dwarf massDM halos (e.g. Beasley & Trujillo 2016; Beasley et al. 2016).An important recent development has been reports ofbaryon dominated UDGs based on globular cluster stud-ies (van Dokkum et al. 2018, 2019b,a) and resolved H i Mancera Piña et al. (2019, 2020). Zaritsky (2017), using scaling relations, suggested that UDGs may span a range of DM halomasses between those typically found in large spirals to dwarfgalaxies, but further observations are needed to confirm this.Deriving an accurate estimate of a UDG’s DM content from H i requires a rotation curve extracted from resolved H i observa-tions. To date only a few UDGs have been mapped in H i (e.g.Leisman et al. 2017; Sengupta et al. 2019; Mancera Piña et al.2019). Interestingly, several of these UDGs with resolved H i show an apparent departure from the McGaugh et al. (2000)baryonic Tully Fisher relation (BTFR). Analysis of the DMcontent of six of the UDGs in the Leisman sample of isolatedH i detected UDGs (Mancera Piña et al. 2019, 2020) indicatesthat those galaxies are baryon dominated within their R HI , seeFigure 4 in Mancera Piña et al. (2019), although their results arebased on a limited number of bbarolo (Di Teodoro & Fraternali2015) 3D kinematic tilted–ring disc model fits for each galaxy. Asimilar, although lower magnitude, departure from the BTFR wasreported in Sengupta et al. (2019) for the UDG, UGC 2162 alsoderived from bbarolo fitting. Confirmation of a baryon dominatedpopulation of UDGs has important implications for the origin ofthose UDGs. Additionally, the Mancera Piña et al. (2019, 2020)galaxies as well as UGC 2162 have a lower V rot compared to nor-mal dwarf galaxies of similar H i mass (Mancera Piña et al. 2020;Sengupta et al. 2019). This is consistent with the results from thelarger Leisman et al. (2017) sample which found the mean velocitywidth, W , for UDGs was significantly narrower ( < > km s − )than for ALFALFA galaxies with a similar selection criteria < > km s − ).As noted in section 3.2 the bbarolo
3D kinematic tilted–ringdisc models to the H i cube for UDG–B1 failed and for SdI–2 thefit provided only a poorly constrained estimate the DM halo prop-erties. So as an alternative, we estimated their dynamical masses( M dyn ) from the available H i properties. We did this using the H i radii from the medium resolution GMRT integrated H i maps forUDG–B1 and SdI–2 ( ∼ rot
148 km s − (UDG–B1) and 41 km s − (SdI–2). Using this method their respective M dyn was estimated at45.9 × M ⊙ and 2.7 × M ⊙ . The baryonic masses ( M baryon )of the galaxies ( M HI × + M ∗ ) from Table 2 equal 1.76 × M ⊙ and 0.35 × M ⊙ for UDG–B1 and SdI–2, respectively. The1.4 factor applied to the H i masses was to adjust for the heliumand molecular gas content of the UDGs. Based on these values thebaryon fractions for UDG–B1 and SdI–2 are 0.05 and 0.14, respec-tively. These fractions indicate that, within the radii where H i wasdetected, both UDG–B1 and the main body of SdI–2 are both darkmatter dominated.However as noted in Section 3.2, the UDG–B1 inclination cor-rected V rot estimate of 148 km s − derived from from the H i axialratio is anomalously high. If instead we use V rot =
48 km s − basedon the optical axial ratio it implies a M dyn = × M ⊙ andbaryon fraction ( M bar / M dyn ) increases to 0.45. In figure 6, origi-nally from Mancera Piña et al. (2020), we show the log of the ratioof baryonic to dynamical mass versus the log of dynamical massfor UDG–B1, SdI–2, and UGC 2162 (Sengupta et al. 2019) as wellas the galaxies from the LITTLE THINGS (Oh et al. 2015), andMancera Piña et al. (2019) as shown in Mancera Piña et al. (2020).For UDG– B1 we show the masses derived using both V rot = − and 48 km s − , with a blue line joining the two sets of de-rived masses. The horizontal lines indicate the DM fraction (f DM )shown at the right–hand margin. For UDG–B1 the baryonic (or in-versely DM) fraction depends on the adopted V rot which in turndepends on whether the H i or optical axial ratio is used. Use of the MNRAS , 1– ????
48 km s − basedon the optical axial ratio it implies a M dyn = × M ⊙ andbaryon fraction ( M bar / M dyn ) increases to 0.45. In figure 6, origi-nally from Mancera Piña et al. (2020), we show the log of the ratioof baryonic to dynamical mass versus the log of dynamical massfor UDG–B1, SdI–2, and UGC 2162 (Sengupta et al. 2019) as wellas the galaxies from the LITTLE THINGS (Oh et al. 2015), andMancera Piña et al. (2019) as shown in Mancera Piña et al. (2020).For UDG– B1 we show the masses derived using both V rot = − and 48 km s − , with a blue line joining the two sets of de-rived masses. The horizontal lines indicate the DM fraction (f DM )shown at the right–hand margin. For UDG–B1 the baryonic (or in-versely DM) fraction depends on the adopted V rot which in turndepends on whether the H i or optical axial ratio is used. Use of the MNRAS , 1– ???? () Scott et al.
Figure 2. UDG-B1: GMRT H i Left column (top to bottom):
Integrated H i contours from the low (44.82 ′′ × ′′ ), medium 26.46 ′′ × ′′ ) and high(22.73 ′′ × ′′ ) resolution maps overlaid on an IAC stripe 82 g, r, i composite images. The column density map contours are at, low resolution; 0.39, 0.78,1.17, 2.10 and 3.00 × atoms cm − , medium resolution; 1.46, 2.73, 3.6 and 5.5 × atoms cm − , high resolution; 1.26, 2.21, 2.84, 3.79, 4.74 and 5.60 × atoms cm − . Right top: H i velocity field from the low resolution cube, with the iso–velocity contours values, separated by 10 km s − , indicated by thecolour scale. Lower right:
GMRT medium resolution (26.46 ′′ × ′′ ) integrated H i map contours on a smoothed GALEX NUV image. The small yellowcircles are the positions and fields of view of the two available SDSS spectra fibres. The ellipses in each panel indicates the size and orientation of the GMRTsynthesized beams. optical axial ratio implies V rot =
48 km s − and a baryon fraction of0.45, which is close to that of the galaxy in the Mancera Piña et al.(2019) sample with the lowest baryon fraction. Alternatively, usingthe H i axes ratio and V rot =
148 km s − gives much lower baryonfraction of 0.05, which would place it near the extreme of DM dom-inated dwarfs. As observed in Section 3.2 there are reasons to thinkthe V rot values from which these baryonic fractions are derived areupper and lower limits. As a result we consider its likely that true baryonic fraction of UDG–B1 lies at an intermediate point in range0.05 to 0.45 and thus is likely be similar to the median value for theLITTLE THINGS dwarfs, see Figure 6.Hence, in terms of DM content, our estimates of M dyn for SdI–2, and the uncertainly for UDG–B1, suggest the DM halos withinthe H i radii resemble normal DM dominated dwarf galaxies. Figure5 shows the M HI / M ∗ v R e for UDG–B1, SdI–2 and a selection ofUDGs from the literature, including the baryon dominated UDGs MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments Figure 3. SdI–2:
Top (left to right)t:
GMRT low(39.95 ′′ × ′′ ) and medium (24.50 ′′ × ′′ ) resolution H i contours overlaid on a Pan–STARRSg,r,i–band composite image. The column density map contours are at, low resolution; 0.15, 0.54, 0.77, 1.55, 2.33, 3.11 and 3.50 × atoms cm , mediumresolution; 0.40, 1.02, 1.83, 3.06, 4.08, 6.10 and 7.14 × atoms cm . Lower left:
SdI–2 smoothed NUV (GALEX) image with the star forming regions fromBellazzini et al. (2017) indicated.
Lower right:
Medium resolution H i velocity field iso–velocity contours separated by 10 km s − overlaid on H i integratedmedium resolution map.The GMRT beam sizes and orientations are shown with ellipses at the bottom of each panel. reported in (Mancera Piña et al. 2019, 2020), which have R e & R e and their uncertainties in the plot are from the literaruresources, principally from Leisman et al. (2017) which are based ontheir own photometry of SDSS images. We note from the Figurethat UDG–B1 has a R e within the R e range where baryon dominatedUDGs have been reported. This suggests that for UDGs outsideclusters their R e is determined by factor other than their baryonfraction and M HI M ∗ ratio. Figure 2 (left middle panel) shows the IAC stripe 82 optical im-age of the UDG–B1 and the two small over plotted yellow circlesindicate the positions and 3 arcsec diameters of the SDSS spectrafibres. Each spectra is from a di ff erent galaxy with the spectrum inthe south from UDG–B1 and the one above it from a backgroundgalaxy. With the aid of the SDSS spectra and a NUV (GALEX) im-age (Figure 2 – lower right panel) we interpret the optical image as MNRAS , 1– ????
Medium resolution H i velocity field iso–velocity contours separated by 10 km s − overlaid on H i integratedmedium resolution map.The GMRT beam sizes and orientations are shown with ellipses at the bottom of each panel. reported in (Mancera Piña et al. 2019, 2020), which have R e & R e and their uncertainties in the plot are from the literaruresources, principally from Leisman et al. (2017) which are based ontheir own photometry of SDSS images. We note from the Figurethat UDG–B1 has a R e within the R e range where baryon dominatedUDGs have been reported. This suggests that for UDGs outsideclusters their R e is determined by factor other than their baryonfraction and M HI M ∗ ratio. Figure 2 (left middle panel) shows the IAC stripe 82 optical im-age of the UDG–B1 and the two small over plotted yellow circlesindicate the positions and 3 arcsec diameters of the SDSS spectrafibres. Each spectra is from a di ff erent galaxy with the spectrum inthe south from UDG–B1 and the one above it from a backgroundgalaxy. With the aid of the SDSS spectra and a NUV (GALEX) im-age (Figure 2 – lower right panel) we interpret the optical image as MNRAS , 1– ???? () Scott et al.
Figure 4. SdI–2 PV diagram: from the low resolution GMRT cube (spatial resolution ∼
40 arcsec). The cut (PA 23.3 ◦ ) is through the centre along the majorH i axis. Positive angular o ff sets are to the SW of the optical centre and negative o ff sets are to the NE and include the H i NE detached feature. The velocityresolution of the cube is 7 km s − . Figure 5. M HI / M ∗ v R e for UDG–B1, SdI–2 and selected UDGs reported in the literature i.e. from Mancera Piña et al. (2019, 2020), Leisman, et al.,(2017), Spekkens & Karunakaran (2018), Sengupta et al. (2019.) MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments Figure 6. M bar / M dyn v M dyn : for UDG–B1 and SdI–2 in comparison to the UDG samples from Mancera Piña et al., 2020), dwarf galaxies LITTLE THINGS(Oh et al. 2015) and UDG 2162 (Sengupta et al. 2019.). For UDG B1 we show the M dyn determined using the two values V rot , 48 km s − and 148 km s − discussed in Section 3.2. The blue line indicates the range these two values cover. The horizontal lines indicate the dark matter fractions shown in right handmargin. showing that UDG–B1 is projected in front of a z = ± .0005background galaxy and extends in faint blue emission ∼
30 arcsec(12.5 kpc) north of the strong star forming region detected in theSDSS UDG–B1 spectrum. This picture of UDG–B1’s elongatedN–S morphology is confirmed by the GALEX NUV emission inFigure 2 – lower left panel.From the UDG–B1 GMRT medium and high resolutionH i maps we see the high column density H i is located Wof the optical / NUV axes (Figure 2). UDG–B1’s R e is 3.7 kpc(Román & Trujillo 2017) and the NED major optical axis of 0.35arcmin indicates the R opt is ∼ i map we esti-mate the R HI at 9 kpc and the R HI / R opt ratio = i distribution and uncertain in-clination the UDG–B1 R HI / R opt ratio is consistent with the typical R HI / R opt of 1.8 for late type galaxies (Broeils & Rhee 1997), i.e,we see no clear indication that the UDG–B1 H i disc is truncated.The low resolution H i velocity field for UDG–B1 (Figure 2– upper right panel) shows an overall rotation pattern running ap-proximately perpendicular to the optical / NUV axes, but the changein iso–velocity angle from NE to W suggesting a possible warpeddisc. The H i / optical axis o ff set, H i warp and anomalous V rot are allsignatures of an interaction well within the H i relaxation time scaleof < i maps show the currentblue star–forming region is o ff set in projection from the H i col-umn density maxima which could be a further signature of a recenttidal interaction. A galaxy’s A flux ratio is a measure of the asymme-try in its integrated H i flux density profile (within its W velocityrange) at velocities above and below the galaxy’s systemic veloc-ity (V HI ). A flux = A flux > A flux for UDG–B1 is1.23 ± A flux values (Watts et al. 2020).UDG–B1’s SDSS spectrum cannot be considered represen-tative of the galaxy as a whole because of the 3 arcsec diame-ter the SDSS spectrum fibre samples only ∼ R e disc area and this region has a bluer color than therest of the galaxy (Román & Trujillo 2017). Measurements derivedfrom the spectrum, including oxygen abundance, < + log(O / H) > = ± α ) ∼ ⊙ yr − are set out in Table3. In this table < + log(O / H) > was obtained as the average fromthe Marino et al. (2013), (Kobulnicky et al. 1999, lower branch)and Pettini & Pagel (2004) calibrators. The starformation rate wascalculated assuming the Kennicutt (1998) conversion formula af- MNRAS , 1– ????
30 arcsec(12.5 kpc) north of the strong star forming region detected in theSDSS UDG–B1 spectrum. This picture of UDG–B1’s elongatedN–S morphology is confirmed by the GALEX NUV emission inFigure 2 – lower left panel.From the UDG–B1 GMRT medium and high resolutionH i maps we see the high column density H i is located Wof the optical / NUV axes (Figure 2). UDG–B1’s R e is 3.7 kpc(Román & Trujillo 2017) and the NED major optical axis of 0.35arcmin indicates the R opt is ∼ i map we esti-mate the R HI at 9 kpc and the R HI / R opt ratio = i distribution and uncertain in-clination the UDG–B1 R HI / R opt ratio is consistent with the typical R HI / R opt of 1.8 for late type galaxies (Broeils & Rhee 1997), i.e,we see no clear indication that the UDG–B1 H i disc is truncated.The low resolution H i velocity field for UDG–B1 (Figure 2– upper right panel) shows an overall rotation pattern running ap-proximately perpendicular to the optical / NUV axes, but the changein iso–velocity angle from NE to W suggesting a possible warpeddisc. The H i / optical axis o ff set, H i warp and anomalous V rot are allsignatures of an interaction well within the H i relaxation time scaleof < i maps show the currentblue star–forming region is o ff set in projection from the H i col-umn density maxima which could be a further signature of a recenttidal interaction. A galaxy’s A flux ratio is a measure of the asymme-try in its integrated H i flux density profile (within its W velocityrange) at velocities above and below the galaxy’s systemic veloc-ity (V HI ). A flux = A flux > A flux for UDG–B1 is1.23 ± A flux values (Watts et al. 2020).UDG–B1’s SDSS spectrum cannot be considered represen-tative of the galaxy as a whole because of the 3 arcsec diame-ter the SDSS spectrum fibre samples only ∼ R e disc area and this region has a bluer color than therest of the galaxy (Román & Trujillo 2017). Measurements derivedfrom the spectrum, including oxygen abundance, < + log(O / H) > = ± α ) ∼ ⊙ yr − are set out in Table3. In this table < + log(O / H) > was obtained as the average fromthe Marino et al. (2013), (Kobulnicky et al. 1999, lower branch)and Pettini & Pagel (2004) calibrators. The starformation rate wascalculated assuming the Kennicutt (1998) conversion formula af- MNRAS , 1– ???? () Scott et al.
Table 3.
UDG-B1: Emission line fluxes and properties from theSDSS spectrum. The average 12 + log(O / H) is obtained from theMarino et al. (2013), Kobulnicky et al. (1999; lower branch) andPettini & Pagel (2004) calibrators, respectively.I( λ ) / I(H β ) × ± ± ± α ± ± ± β ± ± α ) × − erg cm − s − ± β ) × − erg cm − s − ± α / H β ×
100 177.78 ± β ) 0.00 ± α ) Å 115.0EW(H β ) Å 41.0log([OIII] / H β ) 0.41 ± / H α ) -1.66 ± / H α ) -0.53 ± α ) M ⊙ yr − ± × M ⊙ + log(O / H) M ± + log(O / H) KL ± + log(O / H) PP ± < + log(O / H) > ± ter correction for Chabrier IMF. The spectrum was corrected forgalactic extinction and internal extinction was insignificant. Thesub–solar 12 + Log(O / H) values for UDG–B1 indicate that the gas,at least at the position of the spectrum, is unlikely to have been ac-quired from more evolved members of HCG 25. Additionally, anal-ysis of the SDSS spectra by Román & Trujillo (2017) indicates astellar age < i perturbation signa-tures make it more likely the recent enhancement of SF was trig-gered by an interaction. In summary the UDG–B1 H i morphol-ogy and kinematics both show indications of a recent interaction,which has perturbed its H i disc, most likely with another memberof HCG 25 and the blue color and young age of the stellar popu-lation at the position of the SDSS spectrum is consistent with starformation triggered by the interaction.SdI–2’s two most striking features are its high M HI / M ∗ ratioof 28.9 (see Figure 5), which is comparable to the well known ex-tremely gas–rich dwarf DDO 154 ( M HI / M ∗ =
31 Watts et al. 2018),and it’s detached H i extended NE region (Figure 3). Given SdI–2’sisolation, ram pressure stripping can almost certainly be ruled outas an explanation for the SdI–2 detached H i extension. So, the mostlikely explanation for the detached H i extension is that it is debrisfrom the accretion of a smaller gas–rich satellite. There are exam-ples in the literature of isolated galaxies with one-sided H i tails andwarped discs attributed to the accretion of or interactions with satel-lite galaxies (Martínez-Delgado et al. 2009; Sengupta et al. 2012;Scott et al. 2014). However, this proposition is only marginally sup-ported by analysis of the integrated GMRT H i spectrum profile, A flux = ± i profile analy-sis is inconclusive, the resolved integrated H i maps and velocity field maps (from the medium resolution cube) provide signaturescharacteristic of a recent interaction. This demonstrates the highersensitivity to recent H i interactions of resolved mapping comparedto H i profile analysis.Figure 3, lower left panel, shows a smoothed NUV (GALEX)image of SdI–2. SDSS catalogues the clump marked ’D’ in the fig-ure as a star but it could alternatively be a background galaxy. Op-tically, SdI–2 consists of an elongated region of bright starform-ing clumps oriented approxmately N-S. Three of these clumps,SdI–2–a, SdI–2–b and SdI–2–c are identified in the Figure us-ing the nomenclature from Bellazzini et al. (2017). These brightclumps are surrounded by a larger region of low surface bright-ness NUV emission (Figure 3). This low surface brightness emis-sion is more extened in the NE, extending ∼
20 arcsec (3.6 kpc) NEof the optical centre. Bellazzini et al. (2017) reported 12 + log(O / H)of 8.1 ± ff erencebetween them being indistinguishable within the errors. Those au-thors also reported on a spectrum from the optical centre, SdI–2-c, with model fits indicating a stellar metallicity of Z = ff ered from SdI–2–a and SdI–2–b in not having any significant emission lines, presumably be-cause of a lack of ionising photons. The SdI–2–c log(N / O) fromBellazzini et al. (2017) is -1.7 implying a primary production ofnitrogen in a low metallicity environment. However, the observeduncertainties of N / O at 12 + log(O / H) . ffi ciency. So, a possible explanationof the lack of emission line in the SdI–2–c spectrum and the NUVhalo is a recent strong bust of SF, but the di ffi cultly with this sce-nario is the absence of the massive star clusters (M* ∼ M ⊙ )this burst should have produced. We, therefore, conclude that thebrightest observed starforming clumps are unlikely to be the resultof strong recent star burst. SdI–2’s H i morphology, together withits regular rotating, but wrapped, main body H i disc as well as theperturbed H i kinematics in the NE of the main H i body at similarvelocities to those in the detached H i extension (Figure 3 Lowerright panel) are suggestive of a recent tidal interaction. While, SdI–2–a and SdI–2–b could be remnants of an absorbed satellite or starforming regions triggered by such an interaction we do not havedefinitive proof of this. We report on H i mapping using the GMRT for two ultra–di ff usegalaxies (UDGs) from contrasting environments. UDG–B1 is pro-jected ∼
225 kpc SW of the compact group HCG 25 while Secco–dI–2 (SdI–2) is a relatively isolated UDG. These two UDGs alsohave contrasting e ff ective radii with R e of 3.7 kpc (similar to theMilky Way) and 1.3 kpc respectively. The H i morphology andkinematics of UDG–B1 suggest a recent interaction has perturbedboth the morphology and kinematics of its H i . UDG–B1’s subso-lar metallicity suggests it did not acquire a significant gas mass inthe interaction. SdI–2 has two striking features, first its M HI M ∗ ratiois 28.9 and despite its isolation, it displays a one–sided detachedextended H i region to the NW. It’s M HI M ∗ ratio implies a very lowhistoric star formation e ffi ciency. Based on our estimate of M dyn thebaryon fraction within H i radius for SdI–2 is 0.14, indicating thatgalaxy is dark matter dominated, at least within the radius in whichH i is detected. However, the baryon fraction for UDG-B1 is quite MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments uncertain because of large uncertainties about the H i disc’s inclina-tion and the question as to whether its H i is in dynamical equilib-rium. As a result we were unable to reliably estimate the UDG–B1DM fraction. The higher M HI / M ∗ ratio in SdI–2 compared to UDG–B1 is consistent with the result from Papastergis et al. (2017) whichindicated UDGs further from group centres have higher M HI / M ∗ ra-tios. In the case of UDG–B1 morphological evidence of tidal distur-bance and possible stripping (and possibly ram pressure stripping)could explain the smaller stellar mass and redder colour of UDGsat distances closer to group centres as reported in Papastergis et al.(2017). Our study highlights the importance of high spatial andspectral resolution H i observations for the study of the dark mat-ter properties of UDGs. While the narrow range of H i velocities inUDGs argues for a velocity resolution below ∼
10 km s − , goingbeyond this level also requires good signal to noise, which adds afurther burden to already demanding observations. ACKNOWLEDGEMENTS
We thank the sta ff of the GMRT who have made these obser-vations possible. The
GMRT is operated by the National Centrefor Radio Astrophysics of the Tata Institute of Fundamental Re-search. TS acknowledge support by Fundação para a Ciência e aTecnologia (FCT) through national funds (UID / FIS / / / MCTES through national funds (PIDDAC) by this grantUID / FIS / / / BPD / / / FSE (EC). TS additionally acknowl-edges support from DL 57 / / CP1364 / CT0009 from The Cen-tro de Astrofísica da Universidade do Porto. PL is supported bya work contract DL 57 / / CP1364 / CT0010 funded by the FCT.This work was supported by FCT / MCTES through national funds(PIDDAC) by this grant PTDC / FIS-AST / / / IPAC Extragalactic Database(NED) which is operated by the Jet Propulsion Laboratory, Cal-ifornia Institute of Technology, under contract with the NationalAeronautics and Space Administration. This research has made useof the Sloan Digital Sky Survey (SDSS). Funding for the SDSSand SDSS-II has been provided by the Alfred P. Sloan Foundation,the Participating Institutions, the National Science Foundation, theU.S. Department of Energy, the National Aeronautics and SpaceAdministration, the Japanese Monbukagakusho, the Max PlanckSociety, and the Higher Education Funding Council for England.The SDSS Web Site is http: // / . This research madeuse of APLpy, an open-source Python plotting package hosted athttp: // aplpy.github.com, (Robitaille & Bressert 2012). DATA AVAILABILITY
The data underlying this article will be shared on reasonable re-quest to the corresponding author.
REFERENCES
Amorisco N. C., Loeb A., 2016, MNRAS, 459, L51 Baars J. W. M., Genzel R., Pauliny-Toth I. I. K., Witzel A., 1977, A&A,61, 99Baushev A. N., 2016, J. Cosmology Astropart. Phys., 2016, 018Beasley M. A., Trujillo I., 2016, ApJ, 830, 23Beasley M. A., Romanowsky A. J., Pota V., Navarro I. M., Martinez Del-gado D., Neyer F., Deich A. L., 2016, ApJ, 819, L20Bellazzini M., Belokurov V., Magrini L., Fraternali F., Testa V., Beccari G.,Marchetti A., Carini R., 2017, MNRAS, 467, 3751Borthakur S., Yun M. S., Verdes-Montenegro L., 2010, ApJ, 710, 385Broeils A. H., Rhee M.-H., 1997, A&A, 324, 877Burkert A., Forbes D. A., 2020, AJ, 159, 56Carleton T., Errani R., Cooper M., Kaplinghat M., Peñarrubia J., 2018,2018arXiv180506896C,Conselice C. J., 2018, Research Notes of the American Astronomical Society,2, 43Di Cintio A., Brook C. B., Dutton A. A., Macciò A. V., Obreja A., DekelA., 2017, MNRAS, 466, L1Di Teodoro E. M., Fraternali F., 2015, MNRAS, 451, 3021Espada D., Verdes-Montenegro L., Huchtmeier W. K., Sulentic J., VerleyS., Leon S., Sabater J., 2011, A&A, 532, A117Flewelling H. A., et al., 2016, arXiv e-prints, p. arXiv:1612.05243Forbes D. A., Alabi A., Romanowsky A. J., Brodie J. P., Arimoto N., 2020a,MNRAS, 492, 4874Forbes D. A., Dullo B. T., Gannon J., Couch W. J., Iodice E., Spavone M.,Cantiello M., Schipani P., 2020b, MNRAS, 494, 5293Guo Q., et al., 2020, Nature Astronomy, 4, 246Holwerda B. W., Pirzkal N., Cox T. J., de Blok W. J. G., Weniger J.,Bouchard A., Blyth S.-L., van der Heyden K. J., 2011, MNRAS,416, 2426Janssens S. R., Abraham R., Brodie J., Forbes D. A., Romanowsky A. J.,2019, ApJ, 887, 92Kennicutt Robert C. J., 1998, ApJ, 498, 541Kobulnicky H. A., Kennicutt Robert C. J., Pizagno J. L., 1999, ApJ,514, 544Koda J., Yagi M., Yamanoi H., Komiyama Y., 2015, ApJL, 807, L2Lagos P., Telles E., Nigoche-Netro A., Carrasco E. R., 2011, AJ, 142, 162Leisman L., et al., 2017, ApJ, 842, 133Liao S., et al., 2019, MNRAS, 490, 5182Mancera Piña P. E., et al., 2019, arXiv e-prints, p. arXiv:1909.01363Mancera Piña P. E., et al., 2020, ArXiv 180404139L, p. arXiv:2004.14392Marino R. A., et al., 2013, A&A, 559, A114Martínez-Delgado D., Pohlen M., Gabany R. J., Majewski S. R., PeñarrubiaJ., Palma C., 2009, ApJ, 692, 955McGaugh S. S., Schombert J. M., Bothun G. D., de Blok W. J. G., 2000,ApJL, 533, L99Oh S.-H., et al., 2015, AJ, 149, 180Papastergis E., Adams E. A. K., Romanowsky A. J., 2017, A&A, 601, L10Pettini M., Pagel B. E. J., 2004, MNRAS, 348, L59Reynolds T. N., Westmeier T., Staveley-Smith L., Chauhan G., LagosC. D. P., 2020, MNRAS, 493, 5089Robitaille T., Bressert E., 2012, APLpy: Astronomical Plotting Library inPython, Astrophysics Source Code Library (ascl:1208.017)Román J., Trujillo I., 2017, MNRAS, 468, 4039Sánchez-Janssen R., Méndez-Abreu J., Aguerri J. A. L., 2010, MNRAS,406, L65Scott T. C., et al., 2014, A&A, 567, A56Scott T. C., Brinks E., Cortese L., Boselli A., Bravo-Alfaro H., 2018,MNRAS, 475, 4648Sengupta C., et al., 2012, A&A, 546, A95Sengupta C., Scott T. C., Chung A., Wong O. I., 2019, MNRAS, 488, 3222Spekkens K., Karunakaran A., 2018, ApJ, 855, 28Springob C. M., Haynes M. P., Giovanelli R., Kent B. R., 2005, ApJS,160, 149Tanoglidis D., et al., 2020, ArXiv 180404139L, p. arXiv:2006.04294Tovmassian H. M., Martinez O., Tiersch H., 1999, A&A, 348, 693Trujillo I., Roman J., Filho M., Sánchez Almeida J., 2017, ApJ, 836, 191Wang J., Yang K., Zhang Z.-Y., Fang M., Shi Y., Liu S., Li J., Li F., 2020,MNRAS, 499, L26MNRAS , 1– ????
Amorisco N. C., Loeb A., 2016, MNRAS, 459, L51 Baars J. W. M., Genzel R., Pauliny-Toth I. I. K., Witzel A., 1977, A&A,61, 99Baushev A. N., 2016, J. Cosmology Astropart. Phys., 2016, 018Beasley M. A., Trujillo I., 2016, ApJ, 830, 23Beasley M. A., Romanowsky A. J., Pota V., Navarro I. M., Martinez Del-gado D., Neyer F., Deich A. L., 2016, ApJ, 819, L20Bellazzini M., Belokurov V., Magrini L., Fraternali F., Testa V., Beccari G.,Marchetti A., Carini R., 2017, MNRAS, 467, 3751Borthakur S., Yun M. S., Verdes-Montenegro L., 2010, ApJ, 710, 385Broeils A. H., Rhee M.-H., 1997, A&A, 324, 877Burkert A., Forbes D. A., 2020, AJ, 159, 56Carleton T., Errani R., Cooper M., Kaplinghat M., Peñarrubia J., 2018,2018arXiv180506896C,Conselice C. J., 2018, Research Notes of the American Astronomical Society,2, 43Di Cintio A., Brook C. B., Dutton A. A., Macciò A. V., Obreja A., DekelA., 2017, MNRAS, 466, L1Di Teodoro E. M., Fraternali F., 2015, MNRAS, 451, 3021Espada D., Verdes-Montenegro L., Huchtmeier W. K., Sulentic J., VerleyS., Leon S., Sabater J., 2011, A&A, 532, A117Flewelling H. A., et al., 2016, arXiv e-prints, p. arXiv:1612.05243Forbes D. A., Alabi A., Romanowsky A. J., Brodie J. P., Arimoto N., 2020a,MNRAS, 492, 4874Forbes D. A., Dullo B. T., Gannon J., Couch W. J., Iodice E., Spavone M.,Cantiello M., Schipani P., 2020b, MNRAS, 494, 5293Guo Q., et al., 2020, Nature Astronomy, 4, 246Holwerda B. W., Pirzkal N., Cox T. J., de Blok W. J. G., Weniger J.,Bouchard A., Blyth S.-L., van der Heyden K. J., 2011, MNRAS,416, 2426Janssens S. R., Abraham R., Brodie J., Forbes D. A., Romanowsky A. J.,2019, ApJ, 887, 92Kennicutt Robert C. J., 1998, ApJ, 498, 541Kobulnicky H. A., Kennicutt Robert C. J., Pizagno J. L., 1999, ApJ,514, 544Koda J., Yagi M., Yamanoi H., Komiyama Y., 2015, ApJL, 807, L2Lagos P., Telles E., Nigoche-Netro A., Carrasco E. R., 2011, AJ, 142, 162Leisman L., et al., 2017, ApJ, 842, 133Liao S., et al., 2019, MNRAS, 490, 5182Mancera Piña P. E., et al., 2019, arXiv e-prints, p. arXiv:1909.01363Mancera Piña P. E., et al., 2020, ArXiv 180404139L, p. arXiv:2004.14392Marino R. A., et al., 2013, A&A, 559, A114Martínez-Delgado D., Pohlen M., Gabany R. J., Majewski S. R., PeñarrubiaJ., Palma C., 2009, ApJ, 692, 955McGaugh S. S., Schombert J. M., Bothun G. D., de Blok W. J. G., 2000,ApJL, 533, L99Oh S.-H., et al., 2015, AJ, 149, 180Papastergis E., Adams E. A. K., Romanowsky A. J., 2017, A&A, 601, L10Pettini M., Pagel B. E. J., 2004, MNRAS, 348, L59Reynolds T. N., Westmeier T., Staveley-Smith L., Chauhan G., LagosC. D. P., 2020, MNRAS, 493, 5089Robitaille T., Bressert E., 2012, APLpy: Astronomical Plotting Library inPython, Astrophysics Source Code Library (ascl:1208.017)Román J., Trujillo I., 2017, MNRAS, 468, 4039Sánchez-Janssen R., Méndez-Abreu J., Aguerri J. A. L., 2010, MNRAS,406, L65Scott T. C., et al., 2014, A&A, 567, A56Scott T. C., Brinks E., Cortese L., Boselli A., Bravo-Alfaro H., 2018,MNRAS, 475, 4648Sengupta C., et al., 2012, A&A, 546, A95Sengupta C., Scott T. C., Chung A., Wong O. I., 2019, MNRAS, 488, 3222Spekkens K., Karunakaran A., 2018, ApJ, 855, 28Springob C. M., Haynes M. P., Giovanelli R., Kent B. R., 2005, ApJS,160, 149Tanoglidis D., et al., 2020, ArXiv 180404139L, p. arXiv:2006.04294Tovmassian H. M., Martinez O., Tiersch H., 1999, A&A, 348, 693Trujillo I., Roman J., Filho M., Sánchez Almeida J., 2017, ApJ, 836, 191Wang J., Yang K., Zhang Z.-Y., Fang M., Shi Y., Liu S., Li J., Li F., 2020,MNRAS, 499, L26MNRAS , 1– ???? () Scott et al.
Watts A. B., Meurer G. R., Lagos C. D. P., Bruzzese S. M., Kroupa P.,Jerabkova T., 2018, MNRAS, 477, 5554Watts A. B., Catinella B., Cortese L., Power C., 2020, MNRAS, 492, 3672Yagi M., Koda J., Komiyama Y., Yamanoi H., 2016, ApJS, 225, 11Yozin C., Bekki K., 2015, MNRAS, 452, 937Zaritsky D., 2017, MNRAS, 464, L110van Dokkum P. G., Abraham R., Merritt A., Zhang J., Geha M., Conroy C.,2015, ApJL, 798, L45van Dokkum P., et al., 2018, Nature, 555, 629van Dokkum P., Danieli S., Romanowsky A., Abraham R., Conroy C.,2019a, Research Notes of the American Astronomical Society, 3, 29van Dokkum P., Danieli S., Abraham R., Conroy C., Romanowsky A. J.,2019b, ApJL, 874, L5van Dokkum P., et al., 2019c, ApJ, 880, 91van der Burg R. F. J., et al., 2017, A&A, 607, A79
Figure 7. UGC 2690:
GMRT medium resolution (26.46 ′′ × ′′ ) H i velocity integrated map contours on Stripe 82 SDSS g–band image. Thecontours are at 2.3, 4.6, 9.2 and 13.8 × atoms cm − . The GMRT beamsize and orientation is shown with white ellipse at the bottom left of thefigure. APPENDIX AUGC 2690 H i was also detected in the SAc galaxy UGC 2690, within GMRTUDG B1 field of view, at V HI = ± − with W ± − . The H i morphology from the medium resolution cubeis shown in Figure 7 and displays an asymmetric increase in minoraxis diameter NE of the optical centre. A similar asymmetry is seenin optical image. It seems likely both the H i and optical features aredue to a recent interaction with another group member. The V HI = ± − and W = ± − from the GBT spectrum(Springob et al. 2005) agree within the uncertainties to the GMRTvalues. Measured from the GMRT spectrum the A flux = ± ∆ sys , as defined by Reynolds et al. (2020), = ±
6. As was thecase for SdI–2 the resolved mapping is more sensitive to indicationsof perturbation than the H i profile parameters. Figure 8 shows the bbarolo velocity field, data, model and residuals from the mediumresolution cube and the PV diagram (PA = ◦ ) for UGC 2690. The bbarolo model fit indicates an H i V rot =
150 km s − ∼ inclinationof ∼ ◦ . MNRAS , 1– ?? () esolved HI in two ultra–di ff use galaxies from contrasting non-cluster environments Figure 8. UGC 2690: top: bbarolo data, model and residual from the medium resolution (26.46 ′′ × ′′ ) GMRT H i cube. Bottom: bbarolo
PV diagramfor PA = ◦ from the same cube. The blue contours are from the data and the red contours are from the model, with the yellow dots showing the fitted rings.MNRAS , 1– ????