Hubble Space Telescope Imaging of Isolated Local Volume Dwarfs GALFA-Dw3 and Dw4
P. Bennet, D. J. Sand, D. Crnojevi?, D. R. Weisz, N. Caldwell, P. Guhathakurta, J. R. Hargis, A. Karunakaran, B. Mutlu-Pakdil, E. Olszewski, J. J. Salzer, A. C. Seth, J. D. Simon, K. Spekkens, D. P. Stark, J. Strader, E. J. Tollerud, E. Toloba, B. Willman
DDraft version January 22, 2021
Typeset using L A TEX twocolumn style in AASTeX62
Hubble Space Telescope
Imaging of Isolated Local Volume Dwarfs GALFA-Dw3 and Dw4
P. Bennet,
1, 2
D. J. Sand, D. Crnojevi´c, D. R. Weisz, N. Caldwell, P. Guhathakurta, J. R. Hargis, A. Karunakaran, B. Mutlu-Pakdil,
10, 11
E. Olszewski, J. J. Salzer, A. C. Seth, J. D. Simon, K. Spekkens,
15, 9
D. P. Stark, J. Strader, E. J. Tollerud, E. Toloba, and B. Willman Physics & Astronomy Department, Texas Tech University, Box 41051, Lubbock, TX 79409-1051, USA Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA Department of Astronomy/Steward Observatory, The University of Arizona, 933 North Cherry Avenue, Rm. N204, Tucson, AZ85721-0065, USA University of Tampa, Department of Chemistry, Biochemistry, and Physics, 401 West Kennedy Boulevard, Tampa, FL 33606, USA University of California, Berkeley, Department of Astronomy, 501 Campbell Hall Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA UCO/Lick Observatory, University of California, Santa Cruz, 1156 High Street, Santa Cruz, CA 95064, USA Space Telescope Science Institute, 3800 San Martin Drive, Baltimore, MD, 21208, USA Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON K7L 3N6, Canada Kavli Institute for Cosmological Physics, University of Chicago, Chicago, IL 60637, USA Department of Astronomy and Astrophysics, University of Chicago, Chicago IL 60637, USA Department of Astronomy, Indiana University, 727 East Third Street, Bloomington, IN 47405, USA Department of Physics and Astronomy, University of Utah, 115 South 1400 East, Salt Lake City, Utah 84112, USA Observatories of the Carnegie Institution for Science, Pasadena, California 91101, USA Department of Physics and Space Science, Royal Military College of Canada P.O. Box 17000, Station Forces Kingston, ON K7K 7B4,Canada Center for Data Intensive and Time Domain Astronomy, Department of Physics and Astronomy, Michigan State University, EastLansing, MI 48824, USA Department of Physics, University of the Pacific, 3601 Pacific Avenue, Stockton, CA 95211, USA LSST and Steward Observatory, 933 North Cherry Avenue, Tucson, AZ 85721, USA (Received January 20, 2020)
Submitted to ApJABSTRACTWe present observations of the dwarf galaxies GALFA Dw3 and GALFA Dw4 with the AdvancedCamera for Surveys (ACS) on the Hubble Space Telescope (HST). These galaxies were initially dis-covered as optical counterparts to compact HI clouds in the GALFA survey. Both objects resolve intostellar populations which display an old red giant branch, younger helium burning, and massive mainsequence stars. We use the tip of the red giant branch method to determine the distance to each galaxy,finding distances of 7.61 +0 . − . Mpc and 3.10 +0 . − . Mpc, respectively. With these distances we show thatboth galaxies are extremely isolated, with no other confirmed objects within ∼ α observations and are therefore classifiedas dwarf irregulars (dIrrs). The star formation histories of these two dwarfs show distinct differences:Dw3 shows signs of a recently ceased episode of active star formation across the entire dwarf, whileDw4 shows some evidence for current star formation in spatially limited HII regions. Compact HIsources offer a promising method for identifying isolated field dwarfs in the Local Volume, includingGALFA Dw3 & Dw4, with the potential to shed light on the driving mechanisms of dwarf galaxyformation and evolution. Corresponding author: Paul [email protected] a r X i v : . [ a s t r o - ph . GA ] J a n Bennet et al.
Keywords:
Dwarf galaxies (416), Dwarf irregular galaxies (417), Galaxy distances (590), HST photom-etry (756), Red giant tip (1371), Star formation (1569) INTRODUCTIONThe Lambda Cold Dark Matter model for structureformation has been very successful at reproducing ob-servations of large scale structures; however challengesemerge at sub-galactic scales (for a recent review, seeBullock & Boylan-Kolchin 2017, and the referencestherein). Some of these challenges can be examined byswitching focus from dwarf galaxies in nearby groups(McConnachie et al. 2018; Crnojevi´c et al. 2019; Bennetet al. 2019, 2020; Carlsten et al. 2020; Mao et al. 2020)to isolated field galaxies within the Local Volume (Sandet al. 2015; McQuinn et al. 2015b; Tollerud et al. 2016).Examining these isolated, gas rich, dwarf galaxies iscritical to our understanding of dwarf galaxy forma-tion and testing dark matter theories. They are thefaintest/least massive galaxies we know of that havenever interacted with a massive galaxy halo, and thushave never felt the effects of tidal/ram pressure strip-ping (Spekkens et al. 2014; Wetzel et al. 2015). Theyare a more controlled experiment for understandingother mechanisms which drive the star formation his-tory (SFH) and metallicity of a dwarf galaxy, for in-stance supernova-driven winds, or infall of pristine gasfrom the local environment (McQuinn et al. 2013). Bycharacterizing their resolved stellar populations, it be-comes possible both to obtain the present-day structuralparameters for these galaxies and to characterize theirSFHs, providing constraints on their pasts (McQuinnet al. 2015a; Tollerud et al. 2016; McQuinn et al. 2020).Additionally, these gas rich galaxies potentially trace thefull dwarf galaxy population at the outskirts of the Lo-cal Group and other similar low-density environments, aregime where the numbers and properties of these dwarfsare just starting to be compared directly with numericalsimulations (Tikhonov & Klypin 2009; Garrison-Kimmelet al. 2014, 2019; Tollerud & Peek 2018).In this work, we will examine the isolated Local Vol-ume dwarf galaxies GALFA Dw3 and Dw4. These ob-jects were discovered as part of an archival search for op-tical counterparts to HI clouds (Giovanelli et al. 2010)discovered in the ALFALFA (Adams et al. 2013) andGALFA (Saul et al. 2012) surveys by Sand et al. (2015),and were both confirmed to have H α emission at a veloc-ity consistent with the HI detection. The key propertiesof GALFA Dw3 and Dw4 are listed in Table 1.An outline of the paper follows. In Section 2, wedescribe the HST photometry and artificial star tests(ASTs), as well as supplemental observations of the dwarfs. In Section 3, we derive distances to GALFADw3 and Dw4 via the Tip of the Red Giant Branch(TRGB) method. In Section 4, we examine the obser-vational properties of the dwarfs in the
HST imagingand derive their physical properties. In Section 5, wediscuss the star formation histories based on their
HST color-magnitude diagrams (CMDs), as well as supple-mental H α and ultraviolet (UV) images. In Section 6,we discuss the environment of the dwarfs and potentialanalogs within the Local Volume. Finally we summarizeand conclude in Section 7. DATA OVERVIEW2.1.
Hubble Space Telescope Observations
The
HST observations of GALFA Dw3 & Dw4 weretaken as part of program GO-14676 (Cycle 24, PI Sand).Both Dw3 & Dw4 were observed for a single orbit withthe Advanced Camera for Surveys (ACS)/Wide FieldCamera (WFC), using the F606W and F814W filters.We did not dither to fill in the WFC CCD chip gap, aseach dwarf easily fit into one chip. The total exposuretime was 1062 s for each filter on both Dw3 & Dw4.Color composites of these images are shown in Figure 1.We perform PSF-fitting photometry on the provided .flt images using the DOLPHOT v2.0 photometric pack-age (with the ACS module), a modified version ofHSTphot (Dolphin 2000). For this work we use thesuggested input parameters from the DOLPHOT/ACSUser’s Guide , including corrections for charge transferefficiency losses and default aperture corrections basedaround a 0.5” aperture. Quality cuts are then appliedusing the following criteria: the derived photometric er-rors must be ≤ ≤ ≤ http://americano.dolphinsim.com/dolphot/dolphotACS.pdf from blue Main Sequence (MS) features to regions red-ward of the red giant branch (RGB). The fake stars havea similar color-magnitude distribution to that of the ob-served sources, except for a deeper extension at faintmagnitudes (down to ∼ § Other Observations
Data from the Galaxy Evolution Explorer (GALEX;Martin & GALEX Team 2005) were also used to checkfor UV emission from GALFA Dw3, as this can bea strong indicator of recent star formation. Indeed,GALFA Dw3 shows substantial FUV and NUV emis-sion, which we report alongside the
HST data in Fig-ure 3. These data were part of the All-Sky Imaging Sur-vey; see Morrissey et al. (2007) for details. GALFA Dw4is outside the GALEX footprint and therefore no conclu-sions can be drawn about its recent star formation withthis dataset. We thus used UV images from the NeilGehrels
Swif t
Observatory (Gehrels et al. 2004) and the Ultraviolet/Optical Telescope (UVOT; Roming et al.2005), which were taken as part of proposal 1417202(P.I. L. Hagen) in all 3 available UV filters (UVW1,UVM2, UVW2). There is no UV emission detected inthese data, likely due to the high levels of extinctionalong the line of sight to Dw4.Supplemental H α narrow band imaging of GALFADw3 & Dw4 were obtained by our group with the WIYN0.9-m telescope and the Half Degree Imager on 21 July2017 (UT). These images are used to trace HII regionswith active star formation within the last ∼
10 Myrs(Calzetti 2013) and can be seen in Figure 4. TIP OF THE RED GIANT BRANCH DISTANCESTo determine distances to these resolved dwarf galax-ies, we make use of the TRGB technique (e.g., Da Costa& Armandroff 1990; Lee et al. 1993; Makarov et al.2006; Rizzi et al. 2007; Freedman et al. 2020). Thepeak luminosity of the RGB is a standard candle inthe red bands, because it is driven by core helium ig-nition and so it provides a useful distance estimate forgalaxies with an old stellar component which are closeenough that the RGB stars can be resolved. To de-termine TRGB magnitudes, we adopt the methodologydescribed in Crnojevi´c et al. (2019). Briefly, the pho-tometry is first corrected to account for the color de-pendence of the TRGB (Jang & Lee 2017); we also con-sider only RGB stars with colors in the range 0 . < ( F W − F W ) < .
35, so as to exclude possiblecontamination from young red supergiant stars. The lu-minosity function for RGB stars is then computed (notethat the field, background+foreground, contaminationas derived from a dwarf-free region of the ACS field-of-view is not significant for the range of colors andmagnitudes considered here), and a model luminosityfunction (convolved with the appropriate photometricuncertainty, bias and incompleteness function as derivedfrom our ASTs) is fit to it with a non-linear least squaresmethod.Using the
HST data, we find a TRGB magnitudesof 25.37 ± ± ± ± +0 . − . Mpc and 3.10 +0 . − . Mpc, respectively.We mark the position of the TRGB and its uncertaintyin Figure 2, and tabulate our results in Table 1.Anand et al. (2019) used the same dataset presentedhere for GALFA Dw4 to study the peculiar velocities ofgalaxies at the edge of the Local Group, and reported aTRGB distance of 2.97 ± STRUCTURAL PARAMETERS
Bennet et al.
Figure 1.
Color composite of F606W/F814W
HST
ACS imaging of the dwarf galaxies GALFA Dw3 (upper panel) and GALFADw4 (lower panel). The bright objects in the SW of Dw3 are background galaxies. Images are 1.2’x1.2’. North is up, east isleft.
Figure 2.
F606W/F814W CMD for the dwarf galaxies GALFA Dw3 (left panel) and GALFA Dw4 (right panel). Magnitudesare corrected for foreground extinction (see § § § σ uncertainty. We display several Padova isochrones (Bressan et al. 2012), shown as solidlines of varying color, each line representing a stellar population of fixed age, shown in the legend of each panel. The redisochrone (RGB stars) is plotted at [Fe/H]= − − Bennet et al.
Figure 3.
The UV images of GALFA Dw3 from the GALEX All Sky Imaging Survey (AIS) alongside optical images from
HST for illustrative purposes, see Figure 1. This clearly shows the elevated UV emission from Dw3. North is up, east is left. Eachimage is 1.1’x1.1’. The ellipses in this plot are illustrative. Left: HST Optical, Center: GALEX NUV, Right: GALEX FUV.
Utilizing the
HST imaging, we revisit the structuralproperties of these dwarf galaxies, previously reportedin Sand et al. (2015). To constrain the structural pa-rameters, we use the maximum-likelihood technique ofMartin et al. (2008) using the implementation of Sandet al. (2009, 2012). First, we select the stars consistentwith the RGB as seen in Figure 2. We fit a standard ex-ponential profile plus constant background to the data,with the following free parameters: the central position(RA , DEC ), position angle, ellipticity, half-light ra-dius ( r h ) and background surface density. Uncertaintieson structural parameters are determined by bootstrapresampling the data 1000 times, from which 68% con-fidence limits are calculated. The resulting structuralparameters are summarized in Table 1.Note that while the derived parameters describe theolder stellar populations in our targets, both Dw3 andDw4 host young populations that are highly irregular inappearance and are concentrated in the HII regions inthe case of Dw4 (see Figure 5).We derive the absolute magnitude of the dwarfs viadirect aperture photometry using an elliptical aperturewith semi-major axis equal to the half-light radius. Weestimate the flux within this aperture (after backgroundcorrection), and multiply by a factor of two to accountfor the total flux of the dwarf, and then convert toa magnitude. After applying our measured distancemodulus and correction for galactic extinction, we findM V = − . ± . − . ± . M/L ) V = a V + b V · ( V − I ) (1)where a V = − V =1.747 with an assumed so-lar luminosity of M V =4.77. This produces masses of2.1 × M (cid:12) and 2.6 × M (cid:12) for Dw3 and Dw4 respec-tively.Our two targets broadly fit on the Local Group size-luminosity relations with slightly higher than typicalsurface brightness (see Figure 6). These properties arevery similar to those found for Pisces A & B, two othergas-rich dwarf galaxies initially found in the GALFA sur-vey of HI compact objects (Tollerud et al. 2015; Sandet al. 2015). Dw3 fits closer with the Local Group size-luminosity relation and has similar properties to manyobjects within the Local Group that are not satellitesof the MW or M31. Dw4 appears to be higher surfacebrightness than many of these objects and is the mostcompact object at its magnitude (McConnachie et al.2018), but has possible analogues at the edge of the Lo-cal Group such as GR8 (Dohm-Palmer et al. 1998; Tol-stoy 1999). This higher surface brightness when com-pared to Local Group satellites is likely explained bythe recent star formation in both objects. These com-parisons are discussed further in § HI mass
The HI mass for GALFA Dw3 and Dw4 can be cal-culated using the HI flux and the distances derived inSection 3. This is done via the standard equation for anoptically thin gas (Haynes & Giovanelli 1984): M HI = 2 . × ( D HI ) S HI M (cid:12) (2)where D HI is the distance in Mpc and S HI is the fluxin Jy km s − . These values are reported in Table 1. Figure 4.
The H α narrow band images (see §
2) of GALFA Dw3 and Dw4 minus the continuum emission (right column),alongside optical images from
HST for illustrative purposes (left column). We point out the elevated H α emission from thenortheast corner of Dw3. GALFA Dw4 shows more H α emission within two clear regions, one at the southeast end of the dwarfand the other at the northwest end. These regions match with the blue regions seen in the HST imaging. North is up, east isleft. Each image is 1.1’x1.1’. The ellipses in this plot are illustrative.
We use the HI fluxes from (Saul et al. 2012) and thedistances we derive here, along with the standard equa-tion to derive HI masses for GALFA Dw3 and Dw4.We note that these fluxes are likely underestimated dueto spatial and spectral smoothing procedures employedby Saul et al. (2012). An example of this underesti-mation is present in the discrepant fluxes for Pisces Aand B, ∼ ∼ − , respectively, found inTollerud et al. (2015) compared to 0.445 and 0.957 Jykm s − from Saul et al. (2012). Nevertheless, for thepurpose of this work, we carry on using the values fromSaul et al. (2012) for Dw3 and Dw4. We use the revised flux values from the erratum.
Given their optical luminosities, both GALFA dwarfsare relatively gas rich, with gas mass to light ratios of ∼ (cid:12) /L (cid:12) for GALFA Dw3 and ∼ (cid:12) /L (cid:12) forGALFA Dw4. These values are comparable to that ofstar forming objects within the Local Group with simi-lar absolute magnitudes to those of the GALFA dwarfs(McConnachie 2012). When we compare GALFA Dw3and Dw4 to Pisces A and B, we find that the former havesmaller gas mass to light ratios (Pisces A: ∼ (cid:12) /L (cid:12) ,Pisces B: ∼ (cid:12) /L (cid:12) , Tollerud et al. 2016; Tollerud &Peek 2018; Beale et al. 2020), though this may be dueto the underestimation of the HI fluxes discussed above.These gas masses are similar to other isolated field ob- Bennet et al.
Table 1.
Properties of GALFA Dw3 & Dw4GALFA Dw3 GALFA Dw4R.A. (J2000) 02 h :58 m :56 s .5 ± h :45 m :44 s .7 ± ◦ :37 (cid:48) :45 (cid:48)(cid:48) .4 ± ◦ :46 (cid:48) :15 (cid:48)(cid:48) .7 ± − − ± ± +0 . − . +0 . − . m V (mag) a ± ± V (mag) a − ± − ± − I (mag) 0.44 0.72E(B − V) b F W b F W b h ( (cid:48)(cid:48) ) 12.62 ± ± h (pc) 466 ±
46 102 ± ± ± ± ± Hα (erg s − cm − ) 0.514 ± × − ± × − HI v LSR (km s − ) c ± ± α v LSR (km s − ) d ±
35 607 ± tot (Jy km s − ) c (cid:63) (M (cid:12) ) 2.1 × × M HI (M (cid:12) ) 6.9 × × SFR
NUV (M (cid:12) ) 8.7 ± × − –SFR FUV (M (cid:12) ) 8.7 ± × − –SFR Hα (M (cid:12) ) 3.77 ± × − ± × − a VEGA Magnitude, derived from m F W using the conversion from(Sahu et al. 2014) b Based on the Schlafly & Finkbeiner (2011) dust maps c From the GALFA survey, see Saul et al. (2012), using the erratum values d From Sand et al. (2015) jects which are gas rich and star forming (Geha et al.2012; Bradford et al. 2015; McQuinn et al. 2020). STAR FORMATION HISTORIESIt is immediately apparent from the
HST images andthe derived CMDs that GALFA Dw3 & Dw4 are nearbystar-forming dwarf galaxies. They have well-resolvedstellar populations, both show RGBs, asymptotic giantbranch (AGB) stars, red helium burning stars, blue he-lium burning stars, MSs, an overall irregular morphol-ogy, and HII regions with young star clusters.We attempted to use the CMD-fitting code MATCH(Dolphin 2002) to determine the SFHs of GALFA Dw3 and Dw4 similar to other works in the Local Volume (e.g.McQuinn et al. 2010; Weisz et al. 2011, 2014). However,the distance to these dwarfs and the shallow nature ofthe CMDs meant that the results did not provide mean-ingful constraints on the SFH of either dwarf, other thanan indication of active star formation within the past100 Myrs. Therefore we have qualitatively analyzedeach dwarf’s possible SFH via comparison to the Padovaisochrones (Bressan et al. 2012) and multi-wavelengthobservations, similar to other works with Local Volumelow-mass dwarfs where more in depth analysis has notbeen possible (e.g. McQuinn et al. 2015a, 2020).
Figure 5.
Spatial distribution of point sources consistent with stellar populations in GALFA Dw3 and Dw4. Point sourcesconsistent with RGB stars are shown in red; these are selected via matching to the RGB isochrones seen in Figure 2. The bluepoints are those point sources consistent with a color of (F606W -F814W ) < (cid:48) squares. North is up, East is left. GALFA Dw3
Isochrone Comparisons
The CMD of GALFA Dw3 reveals a complex SFH,with both young and old stellar populations. We pointthe reader to the left panel of Figure 2 to guide thisdiscussion, where we denote stars in the main bodyof GALFA Dw3, along with those associated with itsHII region (see discussion below), and plot relevantisochrones of varying age and metallicity.There are several faint, blue stars (with 23 (cid:46)
F814W (cid:46) − F814W ) < − ∼
10 Myrs.Other young MS stars are apparent at fainter magni-tudes. A sequence of stars spanning the same F814W range at slightly redder colors ((F606W − F814W ) ≈ (cid:46) F814W (cid:46) (cid:46) (F606W − F814W ) (cid:46) =25.4 mag), likely correspondingto an ancient and metal poor stellar population ( > ≈− HST data. Star formation may have happened beforethis estimated age, however deeper data would be re-quired to determine this.The size and separation of the helium burningbranches in Dw3 indicate a population with [Fe/H] ≈− .
0, based on an approximate match to isochrones. Ametallicity of [Fe/H]= − . V = − Bennet et al.
Figure 6.
Absolute V-band magnitude as a function of half-light radius for GALFA Dw3 and Dw4 (blue stars) as comparedto satellites of the MW and M31 (Red Inverted Triangles) and other Local Group objects, i.e. those outside the virial radius ofeither the MW or M31 (Black Squares). Pisces A & B are shown for comparison (Cyan Triangles), along with Leo P (GreenCircle). The lines of constant central surface brightness assume an exponential profile and range from 16 mag/arcsec to 30mag/arcsec with a line every ∆2 mag / arcsec . for Local Volume galaxies (Berg et al. 2012). It isalso consistent within 1 σ with the possible luminosity–metallicity relation for void galaxies (Pustilnik et al.2016; Kniazev et al. 2018; McQuinn et al. 2020).Generally, dwarf irregulars form stars in bursts (Weiszet al. 2011), and this is also backed up by simulations(Wheeler et al. 2019). Deeper observations would be re-quired to distinguish between continuous star formationand more episodic, bursty star formation in Dw3. Fi-nally, isochrone fitting in the main body (excluding theHII region) shows a well populated young MS of starsbelow m F W ≈ .
5. If this is the MS turnoff for themajority of the dwarf, it would show that star formationacross most of the dwarf ceased ∼
20 Myrs ago.5.1.2. H α Imaging
The H α imaging of GALFA Dw3 (see Section 2) re-veals a single HII region located at the northeast edgeof the dwarf, this image is shown alongside the HST image in Figure 4. This H α imaging shows a flux of0.514 ± × − erg s − cm − , which combined withthe distance, foreground extinction and the conversionfactor from Kennicutt (1998) implies a star formationrate of 3.77 ± × − M (cid:12) yr − .If we limit the CMD to only those stars with a spatialposition consistent with this HII region, we can see thatthe H α emission may be caused by a single MS O-starwith a maximum age of 5 Myrs (see Figure 2 and 4). Inthis region we also see a population of lower-mass youngMS stars as well as red and blue helium burning stars athigher density than across the main body of the dwarf.The RGB is at a similar density in the HII region whencompared to the rest of the dwarf at a similar radius,indicating the overdensity of younger stars is not simplya result of higher overall stellar density in this region.We also find a point source (F814W =23.2 andF606W -F814W = − V = − α emission, but we see no H α emis-sion from its position, we can draw some conclusions.The first idea would be that this is a blended multiplestar system (see the Leo P analysis in McQuinn et al.2015a). If we assume equally massed component stars,then these components would be O8 (M V = − α emission (even an equally massed triple star systemwould have components large enough to produce H α ).This source may be an evolved helium burning star thatdue to noise has been scattered into the region of theCMD equivalent to the MS.5.1.3. GALEX
As an additional method to determine the level andspatial position of recent star formation in GALFA Dw3,we checked the GALEX archive for the dwarf’s ultravio-let emission. Dw3’s position was observed by GALEX aspart of the All-sky Imaging Survey (AIS, exposure time ∼ HST images in Figure 3.The GALEX data shows diffuse NUV and FUV emis-sion across the body of Dw3, though slightly more con-centrated toward the north. We see some concentra-tion of FUV emission in the HII region found in theH α imaging, however the majority is spread across thedwarf. This significant NUV and FUV emission confirmsthe conclusion from the isochrone fitting that significantstar formation has occurred across the dwarf within thelast 100 Myrs (Calzetti 2013).The detected level of NUV emission indicates thatGALFA Dw3 has had recent star formation at a rate of8.7 ± × − M (cid:12) yr − , whereas the FUV emission indi-cates an order of magnitude lower star formation rate of8.7 ± × − M (cid:12) yr − . Both star formation rates werecalculated using the relevant relations from Iglesias-P´aramo et al. (2006). These relations have been shownto be potentially unreliable in low metallicity galaxies,like GALFA Dw3 (McQuinn et al. 2015a); in which casethe star forming rate maybe up to ∼ ∼
100 Myr. This is reinforced by the SFRderived from the H α imaging above (3.77 ± × − M (cid:12) yr − ) which is comparable to the rate derived fromthe FUV emission but slightly lower. This difference instar formation rates between the tracers examined herecan be explained by their differing sensitivity to different ages of star formation. As NUV is equally sensitive toall star formation across the last 100 Myrs, while FUVis most sensitive to stars formed in the last 10 Myrs(though there is some FUV sensitivity to populationsup to 100 Myrs old, Calzetti 2013), and H α is sensitiveto only star formation within the last 10 Myrs.The UV emission coming from across the dwarf, alongwith the difference between the H α , NUV and FUV, sup-ports the conclusion drawn from the isochrone matching:that star formation was higher and more widespreadin Dw3 in the recent past ( (cid:46)
100 Myr), but has nowquenched across most of the dwarf, and that there isongoing star formation only in the single HII region (inthe last ∼
10 Myr).5.1.4.
Spatial structure
Another diagnostic that we can use to analyze GALFADw3 is spatial maps, see Figure 5. When the stars areplotted on spatial maps we can see that the MS stars areconcentrated in the central regions of the dwarf, have amore elliptical distribution and are preferentially foundtoward the northern end of the galaxy. This is true forall MS stars, aside from the very brightest which areonly found in the HII region. This is in contrast to theRGB stars which are more evenly distributed through-out the galaxy. The helium burning stars are also moreconcentrated towards the center of the dwarf when com-pared to the RGB stars, however the concentration isless pronounced than it is for the MS stars.When we examine the star positions and comparethem to the multi-wavelength observations, we find astrong match between the MS stars and the NUV emis-sion. 5.1.5.
Summary
GALFA Dw3 shows an underlying old ( > ≈− α emission for avery young population ( <
20 Myr) that is spatially lim-ited to a single HII region in the northeast of the dwarf(see Figure 4).The differences in the spatial position and extent ofthe tracers of different ages of star formation can be usedto reconstruct a qualitative SFH for GALFA Dw3: thestar formation was at a higher level and distributed moreevenly throughout the dwarf in the recent past, but isnow restricted to a single HII region. This could indicatethat GALFA Dw3 is concluding an episode of recent star2
Bennet et al. formation that has now been quenched outside of theHII region. This interpretation appears to support themodel that star formation in isolated dwarf galaxies isdriven by a series of ‘bursts’ of intense star formation, in-terspersed with periods of quiescence (Weisz et al. 2011).In this model, galaxies go through intense bursts of ac-tive star formation which expels the HI gas through stel-lar feedback. This expulsion of the neutral gas causesthe star formation to wane and the feedback to decrease.Without feedback, more HI gas falls onto the dwarf, pro-ducing a new episode of star formation (Wheeler et al.2019). In this case, GALFA Dw3 would be in the con-cluding part of such a star forming episode with the lastparts of star formation from an active burst. More de-tailed HI observations may be needed to determine theposition and kinematic properties of the gas, as the ex-isting HI information from the GALFA survey is lowresolution (Saul et al. 2012).5.2.
GALFA Dw4
The position of GALFA Dw4 near the galactic planecomplicates creating a comprehensive SFH due to thehigh extinction (particularly in the UV).5.2.1.
Isochrone Comparisons
The CMD of GALFA Dw4 also reveals a complex SFH,with both young and old stellar populations, howeverthere are substantial differences between Dw3 and Dw4.We point the reader to the right panel of Figure 2 toguide this discussion, where we denote stars in the mainbody of GALFA Dw4, along with those associated withboth of its HII regions (see discussion below), and plotrelevant isochrones of varying age and metallicity.Isochrone matching of the red and blue helium burn-ing branches in GALFA Dw4 indicates a metallicityof [Fe/H] ≈ − σ with the luminosity–metallicity relationship for LocalVolume dwarfs (Berg et al. 2012).Isochrone matching also shows Dw4 has an ancient( > ≈ − ≈ − H α Imaging
The H α imaging (see §
2) shows that Dw4 has twoHII regions, one at each end of the galaxy, which matchthe blue regions seen in the
HST imaging (see Figure4). We find an H α flux of 4.184 ± × − erg s − cm − for the southeast region and 1.037 ± × − erg s − cm − for the northwest region, for a total H α flux of 5.221 ± × − erg s − cm − . Combinedwith the distance, foreground extinction and the con-version factor from Kennicutt (1998) this flux implies astar formation rate of 1.37 ± × − M (cid:12) yr − .When we examine these HII regions in the CMD (seethe right panel of Figure 2), we find that there are noobvious O-stars to drive the H α emission. This could becaused by internal extinction within Dw4, which couldcause the MS O-stars to appear as stars at the upperend of the blue helium burning branch. For this to bethe case, the HII regions would have to be obscuredby enough dust to cause extinction of A F W ≈ F W ≈ α emission are visible, but are not recovered inour point source photometry because they were culledat some stage in our reductions. To test this possibility,a CMD for Dw4 was constructed using the DOLPHOTcatalog, but with the photometric quality cuts severelyrelaxed. This did not detect any sources with color andbrightness consistent with MS O-stars across Dw4. Wehave also used the ASTs to confirm that artificial starswith properties similar to MS O-stars are successfullyrecovered by DOLPHOT in the HII regions of Dw4. Wealso tried a similar reduction in photometric quality cuts3with the crowded photometry discussed in §
2, and thisyielded a few point sources consistent with MS O-stars ofthe spectral classes O7-O9. These could be the source ofthe H α emission, however these poorly recovered sourcesare generally too blue to be MS O-stars. We have con-sidered that these objects may be O-stars with line con-tamination from the HII region sufficient to move it offthe MS in the CMD, however this contamination wouldhave to be larger than expected to have the observedeffect. On the other hand, equivalent point sources arenot found in the parallel field, indicating they are uniqueto the dwarf. ‘ Therefore, it is possible these are the MSO-stars, but they are in areas of the dwarf that precludeclean photometric recovery with the present data.It is also possible that a combination of the abovescenarios are the reason we see no MS O-stars in Dw4despite the presence of H α emission. In this case, inter-nal extinction obscures and blurs the O-stars such thatthey are not recovered clearly by DOLPHOT.The two HII regions also contain most of the lower-mass MS stars seen in Dw4 (see Figure 5). This indicatesthat star formation is currently limited to these two re-gions. We also see overdensities of red and blue heliumburning stars in the HII regions compared to the dwarfas a whole. RGB stars appear to be at a similar densityin the HII regions when compared to other parts of thedwarf with similar radius, indicating the overdensitiesof young stars are genuine and not caused by generalstellar overdensities in these regions.5.2.3. SWIFT UVOT
As stated in §
2, GALFA Dw4 is outside of the GALEXfootprint due to its proximity to the galactic plane.Therefore, to get UV information on this object,
Swift
UVOT (Roming et al. 2005) observations were required.These were taken as part of proposal 1417202 (P.I. L.Hagen) to observe the UV dust extinction properties inGALFA Dw4 (along with 4 other Local Volume dwarfs).Despite these
Swift images with a reasonable total ex-posure time ( ∼ α emissionfrom Dw4, combined with the presence of bright MSstars in the HST imaging, means it is likely that thereis UV emission from Dw4, but that it is not observablewith the present data due to the previously mentionedhigh levels of extinction.5.2.4.
Spatial structure
In Dw4 the RGB stars are spread throughout thedwarf while the young MS and helium burning stars arelargely confined to regions near the HII regions. These younger stellar populations are at higher relative den-sity at either end of the dwarf near the HII regions, seeFigure 5. We find that older helium burning stars aremore evenly spread throughout the dwarf, though stillmore concentrated towards the current HII regions thanthe RGB stars. This may be the result of previous starformation being more evenly distributed, or a result ofthese older stars having had time to mix through thedwarf since they formed.5.2.5.
Summary
GALFA Dw4 has an old ( (cid:38) ≈ − α imaging which shows emission con-centrated in two regions at either end of the dwarf atthe same position as the young stellar populations inthe CMD. Therefore we conclude that star formation inDw4 is limited to the HII regions at either end of thedwarf. We also find that star formation has been ongo-ing for >
500 Myrs, and seems to be more concentratedin the HII regions. This can be seen by the concen-tration of young stars in these regions compared to theRGB stars, along with the H α emission. However, ourconclusions here are less robust than for Dw3. This isdue to the lack of UV information, and the lower totalnumber of stars in Dw4 which makes it difficult to deriveconcrete information via examining stellar populations. DISCUSSIONHaving determined the distance ( § §
4) and qualitative SFHs ( §
5) of both GALFADw3 and Dw4, we are in a position to discuss thesegalaxies in detail. 6.1.
Environment
We began exploring the environment around bothGALFA Dw3 & Dw4 using their newly derived dis-tances and a search of the NASA Extragalactic Database(NED) . We searched for any catalogued objects within ∼ −
400 to +600 km s − (thisrange was chosen to avoid contamination by Galacticobjects with velocities less than the MW escape veloc-ity). This search showed that both GALFA Dw3 & Dw4are extremely isolated, confirming the result from Sandet al. (2015). In addition, catalogs of known galax-ies were searched for objects nearby to either galaxy http://ned.ipac.caltech.edu/ Bennet et al. and we found nothing within 1.5 Mpc of either dwarf(Karachentsev et al. 2013).The closest known object to GALFA Dw3 is NGC1156:NGC1156 has a distance consistent with GALFA Dw3at 7.6 ± − (Karachentsev et al. 2013), we considerdirect association at the present time to be unlikely.GALFA Dw4 is projected near to the Orion Dwarf andA0554, however these objects are more distant at D ∼ ∼ − (Karachentsev et al. 2013). This is aHI source with no detected optical counterpart (Donleyet al. 2005). We find that A0554 is the closest objectwith an optical counterpart, though this is extremelydistant with a radial separation of 2.3 Mpc and a pro-jected separation of 220 kpc. However, as GALFA Dw4is in the ‘zone of avoidance’ around the Galactic plane,there have been relatively few deep wide field opticalsurveys done in the area, and therefore it can not beruled out that there maybe other undetected galaxiescloser than A0554. GALFA Dw4 is also unusual as ithas large peculiar velocity ∼ +350 km s − , which is un-expected for isolated systems, which tend to move withthe Hubble flow (Anand et al. 2019).The isolation of GALFA Dw3 & Dw4 can be seen inFigure 7, where the dwarfs are shown to be ‘below’ thesupergalactic plane in very low density regions of theLocal Volume. Therefore we conclude that both GALFADw3 & Dw4 are truly isolated with no other objectsclose enough to influence them at the current time orin the recent past. This isolation allows us to use themas probes into how star formation and galaxy evolutionoccur in isolated low-mass galaxies.6.2. Local Volume Analogs
We have examined other Local Volume dwarf galaxiesto compare the properties of GALFA Dw3 and Dw4 withother low mass systems.GALFA Dw3 & Dw4 have very similar physical prop-erties to Pisces A & B, which were also found in follow-up to the GALFA survey (Tollerud et al. 2015; Sandet al. 2015). All of these objects are very isolated, how-ever Pisces A and B were theorised to be falling intolocal filamentary structure after spending most of cos-mic time at the edge of the Local Void (Tollerud et al. 2016), which is speculated to have triggered recent starformation in Pisces A & B.The other object from Sand et al. (2015), ALFALFADw1 (also referred to as AGC 226067 or SECCO 1; Bel-lazzini et al. 2015) shows stellar populations that werefound to be approximately consistent with a single burstof star formation with an age range of ∼ ∼ ∼
500 Myrs (Mc-Quinn et al. 2010). GR8 (DDO155/UGC8091) is a starforming dwarf in the Local Volume with a distance of ∼ ∼
100 Myrs before star formation ceases and newregions begin to actively form stars (Dohm-Palmer et al.1998; Tolstoy 1999).The Survey of HI in Extremely Low-mass Dwarfs(SHIELD) galaxies (Cannon et al. 2011) are a selectionof 12 galaxies initially detected in ALFALFA (Giovanelliet al. 2005; Haynes et al. 2018) data. These galaxies wereselected based on low HI and stellar mass estimates. Interms of absolute magnitude and gas mass, the SHIELDgalaxies are in the same range as GALFA Dw3 andDw4. Examination of the SFHs of the SHIELD galax-ies also shows a recent star formation rate consistentwith that derived for Dw3 (see § > Figure 7.
The location of GALFA Dw3 and Dw4 in the Local Volume. GALFA Dw3 and Dw4 are shown as blue stars andlabelled, as are Pisces A and B (Cyan Triangles), while the black dots are a 10 Mpc volume-limited sample of nearby galaxies(Karachentsev et al. 2013). The coordinates are supergalactic Cartesian with Earth at the center, oriented such that the x-axispoints towards the origin and the z-axis points towards the Local Void (Lahav et al. 2000). Bennet et al. tially good analogs, while Dw4 is fainter and physicallysmaller than the typical SHIELD galaxy. As previouslymentioned Dw4 is one of the most compact objects atits luminosity detected. CONCLUSIONSWe have presented
HST imaging of GALFA Dw3 andDw4, two Local Volume dwarf galaxies which were ini-tially discovered as optical counterparts to compact HIclouds in the GALFA survey. Both dwarfs resolve intostars, displaying complex stellar populations, includingan old red giant branch, young helium burning sequencesand main sequence stars. Each system also has youngstar clusters and HII regions which are evident in ourH α imaging. In detail, the two dwarfs appear to haveslightly different star formation histories based on aqualitative assessment of their CMDs and on the avail-able UV data. GALFA Dw3 shows signs of a recentlyceased episode of active star formation; although it isnot well constrained, Dw4 seems to have a more consis-tent level of star formation within spatially limited HIIregions at either end of the dwarf.Using the resolved CMDs, we measure the dis-tance to each dwarf using the TRGB method, finding D =7.61 +0 . − . Mpc and D =3.10 +0 . − . Mpc for GALFADw3 and Dw4, respectively. With this information inhand, we found each dwarf to be extremely isolated,with no known neighbor within ∼ § Facilities:
HST (ACS), WIYN:0.9m, GALEX,SWIFT
Software:
Numpy, Astropy (The Astropy Collabora-tion et al. 2018), DOLPHOT (Dolphin 2000)REFERENCES